Seasonal variability of the W

Elsevier Editorial System(tm) for Journal of Hydrology Manuscript Draft Manuscript Number: Title: Seasonal variability of the Western Siberia wetlands from satellite radar altimetry Article Type: Research Paper Keywords: wetlands, remote sensing, radar altimetry, hydrological cycle, watershed, evapotranspiration, floodplain dynamics, Western Siberia, Ob' river Corresponding Author: Dr. Alexei Kouraev, Corresponding Author's Institution: LEGOS-OMP First Author: Elena A. Zakharova Order of Authors: Elena A. Zakharova; Alexei Kouraev; Frédérique Rémy; Valeri A. Zemtsov; Sergey N. Kirpotin Suggested Reviewers: Terry Prowse University of Victoria, Canada [email protected] Claude Duguay University of Waterloo, Canada [email protected] Charon Birkett University of Maryland, USA [email protected] Annett Bartsch Technical Univerisy of Vienna, Austria [email protected] Jerome Benveniste European Space Agency, ESRIN, Italy [email protected]

*Manuscript Click here to download Manuscript: JOH13 21 wtsh vet 02.doc

Click here to view linked References 1

Seasonal variability of the Western Siberia wetlands from satellite radar altimetry Elena A. Zakharova, CNRS, LEGOS, UMR5566 (CNRS-CNES-IRD-University of Toulouse III), France and State Oceanography Institute, St. Petersburg branch, Russia

Alexei V. Kouraev, University of Toulouse III, LEGOS, UMR5566 (CNRS-CNES-IRD-University of Toulouse III), France and State Oceanography Institute, St. Petersburg branch, Russia

Frédérique Rémy, CNRS, LEGOS, UMR5566 (CNRS-CNES-IRD-University of Toulouse III), France

Valeri A. Zemtsov and Sergey N. Kirpotin, Tomsk State University, Tomsk, Russia

Corresponding author: Elena A. Zakharova, CNRS, LEGOS, UMR5566 (CNRS-CNES-IRDUniversity of Toulouse III), France ([email protected])

2

Key points 21 watersheds of the Western Siberia are studied using radar altimetry Seasonal variability of wetlands extent and timing of specific phases are analyzed Seasonal amplitude of water level variation in lakes and bogs is estimated

Abstract Boreal wetlands play an important role in the global water and carbon cycle. The Western Siberian plain is the most bogged region of the world. Large flooded areas, lakes and wetlands present a temporary storage of water after the spring snow melt, gradually releasing water in summer. Knowledge of wetlands hydrological regime is important to understand the water and carbon cycle in this vast region. Existing historical data provide estimations of peatlands and lakes extent only as static average values, without account of seasonal or interannual variability of wetlands extent. Satellite radar altimetry provide reliable, regular and weather-independent data on surface properties allowing to analyse temporal and spatial changes of the wetlands of Western Siberia. We focus on seasonal variability of wetlands hydrological regime over the 21 mid-size watersheds of the Western Siberia located in different natural conditions. By using ENVISAT RA-2 radar altimeter data we analyse seasonal variability of wetlands extent and water level during different phases of hydrological regime in the context of snowmelt, rains and evapotranspiration. There are three main types of wetland water regime characterised by: 1) spring inundation and following deep drainage with/not secondary peak in autumn; 2) spring inundation and low summer variation; 3) spring inundation with medium summer drainage and second autumnal peak. The floodplain inundation contributes less than 8% to the total wet zones extent, confirming that seasonal extent variation is entirely related to the processes taking place in the wetlands of interfluvial areas. The spring inundation lasts for almost 2 months with a latitudinal gradient of melting dates of 8 days / 2°. The wetland's full freezing occurs in middle of November on north and in middle of December on south. The specific scattering properties of lake ice impact the quality of timing of freezing in the regions with lake fraction more than 20% in Surgut Polesye. No considerable latitudinal gradient have been found for dates of full freezing. The seasonal amplitude of water level variation for northern part of Western Siberia from altimetry is 0.7-1.5 m for lakes and 0.2-0.5 m for bogs. This represents seasonal variation of wetland water storage of 400 mm for non-permafrost and 150 mm for permafrost-affected zones and demonstrates more important role of surface runoff in water balance of permafrost-affected areas and smaller regulation effect of wetlands on basin water outflow.

3

Keywords:

wetlands,

remote

sensing,

radar

altimetry,

hydrological

evapotranspiration, floodplain dynamics, Western Siberia, Ob' river.

cycle,

watershed,

4

Introduction

Boreal wetlands play an important role in the global water and carbon cycle. The Western Siberian plain is the most bogged region of the world - in some parts up to 70-80% of its territory is covered by bogs, in overall 1 million km². In this region modern climate conditions (excess of precipitation over evaporation, short and cold summers) and specific geomorphology (flat relief, low slopes, presence of permafrost in the northern part) favor formation of a multitude of interconnected natural objects - large and small rivers and streams, extensive floodplains, lakes and - especially - peat producing wetlands. The presence of large flooded areas, lakes and wetlands results in a rate of evaporation higher than for any other large boreal watershed (Sokolov, 1952). These natural objects also present a temporary storage of water after the spring snow melt, gradually releasing water in summer. Evaporation and retention of water in wetlands thus significantly influence seasonal redistribution of the river discharge.

With about 40% of the world pristine peatlands located there, Western Siberia also plays an important role in the global cycle of carbon. In the southern part, peatlands act as a terrestrial sink of atmospheric carbon through photosynthesis and carbon accumulation in peat deposits, while in the northern permafrost-affected regions peatlands represent a source of methane emission to the atmosphere. As the role of emission depends on wetness conditions of peatlands, knowledge of wetlands hydrological regime is important to understand the water and carbon cycle in this vast region.

Existing historical Soviet (Vodogretskiy 1973, Ivanov and Novikov, 1976) and global (Lehner and Doll, 2004) data provide estimations of wetlands and lakes extent over the different watersheds of the Western Siberia. These data are based on the cartographic analysis of maps and represent static minimal values, without account of seasonal or interannual variability of wetlands extent. In situ observations on wetlands extent are difficult to realize and thus they are not part of regular observations done at the meteorological or hydrological stations. In this context an evident choice is the use of satellite observations, especially in the microwave range (not affected by the presence of cloud cover).

Satellite observations in the microwave range provide reliable, regular and weather-independent data on surface properties. One way to estimate the presence of water on land is to analyze the backscatter coefficient from active microwave instruments (such as SAR, scatterometer, radar altimeter etc). The backscatter coefficient is the ratio between the power reflected from the surface

5

and the incident power emitted by the onboard radar, expressed in decibels (dB), and this parameter is very sensitive to discriminate between wet or flooded areas from the land.

Side-looking radars and scatterometers have been successfully used for studies of soil wetness (Pathe et al., 2009) and open water identification. Papa et al. (2007) have used a multisatellite method, based on a combination of passive microwave data (SSM/I), ERS-1 scatterometer and AVHRR NDVI to estimate pixel fractional coverage of open water at the 0.25° resolution at the equator (773 km2 pixel size). They have estimated spatial and temporal variations over the Ob' river with monthly resolution for 1993-2000. Bartsch et al (2007a) have used ENVISAT ASAR in global mode to identify and create the static mask of boreal peatlands in the Western Siberia. They have also used QuickScat scatterometer data to analyze spatial and temporal variability of spring freeze/thaw cycles for the region of Siberia II (Central Siberia, mostly Yenisey watershed) (Bartsch et al., 2007b). However, these are the estimates that are not very suitable for the study of seasonal wetlands dynamics at sub-monthly timescale.

Another promising source of active microwave data are observations from the radar altimeters. Due to nadir-looking capabilities they are much more sensitive to variations in the surface type than the side-looking instruments, due to lesser energy losses. Compared to many instruments, the altimetric data are also able to enhance the information content due to their high spatial resolution (400-600 m depending on satellite mission) along the track. While the theoretical footprint of the altimeter data is about 12 km (for the oceanic-type surface), over the land the main part of the backscatter signal comes from a small area with a diameter of 1-2 km, which occurs in the case of the quasi-specular signal over ice (Legrésy and Rémy, 1997) or calm water (rivers and small lakes). High radiometric sensitivity and spatial resolution along the satellite track can be successfully used for estimating the extent of wet and flooded zones, as well as formation and disappearance of lake and river ice.

A combination of simultaneous active and passive data from radar altimeters or passive microwave observations from SMMR-SSM/I has also been successfully used for studies of freezing and melting processes of the five largest Eurasian continental water bodies – the Caspian and Aral seas, and the Baikal, Ladoga and Onega lakes (Kouraev et al., 2008), as well as for the Ob' river (Troitskaja et al., 2013).

Papa et al. (2006) have used TOPEX/Poseidon dual-frequency radar altimeter to study inundated wetland dynamics over boreal regions up to 66°N (TOPEX/Poseidon orbit limitation. Frappart et al. (2012) also used TOPEX/Poseidon altimeter for studies of the inundation dynamic of the Ob River

6

floodplain. But it is still unclear is the water regime of wetlands on interfluvial areas different of that of floodplain. ENVISAT RA-2 radar altimetry data has been also used to estimate seasonal wetland extent variability for the Poluy, Nadym, Pur and Taz rivers (Zakharova et al., 2009, Kirpotin et al., 2009) located on the far North of the Western Siberia plain in the permafrostaffected zone. In this area the flat and big mound bogs are largely developed. They significantly redistribute river runoff, retaining in spring in average 37% of meltwater (Zakharova et al., 2011) by increasing its area on 25-35 % (Zakharova et al., 2009). Does the water regime of other Siberian wetlands express the same seasonal pattern?

In this work we focus on seasonal variability of wetlands extent over the whole Western Siberia. To do this, we use ENVISAT radar altimeter data over the 21 mid-size watersheds that belong to the Ob' river basin or are located in the neighbouring regions (the Taz, Pur, Nadym rivers). Among the selected watersheds is also an area within the Ob' watershed that does not contribute to the Ob' river system (Inner watershed) and represents about 15% of the Ob' watershed area. These watersheds cover the regions of Western Siberia that are located in different natural conditions and thus present a good base to analyze the seasonal dynamics of wetlands as a result of combination different factors (climate, soil, relief, vegetation etc).

First we present natural conditions of these watersheds (Section 1), describing the climate conditions, relief, permafrost conditions and different types of peatlands (bogs and fens) that are largely present in the Western Siberia. Then, using the altimetric data from ENVISAT (Section 2) we analyze seasonal variability of wet zones extent for different selected watersheds and assess their relation with evapotranspiration and liquid precipitation (Section 3). We then analyze how different parts of the watershed react to snowmelt and precipitation by quantitatively estimating for the Taz river watershed the role of the floodplain in the seasonal wetland extent dynamics (Section 4). Wetland extent is an important parameter, but in order to estimate water budget of wetlands we need to complement it by simultaneous estimates of water level. In Section 5 we present the results of water level variability of different types of lakes and wetlands from radar altimetry.

7

1. Natural conditions of the Western Siberian plain and selected watersheds Western Siberian plain extends on more than 2000 km from south to north and on more than 1300 km from west to east. It is vast flat region located between the Urals Mountains and the Middle Siberian Highland (Figure 1). Most of its territory comprises the Ob' river watershed (almost 3 million km2), while in its northern part are located rivers Nadym, Pur and Taz, as well as left (western) tributaries of the Yenisey River. According to the hydrographic conditions and river regime, the Ob' is usually divided into the three main parts– the Upper Ob' (from the confluence of Biya and Katun' rivers in Altai mountains up to the confluence of Ob' and Tom' near Tomsk), the Middle Ob' (from the Tom' mouth to the Irtysh mouth near Khanty-Mansiysk), and the lower Ob' (from the Irtysh mouth to the Ob' bay).

The plain is mainly composed by continental deposits of Quaternary age. The absolute elevation is in general between 105-130 m above sea level. Sibirskiye Uvaly hills (altitude ranging from 170200 m in the west to 230-280 m in the east) is an uplift terminal moraine that divides the plain on two parts: northern and southern . The northern part is covered by marine, glacial and fluvioglacial deposits and is affected by modern periglaciation. Starting from the Sibirskiye Uvaly to the Ob valley (Surgut and Vakh Polesye) the mineral deposits are composed by fluvioglacial sands. Further south the deposits are loams and sands of alluvial and lacustrine origin. On the southern limit of the Western Siberia these deposits are eroded by ancient and modern rivers that have formed the series of sandy SW-NE oriented ridges (Sergeev, 1976).

In our study we use river basin approach. We have selected 21 middle-size watersheds located in different climatic and natural zones of the Western Siberia. Watersheds north of the Sibirskiye Uvaly include the Taz, Pur and Nadym river (not flowing directly to the Ob' river, but to the Ob' and Taz estuaries), the Shuchya and Poluy rivers (joining the Ob' in its lower reaches), and the Kazym river (flowing west from the Sibirskiye Uvaly). South from Sibirskiye Uvaly are the right (the Ket', Tym, Vakh, Tromyugan, Lyamin, Nazym and Kazym) and left (Bol'shoy Yugan and Vasyugan rivers) tributaries of the Middle Ob'. Four watersheds (the Tura, Tavda, Konda and Severnaya Sos'va rivers) are located between the Urals Mountains and Irtysh and Lower Ob'. Two watersheds (the Om' river and Inner watershed) are located in the southernmost part of the Western Siberia between the Irtysh and Ob' rivers. The area of the selected watersheds varies from more than 100 000 km2 for (the Taz and Pur rivers and Inner watershed) down to 10 000 km2 (Pim, Nazym and Shuchya rivers).

8

Most of the Western Siberia has a moderate continental climate (except the region of lower reaches of the Pur and Taz, located in the sub-Arctic climate) with short summers and cold and long winters. The mean annual temperature varies from -10°C in the north to +0.5°C in the south. The duration of the period with the stable negative air temperature changes from 210 to 160 days, correspondingly (Kotlyakov, 1997). However seasonal variability is high and difference between winter minima and summer maxima of air temperature can reach 45°C.

The annual precipitation is 400-600 mm (Kremenetsky et al., 2003), with the maximum values observed in summer. Solid precipitation decreases from more than 200 mm/yr. (watersheds of the Tym, Ket', Vasyugan rivers, region of Sibirskiye Uvaly and rivers located north from it) to 50-100 mm/yr. for the region of Inner watershed. Snow depth changes from 40 cm in north to >100 cm in the central and 80 cm in the south. Precipitation is larger than evaporation on the whole Western Siberia (Ivanov,1975). The runoff decreases from north to south from 300 mm to 10 mm/yr. (Kotlyakov, 1997).

There are three main vegetation biomes (tundra, taiga and steppe) on the Western Siberia plain. They are divided by transitional vegetation (tundra-forest, small-leaf forest and forest-steppe) that can form the belts several hundreds of kilometers wide. Most of the selected watersheds are located in the taiga zone (or transitional zones). Inner watershed and southern parts of the Tura and Om' watersheds are located in the steppe zone, and Shuchya river and lower reaches of the Taz and Pur rivers are located in the tundra zone.

A large part of the Western Siberia is affected by permafrost (Figure 2a). In the northern part (river Shuchya and lower reaches of the Nadym, Pur and Taz rivers) permafrost is continuous This region corresponds to the tundra and forest-tundra zones (Popov, 1989). The depth of the frozen ground here reaches 300-600 m (Kremenetsky et al., 2003) and the soil temperature decreases from -1°C to -9°C northward. The frozen ground is absent only under large rivers and lakes (with more than 2 m depth) that form the taliks. The through taliks comprise the ground waters that support the winter discharge of the big rivers (Vodogretskiy, 1973). A large zone north of the Middle Ob' is affected by discontinuous permafrost with temperature -1°C - 0°C.. The unfrozen ground (talik) here occupies 45-50% of the territory, while frozen ground occurs mainly within the peat deposits or in the moist loam/clay mineral soils. Further south (northern parts of Konda and Tavda watersheds, region of the Middle Ob' and large parts of Vakh, Tym and Ket' basins) lies the zone of sporadic permafrost. Here the modern permafrost is found

9

only on the steep slopes of northern orientation, under the peat heaves or wet peatlands, in the narrow valleys of small rivers (Popov, 1989).

The depth of seasonal melting or active layer in mineral soils increases from 0.5-1 m on the north to 2-2.5 m on the south of cryozone. For organic (peat) soils there is no latitudinal variation in active layer depth due to insulating properties of the peat. The summer melting does not penetrate more than 0.1-0.5 m, thus limiting water storage capacity of soil, as well as carbon and heat exchange. The same depth is found for seasonal freezing of peat soils in permafrost free regions (Popov, 1989).

The water regime of the wetlands is governed by the water budget and wetland storage capacity. If the water budget is determined by climatic factors, the storage capacity is related to wetland morphology (geology, wetland type, vegetation etc.).

A typical feature of the Western Siberia is large extent of peatlands (Figure 2b). Peat producing wetlands form under conditions when organic matter production rate exceed decomposition, allowing the peat accumulation. Peatlands can be subdivided hydrologically into two main types: fens and bogs, that differ by their water sources. Fens receive the water from two sources : atmospheric precipitation and lateral surface and subsurface flow. The bogs are fed exclusively by atmospheric precipitation. If at very wet conditions the bogs receive the surface lateral flow from the headwaters, they form a subtype of transitional bog. From both types of wetlands the water discharges through evapotranspiration, infiltration to groundwater and lateral (surface or subsurface outflow).

In West Siberia lowland both type of peatlands are presented. In northern and central part the fens occur mostly on floodplains. In southern part, south from 57 °N they occur also on terraces and interfluvial surfaces (Ivanov, 1975). However, at the scale of the Western Siberia bogs are dominating type of wetlands. Their landforms are very variable with respect to vegetation and hydrological conditions, and represent different phases of paludification processes (Ivanov, 1975).

Principal types of bogs distinguished in the northern part of Western Siberia are polygonal bogs, flat mound bogs and big mound bogs, typical for the Taz, Pur, Nadym, Poluy and Shuchya river watersheds. Most of the other watersheds are located in the zone of raised mound oligotrophic sphagnum bogs. The fens are divided on a) flat eutrophic/mesotrophic and b) concave eutrophic or brackish fens and are typical for the Tura, Om' and Inner watersheds.

10

The lakes are important element of all bog classes. They can be primary, originated in prepaludification time, and secondary, formed during bog evolution. The latter are most abundant. In the permafrost regions secondary lakes (thermokarst lakes) are formed by thermal erosion and represent arctic ponds that can later evolve into lakes and later to khasyreys (drained lakes) (Kirpotin et al., 2009). In non-permafrost landscapes lakes are formed by the peat erosion (compaction, subsidence etc.) processes.

As a result, combination of different climatic and geomorphological conditions, vegetation and soil type lead to large differences in wetlands functioning between the selected watersheds. Use of satellite radar altimetry data makes it possible to analyze seasonal variability of wetlands extent and water level during different phases of hydrological regime.

11

2. Data used and methodology

This study used altimeter RA-2 data from ENVISAT satellite, operating from 2002 to 2012. The repeat period (cycle) for ENVISAT is 35 days, and 18 Hz sampling rate gives along-track resolution of 370 m. ENVISAT tracks provide a homogeneous coverage of the selected watersheds with the mean number of satellite tracks and observations for each of satellite's 35 days-long cycle presented in Table 1.

In order to discriminate open water and wetlands from land we use a threshold approach for the backscatter data in Ku (13.6 GHz) band that we have already applied for the Poluy, Nadym, Pur and Taz watersheds (Zakharova et al., 2009). Backscatter seasonal variation can be used as an indicator of changes in climatic and surface conditions on watersheds. Backscatter seasonal cycle on the entire Western Siberia plain is characterized by two maxima occurring in spring and autumn and by two minima observed in winter and in mid-summer The largest spring maximum is related to water accumulation in ground pools, lakes, bogs and floodplains due to the snow melting. It takes place in April in the southern and in June in the northern watersheds. In summer the backscatter reduces as water infiltrates, drains out or evaporates from the surface. In September-October backscatter increases again and the second maximum occurs due to replenishment of floodplains and bogs by rains.

The inland waters provide a much higher return signal than land. Birkett (1998) has found that if the ground target is dominated by water the backscatter retrieved by ocean retracker from TOPEX/Poseidon mission varies between 20 and 40 dB for Ku-band (Birkett, 1998).

Our observations for the ENVISAT data show that, depending on the watershed, there is a marked seasonal evolution of the backscatter values. In winter the backscatter is a good indicator of the snow wetness properties (Ulaby, 1983). At this time in Western Siberia backscatter values are low and vary between -5 and 15 dB with an average value about 6 dB. When the melting starts and the liquid water appears on the snow, the backscatter value rises on more than 20 dB in bogs as well as in forested areas. During spring-autumn most of the ENVISAT backscatter values in Ku-band are located in 10-35 dB range (see Figure 3a for the example of Taz watershed). With the start of winter, soon after the air temperature drops below zero, the backscatter decreases. In the beginning of winter it is of about 10 dB. As snow accumulates and increases the scattering of radar echo, the backscatter gradually declines till 6 dB in March-April.

12

However, in the context of absence of good physical model for radar echo over highly heterogeneous surface and of dedicated in situ observations, the selection of a threshold for discriminating land and wet zones for snow-free period represents a certain challenge.

The statistical approach was used for specification of the threshold value. Firstly, we have defined two types of elementary units: forest and lake-dominated wetland landscapes. Data along the ENVISAT 79 and 423 tracks in large (about 1500 km long ) south-north band have been selected for four main bog and vegetation biomes of the Western Siberian plain using GlobCover2000 Landsat mosaics (MrSID image server, 2012) and high-resolution SPOT images (Google Earth). They include: a) flat bogs and scarce forest in forest-tundra (64.5°N - 65.7°N), b) big mound bogs and coniferous forest in the northern taiga (61.7°N - 62.2°N), c) oligotrophic dommed bogs and dense coniferous/deciduous forest in middle taiga (57.9°N-58.8°N) and d) eutrophic/mesotrophic fens in forest-steppe biome (54.7°N - 56.0°N).

Statistical distribution (Fig. 3b) of summer (June-August) backscatter values shows two distinct clusters of values: forest-dominated landscapes have a peak located at 7 dB. A much smaller peak with high values (20-22 dB) is probably related with the late snow-melt in the northern regions. For the lake-dominated bog landscapes the maximal peak is observed at 29 and 32 dB. Several peaks are also identified at lower bins 11, 18 and 21 dB and can be attributed to snowy or drying conditions. To exclude the unusual conditions and remove the outliers related to these situations, we determined 95% and 5% frequency values for forest and wetlands

elementary landscapes

correspondingly. There is no overlapping between 95% frequency (15.7 dB) for forest-dominated and 5% frequency (17.1 dB) for lake-dominated landscapes, that prove the soundness of the criterions. Basing on this, we have used 17 dB as the threshold for discriminating between dry land and water/wet zones.

Then in order to compare watersheds with different size, we have calculated the ratio (%) of altimetric observations with values more than 17 dB to the total number of observations for each cycle and each watershed to serve as an equivalent of wet zones. Here under "wet zones" we denote a multitude of objects that are either constant in time (rivers, lakes, wetlands) or have seasonal variability ( river flooding plain). To estimate the uncertainty related with the selection of threshold, we have done a sensitivity study. For the Taz watershed (Figure 4) a change in 1 dB leads to less than 3-4% differences in the absolute values of mean monthly wet zones ratio. However, neither the character of seasonal evolution nor the timing of specific hydrological phases change. This makes it

13

possible to reliably detect and analyze spatial and seasonal changes in wet zones extent as discussed in the following section.

14

3. Seasonal variability of wet zones extent.

Using the threshold approach described in Section 2, the fraction of wet zones was calculated for the 21 selected watersheds. The ENVISAT temporal resolution (35 days exact repeat period) is not suitable for detecting fraction of wet zones, freezing and melting dates at the interannual scale for a specific target. However, by combining different tracks covering watershed at different dates and by putting together all observations for 2002-2010 we can construct the average seasonal cycle and detect the average (for the mentioned period) dates of specific events with an accuracy of 5-10 days.

Most existing studies have used monthly or annual average values of wetlands extent, river discharge, snow cover and other meteorological parameters to analyze the hydrological cycle. However, for more accurate estimation of the physical processes involved, it is necessary to look at the seasonal changes as one continuous physical process. Specific hydrological processes (such as spring flood etc) can start and end at different dates and its duration may vary significantly from one year to another, and thus some of the seasonal or interannual variability is hidden when monthly or annual averages are used. This issue is also important for selecting satellite data (optical or radar) for some fixed month, as it will not necessary provide a representative estimate.

3.1. Types of seasonal distribution According to the seasonal regime of wet zones the watersheds can be divided into 3 groups and one sub-group (Fig. 5). In the first group (Fig 6a) are the watersheds located on the south and southwest of the plain: Inner watershed, Om, Vasyugan, Tura, Tavda, Severnaya Sos'va. In these watersheds the spring rise of wet zones fraction is not so high as compared to with northern basins. The surface dries out very rapid till June and the summer values are lowest (25-30 %). This group also comprises a sub-group (1a) with watersheds located in the middle and eastern parts of Western Siberia plain (Ket', Tym, Bolshoy Yugan and Konda, Fig 6b). The specific feature of this sub-group is the more pronounced rise of wet zone extent in autumn due to rains.

The watersheds of second group include those situated in the central part of the Western Siberian plain on the right bank of the Ob' river (Lyamin, Pim, Nazym and Tromyugan, Fig 6c) as well as Shuchya River. This territory has exceptionally high fraction of wet zones (>70%) with very high portion of open water surface (inner-bog lakes). The main feature of the wet zone regime here is low seasonal changes in extent (less than 10%). In spite of small seasonal amplitude the second peak in wet zones extent in September is still recognized.

15

The watersheds of third group (Nadym, Kazym, Poluy, Pur, Taz, Vakh, Fig 6d) are located on the north and north-east of the Western Siberia in the zone of permafrost. The wet zones extent in these basins is high (spring values are 60-70 %) and has two maxima: one in spring and another in the early autumn. The second peak in autumn is almost as high as the spring one. In the summer big portion of the watershed dries due to drainage and evaporation (extent ratio is 55-65%).

The seasonal variation of the wet zone ratio in the watersheds of all groups is determined not only by precipitation regime. In spite of severe climate, during 4 month (3 month in the northern part) the evapotranspiration (ET) plays an important role in watershed water budget. According to ISBA simulations (Decharme et al., 2010) in July ET reaches 80-90 mm/month, while precipitation (P) are only of 60-80 mm. The period of rain deficit (P-ET) can last 1-3 month depending on latitude. The absolute values of deficit range from 30-40 in south to 15-20 mm in the watersheds of Middle Ob' and to 3-10 mm in the north of Western Siberia. In the northernmost Shuchya basin the deficit haven't been observed due to both higher precipitation and lower ET rate. The highest deficit corresponds to the beginning of the summer dry period (beginning of period of minimal extent of wet zones) and occurs in July. In August the ET decreases in 2 times on whole Western Siberia plain but precipitation does not change much. It results in positive P-ET term followed by progressive accumulation of water on watershed, increasing of wet zone extent and small rising of river discharge (not shown). Finally, in September when the precipitation decreases, we observe reverse watershed reaction - the second wet zones extent maximum in 1a, 2 and 3 groups basins. In the watersheds of proper first group the rain excess over ET in the end of summer is lowest and not sufficient for the second maximum of wet zones extent to be observed.

3.2. Timing of spring flood start and maximal wet zones extent

Following to the 0° air temperature isotherm propagation from south-west to north-east, the melting on the Western Siberia plain starts in mid-March in most southern watersheds (Fig. 7a). For a large number of watersheds melting starts in the beginning of April except for Tromyugan, Nazym and Kazym where it is observed in mid-April. For the northern watersheds melting starts first in the Pur watershed (end of April) and then it progresses northward. The difference in melting dates amounts up to two months between south (mid-March) and north (beginning and mid-May).

The maximal extent of the wet zones occurs 1-1.5 months after the spring start. First it is observed in the end of April for the southern watersheds, then for left tributaries of the Middle Ob' and Konda

16

watershed. Later on it affects the rest of left Irtysh tributaries and gradually right tributaries of the Middle Ob’, and then in the middle of June - the north-eastern watersheds (Fig. 7b).

The spatial variation of dates of melting start has a functional linear relation with latitude and changes on 8 days per each 2° of latitude. The date of maximal wet zones extent is less related to the latitude (r2=0.68) as it depends also on basin morphology (relief, lake and bog fraction etc), but it has the same spatial gradient (8days/2° latitude).

3.3. Timing of autumnal flood and freezing of wetlands Timing of the second (autumnal) maxima of wet zones extent differs by one month between north and south (Fig 8a). For Inner watershed the second peak is not detectable. Autumnal peak is first observed in Mid-September for northern (Taz, Nadym and Shuchya rivers) and eastern (Vakh, Ket') watersheds, then for most of the watersheds (end of September), then for tributaries of the westernmost stretch of the Middle Ob’ and Om’ (beginning of October), and latest – Tura and Konda on the left bank of Irtysh (mid-October).

The freezing of the wetlands has been defined as time when the ratio of wet zones decreases to values close to 0% (see Fig 6). Freezing starts on the north in the mid-November (Shuchya river), then by the end of November northern watersheds (Taz, Pur, Nadym, Poluy and Severnaya Sos'va), then goes south, the last ones – Vakh and Pim (mid January), and then Tromyugan and Lyamin (mid-February (Fig. 8b). The freezing dates as they are defined here are less dependent on climate conditions than on surface properties, notably the presence of lakes and bogs and its heat storage capacity. In the region of Surgut Polesye (watersheds of Pim, Lyamin, Tromyugan, partly Vakh) the fraction of lakes reaches 20% (Vodogretsky, 1973) and according to our approach this territory freezes in end of December even in January-February.

In situ observations on meteorological stations show that, in the middle of November the mean air temperature in this region reaches already -17°C -20°C. Knowing that there is no significant heat sources (anthropogenic hot wastes or hot springs) this excludes the possibility of having open water present. Apparently such late dates in the full freeze in the region of the Middle Ob' is related with influence of ice on radar backscatter. In winter the ice of big lakes in Surgut region provide a quasispecular echoes as observed for other boreal lakes (Baikal, Onega or Ladoga) (Kouraev et al., 2009), giving the high values of order of 15-25 dB. As a result, use of 17 dB threshold in winter for watershed in Surgut Polesye region potentially leads to some difficulties in discrimination between wet zones and land/ice. When the snow cover on lake ice becomes developed, it reduces the

17

backscatter values in second part of winter we no longer have ambiguity between water and ice discrimination. In other regions of Western Siberia smaller lakes do not manifest this ambiguity either due to mixed (snow-covered land and ice) character of footprint or to more favorable conditions of snow accumulation and thus thicker snow cover on lake ice. Further studies are needed to regional tuning of ice set-up timing using a threshold that is specific for watersheds with high lake fraction.

18

4. Impact of seasonally-flooded areas on wet zones ratio estimates. The seasonal variability addressed in Section 3 considers watersheds as a whole. However, within each watershed there are specific types of permanent (bogs and fens) and seasonally inundated (river floodplains) areas that reacts differently to snowmelt and precipitation. In order to quantitatively estimate how seasonally flooded areas contribute to the total estimates of the fraction of the wet zones, we have done an analysis for the biggest of the selected watersheds - the Taz river. This river has the largest floodplain that is up to 23 km wide in its middle and low reaches.

Using Landsat GlobCover mosaic (MrSID Image server, 2012) and topographic maps of 1/200000 scale, we have digitized the floodplain area, defined as "upper floodplain" (Fig 9). With the total watershed area being 139 600 km2, the floodplain represents 13.5% (18 760 km2) of the total watershed surface. Then as before we have calculated the ratio of wet zones for each ENVISAT cycle, this time separating it onto the whole watershed and floodplain area.

The temporal variability (Fig 10) shows that snowmelt leads to rapid increase of wet zones ratio already in late April - beginning of May. By the beginning of June (line A) the ratio of wet zones for the whole watershed reaches maximal values of 55-60% while for the floodplain maximal values (up to 85%) are observed later - in the second half of June (Line B). Noted time lag (half a month) indicates the meltwater travel time from the watershed to the floodplain that is specific for each river basin. The ratio of wet zones decreases during summer and reaches summer minima: only 35% for the whole floodplain but still high values (60-70%) for the floodplain, reflecting water concentration in the river channel and on the floodplain. If the floodplain dries in the middle of August, from the interfluvial areas the water drains out one month earlier - in July. As a result, the floodplain contribution to the satellite estimates of the wet zone extent is highest (8%) in August and not during the spring flood. The minimal floodplain contribution is observed in September and corresponds to the period of lowest for snow-free season river water level and river discharge. Autumnal rainfall leads to the increasing of wet zones ratio in late September (up to 53% for the whole watershed and 78% for the floodplain (Line D).

Freezing of wet zones starts on the

watershed and on the floodplain at the same time, in the middle of October, and lasts in average one month.

While for the Taz river the floodplain occupies 13.5 % of the total watershed area, its contribution to the ENVISAT estimates of wet zones extent ratio in snow-free season changes from 4 to 8%. For other rivers of the Western Siberia its contribution is expected to be even less, as they do not have such a vast floodplain. It means that for the small and medium size rivers of the Western Siberia the

19

seasonal variation of the wet zones ratio on the basin scale is entirely related to the processes taking place in the wetlands of interfluvial areas and is not related to the seasonal inundation of floodplain due to water level rise in rivers.

20

5. Water level variability in wetlands

Seasonal variability of wet zones extent is an important parameter that shows quite different behavior in various watersheds of Western Siberia. However, in order to correctly estimate water budget of wet zones we need to be able to complement it by estimates of water level variability.

Recently the radar altimetry have been used for observation of water level in the large rivers (Birkett et al., 2002, Kouraev et al., 2004, Birkinshaw et al., 2010), lakes (Becker., et al., 2010., Cretaux et al., 2006 Berry et al., 2005), reservoirs (Zakharova et al., 2007; Kouraev et al., 2007; Birkett., 2010, Pereira-Cardenal et al., 2010), flood plains (Andersen et al., 2010, Frappart et al., 2005). Radar altimetry has been successfully used for the estimates of the Ob' river level near Salekhard (Kouraev et al., 2004), for the Lower Ob' floodplain (Frappart et al., 2010). There are works on estimates of water level of tropical wetlands such as Okawango or Mekong (Frappart et al., 2006). However, so far there are no estimates of surface water level variability of the boreal bogs and lakes.

In West Siberia the large territories between 61°N and 64° that have flat relief. The surface of these bogs represents an alteration of small peaty ridges of 30-50 cm height and wet hollows. As the bog water table in this region is close to the surface (Ivanov and Novikov, 1976) and the peat water holding capacity is very high (40-80%), the surface peat is usually wet even in ridges and radar signal behavior over bogs is similar to that over flood plain. For these bogs the radar altimetry can provide estimations of water level changes for relatively homogeneous (same structure of mesorelief) surfaces.

Another typical feature of bog landscapes are lakes. In the area 61°N - 63°N / 69°E-77°E (Surgut Polesye) they occupy in average about 20% of the surface (Vodogretsky, 1973). The largest lakes have surface of 40-90 km2, while the smallest are of only several hectares. All lakes are hydraulically connected to the bog waters and this whole hydrologic system experiences the seasonal variation in water height (Ivanov and Novikov, 1976).

We studied the water level variation in various wetland landscapes using ENVISAT 18 Hz data (Ice2 retracker). Environmental and geophysical corrections of the altimeter range measurements relevant to the study regions have been applied. These corrections include ionospheric, dry and wet tropospheric, solid Earth tide corrections and correction for the satellite’s center of gravity. We neglect, on the other hand, corrections specific to open ocean environments such as ocean and pole

21

tides, ocean tide loading, inverted barometer effect and sea state bias (Kouraev et al., 2004). The flat bogs in the two regions were selected: 1) Surgut Polesye with no modern permafrost (61°N 63°N and 69°E-77°E) and 2) Nadym region - Nadym and Pur rivers interfluve (63°N-66°N and 7376°E) in the zone of discontinuous permafrost (See Fig 1). In the Nadym region there is usually a frozen ground below the bogs, while the forested area are developed on melted ground (Popov, 1989).

The highest seasonal water level variability from 1 to 1.5 meters is found for middle and big size lakes in Surgut Polesye (Fig. 11). In the bogs of this region the seasonal amplitude is smaller and does not exceed 50 cm. The in situ studies of seasonal water level variation in bogs and lakes in Western Siberia (Ivanov, Novikov, 1976) and Scandinavia (Kellner et al., 2002) support our altimetric estimates.

The lakes of the Nadym region are smaller in size than those of the Surgut Polesye.. They have banks of 1-1.5 m high from the dominant windward side and flat swampy banks from the leeward side that substantially limit the seasonal variation of the water table. The amplitude of water level retrieved by ENVISAT in lakes of Nadym region is less than 0.7 m. On the south of Nadym region the satellite retrieved seasonal amplitude of water level in flat bogs is of 30 cm. On big drained lakes (khasyrey) largely developed in this region (Kirpotin et al., 2008, 2009) the seasonal water level variation is 50-70 cm but can reach as high as 1 m depending on the banks height and morphology of khasyrey's drainage system.

On the north of Nadym region in the zone of continuous permafrost the time series of altimetric height are noisy. We suggest that in this area the radar signal is strongly affected by specific freezeup processes that go in two directions: from up to bottom (seasonal freezing) and from bottom to up (seasonal permafrost front rising). The time series processing for these bogs demands additional methodological developments. However, we can estimate the potential maximal seasonal variation of water level in these systems. The shallow depth of the active layer in flat bogs of this region limits the warm season water table amplitude till to 25-50 cm on the ridges and 50-80 cm in the hollows (authors measurements). Due to close position of frozen ground and constant rising of melted water to the surface the peat moisture content is usually high and precipitation lost for soil saturation is small. The water excess accumulates and fills small depressions. The vertical scale of these depressions (mesorelief scale) is of 30-40 cm and defines, finally, the maximally possible seasonal water amplitude as it can be seen by radar.

22

Using ENVISAT estimates of the seasonal amplitude of the water level and historical estimates of the bog and lake fraction published in issues of Russian Water Cadastre (Vodogretsky, 1973) it is possible to roughly evaluate the seasonal changes in water storage in surfacial pool of the watersheds (Table 2). This calculation was performed for 3 watersheds located in zone of permafrost and flat mound bogs (Nadym, Pur and Taz rivers) and 3 watersheds in Surgut Polesye that belong to non-permafrost big raised mound bogs zone (Lyamin, Pim and Tromyugan rivers). In the three basins of non-permafrost zone the seasonal water storage is about 30 km3, in northern basins it equals to 39 km3. By converting this values to spatial units it gives correspondingly 400 mm and 150 mm. The errors of order of 20 cm in retrieving of altimetric water level amplitude in lakes give 5-10% of errors in water storage estimates on the basin scale. This inaccuracy rises to 20% in the case of 20 cm errors obtained for bog level retrievals. These values of water storage are in agreement with previous estimates made from the in situ gravimetric measurements of water content in the bog profiles in Western Siberia (Smith et al., 2012). They estimated that in Siberian wetlands the seasonal variation of water storage is 120 km3. Extrapolated to the same area (592 400 km2) our results give 140 km3.

In the watersheds of non-permafrost zone the seasonal variation of the water storage in wetlands is 1.5-2.5 times higher than water runoff and compensates the evaporation from the basin surface. In northern watersheds, due to the shallow active layer depth and less wetlands extent, the water stock in surface pool is less and its seasonal variation does not even balance the runoff term. It argues for the more important role of surface runoff in water balance of permafrost affected areas comparing to nearby southern permafrost free areas, and less regulation effect of wetlands on basin water outflow. In Zakharova et al., (2012) we demonstrated that the wetlands of northern watersheds retain about 37% of meltwaters and its holding capacity increased since 1990th. If suggest that it is related to permafrost melting (increasing of active layer depth or retrieving to north) we can expect in future a strong change in water balance of these permafrost-affected territories and its evolution to the budget structure of southern watersheds of Surgut Polesye.

7. Conclusions and discussions

Using the ENVISAT RA-2 radar altimetry data for 2002-2010 we have estimated seasonal variability of the wetlands in the 21 mid-size watersheds of the Western Siberia. The knowledge of the wetlands water regime (extent and water level) is important for hydrological modeling and for better understanding of their role in the water and carbon cycle.

23

By defining and applying a threshold (17 dB value) approach for the backscatter data in Ku (13.6 GHz, Ice2 retracker) we have discriminated open water and wetlands and then, in order to compare watersheds with different size, we have calculated the ratio (%) of altimetric observations classed as open water to the total number of observations for each cycle and each watershed to serve as an equivalent of wet zones extent.

According to the seasonal extent of wet zones the watersheds can be divided into 3 groups and one sub-group. First group represent watersheds located on the south and south-west of the plain. They have a relatively pronounced spring extent maximum and low summer values (this group includes also a subgroup where second autumnal maximum is observed). Second group has very high fraction of wet zones with low seasonal changes in extent and small but discernible second peak in September. The watersheds of third group located in the permafrost-affected zone have two high maxima: one in spring and another in the early autumn, and well pronounced summer minima.

Evapotranspiration, along with precipitation, is an important parameter that determines the wet zone regime in the Western Siberia. If the main spring maximum of wet zone extent occurs due to snowmelt, the second autumnal maximum takes place soon after precipitation exceeds evapotranspiration. In the southern part of the Western Siberia, where ET is high (P-ET is minimal 25 mm), wetlands in summer are getting dry and water table is low, thus no second peak of wet zones extent is observed. In the northern parts wetlands are less drained in summer, thus less water retention capacity and appearance of the autumnal peak.

The

precipitation

excess

results

in

water

accumulation

on

the

watershed

and

replenishment/saturation of surface matrix before freezing. It has an important regional hydrological consequences as it means that flood runoff initial conditions (soil water storage) are not highly different from year to year and the spring runoff rate will depend mostly on volume of snow storage and melting rate.

The difference in melting dates amounts up to two months between southern (mid-March) and northern watersheds (beginning and mid-May). The spatial variation of dates of melting start has a functional linear relation with latitude and changes on 8 days per each 2° of latitude. The date of maximal wet zones extent is less related to the latitude (R2=0.68) as it depends also on basin morphology (relief, lake and bog fraction etc), but it has the same spatial gradient (8days/2° latitude).

24

Timing of the second (autumnal) maximum of wet zones extent differs by one month between north and south. The freezing dates of the wetlands (defined as time when the ratio of wet zones decreases to values close to 0%). show that they are less dependent on climate conditions than on surface properties, notably the presence of lakes and bogs and its heat storage capacity. While snow accumulates on forested areas, on lakes and bogs with low vegetation it is blown away and its accumulation takes more time (backscatter values are elevated until December-January). As a result, use of 20 dB threshold in winter for watershed in Surgut Polesye region potentially leads to some difficulties in discrimination between wet zones and land/ice. Further studies involving complementary data (visible and passive microwave range) are needed to regional tuning of ice setup timing using a threshold that is specific for watersheds with high lake fraction.

To quantitatively estimate the contribution of the floodplain to the total estimates of the fraction of the wet zones, we have done an analysis for the biggest of the 21 selected watersheds with largest floodplain - the Taz river. The Taz river floodplain occupies 13.5 % of the total watershed area, but its contribution to the ENVISAT estimates of wet zones extent ratio in snow-free season is about 4 to 8%. As for other rivers of the Western Siberia (with even smaller floodplain extent) its contribution is expected to be even less, for the small and medium size rivers of the Western Siberia the seasonal variation of the wet zones ratio on the basin scale is entirely related to the processes taking place in the wetlands of interfluvial areas.

Seasonal water level variation in various wetland landscapes from ENVISAT data complement the wetlands extent estimates to provide correct estimate of the water budget of wet zones. Two flat bog zones have been studied: Surgut Polesye with no modern permafrost and Nadym region located in the zone of discontinuous permafrost.

The highest seasonal water level variability from 1 to 1.5 meters is found for middle and big size lakes in Surgut Polesye. In the big raised bogs of this region the seasonal amplitude is smaller and does not exceed 50 cm. For lakes of Nadym region this value is less than 0.7 m and for flat bogs in the south of the Nadym region water level variability is of order of 30 cm. On big drained lakes (khasyrey) largely developed in this region the seasonal water level variation is 50-70 cm but can be as high as 1 m depending on the banks height and morphology of khasyrey's drainage system.

Correct estimation of the seasonal variability of surfacial water level of different wetland types represents a certain methodological challenge, as surface here is usually a mixture of land and water

25

in the footprint. While there exist some in situ data on temporal water changes of water table in bogs, but no estimates have been made of surface water level to compare with our satellite-derived estimations. Nevertheless, radar altimetry provides the first ever results of seasonal water level variability over the boreal wetlands. Our assessment of the first results of the recent radar altimetry mission SARAL/AltiKa (launched in March 2013 on the same orbit as ENVISAT) shows promising potential of AltiKa for continuation of wetlands extent and water level monitoring. Future satellite missions, such as SWOT (Fu et al., 2012) will provide more precise evaluation of both temporal and spatial water level variability.

Using our ENVISAT estimates of the seasonal amplitude of the water level and historical data of the bog and lake fraction it is possible to evaluate the seasonal changes in water storage in surfacial pool of the watersheds. In the three basins of non-permafrost zone the seasonal water storage is about 30 km3, in northern basins it equals to 39 km3 (or 400 mm and 150 mm, correspondingly).

While for the watersheds of non-permafrost zone the seasonal variation of the water storage in wetlands is 1.5-2.5 times higher than water runoff and enough to compensate the evaporation, in northern watersheds, the water stock in surface pool is smaller and its seasonal variation does not even balance the runoff term. This demonstrates more important role of surface runoff in water balance of permafrost-affected areas (comparing to southern permafrost-free areas), and smaller regulation effect of wetlands on basin water outflow. Our previous results show that the wetlands of northern watersheds retain about 37% of meltwaters and its holding capacity have increased since 1990th (Zakharova et al., 2011). If this is related to changes in permafrost (increase of active layer depth or northward propagation of permafrost extent limits), we can expect in future a significant change in water balance of these permafrost-affected territories when they will become more similar to the southern watersheds of Surgut Polesye.

Acknowledgements This research has been done in the framework of the Russian-French cooperation GDRI "CARWET-SIB" and "Franco-Siberian Center for Research and Education", French ANR "CLASSIQUE", PNTS "Permafrost" and CNES TOSCA SWOT projects, Russian FZP 1.5 and FP7 MONARCH-A projects. Altimetric data are provided by the Center for Topographic studies of the Oceans and Hydrosphere (CTOH) at LEGOS. We thank Maria Kolmakova for help with processing altimetric data.

26

References Andersen O. B., P. E. Krogh, P. Bauer-Gottwein, S. Leiriao, R. Smith, P. Berry, Terrestrial Water Storage from GRACE and Satellite Altimetry in the Okavango Delta (Botswana), in: Gravity, Geoid and Earth Observation. International Association of Geodesy Symposia Volume 135, 2010, pp 521-526 Bartsch A., C. Pathe, W. Wagner. 2007a, Wetland mapping in the West Siberian lowlands with ENVISAT ASAR global mode, in: Proceedings of Envisat Symposium 2007, Montreux, Switzerland 23–27 April 2007 (ESA SP-636, July 2007). Bartsch Annett., Richard A. Kidd, Wolfgang Wagner, Zoltan Bartalis. 2007b, Temporal and spatial variability of the beginning and end of daily spring freeze/thaw cycles derived from scatterometer data. Remote Sens Environ 106 (2007), pp. 360–374. Becker M., W. LLovel, A. Cazenave, A. Güntner, J-F. Crétaux. Recent hydrological behavior of the East African great lakes region inferred from GRACE, satellite altimetry and rainfall observations. Comptes Rendus Geoscience, Volume 342, Issue 3, March 2010, Pages 223– 233 Berry P. A. M., J. D. Garlick, J. A. Freeman, E. L. Mathers, Global inland water monitoring from multi-mission altimetry, Geophys Res Lett, Volume 32, Issue 16, August 2005 Birkett C. M., B. Beckley, Investigating the Performance of the Jason-2/OSTM Radar Altimeter over Lakes and Reservoirs, Marine Geodesy, Volume 33, Supplement 1, 2010 Special Issue: OSTM/Jason-2 Calibration/Validation, pages 204-238. Birkett C.M., L.A.K. Mertes, T. Dunne, M.H. Costa, M.J. Jasinski. 3Surface water dynamics in the Amazon Basin: Application of satellite radar altimetry". J Geophys Res, 107 (2002) http://dx.doi.org/10.1029/2001JD000609 Birkinshaw S. J.,, G. M. O'Donnell, P. Moore, C. G. Kilsby, H. J. Fowler, P. A. M. Berry, Using satellite altimetry data to augment flow estimation techniques on the Mekong River, Hydrol Process Volume 24, Issue 26, pages 3811–3825, 30 December 2010 Crétaux J.-F., C. Birkett , Lake studies from satellite radar altimetry. Comptes Rendus Geoscience Decharme, B., R. Alkama, F. Papa, S. Faroux, H. Douville, and C. Prigent (2010), Global off-line evaluation of the ISBA-TRIP flood model, Climate Dynamics, doi:10.1007/s00382-0111054-9, Frappart F., F. Seyler, J-M Martinez, J. Leon, A. Cazenave, Floodplain water storage in the Negro River basin estimated from microwave remote sensing of inundation area and water levels, Remote Sens Environ, Volume 99, Issue 4, 15 December 2005, Pages 387–399 Frappart F.,. F. Papa, A. Güntner, S. Werth, G. Ramillien, C. Prigent, W. B. Rossow, and M.-P. Bonnet. Interannual variations of the terrestrial water storage in the Lower Ob' Basin from a

27

multisatellite approach. Hydrol Earth Syst Sci, 14, 2443-2453, 2010, www.hydrol-earthsyst-sci.net/14/2443/2010/, doi:10.5194/hess-14-2443-2010 Frappart, S. Calmant, M. Cauhopé, F. Seyler, A. Cazenave. Preliminary results of ENVISAT RA-2derived water levels validation over the Amazon basin. Remote Sens Environ, 100 (2006), pp. 252–264 Frappart, F.; Minh, K. Do; L'Hermitte, J.; Cazenave, A.; Ramillien, G.; Le Toan, T.; MognardCampbell, N. (2006) Water volume change in the lower Mekong from satellite altimetry and imagery data, Geoph J Int, Volume 167, Issue 2, pp. 570-584. Fu, L-L., Alsdorf D., Morrow R., Rodriguez E., Mognard N. (Eds). SWOT: The Surface Water and Ocean Topography Mission. Wide-Swath Altimetric Measurement of Water Elevation on Earth. NASA-JPL-CalTech, Pasadena, California, Feburary 2012, 228 pp. Hydro1K, USGS Earth Resources Observation and Science (EROS) Center http://eros.usgs.gov, accessed 2010) Ivanov E.K., Novikov S.M. (Eds). Bogs of the Western Siberia - their structure and hydrological regime.Leningrad, Gidrometeoizdat, 1976, 448 pp. (In Russian). Ivanov K.E. ―Water exchange in bog landscapes‖. Gidrometeoizdat, Leningrad, 281 pp. (In Russian) Kellner E., Halldin S., Water budget and surface-layer water storage in a Sphagnum bog in central Sweden. Hydrol Process, 2002., Vol.16., p.87-103. Kirpotin S, Berezin A., Bazanov V., Polishchuk Yu., Vorobiov S., Mironycheva-Tokoreva N., Kosykh N. Volkova I., Dupre B., Pokrovsky O., Kouraev A.V., Zakharova E., Shirokova L., Mognard N., Biancamaria S., Viers J., Kolmakova M. Western Siberia wetlands as indicator and regulator of climate change on the global scale. Int J Environ Studies, 2009 Vol. 66, No. 4, August 2009, 409–421 Kondratyev K.A., V.A.Kudryavtsev, Map of geocryological regioning of USSR. In Kudryavtsev V.A. (editor) , Merzlotovedeniye Moscow, 1981, Moscow State University publishing, p 240. Kotlyakov V.M. (Ed). World atlas of snow and ice resources ("Atlas snezhno-ledovykh resursov mira"), Moscow Russian Academy of Sciences, Institute of Geography,1997 Kouraev A.V., M.N.Shimaraev, P.I. Buharizin, M.A.Naumenko, J-F Crétaux, N.M. Mognard, B. Legrésy, F. Rémy, Ice and snow cover of continental water bodies from simultaneous radar altimetry and radiometry observations. Survey in Geophysics - Thematic issue Hydrology from space 2008 DOI 10.1007/s10712-008-9042-2 Kouraev A.V., Zakharova E.A., Polikarpov I., Al-Yamani F., Crétaux J-F. Hydrological regime of the Euphrates-Tigris river system using satellite altimetry. The Arabian Seas International

28

Conference on Science and Technology of Aquaculture, Fisheries and Oceanography, 10 13 February, 2007 State of Kuwait Legrésy, B., & Rémy, F. (1997), Altimetric observations of surface characteris-tics of the Antarctic ice sheet. J Glaciology, 43(144), pp. 265-275. Lehner, B. and Döll, P. (2004): Development and validation of a global database of lakes, reservoirs and wetlands. Journal of Hydrology 296/1-4: 1-22. MrSID

Image

Server,

"NASA

Earth

Science

Application

Directorate,"

https://zulu.ssc.nasa.gov/mrsid/, accessed 2012 Papa F., C. Prigent, and W. B. Rossow. 2007, Ob’ River flood inundations from satellite observations: A relationship with winter snow parameters and river runoff. J Geophys Res, Vol. 112, D18103, doi:10.1029/2007JD008451, 2007. Papa, F., Prigent, C., Rossow, W. B., Legresy, B. and Remy, F. 2006, Inundated wetland dynamics over boreal regions from remote sensing: the use of Topex-Poseidon dual-frequency radar altimeter

observations.

Int

J

Remote

Sens,

27:21,4847

—

4866,

DOI:

10.1080/01431160600675887 Pathe C., W. Wagner, D. Sabel, M. Doubkova, and J. B. Basara, Using ENVISAT ASAR Global Mode Data for Surface Soil Moisture Retrieval Over Oklahoma, USA. IEEE Trans Geosci Rem Sens, Vol. 47, No. 2, February 2009 Pereira-Cardenal , N. D. Riegels, P. A. M. Berry, R. G. Smith, A. Yakovlev, T. U. Siegfried, andP. Bauer-Gottwein. Real-time remote sensing driven river basin modeling using radar altimetry. Hydrol Earth Syst Sci, 15, 241–254, 2011 Popov A.I. Regional cryolythology. Moscow State University Publishing, 1989, 255 pp. Sergeev E.M. Ingineering Geology of the USSR.Vol.2. Western Siberia. Moscow State University Publishing, 1976. 496 pp. Sokolov A.A. Hydrography of the USSR (―Gidrografiya SSSR‖). Gidrometizdat, Leningrad, 1952. (In Russian). Troitskaja Yu., S.A. Lebedev, A.G. Kostianoy, A.V. Kouraev, E.A. Zakharova, I. Krylenko, G.V. Rybyshkina, I.A. Soustova, A.A. Panutin. "Rivers and Reservoirs in boreal zone of Eurasia" / In "Inland water altimetry", Editors: J. Benveniste, S. Vignudelli, A. Kostianoy, Springer, 2013 (submitted) Vodogretskiy V.E. (editor). Resources of Surface Waters. Vol 15. Altay and Western Siberia, Issue 3. Lower Irtysh and Lower Ob'. Gidrometeoizdat, Leningrad, 1973, Volume 338, Issues 14– 15, November–December 2006, Pages 1098–1112. Zakharova E.A., A.V. Kouraev, S. Biancamaria, M.V. Kolmakova, N.M. Mognard, V.A. Zemtsov, S.N. Kirpotin, B. Decharme. Snow cover and spring flood flow in the northern part of

29

Western Siberia (the Poluy, Nadym, Pur and Taz rivers). J Hydromet, 2011, Volume 2, p. 1498-1511, DOI: 10.1175/JHM-D-11-017.1 Zakharova E.A., Kouraev A.V., Kolmakova M.V., Mognard N.M., Zemtsov V.A., Kirpotin S.N. The modern hydrological regime of North-Western Siberia from in situ and satellite observations. Int J Environ Studies, Vol. 66, No. 4, August 2009, 447–463 Zakharova E.A., Kouraev A.V., Polikarpov I., Al-Yamani F., Crétaux J-F. "Radar altimetry for studies of large river basins: hydrological regime of the Euphrates-Tigris rivers". Second Space for Hydrology Workshop "Surface Water Storage and Runoff: Modeling, In-Situ data and Remote Sensing" Geneva (Switzerland), 12-14 November 2007

30

TABLES Table 1. Selected watersheds. Watershed surface, ENVISAT coverage (number of tracks and number of 18Hz ENVISAT observations for each watershed) No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Name Taz Vakh Pur Tromyugan Nadym Poluy Shuchya Kazym Nazym Lyamin Pim Tym Ket' Vasyugan Bolshoy Yugan Severnaya Sos'va Konda Tavda Tura Om' Inner Watershed

Watershed, Number of Number of km2 tracks observations 139600 31 23853 76503 26 11097 103329 27 17471 51820 22 7956 54152 16 9192 19557 13 3139 11831 14 1794 34634 17 4663 11169 7 1473 15528 9 2423 12055 7 1830 30902 19 4296 94828 30 11442 63092 18 7756 33366 12 4444 91149 21 10103 70592 21 9387 86629 25 8873 78385 22 8394 62734 24 7974 115524 20 13486

Table 2. Estimation of seasonal water storage variation in wetlands of six watersheds.

H - height variability, W - volume variability River

Area,

% Lake

H lake,m

% Bogs

H bog,m

W,km3

km2 Surgut Polesye Lyamin Pim Tromyugan Northern region Pur Taz Nadym

15528 12055 51820

18 22 20

55 60 45

1.2 1.2 1.2

0.5 0.5 0.5

3.2 2.7 12.2

103329 139600 54152

12 1 10

45 10 45

0.7 0.7 0.7

0.3 0.3 0.3

28.2 33.0 14.5

31

FIGURES

Ob' estuary

II N N SS

60° 60° E E U UN N TT A A

Salekhard

77

Ob' Ob'

Pangody Nadym 33

66

M MO O

55 88 SSib skiy ibirirsk iyee U Uva ly valy 10 10 11 11 99 44

16 16 RA U UR A LL SS

Taz estuary 90°E 90°E

11 Yenisey Yenisey 22

Surgut

12 12

Ob' Ob'

17 17

18 18

15 15

KhantyMansiysk

19 19

60°N 60°N

13 13

14 14

Irtysh Irtysh Tomsk

Russia Russia

20 20 Novosibirsk

Kazakhstan Kazakhstan

21 21 50°N 50°N

Figure 1. Overview map of the study region. For the names of the watersheds indicated as numbers, see Table 1. Watersheds contours are based on data from Hydro1K (Hydro1K web site, 2010) and manual adjustment and verification using GlobCover2000 Landsat mosaics (MrSID image server, 2012). polygon polygon bogs bogs

Continuous Continuous permafrost permafrost

polygonal polygonal bogs bogs

77

77

66

11

33

55

Discontinuous Discontinuous permafrost permafrost

66

88 16 16

18 18

10 11 11 99 10

17 17

flat flat m mound ound bogs bogs

11

33

55

big big m mound ound bogs bogs

88 16 16

44

22 12 12

15 15 14 14

Zone Zone of of constant constant seasonal seasonal soil soil freezing freezing 13 13

18 18

10 11 11 99 10

17 17

44

22 12 12

15 15

13 13

14 14

19 19

raised raised m mound ound oligotrophic oligotrophic bogs bogs

19 19

20 20

concave concave eutrophic eutrophic reed reed or or brackish brackish fens fens

21 21

flat flat eutrophic eutrophic or or m mesotrophic esotrophic fens fens

20 20

21 21

a)

b)

Figure 2. Permafrost coverage after (Kondratyev and Kudryavtsev, 1981) (a) and bogs types in the study area after (Ivanov and Novikov, 1976) (b)

32 10

Month

11 9 7 5 3 1 -10 0 10 20 30 40

Backscatter, dB

20 18 16 14 12 10 8 6 5 4 3 2 1 0

8

Frequency (%)

13

6

4

2 B

A

Forest-dominated landscape

Lake-dominated landscape

0 -12

-8

-4

0

4

a)

8

12

16

20

24

28

32

36

40

44

b)

Backscatter value, dB

Figure 3. Statistical parameters of ENVISAT 18 Hz backscatter values in Ku-band (Ice2 retracker) for 2002-2010. a) Map of seasonal probability distribution (expressed as % of the total monthly number of observations) for the Taz watershed. To get better representation of seasonal sequence, January and February are plotted twice (Month 13 - January, Month 14 - February). b) Frequency (%) histograms of backscatter values (June-August) for forest-dominated (220 observations) and lake-dominated (98 observations) landscapes. Line A - 95% frequency value for forest- and line B 5% frequency value for lake-dominated landscapes.

%

90 80

14 dB

70

16 dB 17 dB

60

18 dB

50

20 dB

40 30 20 10 Month

0 1

2

3

4

5

6

7

8

9

10

11

12

Figure 4. Average values (2002-2010) of seasonal variability of wet zones extent (in %) using different thresholds for the Taz watershed.

33

77

66

11

33

55

88

16 16

44

10 11 11 99 10

18 18

17 17

22 12 12

15 15

13 13 14 14 19 19

20 20

1 1A 2 3

21 21

Figure 5. Watershed type based on seasonal variability of wet zones extent

40 40 20 20

0 -20

100

50

100

150 200 250 Day of year

300

100

60

60

40 40 20 20

0

0

-20 0

50

100

150 200 250 Day of year

300

40 40 20 20

350

0 -20 0

80

80

60

60

350

b)

Tym

80

80

0

100

Wet zones fraction (%)

0

100

Rain (black) and Rain-ET (gray), mm

60

c)

Tromyugan

50

100

150 200 250 Day of year

300

350

d)

Pur

100 80

80

60

60

40 40 20 20

0

0


60

100


80

80

0


100


a)

Vasyugan



100

-20 0

50

100

150 200 250 Day of year

300

350

Figure 6. Average values (2002-2010) of seasonal variability of wet zones extent (in %, black dots), precipitation in mm, black line and difference between liquid precipitation and evapotranspiration (grey line, with negative values filled) after (Decharme et al., 2010). Examples for four types: a) Vasyugan, group 1, b) Tym, group 1a, c) Tromyugan group 2, and d) Pur, group 3.

34 77

77

66

11

33

55

66

88

16 16

18 18

88

16 16 44

10 11 11 99 10

17 17

44

10 11 11 99 10

22 18 18

12 12

15 15

11

33

55

17 17

22 12 12

15 15

13 13

13 13

14 14

14 14

19 19

19 19

20 20 Beginning of May End of April Mid of April Beginning of April End of March Mid-March

20 20 Mid June Beginning of June End of May Mid May Beginning of May End of April

21 21

Timing of the beginning of spring flooding

21 21

Timing of the maximal spring flooding on the watersheds

Fig. 7 Dates of melting start (a) and of maximal extent of wet zones (b).

77

77

66

11

33

55

66

88

16 16

18 18

17 17

88

16 16 44

10 11 11 99 10

10 11 11 99 10

22 18 18

12 12

15 15

11

33

55

17 17

44

22 12 12

15 15

13 13

13 13

14 14

14 14

19 19

19 19

20 20

Mid October Beginning of October End of September Mid September

Autumnal peak

20 20

21 21

February January End of December Mid December Beginning of December Mid November

21 21

Winter start

Figure 8. Dates autumnal peak of wet zones extent (a) and of full freezing (b).

35

a)

b)

Figure 9. Taz river watershed: a) main river channels and lakes, b) upper floodplain

AB

100

C

DE

Difference (%) between whole watershed ratio and that of excluding floodplain

Wet zones extent (%)

a) 80 Floodplain 60

40 Whole watershed

20

0

b)

8 6 4 2 0 J 0

F 1

M 2

A 3

M 4

J 5

J 6

A 7

S 8

O 9

N 10

D 11

12

Month

Figure 10. Seasonal variability (%) of wet zones extent (a) for the whole watershed (black crosses) and for the floodplain only (open circles), and difference (%) between wet zones extent for the whole watershed and that excluding floodplain (b, black circles). Lines A-E denote specific events in the wet zones extent variability.

36 lake_Surgut bog_Surgut lake_Nadym

Height, m

47

46

45

44 0

90

180

270

360

DOY

Figure 11. Seasonal (15-days smoothing) water level variability (2002-2010) for Surgut and Nadym region. For the region location see Figure 1.

Highlights (for review)

21 watersheds of the Western Siberia are studied using radar altimetry Seasonal variability of wetlands extent and timing of specific phases are analyzed Seasonal amplitude of water level variation in lakes and bogs is estimated

Fig01 Click here to download high resolution image

Fig02a Click here to download high resolution image

Fig02b Click here to download high resolution image












