Water balance of short rotation coppice

4 downloads 0 Views 14MB Size Report
0.20 m of water per meter depth of the aquifer while a clay aquifer can hold 0.45 ...... were cut as a whole and subsequently divided into four separated cartloads.
Mendel University in Brno Faculty of Agronomy Institute of Agrosystems and Bioclimatology

Water balance of short rotation coppice „Vodní bilance porostu energetických dřevin“ Ph.D. thesis

Supervisor: prof. Ing. Zdeněk Ţalud, Ph.D.

Author: Ing. Milan Fischer

Brno 2012

ACKNOWLEDGEMENTS I would like to express deep gratefulness to the supervisor prof. Ing. Zdeněk Ţalud, Ph.D. and to my supervisor specialist doc. Mgr. Ing. Miroslav Trnka, Ph.D. for their inspiration, many useful suggestions and ideas, transmitted experience and overall for providing me conditions to work on very interesting research. Special thanks belong to Ing. Jiří Kučera who shared with me his valuable experience and the knowledge of the Bowen ratio and other micrometeorological and ecophysiological measurements techniques and who provided me many fruitful discussions and recommendations. Sincere thanks to Dr. Gaby Deckmyn for introducing me the world of forest ecophysiological modelling and to all members from Laboratory of Plant Ecology (Department of Biology, University of Antwerp) for scientific cooperation, support and inspiration. I would like to admit many thanks to all from the Department of Agrosystems and Bioclimatology and other colleagues from MENDELU who helped me with the planning and realization of the field experiments, regular measurements routines, data handling, work in laboratory and who gave me any other support. Finally, thanks to all of my reviewers for useful and constructive comments. The Ph.D. thesis was supported by the Research plan No. MSM6215648905 ―Biological and technological aspects of sustainability of controlled ecosystems and their adaptability to climate change―, which is financed by the Ministry of Education, Youth and Sports of the Czech Republic. This publication is further an output of the CzechGlobe Centre, that is being developed within OP RDI and co-financed from EU funds and state budget of the Czech Republic (Project: CzechGlobe – Centre for Global Climate Change Impacts Studies, Reg.No. CZ.1.05/1.1.00/02.0073). Finally, the presented study was supported also by In-house Grant Agency at MENDELU by projects no. IP 14/2009, no. IP 19/2010 and no. TP 11/2010.

2

PROHLÁŠENÍ Prohlašuji, ţe jsem disertační práci na téma „Vodní bilance porostu energetických dřevin“ vypracoval samostatně a pouţil jen pramenů, které cituji a uvádím v přiloţeném seznamu literatury. Disertační práce je školním dílem a můţe byt pouţita ke komerčním účelům jen se souhlasem vedoucího disertační práce a děkana AF MENDELU.

dne 13.2.2012 podpis autora……………………………. 3

CONTENTS SOUHRN ............................................................................................................................... 8 1

INTRODUCTION ........................................................................................................ 11

2

LITERATURE ............................................................................................................. 13 2.1

Plant water relations .............................................................................................. 13

2.2

Water balance concept .......................................................................................... 17

2.2.1

Precipitation and infiltration .......................................................................... 18

2.2.2

Surface (subsurface) run-off (lateral inflow) ................................................. 19

2.2.3

Evapotranspiration ......................................................................................... 21

2.2.4

Drainage and capillary rise ............................................................................ 23

2.2.5

Soil moisture .................................................................................................. 23

2.3

Measuring water loss............................................................................................. 24

2.3.1 2.3.1.1

Water balance ......................................................................................... 26

2.3.1.2

Lysimeters .............................................................................................. 30

2.3.2

Micrometeorological methods ....................................................................... 32

2.3.2.1

Bowen ratio and energy balance method ................................................ 40

2.3.2.2

Aerodynamic method ............................................................................. 42

2.3.2.3

Eddy covariance ..................................................................................... 44

2.3.2.4

Scintillometry ......................................................................................... 46

2.3.3

2.4

Hydrological approaches ............................................................................... 26

Plant physiology approach ............................................................................. 50

2.3.3.1

Porometry ............................................................................................... 51

2.3.3.2

Sap flow .................................................................................................. 53

2.3.3.3

Chamber method .................................................................................... 58

Estimating water loss ............................................................................................ 60

2.4.1

Empirical evapotranspiration models ............................................................ 60

2.4.1.1

Penman approach.................................................................................... 60

2.4.1.2

Priestley-Taylor approach ...................................................................... 62

2.4.2 2.4.2.1

Penman-Monteith analytical approach .......................................................... 63 FAO reference and crop evapotranspiration ........................................... 66 4

2.4.3

Soil water balance modelling ......................................................................... 69

3

AIMS ............................................................................................................................ 70

4

MATERIALS AND METHODS ................................................................................. 71 4.1

Locality ................................................................................................................. 71

4.1.1

Poplar plantation ............................................................................................ 71

4.1.2

Climate conditions ......................................................................................... 72

4.1.3

Soil ................................................................................................................. 74

4.2

Measurement scheme, sensors and data processing .............................................. 75

4.2.1 4.2.1.1

Soil moisture calibration......................................................................... 76

4.2.1.2

Soil moisture spatio-temporal variability ............................................... 77

4.2.1.3

Soil moisture temporal variability .......................................................... 78

4.2.2

5

Evapotranspiration ......................................................................................... 78

4.2.2.1

Error analysis .......................................................................................... 80

4.2.2.2

Data rejection.......................................................................................... 81

4.2.2.3

Gap filling ............................................................................................... 82

4.2.2.4

Fetch and footprint ................................................................................. 83

4.2.3

4.3

Soil Moisture.................................................................................................. 76

Other micrometeorological and eco-physiological variables ........................ 83

4.2.3.1

Bowen ratio ............................................................................................ 83

4.2.3.2

Surface resistance ................................................................................... 84

4.2.3.3

Decoupling coefficient ........................................................................... 84

4.2.3.4

Crop coefficient ...................................................................................... 84

4.2.3.5

Sap flow .................................................................................................. 85

4.2.3.6

Biomass growth and water-use efficiency .............................................. 86

Statistical methods ................................................................................................ 88

RESULTS ..................................................................................................................... 90 5.1

Soil Moisture ......................................................................................................... 90

5.1.2

Soil moisture calibration ................................................................................ 90

5.1.2

Soil moisture spatio-temporal variability ...................................................... 93

5.1.3

Soil moisture temporal variability ................................................................. 98

5

5.2

5.2.1

BREB error analysis .................................................................................... 103

5.2.2

Gap filling .................................................................................................... 110

5.2.3

Fetch and footprint ....................................................................................... 113

5.2.4

Daily course of actual evapotranspiration ................................................... 121

5.2.5

Monthly course of actual evapotranspiration............................................... 127

5.2.6

Yearly course of evapotranspiration ............................................................ 136

5.3

6

Evapotranspiration .............................................................................................. 103

Other micrometeorological and eco-physiological variables .............................. 138

5.3.1

Bowen ratio .................................................................................................. 138

5.3.2

Surface resistance ........................................................................................ 140

5.3.3

Decoupling coefficient ................................................................................. 142

5.3.4

Crop coefficient ........................................................................................... 144

5.3.5

Sap flow ....................................................................................................... 149

5.3.6

Biomass growth and water-use efficiency ................................................... 154

DISCUSSION............................................................................................................. 163 6.1

Soil Moisture ....................................................................................................... 163

6.1.1

Soil moisture calibration .............................................................................. 163

6.1.2

Soil moisture spatio-temporal variability .................................................... 164

6.1.3

Soil moisture temporal variability ............................................................... 166

6.2

Evapotranspiration .............................................................................................. 169

6.2.1

BREB error analysis .................................................................................... 169

6.2.2

Gap filling .................................................................................................... 177

6.2.3

Fetch and footprint ....................................................................................... 180

6.2.4

Daily course of actual evapotranspiration ................................................... 184

6.2.5

Monthly course of actual evapotranspiration............................................... 187

6.2.6

Yearly course of evapotranspiration ............................................................ 193

6.3

Other micrometeorological and eco-physiological variables .............................. 197

6.3.1

Bowen ratio .................................................................................................. 197

6

7

6.3.2

Surface resistance ........................................................................................ 199

6.3.3

Decoupling coefficient ................................................................................. 200

6.3.4

Crop coefficient ........................................................................................... 201

6.3.5

Sap flow ....................................................................................................... 204

6.3.6

Biomass growth and water-use efficiency ................................................... 208

CONCLUSIONS ........................................................................................................ 212

REFERENCES .................................................................................................................. 215 APPENDIXIES .................................................................................................................. 240 ANNOTATION ................................................................................................................. 261

7

SOUHRN Produkce a vyuţití obnovitelných zdrojů surovin a energie je klíčovou otázkou udrţitelného hospodářského rozvoje společnosti zejména s ohledem na sniţování emisí CO2 a diverzifikaci energetických zdrojů. Posledním z předpisů EU týkajících se přímo obnovitelných zdrojů energie (OZE) a jejich podílu v energetice a dopravě je směrnice 2009/28/ES, která definuje závazná kritéria pro jednotlivé členské státy. V zásadě se jedná o dva základní cíle pro rok 2020: i) zvýšit podíl OZE nejméně na 20 % v celkové spotřebě energie EU, ii) zvýšit podíl biopaliv v dopravním sektoru nejméně na 10 % pro kaţdý členský stát. Česká republika se v souladu se zmiňovanou směrnicí zavázala zvýšit podíl OZE alespoň na 13 %. Předchozí indikativní cíl 8% k roku 2010 byl splněn jen s mírným překročením o 0,3 %. Dle řady vědeckých studií je velmi významným zdrojem OZE v podmínkách střední Evropy právě energie z biomasy. Nejvyšší potenciál produkce biomasy je spatřován v energetických plodinách pěstovaných na orné půdě, především pak v rychle rostoucích dřevinách (RRD) a vytrvalých travinách. Doposud byl podíl energetických plodin na celkové produkci biomasy poměrně nízký a je charakteristický převahou jednoletých plodin určených především k výrobě biopaliv charakteru pohonných hmot. Přitom víceleté energetické plodiny mají řadu předností, které jsou dány významně vyššími výnosy sušiny, vyšší čistou energetickou bilancí a příznivějším environmentálním vlivem. Závěry analytických studií zabývajících se porovnáním ekonomické efektivity produkce různých druhů energetických plodin navíc potvrzují, ţe nejniţší náklady na jednotku získané energie odpovídají RRD a vytrvalým travinám, zatímco jednoleté plodiny vykazují náklady nejvyšší. Další pokles nákladů u víceletých energetických plodin lze očekávat aţ do výše 50 % v souvislosti s pokrokem ve šlechtění, vývojem technologií pěstování, velkoplošným pěstováním a vyuţitím potenciálu stanoviště. Oproti tomu moţnost sníţení nákladů na produkci energie jednoletých energetických plodin je jiţ velmi nízká. Současně s ekonomickou optimalizací produkce energie se ovšem pěstování energetických plodin potýká také s řadou otázek souvisejících s vlivem na přírodní zdroje, které mají velmi často globální dopady. K nejvýznamnějším patří otázka hospodaření s vodou (uchování zdrojů pitné vody a dostatku vody pro ostatní zemědělskou produkci), efektivita hospodaření s ţivinami (především dusíkem) a obecně pak zajištění dostatečné plochy orné půdy pro produkci potravin. Zcela zásadní je pak otázka výběru vhodného stanoviště a optimalizace technologie pro daný druh a přírodní podmínky. Klíčovým faktorem pro výběr stanoviště RRD jsou bezesporu jejich nároky na vodu ve srovnání s místními klimatickými podmínkami. Hlavním cílem předkládané disertační práce je proto kvantifikace aktuální evapotranspirace ( ) porostu RRD (Populus nigra x P. Maximowiczii) zaloţeného na orné půdě v podmínkách Česko-Moravské vysočiny a dále porovnání RRD a pravidelně sečeného (cca jednou za 10 dnů) trvalého travního porostu, jenţ je obvykle povaţován jako standardní referenční povrch. Mimo fakt, ţe RRD pro podmínky ČR dosud nijak stanovována nebyla, lze jedinečnost této práce, a to ve světovém měřítku, spatřovat i ve vyuţití jednotné metody při měření obou porostů. K tomuto účelu byla zvolena tradiční, a řadou vědeckých prací prověřená, mikrometeorologická metoda 8

Bowenova poměru a energetické bilance principiálně zaloţená na gradientovém měření teploty a vlhkosti vzduchu a současném měření radiační bilance a toku tepla do půdy. Východiskem metody je teorie turbulentní difuze popisující tok vzdušných příměsí a energie ve směru spádu koncentračního a teplotního gradientu a zároveň teorie radiační bilance popisující disipaci dostupné energie na výpar (respektive evapotranspiraci), turbulentní přenos tepla a tok tepla do půdy. Podílem energie turbulentního toku tepla a energie spojené s tokem vodních par (tok latentního tepla) je právě tzv. Bowenův poměr, jenţ lze empiricky stanovit z měření teploty a vlhkosti vzduchu alespoň ve dvou úrovních nad povrchem. Mimo samotné stanovení a její porovnání byly u obou kultur dále sledovány další proměnné jako např. dynamika a prostorová variabilita půdní vlhkosti, Bowenův poměr, povrchový odpor či koeficient spřaţenosti. Pouze pro RRD pak byla navíc sledována produkce biomasy, efektivita vyuţití vody, index listové plochy ( ), transpirační proud či plodinový koeficient. Na základě výsledků zahrnujících celkově tři celé vegetační sezóny lze konstatovat, ţe v daných pedo-klimatických podmínkách dosahuje sezónní spotřeba vody porostu RRD překvapivě velmi srovnatelné či spíše niţší úrovně neţ referenční travní porost. Vývoj odpovídá do určité úrovně míře olistění, a proto v jarních měsících (např. duben), či v období po smýcení během obnovujícího se a ještě nezapojeného porostu (rok 2010) byla RRD statisticky průkazně niţší. V případě vzrostlého porostu RRD v osmém roce po výsadbě během prvního obmytního období (rok 2009) dosahovaly sice některé měsíční úhrny mírně vyšších hodnot neţ referenčního travního porostu, nicméně rozdíly nebyly průkazné. Tyto případy s vyšší měsíční u RRD byly typické pro období s vyššími sráţkovými úhrny, a lze proto předpokládat, ţe se jedná o tzv. ztrátu intercepcí, čemuţ nasvědčují i rozdílné hodnoty LAI obou porostů. Celkové úhrny během vegetační sezóny (1. dubna aţ 30. září) 2009 stanovené na základě metody Bowenova poměru a energetické bilance činily v případě RRD v osmém roce prvního obmytního cyklu 509 mm a v případě referenčního travního porostu 533 mm. Počátkem roku 2010 proběhla v zimním období sklizeň topolové plantáţe, coţ vedlo k celkovému poklesu sezónní na pouhých 264 mm. Pro srovnání, travního porostu představovala 474 mm. Celkové roční sumy v roce 2009 činily 572 mm pro RRD a 619 mm pro travní porost. V následujícím roce pak 314 mm pro RRD a 584 u referenčního travního porostu. Sráţkové úhrny během roku 2009 dosáhly 778 mm a v následujícím roce 694 mm. Za zmínku stojí i rok 2011, který sice v práci nebyl podrobně analyzován, ale jehoţ výsledky, byly stručně zmíněny v diskusi. Během tohoto roku byly na lokalitě nově nainstalovány další dva systémy na měření metodou Bowenova poměru a energetické bilance. Jeden z nich byl umístěn nad typický luční trvalý travní porost a druhý nad topolovou plantáţ ve třetím roce druhého obmytního období. Sezónní úhrny (1. dubna aţ 30. září) dosáhly 484 mm pro referenční travní porost a 496 mm pro typickou luční kulturu. V případě RRD dosáhla 466 mm pro porost ve druhém roce po první sklizni a 447 mm u nově sledovaného porostu ve třetím roce druhého obmytního cyklu.

9

Jak jiţ bylo naznačeno, rozdíly v jednotlivých letech lze vysvětlit především odlišnou dynamikou vývoje a absolutními hodnotami , který byl u RRD měřen jak pomocí ceptometru, tak na základě sběru do spadových sítí. V případě referenčního travního porostu, dostupné metody měření neumoţňovaly, a proto byl pouze odhadnut na základě výšky porostu a informací z literatury na 2–3. V roce 2009 došlo u RRD k zahájení olisťování přibliţně v polovině dubna a na konci srpna a září dosáhl maximální hodnoty 7,5. Během roku 2010 byl výskyt prvních listů na obráţejících pařezech zaznamenán aţ počátkem června, přičemţ maximální úrovně olistění bylo dosaţeno rovněţ na přelomu srpna a září s hodnotou 3,7. Tomuto rozdílnému načasování aktivního růstu a rozsahu olistění odpovídaly i rozdíly v produkci nadzemní dřevní biomasy, která byla odhadnuta na základě destruktivních alometrických měření a inventarizace porostu. Výnosy (v sušině) činily pro rok 2009 16,5 t ha-1, oproti tomu v roce 2010 pouhých 3,9 t ha-1. Pro srovnání, v roce 2008, kdy pozorovaný přírůst na kmenech začal přibliţně o dva týdny později neţ v roce 2009, byla produkce nadzemní dřevní biomasy 13,4 t ha-1 (sušiny). Vztah mezi produkcí biomasy a vytranspirovanou vodou lze vyjádřit pomocí tzv. efektivity vyuţití vody ( ). V předkládané práci byl jako uvaţován podíl celkové produkce pouze nadzemní dřevní biomasy a (tedy transpirace i výpar z půdy a povrchů) v průběhu určitého časového intervalu. Během vegetační sezóny 2008 dosáhla celková hodnoty 3,13 g kg-1, v následujícím roce 3,54 g kg-1. V roce 2010 typickým zvýšeným výparem z půdy v důsledku pozdě a málo vyvinutého představovala -1 2,1 g kg . Dosaţené výsledky představují cenný materiál a poznatky v oblasti výzkumu ekologických nároků RRD a mohou přispět především k detailní představě o produkčním potenciálu RRD na daném či pedo-klimaticky podobných stanovištích. Mimo tyto hlavní výše zmíněné výsledky, byla v rámci disertační práce získána řada výstupů ve formě naměřených nebo následně vypočítaných dat dalších meteorologických a ekofyziologických proměnných a parametrů, které svou obsáhlostí mohou slouţit jako exkluzivní materiál pro kalibraci a verifikaci, či pro samotný vývoj dynamických modelů schopných predikovat růst, vývoj a v neposlední řadě spotřebu vody porostů RRD.

10

1 INTRODUCTION The term short rotation coppice (SRC) is generally used for any high-yielding woody species managed in a coppice system usually grown mainly for energy use on arable land. Typically, these crops are harvested on a 3–7 years long rotation and remain viable for 15– 30 years. The SRC plantations in the conditions of central Europe are usually based on poplar or willow species and have been seen recently as a promising source of bio-energy. The replacement of fossil fuel with biomass in the generation of energy and heat has recently been an important strategy promoted by the European Union (EU) to mitigate the effects of climate change and enhance the security of the supply and diversification of energy sources (IEA, 2003). Moreover, biomass and in particular energy crops have attracted attention as a promising renewable and local energy source which could help the EU reduce its dependency on external energy sources, i.e., the main oil-exporting and gasexporting countries (EU COMMISSION, 2005, GASOL et. al 2007). Within the biomass option, SRC plantations feature several environmental advantages. Of all the raw materials, such as winter rape oil, sugarcane, sorghum, soy and palm oil, wood chips show the best performance as biofuel with respect to the total environmental impact and greenhouse gas emissions (SCHARLEMANN and LAURANCE, 2008). Only biofuels produced from waste materials show a better performance (LASCH et al., 2009). Furthermore, the life cycle assessment for poplar SRC plantations in Germany (RÖDL, 2008) confirms the very low CO2 emissions resulting from energy production using biomass from SRC. It produces just 0.015 kg CO2-equivalent per kWh generated electricity. At the other extreme, lignite-fired power plants discharge 1.1 kg CO2-equivalent per kWh. According to ERICSSON et al. (2009) the calculated energy crop production costs were also found to be consistently lowest for SRC and highest for annual straw crops. The production costs of SRC were calculated to be about 4–5 € GJ-1 under present conditions and 3–4 € GJ-1 under improved future conditions. The production costs for perennial grasses were calculated to be about 6–7 € GJ-1 and 5–6 € GJ-1 under present and improved future conditions, respectively. The production costs for annual straw crops were estimated to be 6–8 € GJ-1under present conditions with little potential for cost reductions in the future. In accordance with the results of many research papers, SRCs has additionally positive impacts on their surroundings; e.g. on water cycle, carbon cycle, biodiversity, liveliness of the countryside, and also beside the energetic independence also higher employment rate (ISEBRANDS and KARNOSKY, 2001). Despite these facts, the areas of SRC plantations in the Czech Republic are still small compared with the neighbouring countries (WEGER et al., 2009), however, higher popularity of SRCs and the resolve to plant them has been recorded during the last few years. For successful establishing and choosing suitable sites, it is necessary to understand well the growth requirements of particular species for the planned plantation. According to empirical and modelled results, the water availability of the locality constitutes one of the main constraints for SRC grown on arable land (BRAATNE et al., 1992, CIENCIALA and LINDROTH, 1995, LINDROTH and BÅTH, 1999, DECKMYN et al., 2004, FISCHER et al., 2010). Therefore, the water balance of SRC in the context of their large scale plantations is a crucial and still open topic. The research carried out in other countries across the Europe 11

suggested that the water use of SRC is substantially higher than that of the traditional agriculture crops or grassland and thus they can even have negative impacts on the regional water budget linked with reduced aquifer recharge and river flows or at least their biomass yields and profitability would be strongly dependent on the water availability (HALL et al., 1996, PERSSON, 1997, HALL et al., 1998, ALLEN et al., 1999, DECKMYN et al., 2004, GUIDI et al., 2008, LASCH et al., 2009, PETZOLD et al., 2010). Due to this options linked with low rates of nitrates leaching SRC were recommended as a suitable crop for nitrate sensitive areas or for groundwater protection zones around water supply boreholes (HALL et al., 1996). In contrast to these result, e. g. MEIRESONNE et al. (1999), BUNGART and HÜTTL (2004), LINDERSON et al. (2007), MIGLIAVACA et al. (2009) reported comparable water consumption of SRC to that of grassland and the reference crop evapotranspiration. With respect to these not disunited conclusions the submitted thesis attempts to contribute to the scientific discussion with new research based on measurement of actual evapotranspiration of poplar SRC and the reference turf grass growing in the identical pedo-climatic conditions in the Czech-Moravian Highlands (Czech Republic). In order to assess the water use these two ecosystems situated at one experimental site, the same measuring techniques – namely the Bowen ratio energy balance (BREB) and the soil water balance method, were used for both covers which makes the study unique. According to TRNKA et al. (2008), the yields of selected poplar clones at the site lie within the range 10–14 t ha-1 which can be judged as an economically effective and sustainable. The knowledge of the water use under such biomass production level could prove very valuable, especially for future site selection and other growing and decision support purposes. Because the BREB method used for the actual evapotranspiration estimates within this study has been recently ―in the shade‖ of a much more widely used and direct micrometeorological method called eddy covariance, the thesis additionally focuses more deeply on the error analysis, gap filling process and fetch requirement of the BREB method itself as well as on the supplementing with other independent methods. Moreover, I attempted to provide relatively broad literature overview of methods for evaluating the evapotranspiration in order to point out where the place of the BREB method is and what is the general meaning of the Bowen ratio itself and what its links to the other methods are. Therefore, the thesis reached to more than is the generally recommended length, nevertheless, I hope that the reader will find that everything has its place and meaning and is to a considerable degree important and useful.

12

2 LITERATURE 2.1 Plant water relations Life evolved in water, and water remains the essential medium in which biochemical processes take place. Protoplasm displays signs of life only when provided with water – if it dries out, it does not necessarily die, but it must at least enter an inactive (anabiotic) state, in which vital processes are suspended (LARCHER, 2003). Plants are composed mainly of water. On average, the protoplasm contains 85–90 % water, and even protein and lipid rich cell organelles such as chloroplast and mitochondria contain 50 % water. The water content of soft leaves is particularly high (80–90 % of the fresh weight), as is that of roots (70–95 %). Freshly cut wood contains about 50 % of water. The parts of plants having the least water are ripe seeds (usually 10–15 %); some fat-storing seeds contain only 5–7 % water (LARCHER, 2003). Water acts as a solvent for biochemical reactions, as a medium for the transport and distribution of polar organic molecules (e.g. sucrose in the phloem), inorganic ions (nutrients from root to leaf in the xylem; CO2 or bicarbonates to the site of photosynthesis fixation in the cell), and atmospheric gases (diffusion of oxygen to sites of respiration) (FITTER and HAY, 2002). In terrestrial plants, the aerial parts continually lose water by evaporation; thus, the establishment of suitable water relations is the first requirement for survival. In the case of terrestrial vascular plants (also known as tracheophytes or higher plants), the water and nutrients are absorbed from the soil by root hairs – specialized tubular cells extending roots (PETERSON and FARQUHAR, 1996, EVERT, 2006). Then the water moves by symplastic and apoplastic way (short-distance transport) to xylem composed of conductive vessels and tracheids and non-conductive parenchyma and fibres (BERG, 2008). Through the xylem (long-distance transport) water moves from fine and coarse roots to stem, branches toward petioles and finally leaves, where in mesophyll substomatal cavity is converted from liquid state to gas state and escapes through stomatal apertures into the bulk air in process of (stomatal) transpiration (ROTH-NEBELSICK et al., 2001, LARCHER, 2003, BRODRIBB and HOLBROOK, 2006). Note that small amount of water is lost by cutinized epidermis (cuticular transpiration) and suberized surfaces (peridermal transpiration) beside the leaves also from other organs like stem and branches (LARCHER, 2003, LEUSCHNER et al., 2003). In principle, water movement in the plant along the conductive pathway is driven according to the water potential gradient, since the water flows towards region with more negative values to reach equilibrium. It means that the plant can withdraw the water from the soil only if the water potential of root hairs is lower than that of water in soil capillaries. Toward the leaves the water potential decreases which enable the upward water flow. The steepest water potential gradient is that between the plant surface and dry air. Here also is the largest transfer resistance from the whole pathway caused due to the epidermal resistance to diffusion and to the high energy requirement for the change of water from the liquid phase to the gas phase known as latent heat of vaporization (2441 J g-1 at 25 °). Thus exceptionally high latent heat of vaporization is caused by unusually strong intermolecular hydrogen bonds which among others result in very high specific heat (4.2 J g-1 K-1). This has important implication for the thermal economy of the plants: its high specific heat buffers plant tissues against rapid fluctuations 13

in temperature, whereas its high latent heat of vaporization facilitates leaf cooling by evaporation of water (FITTER and HAY, 2002). As long as the water potential gradient between the leaves and the bulk atmosphere exists, transpiration is taking place, meaning that xylem sap is under tension, rather than under pressure, due to transpirational pull. When transpiration ceases, as it can at night or under conditions of very high relative humidity in the air surrounding the plant, the tension in the xylem becomes very small and the water potential gradient between the soil and leaves tends to, or is equal to zero – equilibrium (SLATYER, 1967, FITTER and HAY, 2002, NOBEL 2009). The leaf water potential under such steady-state nocturnal conditions when transpirations is suppressed termed as pre-dawn leaf water potential is especially informative, since it corresponds approximately to that of the rhizosphere which further inform us about the soil water availability for the plant (HINCKLEY et al. 1978, RICHTER 1997, RANA and KATERJI, 2000, LARCHER, 2003). However, in case of taller trees, the gravitational component of water potential should result in a standing xylem tension gradient of 0.01 MPa per metre increase in height (SCHOLANDER et al., 1965, HELLKVIST et al., 1974, BAUERLE et al. 1999, TYREE and ZIMMERMANN, 2002). Thus, at the top of a nontranspiring 60 m tall tree, leaf water potential should be at least 0.6 MPa more negative than that of the soil instead of being equal (WOODRUFF et al., 2004). In fact, the water potential can even become positive, e.g. in xylem of guttating plants, whose hydatodes located near the ends of veins on the margins of leaves can release drops of liquid water (FITTER and HAY, 2002, NOBEL, 2009). This is the effect of root pressure which occurs when the xylem parenchyma actively releases ions and other osmotically active substances. Water then diffuses from the soil into the root xylem due to osmosis. Root pressure is caused by this accumulation of water in the xylem pushing on the rigid cells that causes sap to rise through a plant stem to the leaves (TANNER and BEEVERS, 2001). This process requires metabolic energy and thus is controlled by the availability of oxygen and prevalent temperatures (LARCHER, 2003, NOBEL, 2009). Root pressure can play important role in refilling embolized vessels caused by cavitation in herbaceous species as well as in certain woody species because the capillary rise of water is not sufficient to refill most airfilled xylem (NOBEL, 2009). Transpiration is affected by external factors to the extent to which they alter the steepness of the vapour pressure gradient between the plants surface and the surrounding air. When leaf stomata begin to open in response to diurnal rhythms or environmental cues, the air outside the leaf is normally not saturated with water vapour. The vapour pressure deficit ( ), between the substomatal cavity and the bulk air, is the driving force for the outward diffusional movement of water molecules, via the stomatal pores (FITTER and HAY, 2002). Thus the intensity of transpiration rises with decreasing air humidity and with rising temperature, i.e. with higher . Also, warming of the leaf surface by strong irradiance leads to a steeper vapour pressure gradient, so that transpiration can also occur despite high air humidity, even if it reaches saturation. This is important for the transport of water and minerals in plants of humid regions (LARCHER, 2003). Wind removes the highly saturated layer of air from the epidermis and replaces it with fresh, unsaturated air. Structural details of leaves such as prominent veins and decurrent edges have the effect of tripping the air flow and causing turbulence even at very low air motion (GRACE, 1983). 14

The aerodynamic relations of the leaves can be described by term boundary layer, which is the layer in the close vicinity of the surface where the friction stress decreases with the height (FOKEN, 2008a). The average thickness of the boundary layer is related to leaf size. Thus small leaves have thin boundary layers which constitute small boundary layer resistances, and their temperatures are never very different from that of surrounding air. Large leaves have thick boundary layers with large boundary layer resistances and temperatures which may differ substantially from that of the surrounding air. At high wind speeds the boundary layer is thinner than at low speeds and the resistance correspondingly smaller (GRACE, 1983). Inside closed plant stands, dense tree crowns, thorny shrubs, tussocks of grass, and cushion plants the force of the wind is greatly reduced, boundary layers become thicker and transpiration is therefore less intense (LARCHER, 2003). Transpiration is strictly dependent on the physical conditions affecting evaporation only as long as the stomata remain open to a fixed degree or firmly closed. Only under these conditions is the amount of water lost proportional to the evaporative potential of the air. The ability of plant to regulate stomatal opening enables it to modulate the rate of transpiration to the requirements of its water balance. The involvement of physiological regulatory mechanism is apparent when the transpiration rates can no longer carry out the evaporative demand. A temporary reduction in stomatal opening is elicited by a decrease in light, low humidity (particularly in connection with wind), water deficit, extremes of temperature, and toxic gases (LARCHER, 2003). Stomata can be regarded as hydraulically driven valves that operate in the aerial parts of land plants. These pores enable the diffusion of gases between the plant´s interior and the atmosphere (ROELFSEMA and HEYDRICH, 2005). Although fully open stomata occupy only 0.5–5 % of the leaf surface area, almost all the water vapour and CO2 exchanged between leaves and the atmosphere passes through these pores (JONES, 1992). In general, light and drought act in an antagonistic manner on stomatal movement. Light induces the opening of stomata to enhance CO2 uptake, while drought causes stomata to close, thereby limiting water loss through transpiration. In nature, stomata will often receive both signals at the same time, as sunny periods frequently coincide with drought. Under these circumstances, plants will have to make a trade of – either open the stomata for optimal photosynthesis, or close them to save the water (ROELFSEMA and HEYDRICH, 2005). The stomatal aperture is controlled by the conformation of the two guard cells surrounding a pore. These cells are generally kidney-shaped, and unlike ordinary epidermal cells, usually contain chloroplasts (FITTER and HAY, 2002). Like most motions exerted by plants, stomatal movements are hydraulically driven. Stomatal movements, however, differ from many other motions exerted by plants, because they are completely reversible (ROELFSEMA and HEYDRICH, 2005). The increase in turgor and thus in volume of both guard cells causes opening of the stomatal pore (ROELFSEMA and HEYDRICH, 2005). In intact plants, guard cells need to accumulate a large volume of solutes to raise their osmotic potential and force the stomatal pore to open. In the early literature, guard cells were believed to convert starch into sugars, as starch disappears from guard cell chloroplast during stomatal opening in the light (STRUGGER and WEBER, 1926). Later studies showed that stomatal opening depends on the uptake of K+ into guard cells (FISCHER, 1968). Increase in turgor is an osmoregulatory process brought about by the active transport of ions. In particular, when K+ is transported 15

from the subsidiary cells into the guard cells the pores open. The closure of stomata induced by abscisic acid (ABA) and by related changes in the cytoplasmatic Ca2+ and the opening of efflux channels. Transport of ions is dependent on energy supply (ATP) and is influenced by endogenous signal substances and effectors (e.g. phytohormones). Thus, the readiness of the stomata to open and close during the course of the day varies according to the states of activity, development, stress, and adaptation. Phytohormones have a large influence on the responsiveness of the stomatal apparatus. ABA decreases the ability to open stomata, while cytokinins promote stomatal opening (LARCHER, 2003). Although the stomata respond to many influences, their movements appear to be chiefly controlled by two circuits, one involving CO2 and the other H2O. The CO2 control circuit operates in response to the partial pressure of CO2 in the intercellular space. Its influence is most apparent in the dark, when the epidermis is exposed to air of varying CO2 concentrations. When the CO2 and partial pressure of the air is abnormally high, reaching 70–150 Pa, the stomata become narrow; when CO2 partial pressure drops they open again. In the light, the CO2 partial pressure in the intercellular space drops as CO2 is used up during photosynthesis (LARCHER, 2003). Stomatal opening in light is largely caused by an indirect effect of CO2 deficiency, however, the direct effect of blue light and photosynthetic active radiation ( ) play also important role (KINOSHITA et al., 2001, SAKAMOTO and BRIGGS, 2002, ROELFSEMA and HEYDRICH, 2005). The H2O control circuit comes into play during water deficiency (dry soil and dry air), when the ABA concentration in plants rises, a process that inhibits the osmoregulation of the guard cells. As a result, the opening capacity of the pores slowly diminishes, so that, under severe water deficiency, the stomata do not respond to external factors and remain closed (LARCHER, 2003). In plant physiology it is often convenient to express fluxes in units of mol m-2 s-1 and concentration gradients in mol of gas mol-1 of air. Conductance then has the same units as flux and resistance has reciprocal units. For conversion, conductance in mol m-2 s-1 must be multiplied by m3 mol-1 (0.0224 at STP – standard condition for temperature and pressure) to obtain units of m s-1, and resistance in units of mol-1 m2 s must be multiplied by mol-3 (44.6 at STP) to obtain units of s m-1 (MONTEITH and UNSWORTH, 2008). As a convenient characteristic assessing how the water use is coupled with the creation, so called water-use efficiency ( ) was proposed. There are many ways describing depending on the scientific disciplines and different measuring methods and approach. Firstly, can be described as a ratio of CO2 uptake and transpiration in the process of photosynthesis (POLSTER, 1950, BIERHUIZEN and SLATYER, 1965, HOLMGREN, 1965, TANNER and SINCLAIR, 1983). In this case we are talking about so called water-use efficiency of photosynthesis ( ) or sometimes referred to as – instantaneous because of the time resolution (BIERHUIZEN and SLATYER, 1965, ZUR and JONES 1984, BALDOCCHI et al., 1987, DENMEAD et al., 1993, LINDROTH et al., 1994, CIENCIALA and LINDROTH, 1995, LARCHER, 2003). Some authors also use the term intrinsic for given as the ratio of net CO2 assimilation rate to stomatal conductance to water vapour based on gas exchange measurement (FICHOT, et al., 2009, MONCLUS et al., 2006, RIPULLONE et al., 2004). For ecological, agricultural and forestry purpose, however, the ratio of dry matter production to water consumption over longer period is 16

more informative than temporary gas exchange ratios. HELLRIEGEL (1883) and MAXIMOW (1923) are considered to be among the first who carried out calculations on the relationship between increase in dry matter and water requirement. By dividing the biomass productivity, expressed as organic dry matter, with the water lost by transpiration or whole evapotranspiration, water-use efficiency of productivity ( ) or long term ( ) is obtained (DE WIT, 1958, LINDROTH et al., 1994, CIENCIALA and LINDROTH, 1995, LINDERSON et al., 2007, FORRESTER et al., 2010). Sometimes the reciprocal transpiration, evapotranspiration or assimilation ratio is used describing the water use per unit of growth (JONES and MANSFIELD, 1972, MASLE et al., 1992). To estimate the of whole ecosystem geoscientist and ecologist commonly use the ratio of the main ecosystem fluxes such as gross primary production ( ) gross ecosystem production ( or net primary (ecosystem) production ( , ) to the water losses by evapotranspiration and thus is obtained (LAW et al., 2002, REICHSTEIN et al., 2002). Depending on which method is used, factors such diurnal variation in root and soil respiration, relative carbon allocation to roots and shoots, turnover of fine roots and leaves influence in resulting (LINDROTH et al., 1994, DICKMANN et al., 2001).

2.2 Water balance concept Driven mainly by solar heating, water is evaporated from ocean and land surfaces, transported by winds, and condensed to form clouds and precipitation that falls to land and oceans. Precipitation over land may be stored temporarily as snow or soil moisture, while excess rainfall runs off and forms streams and rivers, which discharge the freshwater into the oceans, thereby completing the global water cycle (TRENBERTH et al., 2006). Associated with this water cycle, energy, salt within the oceans, and nutrients and minerals over land are all transported and redistributed within the earth climate system (CHAHINE, 1992, SCHLESINGER, 1997). Thus, water plays a crucial role in earth‘s climate and environment. A quantitative evaluation of all water volumes that enter and leaves a threedimensional space over a specified period of time could be simply described as a water balance (BURT, 1999). The basic processes involved in the water balance of a plant are water uptake, water conduction, and water loss, which are driven by the mechanisms described earlier. In larger scale, water balance of a plant stand, or generally whole landscape, is expressed by the water-balance equation: (1) where is income of water in precipitation form, is income of water due to capillary rise, is surface (subsurface) run-off (lateral inflow), is drainage of water to the deeper layers, is water loss from layer due to evapotranspiration and is change of soil water storage in the defined soil layer (e.g. RANA and KATERJI, 2000, LARCHER, 2003, STRANGEWAYS, 2003, MAVI and TUPPER, 2004, FOKEN, 2008a). All the quantities involved are referred to unit ground area and are given as precipitation equivalents in mm of water column (i.e., litre m-2). Note, that this general 17

equation, due to its simplicity and practical applicability, does not account with contribution of horizontal precipitation like especially fog and dew formation from water vapour in the air above the surface and in the soil (PITACCO et al., 1992, AGAM and BERLINER, 2005). According to global observations and model simulations of water balance over land, mean annual precipitation are estimated to 806–864 mm, mean annual evapotranspiration 535–544 mm and mean annual run-off 274–329 mm (BAUMGARTNER and REICHEL, 1975, UNESCO, 1978, LEVIS et al. 1996). These data suggest that evapotranspiration constitute the main loss component of water balance which is in addition directly influenced by the plants or type of the cover. Together with amounts of precipitation and the retention ability refers to a hydric regime of the field, stand or whole landscape and thus becomes an important issue of hydrology, agriculture, forestry, land ecology and other related fields. 2.2.1 Precipitation and infiltration Precipitation is a special category or subset of conditions under which water exists. Specifically, water particles, in either liquid or solid form, to be defined as precipitation must fulfil two conditions – fall from the atmosphere, and reach the ground (WYNNE, 2008). This definition would then include rain, drizzle, snow, sleet, and freezing rain and exclude other forms of water in the atmosphere such as clouds, fog, dew, rime, or frost in that it must fall from the atmosphere even though they can produce trace readings in a gauge of up to 0.2 mm (STRANGEWAYS, 2003). The precipitation available to the plants for maintenance of their water balance is the amount reaching and penetrating the soil (infiltration). In dense plant stands, not all of the precipitation falling on an area (total precipitation) actually reaches the ground. Instead, the water reaching the ground is only the amount that falls through gaps in the plant canopy (canopy throughfall), drips off the leaves, and runs down the stems (stemflow). The amount of precipitation available to plants is termed the available or net precipitation (LARCHER, 2003). The resulting local unevenness in the distribution of precipitation is particularly pronounced in forests. The greater amounts of water reaching the soil in canopy gaps of trees and the periphery of tree crowns have a considerable influence on the spreading of the roots of trees and the understory. The amount of stemflow is greater, the steeper the angle of the branches and the smoother the bark. Therefore, at the bases of trunks of deciduous trees more than 1.5 times as much water infiltrates the soil as in the open space (LARCHER, 2003). By far, most of the intercepted water (crown interception) is lost by evapotranspiration. Only a negligibly small fraction of the water wetting the trees is taken up directly through the leaves and bark. So, for the practical purposes, all of the water retained by the vegetation can be treated as a loss by interception (LARCHER, 2003). A different situation can arise in regions where mist occurs frequently (montane belts, and coastal area near cold oceans currents). As a result of the moisture ―combed‖ by the vegetation from low clouds and passing billows of mist, the available water income can be even greater than the precipitation in the open filed, yielding an overall gain by interception (LARCHER, 2003). The size of the loss by interception depends on the composition and density of the plant cover and on the meteorological conditions prevailing during and after precipitation. 18

Dense crowns of trees with small, easily wettable leaves or needles retain more precipitation than open crowns with large, smooth leaves. Of course, the stage of leaf development is also very important (LARCHER, 2003). Under changing weather conditions, the interception varies widely depending on the amount and type of precipitation or dew formation, on the prevailing temperature and on the wind. In general, more precipitation is intercepted, the finer the drops and the smaller the total amount of water. A certain amount of water is required to wet the whole plant cover thoroughly, and only after this has happened does the water begin to drip from leaves and twigs. In broadleaved forests rain only penetrates the canopy and reaches the understory when about 1 mm (when the trees are in foliage) or about 0.5 mm (when the trees are bare) has fallen; in coniferous forests, 2 mm are necessary before penetration occurs (interception storage capacity). Heath vegetation and grassland hold back roughly 1–2 mm, a peat moss cover about 15 mm, before precipitation reaches the ground (LARCHER, 2003). Infiltration is that part of the field water cycle which addresses the movement of water into the soil. This may be either from water falling on the surface from precipitation, water moving across the surface in run-off or irrigation, the melting of snow on the surface, or from surface water bodies such as ponds. The infiltration rate of water into the soil during a rain precipitation event, when the precipitation rate is greater than the infiltration rate, is seen to vary with time. In cases where the soil was initially relatively dry the infiltration rate begins at a relatively high rate. It then decreases with time. In cases where the water is continually applied to the surface the infiltration rate approaches a constant value. The sorptivity is proportional to the difference between the initial and final water contents and to the field water diffusivity (RASMUSSEN, 2008). During the initial stages of infiltration, when there is a relatively high infiltration rate, both the water entering the soil and the air contained in the soil move downward. Later, as the infiltration rate decreases, air starts to flow upward. The effects of air compression and its flow counter to that of water are sometimes included in the water transport equations (MOREL-SEYTOUX and KHANJI, 1975). As water infiltrates into a relatively dry soil, a soil moisture front is formed between the increased soil moisture from the applied water and the existing soil moisture. When water is applied to a well-drained soil this wetting front moves generally downward, with some lateral movement. The lateral movement depends on the physical characteristics of the soil as well as vegetative root distributions, cracks in the soil and other factors. In many cases the wetting front is found to not be a smooth surface but to exhibit preferential flow channels or regions where the water enters and moves through the subsurface faster than in surrounding regions (RASMUSSEN, 2008). 2.2.2 Surface (subsurface) run-off (lateral inflow) Not all the water reaching the ground is available for evapotranspiration. Some of it runs off the surface of the ground, and some of it percolates into deeper layers and joins the groundwater, which in humid regions is a subterranean form of drainage, appearing at the surface here and there as springs, maintaining connection with the rivers, and eventually draining to sea level (LARCHER, 2003). Water that leaves a site by surface or subsurface 19

flow is called run-off. It occurs when there is a greater input rate of water to a surface site than can leave by the sum of infiltration, evapotranspiration, and rate of storage. Run-off usually results when the precipitation rate, or water application rate, exceeds the infiltration rate. Many chemicals (e.g., those from fertilization of vegetation or pesticides for insect management) can be dissolved in run-off water or transported on detached soil particles moving with it. As a result many non-point source pollution analyses are closely linked with run-off analyses. The amount of run-off and its timing are important to analyses ranging from flood and erosion prediction on watersheds to predicting the rate of advance of irrigation water over a surface irrigated field (RASMUSSEN, 2008). The surface run-off is relatively easy to measure, particularly when circumscribed catchment areas of a river are studied. Subterranean run-off and deep drainage, however, must be estimated indirectly (LARCHER, 2003). The amount of water drained away depends primarily on the slope of the terrain and the type and density of vegetation. Where there are steep slopes, more than half of the precipitation runs off over the surface, and where the precipitation is heavy and vegetation sparse, as much as two thirds to three-quarters may be lost in this way. The high gravitational energy on the steep areas in mountainous regions, and increase run-off combined with decreased water storage capacity of the soil and vegetation, greatly increase the danger of floods, erosion and landslides (LARCHER, 2003). The opposite to the subsurface run-off is so called lateral inflow which usually occurs due to subsurface runoff and drainage from upslope and causes an augmentation of soil moisture on lower slope versus ridges, or on concave versus convex sites (DYER, 2009). Water in the subsurface moves in response to a number of gradients, including those due to gravity, water-content, and temperature. The potential concept is often used to simplify the analysis of the various forces acting on water. The water responds to the total potential it experiences (MARSHALL et al., 1996). The total potential is comprised of the matric, osmotic, gravity, pneumatic, and overburden potentials. For soil beneath the water table, the matric potential is replaced by a submergence potential. Water content and matric potential are often related by water characteristic curves. These curves are not unique but depend on the soil's wetting and drying history. The curves exhibit hysteresis loops referred to as scanning curves. A number of equations have been developed to describe the movement of water in the unsaturated and in the saturated regions of the soil. Darcy's equation relates velocity of water to permeability of the soil and the hydraulic gradient (HILLEL, 1971). For vertical water entry into the soil, what is often measured rather than velocity is the change in volumetric moisture content with respect to time at various depths. In a sufficiently deep soil with no ground water near the surface, the typical moisture profile after infiltration has ceased consists of a wetted region in the upper part of the profile and a relatively dry zone beneath. The post infiltration movement of water from the more moist to less moist regions is called redistribution. The downward movement of the water becomes increasingly slower with time. When the redistribution rate becomes negligibly small, the soil field capacity is reached. This is often taken to represent the upper limit of moisture available to plants. It is not a constant, for while the redistribution rate becomes relatively slow after two to three days it still has a non zero value after days, weeks, or even months in some cases. Soils vary considerable in their ability to hold water 20

at any given time after cessation of infiltration. Field capacity can range from 4 % in sands to 45 % in heavy clay soils (RASMUSSEN, 2008). 2.2.3 Evapotranspiration The combination of two separate processes whereby water is lost on the one hand from the soil (or from the wet canopy) surface by evaporation and on the other hand from the plants by transpiration is referred to as evapotranspiration ( ). Evaporation is the process whereby liquid water is converted to water vapour (vaporization) and removed from the evaporating surface (vapour removal). Water evaporates from a variety of surfaces, such as lakes, rivers, soils and wet vegetation. Energy is required to change the state of the molecules of water from liquid to vapour. Direct solar radiation and, to a lesser extent, the ambient temperature of the air provide this energy. The driving force to remove water vapour from the evaporating surface is the difference between the water vapour pressure at the evaporating surface and that of the surrounding atmosphere. As evaporation proceeds, the surrounding air becomes gradually saturated and the process will slow down and might stop if the wet air is not transferred to the atmosphere. The replacement of the saturated air with drier air depends greatly on wind speed and turbulent mixing of air which is on the other hand influenced by the aerodynamic properties of the surface, namely its roughness. Hence, solar radiation, air temperature, air humidity and wind speed are climatological parameters to consider when assessing the evaporation process. Where the evaporating surface is the soil surface, the degree of shading by the canopy and the amount of water available at the evaporating surface are other factors that affect the evaporation process. Frequent rains, irrigation and water transported upwards in a soil from a shallow water table wet the soil surface. Where the soil is able to supply water fast enough to satisfy the evaporation demand, the evaporation from the soil is determined only by the meteorological conditions. However, where the interval between rains and irrigation becomes large and the ability of the soil to conduct moisture to near the surface is small, the water content in the topsoil drops and the soil surface dries out. Under these circumstances the limited availability of water exerts a controlling influence on soil evaporation. In the absence of any supply of water to the soil surface, evaporation decreases rapidly and may cease almost completely within a few days (ALLEN et al., 1998). Transpiration consists of the vaporization of liquid water contained in plant tissues and the vapour removal to the atmosphere. Plants predominately lose their water through stomata. The water, together with some nutrients, is taken up by the roots and transported through the plant. The vaporization occurs within the leaf, namely in the intercellular spaces, and the vapour exchange with the atmosphere is controlled by the stomatal aperture. Nearly all water taken up is lost by transpiration and only a tiny fraction is used within the plant. Transpiration, like direct evaporation, depends on the energy supply, vapour pressure gradient and wind. Hence, the same climatological variables should be considered when assessing transpiration. The soil water content and the ability of the soil to conduct water to the roots also determine the transpiration rate, as do water logging and soil water salinity. The transpiration rate is also influenced by physiological characteristics, 21

environmental aspects and cultivation practices. Different kinds of species may show different transpiration rates. Not only the plants, but also their development, environment and management should be considered when assessing transpiration (ALLEN et al., 1998). Evaporation and transpiration occur simultaneously and there is no easy way of distinguishing between the two processes. The combination of these two processes is called evapotranspiration and describes water moving into the atmosphere from the Earth (TAIZ and ZEIGER, 2002). In the input-output model of Eq. 1 is an output from the field water system. The rate at which water is taken from the soil by vegetation and passed to the atmosphere depends on the gradients of potential and the conductivities. These depend on a number of factors, the moisture content of the soil, the texture, the relative humidity and temperature of the atmosphere as well as the type of plant, the physiological condition of the plant and its stage of development. As water evaporates from the soil surface, gradients are set up that move soil water to the surface. The rate of movement decreases as water content decreases. Where there is a water table within 2 m of the surface, this movement can continue. Capillary rise will continue to move water from the water table to the surface for evaporation. The introduction of vegetation on the surface adds another process (transpiration) to move water from the subsurface to the atmosphere. The roots of plants reach out into the soil and allow transpiration to remove more water from the soil, and at greater depths, than from evaporation alone. Note that addition of vegetation to a surface area does not mean the resultant water transfer rate to the atmosphere will be the existing evaporation rate plus the transpiration. Since, besides the water availability, mainly the available energy determines the rate of maximum water loss by , the amount of water that is evaporated from the soil surface is reduced by plants and from them originated accumulated vegetative debris shading the soil (RASMUSSEN, 2008). This is the reason why in case of cropped soil the fraction of evaporation decreases over the growing period as the crop develops and the crop canopy shades more and more of the ground area. When the crop is small, water is predominately lost by soil evaporation, but once the crop is well developed and completely covers the soil, transpiration becomes the main process. At sowing nearly 100 % of comes from evaporation, while at full crop cover more than 90 % of comes from transpiration (ALLEN et al., 1998). As the soil water content decreases the energy with which water is held in the soil increases; the work required increases; the plant cannot obtain the water needed and begins to show stress signs such as a loss of turgor pressure in the leaves. With minor water stressing, the plant will recover if water is added to the root zone. If the plant water stress continues a point will be reached for many plants where the addition of water to the root zone will not allow the plant to recover. This is the water content at the permanent wilting point for that soil (RASMUSSEN, 2008). Water depths can also be expressed in terms of energy received per unit area. The energy refers to the energy or heat required to vaporize free water. This energy, known as the latent heat of vaporization (λ), is a function of the water temperature. For example, at 20 °C, λ is about 2.45 MJ kg-1. In other words, 2.45 MJ are needed to vaporize 1 kg or 0.001 m3 of water. Hence, an energy input of 2.45 MJ per m2 is able to vaporize 0.001 m or 1 mm of water, and therefore 1 mm of water is equivalent to 2.45 MJ m-2. The 22

evapotranspiration rate expressed in units of MJ m-2 day-1 (or W m-2 s-1) is represented by (or sometime signed as or ), the latent heat flux (ALLEN, 1998). 2.2.4 Drainage and capillary rise Some of the water that infiltrates, or has arrived in the subsurface by other paths such as subsurface flow, leaves the region by movement in natural drainage to another lateral region, by movement to artificial drains installed to remove water, and/ or by deep percolation to deeper regions of the subsurface. In all cases, drainage water is considered to be moving away from the near surface region of the subsurface and is thus an output in Eq. 1. In contrast, capillary rise of water toward the soil surface is considered as a negative drainage. For a water table near the surface, the capillary rise of water will continue to transfer water from the water table to the surface where it is evaporated or taken up by plant roots. The movement of water from the saturated regions of the subsurface (aquifer) to the surface (e.g., springs) is also regarded as negative drainage. These situations range from water flowing to the surface from a confined aquifer that is under pressure to the gravitational flow of water out the side of a hill slope (RASMUSSEN, 2008). 2.2.5 Soil moisture Water in the near surface region is stored primarily in the pores of the soil. Other, smaller storage sites are surface depressions and the tissue of vegetative material. The interception of precipitation can also be thought of as a storage as can the accumulation of snow, even though their average residence time may be less than that associated with other forms of storage. The porosity of the subsurface material indicates how much water can be contained per unit volume of saturated material (TODD, 1980). A gravel aquifer can hold 0.20 m of water per meter depth of the aquifer while a clay aquifer can hold 0.45 m of water per meter depth. The specific yield of these aquifers suggest about 88 % of the water held in the gravel aquifer will drain out (can be pumped out) while only 7 % of the clay water will drain. The saturated hydraulic conductivity values for these materials indicate how fast water moves in the aquifer in response to a given gradient. Since hydraulic conductivity values for these two materials differ by roughly five orders of magnitude, a similar difference exists for water flow velocity in these aquifers, and the extraction time for a selected volume of water (RASMUSSEN, 2008). Soil moisture (or soil water content) refers to the water that occupies the space between soil particles. It is at its maximum when the soil is saturated, that is when all the air between soil particles is replaced by water but, if the soil can drain, the spaces will normally also contain air, the water then forming a thin film on and between the soil particles, held by capillary attraction. As the soil dry out this film becomes thinner and progressively less easy for plant root to extract. The water is free to move through the soil, up or down, by gravity and by capillary attraction; it is taken up by plant roots, evaporates at the surface or recharges the groundwater (STRANGEWAYS, 2003). The spaces between the soil particles contain air, water and roots. When the soil pores contain only water then the water can move freely, the largest force acting on it being that of gravity which causes it to move downwards provided there is drainage. It then drains away until an equilibrium 23

is reached known as the field capacity, at which point any further water added cannot be retained and will drain away. As the soil dries out (for whatever reason), the water in the pores decreases in volume and retreats to increasingly small spaces between the particles, becoming held more and more tightly by capillary attraction. Under anything less than field capacity there is said to be soil moisture deficit. As drying continues, the suction that plant roots must apply to extract water increases. This suction or soil tension (or soil water potential or soil matrix potential) is measured in the same units as barometric pressure. At saturation (no air spaces), the tension is zero while at field capacity it is around -0.3 bar (300hPa). As drying continues, the remaining water is eventually held so tightly that is not possible for plants to apply enough tension to extract it, this state being known as the permanent wilting point, occurring at about -15 bar. The exact value differs between plants types, but at this point they die. However, most of the water available to plants is held at pressure of less than -1 bar, although negative pressures can rise as -60 bar (STRANGEWAYS, 2003). Knowing the strength of soil tension is value to agriculture, since it acts as a warning of when it is necessary to start irrigation as well as when it is wasteful to continue. But the significance of soil tension is also due the control it exerts on the movement of water within a soil profile. If combined upward tension exceeds the pull of gravity, it is possible for water to move upwards and be evaporated or transpired. If gravity and downward tension combined are the stronger force, water moves downward, eventually joining the water table. But temperature gradients within the soil also cause water to move (GILMAN, 1977), liquid water moving from colder to warmer zones, vapour from warm to cold (by evaporation and condensation) (STRANGEWAYS, 2003). Groundwater is then that part of the soil moisture which does not rise upwards and evaporate instead percolates downward and joins a region in which the soil or permeable rock is saturated with water without airspaces. From here, in due course, the water finds its way into drainage ditch, stream, river or lake and eventually to the sea. Measuring the depth of the water table, changes in its level and the quality of the groundwater is of importance not just to researchers investigating hydrological processes but to those concerned with the practical matters of water resources, civil engineering, agriculture and pollution (STRANGEWAYS, 2003).

2.3

Measuring water loss

How it was pointed earlier, evapotranspiration is the major part of the global water balance loss over land. ROSENBERG et al. (1983) stated that 70 % of the precipitation striking the earth´s surface returns back to the atmosphere by evaporation or evapotranspiration. Further, in arid regions even 90 % of the precipitation returns back to the atmosphere due to these evaporative processes (BURMAN and POCHOP, 1994). Furthermore, RANA and KATERJI (2000) referred that even 99 % of water used in agriculture is lost by crops as evapotranspiration. Moreover, semi-arid and, mostly, arid climates have a great impact on crop growth in conditioning yield and product quality. Under these weather conditions, rain-fed crops and agricultural fields with limited water resources are often submitted to water stress. Thus, it becomes fundamental to know the 24

exact losses of water by evapotranspiration – i.e. actual evapotranspiration ( ), and the crop water status and its influence on production (RANA and KATERJI, 2000). can be measured (directly or indirectly) or estimated. Due to complex interactions amongst the components of the land-plant-atmosphere system, is perhaps the most difficult of all the components of the hydrologic cycle to assess (XU and SINGH, 2005) and as we will see later, there is actually only one really direct method for measuring . Whereas must be precisely measured for research purposes in plant eco-physiology, for farm irrigation management it can be just estimated. The lower the degree of accuracy in the estimation, the greater will be the water waste by incorrect management of irrigation. Generally, it is convenient to discuss the methods of determining in considering separately the measurement and modelling aspects. There is a great variety of methods for measuring . Some methods are more suitable than others for accuracy or cost or are particularly suitable for given space and time scales. For several applications, needs to be predicted, so it must be estimated by model (RANA and KATERJI, 2000). This chapter is focusing on method of direct or indirect measurements of the evapotranspiration whereas the next chapter deals with the bases of modelling. Measurement is generally the process of determining the value of a physical quantity (length, weight, temperature etc.) experimentally with the help of special technical means called measuring instruments. The value of a physical quantity is the product of a number and a unit adapted for these quantities. It is found as the result of a measurement (RABINOVICH, 2005). In metrology there has been a long-standing tradition to distinguish direct, indirect, and combined measurements. In the last few years, metrologists have begun to divide combined measurements into strictly combined measurements and simultaneous measurements. This classification is connected with a definite method used for processing experimental data to find the result of a measurement and to estimate its uncertainty. In the case of direct measurements, the object of study is made to interact with the measuring instrument, and the value of the measurand is read from the indications of the latter. Sometimes the instrumental readings are multiplied by some factor or corresponding corrections are made in it, etc. In the case of indirect measurements, the value of the measurable quantity is found based on a known dependence between this quantity and its arguments. The arguments are found by means of direct and sometimes indirect or simultaneous or combined measurements (RABINOVICH, 2005). Simultaneous and combined measurements employ close methods for finding the measurable quantities. In both cases, they are found by solving a system of equations, whose coefficients and separate terms are obtained as a result of measurements (usually direct). In both cases, the method of least squares is usually employed. But the difference lies in that in the case of combined measurements, several quantities of the same kind are measured simultaneously, whereas in the case of simultaneous measurements, quantities of different kinds are measured simultaneously (RABINOVICH, 2005). Depending on the properties of the object of study, the model adopted for the object, and the definition of the measurable quantity given in the model as well as on the method of measurement and the properties of the measuring instruments, the measurements in each of the categories mentioned above are performed either with single or repeated observations. If a measurement is performed with repeated observations, then to obtain a 25

result the observations must be analyzed statistically. These methods are not required in the case of measurements with single observations. For this reason, the number of observations is an important classification criterion (RABINOVICH, 2005). The methods of measuring should be divided into different categories, since they have been developed to fulfil very different objectives. One set of methods are primarily intended to quantify the over a long period, from weeks to months and growth season. Another set of methods has been developed to understand the process governing the transfer of energy and matter between the surface and atmosphere. The last set of methods is used to study the water relations of individual plants or part of plants (RANA and KATERJI, 2000). This chapter is organized following ROSE and SHARMA (1984) and RANA and KATERJI (2000), according to convention to place the variety of methods in groups, where the main approach or method depends on concepts from hydrology, micrometeorology and plant physiology. 2.3.1 Hydrological approaches 2.3.1.1 Water balance The first of the here described method comes from the basic water balance equation and applying the well known Lomonosov-Lavoisier law of mass conservation with the assumption that no mass in the space can be lost, but only transformed to another form or change its location. Thus the is obtained as a residual term in the water balance equation which can be applied at different spatial and temporal level – from small plots (~10 m2) to a large catchment (~10 km2) ranging from a week to a year long period. Since it is often very difficult to accurately measure all the terms of Eq. (1), a number of simplifications make this method unsuitable for precise estimation. Often, for operational application, the soil water balance Eq. (1) is expressed in its simplified form (RANA and KATERJI, 2000): (2) This is usual relation used for determining evapotranspiration of watershed or smaller catchment by measuring the precipitation, the streamflow in rivers and the amount of water stored in the ground (RODDA et al., 1976, MAIDMENT, 1993, SAVENIJE, 1997, DOW and DEWALLE, 2000, STRANGEWAYS, 2003). Evaluating water balance of smaller fields is usually based on measuring precipitation and soil moisture. The run-off can be also measured but more ordinarily is simply considered as a certain percentage of precipitation (ALLEN et al., 1998) or precipitation exceeding specific threshold, which is determined by local conditions – slopeness, soil type, etc. (HLAVINKA et al., 2011). More complex approach constitute so-called run-off curve numbers ( ), which result from several empirical parameters used in hydrology for predicting direct run-off or infiltration from rainfall excess. method is based on the area's hydrologic soil group, land use, treatment, hydrologic condition and antecedent run-off condition (MAIDMENT, 1993, PONCE et al., 1996). In arid or semi-arid and non-hilly areas, run-off component can be ignored (HOLMES, 1984, ALLEN et al., 1998). But in fact, it depends on the occurrence and characteristics of the precipitation (amount, duration and intensity) and can only be 26

neglected for a particular type of soil (JENSEN et al., 1990), i.e. coarse (sand and loamy sand) and moderately coarse (sandy loam). Unlike other, precipitation and soil moisture present two water balance components which can be measured with considerably less assumptions and estimations, and in addition, due to their magnitude (especially in most of agriculture areas), is their determination a crucial point for deriving total water balance and thus the estimation of evapotranspiration. In fact, irrigation water supply is, in principle, known and precipitation can be measured by rain gauges, but all the other terms need to be measured or, at least, estimated. According to SEVRUK and KLEMM (1989) there are over 50 different types of rain gauge in common use worldwide, all differing in size, shape, material, colour and the height at which they are exposed above the ground – from 0.2 to 2.0 m (STRANGEWAYS, 2003). To deal with this non-unification, the World Meteorological Organisation (WMO) organised field intercomparisons of the major national gauges of the world, with pit gauges as the reference, at 60 sites in 22 countries (SEVRUK and HAMON, 1984, GOODISON et al., 1998). The WMO experiment has developed bias correction procedures for many precipitation gauges commonly used around the world (GOODISON et al., 1998). However, such intercomparsions lump many rain events together over a period and can only give a rough correction (STRANGEWAYS, 2003). The rain gauge inherent errors include deformation of the wind field above the gauge orifice, wetting losses attributable to wetting of the inner walls of the orifice and container, evaporation of the gauge contents, splash-in and splash-out and blowing snow. Wind speed is the most important environmental factor contributing to the under measurement of precipitation, particularly snow; the proportion of snow in the total precipitation affects the accuracy of seasonal and annual totals (GOODISON et al., 1998). Measurements of the speed-up of wind over collectors have shown that the underestimation is particularly dependent on the size and shape of the orifice rim: smaller the rim, the smaller the increase in wind speed and thus higher catch of rain drops (SEVRUK et al., 1991). SEVRUK (1996) developed mathematical simulation techniques of the wind-induced error, using a three-dimensional numerical procedure. Wind speed and raindrop size are the two key factors in how different rain gauges behave. If measurements of these are available, it becomes possible to correct readings with higher precision (STRANGEWAYS, 2003). Another issue of the point measurement by rain gauges is the spatial inhomogenity of rain events. Hence, it is always more suitable to use several gauges to obtain reasonable average (STRANGEWAYS, 2003). The next measurable and in the same time key issue of water balance is the amount of water stored in the soil. Note that there is also water present in soil which is not free to move or to be taken up by plants but which may nevertheless be detected during measurement – but not differentiated from the free water, depending on the measurement technique used. This is water of crystallisation or water of hydratation – water that is chemically bound to minerals within the soil such as gypsum (CaSO4 2H2O). In addition, water may also be bonded to organic material to varying degrees of strength (STRANGEWAYS, 2003). In making measurements of soil water content, soils of great variety are encountered and any instrument or measurement method must be able to handle them all, from pure peat or sand to silt and clay or mix of them, all varying in pore size or having a variety of pore sizes combined, and all varying in the extent of chemically bonded 27

water. Anomalies, such as stones dispersed at random in the soil profile, can also affect measurements since stones do not usually hold any water and if not detected can give rise to large errors, again partly depending on the method of measurement. The terms ―soil moisture‖ and ―soil water‖ are generally interchangeable although the later could be interpreted as including the chemically bound water, which, as explained above, is not in the form of moisture (STRANGEWAYS, 2003). Soil moisture (the soil water of our interest) can be measured by a wide number of different methods with different probes, however three of the approaches remain the most spread and used all around the world – namely gravimetric, neutron and electromagnetic methods (ROBINSON et al., 2008, STRANGEWAYS, 2003). First of the named, the gravimetric method had been the only way of measuring soil moisture until the electronic techniques were developed. Today it is especially important for calibrating the indirect electronic method and it remains still in routine where costs must be kept at minimum, or where only a few measurements are needed (STRANGEWAYS, 2003). The principle of the gravimetric method lies in taking the soil sample with known volume (e.g. with soil corer), immediate weighing it in the fresh state (loss of water by evaporation must be prevent), drying it in a ventilated oven at 105 °C till a constant weight is reached (it takes usually 12 to 48 hours), and finally reweighing (ROBINSON et al., 2008, STRANGEWAYS, 2003, TOPP and FERRÉ, 2002). Due to the drying process, sometimes more correct term termogravimetric method is used (ROBISON et al., 2008). More practical and generally more used way of this method is to collect the soil samples with unknown volume and then following the same methodology described above. Thereby the moisture dry weight fraction is obtained instead of the volumetric water content which is necessary for further water balance calculation. This shortcoming is simply solved by using the value of bulk density known from some of the previous soil sampling. This simplification make this more applicable, however, it demands the assumption that the bulk density of soil is fairly consistent within the area around the sampling point, but as with other variables there is always some spatial variation which makes this approach more inaccurate (STRANGEWAYS, 2003). The main advantage of the tremogravimetric method is its theoretical simplicity and the fact that it is a direct method of measuring soil moisture. However, this directness is given by the destructive way of measurement which on the other hand brings the inability to investigate the same volume of soil more than once and together with the time consuming character makes this method suitable only for large time scale (weekly or greater). From the list of the indirect method, the neutron probe is arguably the most accurate which is especially due to the linear relationship between the subject of measurement (the ratio of transmitted and returned neutrons) and the soil moisture making the calibration more straightforward. Neutron probe consist of tube with source of fast neutron and the detector of the slowed-down or thermalized neutrons, the access tube permanently installed in the soil to the depth appropriate to needs, and the necessary accessories for data storing or displaying (STRANGEWAYS, 2003). Neutrons carry no charge and are unaffected by electromagnetic fields. Released from a source, they will travel in a straight line unless they collide with the nucleus of an atom. A direct collision between a neutron and a hydrogen nucleus causes the neutron to lose all its energy and become thermalized 28

(ROBINSON et al., 2008). By comparison, direct collisions with oxygen, silicon, aluminium, or iron will lead to energy losses of 22, 14, 13, and 7 %, respectively (GARDNER et al., 2001). Collisions also cause the neutrons to change direction; the more collisions, the greater the likelihood of them returning to a detector near a source. Therefore, a count of the number of thermalized neutrons returning to a detector, over a sufficient period of time, gives a good estimate of the number of hydrogen nuclei in soils (ROBINSON et al., 2008). Because not all but most of the hydrogen atoms in the soil are those contained in nonchemically bonded water molecules (i.e. soil moisture), this method also have to be locally calibrated (STRANGEWAYS, 2003). The sampling volume is dependent on the soil moisture and generally described by a sphere of influence less than 0.15 m in wet soils and extending to as much as 0.5 m in dry soils (ROBINSON et al., 2008). That´s why this method is not suitable for investigating the near surface layer (STRANGEWAYS, 2003) and tends to lose the accuracy in arid conditions (PAYNE and BRUCK, 1996). In addition to increasingly strict rules for using radioactive materials, the need for an operator and relatively slow data acquisition have led to reduced use of the neutron probe, other than for deep borehole work where there is little alternative (YAO et al., 2004). Another indirect approach for assessing soil moisture is the electromagnetic method. This technique is based on determining the relative permittivity, , (dielectric constant) of soil, which inform us about how many times is the permittivity of soil, , (or generally any other material) greater than the permittivity of vacuum, . Since the value of for water ( 80) is significantly greater than that of most soil matrix materials ( 4) and of air ( 1), the bulk permittivity of soil is strongly affected by the soil moisture (ROBINSON et al., 2008). There are three main method, how to measure the permittivity of soil. First of them, the time domain reflectometry (TDR) method involves measuring the propagation of an electromagnetic pulse along the transmission lines (wave guides). By measuring the travel time, the velocity and hence the apparent dielectric constant of the soil can be estimated (LUKANU and SAVAGE, 2006). Usually, the TDR method is not soilspecific (DRNEVICH et al., 2005), and therefore no soil calibration is required. The second one, capacitance method, is based on rapidly charging and discharging a positive and ground electrode (capacitor) in the soil causing that an electromagnetic field is generated whose charge time is related to the capacitance of the soil which is dependent on its dielectric permittivity. Thus, the soil moisture can be determined by measuring the charge time of a sensor buried in the soil (GROOVES and ROSE, 2004, KIZITO et al., 2008). The third method is titled as frequency-domain reflectometry (FDR) method makes use of radio frequencies and the electrical capacitance of a capacitor (formed by using the soil and embedded rods as a dielectric) for determining the dielectric constant and thus the soil water content. The signal reflected by soil combines with the generated signal to form a standing wave with amplitude that is a measure of the soil moisture (GASKIN and MILLER, 1996, LUKANU and SAVAGE, 2006). There is some uncertainty with classing this third method. Sometimes it is taken into the TDR group and titled modified TDR or sometimes it is lumped together with the capacitance probes. In reality, both capacitance probes and TDR overlap in principle because both stem from dielectric constant of the soil, albeit in different ways (STRANGEWAYS, 2003). Because all of the electromagnetic methods are based on 29

measuring soil , they are subjected to similar limitations. Usually, the proper placing of the sensor or its access tube into the soil is the crucial point of installation. Especially air gaps and stones can substantially influence the resulting value of soil moisture, nevertheless the same effect can be evoked by roots, earthworm channels or cracks in shrinking clay later after the installation. The difference between the TDR and capacitance sensor can be also seen in their dimensions – TDR are much robust, which can be advantage due to integrating larger soil volume and thus reducing the influence of the mentioned air gaps. On the other hand, capacitance probes fill an important niche because their probe geometry is more adaptable than TDR for short electrodes and boreholes applications (ROBINSON et al., 2008). Compared with the neutron probe, the electromagnetic methods sample smaller volume of soil (sphere of influence) which makes the method more vulnerable than the neutron probe to local anomalies within the soil such as stones and other (unknowable) spatial variations. In consequence, for applications such as estimating irrigation needs, which do not require precise quantitative measurements, for measurements near to the surface and when automatic logging is necessary, dielectric methods have much in their favour. But for accurate, precisely calibrated data, collected manually, the neutron probe is still difficult to better (STRANGEWAYS, 2003). Sensor calibration of electromagnetic methods is a two-step process, from signal response to permittivity and from permittivity to soil moisture while the calibration of neutron probe is direct – counting ratio to soil moisture (ROBINSON et al., 2008). In dielectric method, there is some uncertainty as to which water molecules are measured (the question of their levels of binding and water of crystallisation), although this is less of a problem if it is changes in water content that matter (STRANGEWAYS, 2003). 2.3.1.2 Lysimeters By isolating the plant root zone from its environment and controlling the processes that are difficult to measure, the different terms in the soil water balance Eq. 1 can be determined with greater accuracy. This is done in so-called lysimeters where the plant grows in isolated tanks filled with either disturbed or undisturbed soil (ALLEN, et al. 1998). By suitable design, the contents of the lysimeter can be made to behave in a similar way to their surroundings, provided that the temperature of the soil is kept the same and the drainage is similar. To produce the correct drainage it may be necessary to install some means of suction at the base of the container, to ensure that the soil moisture tension is the same as in the freely draining ground outside the lysimeter. The bigger the better is generally the case for lysimeters, since then edge effects are reduced and internal and surface differences are smoothed out (STRANGEWAYS, 2003). Lysimeters vary greatly in size and design, but can be subdivided into two types: weighing and non-weighing (BRYE et al., 1999). In non-weighing lysimeters the evapotranspiration for a given time period is determined by deducting the drainage water, collected at the bottom of the lysimeters, from the total water input (ALLEN, et al. 1998). Non-weighing lysimeters, much easier and less disruptive to install, can be designed to collect drainage water under zero or non-zero tension lower boundary conditions. A saturated zone must develop above a zero-tension 30

lysimeter before water can enter the drainage collection device installed at its base. This perched water table may create very unnatural soil water flow conditions in arid regions with deep unsaturated zones (REEDER, 1986). Therefore, zero-tension lysimeters are better suited for collecting drainage under wet soil, shallow groundwater conditions. BRYE et al. (1999) were successful in developing an equilibrium tension lysimeter. They reported lower variability in measured drainage between their sampling locations than that reported for comparably sized zero-tension lysimeters installed in other soil systems. However, their system does require a vacuum pump to apply the tension. A soil water flux meter designed by GEE et al. (2002) appears to be a very promising method of measuring drainage efficiently and accurately. Their flux meter design uses a passive capillary wick system to apply tension to the soil and does not require a vacuum pump, as does the system of BRYE et al. (1999). Drainage is the key component in a soil water budget with regards to quantifying transport of pollutants to groundwater. Therefore, it is imperative that an accurate assessment of the amount and timing of drainage be done to guide such things as agricultural nutrient management strategies (PARKIN, 2008). Weighing lysimeters are more expensive and usually require disturbing and re-packing of the soil profile. In this type of lysimeters a mechanical balance is the best way of measuring the weight because it is the very small changes in weight (due to rainfall and evapotranspiration), not the absolute total dead weight of soil and container, that are wanted. Load cells can be used, but they weight the total mass and so must be able to resolve small changes in weight in excess of the very large overall weight. It is also possible to use a hydraulic method of weighing, in which the lysimeter stands on a flexible container filled with oil connected by tubing to a manometer above the ground where it can be read. The change in weight, with due allowance for rain (which is measured separately) and for any drainage (which is stored and measured below the lysimeter), is the amount lost by evaporation. A large mechanical balance, perhaps supporting several tonnes of container, is expensive, and so too is the structure in the ground that houses them. Weighing lysimeters are not, therefore, for general use, but they are one of the best methods of measuring evapotranspiration. Observing strict terminology, it is more correct to call lysimeters that are sealed and have no drainage evapotranspirometers and only to call those that drain lysimeters (STRANGEWAYS, 2003). In precision weighing lysimeters, where the water loss is directly measured by the change of mass, evapotranspiration can be obtained with an accuracy of a few hundredths of a millimetre, and small time periods such as an hour can be considered (ALLEN, et al. 1998). A basic requirement of lysimeters is that the vegetation both inside and immediately outside of the lysimeter be perfectly matched (same height and leaf area index). This requirement has historically not been closely adhered to in a majority of lysimeter studies and has resulted in severely erroneous and unrepresentative crop evapotranspiration estimates. As lysimeters are difficult and expensive to construct and as their operation and maintenance require special care, their use is limited to specific research purposes (ALLEN, et al. 1998).

31

2.3.2 Micrometeorological methods Micrometeorological methods of measurement determine evaporation as the flux of water vapour through the air from the evaporating water surface, vegetation, or soil. The measurements are made in the atmosphere, within the turbulent air close to the ground, so that the measured vapour flow rate is a very good approximation to the surface evaporation rate. There are two broad classes of micrometeorological evaporation measurement: those based on measurement of gradients and those based on measurements of fluctuations (MAIDMENT, 1993). Both rely on the fact that turbulent exchange is the dominant exchange mechanism within the near-surface atmosphere. Since micrometeorological measurements are necessarily made some distance above the ground, and the atmosphere is almost always moving horizontally, the measurements obtained at a particular location are representative of an area some distance upwind. The area upwind from the measurement point from where the fluxes are registered is called source area or more recently footprint (GASH, 1986). It can be an advantage in that the upwind turbulent mixing helps to produce a value representing the average evaporation over a fairly large area. However, if the measured evaporation is meant to be representative of the particular uniform crop surrounding the instruments, it is necessary that there should be an extensive ―fetch‖ of evaporating surface with essentially identical properties extending upwind from the measurement site for a considerable distance (MAIDMENT, 1993). Unlike the hydrological methods, in micrometeorology is the evapotranspiration investigated directly as the intensity of flux intensity of water vapour or latent heat. The base for these methods is the energy (radiation) balance equation and the theory of turbulent diffusion. The energy balance equation is as follows: (3) where is net heat gained through radiative exchanges, allowing for the absorption of both downward and upward components of long and short wave radiation, is the evapotranspiration rate multiplied by the latent heat of vaporization in order to express evapotranspiration in units of energy – latent heat flux, is the heat lost by convection – sensible heat flux, is conduction of heat in the soil – soil heat flux, is the rate at which heat goes into storage within the vegetation and is the rate at which energy is being trapped in chemical bonds by photosynthesis. All the terms are in appropriate units of energy flux, preferably W m-2 (GRACE, 1983). Because and are negligible compared to other components of energy balance, the Eq. 3 is usually simplified to the following form: (4) The net radiation can itself be broken down into four terms referring to either solar or thermal radiation: (5) where and are the solar (short wave) and atmospheric (long wave) radiation respectively, is the surface albedo (the reflected fraction of incident short wave radiation), and are the surface emissivities for absorption and emission of long wave 32

radiation respectively, is Stephan-Boltzman constant (5.67 x 10-8 W m-2 K-4) and is the surface temperature. and are generally considered to be equal (BRUTSAERT, 1982). The albedo depends on the nature of the surface. For example, MONTEITH and UNSWORTH (2008) referred its values 0.26, 0.24, 0.18 and 0.16 for wheat, grass, deciduous and coniferous forest respectively. Soil albedo varies between 0.1 and 0.35 for most soils depending on their type, moisture and surface roughness (BRUTSAERT, 1982, CELLIER, 1996). Assuming that the heat transfer in soil is only the result of conduction, the soil heat flux can be written as: (6) where is the vertical soil temperature gradient at the soil surface, and (m) the depth in the soil (CELLIER et al., 1996). The soil apparent thermal conductivity, (W m-1 K-1) depends on the soil texture, bulk density and water content (DE VRIES, 1963, BUSSIERE et al., 1992). For further understanding of the latent and sensible heat fluxes, it is necessary firstly to look briefly into the theory of turbulent diffusion. The vertical transport of heat, water vapour, carbon dioxide or any other gas between the surface and the surrounding atmosphere depends on air movement. When air flows over a rough surface such as vegetation, surface layers are retarded by friction, giving rise to a region of locally-reduced wind speed, called the boundary layer. Unless the air is flowing very slowly over a very smooth surface, the stresses set up between adjacent air layers are sufficient to break up the laminar flow, causing chaotic motion of the air which is said to be turbulent. In turbulence, parcels of air (called eddies) moving at random, transport heat, carbon dioxide and water vapour from regions of high concentration to regions of low concentration, the overall process being called turbulent diffusion or just simply turbulent transport (GRACE, 1983). The transition from laminar to turbulent flow depends on the relative magnitudes of inertial forces associated with the horizontal movement of the fluid and viscous forces generated by inter-molecular attraction (sometimes referred to as ―internal friction‖). The ratio of inertial to viscous forces is known as the Reynolds number ( ) after the physicist who first showed that this ratio determined the onset of turbulence when a liquid flows through a pipe. When is small, viscous forces predominate so that the flow tends to remain laminar, but when the ratio increases beyond a critical value , inertial forces dominate and the flow becomes turbulent (MONTEITH and UNSWORTH, 2008). The flux, , of any entity in the vertical directions is proportional to the concentration gradient : (7) where the constant of proportionality, (m2 s-1), is the turbulent transfer coefficient (sometime called the eddy diffusivity or exchange coefficient) (GRACE, 1983). In laminar boundary layers, where the streamlines of flow are almost parallel to the surface, vertical transport depends on molecular diffusion and is one to five orders of magnitude slower, the constant of proportionality then being , the molecular diffusivity. The latter depends on the entity being transported and on the fluid involved. In contrast, is more or less 33

independent of the entity being transported and depends on the characteristics of the turbulence. Near the surface where the wind speeds are low and the eddies small, might lie between 10-4 and 10-1 m2 s-1. Above the vegetation, where wind speeds are high and eddies large, is always much higher (GRACE, 1983). The flux density Eq. 7 representing the Fick´s law of diffusion is often treated as an analogue of Ohm´s law in electrical circuits. Because the gradient of a quantity at a point is often difficult to estimate accurately, Fick´s law is generally applied in the integrated form: (8) The denominator on the right hand side can be replaced with so called aerodynamic resistance (s m-1): (9) and thus the flux of any entity can be written according to Ohm´s law in the final form: (10) In laminar flow, the movement of material is predictable, and results in the development of a boundary layer with well-defined profiles of velocity, concentration, and temperature. In contrast, turbulent flow is unpredictable both spatially and temporally. However, just as the depth of a laminar boundary layer above a flat plate increases with distance from the leading edge, the depth of a turbulent boundary layer can be related to the fetch or distance of traverse (x) across a uniformly rough surface that generates turbulence by shear at the surface (MONTEITH and UNSWORTH, 2008). For practical reasons, the thickness of the boundary layer created above the certain surface is very important, because the upper bound of this equilibrium layer is also the upper limit where the measurements of fluxes from this surface are relevant (FOKEN, 2008a). There are several studies which investigate the thickness of the equilibrium layer as a function of fetch (GARRATT, 1990, SAVELYEV and TAYLOR 2001, SHIR, 1972, BRADLEY, 1968). Because of the large scatter of the results from all the experiments, most authors assume a simplified empirical relationship: (11) originally proposed by (ELIOT, 1958) instead of more complicated functions. For fetch distances ranging between 0.01 and 20 km, RAABE (1983) had obtained values for the parameters and to be 0.30 ± 0.05m and 0.5 ± 0.05 respectively. Over much shorter distances (fetches from about 0.01m to 160 m), WALMSLEY (1989) has adopted the values: = 0.75 m and = 0.8, based on validated experimental data (JEDEGE and FOKEN, 1999). According to FOKEN (2008a), the relation proposed by RAABE (1983): (12)

34

can be classified as exceedingly robust for the determination of the new equilibrium layer. The mentioned relationships are correcting the commonly quoted fetch to height ratios (x/δ) in the range 100 to 200. Especially for taller crops and forests, it is necessary to be aware, that the high of the equilibrium layer is not estimated from the ground level, however, from the imaginary horizontal level in canopy called zero plane displacement, (m), above which in distance of roughness length, (m), the wind speed tends to be zero (IRVINE et al., 1997). STANHILL (1969) found that the average value of zero plane displacement for a range of primarily agricultural vegetation was close to 0.63 of the mean vegetation height, but this can usually vary between 0.6 to 0.8 with respect to the vegetation density (GRACE,1983). The roughness length is one order smaller, usually estimated between 0.08 and 0.12 of the mean canopy height (MONTEITH and UNSWORTH, 2008). In the lowest atmospheric layers, in the so-called surface layer, the turbulent fluxes can be considered as constant with height. Thus, according to the theory of turbulent diffusion (flux gradient theory) the latent heat flux is given by (STEDUTO and HSIAO, 1998, FOKEN, 2008a): (13) where is the density of air (kg m-3), is latent heat of vaporization (J kg-1), the 2 -1 exchange coefficient (m s ) for latent energy flux, is specific humidity difference (kg -1 kg ) and between the two heights and is the difference in height (m), the last two are making together the water specific humidity profile gradient . The specific humidity, , express the mass of water vapour per mass of moist air (kg kg-1) and can be replaced with sufficient accuracy by the mixing ratio which describes the mass of water vapour per mass of dry air (kg kg-1) (FOKEN, 2008a). Providing that the ratio of the molecular weights of water and dry air is equal to 0.622, the relation between water vapour pressure, (kPa), atmospheric pressure, (kPa), and the specific humidity can be written (PEREZ et al., 1999, FOKEN, 2008a, MONTEITH and UNSWORTH, 2008) as follows: (14) By introduction of the psychrometric constant,

(~ 0.066 kPa K-1),

where is the specific heat capacity of air at constant pressure (J kg-1 K-1), the latent heat flux can be determined in the following very common form based on measuring the water vapour pressure profile gradient (PEREZ et al., 1999, PEACOCK and HESS, 2004, GUO et al., 2007, SAVAGE, 2010): (15) Similarly, the sensible heat flux

is given by: (16)

35

where is the exchange coefficient for sensible heat flux (m2 s-1) and is the air temperatures differences (°C) between the two heights enabling to calculate the temperature profile gradient . Within this thesis, energy fluxes are positive if they transport energy away from the earth´s surface (into the atmosphere or into the ground) and net radiation is positive during the day when the incoming short wave radiation from the sun prevails. Both and are positive upward and negative downward, opposite to the gradient directions. is then positive downward to the soil and negative upward when the heat is coming from the deeper layer toward surface. The same energy flux and radiation sign convention was used e.g. in PEREZ et al. (1999), GUO et al. (2007) or MONTEITH and UNSWORTH (2008). Same convention for energy flux direction but opposite for net radiation is used e.g. by FOKEN (2008a). Finally, some authors use sign convection that under normal conditions during daytime hours, surface- to atmosphere or soil-directed fluxes are indicative of energy losses and are considered as negative, whereas the incoming short and long wave radiation from sun and sky respectively is positive (SAVAGE, 2010). It means that there is no strict rule for radiation and energy flux signs, but they have to be always defined by the author. In the same way as the flux of water vapour or heat, flux of any other entity can be described. When the entity is horizontal momentum defined as a product of horizontal wind speed (m s-1) and the mass of air (kg or ), the gradient between two levels above the ground will exist, because the wind speed is decreasing toward the ground due to the drag forces exerting on the air mass blowing over the surface with certain roughness. Provided that the vertical flux is constant with height, this quantity can be identified as the force per unit ground area (N m-2), known as the shearing stress ( ) which can be written in following form (MONTEITH and UNSWORTH, 2008): (17) where is a turbulent transfer coefficient for momentum and is the differences in wind speed in two different heights (m). Often not the shearing stress but the generalized velocity, the so called friction velocity, , is used and described as (FOKEN, 2008a): (18) Consequently, it can be written that: (19) where and represent the vertical and horizontal wind velocity fluctuations respectively, providing the friction velocity is a measure of mean eddy velocities (MONTEITH and UNSWORTH, 2008). From experience, both wind speed and turbulent mixing increase with height which can be express in the following relation: (20)

36

where is a constant – the von Karman constant – and based on measurements it is usually assigned value 0.40±0.01 (HÖNGSTRÖM, 1996, FOKEN, 2008a). By introduction this relation into the Eq. 17, the following relation will be obtained: (21) Consequently, integration of Eq. 21 between the limits

and

yields: (22)

where is a constant termed roughness length, such that when . This does not imply that the real wind speed is zero at height (m), because the assumptions made in deriving Eq. 22 may fail near the boundary, so the limit should be regarded as a mathematical convenience. A more general form of Eq. 22 takes account of the height of the surface elements by assuming the vertical displacement of the zero plane to a height (m) known as zero plane displacement such that the distribution of shearing stress over the elements is aerodynamically equivalent to the imposition of the entire stress at height . The wind profile Eq. 21 then becomes (MONTEITH and UNSWORTH, 2008): (23) or (24) and Eq. 20 becomes: (25) The similarity hypothesis states that, in neutral stratification, the turbulent transfer coefficients of momentum, sensible heat, water vapour or any trace gas are equal which is the basis of Monin-Obukhov similarity theory (MOST) (MONTEITH and UNSWORTH, 2008). In neutral stability the eddy structure can be envisaged as a set of circular eddies with diameters increasing linearly with height by factor (known as Karman´s constant) and given by the mixing length , rotating with tangential speed equal to the friction velocity , i.e. (THOM, 1975): (26) In unstable (lapse) conditions, which occur when the surface is strongly heated, vertical motion is enhanced by buoyancy. The amount of enhancement increases as the wind shear (depending on viscosity) decreases, and this is illustrated in Fig. 1, where the eddies are progressively stretched vertically. Thus exceeds , where is still given by but is greater than . Conversely, in stable conditions (inversion), for example on a clear night with light winds, vertical eddy velocities are damped, and so , with but .

37

Figure 1: Wind speed profiles and simplified eddy structures characteristic of three basic stability states in air flow near the ground (from THOM, 1975).

The qualitative effect of stability on the shape of the wind profiles is apparent in Fig.1 a–c, and is summarized in semi logarithmic form in Fig. 1 d. In each of the examples in this figure the momentum flux transmitted to the surface is assumed the same, so is constant. Since by differentiating Eq. 24: (27) this requires the gradient of each profile at the lowest level to be the same. As height increases, velocity gradients become smaller in unstable conditions and larger in stable conditions than those for neutral case. The differential wind profile (Eq. 24) can therefore be written in generalized form as: (28) where is a dimensionless stability function with value of unity in neutral stability and larger or smaller than unity in stable or unstable conditions, respectively. 38

Using the relation between momentum flux and gradient: (29) it can be readily shown that: (30) Stability functions can be defined for sensible heat (31) and latent heat flux (32) and similarly for other entities by . The relation between , , , and or the equivalent functions , , , and have been a source of considerable argument in micrometeorology and a number of empirical relationships have been proposed. In neutral stability, all entities are transported equally effectively and all profiles logarithmic in the constant flux layer, , and . In unstable conditions, exceeds because there is preferential upward transport of heat. Measurements (reviewed by DYER, 1974) support the view that in unstable conditions. In slightly to moderately stable conditions, DYER (1974) inferred that , but as stability increases turbulence is increasingly damped, and then suddenly changes to a quasi laminar, non-turbulent, flow and the concepts underlying similarity theory become invalid (FOKEN, 2008a, MONTEITH and UNSWORTH, 2008). The dependence of the functions on stability is generally expressed as a function of parameters that depend on the ratio of the production of energy by buoyancy forces to the dissipation of energy by mechanical turbulence. The two best established parameters are the gradient Richardson number , calculated from gradients of temperature and wind speed: (33) and the Obukhov length

, which is a function of fluxes of heat and momentum (34)

where is absolute temperature (K), and the gravitational acceleration (m s-2) (MONTEITH and UNSWORTH, 2008). From the definition of and it can be shown that: (35) The Richardson number can be understood as the ratio of the kinetic energy due to buoyancy to that due to shearing forces. is negative, zero and positive under lapse (unstable), neutral and inversion (stable) conditions, respectively (MALEK, 1993). The Obukhov length scale called Obukhov length (Obukhov, 1946) is then an apparent expansion or contraction of the turbulent boundary layer (WOODWARD and SHEEHY, 1983). 39

A physical interpretation of is that the absolute value of the Obukhov length is equal to the height of an air column in which the production ( < 0) or the loss ( > 0) of turbulent kinetic energy by buoyancy forces is equal to the dynamic production (turbulence generated by wind shear) of turbulent kinetic energy per volume unit at any height multiplied by (FOKEN, 2008a). Initially, the notation Monin-Obukhov length was used, but this is, in historical sense, not exact (BUSINGER and YAGLOM, 1971). In unstable conditions, DYER and HICKS (1970) concluded that: (36) i.e.: (37) for

.

From measurements in stable and slightly unstable conditions, WEBB (1970) deduced the empirical relation: (38) i.e.: (39) for

.

2.3.2.1 Bowen ratio and energy balance method The Bowen ratio and energy balance method (BREB) is one of the most common methods used to determine the fluxes of sensible and latent heat. The method is based on Bowen ratio similarity and the energy balance equation (FOKEN, 2008a). The Bowen ratio (BOWEN, 1926) is defined as the ratio of the sensible to the latent heat flux: (40) which using the Eqs. 15 and 16 can be written as: (41) According to the Bowen ratio similarity principle (e.g. VERMA et al., 1978), for all the type of stratification and thus the Eq. 41 can be written in the more simple form: (42) enabling to calculate the Bowen ratio by measuring the air temperature and water vapour pressure in two different levels in the atmosphere (SAVAGE, 2010). Then combining of the Eqs. 4 and 42 (radiation balance and the Bowen ratio) results in:

40

(43) (44) The Bowen ratio similarity principle is from a certain point of view simplification producing consequently some limitations of the method (FOKEN, 2008a). In addition to this simplification, BREB method does not include the wind velocity and does not prescribe a certain difference between the measurement heights. To ensure that a sufficient turbulent regime exists, FOKEN et al. (1997) recommended that only the measurements with wind velocity at the upper height greater than 1 m s-1 and difference of the wind velocities between both heights greater than 0.3 m s-1 should be used for correct flux determination. This requires additional instrumentation with anemometers. Even though the height difference of the measurements is not included into the equations, an increase of also increases the difference of the temperature and the humidity. Consequently the influence of the measuring errors decreases (FOKEN, 2008a). It is therefore recommended to choose the ratio of measuring heights (taken from zero plane displacement, i.e. aerodynamical height) greater than 4–8 (Foken, 2008a). These requirements are seldom taken into account in practise because measurements over high vegetation have ratios of the aerodynamical heights of about 1.5 (BERNHOFER, 1992, BAR et al., 1994). As a consequence of using the available radiation partitioning by Eqs. 43 and 44, there is the inherent impossibility of BREB method to estimate correctly the fluxes when , since the denominator in these equations approaches 0 and they lose their physical meaning (OHMURA, 1982, PEREZ et al., 1999, FOKEN, 2008a, SAVAGE et al., 2009). These values imply that and therefore that the available energy is near zero. From that reasons, Bowen ratio values within the range of -1.25 < β < -0.75 are usually rejected (TANNER et al., 1987, UNLAND et al., 1996, ORTEGA-FARIAS et al., 1996, FOKEN, 2008a). Based on pioneering work of OHMURA (1982), PEREZ et al. (1999) and later GUO et al. (2007) and SAVAGE et al. (2009) proposed methodology for data discarding without fix range in the vicinity of -1 but variable and dependent on the sensors accuracy. In addition, when heat fluxes change their sign and the differences are within the range of sensors resolution limits, BREB method can provide an incorrect direction to the fluxes so the data must be discarded (OHMURA, 1982, PEREZ et al., 1999). To determine the correct sign, OHMURA (1982) suggested following criteria, rewritten in terms of the water vapor pressure and temperature gradients: – –

(45)

If the criteria are not fulfilled, the fluxes must be deleted. However, all the mentioned uninterpretable values usually occur in the evening and morning hours or during the precipitation when the absolute magnitude of the fluxes is low, so such data exclusion does not have such dramatic practical impact. BREB measurements are still of interest since the method is less expensive than eddy covariance and scintillometry and there is reduced operator skill required with fewer 41

corrections to the data applied (SAVAGE et al., 2009). Furthermore, the method will remain of interest until the issue of lack of closure of eddy covariance measurements will be solved (FOKEN, 2008a, FOKEN, 2008b). For example, WOLF et al. (2008) used a modified BREB method to investigate the effects of different eddy covariance correction schemes on energy balance closure and SAVAGE (2009) used BREB method to validate surface-layer scintillometer measurements. The so-called modified Bowen ratio method based on the concept originally proposed by BUSINGER (1986) has been recently commercially introduced as a new alternative technique (LIU and FOKEN, 2001). In this approach the sensible heat flux is directly determined with a sonic anemometer (10–20 Hz) and a correction for the buoyancy flux. Subsequently, if is known and the Bowen ratio is derived from the gradient of air temperature and humidity measurements (Eq. 42), the can be solved as the only unknown in Eq. 40 without any need for net radiation and soil heat flux sensors. Nevertheless, this method underestimates the energy balance closure in a similar way as do eddy covariance (LIU and FOKEN, 2001). The assumptions made when using the Bowen ratio approach are (ANGUS and WATTS, 1984): Turbulent transfer coefficients for heat and water vapour are identical. A recent study (HEIKINHEIMO et al., 1999) found that the similarity theory applied in all atmospheric stability situations, thus confirming earlier work such as BUSINGER et al. (1971) and CRAWFORD (1965). MCNAUGHTON and LAUBACH (1998), in their similarity studies, conclude that Bowen ratio measurements will be satisfactory except with quite intense inversions and that using the assumption will only result in minor error. The two levels at which temperature and humidity are measured must be within the layer of the airflow that has adjusted to that surface so that there is an absence of horizontal gradients of temperature and humidity. This assumption is assessed by estimating required fetch. Note, that BREB method can be used also for measuring the flux of CO2 if its concentrations in two vertical levels (same as air temperature and humidity) are precisely measured (STEDUTO and HSIAO, 1998). This approach relies on the assumptions that the exchange coefficients for flux of CO2, and are equal. The exchange coefficient can be then obtained from Eq. 15 or Eq. 16 if the and as a result from standard BREB system are known, and finally the general Eq. 7 can be used to derive the flux of CO2. 2.3.2.2 Aerodynamic method The second technique based on gradient measurements is called aerodynamic method. The core of the method is the momentum flux Eq. 17 and the relationship between eddy diffusivity and the general flux-gradient relationship of any entity described in Eq. 7. The similarity hypothesis states that, in neutral stability, the transfer coefficients of momentum, sensible and latent heat flux or flux of any other entity or trace gas are equal (MONTEITH and UNSWORTH, 2008). Taken into account this equality, we can rearrange the Eqs. 17, 7, 15 and 16 into the following relation:

42

(46) which can be subsequently rearranged applying the relationship from Eq. 18: (47) (48) (49) In neutral stability, can be estimated from the wind profile alone, and so the aerodynamic method requires only two sets of profiles: temperature or concentrations of water vapour or any gas measured together with wind speed in series of identical heights. The friction velocity is found from the wind profile, and the gradient is found by plotting values of against (similarly for and other measurable entities) (MONTEITH and UNSWORTH, 2008). Fluxes are then calculated from Eqs. 47 and 49. If a fast response anemometer is available, a hybrid eddy covariance/aerodynamic method measuring directly (Eq. 19) may be used avoiding the empiricism of wind profile analysis and in the same time keeping costs lower due to unnecessity of the gas analyzer (MONTEITH and UNSWORTH, 2008). An alternative way of applying the aerodynamic method eliminates by differentiating the wind profile equation (Eq. 24) obtaining Eq. 27 which if substituted in Eq. 48 and results in: (50) The minimum number of heights over which the gradients may be determined is two. Similar equations can be written for sensible heat flux, or any other entity – e.g. carbon dioxide. However, the main defect of the aerodynamic method using Eq. 49 is the dependence on wind and humidity (or temperature or mass) at two heights only, so that the estimate of flux is sensitive to the error in a single instrument or to local irregularities of the site. More accurate estimates of flux can therefore be obtained from Eqs. 47 to 49 using four or more measuring levels (MONTEITH and UNSWORTH, 2008). In non-neutral conditions it is necessary to know profiles of and or to estimate from wind profile analysis, and the equality of eddy diffusivity coefficients cannot be assumed. Therefore, the Eq. 50 takes the generalized form: (51) where is stability factor defined earlier by Eqs. 36–37. Similar equation may then be written for and generally (MONTEITH and UNSWORTH, 2008). Apart from the necessity to use the stability universal function to derive the fluxes by the vertical profile measurements of wind speed and other scalars there is one more issue which produces limitations of the aerodynamic method. According to the surface layer similarity theory, the flux profile relationships on which it is based are only valid in so 43

called inertial sublayer which is the upper part of the surface boundary layer (CELLIER and BRUNET, 1992). The lower one is called roughness sublayer where the turbulent structure is strongly influenced by the wakes generated by the canopy roughness elements (MÖLDER et al., 1999). Therefore, it was for a long time assumed, that the method is not practically suitable for determining the fluxes over high and aerodynamically rough canopies like forest due to the fact that the measurements would has to be carried out at very high levels where the limitation by fetch are usually encountered (THOM et al., 1975) and is rather suitable for measurements above agricultural crops (CELLIER and BRUNET, 1992). Note that the roughness sublayer can usually extend to heights from several roughness lengths to several canopy heights above the top of roughness elements (THOM et al., 1975, CELLIER and BRUNET, 1992, MÖLDER et al., 1999, MONTEITH and UNSWORTH, 2008). However, CELLIER and BRUNET (1992) and later MÖLDER et al. (1999) introduced a corrections which enables to measure even within roughness sublayer where the eddy diffusivity for momentum and other scalars are enhanced above their inertial layer values. Note that since the BREB method does not rely on the equality of the exchange coefficient of momentum and other scalars, but only on the similarity of transfer processes for latent and sensible heat (or CO2) flux which still holds also in the roughness sublayer, its application close to the roughness elements is not restricted. It practically means that in case of BREB method the lower level can be adjusted just above the top of the canopy (STANNARD, 1997). 2.3.2.3 Eddy covariance As it was mentioned, turbulence in the atmospheric boundary layer ensures mixing and diffusion of latent, sensible heat and trace gas fluxes to and from underlying surface. Using sonic anemometers, turbulent motion is measurable with a high level of precision and with a high degree of spatial and temporal resolution. If the transported entity is measured with equivalent precision and resolution, it is possible to directly measure the fluxes using the eddy covariance method (EC). The vertical component of the fluctuating wind is responsible for the flux across a plane above a horizontal surface. Because there is a net transport of energy across the plane, there is a correlation between the vertical wind component and temperature, water vapour and trace gases. For example, if water vapour is released into the atmosphere from the surface, updrafts will contain more vapour than downdrafts, and vertical velocity (positive upwards) will be positively correlated with vapour content. Therefore, the covariance of vertical wind speed and the investigated entity are used to estimate the flux density (DREXLER et al., 2004). In turbulent flow, vertical flux of any scalar can be presented as: (52) where is the dry air density, is vertical wind velocity component and is the mixing ratio ; i.e. the mass of the substance per unit mass of dry air. By using the Reynolds decomposition method, we can break the Eq. 52 into means and deviations by following way: (53) 44

By opening the parentheses and removing the averaged deviation from the average we obtain form: (54) Then the conventional eddy covariance is characteristic by two main assumptions. First, the density fluctuations are assumed negligible during the averaging time. This assumption may not work for example with strong winds over a mountain ridge when the density fluctuation may be large and should not be ignored. But in most cases when eddy covariance is used conventionally over flat and vast spaces, such as fields or plains, the density fluctuations can be safely assumed negligible. Secondly, the mean vertical air flow is assumed negligible for horizontal homogenous terrain, so that no flow diversions or conversions occur. If these two main assumptions are fulfilled, we can use the relation: (55) saying us that the flux is equal to the product of the mean air density and the mean covariance between instantaneous deviations in vertical wind speed and mixing ratio (BURBA and ANDERSON, 2010). This equation has been already mentioned for description of the square of the friction velocity (Eq. 19) and can be simply applied to express the latent heat flux: (56) as well as sensible heat flux: (57) Except these two energy fluxes, the recent commercially available systems are able to determine the fluxes of carbon dioxide (most usual combination with water vapour analysers) , methane, nitrogen monoxide and ozone. Apart from the decision which fluxes will be evaluated, researchers have to choose between so-called open-path and closed-path analysers design. The former quantify scalar concentrations in situ, close to the point where the vertical wind speed (usually by means of a sonic anemometer) is measured. The latter are situated at some distance from this point (typically in a weather-proof enclosure at the base of the tower) and quantify scalar concentrations of air sucked down through a tube from an intake close to the anemometer (HASLWANTER et al., 2009). According to the Eq. 55, the turbulent fluctuations of the components of the wind vector and the scalar parameter must be measured at a high sampling frequency so that the turbulence spectra can be extended to 10–20 Hz (FOKEN, 2008a). The measuring devices used for such purposes are sonic anemometers for the wind components and sensors that can measure scalars with the required resolution in time. The later are often optical measuring methods. The measuring or sampling time depend on the atmospheric stratification, the wind velocity, and the measuring height. For short sampling periods, the low frequency contributions to the fluxes are missed, and for long sampling times the steady state conditions may not be fulfilled. Generally, there will not be remarkable errors if a sampling time of 30 min is used (FOKEN, 2008a).

45

Flux measurements using the EC method are the only direct technique without any applications of empirical constants (BUSINGER, 1986, KAIMAL and FINNIGAN, 1994, LEE et al., 2004). However, the derivation of the mathematical algorithm is based on a number of simplifications so that the method can be applied only if the given assumptions are exactly fulfilled, otherwise the specific corrections must be used. Because necessary corrections cannot be seen from the recorded data, extensive test must be made. Most of these tests will be done after applying all corrections. Most important are the tests for steady-state conditions and developed turbulence (FOKEN, 2008a). Because some values, like e.g. the stability parameter, must be calculated from the corrected data, but they are also necessary for correction, the correction should be calculated by iterative procedure (FOKEN, 2008a). Among the basic and widely used corrections belongs coordinate rotation (tilt correction) (LEE et al., 2004), spectral correction in the high frequency range and spectral correction in the low frequency range or sensor separation correction (MOORE, 1986), correction of the buyoancy flux measured with the sonic temperature (KAIMAL and GAYNOR, 1991) and density fluctuation correction (WEB et al., 1980). Apart from the corrections, closely connected data quality control and subsequent gap-filling procedure has to be applied in order to obtain reliable uninterrupted data series. While EC method is not dependent on the flux-gradient relationship, there is practically no restriction to measure within the roughness sublayer. However, due to low friction velocity just above the canopy, the EC should be placed usually at least one to two meters above. The EC can be also subjected to error if the size of the instrument sensing path is larger than the dominant eddy size. This can occur close to aerodynamically smooth surfaces (MONTEITH and UNSWORTH, 2008). Despite the elegance and a sound theoretical basis, the EC method has one big shortcoming. Namely the resulting fluxes very often underestimate the energy balance closure even when all of the sophisticated and physically-based corrections were applied (FOKEN, 2008a, FOKEN, 2008b). When measuring the latent heat flux, such underestimation is troublesome but not so important as when measuring CO2 flux. Undermeasurement of the latent heat flux is less worrisome because the extent of loss can be estimated (and perhaps corrected) by calculating the recovery ratio for surface energy fluxes relative to a measured energy budget ( ). However, in the case of CO2, the extent of underestimation cannot easily by quantified by subsidiary measurements. Moreover, the underestimation is more pronounced during the night which is again less important for latent heat flux (which is generally small during that time), however, great care is needed when measuring CO2 flux because the long-term average net carbon exchange between the atmosphere and the ground is measured as the small difference between large daytime and night time fluxes, and the likely under-measurement using the EC is very difficult to estimate at these times (SHUTLEWORTH, 2007). 2.3.2.4 Scintillometry Unlike the previous micrometeorological methods which provide rather point measurement even with respect to their footprint, scintillometry is relatively new method for indirect measuring of path averaged latent and sensible heat fluxes over landscape from 46

100 m up to more than 10 km in one dimension. Scintillometer is an instrument that consists of a transmitter and receiver which are separated by a distance , the path length. The transmitter and the receiver have a certain aperture diameters ( and ). The radiation emitted by the transmitter is scattered by the turbulent medium between the transmitter and the receiver. It is just the fluctuation of the scattered radiation that carries information about the turbulent field along the scintillometer path (MOENE, et al., 2005). The scattering of the emitted radiation is caused by the variation of the refractive index which is rarely absolutely constant. In fact, the refractive index of air, the medium of our interest, is not strictly equal to one but is lightly oscillating in the range 1 ± 10-6 (MOENE, 2011). These small fluctuations of the refractive index of air cause intensity fluctuations of emitted radiation known as scintillations. Some examples that clearly show the distortion of wave propagation by the turbulent atmosphere, which can be seen regularly by the human eye, are the twinkling of stars, image dancing and image blurring above hot surface (MEJNINGER, 2003). Since the radiation source of the scintillometer can be interpreted as a collection of point sources, the propagation of a spherical wave through a random medium may be analysed. TATARSKII (1961) derives the log-intensity fluctuation in a spherical wave due to the passage of that wave through a medium with random refractive index fluctuation. The log-intensity fluctuation can be linearly related to structure parameter of refractive index, by following expression: (58) where is a path averaged value of , is the wave number of the electromagnetic radiation ( ) and L the path length (MOENE, et al., 2005). This proportionality of and play important role for separating the scintillometers in different groups. The first of them are called small aperture scintillometers (SAS). A scintillometer is considered as SAS when its diameter aperture diameter is smaller than the first Fresnel zone ( ). The proportionality between and for SAS is only valid when remains smaller than 0.3, otherwise the signal becomes saturated (CLIFFORD et al., 1974). The saturation occurs if the turbulence becomes too intense and an increase of

no longer

results in an increase of and the proportionality between the two is lost (MOENE, et al., 2005). This means that for near infrared to visible wavelengths the optical path is restricted to short distances of approximately 250 m. In the radio wavelength region saturation is less likely and the distance can be several kilometres. Turbulent scales in the order of the first Fresnel zone primarily cause the scintillations observed by SAS. Depending of the operational wavelengths of the light source varies between several millimetres (near infrared region) to a few meters (radio wave region). A near-infrared SAS is therefore very sensitive to inner scale effects. Outer scale effects are relevant in the propagation statistics of radio waves meaning that non-isotropic conditions can distort the measurements. Another problem is that at radio wavelengths also absorption fluctuations by water molecules influence the intensity statistics (MEJNINGER, 2003).

47

The large aperture scintillometer (LAS) and the extra large aperture scintillometer (XLAS), which operate in near infrared region, are designed to overcome the saturation effect of the near-infrared SAS. Due to the increased aperture small-scales structure are filtered out, which lead to a reduction of the amount of scintillations. As a result the LAS and XLAS can operate over longer distances, i.e., the proportionality between and remains valid under strong turbulent conditions. Although the LAS and XLAS also have a saturation maximum, it has not been thoroughly investigated. Another advantage of the LAS is that the instrument is most sensitive to eddy sizes in the order of its diameter (LAS, = 0.15 m; XLAS, = 0.32 m), which lie far from the inner scale and outer scale. As a result the LAS is less sensitive to inner scale and outer scale effects. Temperature, humidity and to a lesser extend pressure fluctuations cause fluctuation in the refractive index of air . By neglecting pressure fluctuations, can be related to the structure parameters of temperature follows (HILL et al., 1980):

, humidity

, and the covariant term

as

(59) where and are functions of the wavelength and the mean values of temperature, humidity and atmospheric pressure (HILL et al., 1980, ANDREAS, 1989). For visible and near-infrared wavelengths ( between 0.36–3μm) and are defined as follows: (60) (61) where is the specific gas constant for water vapour (461.5 J K-1 kg-1). At radio wavelength ( > 3 mm) and are slightly different (KOHSIEK and HERBEN, 1983, ANDREAS, 1989): (62) (63) To solve the Eq. 59, KOHSIEK (1982a) suggested measuring at three different wavelengths in order to obtain , and . However, the problem is that there is no wavelength where only humidity fluctuations are dominant. To some degree temperature fluctuations always play a role at most wavelengths. KOHSIEK and HERBEN (1983) proposed as an alternative to use two wavelengths plus an extra relation between and fluctuations instead. Following their suggestion, ANDREAS (1989) found that a combination of a visible to near infrared scintillometer and a near-milimetre to radio wave scintillometer, denoted as the ―two-wavelength method‖, is the best option for measuring the fluxes of and . In case is measured at only one wavelength, WESELY (1976) and more recently MOENE (2003) have shown that for scintillometers operating at visible or near-infrared wavelengths, is related to by following, very practical relation: 48

(64) where , the Bowen ratio, is expressed as: (65) The last part in deriving the fluxes from the signal obtained by receiver is based on application of MOST providing that in the surface layer, the vertical fluxes of momentum, heat, humidity and other scalars are nearly constant with the height. Assuming stationary conditions and horizontal homogenous surface MOST describes the relationship between structure parameters of temperature and humidity and the related surface fluxes of sensible and latent heat as follows: (66) with

is the height of the scintillometer above the surface, the displacement height and the Obukhov length and and are temperature and absolute humidity scales respectively, derived in similar way as the friction velocity from momentum flux (Eq. 18) as follows: (67) (68) The universal stability function

is defined as follows for unstable: for

where

and

(69)

and for stable conditions as: for

(70)

where (ANDREAS, 1988, THIERMANN and GRASSL, 1992, DE BRUIN et al., 1993, WYNGARD et al., 1971). Note, that despite that the validity of MOST as a key prerequisite for deriving the fluxes from scintillometer signal rely on the assumptions of homogeneous surfaces, the recent studies reveal that MOST scaling function also holds over heterogeneous surfaces like tall and sparse vegetation (EZZAHAR et al., 2007) In order to solve , the friction velocity is required. The friction velocity can be determined using different techniques. First, measure the inner scale using a small aperture scintillometer (HILL et al., 1992a, THIERMANN and GRASSL, 1992). Once the inner scale is known, and the fluxes can be solved iteratively. Second, an eddy covariance system can be used to measure the friction velocity derived by applying the

. Third,

can be

profile method proposed by HILL et al. (1992b) thereby using 49

two large scintillometers installed at two different heights. Finally, the friction velocity and the temperature scale (and thus ) can be obtained iteratively using additional wind speed data with estimate of the surface roughness put into the following relation: (71) where is the height of wind speed measurement, and applying the Eqs. 34 and 69. The main advantage of the first and third method is that path averaged values for can be derived. The second and later methods are ―traditional‖ point techniques for estimating . For these cases one can question their representativeness over non-homogenous areas (MEJNINGER, 2003). For very unstable atmospheric conditions

the following simple expression

for sensible heat flux can be derived: (72) where (73) varies between 0.48 and 0.57 (KOHSIEK, 1982b, DE BRUIN et al., 1995). This expression is also known as the free convection limit and provides a simple method to determine directly from without knowing . In practical applications the free convection approach can provide accurate fluxes when the scintillometer is installed relatively high above the surface ( 10 m). Despite its simplicity it must be noted that the measurement height of scintillometer should be measured accurately because is linearly related to the measurement height above the displacement plane, which can be complicated especially in non-flat areas (MEJNINGER, 2003). All of the recent commercially available systems are optical scintillometers and therefore follow the above described calculation procedure leading to obtain the path averaged sensible heat flux. In order to derive the latent heat flux the knowledge of the available energy ( ), and in case of the heterogeneous surfaces its spatial variability, is necessary to solve the basic energy balance equation (Eq. 4). For these purposes, one or better a network of net radiometers and soil heat flux plates has to be used. However, deploying a high number of sensors is costly and really not particularly feasible. Therefore, estimates of available energy based on local measurements of upscaled using any physically based land surface model is a suitable and effective solution for large areaaveraged determining by scintillometry (SAMAIN et al., 2011). 2.3.3 Plant physiology approach So far described hydrological and micrometeorological methods are suitable and developed for measuring the evapotranspiration at stand or ecosystem level. This means that they take the area of interest as a complex and they are not able to directly distinguish what is the rate of transpiration by vegetation, evaporation from soil and interception. For 50

that reasons, the approach based on measurement of water loss through particular leaves, whole plant or group of plants can provide detailed view on pure transpiration. On the contrary, the up-scaling such results to level of stand or ecosystems demands again some portion of estimation, which can constitute substantial constraint of these methods. The plant physiology methods can be divided into three main groups: porometry, sap flow and chamber systems. 2.3.3.1 Porometry Porometer is an instrument designated for measuring leaf stomatal conductance ( ) or transpiration itself (DUGAS et al., 1993). Transpiration rates measured in a small chamber clumped on the leaf surface are not useful parameters in themselves since they depend on properties of the particular chamber environment as well as those of the plant. It is usually very difficult to match the chamber environment to that of a leaf outside of the chamber closely enough so that the measured transpiration rates will be representative of the outside leaves. Thus a conductance of a leaf to water vapour loss is derived from the transpiration rate. Conductances are usually presented rather than their inverse, a resistance, since conductances are proportional to the flux and they express the regulatory control exerted by the stomata on transpiration rates (PEARCY et al., 1989). Conductance to water loss, , is derived from Fick´s law of diffusion and can be expressed in its simplest form as the proportionality between the rate of transpiration and the driving force for evaporation, the gradient in water vapour from the intercellular spaces in the leaf to the atmosphere (PEARCY et al., 1989). If the gradient of water vapour is expressed as a mole fraction (mol mol-1) and the transpiration in molar units (mol m-2 s-1) then the units of are mol m-2 s-1. Formerly, if the humidity gradient is expressed as a concentration gradient (g m-3) and the transpiration in mass units (g m-2 s-1) the resulting units of are m s-1 (PEARCY et al., 1989). Leaf transpiration can be calculated from if leaf temperature and air temperature and humidity are known. However, measurements of transpiration using porometers are not likely to be representative of the whole plant (MCDERMITT, 1990) because the attachment of the porometer chamber modifies the leaf micro-environmental conditions of wind speed (and thus the boundary layer conductance) as well as air temperature and humidity (FICHTNER and SCHULZE, 1990). Nevertheless, if the chamber is only applied to the leaf for a short time during which the stomatal aperture does not change, an accurate and representative can be measured (DUGAS et al., 1993). Measurement of on sample of leaves can then be scaled up using total leaf area and the climatic variables to calculate whole-plant transpiration (DUGAS et al., 1993). The accuracy of this whole-plant transpiration calculation depends upon leaf size, canopy aerodynamic conductance, and within-plant gradients of leaf area and vapour pressure (PEARCY et al., 1989). Inaccessibility of leaves and variation between leaves (LEVERENZ et al., 1982) and plants (HATTON and VERTESSY, 1989) may also affect the accuracy of -based transpiration estimates. Nevertheless as there is often no alternative, this method has been widely used for some species (e.g. SCHULZE et al., 1985, MUNRO, 1989). The most widely used instruments for measuring stomatal conductance are diffusion porometers. In principle, all diffusion porometers measure transpiration which is then used 51

to derive a value for the stomatal conductance (PEARCY et al., 1989). According to PEARCY et al. (1989), there are three main types of diffusion porometers for measuring of stomatal conductance and the transpiration – i) transient porometers, ii) constant flow porometers and iii) null balance porometers. The first type called transient porometers, primarily developed by WALLIHAN in 1964, is based on the diffusion of water vapour from a leaf to the humidity sensor inside the chamber. The principle of this type of instrument is that when a leaf is enclosed in a sealed chamber, transpiration will tend to increase the humidity in the chamber at a rate dependent, among other things, on the stomatal diffusion conductance. The time taken for the humidity to increase over a fixed interval can be converted to conductance by the use of a previously obtained calibration curve. This empirical calibration involves replacing the leaf by a wet surface (e.g. wet blotting paper) covered with a calibration plate or microporous membrane that has a known diffusion conductance. Calibration plates with a range of conductances are obtained by varying the number and size of precision drilled holes with the conductance for any plate being obtained from three-dimensional diffusion theory as being proportional to pore radius (JONES, 1992). This makes that the precision of the values obtained with this method is broadly questioned (REIGOSA ROGER and SÁNCHEZ-MOREIAS, 2003). In addition, the transient porometer has the inconvenience that for a good measurement the chamber must totally enclose the leaf. Therefore thin leaves (as those from conifers) cannot be precisely measured by this method. Transient porometers are cheap and easy to use, but its use is limited by the temperature dependence, the complicated calibration or the possible errors (REIGOSA ROGER and SÁNCHEZMOREIAS, 2003). The second type of porometer is that with a constant flow. PARKINSON and LEGG (1972) and later DAY (1977) developed a diffusion porometer that maintain a constant flow of dry air though the chamber where the leaf is placed. After a steady state, has been achieved, the resulting humidity is measured (PEARCY et al., 1989). This method is exposed to several differences between the environmental conditions outside the chamber and the microenvironmental conditions at which the leaf is exposed into the chamber. Furthermore, if a change in the vapour pressure occurs, the stomata can react in front of it and the actual plant status will be different from the starting plant status (REIGOSA ROGER and SÁNCHEZ-MOREIAS, 2003). Instrument based on a constant flow system able to measure CO2/H2O exchange was developed by SCHULZE et al. (1982). This type of porometer has inside an infrared analyser that allows the measurement of the differences in the CO2 decrease and H2O increase. These apparatus are interesting when simultaneous measurements of photosynthesis and transpiration are needed in the experiment, or when the environmental humidity is really high (the infrared gas analyser does not lose its accuracy in high relative humidity compared to the thin-film capacitance sensor commonly used in the two other types of porometer). In any case, they are more precise than the transient porometer but also more expensive and more difficult to use (REIGOSA ROGER and SÁNCHEZ-MOREIAS, 2003). The last type of diffusion porometer is so called null-balance porometer developed by BEARDSELL et al. (1972) based on a dry airflow through the chamber to balance the humidity generated into the chamber by a transpiring leaf reaching a steady-state humidity. 52

Because the humidity of the entering air is known (usually 0% or very close to 0% relative humidity) it can be shown that under isothermal conditions, determination of requires only a humidity measurement of the air within the chamber, the flow rate of dry air entering the chamber and the leaf area. In practice, however, conditions are not isothermal so that leaf temperature must also be measured (PEARCY et al., 1989). With this method it is necessary taking care when measuring the air and leaf temperatures, because differences can occur with the dry airflow. That is one of the advantages of this porometer, that it does not assume as equal the leaf and air temperatures, avoiding important errors in the measurement (REIGOSA ROGER and SÁNCHEZ-MOREIAS, 2003). As it happens with constant flow porometers, the conductance value is calculated in base of the mole fraction gradients of water vapour and the measured transpiration rates (PEARCY et al., 1989). As well in constant flow porometer as in null balance porometer errors in the humidity measurement can occur and it is possible to quantify them (CAMPBELL, 1975). By contrary, errors in transient porometer are more difficult to know. This system (null-balance porometer) is perhaps the most complex of the three presented here, but it is also the most flexible, able for field and laboratory measurements and with a quite easy calibration, which can be made at constant humidity values being also more precise (REIGOSA ROGER and SÁNCHEZ-MOREIAS, 2003). In any case, it is necessary to take in account that the extrapolation of the measurement obtained in these porometers to the whole plant transpiration is not straightforward because the impact of stomatal movements on transpiration is diminished by the diffusive resistance of boundary layers surroundings each leaf and the entire canopy (WULLSCHLEGER et al., 1998). These boundary layers allow transpired water vapour to humidify the air near leaves, uncoupling the vapour pressure at the leaf surface from that in the bulk air (JARVIS and MCNAUGHTON, 1986). The problem is less acute for needleleaved species, such as conifers, than for broad-leaved plants because of the high boundary layer conductances and close coupling to air temperature (PEARCY et al., 1989). Further, the user and the apparatus must be able to maintain undisturbed the leaf during the measurements, to avoid abrupt changes of temperature trying to take the measurement at shade, protect the plant against wind and not changing the external conditions during the measurement (PEARCY et al., 1989). Furthermore, the necessary number of replicates is high and moreover, they have to be taken in a short time period to avoid external environmental changes (REIGOSA ROGER and SÁNCHEZ-MOREIAS, 2003). Some scientists have developed models where the environmental conditions are directly implicated (KÜPPERS and SCHULZE, 1985). 2.3.3.2 Sap flow Transpiration rates for whole plants, individual branches or tillers can be determined by techniques which measure the rate at which sap ascends stems, i.e. sap flow. All of these methods are based on the evaluating of changes in quality or quantity of the employed tracer giving the information about the sap motion across the investigated volume of xylem tissue. Based on the early stage of the studies where the salt or dye were used as a indicator of sap velocity and especially following the pioneering work of HUBBER (1932), many different types of sap flow measurement methods have been 53

developed or described. Although, there is large variety of measurement principles, e.g. magneto-hydrodynamic (SHERIFF, 1972, SHERIFF, 1974), electric potential (MORAT et al., 1994, KOPPÁN et al., 2005, GIBERT et al., 2006), nuclear magnetic resonance (KÖCKENBERGER et al., 1997, ROKITTA et al., 1999, PEUKE et al., 2001, SCHEENEN et al., 2007, WINDT et al., 2007), there are two groups of methods which are the most established and widely used – the thermodynamics methods (HUBBER, 1932, ČERMÁK and DEML, 1974, SAKURATANI, 1981, GRANIER, 1985, NADEZHDINA and ČERMÁK, 1998) and deuterium tracing (CALDER et al., 1986, DYE et al., 1992). The thermodynamics techniques use heat as tracer of sap and have the advantage of being able to continuously track variations in flow at relatively precise locations in the xylem (MEINZER et al., 2006). Till now, there are overall five fundamental types of sap flow measurements using heat as a tracer (ČERMÁK et al., 2004). Chronologically, the first is heat pulse velocity method (HUBBER, 1932, HUBBER and SCHMIDT, 1936, MORIKAWA, 1972, COHEN et al., 1981, GREEN at al., 2003). The heat pulse velocity method measures the rate of sap flow by timing how long it takes short pulses of heat to travel over a known distance in the stem. The pulses of heat (typically one second duration, every 15 minutes) are provided by an electrically-powered line heater, in the form of metal probe, which is implanted radially into a small hole drilled in the stem. Two temperature sensing probes, one 10 mm above (downstream) and one 5 mm below (upstream) are also implanted, parallel with, and to the same depth as the heater probe. The temperature probes are used to monitor stem temperature after the heat pulse released, as it is propagated upwards by the sap stream. The lower (upstream) probe is needed to compensate for the fact that as well as being carried upward by the moving sap stream, the pulse also spreads out, both upstream and downstream, because of heat conduction in the static wood. Although the pulse spreads, it remains symmetrical. This means that when both temperature sensors reach the same temperature after release of the pulse, the peak (i.e. centre) of the pulse must be exactly halfway between them, i.e. (10 - 5)/2 = 2.5 mm downstream of the heater. In this way it is possible to determine the speed of the heat pulse, from the time between releasing the pulse and the measurement of zero temperature difference between the probes, divided by the distance of travel (2.5 mm). Each probest (one heater and two temperature sensors) measure heat pulse velocity at one depth in the stem. As the sap velocity tends to be higher near outside of the stem, and low or zero near the centre, velocity must be measured at several depths, so that the radial variation is properly sampled (HALL et al., 1996). The second method is called trunk segment (or tissue) heat balance. The original tissue heat balance method characterized by direct electric heating and internal sensing of temperature was originally designed for large trees (ČERMÁK and DEML, 1974, ČERMÁK et al., 1973, 1976, 1982, KUČERA, 1977, KUČERA et al., 1977). A section of a large trunk is heated from the inside by an electric current (supplied by electrodes) passing through the tissue. Heat is released more uniformly within the bulk xylem tissue and does not come through the thick bark (ČERMÁK et al., 2004). Both, power (which is proportional to sap flow) or temperature difference (indirectly proportional) can be held constant by electronic circuits, while the other variable recorded. In the recent commercially available technical arrangement, three electrodes, i.e. stainless steel plates (usually 25 mm wide and 1 mm 54

thick) are pounded at short distances (about 2 cm) into approximately the depth of sapwood (therefore they are provided in different length). They are inserted in parallel into sapwood keeping the central electrodes in a radial direction relative to the tree trunk. Another electrode of the same size is hammered 10 cm below to serve as a reference. All of the electrodes are equipped with a needle thermosensor. Via the set of three upper electrodes a variable heating power (up to 1.5 W depending on the sap velocity) is supplied into the sapwood and is automatically controlled in order to keep 1K difference between heated and reference (non-heated) electrodes in the measuring point (KUČERA et al., 1977). This method calculates the heat balance of a defined heated space. Basically, the input energy has to be split between the conductive heat losses and the warming of water passing through the measured segment, according to the following simple equation (KUČERA et al., 1977): (74) where is the heat input power (W), is the sap flow rate (kg s-1), is the temperature difference in the measuring point (K), is the specific heat of water (J kg-1 K-1) and the is the coefficient of the heat losses from the measuring point (W K-1). The amount of water in terms of mass or volume passing through the measuring point in the stem is calculated from the actual power and temperature rise of water passing through the heated space. The calculation of sap flow values derives from the Eq. 74 (KUČERA et al., 1977): (75) The first term of the Eq. 75 describes heat conducted by sap flow. The second term of this formula represents heat losses from the sensor. The heat loss considered in the applied equation is eliminated partially by the technical design of the measuring point (insulation by polyurethane foam and shielded against direct sun radiation), but nevertheless loss does take place. It changes mainly with the heat field patterns as it is changed with sap flow magnitude (ČERMÁK et al., 2004). The heat loss magnitude is clearly evident on the records of sap flow as a certain value of the so-called ―fictitious flow‖ ( ) which is recorded even when the actual flow is zero. When calculating the actual sap flow, it is necessary to subtract (estimated periodically when the actual flow approaches zero, e.g. early morning after prolonged rain) from the recorded flow data ( ): (76) However, it is difficult to distinguish between ―true‖ and plus a certain very low value of re-saturating flow, which almost always occur in early morning and which can cause minor errors and thus more days should be taken into account (ČERMÁK et al., 2004). Although, the described technical arrangement with inserted electrodes is applicable only for trees with more than 12 cm in diameter, there are yet commercially available socalled ―baby‖ sensors (applicable for diameters 0.6–2 cm), which are a significantly improved version of those based on flexible external heating and sensing as described by ČERMÁK et al. (1984) and LINDROTH et al. (1995). Heat is supplied by a resistance wire fastened to a soft resilient insulating tissue, assuring good contact even to non-cylindrical 55

stem surfaces, even if a plant grows in diameter up to about 50 % during the period of measurement. Temperature difference between heated and non-heated part of the measuring point (4 K) is kept constant and measured by thermo-sensors closely connected to xylem and controlled by variable power that is proportional to sap flow magnitude. Sap flow is calculated from the usual equation applied in the tissue heat balance technique, but excludes the term for section size, when flow is measured for the whole small stem in this case. Heat losses are considered periodically as the same way as described above. Close connection of thermocouples with the conductive xylem is important, because insulation properties of bark should be considered in order to get reasonable data (ČERMÁK et al., 2004). For example, MCNABB and HART (1962) found in elm branches differences in temperature above the heater reaching up to several degrees when bark was removed when compared with intact branches. STONE and SHIRAZI (1975) found that when applying surface heating through intact bark (2 mm thick) in small plants the sap velocity was lower by 25 % when compared with the situation when the bark was removed. The third sap flow technique is represented by stem heat balance method developed by SAKURATANI (1981) and BAKER and VAN BAVEL (1987) for intact plant stem sap flow measurement. It is based on the calculation of an energy budget between the energy put into the stem and the energy losses. The stem heat balance gauge comprises a partially flexible heater, a few centimetres in width, which is wrapped around the stem, and secured tightly in position with velcro straps. The walls of the tube are made of thermallyinsulating foam rubber. Around the inner surface, is an electrical heater, through which a constant amount of heat is applied to the surface of the stem. Mounted within the wall of the gauge are several temperature sensors, which measure the rate of heat loss from the gauge to the environment. From the heat balance, the difference between the applied heat and the measured losses is the amount of heat, (W) being dissipated by heating of the sap as it flows through the heated region. The gauge also measures the temperature increase of the sap, (K), so the rate of sap flow can be simply calculated as (HALL et al., 1996): (77) The fourth method based on the heated probe technology called heat dissipation method was developed by VIEWEGH and ZIEGLER (1960), independently also by ITTNER (1968), BALEK and PAVLIK (1977) and later better quantified by GRANIER (1985). It is based on the detection of convective heat transport (heat carried with the sap stream). Usually two thermocouples are mounted in thin needles, one on each. The needles are inserted radially in the sapwood above each other about 10 cm apart. The upper sensor is heated with constant power and the sap flow velocity is calculated from the temperature difference ( ) between the two needles (FERNANDÉZ et al., 2011). This background information goes into the equation using an empirical coefficients derived originally by GRANIER (1985) commonly applied for all species, according which so-called sap flux density is calculated. Since the sap flux density (kg m-2 s-1) express practically the sap flow per sapwood area (m2), it is has to be multiplied by the area of cross-section of the sapwood at the level of the heated probe ( ): 56

(78) The maximum temperature difference ( ) occurs under no flow conditions, and decreases with increasing sap flow velocity. Since the system measures flow along a short part of sapwood depth, only more such sensors installed in different depths (but always within the conducting xylem) can provide values valid for the whole sapwood depth (LU et al., 2004, FERNANDÉZ et al., 2011). The last from the five thermodynamics sap flow techniques is called heat field deformation method. This method, developed by NADEZHDINA and ČERMÁK (1998), is based on the analysis of temperature differences around a linear heater inserted in the sapwood. These temperature differences characterise the deformation of the heat field around the heater caused by the ascent of sap. The heat field deformation method can measure wide size of trees in a wide scale of flow rates including low, zero and reverse flows. The heat field deformation sensors consists of a linear heater and two pairs of differential thermocouples (symmetrical and asymmetrical) measuring the temperature difference in axial and tangential directions around the heater (NADEZHDINA et al., 2002) as raw data. Sap flow is then calculated from the mentioned temperature differences. The multi-point sensor has several thermocouples along each needle and allows measurements of sap flow radial profile. The temperature difference in axial is also known as the sap flow index which can be used as a stress indicator (NADEZHDINA, 1999). Except for techniques based on thermodynamic methods, chemical tracers (GREENIDGE, 1955), radioisotopes like tritium (KLINE et al., 1970, WARING and ROBERTS, 1979) and stable isotopes like deuterium (CALDER et al, 1986, DYE et al., 1992) have been used to determine the sap flow rates. However, the chemical tracing needs irreversible treatments of plants like cuttings and the use of radio-tracers like tritium and 32P are increasingly subject to regulatory constraints and seasonal patterns of water use are difficult to obtain with this method. The use of deuterium overcomes many of these limitations and stable isotopes have largely replaced tritium as a tracer-based technique for quantifying rates of water use (WULLSCHLEGER et al., 1998). This technique involves the injection of a known mass of deuterium oxide into the stem (CALDER et al., 1986, CALDER, 1991, CALDER et al., 1992). The condensate water is then collected from a number of plastic bags attached over the leaves in the tree canopy. This procedure is preferred to the leaf sampling method as it provides time averaging of the concentration rather than instantaneous values. The periodically sampled transpired water and its actual deuterium concentrations are used to construct a curve of tracer concentration in transpired water as a function of time. Assuming that the tracer passed through the plant over the course of the experiment, the plant water use is inversely proportional to the area under the curve of tracer concentration (CALDER, 1991). The total mass of tracer injected, (g), is related to the sap flow by the means conservation principle: (79) where is the sap flow (kg day-1), is the deuterium concentration of the transpired water (g l-1) and and are the first and the last time increment, respectively, in which tracer is 57

present (days) (MARC and ROBINSON, 2003). For practical considerations the equation is rewritten in the finite difference form as follows: (80) where is the concentration in the ith time increment which is measured from the collected condensate, is the duration of the ith time increment and is the last time increment in which the deuterium was present (CALDER, 1991). Besides estimates of individual plant water use, tracing methods can also provide information on absolute sap velocity, internal water storage and retention of water in plants (JAMES et al., 2003, MEINZER et al., 2006). Unlike the thermodynamic methods, the deuterium tracing method neither depends on power source nor does it require specialized equipment in the field, which makes its application at remote sites and for extensive spatial sampling designs feasible (SCHWENDENMANN et al., 2010). In contrast to thermodynamic method, water use cannot be characterized on a time scale smaller than one day or even several days using the tracing method (SMITH and ALLEN, 1996). Further, sampling is labour intensive, especially for taller trees. Still, for specific questions or under certain circumstances, the deuterium tracing method might provide suitable alternative to the thermodynamic method (SCHWENDENMANN et al., 2010). 2.3.3.3 Chamber method Chambers are used to measure directly the flux of gases between earth´s surface and the atmosphere by enclosing a volume and measuring all flux into and out of the volume (DENMEAD et al., 1993). DENMEAD (1984) classified gas exchange canopy chambers into closed (transient-state) and semi-closed (steady-state) systems. Transient-state closed chambers are closed for a short period of time (minutes) while the change in gas concentration, usually by infra-red gas analyser, is measured. Since they are portable, they can be placed on the plant for measurement and can be removed before the plant can respond physiologically to the changing environment (PÉREZ-PRIEGO, et al. 2010). The chambers to rapidly measure water loss by evapotranspiration were described for the first time by REICOSKY and PETERS (1977). They consisted of aluminium conduits covered with polycarbonate or glass sheets. The air within the chamber was mixed continuously with strategically located fans. The first version was portable (by means of a tractor, for example) and the evapotranspiration rates were calculated by a psychrometer before the chamber was lowered on the plot and one minute later, as latent heat storage. The volume of the chamber can be easily adapted to a herbaceous field crop and the accuracy was ~10 %, compared with weighing lysimeters (REICOSKY et al., 1983). The semi-closed (open or open-top) system continuously renews the air inside the chamber while measuring the gas concentration in the entry and exit airstreams. This type of chamber is also suitable for measuring larger plant like trees (CIENCIALA and LINDROTH, 1994, KATERJI et al., 1994, PÉREZ-PRIEGO, et al. 2010) and enables to measure canopy gas exchange with high resolution in time over days or weeks including the assessment of short-term impacts of environmental changes on photosynthesis (BURKART et al., 2007). Since these chambers are always closed (except for the part which is opened – e.g. the top) 58

and under constant forced ventilation, the radiative and aerodynamic microclimate around the tree can be very different from the natural undisturbed conditions (PÉREZ-PRIEGO, et al. 2010). In the most of open chamber systems with air supply, there is a constant overpressure within the chamber to keep the structure inflated. Furthermore, steady airflow is often higher than natural wind speed and suitable ventilation rate is difficult to get (PÉREZ-PRIEGO, et al. 2010). The critical aspect common for both type of chambers is the radiation environment in the chamber, by reducing direct beam solar radiation and increasing the diffuse component (LEUNING AND FORSTER, 1990). The most serious problems of almost all chambers are related to the modification of the microclimate during the measurement period (RANA and KATERJI, 2000). Most of the chambers were designed for simultaneous measurement of both CO2 and water vapour exchanges (DENMEAD, 1984). These dual requirements are not often compatible because wall materials are usually selected for high transmission in the short wavelengths, neglecting the long-wave exchange (which can usually represent ~20 % of the incoming short-wave radiation. As a consequence of the modified radiation balance, the air temperature and vapour pressure deficit rapidly increases and thus the conditions inside the chambers could alter the biological control of leaves transpiration process. In case of the semi-closed chambers which are used for longer measurement periods, the modification of the conditions may naturally alter the growth process. As the result, the biometric features (leaf area, tree height, stem thickness etc.) can be different compared to the surrounding plants and the measurements lose their representativeness. In addition, CIENCIALA and LINDROTH (1994) suggested that trees protected against wind in the stable chamber frame have decreased demand for lignifications and thus can grow faster. Therefore, using of the semi-open chambers for periods longer than approximately one week is not recommended if the aim is to maintain the biometry of an enclosed tree similar to the other specimens (CIENCIALA and LINDROTH, 1994). Finally, in case of the portable chambers, the wind speed inside the chamber is strongly reduced which has a direct effect on the measurement accuracy (RANA and KATERJI, 2000). Further, more serious errors may be introduced if chamber results were extrapolated over time as well as spatially (LIVINGSTON and HUTCHINSON, 1994). Under a semi-arid climate, DUGAS et al. (1991) demonstrated that portable chambers, equipped with an infrared analyser for measuring vapour density differences in differential mode, are not representative of the field, particularly if placed on its edges (where the portable chambers are usually positioned due to their good accessibility), where evapotranspiration measurements are affected by surrounding advective conditions. Furthermore, the last improved version of portable chambers can be very expensive, due to the high costs for key components (data acquisition system and infra-red gas analyser) and accurate chamber manufacturing that may require some heavy engineering. Recent developments of portable chamber systems can be found in WAGNER and REICOSKY (1996); here the equipment also provides CO2 measurements, but its complexity and costs limit its use for research to a small temporal scale (RANA and KATERJI, 2000).

59

2.4 Estimating water loss Although, the measurement of evapotranspiration can brings the most relevant results describing the actual water loss from the area of interest, it is not usually used for routine applications in hydrology, agriculture, forestry, and the related fields. The main reasons are especially the need for expensive equipment, knowledge in micrometeorology and measuring technique, continuous scientific and technical support and also time consuming procedure to obtain reliable data. Therefore, the estimation methods based on climatic data represent common practice and the direct or indirect measurement methods rather the tool for research. All around the world, there are a lot of methods for estimating the evapotranspiration which differ by the input variables, time resolution and their space universality. Generally, we can categorise the evapotranspiration estimates into three groups of methods: empirical (or semi-empirical) methods, analytical approach, and soil water balance modelling. 2.4.1 Empirical evapotranspiration models This group of methods is based on empirical relation between the evapotranspiration and at least one or more basic meteorological variables and alternatively geographic position, day of year etc. Depending on the measured meteorological variable, we can differentiate between radiation based methods (MAKKINK, 1957, TURC, 1961, JENSEN and HAISE, 1963), temperature based methods (THORNTHWAITE, 1948, BLANEY-CRIDDLE, 1962, HARGREAVES and SAMANI, 1985), and humidity and wind speed based methods (DALTON, 1802, ROHWER, 1931, MARCIANO and HARBECK, 1954, HARBECK, 1966). Most of these methods describe potential evaporation from open water or reference grass surface and demand specific calibration. Despite the rich list of the empirical methods, there are two which are the most wide-spread and universal for determining the potential evaporation from open water or other surfaces with no lack of water. They are chronologically Penman method (1948) and Priestley-Taylor method (1972). 2.4.1.1 Penman approach PENMAN (1948) was the first who combined the Bowen ratio and the Dalton approach, i.e. an energy balance and an aerodynamic term to estimate the potential evaporation of open water. His physically-based derivation, recently well-known as Penman combination equation, should be still termed as a semi-empirical approach because the dependence on the wind velocity was solved empirically. The Penman method, applied to open water, can be briefly described by the energy balance at the earth´s surface (Eq. 4), which equates all incoming and outgoing energy fluxes (note that mean heat flux density into the water body in this case). Supposing that and can be measured, one can calculate if the Bowen ratio is known. This ratio can be derived from the transport equation of heat and water vapour in air. By applying the Eq. 4 for open water surface we can derive following equation for latent heat flux: (81)

60

and analogically for the turbulent heat transfer: (82) The new symbols here are as follows: is the saturation vapour pressure at the evaporating water surface and the vapour pressure (kPa) measured at the height z (m) and and are the surface temperature and the above air temperature (°C) at the same height as respectively. The aerodynamic diffusion resistance (s m-1) is the reciprocal turbulent transfer coefficient integrated between height and the surface. Its value is a function of wind and informs us about what is the air resistance for heat or molecules of water vapour to transfer between the surface and the level . Since the Bowen ratio similarity for and analogically for does exist, the Bowen ratio above open water can be described similarly as the previously mentioned air gradient form (Eq. 42) as follows: (83) The problem with the above equation is that the surface temperature is not generally known (not routinely measured). PENMAN (1948) therefore took three intermediate steps. Firstly, he introduced the proportionality constant: (84) where is the saturation vapour pressure (kPa) again at the level . The proportionality constant (kPa/°C) is the first derivation of the function versus known as the saturation vapour pressure curve. Substituting Eq. 84 into Eq. 83 yields: (85) Secondly he replaced the vapour pressure gradient

in Eq. 85 with: (86)

which gives: (87) Thirdly, under isothermal conditions (i.e. if no heat is added to or removed from the system), he assumed that . This implies that . If we introduce this assumption into Eq. 81, the theoretical adiabatic latent heat flux equals: (88) Dividing the Eq. 88 by Eq. 81 yields: (89) so that: 61

(90) Substituting the above information into Eq. 4, and writing (subscript o denoting open water) for the potential evaporation yields the Penman formula in the following form: (91) is and adiabatic evaporation (MJ day-1) and is solved as an empirical wind function. According to Penman original derivation (PENMAN, 1948), the adiabatic evaporation is as follows: (92) In fact, the adiabatic term in the Penman formula is the only empirical part which is just linear regression adjust taking into account the differences between estimated and observed . Apart from the validity for open water, PENMAN (1953) found (as he wrote: ―by a happy accident‖) that this empirical wind function is valid also for short grass. Maybe from that reason, people adopted to use this formula for potential evapotranspiration of grass. Nevertheless, such grass would have to have zero surface resistance which would be possible only in the case of wet surface. Penman originally derived the potential evapotranspiration of grass for England conditions as a proportion of open water evaporation. Namely, it is 0.8 of during the period May to August, 0.75 during March, April, September and October and 0.6 for the rest of the year (PENMAN, 1948, PENMAN, 1956). Note that the Eq. 91 is derived for daily time step where the radiation term and the adiabatic terms are in MJ day-1 (originally in mm of water equivalent assuming all available energy is used for evaporation). While one can use daily averages, the use of 1–60 minutes averages is considerably more meaningful (FOKEN, 2008a). However, in this case the units must be converted. 2.4.1.2 Priestley-Taylor approach It has been thought theoretically that air passing over a very large, homogenous, moist surface (but not necessarily water) will gradually decrease in saturation water vapour deficit until equalize with that of the surface and thus the ―equilibrium‖ evaporation ( ) is reached (SLATYER and MCILROY, 1961, MCNAUGHTON, 1976, MONTEITH, 1981). Consequently, the derivation of can be solved straightforwardly from the Penman open water Eq. 91 assuming that the saturation vapour pressure deficit is equal to zero and as a consequence, the adiabatic part has no contribution to the evaporation in the absence of advection: (93) Note that the conditions of equilibrium evaporation are rarely found in nature. Even over the tropical Pacific Ocean, with 100 km of fetch, a saturated atmosphere is never 62

reached. In fact, the atmosphere is not in an ―equilibrium‖ when the Eq. 93 is valid, since it implies an artificially closed top atmospheric boundary layer (EICHINGER et al., 1996). The measurement reviewed by PRIESTLEY and TAYLOR (1972) convinced them that, on average, the observed latent heat of evaporation from water or from well-watered short vegetation exceed by a factor of about 1.26. Based on these results they proposed empirical relation for potential evaporation (or evapotranspiration) known as the PriestleyTaylor equation which demands less input data compared to the previous Penman combination equation: (94) Since the publication of their paper in 1972, there has been large number of papers published which report measurements of evaporation from wet or well-watered surfaces that are consistent with this value of 1.26 (DAVIES and ALLEN, 1973, STEWART and ROUSE, 1976, 1977, MUKAMMAL and MEUMANN, 1977, PARLANGE and KATUL, 1992). Although there are some reports of deviations from the value of 1.26 (MCNAUGHTON and BLACK, 1973, SHUTTLEWORTH and CALDER, 1979) and some data to indicate that there may be systematic variations in the value of with the time of day (GUO et al., 2009) and the season of the year (DE BRUIN and KEIJMAN, 1979) there has been no explanation of why the value should be essentially constant or under what conditions differences from this value of might be expected (EICHINGER et al., 1996). It has been ascribed to the entrainment of relatively warm, dry air downwards through the top of the convective boundary layer, defined as the turbulent surface layer which develops during daytime, due to the input of sensible heat from the ground. But it is still not clear why the entrained energy should be a conservative fraction of the available energy at the grand (MONTEITH, 1981, 1985). 2.4.2 Penman-Monteith analytical approach Although the PENMAN (1948) in his original work investigated also evaporation from turf grass and bare soil, he solved them empirically as a fraction of the open water evaporation with respect to the seasonal period. In his further work, he also attempted to derive evapotranspiration from vegetation (PENMAN, 1953), but the fully analytically based and universal approach was given later by MONTEITH (1965) described in PenmanMonteith equation. Firstly, in this more advanced approach, the effect of physiological control of transpiration by stomata was expressed and secondly, the aerodynamic term was elaborated on the physical basis of the theory of turbulent diffusion. MONTEITH (1965) noted, that the Penman equation is not valid for a surface where the vapour pressure is less than the saturation vapour pressure at surface temperature. Further, he showed that a leaf with internal resistance can be treated formally as a free water surface if the vapour pressure difference is replaced by following expression: (95)

63

Alternatively, the psychrometric constant obtaining the same effect (MONTEITH, 1965):

can be replaced by a modified value

(96) After that, the Penman-Monteith combination equation can be written in the final form (MONTEITH, 1965): (97) This Penman-Monteith equation is valid for a dry crop completely shading the ground. Note, that for a wet crop covered with a thin water layer caused by rain or dew fall to zero and the wet-crop formulation is obtained (MONTEITH, 1965): (98) This wet-crop form is almost identical with the Penman combination equation (PENMAN, 1948), only the empirically derived wind function is replaced with a physically based aerodynamic resistance. If we assume, that also the saturation vapour pressure is equal to zero, the equilibrium evaporation, which is the core of the Priestley-Taylor equation, will be obtained again. Eq. 97 is, in principle, not able to describe the evapotranspiration from sparsely-cropped surfaces. With a sparsely-cropped surface, the evaporation from the soil might become dominant (FEDDES and LENSELINK, 1994). It appears that the canopy resistance, of a dry crop completely covering the grounds has non-zero minimum value if the water supply in the root zone is optimal. For arable crops, this minimum amounts to ; that of a forest is about 150 (FEDDES and LENSELINK, 1994). The canopy resistance is a complex function of incoming solar radiation, water vapour deficit, and soil moisture. The relationship between and these environmental quantities varies from species to species and also depends on soil type. One of the most diffuse methods of estimating is that proposed by SZEICZ and LONG (1969), in which canopy resistance can be determined as a function of daily mean stomatal resistance of the single leaves ( and of the leaves active (effective) in transpiration ( ). These authors assumed that only the surfacial layer of the canopy participates effectively to the transpiration. Usually, such a canopy resistance is assumed to be half the full crop (RANA and KATERJI, 2000): (99) Since the spatial and temporal variability of the soil water status (and consequently, of the plant water status) is very high and microclimate to which they are exposed is very different (DENMEAD, 1984), to obtain sufficiently accurate canopy resistance values by means of this method, it is necessary to realise a great number of measurements of stomatal resistance in a brief interval time (RANA and KATERJI, 2000). Furthermore, from the 64

theoretical and practical point of view, the problems of scaling-up the canopy resistance from leaf to canopy is not completely solved (JARVIS and MCNAUGHTON, 1986, BALDOCCHI, 1989, BALDOCCHI et al., 1991, RAUPACH, 1995, STEDUTO et al., 1997). An improvement of this estimating method has been proposed by MONTEITH (1965) and tested and applied by KATERJI and PERRIER (1985). In this method, the canopy was divided into layers and canopy resistance was estimated starting from stomatal resistance measured layer by layer and weighted with the of each layer. The expression of the canopy conductance is given, in this case, by relationship: (100) where

is the stomatal conductance measured on the leaves belonging to the layer

characterised by . This approach is difficult and hard to carry out, giving values not always accurate (RANA and KATERJI, 2000). Nevertheless, the estimation of field crop evapotranspiration can be acceptable, at least for well-watered crop (KATERJI and PERRIER, 1985). Since it is very difficult to measure directly, it is usually determined experimentally with the use of the Penman-Monteith equation, where is measured independently (e.g. by the soil water balance or micrometeorological methods). The problem is that, with this approach, the aerodynamic resistance, , has to be known. Owing to the crude description of the vegetation layer, this quantity is poorly defined. This implies that some of the values published in literature are biased because of errors made in (DE BRUIN, 1982). Other possibility is to use of the surface resistance model. Generally, the most complete model of should have the following general form (STEWART, 1988, 1989, ITIER, 1996): (101) Actually, the better way to evaluate the crop water status is by means of measurements of leaf water potential or root abscisic acid (ABA) (ZHANG et al., 1987, ZHANG and DAVIES, 1989). Nevertheless, the plant water status varies almost instantaneously, so that to have acceptable accuracy of the actual crop water status, a great number of leaf water potential or ABA measurements has to be carried out during the entire day. Hence, this method is not practical for application in a field (RANA and KATERJI, 2000). In order to avoid the problem of large number of punctual measurements, a number of simplified models based on the general relationship (Eq. 101) have been developed. For example LOHAMMAR et al. (1980) and later LINDROTH (1985) and HALLDIN (1989) modelled as a function of global radiation and vapour pressure deficit. HATFIELD (1985) proposed model based on global radiation and available soil water as inputs, or FUCHS et al. (1987) alternatively used photosynthetic active radiation ( ) and soil moisture deficit. JARVIS (1976) used regressive experimental functions to model as a result of interactions between , global radiation, , and soil water potential. Unfortunately, all these functions need to species, local calibration and even temporal calibration, i.e. they are not valid for the whole growth season (STEWART, 1988). An original approach to take into account the crop water status to estimate was presented by RANA et al. (1997). In their 65

model, is, beside available energy, vapour pressure deficit and aerodynamic resistance, function of ―predawn leaf water potential‖ which does not need to be measured instantaneously, but just once a day. Such model is valid also for crop under water stress and needs only species specific calibration so it can be generalized (RANA et al., 1997). It is possible to use the available soil water as input instead of predawn water potential, however, thus the model loses its generality and must be locally calibrated. The second resistance term in Penman-Monteith equation, aerodynamic resistance has been already expressed through K-theory (flux-gradient relationship) in Eq. 9. More usual form is derived from Eq. 10 for momentum transfer resulting in expression: (102) Assuming the logarithmic wind profile under neutral stability conditions, the equation can be rewritten into the following more usual forms (MONTEITH, 1965, THOM, 1975, MONTEITH and UNSWORTH, 2008): (103) However, in real conditions wind is blowing over vegetation which never represents flat plane and thus the surface is not in parallel to the airstream which means that the drag is not purely generated by ―skin friction‖ effect of air molecules moving along the surface. For extreme situations, the surface is at right angles to the airstream and there is no frictional force in the direction of flow but normal pressure or ―bluff body‖ drag forces are dominant. In general, however, with leaves at intermediate angles of incidence, momentum transfer is accomplished by a combination of bluff body forces and molecular skin friction drag, roughly in the ratio of three parts to one (THOM, 1968, 1975). As a consequence, the roughness length for momentum transfer, , will include the effect of bluff body forces, whereas heat and vapour pressure transfer will not have similar analogy. As a result the apparent roughness for heat and vapour transfer ( will generally be less than that for momentum (BRUTSAERT, 1982) leading to rewrite the Eq. 103 according to GARRATT and HICKS (1973), and BRUTSAERT and STRICKER (1979) as:

(104) where is height of the wind speed measurement, and height of the air temperature and humidity measurement above the ground surface (m). THOM (1972, 1975), THOM and OLIVIER (1977), and CAMPBELL (1977) suggested estimating the value of as 0.2 , whereas measurement by CHAMBERLAIN (1966), MONTEITH (1973), and BRUTSAERT (1979, 1982) indicated that the value of can be approximated as 0.1 . 2.4.2.1 FAO reference and crop evapotranspiration Since the evapotranspiration constitute the key issue of agriculture water management, it is not surprising that panel of expert from FAO (Food and Agriculture Organization) put years of effort to elaborate methodology for evapotranspiration estimation – firstly FAO 66

Irrigation and Drainage paper No. 24 (DOORENBOS and PRUITT, 1977) and more recently FAO Irrigation and Drainage paper No. 56 (ALLEN et al., 1998). The core of this methodologies is a calculated evapotranspiration of the reference surface which represents nearly all effect of weather, and the empirically obtained relations between the reference evapotranspiration ( – zero denotes ―reference‖, do not change it with for open water potential evaporation) and the potential crop evapotranspiration ( ) known as the crop coefficient predominately varying with specific crop characteristic and only to small extend with climate. In the past, an open water surface has been proposed as a reference surface. However, the differences in aerodynamic, vegetation control and radiation characteristics present a strong challenge in relating to measurements of free water evaporation. Relating to a specific crop has the advantage of incorporating the biological and physical processes involved in from cropped surfaces. In the previous FAO Irrigation and Drainage Paper No. 24 ―Crop water requirements‖, four methods were presented to calculate the reference crop evapotranspiration according to data availability of users: the Blaney-Criddle, radiation, pan evaporation, and modified Penman method. The last mentioned, Penman method, was the most advanced alternative, which was fundamentally the original Penman equation (PENMAN, 1948) with the modified wind function taking into account different roughness of the reference crop and the part of the season (DOORENBOS and PRUITT, 1977, ALLEN et al., 1989). As a reference crop, two possibilities could be chosen. According to JENSEN et al. (1970), alfalfa with 30 to 50 cm in height, and later according to DOORENBOS and PRUITT (1977), grass from 8 to 15 cm high, both with optimal water availability, were provided. Since the modified Penman method was frequently found to overestimate , even by up to 20 % for low evaporative conditions, it usually demanded local calibration of the wind function and was losing the expected generality. After reviewing the FAO methodology in 1990, panel of experts recommended the adoption of the Penman-Monteith combination methods as a new standard for reference evapotranspiration and advised on procedures for calculating the various parameters. As a result, new FAO Irrigation and Drainage Paper No. 56 was elaborated by ALLEN et al. (1998) which has achieved world-wide respected universal validity. According to ALLEN et al. (1998), following unambiguous definition for the reference surface was proposed: “A hypothetical reference crop with an assumed crop height of 0.12 m, a fixed surface resistance of 70 s m-1 and an albedo of 0.23‖. This reference surface closely resembles and extensive surface of green grass of uniform height, actively growing, completely shading the ground and with adequate water. The requirements that the grass surface should be extensive and uniform result from the assumption that all fluxes are one-dimensional upwards. Following this straightforward definition, the aerodynamic resistance of the reference crop with respect to its height can be derived from Eq.104 as a linear function of the wind speed: (105) Further, assuming the constant height according to simple relation:

67

0.12 m, the constant

is expected

(106) where is the crop height (m). Furthermore, two more simplification are taken into account. Firstly, that only a half of is active, and secondly that stomatal resistance of -1 single leaf has value of about 100 s m under well watered conditions. Finally, applying these assumptions, the Eq. 106 can be used obtaining fixed surface resistance of 70 s m-1 as an average during the whole day. Implementing these two resistances derived for reference crop into the Eq. 97, the FAO Penman-Monteith method to estimate in diurnal step can be written in the following rearranged final form (ALLEN, 1998): (107) The application of the equation for short time period (hourly or less) may require the inclusion of correction for stability in aerodynamic resistance. Nevertheless, when predicting in the well-watered reference surface, heat exchanged is small, and therefore stability correction is normally not required. However, in shorter time step, one should take into account that the constant surface resistance does not provide real stomatal behaviour. In this case, the constant resistance should be replaced by variable value, which is usually modelled. One of the well established models for deriving variable surface resistance of the reference crop was proposed by TODOROVIC (1999) and has the advantage that does not need any specific local and temporal calibration and even more any other inputs than those demanded for standard calculation of . The second point of the FAO Penman-Monteith methodology is the so called crop evapotranspiration and the related crop coefficients. Fundamentally, the crop coefficient is defined as the ratio of evapotranspiration from any specific crop or soil surface to some reference evapotranspiration defined by weather data and the properties of reference surface: (108) In FAO-56, values listed for represent evapotranspiration under growing conditions having a high level of management and with little or no water or other evapotranspiration reducing stresses and thus represent what are referred to as potential levels for crop evapotranspiration (ALLEN et al., 2005): (109) Actual can be less than the potential for a crop under non-ideal growing conditions including those having water stress or high soil salinity. According to ALLEN et al. (2005), representing evapotranspiration under any condition, ideal or non-ideal, is termed ―actual ‖ and is denoted as . In FAO-56, the was originally termed ―adjusted ( ) (ALLEN et al., 1998). The terms are synonymous and: (110)

68

where is ―actual‖ crop coefficient that includes any effects of environmental stresses (ALLEN et al., 2005). 2.4.3 Soil water balance modelling An exhaustive survey of the most widely spread water balance models or their general characterization and classification can be found in LASCANO (1991) de JONG and BOOTSMA (1995), LEENHARDT et al. (1995) and ZHANK et al. (2002). According to RANA and KATERJI (2000), two classes of models are generally retained for the simulation of soil water balance: (i) mechanistic models (complex) and (ii) analogue (water reservoirs or bucket) models. In the mechanistic approach, the water flux in the soil is controlled by the existence of soil water potential gradients, by means of Darcy‘s law and continuity principle. The equations are usually solved by different methods, all involving the splitting of the soil in more or less small layers (de JONG, 1981, FEDDES et al., 1988, PERSSON and JANSSON, 1989, PERSSON and LINDROTH, 1994, RODRIGUEZ-ITURBE et al., 2001, PORPORATO et al., 2003, DECKMYN et al., 2008). The difficulties in the application of these models are linked to the accuracy of the used pedo-transfer functions for estimating the water transfer and also the procedures used for estimating the boundary conditions of the soil-plantatmosphere system (RANA and KATERJI, 2000). In the second, analogue approach, the soil is treated as a collection of water reservoirs, filled by rainfall or irrigation and emptied by evapotranspiration and drainage. In general, they can be based on the two following principles (LHOMME and KATERJI, 1991). Firstly, determination of soil water storage is a function of the soil and roots depth (e.g. BRISSON et al., 1992, IRITZ et al., 2001). Secondly, the soil water is split into readily available soil water and total soil water (RITCHIE, 1972, ALLEN et al., 1998, PETZOLD et al., 2010, HLAVINKA et al., 2011). The stress coefficient is then supposed to be ~1 within the range of the readily available water which is the upper part of total available water defined by the upper and lower limits – field capacity and the wilting point respectively. The portion of the readily available water is given by soil specific properties – especially clay/loam/sand ratio, organic matter content. By subsequent drying of the soil profile, the soil moisture level decreases below so called stress (refill) point when the starts to decrease proportionally to the water content in the total available reservoir and the evapotranspiration becomes limited.

69

3 AIMS As it was outlined in the introduction, the admirable vigorous growth of short rotation coppice (SRC) based on poplar or willow species is usually associated with high water consumption. Therefore, it is assumed that large scale plantation management can reduce the water table and thus may have negative influence on water resources of the whole catchment. Further, for the rain-fed areas, the annual course and amounts of precipitation can strongly influence the final yield and thus determine the economic profitability. To contribute to the still open discussion on the water relations of this eventual bio-energy source, the following goals (according to their priority) will be fulfilled with using of results from the own field measurements:  To quantify the actual evapotranspiration ( ) of SRC based on poplar clone (Populus nigra x P. maximowiczii) using the Bowen ratio energy balance (BREB) method in the conditions of the Czech-Moravian Highlands.  To compare the of the poplar plantation with the in the same pedo-climatic conditions at the same site.

of turf grass growing

 To analyze other additional hydrological, micrometeorological and ecophysiological variables (especially soil moisture patterns, crop coefficient, Bowen ratio, surface resistance, decoupling factor, sap flow, leaf area index, and aboveground woody biomass increment) of SRC.  To investigate the relation between and the biomass growth and its possible dependency on the main meteorological factors and its seasonal variation.  To ensure original high quality reliable database for future creating of simple empirical models predicting the water balance and biomass growth, and also for parameterization of more complex semi-mechanistic or mechanistic models providing deep view into the whole ecosystem processes.  To carry out the error analysis of the BREB method and to analyze the limitations in the given fetch conditions.  To provide a literature overview of the fundamental methods applicable in the research of the water balance of SRC or other energy crops.

70

4 MATERIALS AND METHODS 4.1 Locality The research locality Domanínek (Czech Republic, 49° 32´ N, 16° 14´ E and altitude 530 m a.s.l.) is situated in the west part of the city Bystřice nad Pernštejnem (~9,000 inhabitants) in the Czech-Moravian Highlands. The city lies ~180 km on the approximate south-east from the capital Prague and ~60 km north-north-west from Brno, the second largest city in the Czech Republic. The surrounding landscape is typical with the deep valley of the river Svratka and the foot of the Ţďár Peaks. From the agricultural point of view, the area belongs rather among the marginal areas of the potato growing region.

Figure 2: The geographical position of the research locality Domanínek within the Czech Republic.

4.1.1 Poplar plantation In April 2002 a high-density experimental field plantation for verification of the performance of poplar clone J-105 (Populus nigra x P. Maximowiczii Maxfünf, country of origin – Austria) with the total area of 2.85 ha was established in Domanínek. The plantation is a part of the city bio-energy program initiated in 1998 when the biomass district heating plant (9MW capacity) was built. The plantation was established on agricultural land previously cropped predominantly with cereals and potatoes. No irrigation, fertilization and herbicide treatments (except the local application of Roundup on the most vigorous and tenacious weeds) were applied during the whole experiment. Hardwood cuttings were planted in a double row design with inter-row distances of 2.5 m and spacing of 0.7 m within rows accommodating a theoretical density of 9,216 trees ha-1. Subsequent mechanical weeding was carried out two times per growing season till the canopy closure in 2005. The most problematic weed species were Cirsium arvense, Rumex crispus, Rumex obtusifolius, Artemisa vulgaris, Tanacetum vulgare, Elytrigia repens and Arrhenatherum elatius.

71

4.1.2 Climate conditions The site itself is situated on a mild slope of 3° with an eastern aspect and is generally subject to cool and relatively wet temperate climate typical for this part of Central Europe with mingling continental and maritime influences (Tab. 1). The site climate is suitable for SRC based on Populus species clones according to HAVLÍČKOVÁ et al. (2006) as the area belongs to the climate region no. 7. Weather parameters were obtained from the meteorological station Bystřice nad Pernštejnem (Czech Republic, 49° 32´ N, 16° 16´ E and altitude 560 m a.s.l.) of the Czech Hydrometeorological Institute, which is less than 1 km from the experimental plantation. The first 8 years long rotation period was characterized by favourable conditions during the year of establishment but with higher incidence of droughts in the following two years (Fig 3). very warm - exceptionaly warm very cold - exceptionaly cold normal 1961-1990 mean

30

Temperature ( C)

25 20 15

10 5 0

Jul-09

Oct-09

Jan-10

Apr-10

Jul-10

Oct-10

Jan-10

Apr-10

Jul-10

Oct-10

Apr-09 Apr-09

Oct-09

Jan-09 Jan-09

Jul-09

Jul-08

Oct-08

Apr-08 Apr-08

Oct-08

Jan-08 Jan-08

Jul-08

Jul-07

Oct-07

Oct-07

Apr-07 Apr-07

Jul-07

Jan-07 Jan-07

Jul-06

Oct-06

Apr-06

Jan-06

Jul-05

Oct-05

Apr-05

Jan-05

Jul-04

Oct-04

Apr-04

Jan-04

Jul-03

Oct-03

Apr-03

Jan-03

Jul-02

Oct-02

Apr-02

-10

Jan-02

-5

very dry - exceptionaly dry very wet - exceptionaly wet normal 1961-1990 mean

160 120 80

Oct-06

Jul-06

Apr-06

Jan-06

Jul-05

Oct-05

Apr-05

Oct-04

Jan-05

Jul-04

Apr-04

Jan-04

Oct-03

Jul-03

Apr-03

Jan-03

Oct-02

Jul-02

0

Apr-02

40

Jan-02

Precipitation (mm)

200

Figure 3: Dynamics of mean monthly temperatures (°C) and monthly precipitation sums (mm) during the experimental period (2002–2010) in comparison with the 1961–1990 period (visualized as a solid line). The thresholds of World Meteorological Organization for evaluation of the meteorological conditions were used (KOŽNAROVÁ and KLABZUBA, 2002).

As might be seen in Tab. 1, temperatures during summer, winter and the entire vegetation period were higher than during 1961–1990 reference period with the precipitation mostly close to the normal level. The last summer of the first rotation period (2009) was exceptionally wet but in the same time exceptionally warm.

72

Table 1: The overview of the selected climate characteristics of the experimental site. Climate characteristics Parameter

Units

January– April– December September

June– August

December– February

Mean air temperature: 1961–1990

°C

6.6

12.8

15.5

-2.7

Mean air temperature: 2002–2010

°C

7.6

14.2

17.1

-2.2

Precipitation sum: 1961–1990

mm

580.6

359.6

208.3

113

Precipitation sum: 2002–2010

mm

634.0

382.8

229.7

125.5

Except this general survey, the Tabs. 2 and 3 summarized the selected climatological variables just within the three years investigated in this study. From the point of water balance view it is important that the precipitation exceeded the evapotranspiration of the reference grass at the yearly level by 125 mm in average. Table 2: More detailed overview of the selected weather characteristics at the experimental site during the three main years investigated in this study. abbreviates standard deviation (calculated from the daily values) and the reference evapotranspiration according to ALLEN et al. (1998). Summary of the weather during the period 2008 to 2010 Variable Solar radiation (MJ m-2 day-1)

minimum maximum

mean annual total

mean

0.3

29.4

10.4

7.8

-27.7

31.4

7.1

8.1

Vapour pressure deficit (kPa)

0

3.1

0.2

0.2

Wind speed (m s-1)

0

7.5

1.7

0.9

Air temperature (°C)

66.4*

Precipitation (mm) (mm day-1) *

-0.2**

6.1

3787

657 1.5

1.4

532

daily maximum only condensation

**

At the growing season level the precipitation and reference evapotranspiration are relatively balanced which, however, means that during the dryer year compared to the mean of reference period or due to the unevenly distributed precipitation throughout the season, the drought spells might occur within this region.

73

Table 3: The overview of the selected weather characteristics at the experimental site during the three main growing seasons (1st April to 30th September) investigated in this study. (calculated from the daily values) abbreviates standard deviation and the reference evapotranspiration according to ALLEN et al. (1998). Summary of the weather during the growing seasons from 2008 to 2010 Variable

minimum maximum

mean seasonal total

mean

Solar radiation (MJ m-2 day-1)

2.3

29.4

16.4

6.5

Air temperature (°C)

-5.1

31.4

13.5

4.4

Vapour pressure deficit (kPa)

0

3.1

0.4

0.2

Wind speed (m s-1)

0

7

1.7

0.8

66.4*

Precipitation (mm) (mm day-1) *

0.2

6.1

2954

425 2.5

1.2

449

daily maximum

4.1.3 Soil Soil sampling took place prior to planting in August 2001 (see Tab. 4). Soil conditions at the location are representative of the wider region with deep luvic Cambisol influenced by gleyic processes and with a limited amount of stones in the profile. Although the area in general does not provide prime conditions for SRC based on Populus sp. clones, the site itself is highly suitable for planting due to deep soil profile (TRNKA et al., 2008). Except the initial soil sampling in 2001, each mid July, starting from 2007 (still continuing), one soil hole ~1 m deep with top view ~0.5 m2 was dug in the poplar plantation and one at the adjacent reference turf grass. From these holes the soil profile type was determined and the soil samples were collected from the three particular depths (~0.1, ~0.4 and ~0.7 m) into plastic bags (disturbed samples) and into standard stainless steel rings with uniform volume 100 ml (undisturbed samples) for further laboratory analysis. These analysis have been part of a broader research investigating the soil physical, chemical and mechanical properties in details, however, within this thesis only the basic hydrolimts (porosity, field capacity and wilting point) and bulk density are presented in the results because of their importance for water balance in general.

74

Table 4: The overview of the selected soil characteristics of the experimental site. Depth (cm) Component

Units 0–24

24–66

66–94

94–130+

Sand

wt %

34.2

27.6

42.7

67.1

Silt

wt %

50

46.1

38.7

19.6

Clay

wt %

15.8

26.3

18.6

13.3

-3

Bulk density

g cm

1.55

1.64

1.59

1.64

Organic matter

wt %

2.65

0.28

0.14

0.14

Total nitrogen

wt %

0.16

< 0.05

< 0.05

< 0.05

5.9

5.4

4.0

3.4

pH (KCl) Available P

mg kg-1

148

1.3

0.9

24

Available K

mg kg-1

151

91

62

76

Available Mg

mg kg-1

143

230

278

291

Available Ca

mg kg-1

1230

1353

748

652

Fraction of sand (0.05–2 mm), silt (0.05–0.002 mm), clay (< 0.002 mm); organic matter and total nitrogen are expressed in terms of weight fraction (wt %). The concentration of available nutrients was determined by Mehlich III method.

4.2 Measurement scheme, sensors and data processing Fig 4. depicts the aerial image of the experimental site. The core of the whole water balance measurement scheme consists of two Bowen ratio energy balance systems (BREB 1 and BREB 2). The first is placed above the poplar plantation and the second as a reference above short, regularly clipped (~10 days intervals) and fertilized turf grass with mean annual aboveground biomass productivity ~3 t ha-1 year-1 of dry matter content. Due to the height of the grass (~0.05 m) and the sensors configuration of the BREB 2, it is also very suitable for calculation of the so-called reference evapotranspiration according to FAO-56 methodology (ALLEN et al., 1998) and can serve as a typical meteorological station for various purposes. Around these two central measurement points, soil sensors and the access tubes for portable soil moisture measurement are placed. In case of the poplar plantation, there have also taken place the additional measurements like stem biomass and height increments, overall inventory, , sap flow etc. in the close vicinity to the mast with the BREB 1. The technical details of the particular measurements will be given in the following parts.

75

Figure 4: The aerial image showing the investigated poplar plantation with measuring point BREB 1 and the adjacent turf grass with the measuring point BREB 2. In the upper left corner, the plantation with clonal experiment described in TRNKA et al. (2008) is situated. 4.2.1 Soil Moisture To evaluate soil moisture patterns in the poplar plantation and under the reference turf grass, three different methods were used. Two of them are non-destructive and indirect, based on measuring the soil dielectric constant. The direct and destructive thermogravimetric method was used only for calibration of the indirect methods. Note that also the gypsum blocks were used for evaluating the soil water tensions (6 for each measuring points BREB 1 and BREB 2), however, the results are not presented within this thesis. 4.2.1.1 Soil moisture calibration The thermo-gravimetric method was chosen as the reference one for calibration of the electric soil moisture sensors. The soil samples were occasionally collected by using soil core sampler with one meter long gouge auger. At least two sampling points where the gouge auger were hammered next to the each permanently buried soil moisture sensors or the access tubes were chosen for each sampling occasion. In order to not disrupt the physical properties just in the close vicinity of the probes, the distance around one meter was set as the minimum and kept. The soil samples were always withdrawn from four different depths (0–0.15, 0.15–0.25, 0.25–0.35 and 0.35–0.45 m) and put into aluminium 76

boxes. Each of the boxes was filled with approximately 100 g of soil (fresh weight) and closed. After that, the soil samples were immediately taken into the laboratory where they were weighted and dried in the ventilated oven at 105 °C for 24 hours which was judged long enough to reach a constant weight. Then they were cooled and reweighed. The differences between two measurements provided the weight of the free water which the samples contained in the field. The mean dry bulk density determined from soil sampling in mid July 2007, 2008 and 2009 was used for conversion of the weight of dry soil into its volume. Assuming that the density of water is practically equal to unity, the volumetric soil moisture percentage could be established as the ratio of the volume of water and the volume of soil. Since the volume of the samples was not strictly defined, the one big assumption had to be adopted for this method that the bulk density is constant across the research plot and over the seasons which makes the method not direct in the true sense of the word. However, it remains still superior and more straightforward compared to the indirect methods based on the relations between soil dielectric constant and the electric output of the sensor. By relating the obtained volumetric soil moisture values to those from the electronic sensors, the linear and polynomial regression fit was used for the construction of the calibration curves. 4.2.1.2 Soil moisture spatio-temporal variability In April 2008, an array of 16 composite material access tubes was installed into the soil in order to measure soil moisture via portable system of PR1 profile probe and HH2 hand held readout datalogger device (Delta-T Devices Ltd., UK) – a system measuring dielectric properties of soil, which are straight depending on soil water content. One month later, two other access tubes were placed next to the BREB 2 and one access tube at the edge of the reference turf grass close to the border with the poplar plantation for comparison of soil moisture patterns under two different cultures. The layout of the access tubes in the poplar plantation was designed to record differences in soil moisture between double rows and inter-rows and the soil moisture variability itself within the investigated area which is roughly 600 m2 large. Readings were taken usually once a week. The calibration through the gravimetric method was done at least five times for each of the access tubes occasionally during the seasons 2009 and 2010. The frequency domain reflectometry (FDR) PR1 profile probe enables to evaluate volumetric content of soil water (m3 m-3) in different depths (0.1, 0.2, 0.3 and 0.4 m). It consists of a sealed composite rod (~25 mm in diameter) with electronic sensors (in the form of pairs of stainless steel rings) arranged at fixed intervals along its length. The accuracy of the calibrated probe is ±3 % in an access tube and the sampling volume in each of the four measured depths is ~1.5 dm3. When the electric power is applied to the profile probe, it generates a 100 MHz signal which is applied to the pairs of stainless steel rings. They generate an electromagnetic field which extends about 100 mm into the soil passing easily through the access tube walls, but rather less easily through any air gaps. If the dielectric properties of the soil are different from the probe electronics, some of the 100 MHz signal gets reflected back. The reflected part of the signal combines with the applied signal to form a standing wave defined with a certain analogue voltage correlating with the soil water content. Because the electromagnetic field is not propagated uniformly around 77

the whole circumference of the measuring rod, the average of three readings with profile probe rotated through 120° in the access tube was always taken for each measuring point, as recommended by the manufacturer. The thorough description of the measurement technique and principle can be found in the user manual of Delta-T Devices Ltd. (2001). 4.2.1.3 Soil moisture temporal variability In May 2008, the mast with BREB 1 system was erected within the poplar plantation (see Fig. 4). At the same place below ground, three capacitance sensors EC-10 (Decagon Devices, Inc., USA) for measuring volumetric water content of soil and three pairs of resistance gypsum blocks (Delmhorst Instrument Co. Moisture Meters, USA) to measure soil water potential (data not presented) were accommodated in soil within the depths 0.1, 0.3 and 0.9 m at the end of June 2008. All sensors were connected to datalogger ModuLog 3029 (EMS Brno, Czech Republic) and measuring step was adjusted to measure each 2 minutes and to store each 10 minutes averages which enabled to investigate the soil moisture temporal variation at one place in details. The same configuration of soil moisture sensors was used for BREB 2 constructed in July 2008. The accuracy of the EC-10 sensor is ±4 % (with a soil-specific calibration ±2 %). The probe dimensions are 145 x 31.7 x 150 mm, whereas the probe averages the volumetric water content over the entire length of the probe, with about a 20 mm zone of influence with respect to the flat surface along its length. The EC-10 measures the water content of the soil using a capacitance technique (CAMPBELL and GREENWAY, 2005, BOGENA et al., 2007, KIZITO et al., 2008). By rapidly charging and discharging a positive and ground electrode (capacitor) in the soil, an electromagnetic field (5 MHz) is generated whose charge time is related to the capacitance of the soil. Further, for a capacitor with a geometrical factor, the capacitance is related to the dielectric permittivity of the medium between the capacitor electrodes. For more details regarding the measuring technique principles of these capacitance sensors, see the user manual of Decagon Devices, Inc. (2002). 4.2.2 Evapotranspiration In May 2008, 12 m high aluminium mast construction with system for estimating actual evapotranspiration ( ) by measuring Bowen ratio and radiation balance (EMS Brno, Czech Republic) was placed within the poplar plantation at a properly selected place (see Fig. 4) with respect to distances from the edges and the prewailing wind direction – hereinafter BREB 1. The upper arm with combined air temperature and humidity sensor, global radiation sensor and the net radiometer can be adjusted to the maximal height 14 m above the ground. The BREB system 1 consists of three vertically adjustable aluminium arms with three combination thin-film polymer capacitive relative humidity and adjacent temperature sensor instruments EMS 33 (EMS Brno, Czech Republic). EMS 33 is microprocessor-controlled reliable and accurate sensor with standardized linear output attended for long-term measurement of air temperature and relative humidity. The sturdy sensor design protects the sensitive parts against dust and sprinkling water. The sensor body is made from anodized aluminium and the sensor chips are shielded with stainless screen. The electrical connection as well as the mechanical design is based on robust 78

Amphenol C-16 connector. The accuracy of the sensors is ±0.3 °C and ±2 % for the temperature and the relative humidity respectively. However, due to pairing of the sensors, the mutual error expressed as was found to be below 0.3 °C and 0.6 % for the temperature and the relative humidity respectively. All three EMS 33 instruments were placed into the white painted aluminium radiation shields (EMS Brno, Czech Republic) protecting the sensors against the direct sunshine and rain. The two lower shields are AL 070 and the upper one is AL 071 with holder for global radiation sensor EMS 11 (EMS Brno, Czech Republic). The temperature and humidity sensors are essential parts of the BREB enabling to calculate the temperature and vapour pressure vertical differences and subsequently the Bowen ratio itself according to Eq. 42. Nevertheless, prior this step, the relative humidity has to be converted into vapour pressure according to the following relations: (111) (112) where is actual vapour pressure (kPa), relative humidity (%), saturated vapour pressure (kPa), and temperature (°C). Note that three levels for gradient measurement were used although only two are necessary. The third additional level has the advantage over aerodynamically rough surfaces like forest where the gradients are generally smaller. Moreover, it has also the advantage for evaluating the fetch conditions. The other necessary parts are the net radiation and soil heat flux sensors enabling to solve the radiation balance part of the BREB method. The net radiation was measured with properly calibrated net radiometer 8110 (Philipp Schenk GmbH Wien, Austria) with accuracy ±5–10 % and spectral sensitivity 0.3–100 μm. The soil heat flux was assessed by plate sensor HFP01 (Hukseflux, Netherland) with expected typical accuracy -15 % to +5 % in most common soil (for 12 hour totals). All of the mentioned sensors were connected to datalogger ModuLog 3029 (EMS Brno, Czech Republic) with measuring step adjusted to each 2 minutes and storing the averages set to each 10 minutes. The ModuLog 3029 has the storage capacity for 220,000 values and provides 18 voltage channels, 8 A.C. resistance channel and two 16-bit counter channels. For more information see the Datalogger ModuLog User´s Manual (2009). The whole system was fed by rechargeable battery (12 V) connected to solar panel. The stored outputs of the net radiation and the soil heat flux sensors and the knowledge of the Bowen ratio from the temperature and humidity gradient measurements enabled to use the Eqs. 43 and 44 to obtain the rough latent and sensible heat fluxes. For calculation of the fluxes, the halfhourly means of the input variables were used. The latent heat flux resulted from Eq. 43 was subsequently divided by latent heat of vaporisation (after necessary unit conversion) given according to SHUTTLEWORTH (2007) by the following equation: (113) where (MJ kg-1) is the latent heat of vaporisation of water and (°C) the air temperature in, in order to obtain the actual evapotranspiration in mm hour-1. Except the calculation of 79

the actual evapotranspiration, other variables like Penman open water evaporation (Eq. 91), Priestley-Taylor potential evaporation (Eq. 94), Penman-Monteith (Eq. 97) and other variables mostly described in chapter 1 were calculated for various purposes later described in text. Except this fundamental instrumentation of the BREB system, other following supplemental sensors were connected to ModuLog 3029 and used for other purposes. At the radiation shield AL 071, the global radiation sensor EMS 11 (EMS Brno, Czech Republic) was placed at the aluminium horizontally levelled holder. The EMS 11 is low cost photodiode sensor with accuracy ±7 %. As already mentioned, at the same place below ground, three sensors EC-10 (Decagon Devices, Inc., USA) for measuring volumetric water content of soil and six gypsum blocks (Delmhorst Instrument Co. Moisture Meters, USA) to measure soil water potential were buried in the depths 0.1 m, 0.3 m and 0.9 m. In addition to these soil sensors, three PT 100 sensors (accuracy ±0.3 °C) for measuring soil temperature were placed in the same depths in the soil profile. The second system BREB 2 positioned at the reference turf grass consists almost identical sensor configuration however, there are some differences. The first difference is in the length of the mast which is only 2 m height with the fixed arms at 2 and 0.4 m above the ground. Further, there are only two levels for measuring the temperature and humidity gradients. The rest of sensors are absolutely identical. The BREB 2 is additionally equipped with wind sensor MetOne 034B (MetOne Instruments, USA) for measuring the wind speed and direction with accuracy ±0.1 m s-1 and ± 4° for the speed and direction respectively. Beside the BREB 2 is in addition placed the tipping bucket rain gauge MetOne 370 (MetOne Instruments, USA) with the resolution 0.2 mm per pulse and collecting area 0.032 m2. While the arms with temperature-humidity sensors at the BREB 2 were fixed, the arms at the BREB 1 were regularly adjusted so that the lowest level was about 0.2 m above the highest tree at the time of adjustment. In 2008 and 2009 the vertical distances between all consecutive arms were 2 m, whereas in 2010 the distance between the lowest and the middle level were 1 m and the distance between the middle and the upper level 2 m. In April 2009, two low cost digital phenocameras (EMS Brno, Czech Republic) taking colour picture (auto-zoomed) each hour during the light part of day were installed at the mast with the BREB 1 at heights 7 and 13m within and above canopy respectively. During the next year 2010, their heights were adjusted according to the growth of the resprouting shoots with the lower one approximately in the middle of the canopy height and the upper one roughly two meters above the canopy. Although the cameras were placed on the mast primarily for the phenological observations, within this thesis they were used especially for the additional assessment of the leaf are index ( ) and overall canopy conditions. 4.2.2.1 Error analysis To estimate the uncertainty propagated from the particular sensor errors into the resulting flux, the procedure describing the maximal possible errors given by FUCHS and TANNER (1970) was used. As the first step, the following relative uncertainty of the Bowen ratio has to be known:

80

(114) where , and are the errors in the Bowen ratio, temperature and vapour pressure differences respectively given by the resolution limits of the sensors. However, in case of any relative humidity capacitive sensor, the error in water vapour pressure is not constant, but depends on the actual relative humidity, temperature and the accuracy of both sensors (SAVAGE, 2010): (115) where (kPa °C-1) is the slope of the saturation water vapour pressure vs temperature relationship at temperature T: (116) If the error in the Bowen ratio is known, the error analysis can be applied on the final fluxes of latent and sensible heat: (117) (118) where , , and are the errors in the latent heat flux, sensible heat flux, net radiation and soil heat flux respectively. 4.2.2.2 Data rejection Till now, only the fluxes without any filtering and data rejecting were considered. However, by applying the error analysis given above, some part of the data will result in very large errors and unrealistic magnitude and has to be discarded. To solve this issue, two-step procedure was adopted. Firstly, data which did not fulfil the following inequality given by OHMURA (1982) later rewritten just into one equation according to PEREZ et al. (1999) were rejected: (119) If the measured does not meet Eq. 119, the calculated fluxes have the same sign as the temperature and humidity differences, opposed to the flux-gradient relationships in Eqs. 15 and 16. Secondly, BREB method produce the energy fluxes with unreasonably large absolute values when approaches -1 (Eqs. 43 and 44). Traditionally, the data are rejected in a fixed range around -1. With error analysis, PEREZ et al. (1999) and later GUO et al. (2007) proposed and analytical method to express the range defined as: (120) where is variable value depending on both system accuracy and the measured vertical humidity difference (GUO et al., 2007): 81

(121) Although this data rejection method is quite strict, there can still remain some outliers which had to be manually discarded following the graphical check using the Penman open water evaporation (Eq. 91) as a reference variable. Note that the sign convention within this thesis is that the is positive downward during the daytime, same as , and negative upward during the night time. Both latent and sensible heat fluxes are positive upward and negative downward, opposite to the gradient directions (Eqs. 15 and 16) and thus the temperature and humidity vertical differences are always calculated as the upper minus lower level. 4.2.2.3 Gap filling The proper quality control and data rejection ensured that only the reliable values remain within the dataset. However, they could not be easily integrated into daily or longer totals due to the originated gaps without any gap filling procedure. Within this theses the gaps were filled by Penman open water evaporation model (PENMAN, 1948) multiplied by a factor obtained as a slope from linear regression between the actual evapotranspiration resulted from BREB measurement and the Penman potential evaporation (Eq. 91). The similar approach was used by GUO et al. (2007) who used linear regression between actual evapotranspiration and the equilibrium evaporation separated into three flux sign scenarios according to fluxes direction combinations. The linear regression between two variables can be done e.g. for each month or even for longer period, but it was hypothesized that the ratio between the Penman and actual evapotranspiration is more variable in the rain-fed conditions of Czech-Moravian Highlands. For this purposes, simple software for gap filling based on linear regression with a work name ―FluxCorrector‖ was proposed within this study. The advantage of the software is that it can make the linear regression for each day separately if there is enough reliable data. Moreover, it basically split the whole day into two main parts – day and night, according the calculated time of sunrise and sunset based on the input coordinates, and it calculates the linear regression for these parts of day separately. Even these two parts can be further divided and the regression calculated separately if there is enough data. This could contribute to better fitting especially during the water shortage when the course of evapotranspiration has significant trends during the day (noon depression). The reliable data sufficiency can be adjusted by simple percent threshold. For all of the gap filling calculations in this study a threshold 30 % was used which means that when during the particular part of the day more than 30 % had been missing, the data were not used for calculating the linear regression and adjusting the multiplying factor. If there is not enough values, program will find out the temporally closest part of day with sufficient number of values and with the most similar radiation conditions while takes into account the occurrence of precipitation. The similarity of radiation conditions are evaluated based on the ratio between the extraterrestrial (calculated from the coordinates according to ALLEN et al., 1998) and global radiation (measured) whereas user can set the radiation tolerance threshold – the allowed percent of the difference between the two variables. The day with precipitation are taken completely

82

separately. These two last conditions are ensuring that the sunny, cloudy and also rainy days were treated separately. For the purposes of this thesis, the day was split simply only into the diurnal and nocturnal part since no considerable water stress with marked noon depletion of evapotranspiration was observed. 4.2.2.4 Fetch and footprint Flux measurements from small fields are inherently difficult because of the limited fetch. To quantify to which extend are the measured fluxes equilibrated into the area of interest (poplar plantation for BREB 1 and the turf grass for BREB 2) the following rewritten form of the Gash´s formula (GASH, 1986) was used:

(122) where (%) express the percent of the adjustment or percent of so-called effective fetch, (m) is height of sensor above the zero plane displacement, (m) is the roughness length, is von-Karman constant (0.41), (m) is distance from the point of measurement. Note that the percent of adjustment has the same meaning as the cumulative normalized contribution to flux measurement described by SCHUEPP et al. (1990) or BURBA and ANDERSON (2010). In other words, this rewritten form of the Gash´s formula tells us, how much of the flux is coming from the upwind area for a defined distance from the measuring point. Since this Gash´s model is derived for one point measurement which, in practical term, means eddy covariance, the following third order polynomial equation given by STANNARD (1997) was used: (123) This equation enables to estimate the theoretical effective height, , of the hypothetical one point measurement instead of two levels used in BREB method where the symbols and are the heights of the upper and lower sensors (intakes) respectively. Note that all heights (m) are considered to be located above the zero plane displacement. The combination of these two equations enabled to evaluate if the distances between the point of measurement and the border of the investigated cover were sufficient assuming the wind blowing from particular azimuths. For measuring the distances from the BREB system to edge of the area of interest towards the particular azimuths, the simple tool for measuring distances of aerial images in Google Maps was used. 4.2.3 Other micrometeorological and eco-physiological variables 4.2.3.1 Bowen ratio After the proper quality control and data rejection, the diurnal ( > 50 W m-2) medians of the Bowen ratio resulting from the Eq. 42 were used for comparison of the energy partitioning above the poplar plantation and the turf grass. The diurnal values were used due to fact that this data are the most important from the evapotranspiration point of 83

view. Further, using of the medians was selected because of the presence of peaks with outlier character which did not allow to ensure reliable mean values. These two facts are valid also for the following two variables. 4.2.3.2 Surface resistance To compare the differences between the surface resistance of the poplar plantation and the turf grass, the diurnal ( > 50 W m-2) medians of the surface resistance resulting from the following rewritten Penman-Monteith formula according to PEREZ et al. (2008) were used: (124) where (s m-1) is the aerodynamic resistance obtained from Eq. 103 using the wind speed measured at BREB 2 system, (0.066 kPa °C-1) is the psychrometric constant, (1.23 kg m-3) is the density of air, is the specific heat of air (1004 J kg-1°C-1) and (kPa) is the water vapour pressure deficit. Further, a semiempirical relation for stomatal conductance developed originally by LOHAMMAR et al. (1980) was used in order to reveal the dependency of the canopy conductance (reciprocal to the surface resistance) of two different canopies on the vapour pressure deficit and the global radiation. The following expression was parameterized using the input values and the canopy conductance obtained from Eq. 124 for periods with > 200 W m-2 excluding the days with rain: (125) where

(m s-1) is the canopy conductance,

and

(W m-2) is the global radiation and

,

are the species specific parameters.

4.2.3.3 Decoupling coefficient In the same way as in the previous step, the diurnal ( > 50 W m-2) medians of the decoupling factor derived from the rewritten form of the Penman-Monteith equation according to MCNAUGHTON and JARVIS (1983) formula for amphistomatous species: (126) were calculated for the two contrasting surfaces. An omega factor close to one indicates that transpiration is mainly controlled by available energy (de-coupled system) and a small value indicates that transpiration is mainly controlled by vapour pressure deficit and canopy conductance (well-coupled system). 4.2.3.4 Crop coefficient Following the FAO-56 methodology (ALLEN et al., 1998) the crop coefficient for poplars were derived from the Eq. 108 as a ratio of the measured actual evapotranspiration (BREB 1) and the reference evapotranspiration calculated from the daily meteorological 84

data measured above the turf grass (BREB 2) according to Eq. 107. In order to obtain more generally usable the inverted form of the formula given by ALLEN et al. (1998) was used: (127) where is the crop coefficient as the result of the ratio between the measured actual (assuming that the stand did not suffer on any water shortage) and the calculated reference evapotranspiration (Eq. 1008), (m s-1) is the wind speed measured at 2 m above the ground, (%) is the minimal daily relative humidity and (m) is the mean canopy height. 4.2.3.5 Sap flow Sap flow measurement based on current tissue heat balance method design enabled to investigate only the most of the dominant trees. Sap flow of three trees were investigated at the end of the growing season 2008 (September), and four trees were measured from the 18th June till the end of October 2009 using the system EMS 51 (EMS Brno, Czech Republic). Sap flow module EMS 51 is one-channel battery operated system for the field measurement of sap flow in large trunks with the diameter above ~0.1 m. The module cooperates with a measuring point composed of a set of four stainless steel electrodes and needle temperature sensors. The module is designed as a watertight unit containing all electronics necessary for the sap flow measurement based on tissue heat balance method with internal heating and variable power. The unit supplies xylem tissue around three flat stainless electrodes with an A.C. voltage. The temperature of electrodes is measured with a set of needle thermosensors. The heating power is automatically controlled in order to keep 1 K temperature difference between heated and reference (nonheated) electrodes in the measuring point. For brief theory, see the subchapter 2.3.2.2 and the provided references or the manual of the used sap flow system EMS 51 (2006). Prior to the hammering the particular electrodes using the original electrodes inserting tool and electrode bumper (EMS Brno, Czech Republic), the depth of xylem (6 mm) was measured by bark and phloem thickness gauge in order to place the electrodes precisely into the xylem and to avoid measuring in the highly electrically conductive phloem. The sensors were orientated northward towards the inter-row. They were installed following exactly the EMS 51 Sap Flow System Installation Guide (2010) and finally covered by the weather protection set. Between three heating sensors, which cover an effective tangential width of 5.5 cm and a reference sensor, the system kept a constant temperature difference of 1 K. Stainless heating plates and electrodes covering a xylem depth of 35 mm were used assuming that electrodes cover the main part of the conducting xylem depth. The output voltage for heating the stainless plates of the three upper sensors was measured each 1 minute and the averages were logged every 10 min in datalogger Edge Box V 8 (EMS Brno, Czech Republic) with 8 voltages and 2 counters channels and the memory capacity for 110,000 records. The values represent the integral over the heated volume of the trunk segment. They were directly converted to the sap flow rate by a physical-based heat balance Eq. 75. Mini32 software (EMS Brno, Czech Republic) automatically calculated the sap flow rate (kg h-1) for each measured tree according to biometric parameters. Baseline 85

correction was manually performed using Mini32 assuming zero flow before sunrise when during the night dropped to zero. Upscaling of single tree transpiration to stand transpiration was based on relation between the sap flow and the (diameter at breast height) whereas it was assumed that the sap flow is related to stem thickness in the same way as the leaf area. Based on this assumption, all leaves from selected 30 trees (representing the whole social distribution) were manually harvested and the relation between and specific leaf area ( ) was determined for further scaling purposes. In order to get information about the vertical leaf distribution, each tree was split into several levels by one meter where the leaves were collected separately. 4.2.3.6 Biomass growth and water-use efficiency Since April 2009 the was measured regularly (usually weekly) by using SunScan ceptometer system (Delta-T Devices Ltd., Cambridge, UK). This non destructive measurement technique is based on simultaneous measurement of the incident photosynthetic active radiation ( ) in the waveband 400–700 nm above (or outside) and the transmitted below the evaluated canopy. The incident was measured by the Beam Fraction Sensor (BFS) which incorporates multiple photodiodes, of which one is always shaded. This patented design allows the direct and diffuse components of to be separated, which is necessary for the computation of . The BFS was connected by 50 m long cable to the ceptometer in form of a 1 m long measuring probe with 64 equally spaced photodiodes and handle with batteries electronics for converting the photodiode outputs into digital readings. The ceptometer was connected via serial RS232 cable to the portable computer Psion Workabout with the SunData software. This software contains the algorithms approximating functions for real-time estimations of from the measured inputs in the field. The values calculated by the SunData software are within ± 10 % ± 0.1 of the that would have been calculated by the full model which is much more computationally demanding. The full model is based on the Campbell´s derivation for a beam of light passing through the canopy from a single direction CAMPBELL (1986). The authors of the model (see the SunScan Canopy Analysis System User Manual, 1996) integrated the Campbell´s results over the whole sky assuming a point source of direct light only from one given zenith angle and a diffuse radiation coming from the every point in the sky. Moreover, the model considers that the canopy is not a black body and that reemits the radiation from the leaf elements uniformly in all directions. Using this sophisticated model of transmission, the is finally calculated using a rewritten Beer´ law, describing the radiation attenuation by absorption, from the differences of the incident and the transmitted light. During the measuring routine, the BFS was placed on the 1.5 m high tripod at the open grassland south from the plantation and was precisely horizontally levelled. The measurements were carried out regularly approximately one hour before noon at three locations within the plantation. The locations were selected close to the mast with BREB 1 as far from the BFS as the length of the cable allowed. Within each location the measurements with the hand held and horizontally levelled ceptometer were carried out within the ~0.7 m distances diagonally between the trees within the double row and with 86

the same distances and horizontal angle within the inter-row. At each location approximately 40 readings were carried out for the double row and for the inter-row. Additionally, in the middle of each inter-row one point measurement around the vertical axis (360°) was done with two reading for approximately each 30°. Note the result obtained by the SunScan does not actually provide the but whole plant area index ( ). Therefore, at the beginning and at the end of the season when the trees were unfolitaed, the same measurement with SunScan was carried out in order to determine the bark area index ( ). This were subsequently deducted from the measured and the was obtained. To validate the measurement by SunScan a common destructive leaf litter collection was done. The leaves were collected from three traps with the area of 3 m2 positioned in the middle of each location where the in the inter-row was measured assuming that the contents of the traps represent the average amount of leaves falling in the stand. The leaves were collected five times at the end of the growing season 2009 (from the end of September till the last leave fall in November). The leaves were collected continuously throughout the period, not allowing the leaves to lose their original weight by decomposing. The leaves were collected into the paper sack and weight. After that a representative sample of leaves (~50 pieces) from each sack was taken and weight again. Then the selected leaves were spread at the white desk with known dimensions and the photos by camera Sony DSC-W210 (Sony Electronics Inc., USA) were taken from the tripod with a known distance from the object. Then the images were analysed in the software MultiSpec Application v 3.1 (Multicolor Specialties, Inc., USA) and the area of leaves were determined. These leaves were subsequently dried in the ventilated oven at 50 °C until the constant weight was reached. By this procedure the specific leave area index ( ) was obtained which enabled to back calculate the from the weight of the dry leaves. The selected leaves were also considered as a representative sample for determining the ratio of the dry and fresh mass. For estimating the increment of the aboveground woody biomass an array of 15 mechanical – DB 20 and 3 automatic dendrometers – DRL 26 (EMS Brno, Czech Republic) were used. These measurements were updated by adding another 15 DB 20 and 1 DRL 26 dendrometers at start of the season 2009. The DBs and DRLs were fixed to trunks at the breast height and read in a week period while the DRL 26 dataloggers were adjusted to hourly measuring step. The dendrometers used are designed for long-term registration of tree trunk circumference via stainless tape that encircles the tree trunk. The values of the stem circumference or increment are very useful because they could be subsequently converted through the allometric equation to increment of biomass (e.g. CIENCIALA et al., 2005; AL AFAS et al., 2008; FAJMAN et al., 2009). During the winter harvest in March 2010, the allometric measurements were carried out in order to find out the relation between the easily measurable tree parameters and the total aboveground woody biomass. The stem girths in the place of cut (about 0.1 m above the ground), at the original height of 1 m and at original were measured by means of a gauge as well as the stem length (the stump height was not taken into account at measuring the length of the felled stem). Thanks to the gauge, through the girth measurement the stem diameter was directly read and thus, effects of ovality of some 87

stems were reduced. Moreover, the total length of a felled stem was measured and branches of a diameter over 10 mm were counted. Samples from 42 individuals, where the dendrometers and sap flow measuring instruments were placed, were taken to determine dry matter content (DMC). Whole trees (stems and branches separately) were converted to the wood chips samples (~0.75 dm3) representing a mixture of the whole tree branch and stem part and weighed in fresh state. Except these 84 samples of particular 42 trees, the biomass of other 80 trees divided into four groups (n = 20) were used for DMC analysis to increase the number of samples and make the allometric relationships more robust. Overall, 6 random samples (3 from stem and 3 from branch biomass) were taken from each of the four groups of these additionally measured trees. The biomass samples were subsequently drying in ventilated oven at 105 °C till the constant weight was reached. After drying, the DMC of stems and branches was determined. As a result, the curvilinear or non-linear allometric functions describing the relation between the biomass content and the other measured variables were parameterized. The same biometric variables were measured at these additional 80 trees and were used for further fitting of the allometric curves too. Finally, the aboveground woody biomass (AB without leaves) increment expressed in g of biomass (DMC) per m2 of the ground was divided by the amount of (mm) integrated to the periods corresponding to dendrometers readings and thus the long term was obtained. Because in this work, the is defined only from part of above ground biomass (stems and branches – the growth of leaves and roots is not considered) which is divided by (not only transpiration), the term gross is used in order to point out the differences against typical view on . Except the relation of the allometrically derived biomass increment from dendrometers and the from BREB, the results from tree sap flow were related to the tree AWB increment in order to assess at the tree level.

4.3 Statistical methods Data analyses were performed using three different numerical softwares. Firstly, statistical package STATISTICA 9 (StatSoft, Inc., USA) was used especially for an analysis of variance (ANOVA) with a post-hoc Fischer´s least significant difference (LSD) test to evaluate the significance of differences between the particular replicates and treatments at < 0.05 and < 0.01. Prior the ANOVA, the data were checked for normal distribution and normalized using a Box-Cox transformation if necessary. Another widely used process used in STATISTICA 9 was Pearson correlation coefficient (r) and its second order coefficient of determination ( or in case of multiple linear regression). For fitting the parameters of nonlinear functions, the least-squares method using Levenberg-Marquardt algorithm was used. Secondly, the software Mini32 (EMS Brno, Czech republic) was used for collecting the data from dataloggers as well as for in the field data check and finally for the whole calculation procedure of the actual evapotranspiration. The graphically oriented software Mini32 consists basic statistic functions including multiregression analysis up to twenty parameters, histograms of frequency distribution, drawing of vertical profiles etc., always easy accessible from the current graph. Except the basic functions it enables to use 88

iterative parameterization of the complex dynamic equation like Penman-Monteith. In addition, Mini32 enables writing program scripts for rough or already processed data postprocessing in easy and intuitive mathematical language. These scripts may include various type of equations linked in complicated chains in order to derive required variable from the input data. One example of complete script for calculation the actual evapotranspiration and other variables from BREB 2 is shown in appendix. Thirdly, the Microsoft Excel 2007 was used especially for creating figures and tables, for data organizing and the basic statistical calculations like e.g. linear regression. Except the mentioned statistical tools and methods, the mean bias error ( ) and the root mean square error ( ) were used according to PEREZ et al. (2008) for evaluating the systematic and random error respectively: (128) (129) where the

represent the observed values, whereas the

89

the predicted (estimated) one.

5 RESULTS In the first part of this chapter, the particular results from soil moisture measurements based on different methods will be presented. Further, the evapotranspiration measured by the Bowen ratio technique and the analysis of the possible source of errors of the method itself will be investigated. In the last part, the additional micrometeorological and ecophysiological variables based mostly on BREB, sap flow, and allometrically derived biomass measurements will be analysed.

5.1 Soil Moisture In fact, soil moisture is mainly the result of precipitation and evapotranspiration balance and the soil properties. Although, the precipitation is very important part of the water balance, it will be not analysed individually but only as the additional supporting variable. The first mention about precipitation can be found in the methods (Fig. 3) and the second in the results of soil water loss and income (Figs. 18, 19) and then daily (Tab. 10) and yearly course of actual evapotranspiration (Fig. 45). The main mean hydro-pedological properties based on the soil sampling during the summer (always mid July) 2007, 2008 and 2009 are described in the following table. Table 5: The overview of the main hydro-pedological parameters calculated as the mean from three soil sampling (2007, 2008 and 2009) in the poplar plantation and the turf grass.

Cover

Poplar plantation

Turf grass

Depth (m)

Bulk density (kg dm-3)

Porosity (m m-3)

Field capacity (m m-3)

Wilting point (m m-3)

0–25

1.6

0.38

0.34

0.16

25–50

1.7

0.39

0.33

0.15

50–90

1.6

0.43

0.39

0.12

0–25

1.5

0.51

0.33

0.16

25–50

1.6

0.40

0.33

0.16

50–90

1.7

0.42

0.36

0.17

5.1.2 Soil moisture calibration The resulting dry bulk densities (Tab. 5) were subsequently used for calculation of the volumetric soil moisture (m3 m-3) from thermo-gravimetrically determined moisture dryweight fraction (kg kg-1). By this way obtained volumetric soil moisture were considered as the standard to which all other indirect methods were related in this work. Firstly, the outputs of six capacitance probes EC-10 placed within three depths (0.1, 0.3 and 0.9 m) in soil profile in the poplar plantation and in the turf grass were compared with thermogravimetrically determined volumetric soil moisture and the calibration curve for EC-10 was established (Fig. 5).

90

Figure 5: The scatter view on the relationship between soil moisture measured by capacitance EC-10 sensor (horizontal axis) and the thermo-gravimetrically based soil moisture (vertical axis) which serves usually as the reference method. The EC-10 sensors were placed below two contrasting covers (the poplar plantation and the turf grass), in both cases in three depths (0.1, 0.3 and 0.9 m). The resulting 3rd order polynomial function is used as the calibration equation for further analysis. The four points depicted in red colour were judged as the outliers and thus they were not considered for establishing the calibration relationship. The dashed blue line depicts the linear relationship, which were not, however, used as the calibration curve. By considering errors linked with thermo-gravimetric method, then three different depth and two different covers for EC-10 placement, that implies slightly different bulk densities, different soil particle size distribution, different organic and mineral contents, the relation in Fig. 5 is very tight and confirms the robustness and high accuracy of EC-10 after site-specific calibration. The outliers depicted in the Fig. 5 as the red circles can be explained just by the above mentioned errors of the both methods and especially by the soil spatial variability. This variability is not only in the vertical direction, but soil properties can also changed very sharply horizontally. Therefore, it must be taken into account that the soil samples treated by the gravimetric method are not identical with the samples measured by EC-10 (there is always at least 1 m distance in order to not disrupt the continual measurements) which can additionally results in differences of the both methods. Secondly, the output from the FDR sensor PR1 was compared with the thermogravimetrically based soil moisture values, in the same way as the EC-10. However, no very tight and robust relationship was found for PR1. One of the relevant reasons is that there were unfortunately only sporadic occasions to treat the soil by the gravimetric method (also with respect to the time consuming nature of its feasibility) during the low soil moisture levels. It resulted in many points for calibration in the upper level of soil moisture values and at the same time in almost no values which could anchor the calibration curve near the bottom of the y axis. Therefore, the previously verified 91

calibrated EC-10 was used as a reliable tool for calibrating the PR1. The advantage of such more indirect way of calibration is that the sensors were closer one to each other (less than 1 m) and especially they measured in the same time much more often than the gravimetric method was carried out. The Fig. 6 shows the relationship between soil moisture measured by PR1 and those measured by already calibrated EC-10. In order to find one general calibration for our site, the values from the poplar plantation and from the turf grass were grouped together, which resulted in higher scatter, but on the other hand higher robustness for further applications. Again, some small number (less than 5 %) of outliers was excluded to find the dominant and robust calibration curve. The final curve was linear, which agrees to the basic assumption when comparing moisture to moisture. As it is evident at the Fig, 6, there are four calibration curves for each of the particular depths. It was found that each of the four sensors embodied in PR1 profile probe does not measure the same values even for the air or pure water. Namely, there is almost 7 % difference between the sensors in 0.1 and 0.4 m and approximately 1–2 % difference between the sensors in 0.1, 0.2 and 0.3 of measuring depths. The scatter of the points and with this linked occurrence of the outliers can be explained by the same reasons as previously for the EC-10.

Figure 6: The four scatter views (four different depths from 0.1 to 0.4 m) on the relationships between soil moisture measured via portable PR1 profile probe and the thermo-gravimetrically calibrated EC-10 sensor. 5 % of data were judged as outliers which were removed in order to obtain more robust relationship.

92

5.1.2 Soil moisture spatio-temporal variability Assuming that the PR1 is more or less well calibrated, its portability with using of access tubes enable to investigate soil moisture spatial variability within the field at different depths. The following four figures at the next pages show the soil moisture spatial variability within the plot (approximately 600 m2) near the centre of the poplar plantation as an interpolation between 16 access tubes – measuring points. The figures include the image on the spatial variability within the particular depths during the days usually in weekly step. In this way, the evolution of the spatial variability, the drying or wetting spatial anomalies and natural vertical rules are very well illustrated. At the first Fig. 8, the driest period (late summer 2008) from the whole measurement campaign is depicted. Fig. 9 shows the wet period after snow melting and subsequent drying from the April till the May 2009. This drying period at the late spring 2009 is terminated by the strong rainfalls in the second half of June 2009 (Fig.10). And finally Fig. 11 depicts very fast drying of the surface layer under bare soil (plantation after winter coppicing) during the beginning of July 2010. Note, that only the part of the each top view which is situated inside the imaginary rectangle given by the corner points (particular access tubes) might be considered as correct. This imaginary rectangular is depicted by black dashed line (slightly transparent to enable to see the particular access tubes) at Fig. 7. Everything which is out of this rectangular is not correct, because it is only estimated extrapolation based on the interpolation (weighted average) between the closest access tubes. 16

15

14

13

9

10

11

12

8

7

6

5

1

2

3

4

Figure 7: Zoomed top view on the spatial soil moisture variability from the 1st of July 2010 in the 0.2 m depth. The particular access tubes are depicted as white spots. Their numbers describe the usual sequence of measuring routine from 1 to 16. The black dashed line (slightly transparent to not cover the particular white point – access tubes) depicts the imaginary rectangular (defined by particular access tubes 1, 4, 13 and 16) inside which the interpolation are made by weighted averages between particular access tubes and might be considered as correct. However, outside this area, there are only extrapolations based on the interpolations between the access tubes and they have increasing or decreasing trends with unrealistic tendency and meaning.

93

0.2 m depth

0.3 m depth

0.4 m depth

6.9.2008

29.8.2008

21.8.2008

12.8.2008

7.8.2008

1.8.2008

24.7.2008

15.7.2008

9.7.2008

0.1 m depth

0 4 8

16

24

32

Meters Soil moisture (m3 m-3) > 0.40 < 0.40 < 0.38 < 0.36 < 0.34 < 0.32 < 0.30 < 0.28 < 0.26 < 0.24 < 0.22 < 0.20 < 0.18 < 0.16 < 0.14

Figure 8: The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the late summer drying period (from 9th of July to 6th of September 2008). Each column belongs to one depth and each line to the particular date. The small white points depict the 16 access tubes for measuring with PR1.

94

0.2 m depth

0.3 m depth

0.4 m depth

13.5.2009

7.5.2009

1.5.2009

25.4.2009

17.4.2009

10.4.2009

6.4.2009

27.3.2009

19.3.2009

0.1 m depth

0 4 8

16

24

32

Meters Soil moisture (m3 m-3) > 0.40 < 0.40 < 0.38 < 0.36 < 0.34 < 0.32 < 0.30 < 0.28 < 0.26 < 0.24 < 0.22 < 0.20 < 0.18 < 0.16 < 0.14

Figure 9: The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the spring drying period (from 19th of March to 13th of May 2009). Each column belongs to one depth and each line to the particular date. The small white points depict the 16 access tubes for measuring with PR1.

95

0.2 m depth

0.3 m depth

0.4 m depth

22.7.2009

15.7.2009

9.7.2009

2.7.2009

25.6.2009

18.6.2009

3.6.2009

8.5.2009

20.5.2009

0.1 m depth

0 4 8

16

24

32

Meters Soil moisture (m3 m-3) > 0.40 < 0.40 < 0.38 < 0.36 < 0.34 < 0.32 < 0.30 < 0.28 < 0.26 < 0.24 < 0.22 < 0.20 < 0.18 < 0.16 < 0.14

Figure 10: The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the spring drying period terminated by several modest rain events and later with strong summer rainfalls which filled back the soil moisture to the level of the field capacity (from 20th of May to 22nd of July 2009). Each column belongs to one depth and each line to the particular date.

96

0.2 m depth

0.3 m depth

0.4 m depth

10.8.2010

4.8.2010

20.7.2010

12.7.2010

1.7.2010

22.6.2010

15.6.2010

11.6.20 10

4.6.2010

0.1 m depth

0 4 8

16

24

32

Meters Soil moisture (m3 m-3) > 0.40 < 0.40 < 0.38 < 0.36 < 0.34 < 0.32 < 0.30 < 0.28 < 0.26 < 0.24 < 0.22 < 0.20 < 0.18 < 0.16 < 0.14

Figure 11: The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the end of spring and summer 2010 (from 4th of June to 10th of August 2010). There is pronounced drying period initiated after refilling to the filed capacity at the 15th of June. Each column belongs to one depth and each line to the particular date.

97

5.1.3 Soil moisture temporal variability The course of the means of soil moisture measured by PR1 in four different depths is depicted in Fig.12 and Fig. 13 for the poplar plantation (mean of 16 access tubes) and the turf grass (mean of 3 access tubes) respectively. Generally, there are resembling trends of soil moisture dynamics driven mainly by very similar effective precipitation (note that poplars have higher during most of the growing season and thus higher interception can be assumed) and similar atmospheric evaporation demands. Therefore wetting and drying periods have almost identical timing for both of the covers.

Soil moisture - PR1 (m3 m-3)

0.50 0.45

Poplar plantation

0.40 0.35 0.30 0.25 0.20 0.15 0.10

0.1 m

0.2 m

0.05

0.3 m

0.4 m

0.00 3-May-08

20-Sep-08

7-Feb-09

27-Jun-09

14-Nov-09

3-Apr-10

21-Aug-10

8-Jan-11

28-May-11

Figure 12: The temporal variation of soil moisture under the poplar plantation culture during the years 2008 to 2011 measured in four different depths (0.1–0.4 m) by PR1 profile probe. The peaks in curves of soil moisture are linked with occurrence of precipitation. Likewise, the higher levels of soil moisture are caused by precipitation events (during the growing season) and by snow melting during at the beginning of spring or in winter during the thaws.

Soil moisture - PR1 (m3 m-3)

0.50 0.45

Reference lawn

0.40 0.35 0.30 0.25 0.20 0.15 0.10

0.1 m

0.2 m

0.05

0.3 m

0.4 m

0.00 3-May-08

20-Sep-08

7-Feb-09

27-Jun-09

14-Nov-09

3-Apr-10

21-Aug-10

8-Jan-11

28-May-11

Figure 13: The temporal variation of soil moisture below turf grass cover during the years 2008 to 2011 measured in four different depths (0.1–0.4 m) by PR1 profile probe. Mean soil moisture in poplar plantation (Fig. 12) varied usually between the levels close to the field capacity (sometime even reached to saturation) to the point of decrease availability (stress point or refill point). Considering the individual access tubes in the images of spatial variability (Fig. 8–11), there were also the places where the soil moisture 98

Available soil water content (mm)

reached the wilting point in all of the measured depths. Nevertheless, the spatial variability of the hydrolimits (based only on the soil sampling from three places) should be also taken into account because it can practically mean that the red areas do not have to be necessarily below wilting point level. In the case of the turf grass (Fig. 13), the soil moisture in the soil layers 0.1–0.3 m shows higher amplitudes (from near saturation to the levels below the wilting point) compared to the poplar plantation. On the other hand, the deepest layer in 0.4 m below ground was characteristic with very invariable patterns usually very close to field capacity. At the next Fig. 14, only the volumetric soil moisture up the wilting point (see the Tab. 5) was considered as the available water. Further, by this way limited soil moisture was converted into the water column (mm) which was subsequently integrated for all of the measuring depths providing only one value of the water content in the 0.45 m thick soil layer (it was simply assumed that the upper sensor in 0.1 m depth integrates soil moisture of superficial layer 0–0.15 m). Furthermore, the available soil water content patterns in the poplar plantation and the turf grass were compared. 120

Poplar plantation

100

Reference lawn

80

60

40

20

0 3-May-08

20-Sep-08

7-Feb-09

27-Jun-09

14-Nov-09

3-Apr-10

21-Aug-10

8-Jan-11

28-May-11

Figure 14: Seasonal course of mean available soil water content (here simply the water content above the wilting point) under to contrasting covers (the poplar plantation and the turf grass) measured by PR1profile probe in the integrated soil profile 0–0.45 m. At the first glance, they show very similar courses coherently fluctuated according the characteristic seasonal dynamics mentioned above. However, there can be observed, that the soil water content of the turf grass reaches usually higher amplitude. In other words it means that higher amount of water lost under the turf grass compared to the poplar plantation within the soil layer 0–0.45 m. This fact is better depicted at Fig. 15 where the values of soil water contents (not only the available because also the soil moisture below wilting point can evaporate) for 0–0.45 m deep profile were differentiated and subsequently only the reciprocal of the resulting negative values (water loss) were took into account. Finally, these values were cumulated in order to distinguish the trend in differences in the water loss of two contrasting covers. It should be noted, that the total difference between water loss from soils in the poplar plantation and the turf grass are not absolute values, because they come from the measurements carried out usually in week time step. Nevertheless, the mentioned total differences can provide estimation about 99

which cover loose more water from the 0–0.45 m deep soil layer during the course of seasons. As it is obvious, in all years was higher water loss from soil below the turf grass. 100 90

250

Poplar plantation

80

Cumulative water depletion (mm)

70

200

Reference lawn

60

150

50 40 30

100

2008

20

Total difference: 22 mm

10 0 3-May-08

31-May-08

28-Jun-08

26-Jul-08

23-Aug-08

20-Sep-08

18-Oct-08

2009

50 0 27-Mar-09

Total difference: 53 mm 1-May-09

5-Jun-09

10-Jul-09

14-Aug-09

18-Sep-09

23-Oct-09

250

140 200

120 100

150 80 100

2010

60

2011

40 50 0 25-Mar-10

Total difference: 57 mm 29-Apr-10

3-Jun-10

8-Jul-10

12-Aug-10

16-Sep-10

21-Oct-10

Total difference: 23 mm

20 0 24-Mar-11

14-Apr-11

5-May-11

26-May-11

16-Jun-11

7-Jul-11

Figure 15: Cumulative water depletion of two contrasting covers (poplar plantation and turf grass) during the years 2008–2011 measured from the upper soil layer (0–0.45 m) by PR1 profile probe. Note, that the soil moisture was measured usually once per week and therefore the values cannot be considered as an absolute water loss. On the other hand, the picture can provide good image about which cover loose more water. At the Figs. 16 and 17 is depicted the time variation of the volumetric soil moisture measured by calibrated EC-10 placed in three different depths (0.1, 0.3 and 0.9 m) in the poplar plantation for 2009 and 2010 respectively. Due to often repeated malfunctioning of three EC-10 sensors placed in the soil below the turf grass, the data were not evaluated and compared. The pictures show similar temporal variation and relation of soil moisture to the soil depth as in the previous results from PR1. However, since the EC-10 were measuring each 10 minutes, they provide uninterrupted and more detail information about the soil moisture processes; nevertheless only from one place in the plantation.

Figure 16: The seasonal course of soil moisture in poplar plantation measured by three calibrated EC-10 permanently buried into three different soil depths (0.1, 0.3 an 0.9 m) during the year 2009. The winter values are not depicted because of very high inaccuracy of measurement during low soil temperature. 100

By comparing the Figs. 16 and 17, some differences between two years of 2009 and 2010 can be observed. In the first year 2009, there was eight years old stand with maximum around 7.5 at the end of August. In the second year 2010, the plantation after winter coppicing was actually firstly only bare soil with some occasional presence of weeds. From half June the field started to be subsequently covered by foliated shoots from resprouting stumps with consecutive development of to the maximal level around 3.5 at the beginning of September.

Figure 17: The seasonal course of soil moisture in poplar plantation measured by three calibrated EC-10 permanently buried into three different soil depths (0.1, 0.3 an 0.9 m) during the year 2010. The winter values are not depicted because of very high inaccuracy of measurement during low soil temperature. Namely, the most pronounced is the difference in the intensity of the drying of the upper soil layer measured by the EC-10 in the depth of 0.1 m. Bare soil without any coverage seems to logically loose more water by soil evaporation. On the other hand, the moisture in deeper layers is not so much variable during 2010, mainly due to limited water uptake compared to 2009. However, there is some ongoing water loss caused mainly by roots of weeds, subsequently resprouting stumps with later developing shoots and also by the capillary rise which refill the soil deficit in the superficial soil layer exposed to solar radiation and high evaporative demand of the atmosphere. The next Fig.18 and Fig. 19 depicts the precipitation amounts measured by rain gauge and in the opposite the water loss derived from EC-10 during both of the consecutive years 2009 and 2010 respectively. The values of water loss were calculated in the similar way as in case of PR1. In detail, the soil moisture was transformed into the soil water column (mm) integrating the soil profile 0–1 m in depth. Further, the diurnal maxima of soil water content were differentiated. In this way the differences of soil water column between the particular days were obtained. Basically, these differences mean that if the values are positive, it is the soil water income, and if they are negative, it is the soil water loss.

101

70

Water column (mm day-1)

60 50

2009

Water loss

Precipitation

40 30 20 10 0 -10 -20 -30 Jan-2009 Feb-2009 Mar-2009 Apr-2009 May-2009 Jun-2009

Jul-2009 Aug-2009 Sep-2009 Oct-2009 Nov-2009 Dec-2009

Figure 18: Comparison of precipitation measured by rain-gauge (blue columns) and the water loss (red columns) calculated as the total diurnal depletions of soil moisture measured by EC-10 in three depths (0.1, 0.3 and 0.9 m) integrated for the soil profile 0–1 m. The graph depicts the water balance conditions in the poplar plantation during 2009. 70

Water column (mm day-1)

60 50

2010

Water loss

Precipitation

40 30 20 10 0 -10 -20 -30 Jan-2010 Feb-2010 Mar-2010 Apr-2010 May-2010 Jun-2010

Jul-2010 Aug-2010 Sep-2010 Oct-2010 Nov-2010 Dec-2010

Figure 19: Comparison of precipitation measured by rain-gauge (blue columns) and the water loss (red columns) calculated as the total diurnal depletions of soil moisture measured by EC-10 in three depths (0.1, 0.3 and 0.9 m) integrated for the soil profile 0–1 m. The graph depicts the water balance conditions in the poplar plantation during 2009. The last picture of this section (Fig. 20) shows the comparison of water income recorded by rain gauge – precipitation, the soil water income obtained by summing all the water income estimated from EC-10 measurements and finally the analogically estimated water loss. Note that the changes after soil freezing were not took into account because they are linked with changing the relative permittivity of water due to changing phase (from liquid to solid), not with changing soil moisture. Considering the whole year, the slightly differences confirm the validity of the soil moisture measurements and the used approach itself.

102

750 740

680 670

Soil water loss

690

Soil water income

700

Precipitation

710

Soil water loss

720

Soil water income

730

Precipitation

Water column (mm year-1)

760

660

2009

2010

Figure 20: The comparison of yearly totals of precipitation measured by rain-gauge, total soil water income and total soil water loss measured by EC-10 in three depths (0.1, 0.3 and 0.9 m) integrated for the soil profile 0–1 m.

5.2 Evapotranspiration The following section deals with the most important part of the water loss at the investigated locality. Before we start to evaluate the absolute numbers of evapotranspiration of the poplar plantation and the reference turf grass, let´s have a look briefly into the possible source of errors in the employed BREB systems, with this linked data rejecting and subsequent gap filling and of course on the limitation related to the footprint which are inherent for all of the micrometeorological methods. 5.2.1 BREB error analysis Prior the evaluating of the possible errors and carrying out the data discarding, it is useful firstly to look on the magnitudes and frequency distribution of the air temperature and humidity differences as a rough output of the gradient measurements. Firstly, Fig. 21 depicts the frequency distribution of temperature differences during the growing seasons from the beginning of the BREB measuring campaign to the end of 2010. The half hourly data are divided into two categories according to net radiation intensity. Namely, when Rn > 50 W m-2 and when Rn < 50 W m-2. The first category comprises the diurnal data which are the most important for evapotranspiration and the second one includes the rest – mostly the nocturnal periods, mornings and evenings. Note that the vertical distance between the sensors (both air temperature and humidity) was adjusted as 1.6 m for all of the time in case of the turf grass. The vertical distance of sensors above the poplar plantation was 2 m during 2008 and 2009 and 3 m during 2010. The higher vertical distance during 2008 and 2009 was not unfortunately possible due to the technical limitation linked with the height of the mast. At the first glance it can be observed that the data are generally quite normally distributed. By dividing the median value by the vertical distance between the sensors, the particular gradients (the change of variable with height) can be obtained. Although the frequency distributions of the gradients are useful for general idea, they were not intentionally depicted. Instead of that, the real measured differences, which are very informative from measurement point of view, are shown. 103

18 16

18

Turf grass 2008 - 2010

16

14 12

Frequency (%)

10

18

Poplar plantation 2008 - 2009

median = -0.46 n = 10812

Rn > 50W m-2

14

12

Rn > 50W m -2

median = -0.21 n = 6021

10

12

8

8

6

6

6

4

4

4

2

2

2

0

0

18

16

-2.4

-1.6

-0.8

0

0.8

1.6

Turf grass 2008 - 2010

16

14

0 -3.2

18

-2.4

-1.6

-0.8

0

0.8

1.6

median = 0.48 n = 10605

10

Poplar plantation 2008 - 2009

2

2

2

0

0 1.2

1.6

2

2.4

1.6

Rn < 50W m-2

6 4

0.8

0.8

8

Rn < 50W m-2

6 4

0.4

0

median = 0.50 n = 3613

10

4

0

-0.8

12

8

-0.4

-1.6

Poplar plantation 2010

16

median = 0.36 n = 11596

10

Rn < 50W m-2

6

-2.4

14

12

8

-3.2 18

14

12

Rn > 50W m -2

median = -0.77 n = 3639

10

8

-3.2

Poplar plantation 2010

16

14

0 -0.4

0

0.4

0.8

1.2

1.6

2

2.4

-0.4

0

0.4

0.8

1.2

1.6

2

2.4

Temperature difference ( C): upper - lower sensor

Figure 21: Frequency distribution of air temperature differences above different covers. The vertical distance between the sensors above the turf grass was 1.6 m. In case of the adult poplar plantation in 2008 and 2009 the distance was adjusted to 2 m and during the season 2010 with resprouting poplars the distance was 3 m. 18 16

Turf grass 2008 - 2009

18

Rn > 50W m-2 16

14

14

median = -0.08 n = 10812

12

Frequency (%)

10

18

Poplar plantation 2008 - 2009

Rn > 50W m-2 16

median = -0.03 n = 6021

12

10 8

8

6

6

4

4

4

2

2

2

0

0 -0.08

0

0.08

18

16

Turf grass 2008 - 2009

12

16

-0.16

-0.08

0

0.08

16

10

10

8

8

8

6

6

6

4

4

4

2

2

2

0

0 -0.16

-0.08

0

0.08

0.16

-0.08

0

0.08

Rn < 50W m-2

median = 0.03 n = 3613

12

10

-0.16

Poplar plantation 2010

Rn < 50W m-2 14

median = 0.002 n = 11596

12

-0.24 18

Poplar plantation 2008 - 2009

Rn < 50W m-2 14

median = 0.02 n = 10605

14

0 -0.24

18

median = -0.03 n = 3639

10

6

-0.16

Rn > 50W m-2

12

8

-0.24

Poplar plantation 2010

14

0 -0.16

-0.08

0

0.08

0.16

-0.16

-0.08

0

0.08

0.16

Vapour pressuredifference (kPa): upper - lower sensor

Figure 22: Frequency distribution of air humidity differences above different covers during the growing seasons 2008–2010. The vertical distance between the sensors was the same as for the air temperature at Fig. 21. Generally, the highest temperature diurnal (when Rn > 50 W m-2) gradients were typical above the turf grass with median -0.29 °C m-1 and then above the poplar plantation during the growing season 2010 after harvest with median -0.26 ° C m-1. The lowest temperature gradients were observed above the high-grown poplar stand (~12 m) in 2008 and 2009 with median -0.1 °C m-1. The ―nocturnal‖ temperature gradients (Rn < 50 W m-2) were inverse and generally higher. The highest was again above the turf grass with median 0.3 °C m-1. Thereafter, in case of the poplar plantation, there was observed opposite order than in the diurnal gradients. Namely median 0.18 °C m-1 for the poplar plantation in the 104

7th and 8th year of the first rotation and median 0.16 °C m-1 for the poplar plantation during the year 2010. The humidity gradients were greatest above the turf grass with medians 50 and 13 Pa -1 m for diurnal and ―nocturnal‖ periods respectively. Then in case of the poplar stand during the 2008 and 2009 the medians were 15 and 1 Pa m-1 for diurnal and ―nocturnal‖ periods respectively. Finally, for young poplar plantation during the year 2010 were the medians for both time periods 10 Pa m-1. There are many other works, in which the BREB method is used, where in order to avoid serious errors in the estimation of the fluxes, the data within the instrumental errors of the Bowen ratio system are excluded (PEREZ at al., 1999). Considering the maximum mutual errors of pairs of sensors used in this study reaching to 0.3 °C and 50 Pa (expressed as ) for air temperature and humidity respectively, this very conservative approach could have very drastic impacts on the data measured in this study. Taking into account only the growing period, more than 35 % and 75 % of the data from the turf grass would be discarded for diurnal and ―nocturnal‖ periods respectively. In case of the adult poplar plantation, more than 75 % and 85 % of data would be rejected for diurnal and ―nocturnal‖ periods respectively. Finally, for the young resprouting culture, more than 65 % would be discarded for both of the time period. In all case, the high number of deleted data is mostly especially due insufficient accuracy of the air humidity measurements. Nevertheless, PEREZ et al. (1999) and later GUO et al. (2007) proposed analytically based methodology for data rejecting which is based on the knowledge of accuracy of the particular sensors and actual vapour pressure. However, as a preliminary condition, the two step inequality given originally by OHMURA (1982) which ensure the right flux sign should be also always used. Fig. 23 depicts the distribution of rejecting data according to limits given by GUO et al. (2007) already filtered by right flux sign condition. The data which are situated inside the area restricted by the red ―cross‖ have to be rejected using the capacitance humidity and resistance temperature paired sensors EMS33 with the mentioned mutual error below 0.3 °C and 50 Pa for air temperature and humidity respectively. The picture also shows the area of rejections for more precise system based on dew point hygrometer and thermocouples with errors below 0.02 °C and 20 Pa for the gradients of air temperature and humidity respectively. Note that now only the fluxes with the right flux sign according to the condition given by OHMURA (1982) have been already considered. By dividing the Fig. 23 by the cross into four quadrants, it can be noticed that only three situations are real. Namely, evaporation during unstable conditions when both latent ( ) and sensible ( ) heat fluxes are positive (the most of the diurnal conditions). Then evaporation during stable conditions when is positive but negative (situation during the day just after precipitation or during the time of sunset in inversion conditions which can persist throughout the night until sunrise). The last is the period of condensation during the stable condition when both and are negative.

105

β

15 13 11 9 7 5 3 1 -1 -3 -5 -7 -9 -11 -13 -15

LE > 0 H >0

LE < 0 H 0 H 50 W m-2) median in 2008 for the poplar plantation (brown points) and the turf grass (green points).

Figure 47: Seasonal course of Bowen ratio diurnal (Rn > 50 W m-2) median in 2009 for the poplar plantation (brown points) and the turf grass (green points). The last of this group, Fig. 48 shows the situation after the winter coppicing during the growing season 2010. The Bowen ratio of the turf grass without any higher values are indicating no water stress and not reduced stomatal control. Firstly slightly higher values around 0.8 in April are probably just due to beginning of the physiological activity. Then during the rest of the growing season is Bowen ratio fluctuating close around the mean 0.46. The Bowen ratio of the poplar plantation during the first part of the growing season is very variable and strongly dependent on the soil moisture of the surface layer which is closely linked with the precipitation occurrence and the rates of evaporation itself. The mean of the medians within this period is equal to 2.89. After the initiating of the leaf area development, the Bowen ratio become more stable and lower with the mean of medians 1.32.

139

Figure 48: Seasonal course of Bowen ratio diurnal (Rn > 50 W m-2) median in 2010 for the poplar plantation (brown points) and the turf grass (green points).

5.3.2 Surface resistance To obtain better information about the stomatal control, the rewritten form of the Penman-Monteith equation can be used to obtain the surface resistance as the unknown variable if the resulting measured by BREB is given. Moreover, the surface resistance can be easily transformed to stomatal conductance if the is known. The importance of the surface resistance lays also in the fact, that it is a very important parameter for many water balance models based on the Penman-Monteith. Within this context, it is very important to know, that it does not really behave like any constant as we can see in the following pictures. Generally, the temporal variation is driven by the same rules explained for the Bowen ratio values. Due to the similar variable character as Bowen ratio during the day, again only the medians from the diurnal part limited by the net radiation threshold 50 W m-2 were used. First Fig. 49 depicts the second half of the growing season 2008. In case of turf grass, the mean surface resistance for July and August is 96 s m-1 and for September 165 s m-1. The poplar plantation had generally lower values with the overall mean 85 s m-1 and with the lowest monthly mean 72 s m-1 in July. Growing season 2009 is depicted at Fig. 50. The overall mean was 65 and 80 s m -1 for the turf grass and the poplar plantation respectively. By omitting the values from April, the mean surface resistances 61 and 70 s m-1 will be obtained for the turf grass and the poplar plantation respectively. If we omitted also the higher values linked with decrease in of the poplar plantation at the turn of the July and August we will get on the level near 65 s m-1.

140

Figure 49: Seasonal course of surface resistance diurnal (Rn > 50 W m-2) median in 2008 for the poplar plantation (brown points) and the turf grass (green points).

Figure 50: Seasonal course of surface resistance diurnal (Rn > 50 W m-2) median in 2009 for the poplar plantation (brown points) and the turf grass (green points).

The last from the figures focusing on the surface resistance (Fig. 51) shows the year 2010 with its typical differences. The surface resistance of the turf grass was not significantly changed with the mean reaching up to 67 s m-1. There is only noticeable the period at the turn of the June and July characteristic with very hot weather without precipitation causing very fast depletion of the soil moisture. The overall mean for the poplar plantation was 158 s m-1. By taking into account only the second half of the growing season linked with the development and the vigorous growth of shoots, the -1 mean value 148 s m will be obtained.

141

Figure 51: Seasonal course of surface resistance diurnal (Rn > 50 W m-2) median in 2010 for the poplar plantation (brown points) and the turf grass (green points).

5.3.3 Decoupling coefficient Finally, by knowing the surface resistance and the aerodynamic resistance which can be estimated from the wind speed and height of the canopy, so-called decoupling factor (Ω) can be calculated according to MCNAUGHTON and JARVIS (1983) in the form for amphistomatous species (species with stomata situated on both the adaxial and abaxial surfaces of leaves). The omega factor describes the stomatal control over water transpired by vegetation. It ranges between 0–1. An omega factor more close to unity indicates that transpiration is mainly controlled by available energy (decoupled system), whereas a small omega value indicates that transpiration is mainly controlled by canopy conductance and vapour pressure deficit (well-coupled system). Note that we are operating here with the evapotranspiration instead of transpiration and thus the omega factor provides only a rough estimation which on the other hand enables us to compare two different cultures. At Fig. 52 the time course of the omega factor for the second part of the growing season 2008 is depicted. The overall mean for this period was 0.78 and 0.41 for the turf grass and the poplar plantation respectively. The large short time variation can be explained to be driven by the same mechanisms mentioned in the surface resistance and Bowen ratio variation. Note that also the omega factor was calculated as a median from the diurnal part when the net radiation was higher than 50 W m-2.

142

Figure 52: Seasonal course of decoupling coefficient diurnal (Rn > 50 W m-2) median in 2008 for the poplar plantation (brown points) and the turf grass (green points).

Figure 53: Seasonal course of decoupling coefficient diurnal (Rn > 50 W m-2) median in 2009 for the poplar plantation (brown points) and the turf grass (green points). Next Fig. 53 shows the temporal variation of the omega factor in the growing season 2009. The overall mean was 0.85 and 0.43 for the turf grass and the poplar plantation respectively.

143

Figure 54: Seasonal course of decoupling coefficient diurnal (Rn > 50 W m-2) median in 2010 for the poplar plantation (brown points) and the turf grass (green points). The last of these three pictures dealing with the omega factor (Fig. 54.) shows the situation within the characteristic growing season 2010. Note that before foliating the omega factor of the poplar plantation was closer to level of the turf grass. The mean values for 2010 were 0.84 and 0.51 for the turf grass and the poplar plantation respectively. By taking into account only the values from second half of the growing season, the omega factor for the poplar plantation reached up to 0.47 whereas that of the turf grass to 0.86. 5.3.4 Crop coefficient Crop coefficients ( ) provide simple way to estimate crop evapotranspiration ( ) from weather-based reference values. is simply defined as a ratio between the crop evapotranspiration ( ), not limited by available water, and the calculated . One may ask why not to use the measured by BREB above the turf grass which has very similar parameters (height, albedo and probably also stomatal conductance) and taking into account the data without any water limiting conditions it could serve as good source of the reference evapotranspiration values. Moreover, such comparison has been already done and the ratios between of the poplar plantation and the turf grass could be logically used as the crop coefficients (or possibly the slopes of the linear regressions as well). The problem is that the over different grass surfaces based on different species could be quite variable. Another reason could be the differences in used methods for determining the based on measurements. Therefore, it is very suitable to use the calculated reference evapotranspiration as just only one specific method for all cases when identifying the across varying conditions and do not combine it with different approaches. As a good argument, why do not to use the measured of turf grass instead of the calculated according to united methodology is the comparison at Fig. 55. where the measured of the turf grass is generally by 16 % higher compared to the calculated according to FAO-56 methodology.

144

Actual evapotranspiration (mm) - BREB

6

y = 1.16x R² = 0.96 5

4

3

2

1

0 0

1

2

3

4

5

6

Reference evapotranspiration (mm) - FAO-56

Figure 55: The relationship between the daily sums of the reference evapotranspiration calculated according to FAO-56 methodology (ALLEN et al., 1998) and the actual evapotranspiration measured by BREB above the turf grass. The whole dataset from June 2008 to the end of 2010 was used including the winter time values. Fig. 56 shows the varying crop coefficient based on the ratio between measured of the poplar plantation and the according to FAO-56 calculated from the data from standard meteo-station placed at the turf grass (which is practically BREB system 2 at Fig. 4). For determining the daily values of only the data from April to October were used in order to catch the whole period when poplars are foliated. The season 2008 was not evaluated simply because it was not complete. The crop coefficients obtained in this way were consequently transformed according to Eq. 127 – rewritten form of equation given by ALLEN et al. (1998), in order to reflect the difference in relative humidity and wind speed conditions (the validity of suppose that the minimum daily relative humidity is 45 % and mean daily wind speed 2 m s-1 in average). This adjustment made the final values in average by 10 % higher. At the first glance, there can be seen large scatter of the values. Following the hypothesis that the scatter and especially the high positioned values are caused by the effect of precipitation subsequently causing higher evaporation rates of intercepted water, the values were divided into three groups. Namely, the days when any precipitation was measured, one day after precipitation, and the rest – two or more days after precipitation events.

145

3.0

2009 2.5 2.0 1.5 1.0 0.5 0.0

April

May

June

July

August September October

3.5

Daily crop coefficient Kc

2010 3.0

resprouting canopy 2.5

bare soil 2.0

1.5 1.0

0.5 0.0 April day with rain

May

June

July

one day after rain

August September October two and more days after rain

Figure 56: The time course of daily in 2009 and 2010. Different colours indicate the differences between the days with and days after precipitation events. In 2010, the coppice was in the first year of the first rotation and in the first half of season, there was only bare soil with occasional weeds. The beginning of resprouting of new shots is marked by the dashed blue vertical line. From Fig. 56, it seems that the hypothesis was correct and the high values are really the effect of precipitation events. To deal with this issue it was suggested to derive the from days without rain and the effect of precipitation take as an additional amount of evaporated water. It rather seems to be more physically based than just create the from all data which would be very dependent on the precipitation regime of particular period and can vary from year to year. The Fig. 57 gives the picture of varing ten days mean of daily according to recomendation of ALLEN et al. (1998) for the data with irregular wetting. Note that the data from days with or from the first days after precipitation were ommited, nevertheless even here quite substantial time variation can be observed.

146

Ten-day crop coefficient Kc

1.6

1.4

1.2

1.0

0.8

2009 - eight years old stand 2010 - bare soil

0.6

2010 - resprouting canopy 0.4

April

May

June

July

August

September October

Figure 57: Time course of ten-day averaged in 2009 and 2010 determined by leaving out the days with and first days after precipitation events. At both of the above pictures we can follow firstly decreasing trend and from the beginning of July increasing trend of in 2010. This feature is probably linked with high soil moisture levels and lower evaporating demand at the beginning of growing season leading to very comparative evaporation from the bare soil and calculated . The soil moisture level was generally very high during the whole spring, but in combination with increasing solar radiation, day length and air temperature, the atmospheric evaporation demand increased and bare soil was not able further to evaporate as much as the reference cover. In addition, at the end of June, the soil moisture started to deplete which caused further decrease of . Nevertheless, at the turn of the June and July this situation culminated and in the same time the resprouting shoots started to play important role by their transpiration leading to increase of close to unity within one month. The next Fig. 58 shows the constructing of the curve for particular crop growth stages according to ALLEN et al. (1998). For the construction of these curves again only the values from at least two days after precipitation and moreover only the values when the were higher than 2 mm per day were used. This additional filter was used due to the fact that the low values of evapotranspiration from both methods are generally more error prone. The core of the initial and mid-season curve was always the mean of values during the particular period and the crop development stage just simply the connecting line between them. The mean value of the initial period 2009 was 0.66, then it was linearly increasing during the May crop development stage to the level of 1.2 as a mean for the whole mid season. The late season is not evaluated since there were not enough reliable values. But by taking into account the low evaporative demand and high soil moisture we can assume that the value was not far from the unity. If we use simply the value one for the rest of the whole months, it would be not far from the truth. Moreover, taking into account the low magnitude of the absolute values during this time, it cannot cause large errors. In 2010, the mean of the initial period was 0.75. However, the was linearly decreasing from 1.05 to 0.45. After that at the end of June it turned to increase within the crop development stage till the end of July up to the level of 0.97 as the mean of the mid-season period. In case of the late-season 2010, the same reasons of its absence and the same assumptions as in 2009 were used. 147

2.0 1.8

2009 mid-season

late-season

1.6 1.4

1.2

initial

1.0 0.8

0.4 0.2 0.0

April

May

June

July

August

September

2.0

1.8

2010

1.6 1.4

mid-season

late-season

Crop coefficient Kc

0.6

1.2

1.0 0.8 0.6 0.4

0.2 0.0

April

May

June

July

August

September

Figure 58: The curve (solid green line) constructed for particular crop stages according to ALLEN et al. (1998). To construct the curve, only the values at least from 2nd day after rain and only the values when was higher than 2 mm were used. It provides more accurate values of and suppresses the undesirable effects of interception, dew and fog. The brown dashed line expresses the in case that the soil surface would be dry during the initial stage which is characteristic with dominant soil evaporation. In crop development stage, transpiration becomes leading and soil evaporation negligible. To back calculate of the poplar plantation from and above derived curves, two assumptions were used. Firstly, to cover the additional effect of interception, the following relation according to IRITZ et al. (2001): Imax = 0.2 PAI, where Imax is maximal interception capacity (mm) given by plant area index ( ) which (if filled by rain) is added to the product of and , was used. Secondly, the amount of evaporable water given by is not unlimited. By setting as the upper limit the potential evaporation from wet canopy using the Penman-Monteith equation (MONTEITH, 1965) when stomatal resistance was set to zero, it gave unreasonably high values. Therefore, the Priestley-Taylor potential evaporation equation assuming that the evaporation of the intercepted water was governed only by the available energy (IRITZ et al., 2001) was used as the upper limit of daily of the poplar plantation. To validate the assumptions given above, the constructed curves were used and applied together with the assumptions linked with interception on , and precipitation data. The Fig. 59 shows good agreement between and and confirms 148

the correctness of the provided assumptions and whole calculating procedure. However, for more generality of this approach, comparison with independent measurements is necessary. Cumulated evapotranspiration (mm)

600

500

2009

2010

400

300 ETa act ET0

200

ETc c 100

0

Apr

May

Jun

Jul

Aug

Sep

Oct

Apr

Figure 59: The comparison of measured cumulative way.

May

Jun

, calculated

Jul

Aug

and

Sep

Oct

, all expressed in

5.3.5 Sap flow To get more detail image about the transpiration at tree level, the sap flow rates of three sampled trees in 2008 and four sampled trees in 2009 were investigated. Due to the fact that the tissue heat balance method (ČERMÁK et al., 2004) does not enable to analyze the trees practically within the range of 2 to 10 cm in , only the sap flow of the most dominant trees (see Fig. 60) was possible to be measured. Note that KUČERA et al. (2011, unpublished data) have been recently testing and developing the using of tissue heat balance method on poplars with dimensions of 5–10 cm in which could be very crucial for assessing sap flow of SRC cultures like that in this study. in total: 8,610 trees per 1 ha

1200 1000

Sap-flow measurements

Number of trees per 1 ha

1400

800 600 400 200

0 15

25

35

45

55

65

75

85

95

105

115

125

DBH (mm)

Figure 60: The frequency distribution of breast height diameters of trees in the stand expressed in the absolute numbers per ha. The diameter range of trees measured by tissue heat balance method is marked in the dashed red quadrangle. 149

Next two Figs. 61 and 62 show the diurnal variation of sap flow of selected ten days periods during 2008 and 2009 respectively. At Fig. 61 we can see the September patterns of three trees with very similar ranging from 103 mm to 110 mm. Although the dimensions of the trees are almost identical, large variation in their sap flow rates can be observed. The maximum rates of sap flow reached up to almost 5 kg of water per hour and 34 kg of water per day at the 10th of September by tree with 110 mm in . This tree showed generally the highest values during the whole period (September 2008) with 555 kg per month. The total sap flow within the September for the two remaining trees reached up to 496 and 407 kg of water respectively. 5.0 4.5

DBH - 103 mm

2008

DBH - 111 mm

DBH - 110 mm

4.0 3.5 3.0 2.5 2.0 1.5

1.0 0.5 0.0 31.8.08

1.9.08

2.9.08

3.9.08

4.9.08

5.9.08

6.9.08

7.9.08

8.9.08

9.9.08

10.9.08

11.9.08

12.9.08

13.9.08

14.9.08

15.9.08

16.9.08

17.9.08

18.9.08

19.9.08

23.9.08

24.9.08

25.9.08

26.9.08

29.9.08

30.9.08

5.0 4.5

Sap-flow (kg hour -1)

4.0 3.5

3.0 2.5 2.0 1.5 1.0 0.5

0.0 5.0 4.5 4.0 3.5

3.0 2.5 2.0 1.5 1.0 0.5

0.0 21.9.08

22.9.08

27.9.08

28.9.08

Figure 61: Selected periods during 2008 with diurnal courses (half-hourly) of sap flow of three sampled poplar trees belonging into one class (10–11 mm). Note that the given were measured at end of the growing season. The shape of the diurnal curves characteritic with the peak before noon (usually between 9:30 to 11:30) and subsequent decline can indicate water stress and reducing of the stomatal aperture. During the two rainfalls at the 15th and 16th of September there were also zero flux of water upward the stem caused by almost zero vapour pressure deficit and 150

wet leaf surfaces. After these days also the soil moisture increased and the characteristic noon declinig of sap flow had not been already observed. 5.0

4.5

DBH - 122 mm DBH - 106 mm DBH - 123 mm DBH - 119 mm

2009

4.0 3.5 3.0 2.5

2.0 1.5 1.0

0.5 0.0 1.7.08

2.7.09

3.7.09

4.7.09

5.7.09

6.7.09

7.7.09

8.7.09

9.7.09

19.7.09

20.7.09

21.7.09

22.7.09

23.7.09

24.7.09

25.7.09

26.7.09

27.7.09

28.7.09

29.7.09

30.7.09

31.7.09

1.8.09

2.8.09

3.8.09

4.8.09

5.8.09

6.8.09

7.8.09

31.8.09

1.9.09

2.9.09

3.9.09

4.9.09

5.9.09

6.9.09

7.9.09

8.9.09

9.9.09

5.0 4.5 4.0

3.5 3.0 2.5 2.0 1.5

Sap-flow (kg hour -1)

1.0 0.5 0.0 5.0

4.5 4.0 3.5

3.0 2.5 2.0 1.5 1.0

0.5 0.0 5.0 4.5

4.0 3.5 3.0

2.5 2.0 1.5

1.0 0.5

0.0

Figure 62: Selected periods with diurnal courses (half-hourly) of sap flow of four sampled poplar trees with slightly different dimensions ( measured at the end of the growing season) during 2009. Note that the vertical levels of the picture represent ten days time periods which are not consecutive. The tree depicted by red colour is the only tree measured also in the previous year 2008 (the red one in Fig. 61). Fig. 62 shows the diurnal sap flow rates of four sampled trees during the selected ten days periods within 2009. There is one tree with considerably different lower sap flow 151

rates. Looking at the stem inventory scheme, this tree was surrounded within the raw by four trees with 113.8 mm in from the east side, no tree from the south-east, tree with 111.5 mm in from the south side, trees with 94.2 and 43.1 mm in from the nd south-west and west side respectively (taken from summer inventory at 22 of July 2009). The presence of three other big trees in the close vicinity could probably lead to the mutual competition for water, light and nutrients which could cause the systematically lower values of sap flow. The maximum peak was recorded at 27th of July reaching up to 5.34 kg of water per hour by tree with 119 in . The maximum daily sum of sap flow 44 kg per day was st recorded at the 1 of August by the tree with 106 mm in . By contrast, the maximum diurnal sap flow peak of the least transpiring mentioned tree was 2.8 kg per hour (28th of August) and maximum daily sum was 23.9 kg per day (1st of August). The mean daily sap flow of all sampled trees during the whole measuring period (from 18th of June till 31st of October) was 15.6 kg per day. The total amount of water during the same period reached up to 2517 kg for the most transpiring of the sampled trees with the 123 mm. The remaining trees with respective 106, 119 and 122 mm amounted in the same order to 2441, 2300 and 1525 kg of total during the investigated period. The noon decline of sap flow has been observed only sporadically in case of tree with 119 mm in . Looking back to the inventory, this tree was positioned between the other five trees in the closest distances within the raw with 50.7 and 48.3 mm in from the east and south-east side respectively, then with 107.2 mm in from the south side and with 88.2 and 48.4 mm in from the south-west and west side respectively. The smaller trees from the east and south east side could be the possible cause of the high sap flow values during the morning. On the other hand the big trees situated at the south and south-west side from the sampled individual could be the reason why the sap flow declined during the noon and afternoon. For the rest of trees the well bell-shaped diurnal sap flow curves were typical during the nice sunny days indicating well water supply. Just for comparison, the first tree with 123 mm in (the most transpiring) was adjacent within the raw to trees with 58.3 and 74.1 mm in from the east and south-east side respectively. No tree from the south and trees with 85.6 and 49.7 in from the south-west and west side respectively. The second tree with 106 mm in neighboured with tree with 81.6 mm in from the east side. No tree from the south-east, and trees with 93.1 and 64.1 mm in from the south-west and west side respectively within the raw. As it can be seen at Fig. 60 depicting the frequency distribution of and indicating the class treated by sap flow measurement, it is very difficult to upscaled the sap flow to the whole stand from the measuring of only four trees in addition with almost same . To deal with this problem, 30 trees were not cut during the winter harvest in early 2010. Supposing that the leaf mass versus relation would not dramatically change during the five months when the trees grew in different conditions, not surrounded by other trees in high density culture, the destructive measurement of leaf mass of all remaining 30 trees with different (ranging from 38.2 to 152.1 mm) was carried out at the end of August 2010. To get the information about the vertical distribution of the leaf mass, each tree was divided into particular levels with one meter distances and the weights of leaves in 152

these levels was assessed individually. Further, from all of the trees for the previous year 2009 was estimated from annual rings measurement supposing that the thickness of bark remained the same. Finally the power function describing the relation between and leaf mass (Fig. 63-a) was parameterized. For simplicity and straightforwardness the leaf mass was not converted into for these scaling purposes. 12

6.0

LM active = 0.33*10-11 * DBH5.9157 r2 = 0.93

5.5

Active leaf mass (kg of fresh weight)

Leaf mass (kg of fresh weight)

LM = 0.4*10-5 * DBH3.08182 r2 = 0.97 10

a) 8

6

4

2

5.0

b)

4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5

0

0.0 20

40

60

80

100

120

140

DBH (mm)

20

40

60

80

100

120

140

DBH (mm)

Figure 63: a) on the left side the leaf mass (fresh weight) related to the of 30 destructively sampled trees. b) on the right side, the active leaf mass (fresh weight) again related to the of the 30 sampled trees. The solid red curves depict the regression lines and the dashed black lines define the 95 % intervals of confidence. By assuming that not whole leaf area or leaf mass of the tree is transpiring equally, the Beer´s law (LARCHER, 2003) was used to express the radiation attenuation through the canopy leading to the downward decreasing of transpiration from the top. It was simply assumed that transpiration is decreasing in the same way as the solar radiation does. As the extinction coefficient the value of 0.6 from other measurement of by SunScan was used. Additionally, the values of leaf mass from particular levels by 3.28 m2 kg-1 (leaf are divided by the weight of fresh leaves) were converted to obtain the in particular levels just for modelling the radiation attenuation in canopy. The resulting relative values of radiation (the ratio between the radiation penetrated into the canopy and the radiation incident above the canopy) were used to multiply the leaf mass in different height levels to obtain so-called active leaf mass (the leaves which are receiving the solar energy and thus are transpiring and photosynthesizing). Finally, a power function was parameterized to solve the relationship between and active leaf mass (Fig. 63-b). Assuming a linear relation between the weight of leaves and the transpiration based on the very near to linear relationship between and transpiration (VERTESSY et al., 1995), this parameterized function (Fig. 63-b) was used for up-scaling the sap flow from the sampled trees to the whole stand. By using the stem inventory we calculated that the mean active leaf mass of the whole stand is 0.65 kg m-2. Then the of the four sampled trees was used to estiamte their mean active leaf mass as 5.28 kg. By dividing the mean active leaf mass of the stand by the mean active leaf mass of the sampled trees, the scaling coefficient 0.12 was obtained. It enabled to convert the mean sap flow of the four sampled tress to the stand

153

sap flow (kg m-2), or taking into account the daily time step, it can be considered as the stand transpiration in mm (Fig. 64). 7

Transpiration - belt of confidence Estimated transpiration

Water loss (mm day-1)

6

Transpiration

5

Actual evapotranspiration 4 3 2 1 0 01-Apr-09

29-Apr-09

27-May-09

24-Jun-09

22-Jul-09

19-Aug-09

16-Sep-09

14-Oct-09

Figure 64: Temporal variation of the measured by BREB (black dashed line) and the transpiration upscaled from sap flow measurements (blue solid line) with the belt of confidence. Because the measurements of sap flow started on the 18th of June, the previous period was estimated (dashed blue line) by using the linear relation between the upscaled transpiration and the magnitude of the leaf area index. In order to calculate the transpiration during the whole year 2009, the transpiration before the start of the sap flow measurement had to be estimated. To deal with this issue the assumption that the transpiration is the function of leaf area was adopted. It seems that during the period with around 7 and with low soil evaporation (August), the transpiration comprised almost 100 % of the measured by BREB. For the rough estimation of the spring transpiration the linear decrease of transpiration from the maximum 7.5 to zero as 95 % to 0 % of the measured by BREB was used. th Taking into account that the at the 18 of June when the sap flow measurement was initiated was 4.8, the back decreasing from transpiration to evapotranspiration ratio 0.62 was used. At this moment, the effect of changing soil moisture and with this linked soil evaporation was not considered for the transpiration estimation. By using this final estimation total yearly sum of transpiration resulted in 346 mm. The total during the th th transpiration period (13 of April to 14 of November) was 495 mm and thus the transpiration comprised 70 % of the whole . 5.3.6 Biomass growth and water-use efficiency This section starts with variable called leaf are index ( ) which has already been mentioned many times in the previous text in conjunction with , transpiration and other variables. Next Fig. 65 shows the seasonal variation of within two consecutive years 2009 and 2010 respectively. As already mentioned, the season 2010 was very specific, because it was the start of the second rotation period. In 2009, the leaf area development initiated with budburst at the 13th of April. The culminated at the turn of August and September with peak value 7.48. The last leaves were finally shed at the beginning of November. Note that at the end of July and beginning of August there can be observed slight decline of values which could be the effect of the two strong winds 154

(15th July and 23rd July) mentioned previously. During the second year 2010, the first small leaves on the resprouting stumps started to appear at the beginning of June. However, very progressive growth of the shoots took place mainly in July. The again culminated at the turn of August and September reaching up to 3.7. It is evident, that during 2010 the variability of the was very high (expressed as the standard deviation). The main reason is that the canopy was not closed during the whole growing season and therefore there were very big differences between the in rows and inter-rows.

LAI measured by SunScan

9 8

2009

2010

7 6

5 4 3

2 1 0 1-Apr 22-Apr 13-May 3-Jun 24-Jun 15-Jul 5-Aug 26-Aug 16-Sep 7-Oct 28-Oct

1-Jun

22-Jun 13-Jul

3-Aug 24-Aug 14-Sep 5-Oct

26-Oct

Figure 65: Temporal variation of during two contrasting years 2009 and 2010. The green line depicts the mean and the dashed error bars express the standard deviation. To validate these results, the by litter collection method during five consecutive weeks in autumn 2009 was independently measured. It was carried out at three places in the centre of the positions, where the by SunScan was measured regularly. The scatter view with linear regression fit at Fig. 66 shows good agreement between the direct and indirect methods and underlines the validity of the above mentioned results.

LAI measured by SunScan

7 y = 1.07x - 0.14 R² = 0.96

6 5 4 3 2 1 0 0

1

2

3

4

5

6

7

LAI measured by litter collection

Figure 66: The comparison of the measured from litter collection (horizontal axis) and measured by SunScan. The solid dashed line depicts the linear regression line and the dashed black line depicts the slope of 45°. The other compound of the above-ground biomass is wood biomass which comprises stems and branches. During the harvest in 2010 the allometric measurements took place and the relations between many tree and stand parameters were obtained. Firstly, it was 155

found that the biomass is in average allocated from 19 % to branches and from 81 % to stems. Dry matter is in average 41.6 % of the fresh weight. Mean tree contains 8.38 kg of aboveground woody biomass (DMC) and reaches to 9.68 m in length. The total stand yield was 71 tons per ha. By taking into account whole rotation period, the mean annual biomass increment was close to 9 tons per ha. Although, many allometric equations were fitted as the results of this winter harvest campaign (like versus stem biomass, versus branch biomass, versus height) only one fundamental relationship, the versus whole aboveground woody biomass (DMC) is given at this place (Fig.67-a). This equation is crucial, because it enables to estimate the aboveground woody biomass from any tree at the plantation if its is known. Firstly, based on this assumption, the allometric function was validated by applying it at the 702 values obtained from the summer inventory and comparing it with the values yielded from weighting of these trees as a whole during the harvest. The resulted yield estimation based on the application of the allometric function on 702 tress and subsequently upscaled to one hectare reached 70.8 ± 2 tons per ha (DMC). Note that the error range is calculated from the belts of confidence and the systematic error propagation theory (see e.g. TAYLOR, 1997). During the harvest, all of these 702 trees were cut as a whole and subsequently divided into four separated cartloads which were drove separately at the weight bridge (truck scale). The resulting values from the weight bridge were subsequently upscaled again into biomass content per one hectare (they covered an area of 815 m2) obtaining 68.1 tons per ha (DMC). Considering the error of this value two major error sources have to be taken into account. Firstly, errors of the weight bridge itself, say ±1 %. Secondly, the loos of biomass during the transport, harvesting and chipping could cause only one-way error – underestimation. This could be estimated to be say 3 % down. Other errors could originate from the weighting of the samples for dry matter determination and also by the fact that the dry matter slightly varied from tree to tree and thus the chips were note absolutely homogeneous material from that point of view. Taking into account all these facts, the real weight was estimated to be within the range of 69.5 to 71 tons per ha. Comparing it with the allometric estimation 70.8 ± 2 tons per ha it is obvious that using of the allometric function is very reliable method for biomass estimates. Another way how to derive the error of the allometric function is to divide the data of and DMC into calibration and validation dataset. The 120 trees were divided into these two groups as 100:20 respectively, whereas each 6th (starting from 3rd) tree was selected from the data in ascending order. After that the power function (which was not statistically significantly different from the equation created from all 120 samples) was calibrated and the resulting values were compared with the validation dataset using and as the testing indicators. The resulting values were 8.9 and 0.4 % for and respectively. Taking into account this more strict approach, the resulting allometrically defined yield estimation would be 70.8 ± 6.3 tons per ha.

156

16

AB = 1.49*10-4*DBH2.526 r2 = 0.98

30

Reference stem increment (mm of DBH)

Aboveground woody biomass (kg of DMC)

35

a) 25

20

15

10

5

∆DBH ref = 1.06*10-6*DBH 3.397 r2 = 0.83

14

b)

12 10 8 6 4 2 0

0 20

40

60

80

100

120

20

140

40

60

80

100

120

140

DBH (mm)

DBH (mm)

Figure 67: a) allometrically defined relation between and the AB without leaves (kg of DMC) depicted on left side. b) on the right hand, the relation between and – the reference increment in during the period from 3rd Jul to 22nd July 2009). The solid red curves depict the regression lines and the dashed black lines define the 95 % intervals of confidence. In order to scale up the temporal variation of the increments of trees assessed by dendrometers, of 100 randomly chosen trees from all of the diameter and height classes were measured with calliper at the start of June 2009 and after that again during the summer inventory at the end of July 2009. Based on this results power function describing the linkage of stem diameter and the reference increment was determined in form of with parameter a = 1.062 10-6 and b = 3.397 ( = 0.87, = 100). This function shown at Fig. 67-b) was applied on the known of all 702 trees from above mentioned inventory and thus their reference increments were obtained. At the same time the function was applied on the sampled trees where the increment was measured regularly with dendrometers and by dividing the real by the calculated (reference) increments, the scaling coefficient was obtained. It means, that for each record from one dendrometer there is one scaling coefficient making possible to convert reference increment from all the trees to the real change in . Although, there were totally 40 dendrometers, there were only 15 of them with uninterrupted data series for both of the consecutive years. Due to the homogenity and comparability of both years, only these 15 measured trees were used for up-scaling. For each of these 15 sampled trees, one scaling coefficient was obtained resulting in 15 slightly different increment of all 702 tress. This is especially due to the fact, that the relationship at Fig. 67-b) is not completely tight ( = 0.83). The mean of 15 variants of calculated increment of 702 trees at defined area were finally upscaled to the whole stand, assuming that the 702 trees is a sufficient number to provide representative view on social distribution of the whole stand. These all stem increments were subsequently converted by using the allometric equation given at Fig. 67a) to the biomass increment (DMC) per area of one hectare for further purposes. In this way the reconstruction of the biomass increments for 2008 and 2009 was carried out. To estimate the biomass increment during the 2010, the data from the allometric measurement in January 2011 was used to fit the allometric power function for small shots to final form DMC = 26.87 where is the shoot diameter at 20 cm height 157

(distance from the stump where the shoots were not already conic). In combination with late autumn inventory during November 2010 comprised of 560 shoots measured at and by using the allometric function to convert the shoots diameter to DMC their biomass was upscaled to the whole stand level obtaining final yield 3.9 tons per ha. To get some idea about the evolution of the biomass increment during the year 2010, the data from 40 trees where the tree height was regularly measured (each week) from the start of resprouting were used and it was assumed that the biomass was increasing in the same way as the heights of shoots. Just for imagination, the mean tree height at end of season 2010 was 2.08 m. The comparison of the upscaled biomass increments in 2008, 2009 and 2010 is shown at the following Fig. 68. It is obvious there that as the main reason of lower biomass yield in 2008 compared to 2009 (16.5 and 13.4 tons per ha respectively) is especially the later start of the growing period in 2008 with more than two week delay. Biomass productivity (tons of DMC ha -1)

18 16 14

2008 2009 2010

16.5 t

13.4 t

12 10

8 6 4

3.9 t 2 0 1-Apr

1-May

31-May

30-Jun

30-Jul

29-Aug

28-Sep

Figure 68: Aboveground biomass (without leaves) yields (DMC) during three consecutive years. By using the allometric function on the individual trees where the dendrometers were placed together with the sap flow instrumentation, the of the particular trees can be estimated. The application of this simple approach is depicted at Fig. 69. Unfortunately, the stem increment at one sap flow treated tree was not measured. Since the sap flow was not measured before the 18th of June, the estimation of its rates based on measured and and their relation which has been already used also at Fig. 64 for purposes of transpiration estimations was used. By omitting the first high value (caused either by the spring flush or by wrong sap flow estimation – not measured during this period) and considering only the growing season till the end of September, the mean of 3.2 g -1 kg will be obtained. Taking into account only the really measured values measured values from the 18th of June till the end of September, the resulting will be 2.4 g kg-1.

158

WUE (g kg-1)

20

2009

DBH - 106 mm

18

DBH - 132 mm

16

DBH - 119 mm

14 12 10 8 6

4 2 0 14-Apr 28-Apr 12-May 26-May 9-Jun

23-Jun

7-Jul

21-Jul

4-Aug 18-Aug 1-Sep 15-Sep 29-Sep 13-Oct

Figure 69: The long term based on the sap flow measurements of three sampled trees and their simultaneous measurements of increment and thus allometrically defined aboveground woody biomass increments (kg of DMC). The sums of sap flow and the biomass increment were usually integrated in the weekly time-step (see the particular points), and the curves are smoothed by cubic spline fit. The colours of the lines agree with those from the sap flow picture Fig. 62. Another way how to determine the is to use the upscaled biomass increment related to one square meter and measured . Considering that the obtained in this way is not only the ratio of the created biomass to the transpired water, but it is also effected by evaporation from soil, intercepted water at leaves and other surfaces, the term gross was used in order to point out the differences. Fig. 70 depicts the seasonal patterns of gross with means 3.13 and 3.54 g kg-1 in 2008 and 2009, respectively. Apparently, there are systematically higher values during the second investigated season which also showed higher variability. Except the seasonal decreasing trend from spring toward the end of summer in both of the executed years, the linkage with the dynamic of precipitation totals is obvious (especially on the peaks of gross ). Due to the discontinuous character of precipitation dynamic, the differences between precipitation and the were used to depict this relationship.

4

30.00 4

3

10.00 3

2

-10.00 2

1

-30.00 1

0

0 -50.00

gross WUE P (g kg-1)

7

6

gross WUEP 2008

gross WUEP 2009

100 75 50 25 0 -25

Precipitation – ETact (mm)

5

8 90.00 7 WUE p 70.00 6 Precipitation – ETact 50.00 5

8

-50

Figure 70: The seasonal patterns of gross as the ratio of and allometrically defined aboveground woody biomass increments (kg of DMC) compared with the differences in totals of precipitation and actual evapotranspiration. All values are integrated usually in weekly time-step, except the beginning of 2008, and the curves are smoothed by cubic spline fit. 159

At the next Fig. 71, the water loss and change of allometrically defined biomass are correlated. The values are additionally differentiated into the two groups considering the influence of precipitation. The first group ―without rain‖ is comprised only from values for periods of no precipitation events with totals higher than 10 mm. Conversely, the second group ―rain included‖ was created by adding the periods in which precipitation amounts higher than 10 mm were recorded and thus contains all of the measured values. Taking into account all the values, there is narrower relationship in the year 2008 with the coefficient of determination 0.96 compared to lower 0.51 in 2009. On the other hand, the correlation in 2009 was more often based on shorter time span between individual integrations and thus the other factors like the root-shoot carbon allocation and the stem shrinking and swelling linked with the precipitation affected more the estimated biomass increment. Looking at the slopes of both regression lines, note it is approximately 19 % higher in 2009 compared with the 2008 for the variant without precipitation events whereas it is 30 % lower in the case including all the values. This underlines the effect of precipitation on gross , which has to be taken into account.

without rain

Biomass increment (g m-2)

500

without rain y = 4.13x - 23.40; r² = 0.68

y = 3.36x - 18.87; r² = 0.86 200

rain included y = 4.70x - 29.58; r² = 0.96

450

180

400

160

350

140

300

120

250

100

200

80

150

60

100

40

2008

50

rain included y = 3.60x - 0.13; r² = 0.51

2009

20 0

0

0

50

100

150

0

10

20

30

40

50

60

Evapotranspiration (mm)

Figure 71: The relationships between and aboveground woody biomass increments (kg of DMC) during two consecutive seasons 2008 and 2009. The black filled circles depict the periods with precipitation amounts higher than 10 mm, in contrast the empty circles represent the periods with precipitation amounts lower than 10 mm. All values are integrated usually in weekly time-step, except the beginning of 2008 which results in higher values during this period. On the basis of patterns of gross depicted at Figs. 70 and 71 a multiple linear regression model by adding the precipitation and cumulated effective mean air temperature as other independent variables, which explained the variability more satisfactorily (Figs. 72 and 73) was used. The reasonable and statistically best resulted was the separation of the precipitation again simply into the two categories: amounts lower and higher than 10 mm. Further, the characteristic temporal variability of seasonal gross patterns caused by phenology of mobile carbon allocation was simply expressed as indirect proportionality to sums of effective temperatures defined by the common threshold 5 °C. The final form of 160

multiple regression model created from the data of the season 2009 with

0.87 is as

follows: (130)

Modelled biomass increment (g m-2)

where AB is in kg of DMC, in mm, describes number of days with precipitation amounts higher than 10 mm and Tcum is the cumulative mean diurnal temperature (°C) above 5 °C. This empirical model was subsequently compared with the first simple linear model as well as with the measured biomass increments (Figs. 72, 73 and 74). Furthermore, the model was tested on the data from previous year 2008 yielding higher coefficient of determination compared to the simple linear function. 200

450 Multiple regression r²adj = 0.983 400

180

Simple regression r² = 0.957

Mutiple regression

r²adj = 0.872

Simple regression

r² = 0.512

160

350

140

300

120 250 100 200 80 150

60

100

2008

50

40

2009

20 0

0 0

50

100 150 200 250 300 350 400 450

0

20

40

60

80 100 120 140 160 180 200

Mesured biomass increment (g m-2)

Figure 72: The comparison of measured aboveground woody biomass increment (kg of DMC) with simple and multiple linear regression models based on water lost and aboveground biomass increment relationship. All values are integrated usually in weekly time-step, except the beginning of 2008 which results in higher values during this period. The comparison of two linear regression functions with the measured data revealed that the simple linear model overestimated within the range of low values but conversely underestimated the higher values. This is especially more evident in the case of the year 2009. The multiple linear function fitted quite well in both of the executed year but the underestimation during the 2008 turned the scale down whereas in 2009 the under/overestimation ratio was quite balanced. For the simple linear function the was equal to 9.6 and 26.9 g m2; the was equal to -1.6 and -7.4 g m2 of AB in 2008 and 2009, respectively. Multiple linear regression model yielded 23.8 and 14.5 g m2 and 4.8 and -0.2 g m2 of AB for both consecutive years 2008 and 2009.

161

200

Modelled biomass increment (g m-2)

500

2008

450 400

160

350

Measured

140

300

Multiple regression

120

250

Simple regression

100

200

80

150

60

100

40

50

20

0 27-May

2009

180

27-Jun

27-Jul

27-Aug

0 1-May

27-Sep

1-Jun

1-Jul

1-Aug

1-Sep

1-Oct

Biomass increment (g m-2)

Figure 73: The seasonal patterns comparing the measured aboveground woody biomass increment (kg of DMC) with simple and multiple linear regression models based on water loss and aboveground biomass increment relationship. All values are integrated usually in weekly time-step (except the beginning of 2008) which is depicted by diamond points at the grey line expressing measured values. All curves are smoothed by cubic spline fit.

1600

1800

1400

1600

Multiple regression model

1400

Simple regression model

1200

Measured

1200

1000

1000 800 800 600

600

400

2008

200 0 27-May

27-Jun

27-Jul

27-Aug

400

2009

200 0 1-May

27-Sep

1-Jun

1-Jul

1-Aug

1-Sep

1-Oct

Figure 74: Cumulative comparison of measured (grey solid line) aboveground woody biomass increment (kg of DMC) with simple and multiple linear regression models (solid green and dashed black lines respectively) based on water lost and aboveground biomass increment relationship. Fig. 74 depicts the same comparison of modelled and measured AB increments, this time expressed in cumulative way. Quite authentic course of AB increment was modelled by multiple linear function also for the comparative year 2008. On the other hand, the simple linear model shows that the most problematic periods for estimating the biomass increments are at the beginnings and the ends of the seasons when the antinomy of is significant.

162

6

DISCUSSION

6.1 Soil Moisture 6.1.1 Soil moisture calibration The good agreement and explained variability of EC-10 sensors outputs with volumetric soil moisture based on thermo-gravimetric method and dry bulk density value found in this study confirms their universal applicability reported previously by KIZITO et al. (2008) and BOGENA et al. (2007). Although KIZITO et al. (2008) suggested no demand for the site specific calibration, it was found that the EC-10 underestimated by 5 to 10 % and thus the calibration is appropriate. However, it should be also taken into account, that the standard within this thesis was the gravimetric method combined with limited soil samples of the bulk density, which can introduce errors in this approach and does not underline its generality. Remaining errors can be also possibly caused due to soil variability and discrepancies between measurement volume and measurements depths (BOGENA et al., 2007). It should be noted that within this study the new calibration equation to transform the electrical output of EC-10 to soil moisture was not finding, but it was just intended to verify the accuracy of EC-10 calibrated by manufacturer and possibly to provide any kind of correction for further purposes. The next attempt was to carry out similar calibration of PR1 profile probe sensor. In this case, it was obvious that some corrections will be necessary because the outputs based on default calibration produced values as high as 0.70 m3 m-3 or as low as below 0.05 m3 m-3. However, after proper calibration of several access tubes it was found that the relationship is not universally applicable to all of the tubes since it has still resulted in very high or low unrealistic values measured in some of the tubes. Therefore, the specific calibration for each of the tubes individually was necessary. Unfortunately, even though that the gravimetric sampling was done at least five times close to each access tubes, it was not enough to make appropriate calibration relations. The reason was especially that lowest levels of the soil moisture (especially in case of the depths 0.3 and 0.4 m) were not caught and thus the calibration based on the relations within very small range of soil moistures was impossible. By comparing successfully calibrated EC-10 with PR1, there is obvious quite large scatter which can also partially explain the impossibility to calibrate PR1 based on limited number of samples which in addition inherently take always slightly different sampled volume of soil. Moreover, it also support the provisional approach to calibrate the PR1 by simultaneous measurement with calibrated EC-10 as the way how to multiply the repetitions in calibration without any disrupting of surrounding soil linked with destructive gravimetric method. Similar difficulties to calibrate the PR1 was described by MWALE et al. (2005) who tested the PR1 against gravimetric and neutron probe. They found similar overestimation with increasing trend toward the lower depths. Moreover, MWALE et al. (2005) made sensitivity analysis to the bulk density and revealed that the inaccurate and unit value of bulk density is not the reason of unrealistic PR1 outputs. They concluded that PR1 responses well to relative changes in soil moisture but its absolute values are somewhat questionable. Similar conclusions were given by WHALLEY et al. (2004). In addition, they 163

hypothesised two reasons of PR1 disagreement and its depth dependency – the change in bulk density with depth and different electrical conductivity of soil. But both were rejected. In the absence of theoretical explanation for the effect of depth on sensor calibration they suggested that this effect is introduced by the access tube installation procedure and with this linked unavoidable vibrations of the access tube induced by knocking the access tube in soil with a hammer. This is subsequently likely responsible for the shallower part of the access tube fitting more poorly. In my opinion this explanation is correct, but the depth dependency of the used profile probe was in addition given by default shift in voltage output of the particular dielectric sensors when tested on air and immersed in water. WHALLEY et al. (2004) also concluded that when access tubes are left inserted over winter that the maximum water content can be assumed to be the water content of a saturated soil. Then, with this data it will be possible to express any particular water content relative to saturation. The provided calibration led also to the good agreement within the values near the saturation however, the absolute values of soil water loss and with this linked approaching to the wilting point remains again unknown. Finally, WHALLEY et al. (2004) concluded that for high absolute accuracy on-site calibration should be considered which is in agreement with the experiences from this study. 6.1.2 Soil moisture spatio-temporal variability Assuming that the indirect calibration via calibrated EC-10 was successful, the data from PR1 was used for creating the maps of soil moisture spatial variability within the area of the poplar plantation. Note, that in the centre of this approximately 600 m2 large area, the mast with BREB system is placed. These maps provide information about the soil moisture spatial variability in all of three spatial dimensions and by arranging them in the chronological order we can get very clear idea about the evolution of soil moisture and its spatial variability in the course of time. At the first glance, we can observe noticeable temporal persistence of some dry and wet zones within the field. The persistence of drier zones can be explained by the sandy material from the disintegrating gneiss underlying the loamy soil. Naturally, it is well known that the thickness of the loamy layer is quite variable within the region but also within fields (DUDAL, 1953). It is therefore assumed that the dry sampling locations are situated where the loamy soil is shallow, leading to measuring the soil moisture within the transition zone between the loamy soil and the sandy material. Visual observations of the soil profile during the gravimetric sampling confirm this hypothesis. On the other hand, some sampling locations are very moist, especially in the depths 0.3 and 0.4 m, and it is hypothesised that a fine textured layer at this location leads locally to a perched water table. Another obvious feature is that during the wet periods, the soil moisture is spatially very homogeneous. With subsequent drying it becomes more heterogeneous, especially in case of the lower depths. In case of the upper surface layer, it seems that there is not such pronounced relation between the variability and the soil moisture level. The non-homogenity of the soil moisture in the upper layer during the periods with generally high soil moisture levels can be a result of animal burrows, cracks or slits linked with root growth or with shrinking and swelling of clay soils resulting in faster drainage to the deeper layer. Further they in addition can act as air gaps 164

which result in inaccurate measurement of measuring the soil moisture in principle. Another possibility is also the effect of unevenly distributed rain water from canopy throughfall and stemflow which has a considerable influence on the spreading of the roots of trees and of the understory (LARCHER, 2003). Finally we should not omit the presence of stones and air gaps linked with imperfect installation which is rarely managed (MWALE et al., 2005). Looking at the maps in chronological order within any drying period, it is obvious that the upper layers become dryer much faster compared to the lower layers. The soil moisture data set collected in this study allows the investigation of the relationship between the field-scale means and variance of soil moisture content over time depicted on the soil moisture spatial variability maps. If such knowledge is given, it can be of great significance, for example to optimise the number of samples required to estimate the mean value within a specified limit of error. Our previous results (FISCHER, et al., 2010) demonstrated that the spatial variability linearly decreases with the increasing mean soil moisture values. These negative correlations were consistent with the previous findings of FAMIGLIETTI et al. (1999) and HUPET and VANCLOOSTER (2002). Nevertheless, observed results were in contrast with other previous investigations (e.g. HILLS and REYNOLDS, 1969, HENNINGER et al., 1976, FAMIGLIETTI et al., 1998). Note, that most of the disagreeing studies were conducted on experimental sites with much more pronounced topographic features. By continuing in collecting data of soil moisture spatial variability, the contrarily positive correlation between mean of 16 probes and their standard deviation was found in the superficial layer 0.1 m with Pearson´s correlation coefficient = 0.35 and < 0.001. In the two lower layers, the negative correlation (as previously) with Pearson´s = -0.45 and = -0.58 and with < 0.001 for both of the soil layers 0.2 and 0.3 m respectively was found. Although, negative correlation was found also in case of the deepest layer 0.4 m, it was not statistically significant with resulting = -0.16 and > 0.01. Such no good predictive relationship is in accordance e.g. with observations of HAWLEY et al., (1983) or CHARPENTIER and GROFFMAN (1992). This anomaly might be explained especially by the fact that the superficial layer seems to be relatively homogeneous when is dry due to the exposure to free air and thus evaporative demand of atmosphere which is spatially in equilibrium. However, as already mentioned the wetting of the surface layer might be very uneven. The not significant relationship within the lowest layer might be explained by the fact that the soil moisture in this layer was most of the time very high and thus its natural variability linked with texture or systematic errors in measurements was comparable to the variability linked with spatially uneven drying or root extraction. The widest range of variability during the dry period was observed in the layer 0.35–0.45 m. On the grounds of the visual assessment of soil profile during the gravimetric sampling, it was assumed, that the higher variability of soil moisture in deeper layers appears just because there are places with occurrence of sandy material and those with very fine structure remaining clay mentioned above. The soil water from these sandy fields can drain to the deeper layers and thus the soil gets dryer. On the other hand, the fine texture places bound water very tightly and remain moist most of the time. During the wettest days was the layer 0.35–0.45 much more homogenous and also reached the highest soil water content. 165

In the driest periods (the end of summer 2008 and 2009) the soil water content integrated through all measured layers decreased to the level of stress point and in a few rare cases at some places of the investigated plot even near or just to the wilting point (the places where the sandy soil was observed). However, it should be taken into account that if the soil texture varies across the field, also the wilting point and field capacity can spatially differ (GODWINN and MILLER, 2003). Thus the places where the soil moisture is at very low levels do not have to necessarily mean that the water is here unavailable for the poplars. In the same time, we can observe at the same places where the soil moisture decreases to such low levels, there are also lower values during the wettest periods after rainfalls or snow melting. It can in other words mean that the soil is more sandy at these places and the hydrolimts are shifted downwards to the lower soil moisture levels. On the other hand, these places can be also the effect of the systematic errors in measurements caused mainly by air gaps around the access tubes or other above mentioned issues. To solve these open questions, it is necessary to carry out more gravimetric sampling near these suspicious places in order to verify the measurements. Using of the access tubes and portable PR1 profile probe seems to be a good solution for measuring the soil moisture variability within a field. However, HIGNETT and EVETT, (2008) reported that in many cases, the sensed volume is smaller than the representative elemental volume for the soil in which the sensor is tested, leading to an unrealistically large estimate of the variability of soil water content, and thus to a requirement for excessive numbers of access tubes in order to obtain a mean profile water content of acceptable precision. By considering this fact, we can expect that soil moisture can vary absolutely independently on the interpolation between two access tubes. Finally, if we admit the irrelevance of these maps based on this fact, it can be at least assumed that the mean of all 16 probes provides statistically better estimation of the soil moisture than one point measurement of soil moisture albeit with more precise sensors. 6.1.3 Soil moisture temporal variability The seasonal dynamics of soil moisture was typical with a large supply of available water at the beginning of spring after snow melting. The second period of the highest soil moisture values appears usually after summer rainfalls, especially in June and July which are the two months with the highest amounts of precipitation in the conditions of the Czech Republic (KALVOVÁ and NEMEŠOVÁ, 1997). By contrast, shortage of soil water content was recorded at the end of summer in 2008 and 2009. Such decrease of soil moisture occurred due to low precipitation and still relatively high evapotranspiration rates during this period. Another considerable soil moisture depletion was observed at the turn of June and July 2010 due to the low precipitation supply and high evaporative rates. The soil moisture temporal variation were obviously depth dependent and were also more pronounced closer to the soil surface, where soil is subjected to the root water uptake, soil evaporation and on the other hand dew formation and rainfall events. Indeed, the widest range (from the field capacity to the wilting point) of soil moisture values were observed in the upper layers 0–15 and 15–25 m. Such vertical rules are not anything unusual. Similar results had been already observed e.g. by FINCH et al. (2004), HALL et al. 166

(1996) in both poplars and willows stands, ALLEN et al. (1999) in poplar stand, BRÉDA et al. (1995) in oak stand or LOAGUE (1992), HUPET and VANCLOOSTER (2002) and EVETT et al. (2009) in agricultural crops. The amplitude of the temporal variation in the intermediate layers is clearly more attenuated and decrease with depth. Nevertheless, these lower layers get drier in response to the root water uptake combined with the upward flux induced by differences in the soil water potential between the wet low and upper drier layers. By comparing the soil moisture patterns of the poplar plantation and the turf grass we can find different water uptake from the particular soil layers. Namely, most of the water is losing from the upper layers and little from the deepest layer 0.4 m. This holds for both covers. However, in case of the turf grass, there is significantly higher difference between the water loss from the layer 0.1 and 0.4 m. It seems that the turf grass almost does not take any water from the soil layer 0.4 m, whereas the poplars do. Based on the soil sampling, it was found that the most of the root system of the turf grass is situated in the upper mineral layer up to 0.2–0.3 m in depth. QIAN et al. (1997) reported that due to the regular mowing, turf grass root systems are concentrated at surface 0.1 to 0.15 m depth. 0

0.1

0.2

Soil depth (m)

0.3

0.4

0.5

0.6

0.7

0.8

Temperate grassland 0.9

Temperate deciduous forest

1

0

10

20

30

40

50

60

70

80

90

100

Cumulative root distribution (%)

Figure 75: Cumulative root distribution (cumulative percentage) as a function of soil depth for two contrasting terrestrial biomes and for the theoretical model of GALE and GRIGAL (1987). The curve of each biom is the least square fit of parameter β from Gale Grigal s equation Y = 1- βd, where Y is the cumulative root fraction with depth (a proportion between 0–1), d is the soil depth (in m), and β is the fitted parameter with value 0.943 and 0.966 for temperate grassland and deciduous forest respectively according to JACKSON et al. (1996). During the soil sampling in the poplar plantation, it was estimated on the basis of visual evaluation that most of the root system of poplars is concentrated within the upper layer up to 0.4–0.5 m. It is in high agreement with JACKSON et al. (1996) and their results from theoretical Gale and Grigal´s model parameterized for temperate grassland and temperate deciduous forest bioms. The root distribution of temperate grassland and temperate deciduous forest is well illustrated at Fig. 75. Generally, within the upper 0.3 m is concentrated 65 and 85 % of the root system for temperate deciduous forest and 167

temperate grassland respectively (JACKSON et al., 1996). Note that the root to shoot ratio is 0.23 and 3.7 for temperate deciduous forest and grassland respectively (JACKSON et al., 1996). According to DICKMANN et al. (2001), most of the fine roots of poplars are in the top 0.1 m where they form extensive network if there is enough water and nutrients. The main taproot is usually reaching to 0.6 m depth, the same as some oblique roots. The coarse roots growing strongly horizontally from the taproot are the mostly located within 0.05 to 0.2 m from the surface. Horizontal roots can be found several tree lengths away from the base of the stem in a planted stand, although the closer the spacing of trees, the more restricted lateral root growth becomes. Many of the roots from adjacent trees in a clonal stand become grafted together, forming an interconnected network. Vertical ―sinker‖ roots branch from the horizontal roots and explore the soil to depths of 1–3 m or more. Different poplar clones produced from rooted cuttings have a similar root architecture, although clonal variation has been observed (DICKMANN and PREGITZER, 1992). The possible reason of the distinctive water uptake from the lowest layer at 0.4 m in the poplar plantation and the turf grass might be therefore simply the effect of the different root distribution. It is also obvious that turf grass takes more water from the three higher layers compared to poplars. It can be effect of more developed root system in these layers. JACKSON et al. (1997) reported 20 times greater live fine root length (112 km m-2) in grassland than in deciduous forest (5.4 km m-2) for the upper 0.3 m depth of soil. It results in live fine root area index ( as an analogy to aboveground ) 79.1 compared to 9.8 in temperate grassland and deciduous forest respectively. Nevertheless, the reason can be also the fact that the soil under the turf grass is much more exposed to the solar radiation and loose more water by evaporation. By employing water balance model SoilClim (HLAVINKA et al., 2011, ORSÁG et al., 2011), there were found disagreement between simulated (underestimated) and measured soil moisture below turf grass whereas the simulated evapotranspiration agreed with BREB measurements. One possible reason might be some lateral underground water inflow from higher positions (note that the area is on the mild slope), as indirectly confirmed during soil sampling in June 2009, when a flowing water stream at approximately 0.80 m deep was observed. In accordance with the site position, the underground inflow could be anticipated on lower slopes or on concave sites (DYER, 2009). This might be also one of the possible explanations of very stable soil moisture patterns within the layer 0.4 m. With an adoption of the term soil available water, which is by simplification the water up the wilting point, and providing that the majority of the roots are located in the profile 0–0.45 m, we can calculate how much water in this layer could be used by the plants (in mm of water column). Generally, the amount of the available soil water fluctuated between field capacity and the stress point which is usually reported between 0.3 and 0.5 of plant available water (e.g. CHARLESWORTH and STIRZAKER, 2008). The results showed there was still enough available water for the poplars to grow almost during the whole measuring campaign. The water might be limiting factor at the end of summer 2008 and then probably in September 2009, when however, the growing season was ending. In both cases, there were observed decreases in stem increment, but due to the ending of the growing season it was not possible to distinguish what was the real reason of growth decline. Looking back 168

to the summer and the end of summer 2008, there was very possibly some limitation in water supply of the turf grass. By comparing the soil water availability patterns and then especially the cumulative water depletion of both covers measured by PR1, it was found that the turf grass lose approximately 20 % more water from the whole soil layer 0–0.45 than the poplar plantation during the whole growing season. These surprising results are in accordance with JAMES et al. (2003) who found higher soil moisture depletion (0–0.3 m deep soil layer) in grassland compared to aspen stand (Populus tremuloides). They ascribed the ability of grassland to maintain higher water uptake to larger live fine root length and also to higher soil evaporation. To verify the soil moisture measurements, the outputs from three calibrated EC-10 sensors were integrated into the pedon of one cubic meter. Note that the PR1data was not used just because it did not give continual data series. After that, its water balance (water incomes and water losses) was compared with the records from the rain gauge resulting in a very good agreement. However, it might seem suspicious because there should be some effect of interception and surface run-off which should make the water income lower compared to the records from rain gauge. In fact, the interception has been not measured yet. According to HALL et al. (1996) the interception in high density poplars culture varies near 15 % of precipitation per whole season and is 21 % for foliated period. For comparison, PETZOLD et al. (2010) found the seasonal interception of poplar stand to be 24–28 % of precipitation. PERSSON and LINDROTH (1994) reported measured values for willow stand (including stemflow) within the range of 10 to 15 % of precipitation and IRITZ et al. (2001) calculated interception in willow stand as 11 % of the whole seasonal evapotranspiration. Considering these values, it seems that the soil moisture measurement slightly overestimates. However, there should be taken into account other factors like general underestimation of rain gauges (STRANGEWAYS, 2003), lateral inflow and especially already discussed soil moisture spatial variability which make impossible to make conclusions from one point soil profile measurement. Considering these imperfections, it might be concluded, that the verification of soil moisture measurements by rain gauge records provide a certain kind of another at least rough confirmation. Finally, by relating the soil moisture patterns to the particular hydrolimts we can find here several hydric regimes according to KUTÍLEK (1978). The most usual hydric regime for both covers is the semiuvidic regime, which is defined as soil moisture between field capacity and stress point. During the wet conditions after snow melting or after rainfalls, we can find here the uvidic regime (between saturation and filed capacity) and on the other hand during the driest observed conditions, especially in case of the turf grass, we can talk about semiaridic regime defined by interval between stress point and wilting point.

6.2 Evapotranspiration 6.2.1 BREB error analysis The errors introduced by the BREB method in the computed energy fluxes values have been already evaluated by several authors (FUCHS and TANNER, 1970, SINCLAIR et al., 1975, ANGUS and WATTS, 1984, BERTELA, 1989, LINDROTH and HALLDIN, 1990, PEREZ et 169

al, 1999, FOKEN, 2008a, SAVAGE, 2010). As already mentioned, the first conservative rule to avoid the errors in the estimation of the fluxes, is to reject the data within the range of instrumental errors (PEREZ et al., 1999). It is obvious that such kind of filtering would have dramatic impact on the whole gradients measurement dataset of this study, especially on the data from the poplar plantation. The main reason, why the data from the poplar plantation are more error prone, is the height of the canopy and from this resulting aerodynamic roughness length (z0) which is linearly increasing with height. As a result of the high roughness, effective mixing causes that the gradients of temperature and humidity are small (BALDOCCHI et al., 1988). The medians of diurnal temperature gradients within growing seasons were -0.1 and -0.26 °C m-1 for the adult ~12 m height and resprouting ~1 m high poplar canopies respectively. 90 % of all these values were within the range of 0.45 to 0.35 °C m-1 and -0.87 to 0.03 °C m-1 for two types of growth stages respectively. Thereafter, the medians of the humidity gradients within the same periods were 10 Pa and 15 Pa m-1 with 90 % of values within range of -50 to 0 and -63 to 3 Pa m-1 for the adult and resprouting poplar canopies respectively. IRITZ and LINDROTH (1996) reported 90 % of the diurnal temperature gradients fell within the range of -0.2 to 0.23 °C m-1 and the humidity gradients within -65 to 5 Pa m-1 above their high density willow stand with mean three height ~4 m. For comparison, LINDROTH and HALLDIN (1990) reported much lower gradients measured above ~18 m high sparse pine forest. Namely, 90 % of temperature gradients in the range -0.024 to 0.011 °C m-1 and a median value of -0.0164 °C m-1 (converted from potential temperature used by authors). The humidity gradients were then reported within the range -1.5 to 0.18 Pa m-1 and a median -0.5 Pa m-1. Similar gradients were also reported by BERNHOFER (1992) with average -0.025 with maximum around -0.06 °C m-1 (again converted from potential temperature used by author) and average -1 with maximum -2 Pa m-1 for temperature and humidity respectively above a mixed sprucefir-beech forest with mean height ~30 m. Note, that the theoretically more correct specific humidity was not used instead of water vapour pressure gradient (see WEB et al, 1980) in this study or in comparison with other authors, because by assuming that the decrease of atmospheric pressure with height in the altitude 550 m is approximately 11 Pa m-1 (ALLEN et al., 1998), the change of the water vapour pressure would be -5.5 10-3 Pa m-1 and considering the accuracy and the vertical spacing of the sensors, it has very negligible influence on the calculation. In case of the gradients above short canopies, the information in literature is very sparse. Nevertheless, e.g. SAVAGE (2010) reported the vapour pressure gradients above mesic grassland to be within the range -320 to 710 Pa m-1 and very infrequently within ±10 Pa m-1 (occurring usually in the early morning or late afternoon). SAVAGE (2010) did not mentioned the temperature gradients but they are generally considered to be higher in the similar proportion like the gradients of vapour pressure compared to the high canopies (LINDROTH and HALLDIN, 1990). It is in agreement with the gradients measured above very short (~0.06 m) turf grass within the presented study with diurnal medians during the growing seasons -0.29 °C m-1 (90 % within the range -0.8 to 0.13 °C m-1) and 50 Pa m-1 (90 % within the range -106 to 6 Pa m-1) for the temperature and humidity respectively. Comparing the results of the gradients above short and high canopies, it seems that the high density plantations (both poplars and willows) have the gradients somewhere in 170

between those of forest and short canopies, which is very probably due to their lower height and especially very high density. Since the gradients above aerodynamically rough surfaces like forests are one or two orders of magnitude smaller compared to low vegetation, it is necessary to measure them with very high accuracy (LINDROTH and HALLDIN, 1990). There are two possibilities how to deal with this issue. Firstly, we can use extra precise sensors for measuring the gradients. Or alternatively we can enlarge the vertical distance between the sensors as much as possible (FOKEN, 2008a) which on the other hand would enlarge the overall footprint and fetch thus could be the main limitation (LINDROTH and HALLDIN, 1990, STANNARD, 1997, SAVAGE, 2010). FOKEN (2008a) recommended to choose the ratio of the measuring aerodynamical heights (the height above the zero plane displacement; STULL, 1988) of the particular levels greater than 4–8. Looking at the BREB systems in this study, the aerodynamical heights (m) of the particular sensors above the turf grass are almost identical with the real heights (taken from the ground level) due to the low zero plane displacement (~0.04m) leading to the ratio of ~2:0.4 which is 5. In case of the adult poplar plantation, the situation is much more critical. For better estimation of the zero plane displacement of the poplar plantation, the empirical derivation of DOLMAN (1986) with value 0.75 of mean canopy height (~12 m) for dense foliated oak stand (with mean height 9.6 m and density 3000 trees ha-1) can be used. Thereafter, during the seasons 2008 and 2009, the aerodynamical heights were 5 (14 - 9) and 3 (12 - 9) m leading to the ratio of 1.7. Assuming that sparser canopy in the first year of the second rotation comprised the resprouting stumps and later new shoots can aerodynamically act similarly like unfoliated oak stand described by DOLMAN (1986), the zero plane displacement can be estimated as 0.57 of the mean canopy height (~1 m) in 2010. Thereafter the particular aerodynamical heights would be 0.45 (1.2 - 0.75) and 3.45 (4.2 - 0.75) m and their mean seasonal ratio would be 7.7 which is adequate. Note that the zero plane displacement has not substantial role in case of relatively low vegetation and thus the approximation with the unfoliated oak stand is acceptable. The above mentioned practically means that by taking into account the vertical separations within the gradient measurement, the BREB above turf grass and the resprouting poplar plantation are in accordance with the Foken´s recommendation. However, in case of the adult poplar stand the ratio 1.7 is far from the optimum. This deficiency of the adjusting the BREB above the poplar stand was evident on the ratio of rejected data and also on the magnitudes of absolute errors of the final fluxes. Nevertheless, in order to solve this problem, the height of the upper sensor would have to be increased and this will lead to the limitation by fetch. For example, in order to get on the minimum optimal ratio of the measurements height (4) we would have to lift the upper sensor 9 m higher. On the other hand, it would mean that to obtain at least 50 % equilibration of measurement of the fluxes to the surface of interest, at least 140 m long fetch would be required according to the Gash´s model (GASH, 1986) and thus it would make the correct measurement of the Bowen ratio above the poplar plantation impossible. The optimum ratio requirement given by FOKEN (2008a) is seldom taken into account in practise especially in measurements over high vegetation where it is usually around 1.5 (LINDROTH and HALLDIN, 1990, BERNHOFFER, 1992, BAR et al., 1994, IRITZ and 171

LINDROTH, 1996). It could seem that the ratio 1.7 for adult poplar stand is thus not critical, however, it should be noted that the cited authors reported using of sensors with at least two orders of magnitude higher accuracy. For example, IRITZ and LINDROTH (1996) measured Bowen ratio above high density willow culture with very precise fan aspirated wet and dry thermometer interchange system described in details in LINDROTH and HALLDIN (1990). The vertical spacing between the sensors was 0.75 and the ratio of the aerodynamical heights 1.48. By such small vertical separation it was achieved to measure the fluxes without any serious limitation by fetch. The accuracy of the gradient measurement was reported to be better than 0.003 °C and 0.3 Pa. Similar BREB system consisted of two precision psychrometers that were interchanged automatically (to avoid systematic bias) over a distance of 3 m with the error of measured gradients ±0.003 and ±0.5 Pa for the air temperature and humidity respectively was used by BERNHOFER (1992). SPITTLEHOUSE and BLACK (1980) reported similar error ±0.003 and ±1 Pa for the air temperature and humidity gradients respectively and same like the other above mentioned authors confirmed the necessity for using very precise and exclusively reversing psychrometric BREB system for measuring the gradients above aerodynamically rough surfaces like forests. LINDROTH and HALLDIN (1990) concluded that the accuracy of fixed sensor system is probably satisfactory during night-time and for application over low vegetation where the gradients are typically one or two orders of magnitude larger than over forests. Other authors using the fixed sensor BREB systems reported the accuracy of measured gradients to be 0.01 °C and 4 Pa (SINCLAIR et al., 1975), 0.01 °C and 8 Pa (MCNEIL and SHUTTLEWORTH, 1975), or 0.02 °C and 20Pa (TANNER et al., 1987, PEREZ et al., 1999, SAVAGE et al., 2009) for the air temperature and humidity respectively. However, BARR et al. (1994) tested both interchangeable and fixed sensors BREB systems against eddy correlation method above mature 18 m deciduous forest and he surprisingly found better agreement between eddy covariance and BREB with fixed sensors. Finally, FOKEN (2008a) very objectively commented that most of these investigations are based on either single measurements or false assumptions. Often, only the electrical error of sensors is used (about 0.01–0.001 °C), but not the error in the adaptation of the sensors to the surrounding medium and atmosphere with radiation, ventilation and other influences. Only with much effort in measuring technique is it possible that sensors under the same meteorological conditions and mounted close together show differences of only 0.05–1 °C or 5–10 Pa for the air temperature and humidity respectively. Therefore, the errors in temperature and humidity measurements in the atmosphere are significantly higher than the pure electrical error (DUGAS et al., 1991). This range given by FOKEN (2008a) is in addition much closer to the empirically determined accuracy of calibrated pair of sensors EMS33 with root mean square error ( ) below 0.3 °C and 50 Pa for the air temperature and humidity respectively. Similar commercial combination thin-film polymer capacitive relative humidity and adjacent temperature sensor instrument with comparable accuracy were used in BREB systems e.g. in study of CHEN and NOVAK (1997) or SAVAGE (2010). In the in the later, very thorough comparison among several combination relative humidity and adjacent temperature sensor instrument was carried out. Moreover, SAVAGE (2010) tested these sensors with standard 172

cooled dewpoint mirror hygrometer based BREB system and two other independent methods – eddy covariance and scintillometry, with very good results. As the main disadvantage of the combination relative humidity and temperature sensors is considered that two measurements, relative humidity and temperature, are required for estimation of water vapour pressure as opposed to one for a dew-point hygrometer (SAVAGE, 2010). On the other hand, the very big advantage of combination capacitive humidity instrument is that they require no servicing compared to a dewpoint hygrometer which requires a bias adjustment and mirror cleaning each week. SAVAGE et al. (2010) concluded that a combination capacitive humidity instrument, with similar relative humidity and temperature random error magnitude of at most ±2 % and ±0.3 °C, and similar measurement time response, would be an adequate and less expensive substitute for a dewpoint hygrometer. Based on manufacturer information (EMS 33 Air Temperature and Humidity Sensor User´s Manual, 2004), the magnitude of the error of combination sensors used in this study is ±0.3 °C and ±2 % for the air temperature and relative humidity respectively. However, by assuming that the gradients are measured, the errors can either add up or cancel and thus, in the worst case the absolute error ±0.6 °C and ±4 % for the measured variable respectively should expected. By pairing of the sensors (careful selection of the sensors with very similar systematic errors and their precise calibration) the mutual error expressed as was found to be below 0.3 °C and 50 Pa for the air temperature and humidity respectively. Note that the vapour pressure accuracy is approximately one order of magnitude higher around 40 % of the relative humidity when the most important conditions for evaporation generally are. Suggesting that if we are sure, that the errors are random and strictly not systematic, we can assume that the errors during the averaging period (e.g. 30 minutes) are reducing due to repeatability of the measurements, according to the ratio between standard deviation and the square root of the number of measurements which is so called standard deviation of the mean (TAYLOR, 1997). This is well illustrated at Fig. 76 on the example of eliminating the uncertainty in temperature measurement by increasing the number of measurements. Taking into account this basic rule for random error propagation in averaging, it can be assumed that the mutual errors below 0.3 °C and 50 Pa in the measuring the gradients would be reduced by 15 repetitions during half hourly integrations to 0.08 °C and 13 Pa for the air temperature and humidity respectively. This theoretical assumption could also explain the very low errors of the capacitive sensors related to dewpoint hygrometer within 20 minutes integrations in laboratory and filed conditions reported by SAVAGE (2010). However, it should be stressed that to apply this rule, we have to assure that the errors which we are going to reduce by repetitions are strictly not systematic. By deriving the curve at Fig 76, we can also find that the most efficient is to increase the number of repetitions up to 20 or 30, since later increasing of repetitions does not improve the uncertainty significantly.

173

Error in temperature ( C)

0.30 0.25 0.20 0.15 0.10

0.05 0.00 0

10

20

30

40

50

60

70

80

90

100

Repetitions

Figure 76: Example of the theoretical dependency of uncertainty in measuring of the temperature on number of repetitions, assuming sensor with random error ±0.3 °C. As previously mentioned, there are two main situations when the data from BREB method has to be rejected due to the errors in measurements. The first is wrong flux sign firstly demonstrated by OHMURA (1982). For this type of weakness all the errors in net radiation, subsurface heat flux and temperature and humidity profiles are responsible. Improperly adjusted temperature and humidity profiles due to inhomogenity of the surface and the non-steady state can also cause this problem. The problem of incorrect flux sign can be usually encountered with early morning and late afternoon data and during precipitation, when the gradients are small, the boundary layer is not in steady state, or the net radiation and soil heat flux are not accurately measured (OHMURA, 1982). The second case when BREB method fails and is subjected to large errors is linked with approaching of the Bowen ratio to -1 because the denominator of Eqs. 43 and 44 becomes vanishingly small (BLAD and ROSENBERG, 1974, OHMURA, 1982, TANNER et al., 1987, PEREZ et al., 1999, GUO et al., 2007, SAVAGE et al., 2009). This condition occurs in nature for short periods in early morning and late afternoon or precipitation during the conditions when the available energy is close to zero (FUCHS and TANNER, 1970, PEREZ et al., 1999), but at such times latent heat flux rates are so small as to be of little interest (BLAD and ROSENBERG, 1974, LINDROTH and IRITZ, 1993). The situation, however, may also occur especially during an intense foehn wind, when the direction of latent heat flux is opposite to that of sensible heat flux. The problem under the foehn is very serious because evaporation is usually large (OHMURA, 1982). The physical meaning of such condition is that the direction of the temperature gradient changes to be opposite to that of the vapour pressure gradient (PEREZ et al., 1999). When Bowen ratio approaches -1, sensible heat flux is equal to the negative latent heat flux which could be described as pseudoadiabatic and isobaric conditions which can be depicted by the wet-bulb temperature isolines of the psychrometric chart (SAVAGE et al., 2009). The conditions could be also, however, understood as adiabatic, because sum of sensible and latent heat is equal to zero and thus the available energy ( ) is equal to zero too (SAVAGE et al., 2009). The practical impacts of such conditions are unreasonably large absolute values of the turbulent energy fluxes with characteristic spikes which have to be excluded (GUO et al., 2007). 174

Traditionally, the Bowen ratio data are rejected in a fixed range around -1 (TANNER et al., 1987, ORTEGA-FARIAS et al., 1996, FOKEN, 2008a, SAVAGE et al., 2009). However, the data for exclusion is not in actual fact any fixed interval. Assuming theoretically absolutely precise sensors, there would be no interval for exclusion. Only the Bowen ratio exactly equal to -1 would have to be excluded which is in fact very rare situation. Based on the error analysis of the Bowen ratio, OHMURA (1982) and later PEREZ et al. (1999) analytically derived the range of data which have to be rejected. In case of PEREZ et al. (1999) the interval is directly dependent on the vapour pressure gradient measured in each sampling period and on the resolution limits of the sensors. This method is widely used and cited (e.g. PEACOCK and HESS, 2004, GAVILÁN and BERENGENA, 2007, SAVAGE et al., 2009), however, to my knowledge there is small mathematical inconsistency. The original equation given by PEREZ et al., 1999) is as follows: (131) where (°C) and (kPa) are the errors in the air temperature and vapour pressure gradients respectively, (kPa) is actual vapour pressure gradient, (kPa K-1) psychrometric constant and is the error in Bowen ratio which in the same time determines the range in which the data are regarded as suspicious and need to be rejected. But the sign in the numerator does not have physical meaning. It can be simply prove by using e.g. the sensors with theoretical errors in gradients 0.4545 °C and 30 Pa or 0.7575 °C and 50 Pa assuming Pa K-1 when the errors in numerator are cancelling. Note that these theoretical values still lie in real range of errors, although there is quite high systematic error in temperature. The reason why probably the authors overlook this inconsistency is that they use thermometers in order of magnitude more precise and thus the value of temperature error does not change significantly. The mathematical reason of this inconsistency is that PEREZ et al. (1999) removed the sign of the absolute values in their derivation. SAVAGE et al. (2009) derived another analytical solution for the variable range for excluding Bowen ratio around -1 and they concluded that they obtained very similar resulting equation like PEREZ et al. (1999). In fact, they found the absolute analogy of the equation which provides the same results like that of PEREZ et al. (1999) if there would be the right sign in the numerator. However, GUO et al. (2007) used in their article the right form with summing in the numerator which is also the form used in the calculations within this thesis. The hyperbolic Fig. 23 is just based on the correct equation given by GUO et al. (2007) and we can observe here which data have to be rejected due to this analytical condition. The area of the data which should be rejected by using the Campbell Scientific BREB which is very widely used and considered as a standard method is also depicted here. It is noteworthy, that by applying the theory of random error and its propagation to mean described earlier (Fig. 76), the reducing of the errors during halfhourly averaging will result in almost identical area of data exclusion as is expected for the Campbell Scientific BREB system. It could explain very similar results and stability of BREB based on dewpoint hygrometer compared to that based on capacitive relative humidity sensors described in SAVAGE (2010). One of the reasons why the combined relative humidity and temperature sensors instrument show higher random error 175

(and thus noise during short averaging periods) compared to dewpoint hygrometer, but show reliable values during longer time averaging span is probably that they have two sensors for measuring one variable – vapour pressure. Moreover, these two sensors are spatially-separated, each mounted on different chips with slightly different temperature, with both measurement having their own random error, bias and resolution limitation. Ideally the temperature of the capacitive sensor should be known, not that of thermometer. The temperature difference between relative humidity and temperature sensor should decrease with increase in flow rate in the chamber but this would increase the system current drain (SAVAGE, 2010). Finally I would like to demonstrate, what the impacts of the systematic errors are. On the Figs. 25 and 26 the diurnal courses of latent heat flux calculated from original and artificially gradually shifted (from the original value) input values of temperature and relative humidity respectively are depicted. These pictures confirm the finding of SAVAGE (2010) that the main disadvantage of the combined relative humidity and temperature sensor instruments is the fact that temperature is not used only for calculating of the temperature differences but also for the vapour pressure differences. It is probably the main reason why the systematic errors in temperature are more critical than the systematic errors in relative humidity. On the other hand, it is advantage for the manufacturer of combined sensors, because it is much easier to provide sensors with accurate thermometers, whereas the accurate capacitive relative humidity sensors (especially for relative humidity higher than ~ 90 %) are still great challenge. At this pictures, we can also observe the spiky periods around morning and evening which are mostly filtered by the simple fix interval of exclusion . If we used the variable interval, it would be stricter and remove most of the data calculated from the most inaccurate sensors. There also noticeable values during the nights when the relative errors are usually the highest. On the other hand, the absolute errors are not so dramatic because of low absolute values of fluxes. At this place it can be recommended to use always some reference evapo(transpi)ration model like Penman, Penman-Monteith or Priestley-Taylor (however the last is not very suitable into arid and windy areas) in order to graphically check the possible outliers by plotting and screening the data to each other in the same way as described by PAYERO et al. (2003) for net radiation verification or by ALLEN (1996) for other meteorological variables. In this way, most of the data from night time and morning spikes would be safely recognized and rejected. The scatter views at next two pictures (Figs. 27 and 28) are depicting the same situation for the whole one month dataset. It is well depicted here that for high values of latent heat flux, the variability of inaccurate sensors is quite low – the fluxes are only under or overestimated, but remain quite consistent. However, the low values of the fluxes show high variability. Therefore, it should be used as one of the basic rule of thumb. If the data of low fluxes show higher variability and are inconsistent with the modelled data, it is very likely that the sensors are not very accurate and need recalibration or to be replaced. Another simple way, how to get idea about what is the effect of the systematic errors, or how precise are the measurement of the gradients using combining sensors, is to use another pair of sensors with the same type of radiation shields. Fix them into the same height and very close to the positions of the original sensors. Firstly just ensure if they 176

show the same values and after that each week or simply, as often as the locality is attended, interchange the position of the two additional sensors (upper one down and the lower one up) and focus on the possible differences between the calculated fluxes or Bowen ratios resulting from the particular pairs of sensors. By this way, the validity of the measurement can be tested in very simply way. Note that the interchange of two sensors takes ~5 minutes. One might suggest that looking at time course with artificially simulated errors of sensors, the situation for systematic error is ―not too bad‖ and in combination of both sensor errors it would be still never higher than 30 % for most of the diurnal period (the errors can be much lower and can also mutually cancel). Note that such underestimation is quite usual for widely used eddy covariance due to unclosed energy balance (LIU and FOKEN, 2001, WILSON et al., 2002, FOKEN, 2008a, FOKEN, 2008b). However, it must be stressed, that not only the error of Bowen ratio as a consequence of errors in temperature and humidity measurements contributes to the final error of the latent heat flux, but especially the net radiation and to a lower extent also the soil heat flux does so. To conclude this section, it is recommend to all users of BREB to minimize all of the possible systematic errors as much as possible in order to use a reliable method for estimation of turbulent energy fluxes as the BREB method in essence is. 6.2.2 Gap filling After the proper data rejection, some method to fill the gaps has to be used especially for further temporal integrations. If missing or rejected values in a half-hourly (or other time span) data set would be perfectly random distributed, the calculation of an annual sum could be easily performed, i.e. by taking the average of all available data, and converting the unit of the average to the appropriate unit per year. Data gaps, however, do not occur randomly because of system failures or rejection of poor data. The non-randomness of the gaps in the data leads to the need to develop and test a variety of gap-filling methods (FALGE et al., 2001). There are numerous methods for gap-filling of the eddy covariance method but very limited number for that of the BREB method. In case of eddy covariance some methods are based on complex mathematical skills, such as nonlinear regression (FALGE et al., 2001), multiple imputation (HUI et al., 2004) or recursive parameter estimation algorithms with Kalman filtering (ALAVI et al., 2006). Although other semiempirical methods like such as look-up tables and mean diurnal variation (FALGE et al., 2001) are statistically correct, they have very little physical meaning. Whereas, methods using site-specific calibration by simple or multiple linear regression of evapo(transpi)ration model like e.g. Priestley-Taylor (SUMNER and JACOBS, 2005, RIZOU, 2008) are simple but straightforward and it gives physical sense. Comprehensive investigation of 15 gap-filling techniques for eddy covariance method was given by MOFFAT et al. (2007), however, it is especially focused on the carbon fluxes gap-filling not always usable for the energy fluxes. To my knowledge there are only three published (but probably more is used) methods for gap-filling of fluxes obtained by BREB method. The first method is simple linear interpolation of the missing data from the preceding and subsequent values (LINDROTH and 177

IRITZ, 1993, PEREZ et al., 1999) which is naturally temporally limited (say up to two hours). The second described way how to deal with gaps if e.g. vapour pressure gradient data are missing is to use some additional method to estimate exchange coefficient from horizontal wind speed using a method of WANG and BRASS (1998) described by WESSON et al. (2001) or by using the temperature variance method using the standard deviation in air temperature and application of the Monin-Obukhov similarity theory (TILLMAN, 1972, DEBRUIN et al., 1991, GUO et al., 2009, SAVAGE et al., 2009) demanding fast response thermometer (~1–10 Hz). Exchange coefficient or the fluxes themselves can be obtained by combination of BREB with aerodynamic method (VOGT and JAEGER, 1990) demanding additionally measurements of wind speed profiles. However, all of these listed methods need some extra instrumentation but on the other hand they enable to patch the data for which the temperature or vapour pressure differences are invalid, inaccurate, unreliable or for which Bowen ratio is close to -1 (SAVAGE et., 2009). The third possible method was described by GUO et al. (2007) and it is based on site-specific calibration of PriestleyTaylor coefficient using simple linear regression for three possible heat flux combinations ( > 0 and > 0, > 0 but < 0 , < 0 and < 0) individually. Inspired by this paper, four evapo(transpi)ration models were compared in order to judged which is the most suitable for gap-filling. Namely, the Priestley Taylor (with Priestley Taylor coefficient equal to unity which is in fact equilibrium evaporation), Penman and PenmanMonteith with dynamic simulation of surface resistance based on Todorovic and Lohammar approaches were used. Since the study of GUO et al. (2007) was carried out at site with permanent irrigation (a saturated surface) they could use only one Priestley-Taylor coefficient for each flux combination for let say whole season. In our condition the soil moisture is changing in course of season as it is driven by rain events. Therefore, the calibration of each model by linear regression for each day separately was used. To deal with the flux combinations and their individual calibration, only the simplified manner by dividing the data into diurnal and nocturnal part according to calculated sunrise and sunset time was used. As the best method applied at the three different covers, with respect to the resulting coefficient of determination, and , seems to be the Penman equation, but with only very small differences compared to Priestley-Taylor method. Priestley-Taylor model has the advantage in the simplicity of calculation, but on the other hand in windy and arid areas can underestimate (JENSEN et al. 1990, MCANENEY and ITIER, 1996, UTSET et al., 2004) because it does not include the wind speed as the input. From that reason probably the Penman equation was slightly better and would be possibly also more promising solution for gap filling in general. On the contrary, as the least suitable method for gap filling was found the Lohammar equation. However, it should be noted that this method demand the specific calibration by iterative procedure (e.g. by taking the coefficient of determination or as the mark of the best fit) which was done only once for the whole analyzed period leading obviously to the worst results. Likewise, despite the analytical nature of the Penman-Monteith approach with the Todorovic´s model for surface resistance, it did not bring better results in gap filling compared with the semi-empirical and simpler Penman method. The next step of such comparison should be the validation with some independent method. Because all of 178

the tested methods are in principle based on the Bowen ratio and radiation balance equation, the best would be to compare the gap filled data with the eddy covariance, or for long term values with precise lysimeter which does not include radiation balance as the inputs. For further analysis, a method for parameterization of the Lohammar equation each day or for some other short period like e.g. week should be used. If in the same time a proper parameterization of the aerodynamic resistance is done, it might be the most physically based technique for gap filling of the four selected methods within this comparison. On the other hand it could bring similar results like using the PenmanMonteith with variable surface resistance according to Todorovic which probably was not more successful than the simpler semi-empirical methods due to higher number of parameters and assumptions. Thus it seems that more easy method with limited numbers of parameters may be more suitable and robust. Another challenge of this gap filling attempt should be to divide the day not only into diurnal and nocturnal parts but to deal with the three different flux combinations according to GUO et al. (2007). It should not be obstacle in form of any additional data because it will be easily derived from the gradients which are necessary for Bowen ratio calculation. My private target is to develop from the current version of FluxCorrector not only the software for reliable gap filling, but especially software which will be able to calculate the fluxes from the rough data with all the quality checks and data rejections which will be finally gap filled in order to make temporal integrals according to users requirements. Such tool will be very suitable especially for bulk data processing. Generally, it can be concluded that using of current version of FluxCorrector software in combination with Priestley-Taylor or Penman approach is suitable tool for creating uninterrupted curves of latent or sensible heat fluxes with resulting relative within the range of 4–16 % In my personal opinion, the combination of BREB method with the aerodynamic method could provide reliable tool for energy fluxes estimate, at least above relatively short canopies, and in the same time could provide very elegant solution for gap-filling. Firstly, aerodynamic method is stable during the periods when available energy is close to zero and during the time of precipitation which are typical periods when the BREB method fails (HALLIWELL and ROSE, 1989). The aerodynamic method on the other hand will not provide satisfactory results in light, variable winds, when anemometers approach their stall speeds (HALLIWELL and ROSE, 1989). Although this does not directly affect the BREB method in principle, such periods should be take with very high caution, because one of the basic assumption of BREB method can fail. Namely, the equality of the turbulent transfer coefficient looses the validity (FOKEN, 2008a). Therefore, it also recommended to measure the wind speed at least at one level or two levels in order to recognise that the sufficient turbulent regime is developed (FOKEN, 2008a), and thus the combination of these two gradient method itself offers. However, this approach might be complicated above high canopies because the aerodynamic method is very error prone in measuring in the roughness sublayer which usually extents to a few roughness lengths from the tops of the roughness elements. It makes the use of the aerodynamic method in such conditions practically impossible (THOM, 1975) and only one way how to deal with it is to use complicated corrections (CELLIER and BRUNET, 1992, MÖLDER et al., 1999). 179

6.2.3 Fetch and footprint The analysis based on DUGAS et al. (1999) approach showed that the fetch is mostly not critical for BREB method above the poplar plantation, but it is very probably critical above the turf grass. Considering the prevailing wind direction (from west to north-west) and looking at aerial image (Fig. 4) where the particular BREB systems are depicted or looking at the Tab. 6 with the distances from the particular azimuths, it is obvious, that the fetch should be more critical for the turf grass than for the poplar plantation. Another reason might be also the fact that surfaces with larger roughness (as the poplar plantation in our case is) compared to those with smaller one (turf grass) are more rapidly equilibrated within a new internal boundary layer (BRUTSAERT, 1982, HEILMAN and BRITTIN, 1989, STANNARD, 1997). Model calculations indicate that the lowest 10 % of the internal boundary layer is in equilibrium with the surface for an aerodynamically smoothto-rough transition, and the lowest 5 % for a rough-to-smooth transition (MUNRO and OKE, 1975, BRUTSAERT, 1982). The equilibrium sublayer is also defined as the region where the momentum flux density is within 10 % of the value at the surface. This has led to formulations of minimum fetch-to-height ratios to insure that measurements are made in the equilibrium layer. Ratios ranging from 10:1 (PANOFSKY and TOWNSEND, 1964) to 200:1 (DYER, 1965) have been recommended with 100:1 considered adequate for most measurements (ROSENBERG et al., 1983). Based on theory of the development of internal boundary layer, the comparison of measurement in different heights and thus with different footprints is a possible way how to analyze the influence of the limited fetch (HEILMAN and BRITTIN, 1989, STANNARD, 1997). The BREB system above the poplar plantation has in fact three levels for measuring the gradients of temperature and vapour pressure. But from technical reasons (insufficient height of the mast) during the 2008 and 2009 only the two upper levels were adequately above the canopy, whereas the lower level was already inside the roughness sublayer below the top of the canopy and thus was not usable for correct measurement. However, during 2010 when the poplars were harvested, the sensors were pulled down and the measurement with all three levels was possible. The lowest level was usually ~0.2–0.3 m above the surface (firstly bare soil, later canopy), the middle level 1 m higher and the highest level was 3 above the lowest level. Nevertheless, the gradients between the middle and the highest sensor were usually very low (too high above zero plane displacement with respect to the accuracy of the sensors) and data were not interpretable. Therefore, only two combinations of gradient measurement were used for comparison. Namely, first gradient measurement comprises the highest and the lowest levels (3 m vertical distance) and the second gradient measurement comprises the middle and the lowest sensor levels (1 m vertical distance). The Fig. 77 depicts the scatter view of the latent heat fluxes calculated from these two different gradients. The fluxes are relatively in good agreement and there is practically no systematic over/underestimation which confirms the hypothesis that the measurements above the poplar plantation are not negatively influenced by inadequate fetch and are both situated mostly in the equilibrium sublayer of the internal boundary layer. It should be noted here that no sharp discontinuity exists in the profiles of temperature and humidity at 180

the boundary between the equilibrium sublayer and the upper internal boundary layer. The profiles are smooth over the entire internal boundary layer, and the definition of the height of the equilibrium sublayer is strictly arbitrary, based on allowable deviations of profiles from their equilibrium (i.e., unlimited fetch) values (STANNARD, 1997).

LE (W m-2) - upper sensor in 4 m

400

300

y = 0.98x R2 = 0.96 MBE = 1.06 W m -2 RMSE = 16.59 W m -2

200

100

0

-100 -100

0

100

200

300

400

LE (W m-2) - upper sensors in 2 m

Figure 77: The relationship between latent heat fluxes measured from different heights above the poplar plantation. Namely, the upper pair of gradient measurement with the upper arm at 4 m and the lower at 1 m, and the lower gradient measurement with the upper arm at 2 m and the lower arm at 1 m above the ground (mean height during the growing season). The points include all measured data during the growing season 2010 after all quality controls described previously. By using the Stannard equation, the effective height for the lower variant is situated at ~1.45 m above the ground and thus within the distance of 50 m originated more than 85 % of measured fluxes and within the distance of 100 m even more than 90 % of the fluxes according to the simple footprint model proposed by GASH (1986). The upper variant has the effective height at 2.9 m above the ground and thus at the distance of 50 m more than 40 % of the fluxes and within 100 m from the point of measurement approximately 65 % of fluxes comes from the area of interest in near neutral conditions. Above the turf grass, where according to the previous results, the inadequate fetch was the limiting factor, there were only two levels for gradient measurement and thus the similar verification was not possible. However, the Penman-Monteith which should have theoretically different footprint can be used and the resulting fluxes compared with those from BREB system. We have already had the opportunity to see such comparison on diurnal bases where the showed 16 % overestimation compared to FAO reference . However, some authors (e.g. CALIANDRO et al., 1990, RANA et al., 1994, STEDUTO et al., 1996, STEDUTO et al., 2003) described significant underestimation of PenmanMonteith with fixed surface resistance (70 s m-1) calculated on the daily basis. The application of Penman-Monteith with variable surface resistance on hourly or other short 181

time span showed usually better agreement with other methods. One of such approach can be the already used Todorovic´s dynamic model for surface resistance of the reference grass (TODOROVIC, 1999, STEDUTO et al., 2003). The approach of TODOROVIC (1999) introduces a variable surface resistance modelled as a mechanistic function of climatic variables, without additional input variables than those required by the FAO-56-PenmanMonteith method, nor calibration. Moreover, STEDUTO et al. (2003) found better results with Todorovic´s variable resistance also in case of the diurnal input data compared to widely used FAO-56 concept according to ALLEN et al. (1998). Thus it might be worth considering to embody the more physically based variable surface resistance into water balance model based on the reference evapotranspiration concept. Before the comparison with the halfhourly measurement from BREB method let´s have a look briefly on the theoretical basis of different footprints. In fact, the PenmanMonteith model is not nothing else than the gradient measurement of Bowen ratio where the lower level is the so-called ―big leaf‖ with known temperature and vapour pressure lying just at the zero plane displacement. The surface resistance informs us about how easy the ―big leaf‖ transpires water (driven especially by stomatal, plant and root resistance) and the aerodynamic resistance tells us how easy this water in the form of water vapour is going up to equilibrate the upper atmospheric layers so it is the integration of the eddy diffusivity between the ―big leaf‖ and the upper level of measurement. However, such equilibration is rarely achieved because the boundary layer is not cupped as already mentioned with relation to Priestley-Taylor model theory. If we now use the position of the lower level where the imaginary lower sensors are placed, the geometrical mean or the function given by STANNARD (1997) can be used to estimate the effective height of measurement and subsequently the footprint. Considering the upper sensor is at height 2 m above the ground and the zero plane displacement ~0.04 m, the effective height according to the Eq. 123 given by STANNARD (1997) is ~0.3 m. So you can easily imagine what would be the footprint if we placed one-point measurement which is practically only the eddy covariance (if we omit the practical impossibility of such adjustment) at the height 0.3 m. By employing the one-dimensional footprint model given by GASH (1986) it can be found that 60 % of measured fluxes are coming from the distance less than 10 m, 75 % from less than 20 m and 85 % from approximately 40 m. Being more conservative and assuming that the effective height is rather at 0.5 m, 30 % of fluxes from the distance less than 10 m and 55 % and 75 % from the distances ~20 and ~40 m respectively will be obtained. In any case, it is apparent that the footprint of Penman-Monteith might be considerably different and the comparison of two ―measurements‖ makes sense. The Fig. 78 shows very good agreement between the latent heat fluxes measured by BREB and calculating by Penman-Monteith with variable surface resistance according to TODOROVIC (1999). Due to the fact that there is no systematic difference between the two methods with different footprints it is a question, if the fetch is really limiting factor for the measurement above the turf grass. Such optimistic conclusion might be supported by investigations by (HEILMAN and BRITTIN, 1989) who found that when the Bowen ratio is small (which is our case) the method can be used successfully at fetch-to-height ratios as low as 20:1, much less than the often quoted value of 100:1.

182

To conclude this issue, it should be noted here, that the using of Penman-Monteith model is so far just provisional verification and measurement with more levels (e.g. with one more level in 1 m height) would be more valid and acceptable. In the same time, with respect to the magnitude of the gradients above such short canopy as discussed earlier, it should be still possible. However, taking into account such provisional analysis, it shows that the measurements above the turf grass are more or less relevant and should not be ignored. 500

LE (W m-2) - BREB

400

y = 1.00x R2 = 0.97 MBE = 2.26 W m-2 RMSE = 19.92 W m-2

300

200

100

0

-100 -100

0

100

200

300

400

500

LE (W m-2) - PM with Todorovic s rc

Figure 78: The relationship between latent heat fluxes measured by BREB above the turf grass and those calculated from Penman-Monteith equation with dynamic surface resistance model proposed by TODOROVIC (1999). The points include all measured halfhourly data during the growing seasons 2008–2010. At the end of this subchapter one more possible problem should be mentioned. In real world, fetch is often limited, and the measurements of temperature and humidity are somehow affected by fluxes of sensible heat and water vapour from an upwind contaminating surfaces, i.e., some of the measurements are made above the equilibrium sublayer (STANNARD, 1997). But as already mentioned there is no sharp discontinuity in the profiles of the scalars at the boundary between the equilibrium sublayer and the upper internal layer. To deal with the problem of limited fetch, many authors place the lower sensors just above the top of the canopy (like in this thesis, LINDROTH and HALLDIN, 1990, BERNHOFER, 1992, LINDROTH and IRITZ, 1993, BAR et al., 1994 and others). There are several investigations showing that the so-called universal stability correction functions are not valid in the roughness sublayer especially above very rough surfaces (GARRAT, 1978, RAUPACH and LEGG, 1984, HÖNGSTRÖM et al., 1989). It has several implications for the aerodynamic profile method, which utilize the universal stability functions (IRITZ and LINDROTH, 1996). However, RAUPACH and LEGG (1984) argued that for a dense homogeneous canopy the turbulent diffusion coefficient for heat and vapour, respectively, should be the same because the source-sink distributions are similar. This implies that the Bowen ratio based on the assumption of the equality of the eddy diffusivities remain valid 183

also within the roughness sublayer (IRITZ and LINDROTH, 1996). This was also demonstrated by CELLIER and BRUNET (1992) or later confirmed by STANNARD (1997). However, this equality can be disturbed after a rain event causes the soil to be wet. Then most sensible heat will be generated by the sunlit leaves while evaporation will occur from the leaves and the soil and thus the assumption of equality of sources and sinks of scalars is no longer valid (BALDOCCHI, 2004). 6.2.4 Daily course of actual evapotranspiration The of both covers showed very comparable diurnal courses with similar absolute values within the selected ten days periods during the second half of the growing season 2008. However, the turf grass reached usually higher midday peaks during the hot and sunny days with values approximately 0.1 mm per hour higher than the poplar plantation. The peaks amounted more than 0.7 mm per hour but were never higher than the available energy which was during the higher evapotranspiration rates (e.g. from the 27th to 29th of July) close to 600 W m-2 which represent 0.88 mm per hour (recalculated to water column). Thus it is obvious that both latent and sensible heat flux were always upwards during the nice sunny days. They usually changed their signs near the sunset and then again after sunrise, where the sensible heat flux always became positive later in the morning and on the contrary became sooner negative compared to the latent heat flux. During the midday, the sensible heat flux was most of the time positive with only rare exceptions just after precipitation The higher values of of poplars during the day with precipitation might be the effect of higher rates of evaporation from interception which is proportional to (CIENCIALA et al., 1994, IRITZ et al., 2001). There has been only limited number of studies dealing with of turf grass compared to those with poplars. However, e.g. according to BREDE and DUNICH (1984), cool-season grasses can reach and maintain an of 3 or greater after 2 to 3 months of development. KOPEC et al. (1987) reported a range of values for turf grass from ~1 to 3.5. AGATA et al. (1989) reported of turf grass to be ~2, with the value reduced by ~10 % following moving. Later, BREMER (2003) found of turf grass within the range ~2 to ~3 and MILESI et al. (2005) described reaching value 1.5 in time of each cut which reduced its value by 20 %. On the contrary, the maximum seasonal of the poplar plantation is usually two times or three times higher (ZAVITKOVSKI, 1983, CEULEMANS et al., 1996a, HEILMAN et al., 1996, HALL et al., 1998, LIBERLOO et al., 2006) as was already demonstrated in the presented results. Except the effect of interception, there might be one more explanation of the higher of poplars during the days with lower available energy or those with rain event. As has been already discussed, the polar plantation has much more higher aerodynamic roughness compared to the short turf grass canopy. The higher roughness means lower aerodynamic resistance which enhance evaporation. This is especially important during the periods when the surface resistance is zero, just like after precipitation when canopy is wet (HALL, 2002). In addition, the better turbulent mixing within the boundary layer of the poplar plantation increases the relative importance of their stomatal control. This on the other hand can also explain the differences in the diurnal peaks during the nice sunny day with high 184

atmospheric evaporative demand. However, we should bear in mind, that the higher rates of evapotranspiration after the precipitation might be the effect of failure of the basic assumption of the BREB method, namely the equality of the eddy diffusivity coefficients for heat and water vapour disrupted by differential source-sink location. This possible unhomogenity of sources and sinks distribution is generally much more pronounced in tall canopies than in short canopies (BALDOCCHI, 2004). In this case, only the measurement, independent on the Bowen ratio similarity assumption, like eddy covariance would solve this question. Anyway, this is not the case of the lower midday peaks in of poplars during the days with high vapour pressure deficit. Entering the next year 2009, we can observe similar characteristic as mentioned above. Namely the of the turf grass with higher midday peaks during hot sunny days with high vapour pressure deficit. The effect of the high vapour pressure deficit (diurnal peak > 1.6 kPa) was evident e.g. during the free consecutive days from the 24th to 26th of May. On the other hand, the effect of low vapour pressure (diurnal peak ~ 0.3 kPa) and the noon rain event can be observed at 29th of August with characteristically higher of the poplar plantation. At the first of September was recorded higher of the poplar plantation. During this day the vapour pressure deficit reached up to ~1.2 which is not too high and during this period was also reached the maximum which might the possible explanation. We should also take into account that the turf grass was regularly moved (usually in ten days intervals, but the exact dates were note registered) and thus it should be sometime the reason of its lowered evapotranspiration due to reduced . Also at the end of September higher of the poplar plantation during the cloudless days with relatively high temperature were recorded. As a possible reason, the effect of reduced soil moisture during the dryer period at the end of summer, which followed after two exceptionally wet months, was hypothesized. It was assumed that during this water unlimited conditions the turf grass concentrated most of its root system in the upper superficial soil layer. As a consequence, with coming of dryer period, the turf grass was not so well adapted to uptake the water from the lower layers compared to poplars which have generally slightly deeper root system. The diurnal peak of vapour pressure deficit culminated around 1kPa so thus the poplars did not have to reduce the transpiration yet. Nevertheless, these days were gap filled due to relative humidity sensor failure and it is also very probable that the gap filling method were not able to fit the shape with characteristic departure during the midday peaks. From that reason, also the using of correctly parameterized Lohammar equation which takes into account the response of stomata to increasing vapour pressure deficit for specific species might provide better results compared to linear regression between evapotranspiration according to Penman which has in fact fixed zero surface resistance. Following year 2010 revealed several specific features. Firstly, in the cut poplar plantation there was observed typical effect of precipitation on the evaporation from bare soil which is generally very sensitive to the soil moisture of the superficial soil layer ALLEN et al. (2005). It was already obvious during the April and part of May 2009 before foliating and during the low . The variability of energy partioning due to varying evaporation depending on irregular wetting by precipitation is a typical feature of initial 185

crop development stage (ALLEN et al., 1998). With increasing the soil evaporation tends to be more and more negligible and the evaporation to transpiration ratio more stable. Similar patterns were described by LINDROTH and IRITZ (1993) for willow stand in Sweden. Namely, the evapotranspiration was small compared to sensible heat flux especially due to low monthly precipitation and only little developed ( < 1) during the whole May. Later in June, the developed and proportionally the evapotranspiration increased. LINDROTH and IRITZ (1993) also reported negative sensible heat flux and thus latent heat exceeding the net radiation during the day, which, however, has never been observed in here presented study. Generally, the absolute values from their study cannot be compared due to the different latitude. For example, the net radiation in LINDROTH and IRITZ (1993) culminated around 400 W m-2 whereas in our study near 600 W m-2 which represents ~0.6 and ~0.9 mm per hour respectively assuming all energy used for evaporation. Looking into lower latitudes for comparison, similar midday peaks of evapotranspiration above extensive stand of aspen (P. tremuloides) in central Canada measured by eddy correlation reaching up to 0.6 mm hour- 1 were reported by BLACK et al. (1996). MEIRENSONNE et al. (1999) reported midday peak of canopy sap flow in hybrid poplar stand (P. trichocarpa x P. deltoides) in Belgium to be almost 0.5 mm hour-1 during mid August. FINCH et al (2004) showed transpiration of two willow varieties measured by stem heat balance exceeding 0.5 mm hour-1 in south England around mid July 2002. TRICKER et al. (2009) found the June midday peak of evapotranspiration by P. euramericana measured by eddy covariance method close to 0.5 mm hour-1 and slightly lower values of transpiration measured by stem heat balance in central Italy. Another important issue discussed by many authors is exceeding of the potential Penman open water evaporation. 600

600

n = 180

Poplar plantation 500

500

n = 917

LEact (W m-2)

n = 121

Turf grass

n = 1563

400

400

300

300

200

200

100

100

0

0

n = 6813

n = 6531 -100 -100

0

100

200

300

400

500

600

-100 -100

LEPenman (W

0

100

200

300

400

500

600

m-2)

Figure 79: Scatter views on the relationship between latent heat flux calculated by the original Penman (1948) formula and those measured by BREB systems – above the poplar plantation on the left and above the turf grass on the right hand side. The picture includes the diurnal (Rn > 50 W m-2) half hourly data within the growing seasons 2009–2010. The inclined solid line depicts the boundary between exceeding the Penman evaporation. The dashed vertical line differentiates the values exceeding the Penman evaporation into two groups: > 200 W m-2 and < 200 W m-2. 186

Earlier, when most evaporation studies were made over short vegetation, the term ―potential‖ was conceived as a ―maximal‖ evaporation rate which could not be exceed (IRITZ and LINDROTH, 1996). However, later studies have shown that this is not the case and the explanation is simply that the enhanced turbulent diffusion for taller vegetation, compared with that of an open water surface or a short crop, will give rise even four time higher rates than so-called ―potential‖ when the surfaces are wet after rain (LINDROTH, 1993). Moreover, LINDROTH and IRITZ (1996) reported 1.3 higher rates of actual evapotranspiration compared to Penman open water evaporation reached even by dry canopy of willow SRC. They explained these high rates by high canopy and aerodynamic conductance. The Fig. 79 shows the exceeding of the Penman open water evaporation by the actual evapotranspiration for this study. It is obvious, that taking into account the half hourly values, the Penman open water evaporation was exceeded by both of the canopies. Moreover, it is evident that the higher exceeding of the Penman open water evaporation was reached by the poplar plantation (by more than 200 W m-2 in some cases). This priority was linked with the rain events and with the subsequent evaporation of the intercepted water which is more pronounced in case of the poplar canopy due to higher and higher aerodynamic roughness. On the other hand, occurrence of values of above the poplar plantation which are much lower than the Penman evaporation is the effect of the low or zero transpiration and only very limited evaporation from the soil (low soil moisture of the superficial layer) during the time when the poplars were not or only little foliated in the spring or just after coppicing in 2010 when the plantation comprised mainly bare soil till the end of June. Generally, most of the values were lower than the Penman evaporation – namely 86 and 80 % for the poplar plantation and the turf grass respectively. Using only the values when the Penman evaporation was higher than 200 W m-2, which can be considered as the conditions with some substantial evaporative demand including the midday, only ~2.5 and ~1.5 % of values exceeded the Penman open water evaporation in case of the poplar plantation and the turf grass respectively. Finally, considering the low evaporative demand periods defined by the Penman evaporation lower than 200 W m-2, 12 and 18.5 % of values exceeded the ―potential‖ evaporation by the poplar plantation and the turf grass respectively. 6.2.5 Monthly course of actual evapotranspiration The daily sums of of the mature poplar plantation and the turf grass showed very comparable values during the most of the growing season. The only significant difference was recorded in April 2009, when the poplars were not foliated yet and thus the evapotranspiration was driven mostly by soil evaporation. The whole season 2010 was characteristic with significant differences between both of the covers due to zero or low of the poplar plantation and substantial portion of soil evaporation. The comparable values of evapotranspiration of the turf grass (or generally grass or grassland) and the poplar plantation (or short rotation coppice based on willow species) are in strong contrast with some of the previous studies. For example, HALL et al. (1996) and HALL et al. (1998) reported canopy transpiration, based on upscaled sap flow measurements using different 187

techniques, to reach up to more than 10 mm per day which exceeded the reference crop evapotranspiration (SHUTLEWORTH, 1993) and also the Penman open water evaporation more than twice in non-irrigated and unfertilized SRC in the conditions of south England. They found comparable values of such high transpiration for both willow and poplar species and they found them to be highly exceeding the evapotranspiration of other agricultural or tree crops including grassland. They ascribed such high water consumption to a well-established root system and high stomatal conductance and low stomatal control even in case of large atmospheric humidity and soil water deficit. Mean daily transpiration based on sap flow stem heat balance and heat pulse velocity techniques over the period 20th June to 19th July 1995 reported by HALL et al. (1998) was 6±0.5 mm day-1 (mean value ± standard deviation) for poplar clone Beaupré (P. trichocarpa x P. deltoides; three-year-old shoots on four-year-old stools) whereas in this study the mean evapotranspiration (based on BREB method) 3.2±1 and 2.8±1.1 mm day-1 in June and 3.6±1 and 3.6±1 in July 2009 for the poplar plantation and the turf grass respectively was recorded. Moreover, HALL et al. (1997) described even higher transpiration rates for Beaupré (three-year-old shoots on seven-year-old stools) measured by stem heat balance technique reaching 9±2 mm day-1 between 18th to 27th June 1993 at other place in southern England. Similar high transpiration rates for poplar clone Beaupré was reported by ALLEN et al. (1999) with mean over 18-day period with adequate water supply in June 5.0±1.8 mm day-1 and maximum reaching 7.9 mm day-1. In the same time poplar clone Dorschkamp (P. deltoides x P. nigra) showed half transpiration rates 2.4±0.9 mm day-1 suggesting large clonal variation of water consumption which was explained with maximum 4.9 of Beaupré, twice higher than for Dorschkamp. Except the exceeding of the daily Penman open water evaporation, ALLEN et al. (1999) found daily transpiration sums higher than the water equivalent of net radiation by Beaupré leading to the conclusion that some transpiration had to be sustained by the advection of sensible heat from surrounding farmland. Neither exceeding of the net radiation water equivalent nor that of the Penman open water evaporation has been observed in daily totals of during the growing seasons within the here presented study. In the continuing study of FINCH et al. (2004) dealing with water use of other energy crops (Miscanthus, switch grass and willow SRC) the authors stated the water demand of poplar SRC is very high. However, in the same time they also noted that these results should be interpreted with caution as it is likely that varieties could be or are available that would have lower water use which is probably comparable to that of willow SRC. FINCH et al. (2004) attributed the very high transpiration rate of poplars due to the lack of any control on stomata in response to the atmospheric evaporative demand. The same conclusions have been already previously given by HALL et al. (1998) and ALLEN et al. (1999) who found no or weak correlation between stomatal conductance and atmospheric humidity deficit. To express the dependency of surface conductance on the atmospheric demand and to estimate the character of the overall stomatal response the Lohammar equation (LOHAMMAR et al., 1980) was embodied into the Penman-Monteith formula (MONTEITH, 1965) and parameterized by fitting the model to the measured latent heat flux by using data from July to August 2008 for both poplar plantation ( ~ 6.5–7.5) and the turf grass. In fact, the Lohammar equation (Eq. 125) is originally developed for canopy (or 188

stomatal) conductance and when applied on data of evapotranspiration, the soil evaporation and interception should be deduct or replaced (SHUTTLEWORTH and CALDER, 1979, LINDROTH, 1993, GRELLE et al., 1999, SOMMER et al., 2002). Therefore, only the precipitation-free days and only the values when the net radiation was higher than 200 W m-2 were used in order to include mainly the periods when the canopy was dry. Although, the influence of soil evaporation was neglected, the parameterization by last square fit procedure yielded satisfactory results with the correlation coefficients 0.78 and 0.94 and the particular parameters (all statistically significant with < 0.001) gmax = 0.121±0.022 and 0.111±0.023 m s-1, R0 = 363±76 and 1063±223 W m-2, and a = 4.66±0.79 and 1.26±0.24 kPa-1 for the poplar plantation and the turf grass respectively. The graphical visualization of the parameterized Lohammar equation describing the dependency of the surface conductance on the vapour pressure deficit and solar radiation is depicted at Figs. 80 and 81. First thing which is evident is that the poplar plantation with higher has much higher surface conductance during low vapour pressure deficit. As the vapour pressure deficit rise, the surface conductance falls very sharply and around the vapour pressure deficit of 0.5 kPa the surface conductance of the turf grass tends to be higher, if the global radiation is higher than ~400 W m-2. This is an interesting point which can explain many situations when the of the poplar plantation is higher during the days with low evaporative demand and on the other hand is lower during the hot and sunny days, as it was already demonstrated on the diurnal courses of . In the same time, it also suggests that the stomatal control of the poplar plantation is higher than those of the turf grass. However, this conclusion must be taken with high caution, because the lower response of the turf grass to vapour pressure deficit might be also the effect of higher soil evaporation which is in no way physiologically controlled. In the same time, it seems that especially the hybrid clone Beaupré used in the study of HALL et al. (1996) is extremely water demanding compared to clone J-105 used in our study. It should be noted, that HALL et al. (1996) used in their study also neutron probes (the most precise of the indirect methods) to measure soil moisture (2 m in depth) and by calculating the water balance estimated the . However, by comparing this approach with transpiration derived from sap flow measurements they found the from the -1 water balance is maximally only slightly over 5 mm day , i.e. significantly less than the transpiration based on upscaled sap flow and moreover it is quite comparable with Penman-Monteith reference crop (grass) evapotranspiration. They explained this disagreement by the possible extracting the water by roots from lower layers which was supported by roots excavated even in depth o 3 m (the lowest level attainable by the excavator). On the other hand, the soil water profiles indicated that there was no uptake from below 1.6 m depth and further 55 % of the total uptake came from the top 0.5 m layer and 80 % from the top metre.

189

Figure 80: The surface conductance of the poplar plantation and its dependency on global radiation and vapour pressure deficit as a result of parameterization of Lohammar equation during precipitation-free days in July and August 2008.

Figure81: The surface conductance of the turf grass and its dependency on global radiation and vapour pressure deficit as a result of parameterization of Lohammar equation during precipitation-free days in July and August 2008.

190

Moving from high water consuming poplars reported in England to the Scandinavian Peninsula we can use information from very thorough study dealing with water use of intensively managed (irrigated and fertilized) willows SRC (mostly Salix viminalis) near Uppsala in south Sweden. GRIP et al. (1989) using lysimetric measurement reported the highest in July with mean 4 mm day-1 for the normal stand with suboptimal irrigation and fertilization with a production of 12 tonnes (DMC) per hectare, and 4.5 mm day-1 for hypothetical optimally irrigated and fertilized stand with productivity of 20 tonnes (DMC) per hectare. Similarly, PERSSON and LINDROTH (1994) described the highest monthly mean (four years) of daily in July with 4.4 mm day-1 by using physically based model SOIL. As already mentioned, also in the here presented study July was the month with the highest rates with daily mean 3.5±1.2 mm day-1 and 3.6±1 mm day-1 in 2008 and 2009 respectively. Although PERSSON and LINDROTH (1994) and later LINDROTH and BÅTH (1999) reported simulated values of as high as 9.4 mm day-1 it must be stressed that these values were in high disagreement with simultaneous measurements by BREB method (used both for calibration and verification) reaching only to 5.7 mm day1 . The maximum of willow SRC measured by BREB reported in later study by LINDROTH et al. (1995) were usually around 6 mm day-1 with one exceptional high value exceeding 7 mm day-1 during midsummer. Moving to the lower latitudes, PETZOLD et al. (2010) described maximum daily rates of transpiration based on sap flow measurements (tissue heat balance) by poplar clone Max 1 (P. nigra x P. maximowiczii Henry) in Germany reaching 5.7 and 6.7 mm day-1 with seasonal mean (April to September) 2.3 and 2.2 mm day-1 in 2007 and 2008 respectively. These results are comparable to our plantation which reached daily mean of 2.8 mm -1 day within the growing season 2009 assuming that transpiration is ~70 % of the evapotranspiration. Also very similar results were reported by MEIRESONNE et al. (1999) who found mean daily (April to September) transpiration 1.9 with maximum 4.9 mm day-1 in 13 years old stand of P. trichocarpa x P. deltoides cv. Beaupré in Belgium based on sap flow technique (heat field deformation). In another study carried out by TRICKER et al. (2009) in central Italy at the poplar plantation (P. deltoides x P. nigra, clone I-214), transpiration measured by sap flow (stem heat balance) and validated by eddy covariance reached 2.2±0.4 mm day-1 over 40 days from the end of July to the beginning of September with maximum 7.2 mm day-1 during the season with the initial re-growth of the plantation with a closed canopy following coppice. For comparison, in our study within the year 2010 following coppicing the August was 2±1 mm day-1. During the years with completely closed canopy, the August was 3.3±1 and 2.9 ±0.9 in 2008 and 2009 respectively. Note that TRICKER et al. (2009) found 23 and 15 % increase in the period average and maximum respectively in POPFACE (poplar free-air carbon dioxide enrichment) treatment (elevated CO2 to 550±70 ppm). Unique lysimeter-based water use study was carried out by GUIDI et al. (2008) in Mediterranean conditions near Pisa (Italy). The experiment comprised both willows (Salix alba – clone SI62-059) and poplars (P. deltoides – clone Lux) used as irrigated (soil moisture maintained close to the field capacity) vegetation filter under fertilized and unfertilized conditions during two consecutive growing seasons after planting. The study revealed that the high water consumption of poplar and willow SRC is more affected by 191

the availability of nutrients for the plants than by their specific ecological behaviour. In case of the willows, the mean within the first growing season (from May to September) reached 3.6 with maximum 7 mm day-1 and 5.9 with maximum 13.4 mm day-1 for unfertilized and fertilized treatment respectively. In the following season mean of -1 willows increased to 4.5 with maximum 11.4 mm day and to 7 with maximum 21.7 mm day-1 again for unfertilized and fertilized treatment respectively. In case of the poplars, GUIDI et al. (2008) recorded mean of the first growing season 2.9 with maximum 5.0 mm day-1 and mean 4.8 with maximum max 8.4 mm day-1 for the two nutrient treatments respectively. Second growing season brought again higher values with mean 3.7 and maximum 6.2 mm day-1 and with mean 8.9 and maximum 16.4 mm day-1 for non-fertilized and fertilized treatment respectively. For comparison of their results with our data (unfertilized and not irrigated) within the same period (May to September), mean within the first year of the second rotation in 2010 was 1.4 (though 1.7 considering only the period when > 0) with maximum 4.1 mm day-1 and within the second year of the second rotation in 2011 (data not shown) mean was 2.8 with maximum 6.4 mm day-1 (the absolute record value within this study). Taking into account the differences given by geographical location which can be climatologically well described by the reference evapotranspiration with mean (during two compared years) 2.5 and maximum 5 mm day-1 and mean 4.4 and maximum 5.7 mm day-1 in this and the GUIDI et al. (2008) study respectively, we can conclude that the results are quite comparable with the unfertilized albeit irrigated treatment. This might be explained with quite adequate water supply from precipitation in our study where no evident water stress has been observed yet. Further, the results of GUIDI et al. (2008) brought very important conclusion regarding nutrient supply and its effect on the water consumption which is in the same time very probably the main reason of relatively high water use described in many studies from intensively managed (irrigated and fertilized) willow SRC from Sweden. The effect of the nutrient availability and its levels on water consumption was later proved by PISTOCCHI et al. (2009) and it is in agreement with the previous studies by SCHULZE et al. (1994) and LINDROTH and CIENCIALA (1996) describing positively affected maximum stomatal conductance and assimilation by foliar nitrogen status. It must be noted that the extremely high water consumption described by GUIDI et al. (2008) would not be possible without relatively strong effect of advection of sensible heat from adjacent fields with negative Bowen ratio which was not considered. Although the lysimeters used by GUIDI et al. (2008) were surrounded by other trees simulating the density 10,000 tress ha-1 in order to avoid clothesline effect, it was probably not sufficient – similarly like in 2.6 ha large SRC plantation described by LINDROTH and IRITZ (1993) or 1.8 large SRC in ALLEN et al. (1999). I calculated for our locality that if the whole net radiation would be used to evaporate water, rather than heating the air, soil and crop, then the maximum evapotranspiration can be in average (calculated from all of the growing seasons) 1.6 higher than the (reference evapotranspiration). However, GUIDI et al. (2008) reported values of even as high as 4.3 and 5.3 times higher than for poplars and willows respectively. The same explanation is probably valid for very high values 16.4±1.4 -1 (25 June to 17 August) with maxima between 20 to 23 mm day reported by PAULIUKONIS and SCHNEIDER (2001) for willow (Salix babylonica) grown in lysimeters along the 192

shoreline of Oneida lake (New York, USA), with plants intentionally exposed to wind and sun in order to maximize rates and find the differences between species. Looking into other overseas countries, e.g. HANSEN (1988) reported of 2 to 5 years old stand of P. tristis x P. Balsamifera in Wisconsin (USA) during the three months long part of the growing seasons (June to August) to be within the range of 4.4 to 4.8 mm day-1. HANSEN (1988) used lysimeters where the soil moisture was kept through irrigation near the level of field capacity, however, the water loss due to deep seepage was not measured and thus the water consumption served rather as an upper bound on water needs of hybrid poplars. HINCKLEY et al. (1994) reported mean transpiration 3.6 with maximum 4.8 mm day-1 during week period at the beginning of August in four years old stand of P. trichocarpa x P. deltoides in Washington State (USA). BLACK et al. (1996) described in 70 years old aspen forest (P. tremuloides) with peaks within 5–6 mm day-1 measured by eddy covariance in Saskatchewan (Canada). SAMUELSON et al. (2007) investigated two native male P. deltoides clone (S7C15 and ST66) within the factorial experiment combining irrigation and fertilization in plantation with density 1333 trees ha-1 in South Carolina (USA). They found that transpiration in the fertilized and irrigated treatment measured by sap flow (heat dissipation method) ranged between 0.3 and 1.8 mm day-1 which is comparable e.g. to results reported by HINCKLEY et al. (1994) or MEIRESONNE et al. (1999) if we consider the lower density and relate the transpiration to the basal area (to total stem cross sectional area at breast height per ha). However, more important result of SAMUELSON et al. (2007) was that the fertilization has a greater effect on growth and transpiration than the irrigation. By summarizing the above mentioned, it seems that the results of this thesis are relatively well within the range described by many authors with respect to extensive management without fertilization and irrigation and taking into account the climatic conditions. In the same time, it is evident that the lysimetric study by GUIDI et al. (2008) represents good material regarding the physiological potential of transpiration of SRC, however, it is not likely that the advection of sensible heat can sustain such high rates of evapotranspiration in larger scale plantation with extended fetch. By comparing all the mentioned studies, it seems that the SRC water use reported by HALL et al. (1996) and later in HALL et al. (1998) and ALLEN et al. (1999) represent the upper bound of transpiration of SRC grown at the larger scale areas. Within this context, it is noteworthy that the site described by HALL et al. (1996) where the such high rates of transpiration was recorded was previously permanent pasture, non-irrigated and even unfertilized. 6.2.6 Yearly course of evapotranspiration The comparison between water consumption of SRC and grassland based on the same measurement technique is quite rare throughout the available literature. The main reason is that due to the small areas of SRC and with this linked fetch limitations they are usually treated more often by sap flow (HINCKLEY et al., 1994, HALL et al., 1998, ALLEN et al., 1999, MEIRESONNE et al., 1999, PETZOLD et al., 2010), which is, however, not usable for grassland, than by micrometeorological methods like BREB (LINDROTH and IRITZ, 1993, IRITZ and LINDROTH, 1996) or eddy covariance used only in the most recent studies by 193

MIGLIAVACA et al. (2009), TRICKER et al. (2009) and ZONA et al. (2011). Therefore, many studies dealing with water balance of SRC usually compare their water use with some kind of reference evapotranspiration (which can be de facto considered as the potential evapotranspiration of grassland or its linear function) or water balance model. For explanation reference evapotranspiration, , as a nowadays standard given by ALLEN et al. (1998) should refer to evapotranspiration of grass with not short of water and it is practically the same like the reference crop evaporation given by SHUTTLEWORTH (1993). Further, this recently widely used reference evapotranspiration is approximately 0.8 of the Penman open water evaporation during the period May to August (PENMAN, 1948, PENMAN, 1953). For example PERSSON and JANSSON (1989) reported modelled of willow SRC (validated by soil water potential measurement) in Sweden within the range 370 and 420 mm which exceeded Penman open water evaporation 335 mm during the summer period June to September. Also GRIP (1981) found of willow stand measured by small lysimeters higher than calculated Penman open water evaporation, especially in the late summer and early autumn with 5–40 % excess, which in other word means exceeding of potential of grassland by more than 25 %. PERSSON and LINDROTH (1994) described simulated cumulative seasonal (May to October) of irrigated and fertilized willow SRC to be within the range of 420 to 580 mm whereas the Penman open water ranged between 396 to 466 mm and was exceeded in three of four simulated seasons. Although the model SOIL used by PERSSON and LINDROTH (1994) overestimated in some periods the measured by BREB as mentioned earlier, within the longer time scale it provided reliable and relevant values. Taking into account that the potential evapotranspiration of grassland is 0.8 of the Penman open water evaporation, than the of willows was by 77 to 225 mm higher. In later study by PERSSON (1995), including simulations at different and also unirrigated sites across Sweden, the within the range of 365 to 495 mm per season (May to October) was reported with the conclusion that from low productive stand agrees more with water consumption of pine forest and grass (365 to 385 mm per defined period). In other simulation study by PERSSON (1997) focused on different crops growing on a clay soil (non-irrigated) for six growing seasons (April to October) under two different precipitation regimes (―dry‖ east coast and ―humid‖ west coast in southern Sweden), the highest water consumption was found for spruce (516 mm), slightly lower for willow (497 mm) and much lower for grass ley (419 mm) and barley (347 mm). The differences in between the trees (spruce and willow) and grasses (grass ley and barley) linearly decreased with precipitation deficit (precipitation < ) similarly like later described by BUSCH (2009). In more recent study by LINDERSON et al. (2007), the authors reported much lower values of seasonal (April to October) transpiration of willows measured by sap flow (heat dissipation method) varying within the range of 100 to 325 mm for 6 different clones. However, there are many other studies from willow SRC in Sweden based on modelling in combination with measurements which confirm the high water use of intensively managed willow SRC commonly exceeding Penman open water evaporation and thus that of grassland at the seasonal or yearly temporal scale (e.g. GRIP et al., 1989, LINDROTH et al., 1994, LINDROTH and BÅTH, 1999). On the other hand, 194

commercial willow plantations are not irrigated today since it was not regarded as profitable and thus the differences in are dependent mostly on nutrient availability and the precipitation regime (PERSSON, 1997). Other modelling study dealing with comparison of water use by SRC and other crops was carried out by HALL et al. (1996) for conditions of southern England. Although the sap flow measuring campaign was not sufficiently long to provide seasonal sums, it was enough to provide good material for creating and calibrating physically-based model WUCOP based on Penman-Monteith and running in ten minutes step. The seasonal of poplar clone Beaupré simulated by WUCOP for water non-limited scenario was 772 mm per season (16th May to 26th November) and 498 mm (160 mm lower than Penman open water evaporation) in case of real soil water supply. In order to compare the water consumption of SRC with other crops, HALL et al. (1996) developed similar, though more empirically sound, model running at daily meteorological data which were more easily available for other sites. The simulation over 17 years (1967–1983) revealed that mean annual of poplar SRC growing on clay soil is 592 mm whereas that of grass 432 mm. Assuming chalk soils the mean annual was 676 and 467 mm for SRC and grass respectively. In both cases the mean annual precipitation was 686 mm and Penman open water evaporation 633 mm. However, these studies from Sweden and England are in contrast with the results of this thesis. Firstly, the of poplar SRC exceeding Penman open water evaporation at diurnal or even monthly level during the growing seasons has been never observed within this study. The of poplar plantation in 7th and 8th years of the first rotation represented usually 70 to 90 % of the Penman evaporation. Lower ratio (57 %) was during April due to low and on the other hand the highest proportion (98 %) was recorded in June 2009 when in the same time 138 mm of precipitation fell and consequently high interception loss very probably occurred. During the first year of the second rotation the ratio of to the Penman evaporation varied between 25 to 63 %. Secondly, as an analogy to no exceeding of the Penman evaporation, the results presented within this thesis do not agree with the reported differences between water use of grassland and more consumptive SRC which were either comparable or lower in case of the poplar plantation during the periods with low and also during the whole season after coppicing. The big advantage of the results from this study is that both covers were investigated by the same method, BREB, and thus any assumptions were not necessary. There might be questionable the non-optimal fetch conditions, however, no differences between of both covers with respect to wind direction were found and further, the relationship between their water use was relatively the same during the periods with favourable wind direction. Another issue might be the differences between the water use of typical grassland and the turf grass analyzed in this study. However, the new results from 2011 (Fig. 82) when another two BREB systems were employed above typical grassland (unfertilized with one to two cuts per year and mean annual productivity of hay ~3–4 t ha-1 year-1 DMC) and other poplars SRC (3rd year of the 2nd rotation with fertilization after harvest) ~500 m far from the original site dispute this hypothesis.

195

140

Water column (mm month -1)

120 100 80 60 40 20

0 Apr-11

May-11

Jun-11

Jul-11

Aug-11

Poplars 1

Grass 1

Poplars 2

Grass 2

Precipitation

Sep-11 PET

Figure 82: Overview of monthly totals of the basic water balance variables above four different covers in 2011. The label 1 means the main plots investigated within this study and 2 is for the new measurement. expresses the potential open water evaporation according PENMAN (1948). As it is obvious, in no case the of poplars exceeded the Penman evaporation and although in some months the of poplars were higher than that of the grass, in seasonal sums was still lower. Namely, the seasonal sum (April to September) of was 484 and 496 mm for the turf grass and the typical grassland respectively, and 466 and 447 mm for the poplar plantation (the main of this study) in the second and third year of the second rotation period. Finally, the precipitation was 380 mm (below normal) and the Penman open water evaporation 625 mm per growing season. Comparing our data with other studies outside the Sweden and England, we can find relatively good agreement with BUNGART and HÜTTL (2004) who used the same model SOIL like previously PERSSON and LINDROTH (1994) in conditions of Germany applied on poplar clones Beaupré (P. trichocarpa x P. deltoides) and Androscoggin (P. maximowiczii Henry x P. trichocarpa Torr. et Gray) grown on the clayey-sandy mining substrate. In the first seven years of growth the model yielded the mean annual 520 and 542 mm for Androscoggin and Beaupré respectively. PETZOLD et al. (2010) reported seasonal (April to September) transpiration for poplar clone Max 1 (P. nigra x P. maximowiczii Henry) as 486 and 463 for two consecutive years 2007 and 2008 respectively. PETZOLD et al. (2010) further modelled the interception during the same periods as 163 and 113 mm which by adding to the transpiration provide higher values compared to those presented in this study. Good agreement was reached with the results by MEIRESONNE et al. (1999) who found seasonal (April to October) transpiration in 13 years old stand of P. trichocarpa x P. deltoides cv. Beaupré in Belgium reaching 320 mm which was 70 % of potential evapotranspiration over grass. Other comparable values were described by MIGLIAVACA et al. (2009) with mean seasonal 450 mm over 3 growing seasons measured by eddy covariance in poplar plantation (P. x canadensis Moench, Clone I-214) in northern Italy. 196

The lysimetric-based study by GUIDI et al. (2008) found sums in irrigated treatment 620 and 890 mm for willow and 590 and 710 mm for poplars in two successive growing seasons respectively. For the same periods the values of in additionally fertilized treatment reached up to 1190 and 1790 mm and 725 and 1100 mm for willows and poplars respectively. Finally, DECKMYN et al. (2004) using model SECRETS simulated of general hybrid poplar yielding a mean annual value of 588 mm under normal soil water supply typical for Belgium and 750 mm under irrigation. To conclude this subchapter, according to many studies it seems that SRC based on willow or poplars have higher water consumption than traditional agricultural crops or grasslands. However, all results are not united and some of them indicate comparable or even lower water consumption of SRC compared to grassland or the reference evapotranspiration (here presented study, MEIRESONNE et al., 1999, BUNGART and HÜTTL, 2004, LINDERSON et al., 2007, MIGLIAVACA et al., 2009). The reason might be found in different options of different clones and probably in different nutrient availability. The studies from Sweden (e.g. PERSSON, 1997, LINDROTH and BÅTH, 1999) and Italy (GUIDI et al., 2008, PISTOCCHI et al., 2009) concluded that the lower water consumption is linked with the lower productivity caused especially by limitations with water or nutrient supply, where the second might be our case. One of the possible sources of errors might be also the application of different methods (upscaled sap flow, BREB, eddy covariance, water balance etc.) and thus their combination and cross-validation would be the best way to obtain deeper insight into the still open questions of water use of SRC. To deal with this issue the eddy covariance system as an independent method to validate the BREB measurements was employed newly in July 2011 as well as both BREB systems were doubled (as shown at Fig. 82). Very promising (not only from this point of view) is also new study from Belgium (e.g. CEULEMANS et al., 2010, FICHOT et al., 2011, ZONA et al. 2011), where simultaneous measurement of water vapour, energy and other fluxes by eddy covariance in combination with sap flow and traditional water balance technique are carried out over extensive poplar and willow SRC.

6.3

Other micrometeorological and eco-physiological variables

6.3.1 Bowen ratio Accordingly to the agreement between actual evapotranspiration of the turf grass and the poplar plantation, quite similar behaviour in the course of the Bowen ratio with generally higher values in case of the poplar plantation has been found. However, the same Bowen ratio of two covers does not necessarily mean the same because there might be differences in net radiation and thus in the amount of available energy. During the growing seasons 2008, 2009 and 2010 (when > 1) measurements it was found that the net radiation over poplar plantation is approximately 12 % higher than that over the turf grass. It means that if the Bowen ratio is the same, the poplars would have higher . By parameterization of the model given by ALLEN et al. (1998) the mean albedo of the poplar plantation was estimated as ~14 whereas that of the turf grass as ~22 which is the main reason of differences in the net radiation. Generally, the slightly higher Bowen ratio of the poplar plantation is balanced by higher energy availability due to lower albedo and 197

therefore the latent heat fluxes for both covers are comparable when of poplars is developed (say higher than 2 to 3) but in the same time the sensible heat from the poplar plantation is accordingly higher. The rough threshold of 2 to 3 can be observed by looking at the Figs. 47 and 48 depicting the Bowen ratio patterns and Fig. 65 depicting the evolution. Similar results were described by IRITZ and LINDROTH (1996) for willows who found that the Bowen ratio rapidly decreased with increasing of from zero to 1.5 and thereafter the dependence of leaf area was much smaller and driven mainly by other factors. The Bowen ratio of willow stand described by IRITZ and LINDROTH (1996) started from value around five at the beginning of season and after the development the typical hourly values ranged from -0.5 to 0.5. These negative values of the Bowen ratio indicate the advection of the sensible heat and from their data it is also evident that the median was very close to zero. Here presented results of the Bowen ratio diurnal medians are significantly higher with no negative values which generally mean higher production of sensible heat without its advection and lower available energy portion used for evapotranspiration. The higher values at the beginning of season with low or zero are in well agreement with the Bowen ratio measured by IRITZ and LINDROTH (1996) over the willow stand. The same is valid for the unfoliated period after coppicing during 2010 when the Bowen ratio rapidly increased with the dryness of the superficial soil layer. During the period when the soil evaporation is the driving process, the soil layer responsible for the energy partitioning is approximately 0.1 to 0.15 deep depending on the physical soil specific properties and its water capacity is given by the saturation level and approximately the half of the wilting point (ALLLEN et al., 2005). The Bowen ratio values from 2008 and 2009 (except the unfoliated period) fall well into the range for deciduous forest given by WILSON et al. (2002) where the lowest values 0.11 were recorded for poplar forest in Iceland and the highest 0.73 for beech forest in Belgium. Some of the higher values linked with the lower compared to that of the turf grass or the reference one fell rather among the range of values described for coniferous species varied mostly between about 0.5 to and little more than 1.0. WILSON et al. (2002) further reported the Bowen ratio of the grasslands varying between 0.34 to 1.91 depending especially on the soil moisture level and the management practices. Except the driest grassland reported in Little Washita (Oklahoma, USA) the daytime Bowen ratio of 3 other investigated grasslands ranged between 0.34 to 0.76. The Bowen ratio of the turf grass described in this study generally fluctuated near 0.5 which ranks it among the lower level of the values reported by WILSON et al. (2002) meaning relatively high evapotranspiration. Our results are in good agreement with TEULING et al. (2010) who reported average Bowen ratio 0.54 and 0.89 for grassland and forest respectively. Moreover, TEULING et al. (2010) described that forest absorbed more short wave energy and as a consequence produce more sensible heat compared to grassland which agrees with the here presented observations. This situation is even more pronounced during the heat waves at the initial stage of ample supply of soil moisture when the Bowen ratio of grassland decreased to 0.41 whereas that of the forest increased to 1.6. It in other words means that forests are generally higher producer of heat compared to grassland and this feature is even more enhanced during heat wave conditions, at least at the initial stage with no short of water, when the extra energy, due to low cloud cover typical for anticyclones (BLACK et al., 2004, 198

REBETZ et al., 2006), is used more for evaporating in case of grassland and more for heating in case of forests. As one of the reasons, beside the different stomatal response of the forests, is also the rougher surface which provides more efficient turbulent heat exchange with the boundary layer, such that convective cooling relaxes the need for strong evaporative cooling during heat wave conditions. In the long term, however, elevated evaporative cooling expedites soil moisture depletion, and grassland rather than forest becomes the main heat source (TEULING et al., 2010). 6.3.2 Surface resistance The second part of the growing season 2008 was characteristic with higher surface resistance of the turf grass compared to the poplar plantation linked with the limited soil water availability in the shallower root zone. However, the surface resistance over the turf grass with the mean 96 s m-1 for period July to August and the monthly mean 165 s m-1 in September are still in the lower level of the values reported by WILSON et al. (2002) for four grasslands ranging between 97 to 563 s m-1 with the mean 244 s m-1. The surface resistance of the turf grass fell well within the range of the surface resistance (100–300 s m-1) measured by WEVER et al. (2002) over the grassland during the reduced soil water availability. The mean surface resistance of the poplar plantation within the same period was 85 s m-1 with the lowest monthly mean 72 s m-1 during July 2008. This is in very good agreement with the values reported by WILSON et al. (2002) for 16 deciduous forests ranging between 63 and 101 with the mean 72 s m-1. Similarly like in the case of the Bowen ratio, also the same surface resistance of both covers does not mean the same . The main reason is that the canopies have different aerodynamic resistance given by the differences in their height and thus their roughness length. Theoretically, if the net radiation over short and tall canopy would be the same, the short canopy must have lower surface resistance than the taller one with inherently lower aerodynamic resistance in order to sustain the same . Entering the next year 2009, the overall mean of the surface resistance within the growing season was 61 and 80 s m-1 for the turf grass and the poplar plantation respectively. By omitting the high April values of the poplars when the was -1 not developed enough, the mean surface resistance would decrease to 70 s m . In case of the turf grass, the values from 2009 and 2010 (67 s m-1) are in very good agreement with those reported for the reference grass (~70 s m-1) with adequate water (and nutrient) supply (SHUTTLEWORTH, 1993, ALLEN et al. 1998) which indicates the optimal soil water regime during these years. The comparable values for the poplar plantation are again in a good agreement with the range valid for deciduous forests described by WILSON et al. (2002). Namely, the poplar forest in Iceland mentioned by WILSON et al. (2002) had only slightly lower surface resistance 68 s m-1. The values from the mature poplar plantation also fit well with the results of PERSSON and JANSSON (1989) who estimated daily means of the surface resistance on the SOIL model basis within the range 90–160 s m-1 for unirrigated and 50–120 s m-1 for the irrigated willow stand during four months (June to September) of the growing season. In contrast, the values of surface resistance of our poplar plantation are generally higher than the measured canopy resistance (without soil evaporation) of willow SRC reported by LINDROTH (1993) within the range 40–100 s m-1 or the modelled one 199

(SOIL) described by PERSSON and LINDROTH (1994) ranging between 30–50 s m-1, in both cases after the stand has reached of about 2–3. In 2010, bare soil was firstly the dominant feature at the coppiced poplar plantation and thus the surface resistance was strongly dependent on the wetting regime of the superficial soil layer, the same like the Bowen ratio. Therefore, the surface resistance was often rapidly fluctuating from the low values near zero (wet soil) to as high as almost 400 s m-1 (dry soil). Such rapid changes of the surface resistance and its magnitude over bare soil are in a good agreement with model results reported by DAAMEN and SIMMONDS (1996). 6.3.3 Decoupling coefficient The decoupling coefficient (factor) according to MCNAUGHTON and JARVIS (1983) revealed different coupling with atmosphere for two contrasting canopies investigated in this study. Namely the turf grass showed high decoupling with the seasonal means within the range 0.78–0.85 which is in accordance with the value 0.8 originally given by JARVIS and MCNAUGHTON (1986) for the grasslands. The poplar plantation had significantly lower decoupling factor with means reaching to 0.41 and 0.43 in 2008 and 2009 respectively. It places the poplar stand between the group of cotton as summarized by JARVIS and MCNAUGHTON (1986). Similar results were obtained by CIENCIALA (1994) for willow stand in south Sweden with decoupling coefficient equal to 0.4. However, LINDROTH (1993) found that decoupling coefficient is low (0–0.4) with < 1 and after that increases with into the range between 0.6 and 0.8 which means that stand is thus only well-coupled to the atmosphere for very small values of leaf area indices and it is practically de-coupled for leaf area indices above about two. According to JARVIS and MCNAUGHTON (1986), low decoupling coefficients are usually found for forests and high ones for lower crops like wheat, barley, etc. The willow stand described by LINDROTH (1993) acts like a forest from an evaporation control point of view, when leaf area is small, and like a traditional agricultural crop when leaf area is large. The similar conclusion for willow SRC was found in the later modelling study carried out by PERSSON and LINDROTH (1994) who additionally found that for a given leaf area index, the system is better coupled at times when leaf area is declining as compared with when it is increasing. This difference was judged as an effect of height increment of the stand and the associated increase in aerodynamic roughness. HINCKLEY et al. (1994) reported decoupling coefficient with an average 0.66 for hybrid poplars (P. trichocarpa x P. deltoides) which means also poorly coupling to the atmosphere. The reason of the high decoupling coefficients of willows in Sweden might be the high density of the stand with approximately 20,000 trees ha-1 whereas the sparser stand (~ 2,000 trees ha-1) reported by HINCKLEY et al. (1994) was relatively smooth and uniform. In contrast, ZHANK et al. (1999) found decoupling coefficient of poplar plantation (P. trichocarpa x P. tacamahaca – clone TT32) within the range of 0.51 and 0.22 with a mean of 0.35 suggesting that transpiration rates of the trees were to a great extent controlled by changes in vapour pressure deficit and stomatal conductance, and were not ‗‗imposed‘‘ by radiation which was explained by relatively high roughness of the stand. The low decoupling coefficient 0.31 was also found by BLANKEN et al. (1997) for boreal 200

aspen (P. tremuloides) forest in Canada. Likewise, KHAMZINA et al. (2009) reported values of decoupling coefficient never exceeded 0.3 for P. euphratica grown in plantation on degraded crop land with increased salinity in Uzbekistan. For vegetation freely supplied with water, the value of the decoupling coefficient of canopy depends mainly on wind speed, aerodynamic roughness, leaf sizes and the canopy density. From that reason, the most coupled are tall coniferous species and the least coupled are the short and smooth covers like grassland or the dense canopy of alfalfa (MCNAUGHTON and JARVIS, 1983). A fractional change in canopy conductance can therefore be expected to cause an almost proportional change in transpiration in case of the well coupled species. In contrast, transpiration from grassland and other smooth low vegetation is largely dominated by equilibrium rather than imposed (driven by vapour pressure deficit) evaporation, with the consequence that a similar fractional change in stomatal conductance has very little impact on transpiration (JARVIS and MCNAUGHTON, 1986). This means that the physiological control of transpiration dominates at the poplar plantation investigated in this study whereas the climatological control dominates in case of the turf grass. However, during the year 2009, it was also obvious that the coupling to the atmosphere was partially dependent on the . At the beginning of the season, the decoupling coefficient was very low and it increased with the developing to the average level near 0.5 (excepting the very high values linked with wet canopy). At the turn of July and August, however, the decoupling coefficient strongly decreased which indicate higher stomatal control which generally increases when plants are grown in stressed environments (MORRIS et al., 1998, LARCHER, 2003). The possible reason might be linked with the reduction of the in the upper part of canopy due to the strong winds. One of the theoretical reasons might be also the mechanical disturbance of the sapwood induced by the wind gust which were described e.g. by UEDA and SHIBATA (2004) for cypress in Japan. Another possible reason might be the nitrogen deficiency linked with exceptionally humid and rainy weather during June and July. In 2010, the mean seasonal decoupling coefficient for the poplar plantation was 0.51. For the second half of the growing season it decreased only slightly to 0.47 which ranks the resprouting plantation according to JARVIS and MCNAUGHTON (1986) among the group of cotton and prairie. 6.3.4 Crop coefficient The concept of crop coefficient is an alternative way how to use the Penman-Monteith equation with its possibility to theoretically derive all the factors driving the evapotranspiration. Originally PENMAN (1953) and later MONTEITH (1965) came with the idea to include into the Penman formula the influence of stomatal control and thus put into the equation completely physical meaning without any portion of empiricism. However, as it was mentioned previously, measuring stomatal conductance continuously is so far impossible and thus it has to be modelled. One way how to escape from this difficult task is to fixed the aerodynamic and the surface resistance in the Penman-Monteith equation and then to empirically derive the ratio between the resulting (calculated from the main five meteorological variables) and the measured of the crop of interest. The PenmanMonteith equation with the fixed parameters is so-called reference evapotranspiration, , 201

where the reference crop has been recently short grass with ample of soil water (ALLEN et al., 1998). However, this approach omits the natural variability of the surface resistance which is caused by the stomatal response and also by the differences in the wetness of the surface. The first issue can be partially solved with the assumption that the reference crop has no short of water. However, such kind of fixing of the surface resistance can work only at the diurnal level due to the stomatal response to solar radiation and the vapour pressure deficit (JARVIS, 1976, LOHAMMAR et al., 1980, LINDROTH et al., 1985, ITIER, 1996). Moreover, this approach does not respect that beside the stomatal control, the surface resistance is also driven by the wetness of the canopy and thus the of wet and dry surfaces should be solved separately (SHUTTLEWORTH, 1993, IRITZ et al., 2001). This necessity is especially more pronounced for the tall aerodynamically rough canopies with high with the potential to intercept and subsequently evaporate substantial amount of water. of such canopies after rain can exceed even four times the Penman open water evaporation which can be also understood as a kind of reference evapo(transpi)ration with the surface resistance equal to zero (LINDROTH, 1993). This fact is the main explanation why there is so high scatter between the calculated and the measured for our poplar plantation and it confirms suitability to solve the rain days and the effect of precipitation separately. The same impossibility to find any stable crop coefficient, , without the omitting the days with rain was reported by PEACOCK and HESS (2004). To use the simple approach including the effect of wet surface (albeit empirically) can improve the estimation of models based on the concept of the crop coefficient and the reference evapotranspiration (at least for the taller vegetation) in case that the is modelled realistically and its maximum interception capacity is well defined. However, the evaporation of the intercepted water can enhance the of short canopies which is probably one of the reasons why totally 16 % higher values of the above the short turf grass compared to the calculated were measured. The other reason of this disagreement reported also by other authors has been already mentioned in context with fetch and footprint over the turf grass in section 6.2.3. By comparing the values of the poplar plantation from our study with those from literature I have found that there is quite lack of the available information despite the widely use for use of approach for agriculture crops. Moreover, these studies are carried out across various geographic conditions, use different technique of determining of SRC and even different concepts of , which makes the comparison very difficult with low possibility to make general conclusions applicable for condition of Central Europe. For example, one of the studies dealing with the of both SRC species, poplars and willows, is the already mentioned lysimteric-based research from SRC grown as a vegetation filters in Mediterranean conditions of central Italy described by GUIDI et al. (2008) and PISTOCCHI et al. (2009). In the first of the named study, they analysed the effect of fertilization (fertilized and unfertilised) on of willow and poplars with ensuring no short of water during the first two years of the first rotation. The resulting mean seasonal of fertilized willows were 1.37 and 2.77 with maximal ten-days values 2.84 and 5.3 for the first and the second years of the first rotation period respectively. The mean values of fertilized poplars within the same period were 0.96 and 2.23 with ten-day maximums 202

1.9 and 4.28 for two consecutive years respectively. In case of the unfertilized willows the mean were 0.75 and 1.09 with ten-day maximums 1.25 and 1.97 and for unfertilized poplars the were 0.6 and 0.89 with ten-day maximums 1.06 and 1.71. In the later study by PISTOCCHI et al. (2009) describing the first year of the following rotation they investigated the influence of the different fertilization (low and high level again for both willows and poplars. PISTOCCHI et al. (2009) found slightly higher values in the first year of second rotation period compared to the first year in the first rotation with a peak for both species within 1.23–2.95 and 2.54–4.03 in low fertilized and high fertilized treatment respectively. Beside that they concluded that the seems to be related more to the availability of nutrients than the differences between the two species. For comparison, our mean seasonal (April to September) values (after the filtering of the days with and just the first days after the precipitation) were 1.05 with ten-day maximum 1.45 and 0.85 with ten-day maximum 1.19 in the eight year of the first rotation and the first year of the second rotation period respectively. By including also the days with rain and thus the effect of interception, the mean seasonal will be 1.28 with the ten day maximum 1.86 and 0.98 with the ten-day maximum 1.27 in the same order of years respectively. Thus it seems that the results of this thesis are in a good agreement with the unfertilized treatment described by GUIDI et al. (2008). MEIRESONNE et al. (1999) reported transpiration based on sap flow up-scaling to represent 70 % of the reference crop evapotranspiration. If we consider that transpiration can roughly represents 60–70 % of the whole (PERSSON and LINDROTH, 1994, HALL et al., 1996) then the seasonal crop coefficient would be very close to one, the same as found in the here presented study. The of poplars very close to one (0.9–1.1) was also confirmed by VERSTRAETEN et al. (2005) based on the results of water balance model WAVE. Further, in other study of GAZAL et al. (2006) focused on sap flow and transpiration in a semiarid riparian cottonwood forest (P. fremontii) in south-eastern Arizona (USA) the reached to maximal values around 1.2. HOU et al. (2010) reported mean seasonal around 0.4 with fifteen-day maximum 0.62. Moving back to the southern England, HALL et al. (1998) found the transpiration ratio (transpiration divided by ) of poplar clone Beaupré and willow clone Germany (Salix burjatica) within the range of maximums between 2–2.5 during the period with adequate water supply. If we again assume that the transpiration is roughly 70 % of , it enables us to estimate the within the range of 2.85 to 3.57. By taking the same assumption, we can recalculate the transpiration ratios close to 1 and 0.5 respectively for two poplar clones Beaupré and Dorschkamp reported later by ALLEN et al. (1999). The resulting then would be 1.43 and 0.71. Finally, PERSSON and LINDROTH (1994) and later PERSSON (1995) reported the crop coefficient for willow stand in Sweden derived by using the physically-based model SOIL. Since they used the Penman open water (PENMAN, 1948) formula instead of the , the presented are not comparable with those according to ALLEN et al. (1998) used in this study. However, the empirical relation given by PENMAN (1948) or PENMAN (1956) says that the potential of grass, , is equal to 0.8 of the Penman open water evaporation during the period May to August, 0.75 during March, April, September and October and 0.6 for the rest of the year. By taking this assumption, the crop coefficients described by PERSSON and LINDROTH (1994) will result in 0.88–1.25 at the initial crop 203

development stage to 1.5–2.0 at midseason and about 2.7 late in the season. These values are considerably higher than those found in this study, however, as it was already pointed out and also reported by the authors, such high linked with enormous water consumption of willow in Sweden or both poplars and willows from southern England would be not possible without the advection of the sensible heat from adjacent fields. The same is naturally valid for the fertilized SRC in the lysimetric-based study described in GUIDI et al. (2008). As concluded previously within the context of rates, the higher reported by other authors are probably mainly due to different soil fertility, level of intensification in management and possibly the clonal specific variation. 6.3.5 Sap flow Although, the trees with very similar dimensions from the most dominant classes were investigated, their sap flow showed quite substantial mutual variation. The design of the tissue heat balance probe is robust against radial variation of sap flow density and does not require any calibration (ČERMÁK et al., 2004), however, one of the possible reasons might be the applied up-scaling procedure from measured tree trunk segment to tree level by multiplying the circumference assumed a uniform azimuthal flow pattern (PETZOLD et al., 2010). Therefore some authors use more than just only one measured segment (CIENCIALA et al., 1999, ČERMÁK et al., 2004). The azimuthal variations of sap flow are usually explained by the position of the large branches relative to the trunk (LU et al., 2000), asymmetrical shapes of the crown and root system and the variable conditions around the trees causing competition for light, water and mineral elements (GRANIER et al., 1994). Additionally, varying soil structure and thus its retention capacity and soil water tensions affecting root distribution in the soil leading to varying azimuthal sap flow (NADEZHDINA et al., 2007). Moreover, COHEN et al. (2008) studied the radial pattern of sap flow in trunks around circumference of six species and found that amplitude of the azimuthal variation changed during the day. The same explanation as for the azimuthal variation, especially the competition for light, water and nutrients or the effect of other limiting and stress factors might be applied at the stand level to the variability among the individual tress (ČERMÁK et al., 1995, KÖSTNER et al., 1998). COHEN et al. (2008) concluded that the variation of the azimuthal sap flow pattern within a single tree may have the same magnitude to that between the trees. Within the presented study, the azimuthal variation might result from the irregular spacing of the trees in the plantation comprised the double rows and inter rows where different competition can be encountered. The measured segments were positioned toward the inter rows and thus the sap flow of the measured segment recalculated to the whole stem could be overestimated. However, the Populus belongs among the diffuse porous species (DICKMANN, 2001) which means that the sap flow in stem should be more homogeneous and thus the variability between trees plays probably more important role than the azimuthal variation. ČERMÁK et al. (2004) showed that the upscaling error in homogeneous stands sharply decrease with increasing number of sample trees and reaches approximately 15 % and 10 % for 4 and 12 sample trees respectively. In contrast, PETZOLD et al. (2010) reported that six sample trees in their particular poplar stand were the minimum in order to match a measurement error less than 204

20 %. Similarly, CIENCIALA et al. (1999) concluded that with respect to the heterogenous conditions at the site, the set of 12 trees appeared to be the minimum reasonable sample size to extrapolate tree fluxes into stand transpiration and monitor simultaneously two tree species. From the above mentioned reasons it is very difficult and questionable to correctly upscale the sap flow from four sample trees to the whole stand. The situation is even more critical by considering that the sample trees were limited only by the two most dominant classes, which was predetermined by the spectrum of stem thickness measurable with tissue heat balance method. Therefore, the upscaled transpiration by using the leaf mass should be perceived rather than very rough estimation but probably one of the best way of the available scaling methods. For the future research, at least double number of sample trees have to be investigate and a way (e.g. different sap flow technique) how to measure the sap flow across the whole social distribution (dominant, co-dominant and suppressed) have to be found. The relation between the and the active leaf mass might be also expressed by the linear function (just to the or rather to basal area) with omitting the values below of approximately 65 mm which have very low transpiration rates in order to anchor the regression line to its origin similarly like used e.g. in ČERMÁK et al. (1980), ČERMÁK and KUČERA, 1990, or HINCKLEY et al. (1994). This manner would provide similar coefficient of determination, however, using the power function to derive stand transpiration fit better with the total evapotranspiration from BREB, whereas the linear function provide almost two time higher values. As compared with the power function, the linear regression line supported more the trees around middle classes (75 and 85 mm) which had the highest number of trees and therefore significantly overestimated the canopy transpiration. Also ČERMÁK et al. (2004) reported overestimation of true values in small classes and underestimation in large ones when using the linear function between and the basal area. In contrast, the power function lowered the transpiration of the smaller trees and boosted the transpiration of the most dominant ones. In fact, even the relation between the sap flow and the basal area, which is naturally the power function of the , is rarely linear, but usually sigmoidal (e.g. Gompertz), or similar function (ČERMÁK et al., 2004). It can be assumed, that the using of the exponentially shaped function which emphasize the sap flow of the dominant trees at the expense of the suppressed ones has its physical meaning for the young stand, however, with the increasing age of the trees the dominance of the trees cannot naturally increase incessantly exponentially but it should approach to some asymptote as expressed in the Gompertz function. One way or another, it is obvious that the scaling up the sap flow from only four dominant trees to the whole stand is a little bit alchemy and it can result in large errors. For example, the finally used power function result in scaling coefficient 0.12 ( =0.93, = 30), whereas the linear (with anchoring the line to its origin at of 65 mm) 0.19 ( =0.83, = 21), the logistic 0.21 ( =0.90, = 30) and the Gompertz function 0.20 ( = 0.90, = 30) which suggest almost hundred percent difference in transpiration between the first and the rest three functions. Although, the last three scaling curves provided similar results, the first one in turn agreed well with the measured by BREB yielding total seasonal transpiration of 346 mm which constitutes 70 % of the . The significant departure of transpiration from the during the last decade of 205

June and first two decades of July 2009 might be probably the effect of the rainy weather linked with higher amount of evaporation from soil and interception and also high relative humidity (low vapour pressure deficit). Other theoretical reason (which, however, cannot be proved from material within this study) of the underestimation of transpiration derived from the sap flow of the most dominant individuals is that during the period with lower (spring) the suppressed and co-dominant trees could act as an understory which received more light and thus their relative contribution to the stand transpiration were higher. The estimated seasonal transpiration as the 70 % of the total seems to be a reasonable result. For example, PERSSON and LINDROTH (1994) found in their modelling study for willow stand the mean seasonal basis transpiration 66 %, soil evaporation 23 % and interception evaporation 11 % of the whole . The resulting interception was usually 5–23 % of precipitation. HALL et al. (1996) found on sap flow based measurement in combination with modelling that transpiration was 86 % and interception 14 % of the whole (they did not consider soil evaporation) during the growing season, however, authors expected that in more typical summers with more frequent and plentiful rain the ratio of the transpiration to interception loss would be not so large. The applying of the scaling curve on the stem inventory brings also one more interesting conclusion. From resulting active leaf mass 0.65 kg m-2 as the mean of the whole stand, it can be estimated the mean active leaf area by multiplying with 3.28 ( in 2 -1 m kg ) yielding to be 2.1 which in other words mean that between 3 to 4 (not exactly 2.1 because some leaves will be shaded by the upper leaf area) is effective with respect to radiation in the same way like around 7 which is in accordance with Iritz and Lindroth (1996) who found that the threshold of after which the surface resistance and Bowen ratio do not decrease significantly lies between 2 and 3. In later modelling study by IRITZ et al. (2001) they found the best fit of the model based on the Shuttleworth-Wallace approach (SHUTTLEWORTH and WALLACE, 1985) with measured if only the top of the canopy with 3 contributed significantly to canopy transpiration. It might be illustratively explained on the application of the Beer´s law, where assuming the extinction coefficient equal to 0.6, incident solar radiation 1000 W m-2 and 3, only 165 W m-2 will theoretically penetrate through such canopy. Since the works based on sap flow measurement upscaled to the stand transpiration have been already discussed in subchapters focused on evapotranspiration, now only those dealing with sap flow related to the individual trees will be emphasized. For example, HINCKLEY et al. (1994) reported the daily sums of sap flow (measured by tissue heat balance method) for the four-year-old poplar plantation (P. trichocarpa x P. deltoides) in Washington State (USA) during midsummer period ranging between 20–26 for the smallest tree (120 mm class), and 29–38 kg day-1 (140 mm class) and 39–51 kg -1 day (160 mm class) for the largest trees. The diurnal peak of sap flow in HINCKLEY et al. (1994) reached almost up to 4.5 kg hour-1 for the most dominant tree (151 mm in ) with mean 3.6 kg hour-1 for 6 investigated trees during the one warm, dry and bright August day. These values are in a relatively good agreement with the poplars at our plantation where the maximal peak of sap flow 5.34 kg hour-1 was recorded at tree with 106 mm in while the usually reached peaks ranged between 3.5 to 4.5 kg hour-1. For 206

comparison, the maximal water use of the most consumptive tree with 106 mm was -1 44 kg day which agrees well with the values measured by HINCKLEY et al. (1994). In another ecophysiological study by ZHANK et al. (1999) carried out at poplar plantation (P. trichocarpa x P. tacamahaca) in the southeastern England, the sap flow (surface heat balance – analogy to the stem heat balance method) of tree branches was measured and extrapolated in to the sap flow of the whole trees. The of measured trees was within the range 42–46 mm, i.e. less than half compared to the presented study, which resulted also in lower sap flow rate reaching up to 26 kg day-1. However, the daily transpiration rates with highest values between 4 and 5 (absolute maximum 5. 4) mm day-1 (never exceeding the Penman potential ) were very comparable to those estimated in this thesis. The peaks of the diurnal sap flow rates reported by ZHANK et al. (1997) and ZHANK et al. (1999) were close to 3 kg hour-1, i.e. approximately one kg less than in our or HINCKLEY et al. (1994) study. Incomparable higher sap flow rates of hybrid poplar clone NM6 (P. nigra x P. maximowiczii) were reported by ZALESNY et al. (2006) for the trickleirrigated plantation in northern Wisconsin (USA) grown for phytoremediation on soil contaminated with NH3 . The stem diameters of the sampled trees in their study were measured in 0.15 m above the ground and ranged from 46 to 71 mm at the beginning of June and 78 to 107 mm in mid September 2002. During the following year the diameters measured during the same dates were 100 to 131 mm and 119 to 157 mm. Sap flow was measured only during the 18 days period from last week in June to the second week in August for both of the investigated seasons representing fourth and fifth years after planting. During this period in 2002 the mean hourly sap flow per tree ranged from 0.49 to 2.5 kg hour-1 and the mean daily sum was 34 kg day-1 which is still comparable, however, higher than the values recorded in the here presented study or those reported by HINCKLEY et al. (1994). Nevertheless, in the following season the mean hourly sap flow per tree within the same period reached up to 4.56 and 6.96 kg hour-1 with mean daily sum 136 kg day-1 which exceedingly higher than the maxima measured in this or the HINCKLEY et al. (1994) studies. The mean upscaled transpiration was then 2.8 and 11.3 mm day-1 over the 18 days summer period in 200 and 2003 respectively. The authors explained the increase of sap flow rates between the seasons by the three-fold increase of the cross sectional area of sapwood. However, these numbers ranks the poplars described by ZALESNY et al. (2006) among the most water consumptive poplar plantations at all. Just for imagination, during the September 2010 the sap flow (tissue heat balance applicable for diameters 6–20 mm) of the small shoots was measured (data not shown). The shoot diameters at the measuring point (~0.4 m above the ground) ranged from 8 to 19 mm and the mean canopy height ~2 m. The maximal hourly sap flow ranged from 45 to 117 g hour-1 in the same order of shoot diameters and their respective daily sums and 145 to 595 g day-1. The sap flow from the four shoots was upscaled to the whole stand (based on the cross sectional area and the sap flow relation) comprised 168,920 shoots (representing quite homogeneous material) yielding the total stand transpiration with maxima around 2 mm fitting well with BREB measurement. TRICKER et al. (2009) measured sap flow of the one-year-old resprouting shoots (P. deltoides x P. nigra, clone I214) with approximately two times higher dimensions and recorded more than two times higher peaks in hourly values of the shoot sap flow and the daily stand transpiration. 207

TRICKER et al. (2009) also demonstrated that under elevated CO2 concentration (550 ±70 ppm) in POPFACE, the sap flow showed 15 % higher values compared to the mentioned control. 6.3.6 Biomass growth and water-use efficiency The of the poplar plantation measured in this study during the 2009 with maximum 7.5 fell well within the range 4–12 for well established high density stands based on poplar or willow species with closed canopy reported by other authors (ZAVITKOVSKI, 1983, CEULEMANS et al., 1996a, LINDROTH et al., 1994, HEILMAN et al., 1996, HALL et al., 1998, GIELEN et al., 2001, HENDERSON AND JOSE, 2005, LIBERLOO et al., 2006, PETZOLD et al., 2010). For the open canopies which are usual in the year after planting or coppicing, just like the year 2010 in this study, the maximal usually varied within the range of 0.5–6 (LANDSBERG and WRIGHT, 1989, MILNE et al., 1992, CEULEMANS et al., 1996a, BROECKX et al., 2010) while the time of the canopy closure and the development is predetermined especially by the planting density, clone variety, management, weed competition etc. Similar dynamic of development with the peak in the last part of summer characteristic for poplars and willows were described e.g. by CHEN et al. (1994), CEULEMANS et al. (1996a), PELLIS et al. (2004), or GUIDI et al. (2008). The canopy closure in the first rotation occurred very late, even in the fourth year after planting (2005) because of the very strong competition with weeds which had to be intensively controlled. This had negative impact on total biomass yields and it is also the reason why the longer rotation period (eight years) was chosen. By considering the success rate of the grow of young poplars during the year of establishment in other studies (e.g. TRNKA et al., 2008, BROECKX et al., 2010), it is reasonable to assume that if the weeds would be well managed within the first year 2002, at the end of the season the canopy could be at the level of the year 2004 which means that the competition with weeds delayed the development of the stand and the final attained yield approximately by two years. Thus the real yield 71 tons per hectare could be theoretically reached within 6 years and the mean annual yield would be almost 12 instead of 9 tons ha-1 year-1. This would be comparable with the yield 13.9 tons ha-1 year-1 upscaled from small-scale plot clonal experiment at the same locality reported by TRNKA et al. (2008) and lie within the higher range of yields in the Czech Republic estimated by LEWANDOWSKI et al. (2006) or HAVLÍČKOVÁ et al. (2010). However, even by using the real yield with considering the whole 8 years long rotation, the mean annual yield would be comparable with that reported by other authors. For example BENETKA et al. (2007) reported the mean annual yield 9.4 tons ha-1 year-1 for the poplar clone NE-42 (P. maximowiczii x P. trichocarpa) grown in very similar conditions. According to the model based on production soil-ecological units (BPEJ – specific classification within the Czech Republic) described e.g. by WEGER et al. (2009) or HAVLÍČKOVÁ et al. (2010), our site lies at the boundary of the average and above-average production zones with the respective predicted yields 6–8.5 and 8–12 tons ha-1 year-1. As mentioned by HANSEN (1991), CEULEMANS et al. (1996b) and VAN DE WALLE et al. (2007), the small plots yields are in order of 10–15 ha-1 year-1, however, most reported

208

yields are much less, apart from some exceptions grown under intensive irrigation and fertilization or those situated in favourable conditions. By relating the aboveground woody biomass (AB) production to the water use, the gross will be obtained. Gross based on synchronous measuring of AB increment (alometrically derived from the stem circumference increments) and the canopy (by BREB method) showed high seasonal variability in both of the evaluated years 2008 and 2009. The seasonal patterns were typical with the highest rates of gross at beginning of the season and with the decreasing at the end of summer. Very similar seasonal behaviour was observed also at the based the sap flow (and so transpiration) and the stem increment measurements. LINDROTH et al. (1994) described the similar seasonal trends of in intensively managed (irrigated and fertilized) willow SRC (Salix viminalis) in Sweden, where the last most marked fall of was linked with reducing at the end of summer. The results of the presented thesis confirmed this coherency with dynamic since the pronounced decline of gross was observed around the mid August, when the period of culmination in the investigated poplars culture was recorded. Further, LINDROTH et al. (1994) explained the maximal peaks in , which was defined as the ratio of measured AB increment (alometrically derived from shoot diameter) and modelled transpiration (physically-based model KAUSHA tuned against BREB measurements), as an effect of rainy weather and thus with considerably amounts of evaporation of intercepted water and thus lower transpiration rates. However, in our research which took into account with the gross based on total evapotranspiration, the maxima were also found during the rainy events. This is also in contradiction with results of GRELLE et al. (1997) maintaining that precipitation days and the days after, lower the due to enhanced surface evaporation of water that has never been part of the plant metabolism. The higher gross as a consequence of precipitation could be explained by a few reasons. Firstly, by using the long-term water use efficiency based on AB increment and not on CO2 uptake and transpiration ratio, the relative carbon allocation to roots or AB can play important role (HÖLL, 1985, LINDROTH et al., 1994, DICKMANN et al., 2001, HOCH et al., 2003). During the period with reduced soil water availability, the assimilated carbohydrates are directed away from shoots and towards root growth. After alleviated of such conditions by replenish the soil water status with rains or irrigation, the temporary carbon allocation toward the roots is compensated by a later increase in shoot growth (HSIAO and ACEVADO, 1974, KRAMER, 1983, LINDROTH et al., 1994, BARBAROUX et al., 2003). The similar effect caused by mobile carbon pools influences also the seasonal variation with high in the spring characteristic with so called spring flush (strong upward translocation of non-structural carbohydrates produced in assimilation during the end of previous season), and conversely with the drop to zero at the end of season linked with downward accumulation (TESKEY and HINCKLEY, 1981, DEANS and FORD, 1986, DICKMANN et al., 2001, BARBAROUX et al., 2003, LACOINTE et al., 2004). Here is noteworthy that the high accumulation of root reserves has great significance to coppicing. Secondly, growth is the biological phenomenon of increase in size with time. Growth involves the formation, differentiation and expansion of new cells, tissues or organs. The 209

sudden increase in tree diameter often observed after rain is not necessary due to growth but reflects the saturation of shrunk xylem and other stem tissues with water after previous drier period (HERZOG et al., 1995, OFFENTHALER et al., 2001). Moreover, e.g. HSIAO and ACEVADO (1974), HINCKLEY and LASSOIE (1981), BARBAROUX and BREDA (2002), STEPPE et al. (2006), ZWEIFEL et al. (2006) explain that during drier periods when the stem water deficit occurs, new cells which are still created do not immediately expand, but a release of the low pressure conditions in the cambium suddenly enlarges the already existing cells to their mature sizes. This means that, within a certain period of time and for a range of water deficit, growth is not inhibited but just delayed. Within this context, the term is very disputable and using gross (either based on total evapotranspiration or just on pure transpiration) seems to be more reasonable and relevant. On the other hand, research of based on measuring the fluxes of CO2 and by eddy covariance method across European forest ecosystems as described in KUGLITSCH et al. (2008) provided that increased with a rising monthly precipitation sum and rising average monthly temperatures. Finally, the higher productivity during these periods can be result of an increase of diffuse radiation that might stimulate assimilation (ALTON et al., 2007, KNOHL and BALDOCCHI, 2008) and particularly low vapour pressure deficit under warm and humid conditions resulting in open stomata and thus higher CO2 uptake to water lost ratio (KUGLITSCH et al., 2008). Comparing the both of executed years, there was notably higher gross during 2009. Similar situation was described also by LINDROTH et al. (1994) where the authors explained the contrast in the consecutive seasons by different age of culture and with this linked different root-to-shoot ratio. The decreasing root-to-shoot ratio with ontogenically aging is well known phenomenon in SRC and also other tree species culture (OVINGTON, 1957, REYNOLDS and D‘ANTONIO, 1996, COLEMAN et al., 2004) and could provide an interpretation of higher AB increment during 2009. However, the other reasons might be also the facts that the 2009 was abnormally wet (yearly precipitation sum was 778 mm compared to 520 mm in 2008) and the growing season in 2009 started approximately two weeks earlier due to the abnormally warm and dry April. The heterogeneity in carbon allocation during the particular ontogenetic phases and also during the particular parts of the season causes difficulties to predict the yields with some simplified method based on evapotranspiration and biomass relation, but on the other hand also such information could provide some general and gross estimation of the SRC production. By including the influence of precipitation and the decreasing seasonal trend of expressed as a phenological factor, more precise method has been proposed in this work. Nevertheless, this is a simplified empirical relationship which could be valid presumably only for SRC grown on arable land with similar fertility, water holding capacity and other soil properties in the comparable condition of Czech-Moravian Highlands. Modelling the seasonal and yearly dynamic of root-to-shoot allocation could be key point to improve such approach. Generally, the average of most coniferous and broadleaved trees of temperate zone range between 3–5 g of dry biomass per kg of transpirated water (LARCHER, 2003). The research on of three different poplar clones (Beaupré, Trichobel and Ghoy) growing in weighting pots placed in greenhouse indicated relatively constant values 210

varying from 3.5–4.4 g of dry biomass per kg of water despite strongly fluctuated soil moisture during the season leading to marked variation in root-to-shoot ratios (SOUCH and STEPHENS, 1998). However, GUIDI et al. (2008) demonstrated great influence of fertilization on where the effect of fertilizers increased from 0.43 to 2.14 and from 0.68 to 2.4 g of dry biomass per kg of evapotranspirated water, respectively for willows and poplars used as a vegetation filter. LINDROTH et al. (1994) investigated of fertilized and irrigated high-density willow stand and they found relatively high values of mean seasonal reaching to 4.1 and 5.5 g of dry biomass per kg of transpirated water for two consecutive years. Later study by LINDROTH and CIENCIALA (1996) found even higher mean seasonal values of reaching to 6.3 of dry biomass per kg of transpirated water which they associated with high foliar nitrogen concentrations. Although, the gross of the resprouting shoots after coppicing during 2010 was not evaluated, it could be roughly estimate from the total annual biomass yield and the measured by BREB. By taking into account that the resprouting was initiated in June, the gross can be calculated as the ratio of the total yield 3.9 tons ha-1 and the from June to September which amounted 186 mm. After the necessary unit conversion the resulting gross is 2.1 g kg-1. Considering that due to the unclosed canopy the soil evaporation played important role, the real will be much higher. If we stay very conservative and estimate the transpiration to be 70 % of , the -1 will result in 3 g kg . Considering that resulted values of gross with means 3.13 and 3.54 g kg-1 in 2008 and 2009 respectively are calculated from total evapotranspiration which is in mature closed canopy SRC usually around 30 % higher than pure transpiration, the of poplars is comparable and rather higher than other broadleaved tree species of temperate climate zone. The estimated economically profitable yields range at least from 10 to 12 Mg ha-1 year-1 of DMC which, according to the presented results, will consume more than 450– 500 mm per growing season. Therefore, it is assumed that a locality with higher and adequately temporally distributed amount of precipitation is necessary prerequisite for right site selection, especially in rain fed areas such as the discussed Czech-Moravian Highlands.

211

7

CONCLUSIONS

High density poplar stands are being introduced in the Czech Republic as an alternative crop for energy purposes. Due to the recent EU policy it might be expected that areas with short rotation coppice (SRC) and energy crops in general will spread. Within this context it is necessary to know the ecological demands of the particular species in detail in order to ensure the profitable and sustainable utilization of this bioenergy source. Adequate water availability during the growing season is one of the most important prerequisites for attaining the successful growth and a favourable yield level of SRC established on arable land and therefore the water balance of SRC culture represents the main topic of this thesis. The results of this study reveal that poplar SRC grown in the conditions of the CzechMoravian Highlands has comparable or even slightly lower water consumption than the turf grass or typical grassland while the yield level is reasonably satisfactory. Moreover, the ecophysiological measurements using sap flow and dendrometers in combination with the measurement of soil water availability suggest that the yield of an extensively managed unfertilized poplar stand was not limited by water availability but rather by the length of the grooving season and other growth-determining conditions during the investigated period. The total sum of actual evapotranspiration ( ) in the growing season (1st April to 30th September) 2009 measured by Bowen ratio energy balance method (BREB) was 509 mm and 533 mm for the poplar plantation in the eighth years of the first rotation period and the turf grass respectively. Thereafter, in 2010 after the winter harvest it was 264 mm and 474 mm for the coppiced plantation and the turf grass respectively. The total yearly amounts of in 2009 were 572 mm and 619 mm and in 2010 they were 314 mm and 584 mm for the poplar plantation and the turf grass respectively. The precipitation amounts during the whole year 2009 reached up to 778 mm and to 694 mm within the next year of 2010. During the following year 2011 which was not directly evaluated in this study, two other BREB systems were employed above the second poplar plantation and above the neighbouring typical grassland (unfertilized with one to two cuts per year and mean annual productivity of hay ~3–4 t ha-1 year-1 of dry matter content). The resulting sums of in the growing season 2011 were 484 and 496 mm for the turf grass and the grassland respectively. In the case of the poplars it amounted to 466 and 447 mm for plantation in the second year (the previously mentioned stand) and plantation in the third year (the newly measured plantation) of the second rotation period respectively. Although in some months the of poplars was higher than that of the grass, in seasonal sums was always lower and in no case the of poplars exceeded the available energy expressed in water column or the Penman open water evaporation as reported by several authors. Within this context, it remains an open question how large the plantation has to be, or what the effect of land-use is on the of high density stands of poplars with characteristically high stomatal conductance, to show such excessive evapotranspiration rates. It seems that in the conditions of the experimental site surrounded mostly by grassland, other agricultural crops and SRC, there is no noticeable producer of sensible heat which could enhance the of poplars through advection. The differences between of small versus large 212

scale plantations, of the edges versus locations towards the centre of the stands, and the influence of adjacent cultures might be next tasks of the SRC water use research. Taking into account similar water use of both contrasting cultures, indicated also by comparable soil moisture depletion, it was suggested that successful grassland production might be a good indicator for favourable SRC production if the factors like temperature sums and length of the season along with suitable soil fertility are met. Comparing the diurnal courses, the of turf grass showed typically higher rates in midday peaks during hot and sunny days, whereas the poplars stand showed slightly higher rates during the cloudy days with lower radiation. These characteristic features are caused mainly by different coupling to the atmosphere and might be the effect of different stomatal control. Greater water loss from poplar SRC was also typical after precipitation events when the canopy is wet and has near zero surface resistance which suggests higher interception capacity of SRC given by at least two or three times larger leaf are index ( ). Indeed, the maximum mean canopy measured regularly by ceptometer and validated by litter collection reached up to 7.5 at the end of summer 2009. During the next years following the winter harvest, the mean of the unclosed canopy reached up to 3.7. The of the turf grass was not measured but only estimated from the available literature with respect to its height to be equal to ~2–3. The total stand above-ground woody biomass production accumulated during the 8 years of the first rotation period was 71 t ha-1 of dry matter content (DMC). However, it should be taken into account that the canopy closure occurred even in the fourth year after the planting due to very strong initial competition with weeds. Based on the results from other plots, it was speculated that under well-managed weed control, the canopy closure could be achieved two years earlier and accordingly the same yield might be attained within six years only which would result in more satisfactory mean annual productivity ~12 t ha-1 year-1 instead of ~9 t ha-1 year-1 (DMC). The annual yield based on the allometric stem diameter increment measurements in the last two years of the first rotation was estimated to be 13.4 and 16.5 t ha-1 year-1 (DMC) in 2008 and 2009 respectively. The productivity during the first year after winter harvest was estimated to be equal to 3.9 t ha -1 year-1 (DMC). By relating the above ground woody biomass yield with the total stand evapotranspiration, the so called gross was determined. Mean seasonal stand gross -1 reached up to 3.13 and 3.54 g kg within the last two years of the first rotation respectively and to 2.1 g kg-1 in 2010 following coppice. It can be concluded that it is possible to provide site specific simple models predicting the biomass yield and the growth dynamic based on the concept of . It was demonstrated that measuring sap flow on the four dominant trees only cannot provide relevant estimates of the whole canopy transpiration and even if more sophisticated up-scaling approach is used, the error of the result would be very large. It is therefore necessary for future investigations to evaluate at least double or triple number of sampled trees across the whole social distribution (dominant, co-dominant and suppressed tree classes). However, these sap flow measurements e.g. revealed that the water use of the

213

most vigorous trees reached up to 44 kg day-1 and their integrated over season was -1 3.2 g kg . Except for these main results, it was shown that BREB method is suitable for the calibration of crop coefficient as well as other parameters important for the water balance modelling. In the case of the crop coefficient, it was revealed that it is important and therefore recommended to treat the days with and without rain separately due to the substantial effect of the interception, in order to predict more accurately. Further, BREB method supplemented with wind speed measurement enables obtaining surface resistance by inverting the Penman-Monteith formula and also to reveal the differences between cultures in relation to the coupling to the atmosphere. Furthermore, it was found that BREB method is also suitable for parameterizing the Lohammar equation describing the non-linear dependency of surface resistance on global radiation and water vapour deficit which is crucial for evaluating the differences in the stomatal control of different cultures. Generally, the thesis provides suitable material for calibration and validation complex ecophysiological models applicable on SRC and as well as provides suitable material for establishing simple algorithms for the yield or water use estimates. The application of the mechanistic growth model on SRC will be one of the main aims of my further research. From a technical point of view, the results showed that BREB with the use of commercial combination thin-film polymer capacitive relative humidity and adjacent temperature sensor instruments is applicable for relevant estimates as well as for comparative studies while the costs of the system are much cheaper than the eddy covariance. Besides the price, the system has also much lower energy consumption which makes it suitable for using at remote sites. However, it should be noted that the method with such instrumentation is more accurate over shorter, dense and uniform canopies like grassland or agricultural crops, while it is less accurate towards the taller ones and especially those with patchy structure. This is due to the fact that the taller and sparser stands have larger roughness length and the sensors which have to be placed above the top of the canopy will be far from the zero plane displacement leading to measuring small gradients of any entity – often close to or within the range of instrumental errors. From this point of view, SRC characteristic with high density and short rotation period, and thus not very tall, acts as a transition between the traditional agriculture crops and forests. The analysis proved that the BREB might have the advantage of measuring with adequately smaller footprint by positioning both vertical sensors as low as possible while keeping their sufficient mutual distance. This distance would be dependent on the magnitude of gradients given by the canopy roughness and on the accuracy of the temperature and humidity sensors where some future technical progress may be expected. The object of the next research will be to verify the estimates by BREB method with the independent methods like eddy covariance and scintillometry and to test such type of BREB systems extended by CO2 gas analyser for CO2 flux measurement.

214

REFERENCES AGAM, R., BERLINER, P. R. 2005: Dew formation and water vapor adsorption in semiarid environments – a review. Journal of Arid Environments, Vol. 65: 572–590. AGATA, W., N. AOKI, M. MOROKUMA, AND H. NABESHIMA. 1989: An actual state of photosynthesis and transpiration in zoysiagrass and bentgrass greens. In: Takatoh, H. (eds.): Proceeding 6 th International Turfgrass Resesearch Conference, Tokyo, Japan, 31 July–5 August 1989. Japanese Society of Turfgrass Science and the International Turfgrass Society, Tokyo, p. 133–135. AL AFAS, N., MARRON, N., VAN DONGEN, S., LAUREYSENS, I., CEULEMANS, R. 2008: Dynamics of biomass production in a poplar coppice culture over three rotations (11 years). Forest Ecology and Management, Vol. 255: 1883–1891. ALAVI, N., WARLAND, J. S., BERG, A. A. 2006: Filling gaps in evapotranspiration measurements for water budget studies: Evaluation of a Kalman filtering approach. Agricultural and Forest Meteorology, Vol. 141: 57–66. ALLEN, R. G., PEREIRA, L. S., RAES, D., SMITH, M. 1998: Crop evapotranspiration (Guidelines for computing crop water requirements). FAO Irrigation and Drainage Paper No. 56, 290 pp. ALLEN, R. G., PEREIRA, L. S., SMITH, M., RAES, D., Wright, J. L. 2005: FAO-56 Dual Crop Coefficient Method for Estimating Evaporation from Soil and Application Extensions. Journal of Irrigation and Drainage Engineering, Vol. 13: 2–13. ALLEN, R., G. 1996: Assessing the integrity of weather data for reference evapotranspiration estimation. Journal of irrigation and drainage engineering, Vol. 122: 97–106. ALLEN, S. J., HALL, R. L., ROSIER, P. T. W. 1999: Transpiration by two poplar varieties grown as coppice for biomass production. Tree Physiology, Vol. 19: 493–501. ALTON, P. B., NORTH, P. R., LOS, S. O. 2007: The impact of diffuse sunlight on canopy light-use efficiency, gross photosynthetic product and net ecosystem exchange in three forest biomes. Global Change Biology, Vol. 13: 776–787. ANDREAS, E. L. 1988: Estimating C2n over snow and sea ice from meteorological data. Journal of Optical Society of America, Vol. 5: 481–485. ANDREAS, E. L. 1989: Two-wavelength method of measuring path-averaged turbulent surface heat fluxes. Journal of Atmospheric and Oceanic Technology, Vol. 6: 280–292. ANGUS, D. E., WATTS, P. J. 1984: Evapotranspiration – how good is the Bowen ratio method? Agricultural Water Management, Vol. 8: 133–150. BAKER, J. M., VAN BAVEL, C. H. M. 1987: Measurement of mass flow of water in the stems of herbaceous plants. Plant, Cell and Environment, Vol. 10: 777–782. BALDOCCHI, D. 2004: Advanced topics in biometeorology and micrometeorology. Lecture on micrometeorology flux measurement methods. Department of Environmental Science, Policy and Management. University of California, Berkley (USA), 18 pp. BALDOCCHI, D. D. 1989: Canopy–atmosphere water vapour exchange: can we scale from leaf to a canopy? In: Black, T. A. (ed.): Estimation of Areal Evapotranspiration. Proceeding of an International workshop held at Vancouver, Canada, IAHS Publication No. 177, p. 21–41. BALDOCCHI, D. D., HICKS, B. B., MEYERS, T. P. 1988: Measuring biosphere-atmosphere exchanges of biologically related gases with micrometeorological methods. Ecology, Vol. 69: 1331–40. BALDOCCHI, D. D., LUXMOORE, R. J., HATFIELD, J. L. 1991: Discerning the forest from the trees: an essay on scaling canopy stomatal conductance. Agricultural and Forest Meteorology, Vol. 54: 197–226. BALDOCCHI, D. D., VERMA, S. B., ANDERSON, D. E. 1987: Canopy photosynthesis and water use efficiency in a deciduous forest. Journal of Applied Ecology, Vol. 24: 251–260. BALEK, J., PAVLIK, O. 1977: Sap stream velocity as an indicator of the transpirational process. Journal of Hydrology, Vol. 34: 193–200. BARBAROUX, C., BREDA, N. 2002: Contrasting distribution and seasonal dynamics of carbohydrate reserves in stem wood of adult ring-porous sessile oak and diffuse-porous beech trees. Tree Physiology, Vol. 22: 1201–1210. BARBAROUX, C., BREDA, N., DUFRENE, E. 2003: Distribution of aboveground and belowground carbohydrate reserves in adult trees of two contrasting broad-leaved species (Quercus petraea and Fagus sylvatica). New Phytologist, Vol. 157: 605–615. BARR, A. G., KING, K. M., GILLESPIE, T. J., HARTOG, G., NEUMANN, H. H. 1994: A comparison of Bowen ratio and eddy correlation sensible and latent heat flux measurements above deciduous forest. Boundary-Layer Meteorology, Vol. 71: 21–41.

215

BAUERLE W. L., HINCKLEY T. M., ČERMÁK J., KUČERA J., BIBLE, K. 1999: The canopy water relations of old-growth Douglas-fir trees. Trees, Vol. 13, 211–217. BAUMGARTNER, A., REICHEL, E. 1975: The world water balance: Mean annual global continental and maritime precipitation, evaporation, and runoff. Elsevier, New York, 179 pp. BEARDSELL, M. F., JARVIS, P. G., DAVIDSON, B. 1972: A null-balance diffusion porometer suitable for use with leaves of many shapes. Journal of Applied Ecology, Vol. 9: 677–690. BENETKA, V., VRÁTNÝ, F., ŠÁLKOVÁ, I. 2007: Comparison of productivity of Populus nigra L. With an interspecific hybrid in a short rotation coppice in marginal areas. Biomass and Bioenergy, Vol. 31: 267–274. BERG, L. 2008: Introductory botany: Plants, people, and the environment, 2nd edition. Thomson Brooks/Cole, USA, 622 pp. BERNHOFER, C. 1992: Estimating forest evapotranspiration at a non-ideal site. Agricultural and Forest Meteorology, Vol. 60: 17–32. BERTELA, M. 1989: Inconsistent surface flux partitioning by the Bowen ratio method. Boundary-Layer Meteorology, Vol. 49: 149–167. BIERHUIZEN, J. F., SLATYER, R. O. 1965: Effect of atmospheric concentration of water vapour and CO 2 in determining transpiration-photosynthesis relationships of cotton leaves. Agricultural Meteorology, Vol. 2: 259–270. BLACK, E., BLACKBURN, M., HARRISON, G., HOSKINS, B., METHVE, J. 2004: Factors contributing to the summer 2003 European heatwave. Weather, Vol. 59: 217–223. BLACK, T. A., DEN HARTOG, G., NEUMANN, H. H., BLANKEN, P. D., YANG, P. C., RUSSELL, C., NESIC, Z., LEE, X., CHEN, S. G., STAEBLER, R., NOVAK, M. D. 1996: Annual cycles of water vapour and carbon dioxide fluxes in and above a boreal aspen forest. Global Change Biology, Vol. 2: 219-229. BLAND, B. L., ROSENBERG, N. J. 1974: Lysimetric calibration of the Bowen ratio-energy balance method for evapotranspiration estimation in the central great plains. Journal of Applied Meteorology, Vol. 13: 227–236. BLANEY, H. F., CRIDDLE, W. D. Determining consumptive use and irrigation water requirements. Utah State Engineer, Salt Lake City, UT, USA, 1962, 59 pp. BLANKEN, P. B., BLACK, T. A., NEUMANN, H. H., DEN HARTOG, G., YANG, P. C., NESIC, Z., LEE, X. 2001: The seasonal water and energy exchange above and within a boreal aspen forest. Journal of Hydrology, Vol. 245: 118–136. BLANKEN, P. D., BLACK, T. A., ZANG, P. C., DEN HARTOG, G., NEUMANM, H. H., NESIC, Z., STAEBLER, R., NOVAK, M. D., LEE, X. 1997: Energy balance and canopy conductance of boreal aspen forest: partitioning overstory and understory components. Journal of Geophysical Research, Vol. 102: 28915–28927. BOGENA, H. R., HUISMAN, J. A., OBERDÖRSTER, C., VEREECKEN, H. 2007: Evaluation of a lowcost soil water content sensor for wireless network applications. Journal of Hydrology, Vol. 344: 32–42. BOWEN, I. S., 1926: The ratio of heat looses by conduction and by evaporation from any water surface. Physical Review, Vol. 27: 779–787. BRAATNE, J. H., HINCKLEY, T. M., STETTLER, R. F. 1992: Influence of soil water on the physiological and morphological components of plant water balance in Populus trichocarpa, Populus deltoides and their F1 hybrids. Tree Physiology, Vol. 11: 325–339. BRADLEY, E. F. 1968: A micrometeorological study of velocity profiles and surface drag in the region modified by change in surface roughness. Quarterly Journal of the Royal Meteorological Society, Vol. 94: 361–379. BRÉDA, N., GRANIER, A., AUSSENAC, G. 1995: Effects of thinning on soil and tree water relations, transpiration and growth in an oak forest (Quercus petraea (Matt.) Liebl.). Tree Physiology, Vol. 15: 295–306. BREDE, A. D., DUICH, J. M. 1984: Establishment characteristics of Kentucky bluegrass-perenial ryegrass turf mixtures as affected by seeding rate and ratio. Agronomy Journal, Vol. 76: 875–879. BREMER, D., J. 2003: Evaluation of microlysimeters used in turfgrass evapotranspiration studies using the dual-probe heat-pulse technique. Agronomy Journal, Vol. 95: 1625–1631. BRISSON, N., SEGUIN, B., BERTUZZI, P. 1992: Agrometeorological soil water balance for crop simulation models. Agricultural and Forest Meteorology, Vol. 59: 267–287. BRODRIBB, T. J., HOLBROOK, N. M. 2006: Declining hydraulic efficiency as transpiring leaves desiccate: two types of response. Plant, Cell and Environment, Vol. 29: 2205–2215.

216

BROECKX, L., S., VERLINDEN, M., S., ZONA, D., CEULEMANS, R. 2010: Clonal variation in growth and productivity of Populus in a bio-energy plantation (POPFULL). In: Scarascia-Mugnozza, G., Stanton, B. J. (eds.): Proceeding of 5th International Poplar Symposium, Poplars and willows: from research models to multipurpose trees for a bio-based society, International Union of Forest Research Organizations, Orvieto, Italy, September 20–25, p. 179. BRUTSAERT, W. 1982. Evaporation into the atmosphere. Theory, history and applications. D. Reidel Publishing Company, Dordrecht, Netherlands, 299 pp. BRUTSAERT, W., STRICKER, H. 1979: An advection–aridity approach to estimate actual regional evapotranspiration. Water Resources Research, Vol. 15: 443–450 BRYE, K. R., NORMAN, J. M., BUNDY, L. G. 1999: An equilibrium tension lysimeter for measuring drainage through soil. Soil Science Society of America Journal, Vol. 63: 536–543. BUNGART, R., HÜTTL, R. 2004: Growth dynamics and biomass accumulation of 8-year-old hybrid poplar clones in a short-rotation plantation on a clayey-sandy mining substrate with respect to plant nutrition and water budget. European Journal of Forest Research, Vol. 123: 105–115. BURBA, G., ANDERSON, D. 2010: A brief practical guide to eddy covariance flux measurements. Principles and workflow examples for scientific and industrial applications. LI-COR Biosciences, Nebraska, USA, 212 pp. BURKART, S., MANDERSCHEID, R., WEIGEL, H-J. 2007: Design and performance of a portable gas exchange chamber system for CO2- and H2O-flux measurement in crop canopies. Environmental and Experimental Botany, Vol. 61: 25–34. BURMAN, R. POCHOP, L., O. 1994: Evaporation, evapotranspiration and climatic data. Elsevier Science, Amsterdam, Netherlands, 278 pp. BURT, C. M. 1999: Irrigation water balance fundamentals. Proceeding of the Conference on Benchmarking Irrigation System Performance using Water Measurement and Water Balance, March 1999, San luis Obispo, CA. USCID, Denever, Colo, p: 1–13. BUSCH, G. 2009: The impact of Short Rotation Coppice cultivation on groundwater recharge – a spatial (planning) perspective. In: Isermeyer, F. (ed.): Landbauforschung vTI Agriculture and Forestry Research, Johann Heinrich von Thünen Institute, Vol. 59: No. 3, p. 207–222. BUSINGER, J. A. 1986: Evaluation of the accuracy with which dry deposition can be measured with current micrometeorological techniques. Journal of Applied Meteorology, Vol. 25: 1100–1124. BUSINGER, J. A., WYNGAARD, J. C., IZUMI, Y., BRADLEY, E. F. 1971: Flux-profile relationship in the atmospheric surface layer. Journal of Atmospheric sciences, Vol. 28: 189–191. BUSINGER, J. A., YAGLOM, A. M. 1971: Introduction to Obukhov‘s paper on ‗Turbulence in an atmosphere with a non-uniform temperature‘. Boundary-Layer Meteorology, Vol. 2: 3–6. BUSSÈRE, F., CELLIER, P., DORIGNY, A. 1992: Estimation de la conductivité thermique d‘un sol in situ, à l‘aide d‘une sonde à choc thermique. Agronomie, Vol. 12: 69–78. CALDER, I. R. 1991: Implications and assumptions in using the 'total counts' and convective-dispersion equation for tracer flow measurements in trees. Journal of Hydrology, Vol. 125: 149–158. CALDER, I. R., KARIYAPPA, G. S., SRINIVASALU, N. V., SRINIVASA MURTY, K. V. 1992: Deuterium tracing for the estimation of transpiration from tress, Part 1. Field calibration. Journal of Hydrology, Vol. 130: 17–25. CALDER, I. R., NARAYANSWAMY, M. N., SRINIVASALU, N. V., DARLING, W. G., LARDNER, A. J. 1986: Investigation into the use of deuterium as a tracer for measuring transpiration from eucalypts. Journal of Hydrology, Vol. 84: 345–351. CALIANDRO, A., CATALANO, M., RUBINO, P., BOARI, F. 1990: Research on the suitability of some empirical methods for estimating the reference evapotranspiration in Southern Italy. Proceedings of European Society of Agronomy 1st Conference, Paris, 5–7 December 1990, session 2, p 65. CAMPBELL, G. S. 1975: Steady state diffusion porometers. In: Kanemasu, E. T. (ed.): Measurements of stomatal aperture and diffusive resistance, Washington. Agricultural Experiment Station Bulletin 809, p. 20–23. CAMPBELL, G. S. 1977: An Introduction to Environmental Biophysics. Springer, New York, 159 pp. CAMPBELL, G. S. 1986: Extinction coefficients for radiation in plant canopies using an ellipsoidal inclination angle distribution. Agricultural and Forest Meteorology, Vol. 36: 317–321. CAMPBELL, G. S., GREENWAY, W. C. 2005: Moisture detection apparatus and method. US Patent 6904789, Date Issued: 14 June, 2005. CELLIER, P., BRUNET, Y. 1992: Flux-gradient relationships above tall plant canopies. Agricultural and Forest Meteorology, Vol. 58: 93–117.

217

CELLIER, P., OLIOSO, A. 1993: A simple system for automated long-term Bowen ratio measurement. Agricultural and Forest Meteorology, Vol. 66: 81-92. CELLIER, P., RICHARD, G., ROBIN, P. 1996: Partition of sensible heat fluxes into bare soil and the atmosphere. Agricultural and Forest Meteorology, Vol. 82: 245–265. ČERMÁK, J, DEML, M. 1974: Method of water transport measurements in woody species, especially in adult trees (in Czech). Patent (Certification of authorship) CSFR, No. 155622 (P.V. 5997- 1972). Bureau for Inventions and Discoveries, Prague. ČERMÁK, J. 1998: Solar equivalent leaf area: an efficient biometrical parameter of individual leaves, trees and stands. Tree Physiology, Vol. 5: 269–279. ČERMÁK, J., CIENCIALA, E., KUČERA, J., LINDROTH, A., BEDNÁŘOVÁ, E. 1995: Individual variation of sap flow rate in large pine and spruce trees and stand transpiration: a pilot study at the central NOPEX site. Journal of Hydrology, Vol. 168:17–27. ČERMÁK, J., DEML, J., PENKA, M. 1973: A new method of sap flow rate determination in trees. Biologia Plantarum, Vol. 15:171–178. ČERMÁK, J., HUZULAK, J., PENKA, M. 1980: Water potential and sap flow rate in adult trees with moist and dry soil as used for the assessment of the root system depth. Biologia Plantarum. Vol. 22: 34–41. ČERMÁK, J., JENÍK, J., KUČERA, J., ŢIDEK V. 1984: Xylem water flow in a crack willow tree (Salix fragilis L.) in relation to diurnal changes of environment. Oecologia, Vol. 64: 145–151. ČERMÁK, J., KUČERA, J. 1990: Scaling up transpiration data between trees, stands and watersheds. Silva Carelica, Vol. 15: 101–120. ČERMÁK, J., KUČERA, J., NADEZHDINA, N. 2004: Sap flow measurements with some thermodynamic methods, flow integration within trees and scaling up from sample trees to entire forest stands. Trees, Vol. 18: 529–546. ČERMÁK, J., KUČERA, J., PENKA, M. 1976: Improvement of the method of sap flow rate determination in adult trees based on heat balance with direct electric heating of xylem. Biologia Plantarum, Vol. 18: 105–110. ČERMÁK, J., ÚLEHLA, J., KUČERA, J., PENKA, M. 1982: Sap flow rate and transpiration dynamics in the full-grown oak (Quercus robur L.) in floodplain forest exposed to seasonal floods as related to potential evapotranspiration and tree dimensions. Biologia Plantarum, Vol. 24: 446–460. CEULEMANS, R. SHAO, B. Y., JIANG, X. N., KALINA, J. 1996a: First- and second-year aboveground growth and productivity of two Populus hybrids grown at ambient and elevated CO2. Tree Physiology, Vol. 16: 61–68. CEULEMANS, R., BROECKX, L. S., COOLS, J, DE GROOTE, T., EL KASMIOUI, O., NJAKOU DJOMO, S. VERLINDEN, M., S., ZONA, D. 2010: POPFULL: Toward a full balance of all greenhouse gases, of energy and of the economic profitability of an operational SRC plantation of poplar and willow. In: Scarascia-Mugnozza, G., Stanton, B. J. (eds.): Proceeding of 5 th International Poplar Symposium, Poplars and willows: from research models to multipurpose trees for a bio-based society, International Union of Forest Research Organizations, Orvieto, Italy, September 20–25, p. 179. CEULEMANS, R., McDONALD, A. J. S., PEREIRA, J. S. 1996b: A comparison among eucalypt, poplar and willow characteristics with particular reference to a coppice growth-modelling approach. Biomass and Bioenergy, Vol. 11: 215–231. CHAHINE, M. T. 1992: The hydrological cycle and its influence on climate. Nature, Vol. 359, 373–380. CHAMBERLAIN, A. C. 1966: Transport of gases to and from grass and grass-like surfaces. Proceedings of the Royal Society, Vol. 290: 236–265. CHARLESWORTH, P., STIRZAKER, R. J. 2008: Irrigation scheduling: Soil water status. In: Stanley, W. T. (ed.): Encyclopedia of water science, 2nd edition. Boca Raton: Taylor & Francis Group, p. 611–614. CHARPENTIER, M. A., GROFFMAN, P. 1992: Soil moisture variability within remote sensing pixels. Journal of Geophysical Research, Vol. 97: 18987–18995. CHEN, S. G., CEULEMANS, R., IMPENS, I. 1994: A fractal-based Populus canopy structure model for the calculation of light interception. Forest Ecology and Management, Vol. 69: 97–110. CHEN, W., J., NOVAK, M., D. 1997: An improved tension-plate system for measuring first-stage evaporation under straw mulch. Agricultural and Forest Meteorology, Vol. 85: 1–17. CIENCIALA, E. 1994: Sap flow, transpiration and water use efficiency of spruce and willow in relation to climatic factors. Swedish University of Agricultural Sciences, Uppsala, Sweden, PhD Thesis, 120 pp. CIENCIALA, E., LINDROTH, A. 1994: Gas-exchange and sap flow measurements of Salix viminalis trees in short-rotation forest. I. Transpiration and sap flow. Trees, Vol. 9: 289–294.

218

CIENCIALA, E., ČERNÝ, M., APLTAUER, J., EXNEROVÁ, Z. 2005: Biomass functions applicable to European beech. Journal of Forest Science, Vol. 51: 147–154. CIENCIALA, E., ECKERSTEN, H., LINDROTH, A., HÄLLGREN, J-E. 1994: Simulated and measured water uptake by Picea abbies under non-limiting soil water conditions. Agricultural and Forest Meteorology, Vol. 71: 147–164. CIENCIALA, E., KUČERA, J., LINDROTH, A. 1999: Long-term measurements of stand water uptake in Swedish boreal forest. Agricultural and Forest Meteorology, Vol. 98–99: 547–554. CIENCIALA, E., LINDROTH, A. 1995: Gas-exchange and sap flow measurements of Salix viminalis trees in short-rotation forest. II: Diurnal and seasonal variations of stomatal response and water use efficiency. Trees, Vol. 9: 295–301. CLIFFORD, S., OCHS, G., LAWRENCE, R. 1974: Saturation of optical scintillation by strong turbulence. Journal of the Optical Society of America, Vol. 64: 148–154. COCHARD, H., CASELLA, E., MENCUCCINI, M. 2007: Xylem vulnerability to cavitation varies among poplar and willow clones and correlates with yield. Tree Physiology, Vol. 27: 1761–1767. COHEN, Y., COHEN, S., CANTUARIAS-AVILES, T., SCHILLER, G. 2008: Variations in the radial gradient of sap velocity in trunks of forest and fruit trees. Plant and Soil, Vol. 305: 49–59. COHEN, Y., FUCHS, M., GREEN, G. C. 1981: Improvement of the heat-pulse method for determining sap flow in trees. Plant, Cell and Environment, Vol. 4: 391–397. COLEMAN, M. D., DICKSON, R. E., ISEBRANDS, J. G. 2000: Contrasting fine-root production, survival and soil CO2 efflux in pine and poplar plantations. Plant and Soil, Vol. 225: 129–139. COLEMAN, M. D., FRIEND, A. L., KERN, C. C. 2004: Carbon allocation and nitrogen acquisition in a developing Populus deltoids plantation. Tree Physiology, Vol. 24: 1347–1357. CRAWFORD, T. V. 1965: Moisture transport in free and forced convection. Quarterly Journal of the Royal Meteorological Society, Vol. 91: 18–27. DAAMEN, C., C., SIMMONDS, L., P. 1996: Measurement of evaporation from bare soil and its estimation using surface resistance. Water resources research, Vol. 32: 1393–1402. DALTON, J. 1802: Experimental essays on the constitution of mixed gases: on the force of steam or vapor from water or other liquids in different temperatures, both in a Torricelli vacuum and in air; on evaporation; and on expansion of gases by heat. Manchester Literary and Philosophical Society, Vol. 5: 536–602. DAVANSO, V. M., DE SOUZA, L. A., MEDRI, M. E., PIMENTA, J. A., BIANCHINI, E. 2002: Photosynthesis, growth and development of Tabebuia avellanedae Lor. ex Griseb. (Bignoniaceae) in flooded soil. Brazilian Archives of Biology and Technology, Vol. 45: 375–384. DAVIES, J. A., ALLEN, C. D. 1973: Equilibrium, potential, and actual evaporation from cropped surfaces in southern Ontario. Journal of Applied Meteorology, Vol. 12: 649–647. DAY W. 1977: A direct reading continuous flow porometer. Agricultural Meteorology, Vol. 18: 81–89. DE BRUIN, H. A. R. 1982. The energy balance of the earth‘s surface: a practical approach. Thesis, Agricultural University, Wageningen, 177 p. DE BRUIN, H. A. R., BINK, N. J., KRON, L. J. M. 1991: Fluxes in the surface layer under advective conditions. In: Schmugge, T. J., Andre J. –C., (.s): Land surface evaporation: measurement and parameterization, New York: Springer-Verlag, p. 157–169. DE BRUIN, H. A. R., KEIJMAN, J. Q. 1979: The Priestley-Taylor model applied to a large shallow lake in the Netherlands. Journal of Applied Meteorology, Vol. 18: 898–903. DE BRUIN, H. A. R., KOHSIEK, W., VAN DER HURK, B. J. J. M. 1993: A verification of some methods to determine the fluxes of momentum, sensible heat water vapour using standard deviation and structure parameter of scalar meteorological quantities. Boundary-Layer Meteorology, Vol. 63: 231–257. DE BRUIN, H. A. R., VAN DER HURK, B. J. J. M., KOHSIEK, W. 1995: The scintillation method tested over a dry vineyard area. Boundary-Layer Meteorology, Vol. 76: 25–40. DE JONG, R. 1981: Soil water models: a review. Research Branch, Land Resource Research Institute, Contribution No. 123, Ottawa, Canada, 39 pp. DE JONG, R., BOOTSMA, A. 1995: Review of recent developments in soil water simulation models. Canadian Journal of Soil Science, Vol. 75: 263–273. DE VRIES, D. A. 1963. Thermal properties of soils. In: Van Wijk, W. R. (ed.): Physics of Plant Environment. North-Holland, Amsterdam, p. 210–235. DE WIT, C. T., 1958: Transpiration and crop yields. Institute of Biological and Chemical Research on Field Crops and Herbage, No. 64.6, Wageningen, The Netherlands, 88pp.

219

DEANS, J. D., FORD, E. D. 1986: Seasonal patterns of radial root growth and starch dynamics in plantationgrown Sitka spruce trees of different ages. Tree Physiology, Vol. 1: 241–251. DECAGON DEVICES 2002: ECH2O dielectric aquameter user´s manual—For models EC-20 and EC-10, version1.4. Decagon Devices, Inc., Pullman, Washington, USA, 20 pp. DECKMYN, G., LAUREYSENS, I., GARCIA, J., MUYS, B., CEULEMANS, R. 2004: Poplar growth and yield in short rotation coppice: model simulations using the process model SECRETS. Biomass and Bionergy, Vol. 26: 221–227. DECKMYN, G., VERBEECK, H., OP DE BEECK, M., VANSTEENKISTE, D., STEPPE, K., CEULEMANS, R. 2008: ANAFORE: A stand-scale process-based forest model that include wood tissue development and labile carbon storage in trees. Ecological modelling, Vol. 215: 345–368. DELTA-T DEVICES (2001): User manual for the profile probe type PR1, version 1.2. Delta-T Devices Ltd., Cambridge, UK, 36 pp. DELTA-T DEVICES 1996: SunScan canopy analysis system user manual, version 1.05. Delta-T Devices Ltd., Cambridge, UK, 86 pp. DENMEAD, O. T. 1984: Plant physiological methods for studying evapotranspiration: problems of telling the forest from the trees. Agricultural Water Management, Vol. 8: 167–189. DENMEAD, O. T., DUNIN, F. X., WONG, S. C., GREENWOOD, E. A. N. 1993: Measuring water use efficiency of eucalypt trees with chambers and micrometeorological techniques. Journal of Hydrology, Vol. 150: 649–664. DEWALLE, D. R., SWISTOCK, B. R., JOHNSON, T. E., McGUIRE, K. J. 2000: Potential effects of climate change and urbanization on mean annual streamflow in the United States, Water Resources Research, Vol. 36: 2655–2664. DICKMANN, D. I 2001: An overview of the genus Populus. In: Dickmann, D. I., Isebrands, J. G., Eckenwalder, J. E., Richardson, J. (eds.): Poplar Culture in North America, Ottawa: National Research Council of Canada, p. 1–42. DICKMANN, D. I., ISEBRANDS, J. G., TERENCE, J. B., KOSOLA, K., KORT, J. 2001: Physiological ecology of poplars. In: Dickmann, D. I., Isebrands, J. G., Eckenwalder, J. E., Richardson, J., (eds.): Poplar Culture in North America, Ottawa: National Research Council of Canada, p. 77–118. DICKMANN, D. I., PREGITZER, K. S. 1992: The structure And dynamics of woody plant root systems. In: Mitchell, C. P., Ford-Robertson, J.B., Hinckley, T., Sennerby-Forsse L. (eds.): Ecophysiology of Short Rotation Forest Crops. New York: Elsevier Applied Science, p. 95–123. DOLMAN, A. J. 1985: Estimates of roughness length and zero plane displacement for a foliated and nonfoliated oak canopy. Agricultural and Forest Meteorology, Vol. 36: 241–248. DOORENBOS, J., PRUITT, W. O. 1977: Guidelines for predicting crop water requirements. FAO irrigation and drainage paper No. 24. Food and Agriculture Organization of the United Nations, Rome, Italy, 156 pp. DREXLER, J. Z., SNYDER, R. L., SPANO, D., PAW, U, K, T. 2004: A review of models and micrometeorological methods used to estimate wetland evapotranspiration. Hydrological Processes, Vol. 18: 2071–2101. DRNEVICH, V. P., ASHMAWY, A. K., YU, X., SALLAM, A. M. 2005: Time domain reflectometry for water content and density of soils: study of soil-dependent calibration constants. Canadian Geotechnical Journal, Vol. 42: 1053–1065. DUDAL, R. 1953: Etude morfologique et génétique d´une séquence de sols sur limon loessique [The study of morphology and genetic evolution of loess soils on silt]. Agricultura, Vol. 2: 119–163. DUGAS, W. A., HEUER, M. L., MAYEUX, H. S. 1999: Carbon dioxide fluxes over bermudagrass, native prairie, and sorghum. Agricultural and Forest Meteorology, Vol. 93: 121–139. DUGAS, W. A., WALLACE, J. S., ALLEN, S. J., ROBERTS, J. M. 1993: Heat balance, porometer, and deuterium estimates of transpiration from potted trees. Agricultural and Forest Meteorology, Vol. 64: 47–62. DUGAS, W., A., FRITSCHEN, L., J., GAY, L., W., HELD, A., A., MATTHIAS, A., D., REICOSKY, D., C., STEDUTO, P., STEINER, J., L. 1991: Bowen ratio, eddy correlation, and portable chamber measurements of sensible and latent heat flux over irrigated spring wheat. Agricultural and Forest Meteorology, Vol. 56: 1–20. DYE, P. J., OLBRICH, B.W., CALDER, I. R. 1992: A comparison of the heat pulse method and deuterium tracing method for measuring transpiration from Eucalyptus grandis trees. Journal of Experimental Botany, Vol. 43: 337–343.

220

DYER, A. J. 1965. Discussion on Change of terrain roughness and the wind profile by H. A. Panofsky and A. A. Townsend. Quarterly Journal of the Royal Meteorological Society, Vol. 91: 241. DYER, A. J. 1974. A reviewof flux-profile relationships. Boundary-Layer Meteorology, Vol. 7: 363–372. DYER, A. J., HICKS, B. B. 1970: Flux-gradient relationships in the constant flux layer. Quarterly Journal of the Royal Meteorological Society, Vol. 96: 715–721. DYER, J. M. 2009: Assessing topographic patterns in moisture use and stress using a water balance approach. Landscape Ecology, Vol. 24: 391–403. EICHINGER, W. E., PARLANGE, M. B., STRICKER, H. 1996: On the concept of equilibrium evaporation and the value of the Priestley-Taylor coefficient. Water Resources Research, Vol. 32: 161–164. ELIOTT, W. P. 1958: The growth of the atmospheric internal boundary layer. Transactions, American Geophysical Union, Vol. 39: 1048–1054. EMS 2004: EMS 33 air temperature and humidity sensor user´s manual. Environmental Measuring System, Brno, Czech Republic, 5 pp. EMS 2009: Datalogger ModuLog user´s manual. Environmental Measuring System, Brno, Czech Republic, 11 pp. EMS 2010: EMS 51 measuring module—Installation guide. Environmental Measuring System, Brno, Czech Republic, 12 pp. EMS 2010: EMS 51 sap flow system for large tree trunks—Instruction Manual. Environmental Measuring System, Brno, Czech Republic, 11 pp. EMS 2010: EMS 51Sap flow system – instruction manual. Brno, 11 pp. ERICSSON, K., ROSEQVIST, H., NILSSON, L. J. 2009: Energy crop production costs in the EU. Biomass and Bioenergy, Vol. 33: 1577–1586. EUROPEAN COMMISSION 2005: Biomass action plan. 6282005, Brussels. EVERT, R. F. 2006: Esau‘s plant anatomy, meristems, cells, and tissues of the plant body: their structure, function, and development, 3rd edition. Wiley, New Jersey, 601 pp. EVETT, S. R., SCHWARTZ, R. C., TOLK, J. A., HOWELL, T. A. 2009: Soil profile water content determination: spatiotemporal variability of electromagnetic and neutron probe sensors in access tubes. Vadose Zone Journal, Vol. 8: 926–941. EZZAHAR, J., CHEHBOUNI, A., HOEDEJES, J. C. B., CHEHBOUNI, AH. 2007: On the application of scintillometry over heterogeneous surfaces. Journal of Hydrology, Vol. 34: 493–501. FAJMAN, M., PALÁT, M., SEDLÁK, P. 2009: Estimation of the yield of poplars in plantations of fast– growing species within current results. Acta Universitatis Agriculture et Silviculturae Mendelianae Brunensis, Vol. 2: 25–36. FALGE, E., BALDOCCHI, D., OLSON, R., ANTHONI, P., AUBINET, M., BERNHOFER, C., BURBA, G., CEULEMANS, R., CLEMENT, R., DOLMAN, H., GRANIER, A., GROSS, P., GRÜNWALD, T., HOLLINGER, D., JENSEN, N. O., KATUL, G., KERONEN, P., KOWALSKI, A., LAI, C. T., LAW, B. E., MEYERS, T., MONCRIEFF, J., MOORS, E., MUNGER, J. W., PILEGAARD, K., RANNIK, Ü., REBMANN, C., SUYKER. A., TENHUNEN, J., TU, K., VERMA, S., VESALA, T., WILSON, K., WOFSY, S. 2001: Gap filling strategies for defensible annual sums of net ecosystem exchange. Agricultural and Forest Meteorology, Vol. 107: 43–69. FAMIGLIETTI, J. S., DEVEREAUX, J. A., LAYMON, C. A., TSEGAYE, T., HOUSER, P. R., JACKSON, T. J., GRAHAM, S. T., RODELL, M., VAN OEVELEN, P. J. 1999: Ground–based investigation of soil moisture variability within remote sensing footprint during the Southern Great Plains 1997 (SGP97) Hydrology Experiment. Water Resources Research, Vol. 6: 1839–1851. FAMIGLIETTI, J. S., RUDNICKI, J.W., RODELL, M. 1998: Variability in surface moisture content along a hillslope transect: Rattlesnake Hill, Texas. Journal of Hydrology, Vol. 210: 259–281. FEDDES, R. A., LENSELINK, K. J. 1994: Evapotranspiration. In: H.P. Ritzema (ed.): Drainage Principles and Applications. International Institute for Land Reclamation and Improvement Publication, No. 16, Wageningen, p. 145–173. FEDDES, R.A., DE GRAAF, M., BOURMA, J., VAN LOON, C. D. 1988: Simulation of water use and production of potatoes as affected by soil compaction. Potato Research, Vol. 31: 225–239. FERNANDEZ, E., ČERMÁK, J., COHEN, Y., FERREIRA, I., NADEZHDINA, N., TESTI, L. 2011: Methods to Estimate Sap Flow. ISHS Working Group on Sap Flow, 1–7. FICHOT, R., BROECKX, L. S., VERLINDEN, M. S., BERHONGARAY, G., ZONA, D., DILLEN, S. Y., KING, J. S., STEPPE, K., CEULEMANS, R. 2011: Water balance and water-use efficiency of a poplar bio-energy plantation. In: Merino, A., Gallardo, J. F., Bredemeier, M. (eds.): Proceeding of

221

International Year of Forests, Managed forests in future landscapes – Implications for water and carbon cycles. Santiago de Compostela, Spain, May 8–11, p. 160. FICHOT, R., LAURANS, F., MONCLUS, R., MOREAU, A., PILATE, G., BRIGNOLAS, F. 2009: Xylem anatomy correlates with gas exchange, water-use efficiency and growth performance under contrasting water regimes: evidence from Populus deltoides x Populus nigra hybrids. Tree Physiology, Vol. 29: 1537–1549. FICHTNER, K., SCHULZE, E-D. 1990: Xylem water flow in tropical vines as measured by a steady state heating method. Oecologia, Vol. 82: 355–361. FINCH, J. W., HALL, R. L., ROSIER, P. T. W., CLARK, D. B., STRATFORD, C., DAVIES, H. N., MARSH, T. J., ROBERTS, J. M., RICHE, A., CHRISTIAN, D. 2004: The hydrological impacts of energy crop production in the UK. Final report. Department of Trade and Industry, London, 151 pp. FISCHER, M., TRNKA, M., KUČERA, J., ŢALUD, Z. 2010: Soil water availability in a short rotation poplar coppice (Populus nigra x P. maximowiczii) in Czech-Moravian Highlands. Folia Oecologica, Vol. 37: 23–34. FISCHER, R. A. 1968: Stomatal opening: role of potassium uptake by guard cells. Science, Vol. 160: 784– 785. FITTER, A. H., HAY, R. K. M. 2002: Environmental Physiology of Plants, 3 rd edition. Academic Press, San Diego, CA, USA, 397 pp. FOKEN, T. 2008a: Micrometeorology. Springer, Berlin, 306 pp. FOKEN, T. 2008b: The energy balance closure problem: An overview. Ecological Applications, Vol. 18: 1351–1367. FOKEN, T., RICHTER, S. H., MÜLLER, H. 1997: Zur Genauigkeit der Bowen-Ratio-Methode. Wetter und Leben, Vol. 49: 57–77. FORRESTER, D. I., THEIVEYANATHAN, S., COLLOPY, J. J., MARCAR, N. E. 2010: Enhanced water use efficiency in a mixed Eucalyptus globulus and Acacia mearnsii plantation. Forest Ecology and Management, Vol. 259: 1761–1770. FUCHS, M., COHEN, Y., MORESHET, S. 1987: Determining transpiration from meteorological data and crop characteristics for irrigation management. Irrigation Science, Vol. 8: 91–99. FUCHS, M., TANNER, C. B. 1970: Error analysis of Bowen ratios measured by differential psychrometry. Agriculture and Meteorology, Vol. 7: 329–334. GARDNER, C. M. K., ROBINSON, D. A., BLYTH, K., COOPER, J. D. 2001: Soil water content. In: K. A. Smith and C. E. Mullins (ed.): Soil and environmental analysis: Physical methods, 2 nd edidition. Marcel Dekker, New York, p. 1–64. GARRAT. J. S. 1972: Momentum, mass and heat exchange of vegetation. Quarterly Journal of the Royal Meteorological Society, Vol. 98: 124–134. GARRATT, J. R. 1990: The internal boundary layer – A review. Boundary-Layer Meteorology, Vol. 50: 171–203. GARRATT, J. R., HICKS, B. B. 1973: Momentum, heat and water vapour transfer to and from natural and artificial surfaces. Quarterly Journal of the Royal Meteorological Society, Vol. 99: 680–687. GARRATT, J. R., HICKS, B. B. 1973: Momentum, heat and water vapour transfer to and from natural and artificial surfaces. Quarterly Journal of the Royal Meteorological Society, Vol. 99: 680–687. GASH, J. H. C. 1986: A note on estimating the effect of a limited fetch on micrometeorological evaporation measurements. Boundary-Layer Meteorology, Vol. 35: 409–413. GASKIN, G. J., MILLER, J. D. 1996: Measurement of soil water content using a simplified impedance measuring technique. Journal of Agricultural Engineering Research, Vol. 63: 153–160. GASOL, C. M., GABARRELL, X., ANTON, A., RIGOLA, M., CARRASCO, J., CIRIA, P. 2007: Life Cycle, Assessment of a Brassica carinata bioenergy system in southern Europe. Biomass and Bioenergy, Vol. 31: 543–555. GAVILÁN, P., BERENGENA, J. 2007: Accuracy of the Bowen ratio-energy balance method for measuring latent heat flux in a semiarid advective environment. Irrigation Science, Vol. 25: 127–140. GAZAL, R., M., SCOTT, R., L., GOODRICH, D., C., WILLIAMS, D., G. 2006: Controls on transpiration in a semiarid riparian cottonwood forest. Agricultural and Forest Meteorology, Vol. 137: 56–67. GEE, G. W., WARD, A. L., CALDWELL, T. G., AND RITTER, J. C. 2002: A vadose zone water flux meter with divergence control. Water Resources Research, Vol. 38: 35–41. GIELEN, B., CALFAPIETRA, C., SABATTI, M., CEULEMANS, R. 2001: Leaf area dynamics in a closed poplar plantation under free-air carbon dioxide enrichment. Tree physiology, Vol. 21: 1245–1255.

222

GILBERT, D., LE MOUËL, J.-L., LAMBS, L., NICOLLIN, F., PERRIER, F., 2006: Sap flow and daily electric potential variations in a tree trunk. Plant Science, Vol. 171: 572–584. GILLMAN, K. 1977: Movement of heat in soils. Institute of Hydrology Report No. 44, 44 pp. GODWIN, R. J., MILLER, P. C. H. 2003: A review of the technologies for mapping within-field variability. Biosystems Engineering, Vol. 84: 393–407. GOODISON, B. E. (1978): Accuracy of Canadian snow gage measurements. Journal of Applied Meteorology Vol. 27: 1542– 1548. GOODISON, B. E., LOUIE, P. Y. T., YANG, D. 1998: WMO solid precipitation measurement intercomparison. World Meteorological Organisation, Geneva, WMO/TD 872, 212 pp. GOODISON, B. E., VET, R. J. 1989: Precipitation data compatibility in north America and the impact on studies of acid deposition. In: Delleur, J.W. (eds.): Atmospheric Deposition (Proceedings of the Baltimore Symposium, May 1989). IAHS Publ. No. 179, p. 47–55. GRACE, J. 1983: Plant–atmosphere relationships (Outline studies in ecology). Chapman and Hall, London, 92 pp. GRANIER, A. 1985: Une nouvelle méthode pour la mesure du flux de sève brute dans le tronc des arbres. Annales des Sciences Forestières, Vol. 42: 193–200. (in French). GRANIER, A. 1987: Evaluation of transpiration in a Douglas-fir stand by means of sap flow measurements. Tree Physiology, Vol. 3: 309–320. GRANIER, A., ANFODILLO, T., SABATTI, M., COCHARD, H., DREYER, E., TOMASI, M., VALENTINI, R., BRÉDA, N. 1994: Axial and radial water flow in the trunks of oak trees: a quantitative and qualitative analysis. Tree Physiology, Vol. 14: 1383–1396. GREEN, S. R., CLOTHIER, B. E., JARDINE, B. 2003: Theory and practical application of heat-pulse to measure sap flow. Agronomy Journal, Vol. 95: 1371–1379. GRELLE, A., LUNDBERG, A., LINDROTH, A., MOREN, A. S., CIENCIALA, E. 1997: Evaporation components of a boreal forest: variations during the growing season. Journal of Hydrology, Vol. 197: 70–87. GRENEDIGE, K. N. H. 1955: Observations on the movement of moisture in large woody stems. Canadian Journal of Botany, Vol. 33: 202–221. GRIP, H. 1981: Evapotranspiration experiments in Salix stands. Technical Report 15, Swedish University of Agricultural Sciences, Section of Energy Forestry, Uppsala, 29 pp. GRIP, H., HALLDIN, S., LINDROTH, A. 1989: Water use by intensively cultivated willow using estimated stomatal parameter values. Hydrological Processes, Vol. 3: 51–63. GROVES, S. J., ROSE, S. C. 2004: Calibration equations for Diviner 2000 capacitance measurements of volumetric soil water content of six soils. Soil Use and Management, Vol. 20: 96–97. GUIDI, W., PICCIONI, E., BONARI, E. 2008: Evapotranspiration and crop coefficient of poplar and willow short-rotation coppice used as vegetation filter. Bioresource Technology, Vol. 99: 4832—4840. GUO, X., ZHANG, H., CAI, X., KANG, L., ZHU, T., LECLERC, M., Y. 2009: Flux-variance method for latent heat and carbon dioxide fluxes in unstable conditions. Boundary-Layer Meteorology, Vol. 131: 363–384. GUO, X., ZHANG, H., KANG, L. DU, J., LI, W., ZHU, Z. 2007: Quality control and flux gap filling strategy for Bowen ratio method: revisiting the Priestley–Taylor evaporation model. Environmental Fluid Mechanics, Vol. 7: 421–437. HALL, R. L. 2002: Aerodynamic resistance of coppiced poplar. Agricultural and Forest Meteorology, Vol. 114: 83–102. HALL, R. L. 2003: Short rotation coppice for energy production Hydrological guidelines. Department of Trade and Industry, Westminster, 21 pp. HALL, R. L., ALLEN, S. J. 1997: Water use of poplar clones grown as short-rotation coppice at two sites in the United Kingdom. In: Bullard, M.J., Ellis, R.G., Heath, M.C., Knight, J.D., Lainsbury, M.A., Parker, S.R. (eds.): Aspects of Applied Biology 49, Biomass and Bioenergy Crops. Association of Applied Biologists. Wellesbourne, Warwick, UK, p. 163–172. HALL, R. L., ALLEN, S. J., ROSIER, P. T. W., HOPKINS, R. 1998: Transpiration from coppiced poplar and willow measured using sap flow methods. Agricultural and Forest Meteorology, Vol. 90: 275–290. HALL, R. L., ALLEN, S. J., ROSIER, P. T. W., SMITH, D. M., HODNETT, M. G., ROBERTS, J. M., HOPKINS, R., DAVIES, H., KINNIBURGH, D. G., GOODDY, D. C. 1996: Hydrological Effects Of Short Rotation Energy Coppice. ETSU B/W5/00275/REP. ETSU for DTI. Harwell., 226 pp.

223

HALLDIN, S. 1989: Willow stand evaporation: simulation of diurnal distribution using synoptic weather data. In: Perttu, K. L., Kowalik, P. J. (eds.): Modelling of energy forestry: growth, water relations and economics. Simulation Monographs 30, Pudoc, Wageningen, p. 121–146. HALLIWELL, D., H., ROUSE, W., R. 1989: A Comparison of Sensible and Latent Heat Flux Calculations Using the Bowen Ratio and Aerodynamic Methods. American Meteorological Society, Vol. 6: 563–574. HANSEN, E. A. 1988: Irrigating short rotation intensive culture hybrid poplars. Biomass, Vol. 16: 237–250. HANSEN, E. A. 1991: Poplar woody biomass yields: A look to the future. Biomass and Bioenergy, Vol. 1: 1–7. HARBECK, G. E, Jr. 1966: Annual variability in lake evaporation. International Association of Scientific Hydrology Publication, Vol. 70: 294–303. HARGREAVES, G. H., SAMANI, Z. A. 1985: Reference crop evapotranspiration from temperature. Applied Engineering in Agriculture, Vol. 1(2): 96–99. HASLWANTER, A., HAMMERLE, A., WOHLFAHRT, G. 2009: Open-path vs. closed-path eddy covariance measurements of the net ecosystem carbon dioxide and water vapour exchange: A long-term perspective. Agricultural and Forest Meteorology, Vol. 149: 291–302. HATFIELD, J. L. 1985: Estimation of crop water requirements: present and future. In: Perrier, A., Riou, C. (eds.): Les Besoins en Eau des Cultures. Paris, p. 3–21. HATTON, T. J., VERTESSY R. A. 1989: Variability of sapflow in a Pinus radiata plantation and the robust estimation of transpiration. In: Hydrology and water resources symposium, Australian Institution of Engineers, Christchurch, 23–30th November 1989, New Zealand, p. 6–10. HAVLÍČKOVÁ, K., WEGER, J., HAVLÍČKOVÁ, B., VRÁTNÝ, F., TÁBORSKÁ, M., KNÁPEK, J., VAŠÍČEK, J., SOUČKOVÁ, H., GALLO, P., HEJDA, J., KAJAN, M. 2006: Methodology for analysis of biomass potential as renewable source of energy. Acta Pruhoniciana, Vol. 83: 1–96. HAVLÍČKOVA, K., WEGER, J., ŠEDIVÁ, J. 2010: Methodology of analysis of biomass potential using GIS in the Czech Republic. Acta Universitatis Agriculture et Silviculturae Mendelianae Brunensis, Vol. 58: 161–170. HAWLEY, M., JACKSON, T., McCUEN, R. 1983: Surface soil moisture variation in small agricultural watersheds. Journal of Hydrology, Vol. 62: 179–200. HEIKINHEIMO, M., KANGAS, M., TOURULA, T., VENELAINEN, A., TATTARI, S. 1999: Momentum and heat fluxes over lakes Tamnaren and Raksjo determined by the bulk-aerodynamic and eddycorrelation methods. Agricultural and Forest Meteorology, Vol. 98–99: 521–534. HEILMAN, J., L., BRITTIN, C., L., NEALE, C., M., U. 1989: Fetch requirements for Bowen ratio measurements of latent and sensible heat fluxes. Agricultural and Forest Meteorology, Vol. 44: 261– 273. HEILMAN, P. E., HINCKLEY, T. M., ROBERTS, D. A., CEULEMANS, R. 1996: Production physiology. In: Stettler, R. F., Bradshaw, H. D., Heilman, P. E., Hinckley, T. M. (eds.): Biology of populus and its implications for management and conservation, Otawa: National Research Council of Canada, p. 459– 488. HELLKVIST J., RICHARDS, G. P. JARVIS, P. G. 1974: Vertical gradients of water potential and tissue water relations in Sitka spruce trees measured with the pressure chamber. Journal of Applied Ecology, Vol. 11: 637–667. HELLRIEGEL, H. 1883: Beiträge zu den naturwissenschaftlichen Grundlagen des Ackerbaus mit besonderer Berücksichtigung der agrikultur-chemischen Methode der Sandkultur. Friedrich Vieweg und Sohn, Braunschweig, 796 pp. HENDERSON, D. E., JOSE, S. 2005: Production physiology of free fast-growing hardwood species along a resource gradient. Tree Physiology, Vol. 25: 1487–1494. HENNINGER, D. L., PETERSON, G. W., ENGMAN, E. T. 1976: Surface soil moisture within a watershed– variations, factors influencing, and relationships to surface runoff. Soil Science Society of America Journal, Vol. 40: 773–776. HERZOG, K. M., HÄSLER, R., THUM, R. 1995: Diurnal changes in the radius of a subalpine Norway spruce stem: their relation to the sap flow and their use to estimate transpiration. Trees, Vol. 10: 94– 101. HIGNETT, C., EVETT, S. 2008: Capacitance sensors for use in access tubes. In: Evett, S. R., Heng, L. K., Moutonnet, P., Nguyen, M. L. (eds.): Field Estimation of Soil Water Content: A practical guide to methods, instrumentation, and sensor technology. IAEA-TCS-30. International Atomic Energy Agency, Vienna, Austria, p. 73–90.

224

HILL, R. J., CLIFFORD, S. F., LAWRENCE, R. S. 1980: Refractive index and absorption fluctuation in the infrared caused by temperature, humidity and pressure fluctuations. Journal of the Optical Society of America, Vol. 70: 1129–1205. HILL, R. J., OCHS, G. R., WILSON, J. J. 1992a: Measuring surface-layer fluxes of heat and momentum using optical scintillation. Boundary-Layer Meteorology, Vol. 58: 391–408. HILL, R. J., OCHS, G. R., WILSON, J. J. 1992b: Surface-layer fluxes measured using the C2T -profile method. Journal of Atmosphere and Oceanic Technology, Vol. 9: 526–537. HILLEL, D. 1971: Soil and Water. Academic, New York, USA, 288 pp. HILLS, T. C., REYNOLDS, S. G. 1969: Illustrations of soil moisture variability in selected areas and plots of different sizes. Journal of Hydrology, Vol. 8: 27–47. HINCKLEY, T. M., ASLIN, R.G., AUBUCHON, R. R., METCALF, C. L., ROBERTS, J. E. 1978: Leaf conductance and photosynthesis in four species of the oak-hickory forest type. Forest Science, Vol. 24:73–84. HINCKLEY, T. M., BROOKS, J. R, ČERMÁK, J., CEULEMANS, R., KUČERA, J., MEINZER, F. C., ROBERTS, D. A. 1994: Water flux in a hybrid poplar stand. Tree Physiology, Vol. 14: 1005–1018. HINCKLEY, T. M., LASSOIE, J. P. 1981: Radial growth in conifers and deciduous trees: a comparison. Mitteilungen der forstlichen Bundesversuchsanstalt Wien, Vol. 142: 17–56. HLAVINKA, P., TRNKA, M., BALEK, J., SEMERÁDOVÁ, D., HAYES, M., SVOBODA, M., EITZINGER, J., MOŢNÝ, M., FISCHER, M., HUNT, E., ŢALUD, Z. 2011: Development and evaluation of the SoilClim model for water balance and soil climate estimates. Agricultural Water Management, Vol. 98: 1249–1261. HOCH, G., RICHTER, A., KORNER, C. 2003: Non-structural carbon compounds in temperate forest trees. Plant, Cell and Environment, Vol. 26: 1067–1081. HÖGSTRÖM, U., BERGSTRÖSM, H., SMEDMAN, A-S., HALLDIN, S., LINDROTH, A. 1989: Turbulent exchange above a pine forest, I: Fluxes and gradients. Boundary-Layer Meteorology, Vol. 49: 197–217. HÖLL, W. 1985: Seasonal fluctuation of reserve materials in the trunkwood of spruce Picea abies (L.) Karst. Journal of Plant Physiology, 117: 355–362. HOLMES, J. W. 1984: Measuring evapotranspiration by hydrological methods. Agricultural Water Management, Vol. 8: 29–40. HOLMGREN, P., JARVIS, P. G., JARVIS, M. S. 1965: Resistance to carbon dioxide and water vapour transfer in leaves of different plant species. Physiologia Plantarum, Vol. 18: 557–73. HÖNGSTRÖM, U. 1996: Review of some basic characteristics of the atmospheric surface layer. BoundaryLayer Meteorology, Vol. 78: 215–246. HORST, T. W., WEIL, J. C. 1993: How far is far enough? The fetch requirements for micrometeorological measurement of surface fluxes. Journal of atmospheric and oceanic technology, Vol. 11: 1018−1025. HOU, L. G., XIAO, H. L. SI, J. H. XIAO, S. C., ZHOU, M. X. YANG, Y. G. 2010: Evapotranspiration and crop coefficient of Populus euphratica Oliv forest during the growing season in the extreme arid region northwest China. Agricultural Water Management, Vol. 97: 351–356. HSIAO, T. C., ACEVEDO, E. 1974: Plant responses to water deficits, water-use efficiency, and drought resistance. Agricultural Meteorology, Vol. 14: 59–84. HUBER, B. 1932: Beobachtung und Messung pflanzlicher Saftstrome. Berichte der Deutschen botanischen Gesellschaft, Vol. 50: 89–109. HUBER, B., SCHMIDT, E. 1936: Weitere thermoelektrische Untersuchungen uber den Transpirationsstrom der Baume. Tharandter Forstliche Jahrsblad, Vol. 87: 369–412. HUI, D., WAN, S., SU, B., KATUL, G., MONSON, R., LUO, Y. 2004: Gap-filling missing data in eddy covariance measurements using multiple imputation (MI) for annual estimations. Agricultural and Forest Meteorology, Vol. 121: 93–111. HUPET, F., VANCLOOSTER, M. 2002: Intraseasonal dynamics of soil moisture variability within a small agricultural maize cropped field. Journal of Hydrology, Vol. 261: 86–101. INTERNATIONAL ENERGY AGENCY 2003: IEA statistics-renewable in- formation, IEA/OECD, Paris. IRITZ, Z., LINDROTH, A. 1996: Energy partitioning in relation to leaf area development of short-rotation willow coppice. Agricultural and Forest Meteorology, Vol. 81: 119-130. IRITZ, Z., TOURULA, T., LINDROTH, A., HEIKINHEIMO, M. 2001: Simulation of willow short-rotation forest evaporation using a modified Shuttleworth–Wallace approach. Hydrological Processes, Vol. 15: 97–113.

225

IRVINE, M. R., GARDINER, B. A., HILL, M. K. 1997: The evolution of turbulence across a forest edge. Boundary-Layer Meteorology, Vol. 84: 467–496. ISEBRANDS, J. G., KARNONSKY, D. F. 2001: Environmental benefits of poplar culture. In: Dickmann, D. I., Isebrands, J. G., Eckenwalder, J. E., Richardson, J., (eds.): Poplar Culture in North America, Ottawa: National Research Council of Canada, p. 207–218. ITIER, B. 1996: Measurement and estimation of evapotranspiration. In: Pereira, L. S. (ed.): Sustainability of Irrigated Agriculture. Kluwer Academic, Dordrecht, p. 171–191. ITTNER, E. 1968: Der Tagesgang der Geschwindigkeit des Transpirationsstromes im Stammeiner 75jahringer Fichte. Oecologia Plantarum, Vol. 3: 177–183. JACKSON, A. B., CANADELL, J., EHLERINGER, J. R., MOONEY, H. A., SALA, O. E., SCHULZE, E. D. 1996: A global analysis of root distributions for terrestrial biomes. Oecologia, Vol. 108: 389–411. JACKSON, R. B., MOONEY, H. A., SCHULZE, E. D. 1997: A global budget for fine root biomass, surface area, and nutrient contents. Proceedings of the National Academy of Sciences, Vol. 94: 7362–7366. JAMES, S. A., MEINZER, S. A., GOLDSTEIN, G., WOODRUFF, D., JONES, T., RESTOM, T., MEJIA, M., CLEARWATER, M., CAMPANELLO, P. 2003: Axial and radial water transport and internal water storage in tropical forest canopy trees. Oecologia, Vol. 134: 37–45. JARVIS, P. G., McNAUGHTON, K. G. 1986: Stomatal control of transpiration: scaling up from leaf to region. Advances in Ecological Research, Vol. 15: 1–49. JARVIS, P.G. 1976: The interpretation of the variation in leaf water potential and stomatal conductance found in canopies. Philosophical Transactions of the Royal Society of London Series B., Vol. 273: 593– 610. JEDEGE, O. O., FOKEN, T. 1999: A study of the internal boundary layer due to a roughness change in neutral conditions during the LINEX field campaigns. Theoretical and Applied Climatology, Vol. 62: 31–41. JENSEN, M. E. HAISE, H. R. 1963: Estimating evapotranspiration from solar radiation. Proceedings of the American Society of Civil Engineers, Journal of Irrigation and Drainage Division, Vol. 89: 15–41. JENSEN, M. E., BURMAN, R. D., ALLEN, R. G. 1990. Evaporation and irrigation water requirements. ASCE Manual and Reports on Engineering Practice No. 70. American Society of Civil Engineers, New York, USA, p. 360. JENSEN, M. E., D. C. N. ROBB, FRANZOY, C. E. 1970: Scheduling irrigations using climate-crop-soil data. Proceeding of the American Society of Civil Engineers, Journal of Irrigation and Drainage Division, Vol. 96: 25–38. JONES, H. G. 1992: Plants and microclimate: a quantitative approach to environmental physiology, 2 nd edition. Cambridge University Press, Great Britain, 428 pp. JONES, R. J., MANSFIELD, T. A. 1972: Effects of abscisic acid and its esters on stomatal aperture and the transpiration ratio. Physiologia Plantarum, Vol. 26: 321–327. KAIMAL, J. C., FINNIGAN, J. J. 1994: Atmospheric boundary layer flows: Their structure and measurement. Oxford University Press, New York, USA, 289 pp. KAIMAL, J. C., GANYOR, J. E. 1991: Another look to sonic thermometry. Boundary-Layer Meteorology, Vol. 56: 401–410. KAKANE, V., C., K., AGYEI, E., K. 2006: Determination of surface fluxes using a Bowen ratio system. West Africa Journal of Applied Ecology, Vol. 9: 1–9. KALVOVÁ, J., NEMEŠOVÁ, I. 1997: Projection of climate change for the Czech Republic. Climatic Change, Vol. 36: 41–64. KATERJI, N., DAUDED, F.A., CARBONNEAU, A., OLLAT, N. 1994: Etude à l‘échelle de la plante entière du fonctionnement hydrique et photosynthétique de la vigne: comparaison des système de conduite traditionnel et en Lyre. Vitis, Vol. 33: 197–303. (in French). KATERJI, N., PERRIER, A. 1985: Détermination de la résistance globale d‘un couvert végéal à la diffusion de vapeur d‘eau et de ses differentes composantes. Approche théorique et vérification expérimentale sur une culture de luzerne. Agricultural and Forest Meteorology, Vol. 34: 105–120. (in French). KHAMZINA, A., SOMMER, R., LAMERS, J. P. A., VLEK, P. L. G. 2009: Transpiration and early growth of tree plantations established on degraded cropland over shallow saline groundwater table in northwest Uzbekistan. Agricultural and Forest Meteorology, Vol. 149: 1865–1874. KINOSHITA T, DOI, M., SUETSUGU, N., KAGAWA, T., WADA, M., SHIMAZAKI, K. I. 2001: Phot1 and phot2 mediate blue light regulation of stomatal opening. Nature, Vol. 414: 656–660.

226

KIZITO, F., CAMPBELL, C. S., CAMPBELL, G. S., COBOS, D. R., TEARE, B. L., CARTER, B., HOPMANS, J. W. 2008: Frequency, electrical conductivity and temperature analysis of a low-cost capacitance soil moisture sensor. Journal of Hydrology, Vol. 352: 367– 378. KLINE, J. R., MARTIN, J. R., JORDAN, C. F., KORANDA, J. J. 1970: Measurement of transpiration in tropical trees with tritiated water. Ecology, Vol. 5: 1068–1073. KNOHL, A., BALDOCCHI, D. D. 2008: Effects of diffuse radiation on canopy gas exchange processes in a forest ecosystem. Journal of Geophysical Research, Vol. 113, G02023, doi:10.1029/2007JG000663. KÖCKENBERGER, W., POPE, J.M., XIA Y., JEFFREY, K. R., KOMOR, E., CALLAGHAN P. T. 1997: A non-invasive measurement of phloem and xylem water flow in castor bean seedlings by nuclear magnetic resonance microimaging. Planta, Vol. 201: 53–63. KOHSIEK, W. 1982a: Optical and in situ measuring of structure parameters relevant to temperature and humidity, and their application to the measuring of sensible and latent heat flux. National Oceanic and Atmospheric Administration Technical Memorandum, Environmental Research Laboratories, Wave Propagation laboratory-96, Boulder, CO, USA, 64 pp. KOHSIEK, W. 1982b: Measuring C2T , C2q and CTq in the unstable surface layer, and relations to the vertical fluxes of heat and moisture, Boundary-Layer Meteorology, Vol. 24: 89–107. KOHSIEK, W., HERBEN, M. H., A., J. 1983: Evaporation derived from optical and radio wave scintillation. Applied Optics, Vol. 22: 2566–2569. KOPEC, D. M., J. M. NORMAN, R. C. SHEARMAN, PETERSON, M. P. 1987: An indirect method for estimating turfgrass leaf area index. Crop Science, Vol. 27: 1298–1301. KOPPÁN, A., SZARKA, L., WESZTERGOM, V. 2005: Local variability of electric potential differences on the trunk of Quercus cerris L. Acta Silvatica et Lignaria Hungarica, Vol. 1: 73–81. KÖSTNER, B., GRANIER, A., ČERMÁK, J. 1998: Sap flow measurement in forest stands-methods and uncertainties. Annals of Forest Science, Vol. 55: 13–27. KOŢNAROVÁ, V., KLABZUBA, J. 2002: Recommendation of World Meteorological Organization to describing meteorological or climatological conditions. Plant Production, Vol. 48: 190–192. (In Czech). KRAMER, P. J. 1983: Water relation of plants. Academic Press. San Diego, CA. 489 pp. KUČERA, J. 1977: A system for water flux measurements in plants. Patent (Certificate of authorship) CSFR No. 185039 (P. V. 2651-1976). Bureau for Inventions and Discoveries, Prague. (in Czech). KUČERA, J., ČERMÁK, J., PENKA, M. 1977: Improved thermal method of continual recording the transpiration flow rate dynamics. Biologia Plantarum, Vol. 19: 413–420. KUGLITSCH, F. G., REICHSTEIN, M., BEER, C., CARRARA, A., CEULEMANS, R., GRANIER, A., JANSSENS, I. A., KOESTNER, B., LINDROTH, A., LOUSTAU, D., MATTEUCCI, G., MONTAGNANI, L., MOORS, E. J., PAPALE, D., PILEGAARD, K., RAMBAL, S., REBMANN, C., SCHULZE, E. D., SEUFERT, G., VERBEECK, H., VESALA, T., AUBINET, M., BERNHOFER, C., FOKEN, T., GRÜNWALD, T., HEINESCH, B., KUTSCH, W., LAURILA, T., LONGDOZ, B., MIGLIETTA, F., SANZ, M. J., VALENTINI, R. 2008: Characterisation of ecosystem water-use efficiency of European forests from eddy covariance measurements. Biogeosciences Discussion, Vol. 5: 4481–4519. KÜPPERS, M., SCHULZE, E.-D. 1985: An empirical model of net photosynthesis and leaf conductance for the simulation of diurnal courses of and exchange. Australian Journal of Plant Physiology, Vol. 12: 512–526. KUTÍLEK, M. 1978: Vodohospodářská pedologie, 2 nd edition. Státní nakladatelství technické literatury, Praha, Czechoslovakia, 296 pp. (in Czech). LACOINTE, A., DELEENS, E., AMEGLIO, T., SAINT-JOANIS, B., LELARGE, C., VANDAME, M., SONG, G. C., DAUDET, F. A. 2004: Testing the branch autonomy theory: a 13C/14C double labelling experiment on differentially shaded branches. Plant, Cell and Environment, Vol. 27: 1159–1168. LANDSBERG, J. J., WRIGHT, L. L. 1989: Comparison among Populus clones and intensive culture conditions, using an energy-conversion model. Forest Ecology and Management, Vol. 27: 129–147. LARCHER, W. 2003: Physiological Plant Ecology, 4th edition. Springer, Berlin, 513 pp. LASCANO, R. J. 1991: Review of models for predicting soil water balance. Soil water balance in the Sudano-Sahelian, proceedings of the Niamey workshop, zone, February, IAHAS publication no. 199: 443–458.

227

LASCH, P., KOLLAS, C., ROCK, J., SUCKOW F. 2009: Potentials and Impacts of Short Rotation Coppice Plantation with Aspen in Eastern Germany under Climate Change. Regional Environmental Change, DOI 10.1007/s10113-009-0095-7. LAW, B. E., FALGE, E., GU, L., BALDOCCHI, D. D., BAKWIN, P., BERBIGIER, P., DAVIS, K., DOLMAN, A. J., FALK, M., FUENTES, J. D., GOLDSTEIN, A., GRANIER, A., GRELLE, A., HOLLINGER, D., JANSSENS, I. A., JARVIS, P., JENSEN, N. O., KATUL, G., MAHLI, Y., MATTEUCCI, G., MEYERS, T., MONSON, R., MUNGER, W., OECHEL, W., OLSON, R., PILEGAARD, K., PAW, K. T., THORGEIRSSON, H., VALENTINI, R., VERMA, S., VESALA, T., WILSON, K., WOFSY, S. 2002: Environmental controls over carbon dioxide and water vapor exchange of terrestrial vegetation. Agricultural and Forest Meteorology, Vol. 113: 97–120. LEE, X., MASSMAN, W. J., LAW, B. 2004: Handbook of micrometeorology: A guide for surface flux measurement and analysis. Kluwer, Dordrecht, 250 pp. LEENHARDT, D., VOLTZ, M., RAMBAL, S. 1995: A survey of several agroclimatic soil water balance models with the reference to their spatial application. European Journal of Agronomy, Vol. 1: 1–14. LEUNING, R., FOSTER, I. J. 1990: Estimation of transpiration by single trees: comparison of a ventilated chamber, leaf energy budgets and a combination equation. Agricultural and Forest Meteorology, Vol. 51: 63–86. LEUSCHNER, CH., CONERS, H., ICKE, R., HARTMANN, K., EFFINGER, N. D., SCHREIBER, L. 2003: Chemical composition of the periderm in relation to in situ water absorption rates of oak, beech and spruce fine roots. Annals of Forest Science, Vol. 60: 763–772. LEVERENZ, J, DEANS, J. D, FORD, E. D., JARVIS, P. G., MILNE, R., WHITEHEAD, D. 1982: Systematic spatial variation of stomatal conductance in a Sitka spruce plantation. Journal of Applied Ecology, Vol. 19: 835–51. LEVIS, S., COE, M., T., FOLEY, J., A. 1996: Hydrologic budget of a land surface model: A global application. Journal of Geophysical Research, Vol. 101: 16291–16930. LEWANDOWSKI, I., WEGER, J., VAN HOOIJDONK, A., HAVLÍČKOVÁ, K., VAN DAM, J., FAAIJ, A. 2006: The potential biomass for energy production in the Czech Republic. Biomass and Bioenergy, Vol. 30: 405–421. LHOMME, J. P., KATERJI, N. 1991: A simple modelling of crop water balance for agrometeorological applications. Ecological Modelling, Vol. 57: 11–25. LIBERLOO, M., CALFAOIFETRA, C., LUKAC, M., GODBOLD, D., LUO, Z., POLLE, A., HOOSEBEEK, M.R., KULL, O., MAREK, M., RAINES, CH., RUBINO, M., TAYLOR, G., SCARASCIA-MUGNOZZZA, G., CEULEMANS, R. 2006: Woody biomass production during the second rotation of a bio-energy Populus plantation increases in a future high CO2 world. Global Change Biology, Vol. 12: 1094–1106. LINDERSON, M. L., IRITZ, Z., LINDROTH, A. 2007: The effect of water availability on stand-level productivity, transpiration, water use efficiency and radiation use efficiency of field-grown willow clones. Biomass and Bioenergy, Vol. 31: 460–468. LINDROTH, A. 1985: Canopy conductance of coniferous forests related to climate. Water Resources Research, Vol. 21: 297–304. LINDROTH, A. 1985: Seasonal and diurnal variation of energy budget components in coniferous forests. Journal of Hydrology, Vol. 82: 1–15. LINDROTH, A., 1993: Aerodynamic and canopy resistance of short-rotation forest in relation to leaf area index and climate. Boundary-Layer Meteorology, Vol. 66: 265–279. LINDROTH, A., BÅTH, A. 1999: Assessment of regional willow coppice yield in Sweden on basis of water availability. Forest Ecology and Management, Vol. 121: 57–65. LINDROTH, A., ČERMÁK, J., KUČERA, J., CIENCIALA, E., ECKERSTEN, H. 1995: Sap flow by heat balance method applied to small size Salix-trees in a short rotation forest, Biomass and Bioenergy, Vol. 8: 7–15. LINDROTH, A., CIENCIALA, E. 1996: Water use efficiency of short-rotation Salix viminalis at leaf, tree and stand scales. Tree Physiology, Vol. 16: 257–262. LINDROTH, A., HALLDIN, S. 1990: Gradient measurement with fixed and reversing temperature and humidity sensors above thin forest. Agricultural and Forest Meteorology, Vol. 53: 81–103. LINDROTH, A., IRITZ, Z. 1993: Surface energy budget dynamics of short-rotation willow forest. Theoretical and Applied Climatology, Vol. 47: 175–185. LINDROTH, A., VERWIJSTA, T., HALLDIN, S. 1994: Water-use efficiency of willow: Variation with season, humidity and biomass allocation. Journal of hydrology, Vol. 156: 1–19.

228

LIU, H., FOKEN, T. 2001: A modified Bowen ratio method to determine sensible and latent heat fluxes. Meteorologische Zeitschrift, Vol. 10: 71–80. LIVINGSTON, G. F., HUTCHINSON, G. L. 1994: Enclosure-based measurement of trace gas exchange: application and source of error. In: Matson, P. A., Herriss, R. C. (eds.), Methods in Ecology: Biogenic Trace Gas Emissions from Soil and Water, Blackwell Science, London, UK, p. 14–51. LOAGUE, K. 1992: Soil water content at R-5. Part 1. Spatial and temporal variability. Journal of Hydrology, Vol. 139: 233–251. LOHAMMAR, T., LARSSON, S., LINDER, S., FALKS, S. 1980: FAST – Simulation models of gaseous exchange in Scots pine. In: Persson, T. (ed.): Structure and function of northern coniferous forests – an ecosystem study. Ecological Bulletins (Stockholm), No. 32, Swedish Natural Science Research Council, p. 505–523. LU, P., URBAN, L., ZHAO, P. 2004: Granier‘s thermal dissipation probe (tdp) method for measuring sap flow in trees: Theory and practise. Acta Botanica Sinica, Vol. 46: 631–646. LUKANU, G., SAVAGE, M. J. 2006: Calibration of a frequency-domain reflectometer for determining soilwater content in a clay loam soil. Water SA, Vol. 32: 37–42. MAIDMENT, D. R. 1993: Hydrology. In: Maidment, D. R. (ed.): Handbook of Hydrology. McGraw-Hill, New York, p. 1.1–1.15. MAKKINK, G. F. 1957: Testing the Penman formula by means of lysimeters. Journal of the Institution of Water Engineers, Vol. 11: 277–288. MALEK, E. 1993: Comparison of the Bowen ratio-energy balance and stability-corrected aerodynamic methods for measurement of evapotranspiration. Theoretical and Applied Climatology, Vol. 48: 167– 178. MARC, V., ROBINSON, M. 2004: Application of the deuterium tracing method for the estimation of tree sap flow and stand transpiration of beech forest (Fagus silvatica L.) in mountainous Mediterranean region. Journal of Hydrology, Vol. 285: 248–259. MARSHALL, T. J., HOLMES, J. W., ROSE, C. W. 1996: Soil Physics, 3rd edition. Cambridge University Press, Cambridge, UK 453 pp. MASLE, J., FARQUHAR, G. D., WOGN, S. C. 1992: Transpiration ratio and plant mineral content are related among genotypes of a range of species. Australian Journal of Plant Physiology, Vol. 19: 709– 721. MAVI, H., S., TUPPER, G., J. 2004: Agrometeorology - Principles and Applications of Climate Studies in Agriculture. Food Products Press, An Imprint of The Haworth Press, Inc., New York, London, Oxford. 364 pp. MAXIMOV, N. A. 1923: Physiologisch-ökologische Untersuchungen über die Dürreresistenz der Xerophyten. Jahrbücher für wissenschaftliche Botanik, Vol. 62: 128–144. McANENEY, K. J., ITIER, B. 1996: Operational limits to the Priestley–Taylor formula. Irrigation Science, Vol. 17: 37–43. McDERMITT, D. K. 1990: Sources of error in the estimation of stomatal conductance and transpiration from porometer data. HortScience, Vol. 25: 1538–1548. McNABB, N. S., HART, J. H. 1962: Measurements of xylary fluid movement in elm by the thermoelectric method. Proceeding Iowa Academy of Science, Vol. 69: 132–137. McNAUGHTON, K. G. 1976: Evaporation and advection. Quarterly Journal of the Royal Meteorological Society, Vol. 102: 181–191. McNAUGHTON, K. G., BLACK. T. A. 1973: A study of evapotranspiration from a Douglass fir forest using the energy balance approach. Water Resources Research, Vol. 9: 1579–1590. McNAUGHTON, K. G., JARVIS, P. G. 1983: Predicting effects of vegetation changes on transpiration and evaporation. In: Kozlowski, T. T. (ed.), Water Deficits and Plant Growth. Academic Press, New York, p. 1–47. McNAUGHTON, K. G., LAUBACH, J. 1998: Unsteadiness as cause of non-equality of eddy diffusivities for heat and vapour at the base of an advective inversion. Boundary-Layer Meteorology, Vol. 88: 479–504. McNEIL, D. D, SHUTTLEWORTH, W. J. 1975: Comparative measurements of the energy fluxes over a pine forest. Boundary-Layer Meteorology, Vol. 9: 297–313. MEIJNINGER, W., M., L. 2003: Surface fluxes over natural landscapes using scintillometry. Wageningen University. PhD thesis, 164 pp.

229

MEINZER, F., C., BROOKS, J., R., DOMEC, J.-C., GARTNER, B., L., WARREN, J., M., WOODRUFF, D., R., BIBLE, K., SHAW, D., C. 2006: Dynamics of water transport and storage in conifers studied with deuterium and heat tracing techniques. Plant, Cell and Environment, Vol. 29: 105–114. MEIRESONNE, L., NADEZHDINA, N., ČERMÁK, J., VAN SLYCKEN, J., CEULEMANS, R. 1999: Measured sap flow and simulated transpiration from a poplar stand in Flanders (Belgium). Agricultural and Forest Meteorology, Vol. 96: 165–179. MIGLIAVACCA, M., MERONI, M., MANCA, G., MATTEUCCI, G., MONTAGNANI, L., GRASSI, G., ZENONE, T., TEOBALDELLI, M., GODED, I., COLOMBO, R., SEUFERT, G. 2009: Seasonal and interannual patterns of carbon and water fluxes of a poplar plantation under peculiar eco-climatic conditions. Agricultural and Forest Meteorology, Vol. 149: 1460–1476. MILESI, C., RUNNING, S. W., ELVIDGE, C. D., DIETZ, J. B. 2005: Mapping and modeling the biogeochemical cycling of turf grasses in the United States. Environmental Management, Vol. 36: 426– 438. MILNE, R., SATLIN, M., DEANS, J. D., JARVIS, P. G., CANNELL, M. G. R. 1992: The biomass production of three poplar clones in relation to intercepted solar radiation. Forest Ecology and Management, Vol. 55: 1–14. MOENE, A., F. 2011: First steps in scintillometry: from refractive index fluctuations to fluxes. 3 rd Scintillometer Workshop, Wageningen University, 18–19 April, (oral presentation). MOENE, A., F., MEIJNINGER, W., M., L., HARTOGENSIS, O., K., KOHSIEK, W., DE BRUIN, H., A., R. 2005: A review of the relationships describing the signal of a Large Aperture Scintillometer, Internal report 2004/2, Meteorology and Air Quality Group, Wageningen University, 39 pp. MOFFAT, A., M., PAPALE, D., REICHSTEIN, M., HOLLINGER, D., Y., RICHARDSON, A., D., BARR, A., G., BECKSTEIN, C., BRASWELL, B., H., CHURKINA, G., DESAI, A., R., FALGE, E., GOVE, J., H., HEIMANN, M., HUI, D., JARVIS, A., J., KATTGE, J., NOORMETS, A., STAUCH, V., J. 2007: Comprehensive comparison of gap-filling techniques for eddy covariance net carbon fluxes. Agricultural and Forest Meteorology, Vol. 147: 209–232. MÖLDER, M., GRELLE, A., LINDROTH, A., HALLDIN, S. 1999: Flux-profile over boreal forest – roughness sublayer corrections. Agricultural and Forest Meteorology, Vol. 98–99: 645–658. MONCLUS, R., DREYER, E., VILLAR, M., DELMOTTE, F. M., DELAY, D., PETIT, J. M., BARBAROUX, C., THIEC, D. LE, BRÉCHET, C., BRIGNOLAS, F. 2006: Impact of drought on productivity and water-use efficiency in 29 genotypes of Populus deltoides x Populus nigra. New Phytologist, Vol. 169: 765–777. MONTEITH, J. 1965: Evaporation and environment. 19 th Symposia of the Society for Experimental Biology, University Press, Cambridge, Vol. 19: 205–234. MONTEITH, J. L. 1973: Principles of environmental physics. American Elsevier Publishing Company, New York, USA, 241 pp. MONTEITH, J. L. 1981. Evaporation and surface temperature. Quarterly Journal of the Royal Meteorological Society, Vol. 107: 1–27. MONTEITH, J. L. 1985: Evaporation from Land Surfaces: Progress in Analysis and Prediction since 1948. In: Advances in Evapotranspiration, Proceedings of the ASAE Conference on Evapotranspiration, Chicago, Ill. ASAE, St. Joseph, Michigan, USA, p. 4–12. MONTEITH, J. L., UNSWORTH, M. H. 2008: Principles of environmental physics, 3 rd edition. Academic Press, London, 440 pp. MOORE, C. J. 1986: Frequency response corrections for eddy correlation systems. Boundary-Layer Meteorology, Vol. 37: 17–35. MORAT, P., LE MOUEL, J. L., GRANIER, A. 1994: Electrical potential on tree. A measurement of the sap flow? Comptes rendus de l'Académie des sciences. Série 3, Sciences de la vie/Life science, Vol. 317: 98–101. MOREL-SEYTOUX, H. J., KHANJI, J. 1975: Equation of infiltration with compression and counterflow effects. Hydrological Sciences Bulletin, Vol. 20: 505–517. MORIKAWA, Y. 1972: The heat pulse method and apparatus for measuring sap flow in woody plants. Journal of Japanese Forest Society, Vol. 54:166–171. (in Japanese). MORRIS, J., MANN, L., COLLOPY, J. 1998: Transpiration and canopy conductance in a eucalypt plantation using shallow saline groundwater. Tree Physiology, Vol. 18: 547–555.

230

MUKAMMAL, E. B., NEUMAN, H., H. 1977: Application of the Priestley-Taylor evaporation model to assess the influence of soil moisture on the evaporation from a large weighting lysimeter and class A pan. Boundary-Layer Meteorology, Vol. 14: 243–256. MUNRO, D. S. 1989: Stomatal conductances and surface conductance modelling in a mixed wetland forest. Agricultural and Forest Meteorology, Vol. 48: 235–249. MUNRO, D. S., OKE, T. R. 1975: Aerodynamic boundary-layer adjustment over a crop in neutral stability. Boundary-Layer Meteorology, Vol. 9: 53–61. MWALE, S. S., AZAM-ALI, S. N., SPARKES, D. L. 2005: Can the PR1 capacitance probe replace the neutron probe for routine soil-water measurement? Soil Use and Management, Vol. 21: 340–347. NADEZHDINA, N. 1999: Sap flow index as an indicator of plant water status. Tree Physiology, Vol. 19: 885–891. NADEZHDINA, N., ČERMÁK, J. 1998: The technique and instrumentation for estimation the sap flow rate in plants (in Czech). Patent No. 286438 (PV-1587-98). Bureau for Inventions and Discoveries, Prague. NADEZHDINA, N., ČERMÁK, J., CEULEMANS, R. 2002: Radial pattern of sap flow in woody stems of dominant and understory species: scaling errors associated with positioning of sensors. Tree Physiology, Vol. 22: 907–918. NADEZHDINA, N., ČERMÁK, J., MEREISONNE, L., CEULEMANS, R. 2007: Transpiration of Scots pine in Flanders growing on soil with irregular substratum. Forest Ecology and Management, Vol. 243: 1–9. NADEZHDINA, N., NADEZHDIN, V., FERREIRA, M., I., PITACCO, A. 2007: Variability with xylem depth in sap flow in trunks and branches of mature olive trees. Tree Physiology, Vol. 27: 105–113. NOBEL, P., S., 2009: Plant Physiology, 4 th edition. Elsevier, Oxford, UK, 582 pp. OBUKHOV, A. M. 1946: Turbulentnost´v temperaturnoj – neodnorodnoj atmosphere (Turbulence in an atmosphere with a non-uniform temperature. Trudy Instituta Teoreticheskio Geofiziki AN SSSR, Vol. 1: 95–115. (in Russian). OFFENTHALER, I., HIETZ, P., RICHTER, H. 2001: Wood diameter indicates diurnal and long-term patterns of xylem water potential in Norway spruce. Trees, Vol. 15: 215–221. OHMURA, A. 1982: Objective criteria for rejecting data for Bowen ratio flux calculations. Journal of applied meteorology, Vol. 21: 595–598. ORSÁG, M., FISCHER, M., TRNKA, M., HLAVINKA, P., KUČERA, J., BALEK, J., ŢALUD, Z. 2011: Water balance in short rotation poplar coppice and reference grass-plot in conditions of CzechMoravian Highlands. In: Šír, M., Tesař, M. (eds.): Hydrologie malého povodí 2011, Proceedings of the International Conference, March 8–10, Praha, p. 363–370. (in Czech). ORTEGA-FARIAS, S., O., CUENCA, R., H., EK, M. 1996: Daytime variation of sensible heat flux estimated by the bulk aerodynamic method over a grass canopy. Agricultural and Forest Meteorology, Vol. 81: 131–143. OVINGTON, J. D., 1957: Dry-matter production by Pinus sylvestris L. Annals of Botany, Vol. 21: 287–314. PANOFSKY, H., A., TOWNSEND, A., A. 1964: Change of terrain roughness and the wind profile. Quarterly Journal of the Royal Meteorological Society, Vol. 90: 147–155. PARKIN, G. W. 2008: Water budget in soil. In: Chesworth, W. (ed.): Encyclopedia of soil science (Encyclopedia of Earth Sciences Series). Springer, Dordrecht, Netherlands, p. 811–813. PARKINSON K. J., LEGG, B. J. 1972: A continuous flow porometer. Journal of Applied Ecology, Vol. 9: 669–675. PARLANGE, H. L., KATUL, G. G. 1992: Estimation of the diurnal variation of potential evaporation from a wet bare-soil surface. Journal of Hydrology, Vol. 132: 71–89. PAULIUKONIS, N., SCHNEIDER, R. 2001: Temporal patterns in evapotranspiration from lysimeters with three common wetland plant species in the eastern United States. Aquatic Botany, Vol. 71: 35–46. PAYERO, J., O., NEALE, C., M. U., WRIGHT, J., L., ALLEN, R., G. 2003: Guidelines for validating Bowen ratio data. American Society of Agriculture Engineers, Vol. 44: 1051–1060. PAYNE, W. A., BRUCK, H. 1996: Evapotranspiration and other soil water balance terms estimated from neutron and capacitance probe measurements. In: Camp, C. R., Sadler, E. J., Yoder, R. E. (eds.): Evapotranspiration and Irrigation Scheduling. Proceedings of the International Conference, November 3–6, San Antonio, TX, p. 455–461. PEACOCK, C. E., HESS, T. M. 2004: Estimating evapotranspiration from a reed bed using the Bowen ratio energy balance method. Hydrological Processes, Vol. 18: 247–260.

231

PEARCY, R. W., SCHULZE, E.-D., ZIMMERMANN, R. 1989: Measurement of transpiration and leaf conductance. In: Plant physiological ecology. Pearcy, R. W., Ehleringer, J. R., Mooney, H. A., Rundel, P. W. (eds.): London: Chapman and Hall, p. 137–160. PELLIS, A., LAUREYSENS, I., CUELEMANS, R. 2004: Growth and production of a short rotation coppice culture of poplar I. Clonal differences in leaf characteristics in relation to biomass production. Biomass and Bioenergy, Vol. 27: 9–19. PENMAN, H. L. 1948: Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society (series A), Vol. 193: 120–145. PENMAN, H. L. 1953: The physical bases of irrigation control. In: Synge, P. M. (ed.): Report of the Thirteenth International Horticultural Congress, Vol. II., The Royal Horticultural Society, London, p. 913–924. PENMAN, H. L. 1956: Evaporation–an introductory survey. Netherlands Journal of Agricultural Science, Vol. 4: 9–29. PEREZ, P. J., CASTELLVI, F., IBAÑEZ, M., ROSELL, J. I. 1999: Assessment of reliability of Bowen ratio method for partitioning fluxes. Agricultural and Forest Meteorology, Vol. 97: 141–150. PEREZ, P. J., CASTELLVI, F., MARTÍNEZ-COB, A. 2008: A simple model for estimating the Bowen ratio from climatic factors for determining latent and sensible heat flux. Agricultural and Forest Meteorology, Vol. 148: 25–37.¨ PERÉZ-PRIEGO, O., TESTI, L., ORGAZ, F., VILLALOBOS, F. J. 2010: A large closed canopy chamber for measuring CO2 and water vapour exchange of whole trees. Environmental and Experimental Botany, Vol. 68: 131–138. PERSSON, G. 1995: Willow stand evapotranspiration simulated for Swedish soils. Agricultural water management, Vol. 28: 271–293. PERSSON, G. 1997: Comparison of simulated water balance for willow, spruce, grass ley and barley. Nordic Hydrology, Vol. 28: 85–98. PERSSON, G., JANSSON, P.-E. 1989: Simulated water balance of a willow stand on clay soil. In: Perttu, K. L., Kowalik, P. J. (Eds.): Energy Forestry Modelling - Growth, Water Relations and Economics. Simulation Monographs. 30. Pudoc, Wageningen, p. 147–162. PERSSON, G., LINDROTH, A. 1994: Simulating evaporation from short-rotation forest: variations within and between seasons. Journal of Hydrology, Vol. 156: 21–45. PETERSON, R. L., FARQUHAR, M. L. 1996: Root hairs: Specialized tubular cells extending root surfaces. Botanical Review, Vol. 62: 1–40. PETZOLD, R., SCHWÄRZEL, K., FEGER, K. H. 2010: Transpiration of a hybrid poplar plantation in Saxony (Germany) in response to climate and soil conditions. European Journal of Forest Research, Vol. 130: 695–706. PEUKE, A. D., ROKITTA, M., ZIMMERMANN, U., SCHREIBER, L., HAASE, A. 2001: Simultaneous measurement of water flow velocity and solute transport in xylem and phloem of adult plants of Ricinus communis over a daily time course by nuclear magnetic resonance spectrometry. Plant, Cell and Environment, Vol. 24: 491–503. PISTOCCHI, C., WERTHER, G., PICCIONI, E., BONARI, E. 2009: Water requirements of poplar and willow vegetation filters grown in lysimeter under Mediterranean conditions: Results of the second rotation. Desalination, Vol. 246: 137–146. PITACCO, A., GALLINARO, N., GIULIVO, C. 1992: Evaluation of actual evapotranspiration of a Quercus ilex L. stand by the Bowen Ratio–Energy Budget method. Plant Ecology, Vol. 99/100: 163–168. POLSTER, H. 1950: Die Physiologischen Grundlagen der Stofferzeugung im Walde. Bayericher Landwirtchaftsverlag Gmbh, Munchen, 96 pp. PONCE, V. M., HAWKINS, R. H. 1996: Runoff curve number: Has it reached maturity? Journal of Hydrologic Engineering, Vol. 1: 11–19. PORPORATO, A., D‘ODORICO, P., LAIO, F., RODRIGUEZ-ITURBE, I. 2003: Soil moisture controls on the nitrogen cycle I. Modeling scheme. Advances in Water Research, Vol. 26: 45–58. PRIESTLEY, C. H. B., TAYLOR, R. J. 1972: On the assessment of surface heat flux and evaporation using large-scale parameters. Monthly Weather Review, Vol. 100:81–92. QIAN, Y. L., FRY, J. D., UPHAM, W. S. 1997: Rooting and drought avoidance of warm-season turfgrasses and tall fescue in Kansas. Crop Science, Vol. 37: 905–910. RAABE, A. 1983: On the relation between the drag coefficient and fetch above the sea in case of off- shore wind in the near shore zone. Zeitschrift für Meteorologie, Vol. 33: 363–367.

232

RABINOVICH, S. 2005. Measurement errors: Theory and practice, 2 nd edition. Springer, New York, 308 pp. RANA, G., KATERJI, N. 2000: Measurement and estimation of actual evapotranspiration in the field under Mediterranean climate: a review. European Journal of Agronomy, Vol. 13: 125–153. RANA, G., KATERJI, N., MASTRORILLI, M., EL MOUJABBER, M. 1994: Evapotranspiration and canopy resistance of grass in a Mediterranean region. Theoretical and Applied Climatology, Vol. 50: 61–71. RANA, G., KATERJI, N., MASTRORILLI, M., EL MOUJABBER, M. 1997: A model for predicting actual evapotranspiration under water stress conditions in a Mediterranean region. Theoretical and Applied Climatology, Vol. 56: 45–55. RASMUSSEN, W. O. 2008: Field water cycle. In: Chesworth, W. (ed.): Encyclopedia of soil science (Encyclopedia of Earth Sciences Series). Springer, Dordrecht, Netherlands, p. 272–275. RAUPACH, M. R. 1995: Vegetation-atmosphere interaction and surface conductance at leaf, canopy and regional scale. Agricultural Forest Meteorology, Vol. 73: 151–179. RAUPACH, M. R., LEGG, B. J. 1984: The uses and limitations of flux-gradient relationships in micrometeorology. Agricultural Water Management, Vol. 8: 119–131. REBETEZ, M., MAYER, H., DUPONT, O., SCHINDLER, D., GARTNER, K., KROPP, J. P., MENZEL, A. 2006: Heat and drought 2003 in Europe: a climate synthesis. Annals of Forest Science, Vol. 63: 569– 577. REEDER, J. D. 1986: A non-weighing lysimeter design for field studies using nitrogen-15. Soil Science Society of America Journal, Vol. 50: 1224–1227. REICHSTEIN, M., TENHUNEN, J. D., OURCIVAL, J. M., ROUPSARD, O., RAMBAL, S., MEGLIETTA, F., PERESSOTTI, A., PECCHIARI, M., TIRONE, G., VALENTINI, R. 2002: Severe drought effects on ecosystem CO2 and H20 fluxes at three Mediterranean evergreen sites: revision of current hypothesis? Global Change Biology, Vol. 8: 999–1017. REICOSKY, D. C., PETERS, D. B. 1977: A portable chamber for rapid evapotranspiration measurements on field plots. Agronomy Journal, Vol. 69: 729–732. REICOSKY, D. C., SHARRATT, B. S., LJUNGKULL, J. E., BAKER, D. G. 1983: Comparison of alfalfa evapotranspiration measured by a weighing lysimeter and a portable chamber. Agricultural Meteorology, Vol. 28: 205–211. REIGOSA ROGER, M. J., SÁNCHEZ-MOREIRAS, A. M. 2003: Determination of transpiration using steady state-porometer. In: REIGOSA ROGER, M. J. (ed.): Handbook of Plant Ecophysiology Techniques, Springer, Netherlands, p. 223–233. REYNOLDS, H. L., D‘ANTONIO, C. 1996: The ecological significance of plasticity in root weight ratio in response to nitrogen: opinion. Plant and Soil, Vol. 185: 75–97. RICHTER, H. 1997: Water relations of plants in the field: some comments on the measurement of selected parameters. Journal of Experimental Botany, Vol. 48: 1–7. RIPULLONE, F., LAUTERI, M., GRASSI, G., AMATO, M., BORGHETTI, M. 2004: Variations in nitrogen supply changes wateruse efficiency of Pseudotsuga menziesii and Populus x euroamericana; a comparison of three approaches to determine water-use efficiency. Tree Physiology, Vol. 24: 671–679. RIZOU, M. 2008: Evaluation of climatic and ecohydrological effects on longwave radiation and evapotranspiration. PhD thesis, Department of Civil, University of Central Florida, Orlando, Florida, 191 pp. ROBINSON, D. A., CAMPBELL, C. S., HOPMANS, J. W., HORNBUCKLE, B. K., JONES, S. B., KNIGHT, R., OGDEN, F., SELKER, J., WENDROTH, O. 2008: Soil moisture measurement for ecological and hydrological watershed-scale observatories: A review. Vadose Zone Journal, Vol. 7: 358–389. RODDA, J. C., DROWNING, R. A., LAW, F. M. 1976: Systematic Hydrology. Newnes-Butterworth, London, 399 pp. RÖDL, A. 2008: Ökobilanzierung der Holzerzeugung im Kurzumtrieb. Arbeitsbericht des Instituts für Ökonomie der Forst- und Holzwirtschat des vTI, Hamburg. RODRIGUEZ-ITURBE, I., PORPORATO, A., LAIO, F., RIDOLFI, L. 2001: Plants in water-controlled ecosystems: active role in hydrological processes and response to water stress. I Scope and general outline. Advances in Water Research, Vol. 24: 695–705. ROELFSEMA, M. R. G., HEIDRICH, R. 2005: In the light of stomatal opening: new insights into ‗the Watergate‘. New Phytologist, Vol. 167: 665–691.

233

ROHWER, C. 1931: Evaporation from free water surfaces. USDA Technical Bulletin No. 271, United States Department of Agriculture, Washington, D.C., 96 pp. ROKITTA, M., PEUKE, A. D., ZIMMERMANN, U., HAASE, A. 1999: Dynamic studies of phloem and xylem flow in fully differentiated plants by fast nuclear-magnetic-resonance microimaging. Protoplasma, Vol. 209: 126–131. ROSE, C. W., SHARMA, M. L. 1984: Summary and recommendations of the workshop on evapotranspiration of plant communities. Agricultural Water Management, Vol. 8: 325–342. ROSENBERG, N. J., BLAD, B. L., VERMA, S. N. 1983: Microclimate—the biological environment, 2nd edition. John Wiley & Sons, New York, 495 pp. ROTH-NEBELSICK, A., UHL, D., MOSBRUGGER, V., KERP, H. 2001: Evolution and function of leaf architecture: a review. Annals of Botany, Vol. 87: 533–566. SAKAMOTO, K., BRIGGS, W. R. 2002. Cellular and subcellular localization of phototropin 1. The Plant Cell, Vol. 14: 1723–1735. SAKURTANI, T. 1981: A heat balance method for measuring water flux in the stem of intact plants. Journal of Agricultural Meteorology, Vol. 37: 9–17. SAMAIN, B., SIMONS, G. W. H., VOOGT, M. P., DEFLOOR, W., BINK, N.-J. , PAUWELS, V. R. N. 2011: Consistency between hydrological model, large aperture scintillometer and remote sensing based evapotranspiration estimates for a heterogeneous catchment. Hydrology and Earth System Sciences, Vol. 8: 10863–10894. SAMUELSON, L., J., STOKES, T., A., COLEMAN, M., D. 2007: Influence of irrigation and fertilization on transpiration and hydraulic properties of Populus deltoides. Tree Physiology, Vol. 27: 765–774. SAVAGE, M. J. 2010: Field evaluation of polymer capacitive humidity sensors for Bowen ratio energy balance flux measurements. Sensors, Vol. 10: 7748–7771. SAVAGE, M. J., EVERSON, C. S., METELERKAMP, B. R. 2009: Bowen ratio evaporation measurement in a remote montane grassland: Data integrity and fluxes. Journal of Hydrology, Vol. 376: 249–260. SAVELYEV, S. A., TAYLOR, P. A. 2001: Notes on internal-boundary layer height formula. BoundaryLayer Meteorology, Vol. 101: 293–301. SAVENIJE, H. H. G. 1997: Determination of evaporation from a catchment water balance at a monthly time scale. Hydrology and Earth System Sciences, Vol. 1: 93–100. SCHARLEMANN, J. P. W., LAURANCE, W. F. 2008: How green are biofuels? Science, Vol. 4: 43–44. SCHEENEN, T. W. J., VERGELDT, F. J., HEEMSKERK, A. M., VAN AS, H. 2007: Intact plant Magnetic Resonance Imaging to study dynamics in long-distance sap flow and flow-conducting surface area. Plant Physiology, Vol. 144: 1157–1165. SCHLESINGER, W. H. 1997: Biogeochemistry: An analysis of global change, 2nd edition. Academic Press, 588 pp. SCHOLANDER, P. F., HAMMEL, H. T., BRADSTREET, E. D., HEMMINGTON, E. A. 1965: Sap pressure in vascular plants. Science, Vol. 148: 339– 346. SCHUEPP, P., H., LECLERC, M., Y., MACPHERSON, J., I., DESJARDINS, R., L. 1990: Footprint prediction of scalar fluxes from analytical solutions of the diffusion equation. Boundary-Layer Meteorology, Vol. 50: 355–373. SCHULZE, E. D., KELLIHER, F. M., KÖRNER, C., LLOYD, J., LEUNING, R. 1994: Relationships among maximum stomatal conductance, ecosystem surface conductance, carbon assimilation rate, and plant nitrogen nutrition: a global ecology scaling exercise. Annual Review of Ecology and Systematics, Vol. 25: 629–660. SCHULZE, E.-D., HALL, A. E., LANGE, O. L., WALZ, H. 1982: A portable steady-state porometer for measuring the carbon dioxide and water vapor exchanges of leaves under natural conditions. Oecologia, Vol. 53: 141–145. SCHWENDENMANN, L., DIERICK, D., KÖHLER, M., HÖLCHER, D. 2010: Can deuterium tracing be used for reliably estimating water use of tropical trees and bamboo? Tree Physiology, Vol. 30: 886–900. SEVRUK, B. 1996: Adjustment of tipping-bucket precipitation gauge measurements. Atmospheric Research, Vol. 42: 237–246. SEVRUK, B., HAMON, W. R. 1984: International comparison of national precipitation gauges with a reference pit gauge. World Meteorological Organisation, Geneva, WMO instruments observing methods, report No. 17, 111 pp. SEVRUK, B., HETIG, J. A, SPIESS, R. 1991: The effect of precipitation gauge orifice rim on the wind field deformation as investigated in a wind tunnel. Atmospheric Environment, Vol. 25: 1173–1179.

234

SEVRUK, B., KLEMM, S. 1989: Types of standard precipitation gauges. In: Sevruk, B. (ed.): Proceeding of WMO/IAHS/ETH International Workshop on Precipitation Measurement, December 3–7, St. Moritz, Switzerland, p. 227–232. SHERIFF, D. W. 1972: A new apparatus for the measurement of sap flux in small shoots with the magnetohydrodynamic method. Journal of Experimental Botany, Vol. 23: 1086–1095. SHERIFF, D. W. 1974: A model of plant hydraulics under non-equilibrium conditions I. Stems. Journal of Experimental Botany, Vol. 25: 552–561. SHIR, C. C. 1972: A numerical computation of the air flow over a sudden change of surface roughness. Journal of the Atmospheric Sciences, Vol. 29: 304–310. SHULZE, E. D., ČERMÁK, J., MATYSSEK, R., PENKA, M., ZIMMERMAN, R., VAŠIČEK, F., GRIES, W., KUČERA, J. 1985: Canopy transpiration and water fluxes in the xylem of the trunk of Larix and Picea trees—A comparison of xylem flow, porometer and cuvette measurements. Oecologica, Vol. 66: 475–483. SHUTTLEWORTH, W. J. 1993: Evaporation. In: Maidment, D. R. (ed.): Handbook of Hydrology. McGrawHill, New York, p. 4.1–4.53. SHUTTLEWORTH, W. J. 2007: Putting the ‗vap‘ into evaporation. Hydrology and Earth System Sciences, Vol. 11: 210–244. SHUTTLEWORTH, W. J., CALDER, I. R. 1979: Has the Priestley-Taylor equation any relevance to forest evaporation? Journal of Applied Meteorology, Vol. 18: 639–646. SHUTTLEWORTH, W. J., WALLACE, J. S. 1985: Evaporation from sparse crops—an energy combination theory. Quarterly Journal of the Royal Meteorological Society, Vol. 111: 839–855. SINCLAIR, T. R., ALLEN, L. H., LEMON, E. R. 1975: An analysis of errors in the calculation of energy flux densities above vegetation by a Bowen ratio profile method. Boundary-Layer Meteorology, Vol. 8: 129–139. SLATYER, R. O. 1967: Plant-water Relationships. London, Academic Press, 366 pp. SLATYER, R. O., McILROY, I. C. 1961: Practical microclimatology, Commonwealth Scientific and Industrial Research Organisation, Melbourne, Victoria, Australia, 300 pp. SMITH, D., M., ALLEN, S., J. 1996: Measurement of sap flow in plant stems. Journal of Experimental Botany, Vol. 47, 1833–1844. SOMMER, R., DE ABREU SÁ, T. D., VIELHAUER, K., DE ARAÚJO, A. C., FÖLSTER, H., VLEK, P. L. G. 2002: Transpiration and canopy conductance of secondary vegetation in the eastern Amazon. Agricultural and Forest Meteorology, Vol. 112: 103–121. SOUCH, C. A., STEPHENS, W. 1998: Growth, productivity and water use in three hybrid poplar clones. Tree Physiology, Vol. 18: 829–835. SPITTLEHOUSE, D. A. BLACK, T. A. 1980: Evaluation of the Bowen ratio/energy balance method for determining forest evapotranspiration. Atmosphere-Oceans, Vol. 18: 98–116. STANHILL, G. 1969: A simple instrument for the field measurement of turbulent diffusion flux. Journal of Applied Meteorology, Vol. 8: 509–513. STANNARD, D. I. 1997: A theoretically based determination of Bowen-ratio fetchs requirements. Boundary-Layer Meteorology, Vol. 83: 375–406. STEDUTO, P., CALIANDRO, A., RUBINO, P., BEN MECHLIA, N., MASMOUDI, M., MARTINEZCOB, A., JOSE FACI, M., RANA, G., MASTRORILLI, M., EL MOURID, M., KARROU, M., KANBER, R., KIRDA, C., EL-QUOSY, D., EL-ASKARI, K., AIT ALI, M., ZAREB, D., SNYDER, R. L. 1996. Penman-Monteith reference evapotranspiration estimates in the Mediterranean region. In: Camp, C. R., Sadler, E. J., Yoder, R. E. (eds.): Evapotranspiration and Irrigation Scheduling. Proceedings of the International Conference, November 3–6, San Antonio, TX, p. 357–364. STEDUTO, P., HSIAO, T. C. 1998: Maize canopies under two soil water regimes IV. Validity of Bowen ratio-energy balance technique for measuring water vapour and carbon dioxide fluxes at 5-min intervals. Agricultural and Forest Meteorology, Vol. 89: 215–228. STEDUTO, P., KATERJI, N., PUERTOS-MOLINA, H., UNLU, M., MASTRORILLI, M., RANA, G. 1997: Water-use efficiency of sweet sorghum under water stress conditions, gas-exchange investigations at leaf and canopy scale. Field Crops Research, Vol. 54: 221–234. STEDUTO, P., TODOROVIC, M., CALIANDRO, A., RUBINO, P. 2003: Daily reference evapotranspiration estimates by the Penman-Monteith equation in southern Italy. Costant vs. variable canopy resistance. Theoretical and Applied Climatology, Vol. 74: 217–225.

235

STEPHENS, W., HESS, T., KNOX, J. 2001: Review of the effects of energy crops on hydrology. Institute of Water and Environment, Cranfield University, Bedford, 59 pp. STEPPE, K., DE PAUW, D. J. W., LEMEUR, R., VANROLLEGHEM, P. A. 2006: A mathematical model linking tree sap flow dynamics to daily stem diameter fluctuations and radial stem growth. Tree Physiology, Vol. 26: 257–273. STEWART, J. B. 1988: Modelling surface conductance of pine forest. Agricultural and Forest Meteorology, Vol. 43: 19–35. STEWART, R. B., ROUSE, W. R. 1976: A simple method for determining the evaporation from shallow lakes and ponds. Water Resources Research, Vol. 12: 623–628. STEWART, R. B., ROUSE, W. R. 1977: Substantiation of the Priestley-Taylor parameter α = 1.26 for potential evaporation in high latitude. Journal of Applied Meteorology, Vol. 16: 649–650. STONE, J. F., SHIRAZI, G. A. 1975: On the heat-pulse method for the measurement of apparent sap velocity in stems. Planta, Vol. 122: 169–177. STRANGEWAYS, I. 2003: Measuring the natural environment, 2 nd edition. Cambridge University Press, United kingdom, Cambridge, 534 pp. STRUGGER, S., WEBER, F. 1926: Zur physiologie der stomata-nebenzellen. Berichte der Deutschen Botanischen Gesellschaft, Vol. 44: 272–278. STULL, R. B. 1988: An introduction to boundary layer meteorology. Springer, London, 666 pp. SUMNER, D. M., JACOBS, J. M. 2005: Utility of Penman–Monteith, Priestley–Taylor, reference evapotranspiration, and pan evaporation methods to estimate pasture evapotranspiration. Journal of Hydrology, Vol. 308: 81–104. SZEICZ, G., LONG, I. F. 1969: Surface resistance of crop canopies. Water Resources Research, Vol. 5: 622–633. TAIZ, L., ZEIGER, E. 2002: Plant Physiology, 3 rd edition. Sunderland, Mass: Sinauer Associates, Inc., 690 pp. TANNER, B. D., GREENE, J. P., BINGHAM, G. E. 1987: A Bowen-ratio design for long term measurements, ASAE Paper No. 87-2503. American Society Agricultural Engineers. St. Joseph, MI, USA, 6 pp. TANNER, C. B., SINCLAIR, T. R. 1983: Efficient water use in crop production: Research or re-search? In: Taylor, H. M., Jordan, W. R., Sinclair, T. R. (eds.): Limitations of efficient water use in crop production. American Society of Agronomy, Madison, WI, p. 1–27. TANNER, W., BEEVERS, H. 2001: Transpiration, a prerequisite for long-distance transport of minerals in plants? Proceedings of the National Academy of Sciences of the United States of America, Vol. 98: 9443–9447. TATARSKII, V. I. 1961: Wave propagation in a turbulent medium, translated from the Russian by R. S. Silverman, McGraw-Hill, New York, USA, 285 pp. TAYLOR, J. R. 1997: An introduction to error analysis — The study of uncertainties in physical measurements, 2nd edition. University Science Books, Sausalito, California, USA, 328 pp. TESKEY, R. O., HINCKLEY, T. M. 1981: Influence of temperature and water potential on root growth of white oak. Physiologia Plantarum, Vol. 52: 363–369. TEULING, A., J., SENEVIRATNE, S., I., STÖCKLI, R., REICHSTEIN, M., MOORS, E., CIAIS, P., LUYSSAERT, S., VAN DEN HURK, B., AMMANN, CH., BERNHOFER, CH., DELLWIK, E., GIANELLE, D., GIELEN, B., GRÜNWALD, T., KLUMPP, K., MONTAGNANI, L., MOUREAUX, CH., SOTTOCORNOLA, M., WOHLFAHRT, G. 2010: Contrasting response of European forest and grassland energy exchange to heatwaves. Nature Geoscience, Vol. 3: 722–727. THIERMANN, V., GRASSL, H. 1992: The measurement of turbulent surface-layer fluxes by use of bichromatic scintillation. Boundary-Layer Meteorology, Vol. 58: 367–389. THOM, A. S. 1968: The exchange of mo+mentum, mass and heat between an artificial leaf and the airflow in wind-tunnel. Quarterly Journal of the Royal Meteorological Society, Vol. 94: 44–55 THOM, A. S. 1972: Momentum, mass and heat exchange of vegetation. Quarterly Journal of the Royal Meteorological Society, Vol. 98: 124–134. THOM, A. S., OLIVIER, H. R. 1977: On Penman's equation for estimating regional evaporation. Quarterly Journal of the Royal Meteorological Society, Vol. 103: 345–357. THOM, A., S., STEWART, J., B., OLIVER, H., R., GASH, J., H., C. 1975: Comparison of aerodynamic and energy budget estimates of fluxes over a pine forest. Quarterly Journal of the Royal Meteorological Society, Vol. 101: 93–105.

236

THORTHWAITE, V. W. 1948: An approach toward a rational classification of climate. Geographical Review, Vol. 38: 55–94. TILLMAN, J. E. 1972: The indirect determination of stability, heat and momentum fluxes in the atmospheric boundary layer from simple scalar variable during dry unstable conditions. Journal of Applied Meteorology, Vol. 11: 783–792. TODD, D. K. 1980: Groundwater Hydrology. New York, Wiley, 539 pp. TODOROVIC, M. 1999: Single-layer evapotranspiration model with variable canopy resistance. Journal of Irrigation and Drainage Engineering, ASCE, Vol. 125: 235–245. TOPP, G. C., FERRÉ, P. A. 2002: Thermogravimetric method using convective oven-drying. In: Dane, J. H., Topp, G.C. (eds): Methods of Soil Analysis: Part 4. Physical methods, SSSA, Madison, WI, p. 422– 424. TRENBERTH, K. E., SMITH, L., QIAN, T., DAI, A., FASULLO, J. 2006: Estimates of the global water budget and its annual cycle using observational and model data. Journal of Hydrometeorology, Vol. 8: 758–769. TRICKER, P. J., PECCHIARI, M., BUNN, S. M., VACCARI, F. P., PERESSOTTI, A., MIGLIETTA, F., TAYLOR, G. 2009: Water use of a bioenergy plantation increases in a future high CO 2 world. Biomass and Bioenergy, Vol. 33: 200–208. TRNKA, M., TRNKA, M., FIALOVÁ, J., KOUTECKÝ, V., FAJMAN, M., ŢALUD, Z., HEJDUK, S. 2008: Biomass production and survival rates of selected poplar clones grown under a short-rotation on arable land. Plant, Soil and Environment, Vol. 54: 78–88. TURC, L. 1961: Estimation of irrigation water requirements, potential evapotranspiration: a simple climatic formula evolved up to date. Annals of Agronomy, Vol. 12: 13–49. TYREE, M. T., ZIMMERMANN, M. H. 2002: Xylem Structure and the Ascent of Sap, 2 nd edition. Springer, Berlin, 293 pp. UEDA, M., SHIBATA, E. 2004: Why do trees decline or dieback after a strong wind? Water status of Hinoki cypress standing after a typhoon. Tree Physiology, Vol. 24: 701–706. UNESCO 1978: World water balance resources of the Earth. United Nations Educational, Scientific and Cultural Organization, Paris, France, 633 pp. UNLAND, H. E., HOUSER, P. R., SHUTTLEWORTH, W. J., YANG, Z. L. 1996: Surface flux measurement and modelling at a semi-arid Sonoran Desert site. Agricultural and Forest Meteorology, Vol. 82: 119–153. UTSET, A., FARRE, I., MARTINEZ-COB, A., CAVERO, J. 2004: Comparing Penman–Monteith and Priestley–Taylor approaches as reference-evapotranspiration inputs for modeling maize water-use under Mediterranean conditions. Agricultural Water Management, Vol. 66: 205–219. VAN DE WALLE, I., VAN CAMP, N., VAN DE CASTEELE, L., VERHEYEN, K., LEMEUR, R. 2007: Short-rotation forestry of birch, maple, poplar and willow in Flanders (Belgium). Biomass and Bioenergy, Vol. 31: 267–275. VERMA, S. B., ROSENBERG, N. J., BLAD, B. L. 1978: Turbulent exchange coefficients for sensible heat and water vapor under advective conditions. Journal of Applied Meteorology, Vol. 17: 330–338. VERSTRAETEN, W. W., MUYS, B., FEYEN, J., VEROUSTRAETE, F., MINNAERT, M., MEIRESONNE, L., SCHRIJVER, A. D. 2005: Comparative analysis of the actual evapotranspiration of Flemish forest and cropland, using the soil water balance model WAVE, Hydrology and Earth System Sciences, Vol. 9: 225–241. VERTESSY, R., A., BENYON, R., G., O‘SULLIVAN, S., K., O., GRIBBEN, P., R. 1995: Relationships between stem diameter, sapwood area, leaf area and transpiration in a young mountain ash forest. Tree Physiology, Vol. 15, 559–567. VIEWEG, G. H., ZIEGLER, H. 1960: Thermoelektrische Registrierung der Geschwindigkeit des Transpirationsstromes I. Deutsche Botanical Geselschaft Berlin, Vol. 73: 221–226. VOGT, R., JAEGER, L. 1990: Evaporation from a pine forest - using the aerodynamic method and Bowen ratio method. Agricultural and Forest Meteorology, Vol. 50: 39–54. WAGNER, S. W., REICOSKY, D. C. 1996: Mobile research gas exchange machine-MRGEM instrumentation update. In: Camp, C. R., Sadler, E. J., Yoder, R. E. (eds.): Evapotranspiration and Irrigation Scheduling. Proceedings of the International Conference, November 3–6, San Antonio, TX, p. 781–786. WALLIHAN E. F. 1964: Modification and use of an electric hygrometer for estimating relative stomatal apertures. Plant Physiology, Vol. 39: 86–90.

237

WALMSLEY, J. L. 1989: Internal boundary-layer height formulae — A comparison with atmospheric data. Boundary-Layer Meteorology, Vol. 47: 251–262. WANG, J., BRAS, R. L. 1998: A new method for estimation of sensible heat flux from air temperature. Water Resources Research, Vol. 34: 2281–2288. WARING, R. H., ROBERTS, J. M. 1979: Estimating water flux through stems of Scots pine with tritiated water and phosphorus-32. Journal of Experimental Botany, Vol. 30: 459–471. WEBB, E. K. 1970: Profile relationships: the log-linear range, and extension to strong stability. Quarterly Journal of the Royal Meteorological Society, Vol. 96: 67–90. WEBB, E., K., PEARMAN, G., I., LEUNING, R. 1980: Correction of flux measurements for density effects due to heat and water vapour transfer. Quarterly Journal of the Royal Meteorological Society, Vol. 106: 85–100. WEGER, J. et al. 2009: Rámcová typologie zemědělských půd pro pěstování vybraných klonů topolů a vrb k energetickému vyuţití v České republice. In: Kubačka, J. (ed.): Rychle rostoucí dřeviny zdroj biomasy pro energetiku. Hradec Králové: Česká lesnická společnost, 21 st May 2009, p. 10–15. (in Czech). WESELY, M. L. 1976: The combined effect of temperature and humidity fluctuations on refractive index. Journal of Applied Meteorology, Vol. 15: 43–49. WESSON, K. H., KATUL, G., LAI, C.-T. 2001: Sensible heat flux estimation by flux variance and halforder time derivative methods. Water Resources Research, Vol. 37: 2333–2343. WEVER, L. A., FLANAGAN, L. B., CARLSON, P. J. 2002: Seasonal and interannual variation in evapotranspiration, energy balance and surface conductance in a northern temperate grassland. Agricultural and Forest Meteorology, Vol. 112: 31–49. WHALLEY, W. R., COPE, R. E., NICHOLL, C. J., WHITMORE, A. P. 2004: In-field calibration of a dielectric soil moisture meter designed for use in an access tube. Soil Use and Management, Vol. 20: 203–206. WILSON, K., B., BALDOCCHI, D., D., AUBINET, M., BERBIGIER, P., BERNHOFER, CH., DOLMAN, H., FALGE, E., FIELD, CH., GOLDSTEIN, A., GRANIER, A., GRELLE, A., HALLDOR, T., HOLLINGER, D., KATUL, G., LAW, B., E., LINDROTH, A., MEYERS, T., MONCRIEFF, J., MONSON, R., OECHEL, W., TENHUNEN, J., VALENTINI, R., VERMA, S., VESALA, T., WOFSY, S. 2002: Energy partitioning between latent and sensible heat flux during the warm season at FLUXNET sites. Water resources research, Vol. 38: 1–11. WINDT, C., W. 2007: Nuclear Magnetic Resonance imaging of sap flow in plants, PhD thesis. Wageningen University, Netherlands, 172 pp. WOLF, A., SALIENDRA, N., AKSHALOV, K., JOHNSON, D. A., LACA, E. 2008: Effects of different eddy covariance correction schemes on energy balance closure and comparisons with the modified Bowen ratio system. Agricultural Forest Meteorology, Vol. 148: 942–952. WOODRUFF, D. R., BOND, B. J., MEINZER, F. C. 2004: Does turgor limit growth in tall trees? Plant, Cell and Environment, Vol. 27: 229–236. WOODWARD, F. I., SHEEHY, J. E. 1983: Principles and measurements in environmental biology. Butterworths, London, 276 pp. WULLSCHLEGER, S. D., MEINZER, F. C., VERTESSY, R. A. 1998: A review of whole-plant water use studies in trees. Tree Physiology, Vol. 182: 499–512. WYNGAARD, J. C., IZUMI, Y., STUART, A., COLLINS, J. R. 1971: Behaviour of the refractive-index structure parameter near the ground. Journal of Optical Society of America, Vol. 61: 1646–1650. WYNNE, R. T. 2008: Precipitation: Forms. In: Trimble, S. W. (ed.): Encyclopedia of water science, 2nd edition. Taylor and Francis Group, New York, 1370 pp. XU, C., SINGH, V. 2005: Evaluation of three complementary relationship evapotranspiration models by water balance approach to estimate actual regional evapotranspiration in different climatic regions. Journal of Hydrology, Vol. 308: 105–121. YAO, T., WIERENGA, P. J., GRAHAM, A. R., NEUMAN, S. P. 2004: Neutron Probe Calibration in a Vertically Stratified Vadose Zone. Soil Science of America, Vol. 3: 1400–1406. ZALESNY Jr., R., S., WIESE, A., H., BAUER, E., O., RIEMENSCHNEIDER, D., E. 2006: Sapflow of hybrid poplar (Populus nigra L. x P. maximowiczii A. Henry ‗NM6‘) during phytoremediation of landfill leachate. Biomass and Bioenergy, Vol. 30: 784–793. ZAVITKOVSKI, J. 1983: Projected and actual biomass production of 2- to 10-year-old intensively cultured Populus 'Tristis #1'. In: Hansen, E. A. (ed.): Intensive Plantation Culture: 12 Years Research. General Technical Report NC-91. USDA Forest Service North Central Forest Experiment Station, St. Paul, MN, p. 72–76.

238

ZHANG, H., MORISON, J., I., L., SIMMONDS, L., P. 1999: Transpiration and water relations of poplar trees growing close to the water table. Tree Physiology, Vol. 19: 563–573. ZHANG, J., DAVIES, W. J. 1989: Abscisic acid produced in dehydrating roots may enable the plant to measure the water status of the soil. Plant, Cell and Environment, Vol. 12: 73–81. ZHANG, J., SCHURR, U., DAVIES, W. J. 1987: Control of stomatal behaviour by abscisic acid which apparently originates in roots. Journal of Experimental Botany, Vol. 38: 1174–1181. ZHANK, H., MORISON, J., I., L., SIMMONDS, L. 1999: Transpiration and water relations of poplar trees growing close to the water table. Tree Physiology, Vol. 19: 563–573. ZHANK, H., SIMMONDS, L., P., MORISON, J., I., L., PAYNE, D. 1997: Estimation of transpiration by single trees: comparison of sap flow measurements with a combination equation. Agricultural and Forest Meteorology, Vol. 87: 155–169. ZHANK, L., WALKER, G. R., DAWES, W. R. 2002: Water balance modelling: concepts and applications. In: McVicar, T. R. Rui, L., Walker, J., Fitzpatrick, R. W., Changming, L. (eds.): Regional water and soil assessment for managing sustainable agriculture in Cina and Australia, ACIAR monograph No. 84, 31– 47. ZONA, D., JANSSENS, I., A., VERLINDEN, M., S., BROECKX, L., S., COOLS, J., GIOLI, B., ZALDEI, A., CELULEMANS, R. 2011: Impact of extreme precipitation and water table change on N 2O fluxes in a bio-energy poplar plantation. Biogeosciences Discussion, Vol. 8: 2057–2092. ZUR, B., JONES, J. W. 1984: Diurnal changes in the instantaneous water-use efficiency of soybean crop. Agriculture and Forest Meteorology, Vol. 33: 41–51. ZWEIFEL, R., ZIMMERMANN, L., ZEUGIN, F., NEWBERY, D. M. 2006: Intra-annual radial growth and water relations of trees – implications towards a growth mechanism. Journal of Experimental Botany, Vol. 57: 1445–1459.

239

APPENDIXIES APPENDIX A: List of abbreviations and symbols used in the text Abbreviations AB ABA ANOVA BREB DMC EC EC-10 EU FDR LAS MOST PR1 SAS SRC STP TDR UNESCO WMO XLAS

aboveground biomass abscisic acid analysis of variance Bowen ratio energy balance dry matter content eddy covariance capacitance soil moisture sensor European Union frequency domain reflectometry large aperture scintillometer Monin-Obukhov similarity theory soil moisture profile probe small aperture scintillometer short rotation coppice standard condition for temperature and pressure time domain reflectometry United Nations Educational, Scientific and Cultural Organization World Meteorological Organisation extra large aperture scintillometer

Symbols of variables or constants Greek alphabet Priestley-Taylor coefficient Bowen ratio psychrometric constant modified psychrometric constant assuming non-saturated surface sensor or variable error thickness of the boundary layer created above the certain surface saturation vapour pressure curve change in soil water storage in the defined soil layer duration of the ith time increment of deuterium concentration temperature increase of sap ratio of molecular weights of water and dry air variable value of Bowen ratio around -1 which has to be rejected permittivity of vacuum relative permittivity (dielectric constant) the ratio of molecular weights of water and dry air surface emissivity for emission of long wave radiation ´ surface emissivity for absorption of long wave radiation coefficient of the heat losses from the measuring point (sap flow) 240

latent heat of vaporisation of water theoretical adiabatic latent heat flux latent heat flux from open water Priestley-Taylor potential evaporation latent heat flux air density dry air density mass of the substance in air, partial density Stephan-Boltzman constant shearing stress dimensionless stability universal function for sensible heat flux dimensionless stability universal function for latent heat flux dimensionless stability universal function for momentum exchange dimensionless stability universal function for any entity log-amplitude fluctuation in a spherical wave along the scintillometer path decoupling factor (coefficient) Roman alphabet albedo empirically derived parameter function of the wavelength and air humidity function of the wavelength and air temperature and atmospheric pressure cross-sectional area of sapwood empirically derived parameter bark area index deuterium concentration of the transpired water concentration in the ith time increment of deuterium concentration structure parameter of refractive index specific heat capacity of air structure parameter of air humidity specific heat of water structure parameter of temperature covariant term of structure parameters of temperature and humidity curve number zero plane displacement drainage molecular diffusivity aperture diameter of receiver aperture diameter of transmitter diameter at breast height is the shoot diameter at 20 cm height water vapour pressure 241

equilibrium evaporation open water evaporation Priestley-Tylor potential evaporation saturated vapour pressure saturation vapour pressure at the evaporating water surface vapour pressure measured at the height saturation vapour pressure at level above ground evapotranspiration actual evapotranspiration crop evapotranspiration reference evapotranspiration function (general) measure of adjustment into new equilibrium layer first Fresnel zone flux of any entity gravitational acceleration soil heat flux canopy conductance stomatal conductance gross ecosystem exchange gross primary production mean canopy height sensible heat flux von Karman constant wave number of the electromagnetic radiation turbulent transfer coefficient (eddy diffusivity) crop coefficient actual crop coefficient turbulent transfer coefficient for sensible heat flux turbulent transfer coefficient for latent heat flux turbulent transfer coefficient for momentum turbulent transfer coefficient for any entity soil thermal conductivity mixing length path length between transmitter and receiver in scintillometry Obukhov length leaf area index leaf are index of the leaves active (effective) in transpiration latent heat flux total mass of the injected tracer mean bias error dimensionless frequency refractive index of air 242

net ecosystem production net primary production observed values p-value is a measure of how much evidence is against the null hypothesis atmospheric pressure energy used for photosynthesis heat input power (sap flow) precipitation predicted values plant area index photosynthetic active radiation potential evapotranspiration specific humidity sap flow rate absolute humidity scale heat dissipated by heating of the sap flowing through the heated region ―fictitious flow‖ surface (subsurface) run-off (lateral inflow) Pearson correlation coefficient coefficient of determination adjusted coefficient of determination atmospheric (long wave) radiation aerodynamic resistance canopy resistance global radiation net radiation solar (short wave) radiation surface resistance specific gas constant for water vapour root area index Reynolds number critical Reynolds number relative humidity gradient Richardson number root mean square error mixing ratio heat storage by vegetation standard deviation specific leaf area last time increment in which the deuterium was present absolute temperature temperature scales surface temperature 243

surface temperature air temperature measured at level first time increment of deuterium concentration of the transpired water last time increment of deuterium concentration of the transpired water fluctuation of the horizontal wind velocity component horizontal wind velocity component, wind speed at level z friction velocity vapour pressure deficit vertical wind velocity component fluctuation of the vertical wind velocity component capillary rise water-use efficiency instantaneous water-use efficiency (intrinsic) long term water-use efficiency water-use efficiency of production water-use efficiency of photosynthesis fetch distance from the point of measurement distance, height above earth´s surface effective height height of the air temperature and humidity measurement height of the lower sensors (intakes) height of the wind speed measurement height of the upper sensors (intakes) roughness length (parameter, height) apparent roughness for heat and vapour transfer roughness length for momentum transfer

244

APPENDIX B: List of tables Table 1: The overview of the selected climate characteristics of the experimental site. Table 2: More detailed overview of the selected weather characteristics at the experimental site during the three main years investigated in this study. abbreviates standard deviation (calculated from the daily values) and the reference evapotranspiration according to ALLEN et al. (1998). Table 3: The overview of the selected weather characteristics at the experimental site during the three main growing seasons (1st April to 30th September) investigated in this study. (calculated from the daily values) abbreviates standard deviation and the reference evapotranspiration according to ALLEN et al. (1998). Table 4: The overview of the selected soil characteristics of the experimental site. Table 5: The overview of the main hydro-pedological parameters calculated as the mean from three soil sampling (2007, 2008 and 2009) in the poplar plantation and the turf grass. Table 6: The list of particular azimuths (0–345°) and with them related fetches (the upwind distance from the measuring point to the edge of area of interest) for two contrasting covers (poplar plantation and turf grass). Table 7: Illustration of the percentage number of wind directions in particular months during the whole course of measurement. Again, only the values with wind speed at 2 m above the reference grass higher than 1 m s-1 were took into account ensuring the right eddy diffusivity conditions with sufficient friction velocity. Moreover, only the values when the global radiation was higher than 20W m-2 in order to filter the values which are the most important for evapotranspiration were used. From white through green to blue, the number of wind blowing from the particular direction is increasing. Each column of numbers belongs to the class covering 15°of radius between particular azimuths. Table 8: Results from ANOVA post-hoc Fischer s LSD test showing the interactions between particular groups and the statistical significance of their relation expressed as the p-value. Only the interactions between the groups A, B and C within the particular months during the three years were within the scope of the analysis and therefore, only the interactions within each box are taken into account. The statistical significance of particular relations is marked with red colour and red colour and bold font for the alpha levels 0.95 and 0.99 respectively. In group A (the worst fetch conditions), less than 70 % of the total measured flux is theoretically coming from the area of interest (the poplar plantation) according to the Gash s model. Group B include 70–80 % and group C (the best fetch conditions) 80–90 % of flux from the area of interest. Table 9: The same ANOVA post-hoc analysis as in the previous tab. 3, this time made for the turf grass. Due to different upwind distances from the measuring point to the edge of the cover, the groups A, B and C have different parameters. Namely, A (the worst) includes less than 50 % of the measured flux coming from the area of interest (turf grass). Group B has 50–75 % and C (the best) more than 75 % of the fluxes theoretically coming only from the turf grass. Again, only the interactions inside the particular boxes are taken into account – namely the interactions between A, B and C just only within one month. The statistical significance of particular relations is marked with red colour and red colour and bold font for the alpha levels 0.95 and 0.99 respectively. 245

Table 10: The days with the precipitation events and the specific amounts of the fallen rain water during three months periods from 2008 to 2010. The first column always depicts the days when any precipitation were measured (note that the days are not consecutive) and the second column the amount of water. The daily precipitation totals are marked by different colours in order to enable easily follow their intensities (from white through light blue to azure – the higher amounts). In addition, each month is for clarity bordered by the dashed line. At the bottom of table the days and also the amounts of precipitation within each three months long period are summed. Table 11: Results from ANOVA post-hoc Fischer s LSD test showing only the interactions and their statistical significance (p-values) between of contrasting covers just within particular months belonging into particular years. The statistical significance of the individual relations is marked with red colour and red colour and bold font for the alpha levels 0.95 and 0.99 respectively.

246

APPENDIX C: List of figures Figure 1:

Wind speed profiles and simplified eddy structures characteristic of three basic stability states in air flow near the ground (from THOM, 1975).

Figure 2:

The geographical position of the research locality Domanínek within the Czech Republic.

Figure 3:

Dynamics of mean monthly temperatures (°C) and monthly precipitation sums (mm) during the experimental period (2002–2010) in comparison with the 1961–1990 period (visualized as a solid line). The thresholds of World Meteorological Organization for evaluation of the meteorological conditions were used (KOŽNAROVÁ and KLABZUBA, 2002).

Figure 4:

The aerial image showing the investigated poplar plantation with measuring point BREB 1 and the adjacent turf grass with the measuring point BREB 2. In the upper left corner, the plantation with clonal experiment described in TRNKA et al. (2008) is situated.

Figure 5:

The scatter view on the relationship between soil moisture measured by capacitance EC-10 sensor (horizontal axis) and the thermo-gravimetrically based soil moisture (vertical axis) which serves usually as the reference method. The EC-10 sensors were placed below two contrasting covers (the poplar plantation and the turf grass), in both cases in three depths (0.1, 0.3 and 0.9 m). The resulting 3rd order polynomial function is used as the calibration equation for further analysis. The four points depicted in red colour were judged as the outliers and thus they were not considered for establishing the calibration relationship. The dashed blue line depicts the linear relationship, which were not, however, used as the calibration curve.

Figure 6:

The four scatter views (four different depths from 0.1 to 0.4 m) on the relationships between soil moisture measured via portable PR1 profile probe and the thermo-gravimetrically calibrated EC-10 sensor. 5 % of data were judged as outliers which were removed in order to obtain more robust relationship.

Figure 7:

Zoomed top view on the spatial soil moisture variability from the 1st of July 2010 in the 0.2 m depth. The particular access tubes are depicted as white spots. Their numbers describe the usual sequence of measuring routine from 1 to 16. The black dashed line (slightly transparent to not cover the particular white point – access tubes) depicts the imaginary rectangular (defined by particular access tubes 1, 4, 13 and 16) inside which the interpolation are made by weighted averages between particular access tubes and might be considered as correct. However, outside this area, there are only extrapolations based on the interpolations between the access tubes and they have increasing or decreasing trends with unrealistic tendency and meaning.

Figure 8:

The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the late summer drying period (from 9th of July to 6th of September 2008). Each column belongs to one depth and each line to the particular date. The small white points depict the 16 access tubes for measuring with PR1.

Figure 9:

The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the spring drying period (from 19th of March to 13th of May 2009). Each column belongs to one depth and each line to the particular date. The small white points depict the 16 access tubes for measuring with PR1.

247

Figure 10: The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the spring drying period terminated by several modest rain events and later with strong summer rainfalls which filled back the soil moisture to the level of the field capacity (from 20th of May to 22nd of July 2009). Each column belongs to one depth and each line to the particular date. Figure 11: The top views on the soil moisture variability in four different depths (0.1–0.4 m) during the end of spring and summer 2010 (from 4th of June to 10th of August 2010). There is pronounced drying period initiated after refilling to the filed capacity at the 15th of June. Each column belongs to one depth and each line to the particular date. Figure 12: The temporal variation of soil moisture under the poplar plantation culture during the years 2008 to 2011 measured in four different depths (0.1–0.4 m) by PR1 profile probe. Figure 13: The temporal variation of soil moisture below turf grass cover during the years 2008 to 2011 measured in four different depths (0.1–0.4 m) by PR1 profile probe. Figure 14: Seasonal course of mean available soil water content (here simply the water content above the wilting point) under to contrasting covers (the poplar plantation and the turf grass) measured by PR1profile probe in the integrated soil profile 0–0.45 m. Figure 15: Cumulative water depletion of two contrasting covers (poplar plantation and turf grass) during the years 2008–2011 measured from the upper soil layer (0–0.45 m) by PR1 profile probe. Note, that the soil moisture was measured usually once per week and therefore the values cannot be considered as an absolute water loss. On the other hand, the picture can provide good image about which cover loose more water. Figure 16: The seasonal course of soil moisture in poplar plantation measured by three calibrated EC-10 permanently buried into three different soil depths (0.1, 0.3 an 0.9 m) during the year 2009. The winter values are not depicted because of very high inaccuracy of measurement during low soil temperature. Figure 17: The seasonal course of soil moisture in poplar plantation measured by three calibrated EC-10 permanently buried into three different soil depths (0.1, 0.3 an 0.9 m) during the year 2010. The winter values are not depicted because of very high inaccuracy of measurement during low soil temperature. Figure 18: Comparison of precipitation measured by rain-gauge (blue columns) and the water loss (red columns) calculated as the total diurnal depletions of soil moisture measured by EC-10 in three depths (0.1, 0.3 and 0.9 m) integrated for the soil profile 0–1 m. The graph depicts the water balance conditions in the poplar plantation during 2009. Figure 19: Comparison of precipitation measured by rain-gauge (blue columns) and the water loss (red columns) calculated as the total diurnal depletions of soil moisture measured by EC-10 in three depths (0.1, 0.3 and 0.9 m) integrated for the soil profile 0–1 m. The graph depicts the water balance conditions in the poplar plantation during 2009. Figure 20: The comparison of yearly totals of precipitation measured by rain-gauge, total soil water income and total soil water loss measured by EC-10 in three depths (0.1, 0.3 and 0.9 m) integrated for the soil profile 0–1 m.

248

Figure 21: Frequency distribution of air temperature differences above different covers. The vertical distance between the sensors above the turf grass was 1.6 m. In case of the adult poplar plantation in 2008 and 2009 the distance was adjusted to 2 m and during the season 2010 with resprouting poplars the distance was 3 m. Figure 22: Frequency distribution of air humidity differences above different covers during the growing seasons 2008–2010. The vertical distance between the sensors was the same as for the air temperature at Fig. 16. Figure 23: Example of data rejection. Inside the red cross are depicted values which does not fulfil the conditions given by GUO et al (2007) for pairs of sensors with the mutual error below 0.3 °C and 50 Pa for air temperature and humidity respectively. By using more precise sensors the cross would tend to be closer. The narrower dashed black cross depicts the situation for standard sensors of Cambell Scientific Bowen-ratio system with features described e.g. by TANNER et al. (1987). Conversely, if someone uses more inaccurate sensors, the area inside the cross would be larger. Figure 24: Example of data rejection. Suspicious and excluded red curves and spikes depict the values which do not fulfil the two conditions mentioned above for pairs of sensors with the mutual error below 0.3 °C and 50 Pa for temperature and humidity respectively. Figure 25: Six sample days with latent heat flux calculated from the original input data (thick red line) and from data with air temperature measured by the upper sensor gradually artificially shifted down by 0.05 °C from the original value. The shift in the temperature of one sensor simulates the potential error in final latent heat flux caused by imprecise temperature sensors. Note that data of Bowen ratio within the range from -1.5 to -0.5 were excluded which cause the gaps in the final latent heat flux. Figure 26: Six sample days with latent heat flux calculated from the original input data (thick red line) and from data with air humidity measured by the upper sensor gradually artificially shifted down by 0.1 % up to 1 % and then by 0.5 % up to 2 % from the original value. The shift in the humidity of one sensor simulates the potential error in final latent heat flux caused by imprecise humidity sensors. Note that data of Bowen ratio within the range from -1.5 to -0.5 were excluded which cause the gaps in the final latent heat flux. Figure 27: The scatter view on model relation between calculated latent heat flux with original input data (horizontal axis) and those with changed input air temperature data (vertical axis). The input data comprise the complete dataset from August 2008. The input air temperature measured by the upper sensor was gradually shifted down from the original value (from 0.05 to 0.60 °C) to simulate the potential error in final latent heat flux caused by imprecise temperature sensors. Figure 28: The analogy to Fig. 27 – scatter views on model relation between calculated latent heat flux with original input data (horizontal axis) and those with changed input air humidity data (vertical axis). The input air humidity measured by the upper sensor was gradually artificially shifted down from the original value (from 0.1 % to 2 %) to simulate the potential error in final latent heat flux caused by imprecise humidity sensors. Figure 29: Scatter views on the relationship between measured latent heat flux (horizontal axis) and the gap filled one (vertical axis). At the beginning, there were made 249

more than 20 % of artificial random gaps. These gaps were subsequently filled by using FluxCorrector using different model approaches. The graphs with the same colour depict the same method used to provide reference level to which the linear regression between actual latent heat flux was made within the gap filling process (PM is abbreviation Penman-Monteith). The three levels of the pictures express three different covers, namely turf grass, poplar plantation and winter rape respectively. Figure 30: Wind rose depicting the percentage number (0–10 %) of wind directions. This graph depicts only the values from April to September (most important period for evapotranspiration) when the wind speed at 2 m above reference grass was higher than 1 m s-1 (necessary condition for application of theory of turbulent diffusion and thus micrometeorological methods – almost all of the diurnal cases). Figure 31: The scatter views on relationship between the fetch (m) and the ratio of latent heat flux and net radiation for particular months in case of poplar plantation. This should demonstrate whether is there or not any influence on flux magnitude caused by wind flowing from different directions which are characteristic by different fetches (see the Tab. 6). Figure 32: The scatter views on relationship between the fetch (m) and the ratio of latent heat flux and net radiation for particular months in case of turf grass cover. This should demonstrate whether is there or not any influence on flux magnitude caused by wind flowing from different directions which are characteristic by different fetches (see the Tab. 6). Figure 33: Relative flux contribution for upwind distance away from the measuring point (0–200 m) according to Gash s footprint model (GASH, 1986). Two lines express two different situations for two aerodynamically different covers. Figure 34: Cumulative flux contribution for upwind distance away from the measuring point (0–1000 m) according to Gash s footprint model (GASH, 1986). Two lines express two different situations for two aerodynamically different covers. The horizontal axis depicts the upwind distance in logarithmic scale in order to show the mathematically given trend in far upwind. Figure 35: Selected periods with diurnal courses (half-hourly) of of poplar plantation (brown) and the turf grass (green) during 2008. Particular vertical levels of the picture represent ten days periods which are not always consecutive. Figure 36: Selected periods with diurnal courses (half-hourly) of of poplar plantation (brown) and the turf grass (green) during 2009. Particular vertical levels of the picture represent ten days periods which are not consecutive. Figure 37: Selected periods with diurnal courses (half-hourly) of of poplar plantation (brown) and the turf grass (green) during 2010. Particular vertical levels of the picture represent ten days periods which are not consecutive. Figure 38: The comparison of the of the poplar plantation and the turf grass. On left side, the time course of daily sums of two contrasting (the error bars depict the absolute errors). In the middle, the same daily sums of of poplars divided by of the turf grass to show their relative differences (dashed red line is equal to one which means relative difference). On the right hand side, the regression between two contrasting daily sums of . Red solid line depicts the regression line and the dashed black line linear slope of 250

45° (sometimes is completely hidden behind the red regression line). These three types of graphs describe the situation during one month in each of three vertical levels (July 2008–September 2008). Figure 39: The comparison of the of the poplar plantation and the turf grass. On left side, the time course of daily sums of two contrasting (the error bars depict the absolute errors). In the middle, the same daily sums of of poplars divided by of the turf grass to show their relative differences (dashed red line is equal to one which means relative difference). On the right hand side, the regression between two contrasting daily sums of . Red solid line depicts the regression line and the dashed black line linear slope of 45° (sometimes is almost hidden behind the red regression line). These three types of graphs describe the situation during one month in each of three vertical levels (April 2009–June 2009). Figure 40: The comparison of the of the poplar plantation and the turf grass. On left side, the time course of daily sums of two contrasting (the error bars depict the absolute errors). In the middle, the same daily sums of of poplars divided by of the turf grass to show their relative differences (dashed red line is equal to one which means relative difference). On the right hand side, the regression between two contrasting daily sums of . Red solid line depicts the regression line and the dashed black line linear slope of 45° (sometimes is completely hidden behind the red regression line). These three types of graphs describe the situation during one month in each of three vertical levels (July 2009–September 2009). Figure 41: The comparison of the of the poplar plantation and the turf grass. On left side, the time course of daily sums of two contrasting (the error bars depict the absolute errors). In the middle, the same daily sums of of poplars divided by of the turf grass to show their relative differences (dashed red line is equal to one which means relative difference). On the right hand side, the regression between two contrasting daily sums of . Red solid line depicts the regression line and the dashed black line linear slope of 45° . These three types of graphs describe the situation during one month in each of three vertical levels (April 2010–June 2010). Figure 42: The comparison of the of the poplar plantation and the turf grass. On left side, the time course of daily sums of two contrasting (the error bars depict the absolute errors). In the middle, the same daily sums of of poplars divided by of the turf grass to show their relative differences (dashed red line is equal to one which means relative difference). On the right hand side, the regression between two contrasting daily sums of . Red solid line depicts the regression line and the dashed black line linear slope of 45°. These three types of graphs describe the situation during one month in each of three vertical levels (July 2010–September 2010). Figure 43: The graphical output of factorial ANOVA where daily measurements were included as replicates of the dependent variable and the type of cover (poplar plantation and turf grass by brown and green line respectively) and particular months were treated as the independent variables. The error bars indicate the 95 % confidence intervals. Figure 44: Seasonal variation of monthly sums of of the two contrasting covers (the poplar plantation and the turf grass by brown and green colours respectively).

251

The error bars result from the absolute random error of the measurements and their combination and propagation into the monthly sums. Figure 45: Seasonal variation of monthly sums of precipitation (dark blue) and the calculated standard open water potential evaporation (light blue) according to PENMAN (1948). Figure 46: Seasonal course of Bowen ratio diurnal (Rn > 50 W m-2) median in 2008 for the poplar plantation (brown points) and the turf grass (green points). Figure 47: Seasonal course of Bowen ratio diurnal (Rn > 50 W m-2) median in 2009 for the poplar plantation (brown points) and the turf grass (green points). Figure 48: Seasonal course of Bowen ratio diurnal (Rn > 50 W m-2) median in 2010 for the poplar plantation (brown points) and the turf grass (green points). Figure 49: Seasonal course of surface resistance diurnal (Rn> 50 W m-2) median in 2008 for the poplar plantation (brown points) and the turf grass (green points). Figure 50: Seasonal course of surface resistance diurnal (Rn> 50 W m-2) median in 2009 for the poplar plantation (brown points) and the turf grass (green points). Figure 51: Seasonal course of surface resistance diurnal (Rn> 50 W m-2) median in 2010 for the poplar plantation (brown points) and the turf grass (green points). Figure 52: Seasonal course of decoupling coefficient diurnal (Rn> 50 W m-2) median in 2008 for the poplar plantation (brown points) and the turf grass (green points). Figure 53: Seasonal course of decoupling coefficient diurnal (Rn> 50 W m-2) median in 2009 for the poplar plantation (brown points) and the turf grass (green points). Figure 54: Seasonal course of decoupling coefficient diurnal (Rn> 50 W m-2) median in 2010 for the poplar plantation (brown points) and the turf grass (green points). Figure 55: The relationship between the daily sums of the reference evapotranspiration calculated according to FAO-56 methodology (ALLEN et al., 1998) and the actual evapotranspiration measured by BREB above the turf grass. The whole dataset from June 2008 to the end of 2010 was used including the winter time values. Figure 56: The time course of daily in 2009 and 2010. Different colours indicate the differences between the days with and days after precipitation events. In 2010, the coppice was in the first year of the first rotation and in the first half of season, there was only bare soil with occasional weeds. The beginning of resprouting of new shots is marked by the dashed blue vertical line. Figure 57: Time course of ten-day averaged in 2009 and 2010 determined by leaving out the days with and first days after precipitation events. Figure 58: The curve (solid green line) constructed for particular crop stages according to ALLEN et al. (1998). To construct the curve, only the values at least from 2nd day after rain and only the values when was higher than 2 mm were used. It provides more accurate values of and suppresses the undesirable effects of interception, dew and fog. The brown dashed line expresses the in case that the soil surface would be dry during the initial stage which is characteristic with dominant soil evaporation. In crop development stage, transpiration becomes leading and soil evaporation negligible. Figure 59: The comparison of measured cumulative way.

, calculated 252

and

, all expressed in

Figure 60: The frequency distribution of breast height diameters of trees in the stand expressed in the absolute numbers per ha. The diameter range of trees measured by tissue heat balance method is marked in the dashed red quadrangle. Figure 61: Selected periods during 2008 with diurnal courses (half-hourly) of sap flow of three sampled poplar trees belonging into one class (10–11 mm). Note that the given were measured at end of the growing season. Figure 62: Selected periods with diurnal courses (half-hourly) of sap flow of four sampled poplar trees with slightly different dimensions ( measured at the end of the growing season) during 2009. Note that the vertical levels of the picture represent ten days time periods which are not consecutive. The tree depicted by red colour is the only tree measured also in the previous year 2008 (the red one in Fig. 61). Figure 63: a) on the left side the leaf mass (fresh weight) related to the of 30 destructively sampled trees. b) on the right side, the active leaf mass (fresh weight) again related to the of the 30 sampled trees. The solid red curves depict the regression lines and the dashed black lines define the 95 % intervals of confidence. Figure 64: Temporal variation of the measured by BREB (black dashed line) and the transpiration upscaled from sap flow measurements (blue solid line) with the belt of confidence. Because the measurements of sap flow started on the 18th of June, the previous period was estimated (dashed blue line) by using the linear relation between the upscaled transpiration and the magnitude of the leaf area index. Figure 65: Temporal variation of during two contrasting years 2009 and 2010. The green line depicts the mean and the dashed error bars express the standard deviation. Figure 66: The comparison of the measured from litter collection (horizontal axis) and measured by SunScan. The solid dashed line depicts the linear regression line and the dashed black line depicts the slope of 45°. Figure 67: a) allometrically defined relation between and the AB without leaves (kg of DMC) depicted on left side. b) on the right hand, the relation between and – the reference increment in during the period from 3rd Jul to 22nd July 2009). The solid red curves depict the regression lines and the dashed black lines define the 95 % intervals of confidence. Figure 68: Aboveground biomass (without leaves) yields (DMC) during three consecutive years. Figure 69: The long term based on the sap flow measurements of three sampled trees and their simultaneous measurements of increment and thus allometrically defined aboveground woody biomass increments (kg of DMC). The sums of sap flow and the biomass increment were usually integrated in the weekly time-step (see the particular points), and the curves are smoothed by cubic spline fit. The colours of the lines agree with those from the sap flow picture Fig. 62. Figure 70: The seasonal patterns of gross as the ratio of and allometrically defined aboveground woody biomass increments (kg of DMC) compared with the differences in totals of precipitation and actual evapotranspiration. All 253

values are integrated usually in weekly time-step, except the beginning of 2008, and the curves are smoothed by cubic spline fit. Figure 71: The relationships between and aboveground woody biomass increments (kg of DMC) during two consecutive seasons 2008 and 2009. The black filled circles depict the periods with precipitation amounts higher than 10 mm, in contrast the empty circles represent the periods with precipitation amounts lower than 10mm. All values are integrated usually in weekly time-step, except the beginning of 2008 which results in higher values during this period. Figure 72: The comparison of measured aboveground woody biomass increment (kg of DMC) with simple and multiple linear regression models based on water lost and aboveground biomass increment relationship. All values are integrated usually in weekly time-step, except the beginning of 2008 which results in higher values during this period. Figure 73: The seasonal patterns comparing the measured aboveground woody biomass increment (kg of DMC) with simple and multiple linear regression models based on water loss and aboveground biomass increment relationship. All values are integrated usually in weekly time-step (except the beginning of 2008) which is depicted by diamond points at the grey line expressing measured values. All curves are smoothed by cubic spline fit. Figure 74: Cumulative comparison of measured (grey solid line) aboveground woody biomass increment (kg of DMC) with simple and multiple linear regression models (solid green and dashed black lines respectively) based on water lost and aboveground biomass increment relationship. Figure 75: Cumulative root distribution (cumulative percentage) as a function of soil depth for two contrasting terrestrial biomes and for the theoretical model of GALE and GRIGAL (1987). The curve of each biom is the least square fit of parameter β from Gale Grigal s equation Y = 1- βd, where Y is the cumulative root fraction with depth (a proportion between 0–1), d is the soil depth (in m), and β is the fitted parameter with value 0.943 and 0.966 for temperate grassland and deciduous forest respectively according to JACKSON et al. (1996). Figure 76: Example of the theoretical dependency of uncertainty in measuring of the temperature on number of repetitions, assuming sensor with random error ±0.3 °C. Figure 77: The relationship between latent heat fluxes measured from different heights above the poplar plantation. Namely, the upper pair of gradient measurement with the upper arm at 4 m and the lower at 1 m, and the lower gradient measurement with the upper arm at 2 m and the lower arm at 1 m above the ground (mean height during the growing season). The points include all measured data during the growing season 2010 after all quality controls described previously. Figure 78: The relationship between latent heat fluxes measured by BREB above the turf grass and those calculated from Penman-Monteith equation with dynamic surface resistance. model proposed by TODOROVIC (1999). The points include all measured halfhourly data during the growing seasons 2008–2010. Figure 79: Scatter views on the relationship between latent heat flux calculated by the original Penman (1948) formula and those measured by BREB systems – above the poplar plantation on the left and above the turf grass on the right 254

hand side. The picture includes the diurnal (Rn > 50 W m-2) half hourly data within the growing seasons 2009–2010. The inclined solid line depicts the boundary between exceeding the Penman evaporation. The dashed vertical line differentiates the values exceeding the Penman evaporation into two groups: > 200 W m-2 and < 200 W m-2. Figure 80: The surface conductance of the poplar plantation and its dependency on global radiation and vapour pressure deficit as a result of parameterization of Lohammar equation during precipitation-free days in July and August 2008. Figure81: The surface conductance of the turf grass and its dependency on global radiation and vapour pressure deficit as a result of parameterization of Lohammar equation during precipitation-free days in July and August 2008. Figure 82: Overview of monthly totals of the basic water balance variables above four different covers in 2011. The label 1 means the main plots investigated within this study and 2 is for the new measurement. expresses the potential open water evaporation according PENMAN (1948).

255

APPENDIX D: Example of the script applicable in Mini32 for calculation of actual evapotranspiration from BREB 2 Var1 = Chan17 ; Global radiation [W/m2 Var2 = Chan18 ; Net radiation [W/m2] Var3 = Chan19 ; Soil heat flux [W/m2] Var4 = Chan1 ; Air temperature [deg.C] @2 Var5 = Chan2 ; Air humidity [%] @2 Var6 = Chan3 ; Air temperature [deg.C] @0.5 Var7 = Chan4 ; Air humidity [%] @0.5 esup = 0.6108*exp(17.2694*chan1/(237.3+chan1)) ; Savage (2010) eup = esup*chan2/100 esdn = 0.6108*exp(17.2694*chan3/(237.3+chan3)) ; Savage (2010) edn = esdn*chan4/100 dt = chan1 - chan3 Var8 = dt ; dT 2-0.5m [deg.C] de = eup - edn var9 = de ; de 2-0.5m [kPa] rough_Bowen = dt/de*0.066 ; 0.066 is psychrometric constant kPa/K Var10 = rough_Bowen ; Bowen ratio (not corrected) Rn = fill(chan18,chan17*0.6819341048); substitution from global radiation regression relation G = fill(chan19,1) pre1 = -1*(de+0.066*dt)/(Rn-G) pre2 = limmin(pre1,0) pre3 = HEAV(pre2) pre4 = limmin(pre3,0.5) Bowen = dt/de*0.066*pre4 ; Right flux sign limit - Perez et al. (1999) Var11 = Bowen ; Bowen ratio - Perez et al. (1999) Bowen_a = limmax(Bowen,-1.3) ; Ortega-Farias et al. (1996) Bowen_b = limmin(Bowen,-0.7) ; Ortega-Farias et al. (1996) Bowen_ab = fill(Bowen_a,Bowen_b) Var12 = Bowen_ab ; Bowen ratio-Ortega-Farias et al. (1996) T_error = 0.08 ; for halfhourly integration RH_error = 0.15 ; for halfhourly integration slope = 4098.02862*esup/(237.3+chan1)^2 ; Savage (2010) e_error = SQRT((esup/100)^2*(RH_error)^2+(chan2/100)^2*(slope)^2*(T_error)^2) epsilon = (0.066*T_error+e_error)/ABS(de) ; Guo et al. (2007) low = (-1)-epsilon ; Guo et al. (2007) up = (-1)+epsilon ; Guo et al. (2007) Bowen_c = limmax(Bowen,low) Bowen_d = limmin(Bowen,up) Bowen_cd = fill(Bowen_c,Bowen_d) Var13 = Bowen_cd ; Bowen ratio-Guo et al. (2007) prewind = chan21 wind = fill(prewind,1.58) pre1 = limmin(chan21,0.9999) ; Foken (2008a) pre2 = HEAV(pre1) wind_limit = limmin(pre2,0.5) Bowen_e = Bowen_cd*wind_limit ; Foken (2008a) Var14 = Bowen_e ; Bowen ratio-Foken (2008a) LE_a = (Rn-G)/(1+rough_Bowen) ; LE (not corrected) H_a = (Rn-G)/(1+rough_Bowen)*rough_Bowen ; H (not corrected) LE_b = (Rn-G)/(1+Bowen) ; Right flux sign limit - Perez et al. (1999) H_b = (Rn-G)/(1+Bowen)*Bowen ; Right flux sign limit - Perez et al. (1999) LE_c = (Rn-G)/(1+Bowen_ab) ; Ortega-Farias et al. (1996) H_c = (Rn-G)/(1+Bowen_ab)*Bowen_ab ; Ortega-Farias et al. (1996) LE_d = (Rn-G)/(1+Bowen_cd) ; Guo et al. (2007) H_d = (Rn-G)/(1+Bowen_cd)*Bowen_cd ; Guo et al. (2007) LE_e = (Rn-G)/(1+Bowen_e) ; Foken (2008a) H_e = (Rn-G)/(1+Bowen_e)*Bowen_e ; Foken (2008a) Var15 = LE_a ; LE (not corrected) Var16 = H_a ; H (not corrected) Var17 = LE_b ; LE-Perez et al. (1999) Var18 = H_b ; H-Perez et al. (1999) Var19 = LE_c ; LE-Ortega-Farias et al. (1996) Var20 = H_c ; H-Ortega-Farias et al. (1996) Var21 = LE_d ; LE-Guo et al. (2007) Var22 = H_d ; H-Guo et al. (2007) Var23 = LE_e ; LE-Foken (2008a) Var24 = H_e ; H-Foken (2008a) Bowen_error = (T_error/ABS(dt)+e_error/ABS(de))*Bowen_cd ; Fuchs and Tanner (1970) - in case of Guo lim. Var25 = Bowen_error ; Bowen ratio error Var26 = Bowen_error/Bowen_cd ; Relative Bowen ratio error Rn_G_error = 0.05*ABS(Rn)+0.5*ABS(G) LE_error = ((Rn_G_error)/ABS(Rn-G)+Bowen_error/ABS(1+Bowen_cd))*LE_d ; Fuchs and Tanner (1970) - in case of Guo lim. Var27 = LE_error/LE_d ; Relative LE error LE_low = (LE_d)-(LE_error)

256

Var28 = LE_low ; LE low limit LE_up = (LE_d)+(LE_error) Var29 = LE_up ; LE up limit ET = 1800*LE_d/((2.501-0.002361*chan1)* 10^6) ; Shuttleworth (2007) Var30 = ET ; ET_actual [mm] Var31 = 1800*LE_low/((2.501-0.002361*chan1)* 10^6) ; ET_low [mm] Var32 = 1800*LE_up/((2.501-0.002361*chan1)* 10^6) ; ET_up [mm] alfa = (slope+0.066)/(slope*(1+Bowen_cd)) Var33 = alfa ; alfa (Pristley-Taylor) equilibrium_LE = slope/(slope+0.066)*(Rn-G) Var34 = equilibrium_LE ; Equilibrium_LE [W/m2] Var35 = 1800*equilibrium_LE/((2.501-0.002361*chan1)* 10^6) ; Equilibrium_ET [mm] Var36 = 1.26*equilibrium_LE ; LE Priestley-Taylor Var37 = 1.27*1800*equilibrium_LE/((2.501-0.002361*chan1)* 10^6) ; PET [mm] Priestley-Taylor vpd = esup*(1-chan2/100) ra = 208/wind ; aerodynamic resistance [s/m] Allen et al. (1998) rc = ABS((1.22*1004*vpd)/(slope*(Rn-G))) ; surface resistance [s/m] Ortega-Farias et al. (1996) PM_numerator = slope*(Rn-G)+1004*1.22*vpd*ra^-1 PM_denominator = slope+0.066*(1+rc*ra^-1) Var38 = (PM_numerator/PM_denominator) ; LE_PM [W/m2] Ortega-Farias et al. (1996) var39 = (PM_numerator/PM_denominator)/((2.501-0.002361*chan1)*10^6)*1800 ; ET_PM [mm] Ortega-Farias et al. (1996) ra_todor = 208/wind ; aerodynamic resistance [s/m] t = 0.066/slope*vpd/(slope+0.066) ri = 1.22*1004*vpd/(0.066*(Rn-G)) Y = ri/ra_todor a = (slope+0.066*Y)/(slope+0.066)*Y*vpd b = -0.066*Y*t c = -(slope+0.066)*t X = (a+SQRT(b^2-4*a*c))/(2*a) rc_todor1 = ABS(X*ri) rc_todor2 = -1*(limmax(Rn-G,0.00001)) rc_todor3 = HEAV(rc_todor2) rc_todor4 = limmin(rc_todor3,0.5) rc_todor5 = rc_todor4*200 ; Perez et al. (2006) - 200 s/m during nighttime rc_model = fill(rc_todor1,rc_todor5) Var40 = rc_model ; rc_model [s/m] Todorovic (1999) PM_numerator = slope*(Rn-G)+1004*1.22*vpd*ra_todor^-1 PM_denominator = slope+0.066*(1+rc_model*ra_todor^-1) PM = PM_numerator/PM_denominator ; LE [W/m2] Todorovic (1999) Var41 = PM ; LE [W/m2] Todorovic (1998) ETtodor = PM/((2.501-0.002361*chan1)* 10^6)*1800 ; ET_PM [mm/h] Todorovic Var42 = ETtodor; ET_PM [mm/h] Todorovic (1999) PenmanLE_1 = slope*(Rn-G)/(slope+0.066) PenmanLE_2 = (0.066*(6.43*(1+0.536*wind)*vpd/24/3600*10^6))/(slope+0.066) PenmanLE = (PenmanLE_1+PenmanLE_2) Var43 = PenmanLE ; LE_Penman (1948) Var44 = PenmanLE/((2.501-0.002361*chan1)* 10^6)*1800 ; ET_Penman (1948)

257

APPENDIX E: Pictures from the experimental site

View from the poplar plantation on the biomass district heating plant in the background.

The plantation after the winter harvest – double row design with inter-row distances of 2.5 m and spacing of 0.7 m within rows accommodating a theoretical density of 9,216 trees ha-1.

Characteristically large leaves of Populus nigra x P. Maximowiczii (clone J-105). 258

The BREB system at the turf grass with the poplar plantation in the background.

Vertically adjustable BREB system above the poplars placed at 12 m high aluminium mast.

Detail view on the BREB system recently supplemented also with eddy covariance.

259

Measuring leaf area index by SunScan canopy analyser system.

Allometric measurements: determining the relations between biomass content and other easily measurable parameters like or tree length.

Automatic dendrometer measuring the value of stem circumference each hour. 260

ANNOTATION The potential for increasing the areas of energy crops is mainly the result of a recent EU policy oriented towards the energy source diversity, security, independence from the main world oil suppliers and environmental friendliness. According to many studies that compare the particular economic and environmental features of energy crops, the perennial grasses and especially high density woody cultures grown in a short rotation period, called short rotation coppice (SRC), show the best performance. Prior to the large-scale introduction of these new cultures it is necessary to thoroughly assess what the specific impacts of their growing are. Besides the issues like carbon footprint, economic effectiveness, soil properties and biodiversity, the water balance remains a very crucial question. Therefore, the main aim of this study is to quantify the water use of high density poplar stand grown as SRC for energy use on arable land in the conditions of CzechMoravian Highlands. For assessing the water balance of this SRC several independent methods were used. However, as the key technique for the estimations of the actual evapotranspiration ( ) the Bowen ratio energy balance (BREB) method was chosen. BREB is based on the gradient measurements of air temperature and humidity and simultaneous measurements of the radiation balance and soil heat flux. Although BREB may be considered as a point measurement, it actually integrates larger surroundings based on the instruments vertical adjustment, wind conditions and the atmospheric stability. In order to compare the of SRC with other cultures, the other BREB system was placed at the adjacent turf grass that served as a reference surface. Apart from the BREB method, other techniques like soil moisture measurements, sap flow, leaf area development and biomass increments were assessed. Further, the BREB method was not used only for estimates, but it enabled to determine other variables like surface resistance, coupling to the atmosphere, crop coefficient and water use efficiency which serve as suitable inputs in many, mainly water balance, models. As the main result of the research it was found, surprisingly, that water use of SRC and the turf grass (or grassland in general) are more or less comparable and moreover, in some cases the water use of SRC is smaller. The during the growing season 2009 (1st April to 30th September) amounted 509 and 553 mm for the SRC in the eighth year of the first rotation period and the turf grass respectively. Although the of SRC exceeded the of the turf grass in several moths (usually those with high precipitation totals), the differences were not statistically significant and generally lay within the range of the errors of the method itself. However, the of SRC was significantly lower during April for which no or small leaf area index (LAI) is typical. Also within the next growing season 2010 following the winter harvest SRC showed significantly lower rates with the seasonal sum 264 mm compared to that of turf grass reaching up to 474 mm. The maximum mean canopy LAI in 2009 culminated at 7.5 whereas in 2010 only at 3.7. The above-ground woody biomass production estimated on the allometric and stem increments measurements basis was 13.5 and 16.5 t ha-1 (DMC) in 2008 and 2009 respectively. In 2010 the production of the resprouting shoots attained almost to 4 t ha-1 (DMC) per year. To conclude, the results show that in the recent climate conditions the SRC productivity in the Czech-Moravian Highlands is not limited by water while attains to a satisfactory level. 261