The
International Forestry Review
Contents PAPERS 1
Analysis of the effective stakeholders’ involvement in the development of National Forest Programmes in Europe J. BALEST, M. HRIB, Z. DOBṠINSKÁ and A. PALETTO
13
Potentials of REDD+ in supporting the transition to a Green Economy in the Congo Basin K.E. ENONGENE and K. FOBISSIE
29
Evaluating the impacts of plantations and associated forestry operations in Africa—methods and indicators V. INGRAM, E. VAN DER WERF, E. KIKULWE and J.H.H. WESSELER
44
Higher forestry education in Kenya: bridging the gap between educational training and job market competencies S. RAMCILOVIC-SUOMINEN, Y. PUENTES RODRIGUEZ, B. KIRONGO and S. PITKÄNEN
56
The potential of rubber and acacia plantations for forest carbon stocks in Malaysia J. RATNASINGAM, K. THIRUSELVAM and F. IORAS
68
Decentralisation policy as recentralisation strategy: forest management units and community forestry in Indonesia M.A.K. SAHIDE, S.SUPRATMAN, Ahmad MARYUDI, Y.-S. KIM and L. GIESSEN
78
Financing sustainable forest management in developing countries: the case for a holistic approach B. SINGER
96
Exploring the impact of social norms and perceptions on women’s participation in customary forest and land governance in the Democratic Republic of Congo— implications for REDD+ L. STIEM and T. KRAUSE
110
Roles of forests in food security based on case studies in Yunnan, China S. TAKAHASHI and L. LIANG
123
BOOK REVIEWS 133
THE INTERNATIONAL FORESTRY REVIEW Vol. 18 (1), 2016
What drives consumption of wood energy in the residential sector of small cities in Europe and how that can affect forest resources locally? The case of Bragança, Portugal J.C. AZEVEDO, M.C. FERREIRA, L.F. NUNES and M. FELICIANO
EDITOR: A.J. POTTINGER International Forestry Review (print) ISSN 1465-5498 International Forestry Review (online) ISSN 2053-7778 PUBLISHED BY THE COMMONWEALTH FORESTRY ASSOCIATION
Vol.18(1), 2016 www.cfa-international.org
The International Forestry Review Editor Alan Pottinger
[email protected]
Chairman of the Editorial Advisory Board Jeff Sayer James Cook University, Australia
[email protected]
Editorial Board Fred Babweteera Budongo Conservation Field Station, Uganda
John Innes University of British Colombia, Canada
Eberhard Bruenig University of Hamburg, Germany
Peter Kanowski Australian National University, Australia
Neil Byron Australian Productivity Commission Melbourne, Australia
Roger Leakey James Cook University, Australia
José Joaquin Campos CATIE, Costa Rica Jim Carle Independent, New Zealand Ebby Chagala Kenya Forestry Research Institute (KEFRI), Nairobi, Kenya Ben Chikamai Kenya Forestry Research Institute (KEFRI), Kenya Mafa Chipeta FAO, Rome, Italy Jonathan Cornelius World Agroforestry Centre (ICRAF), Peru Verina Ingram CIFOR, Indonesia and LEI Wageningen UR, The Netherlands
Bill Mason Forestry Research, Edinburgh, UK Jack Putz University of Florida, USA Lee Su See Forestry Research Institute Malaysia, Malaysia Changyou Sun Mississippi State University, USA Jerome Vanclay Southern Cross University, Australia Claire Williams National Evolutionary Synthesis Center & Forest History Society, Durham NC USA
Guidelines for authors Please send the Summary of the paper to the Editor at
[email protected]. If it is considered suitable for consideration you will be asked to send the complete manuscript. Manuscripts submitted for consideration must conform to the following points. Any deviation will result in the manuscript being returned to the author. COMPOSITION • Contributions must be original* and not have been submitted for publication elsewhere. (Note: Plagiarism is evaluated by use of electronic software. For more information on what constitutes plagiarism, and why it is important please click here). • The text, excluding tables, references and appendices, should not exceed 7000 words, although exceptions may be permitted in special cases. • A SUMMARY of not more than 150 words must be supplied, together with 5 keywords. • All spelling must conform to UK/international English. • The layout of the text and style of table and figure legends and references must conform to that of the International Forestry Review. This means: o Main title in Arial, Text in Times New Roman o The hierarchy of headings is: CAPITALS, bold lower case, italics lower case. TEXT • • • • • • •
Manuscripts should be produced in Microsoft Word, written in Times New Roman typeface (size 12 pt), with single row spacing, left justification and without hyphenation. The manuscript should be written in the passive voice, e.g. ‘The experiment was carried out’ is correct; ‘We carried out the experiment’ is incorrect. Manuscripts should be submitted with UK English spellings. Make sure that a spell check with UK English spellings is used prior to submission. Paragraphs should not be separated by any additional line spacing. The first paragraph in a section should not be indented. The first line of each subsequent paragraph should be indented to 1.27cm. Use quotation marks (“ “) only around quotations or titles. Do not use them to highlight or emphasise text. et al. is correct, et al is incorrect. The document should be saved as a Microsoft Word file with *.doc file extension.
TABLES • Tables should clear and simple, with a maximum of 8 columns. (Note that tables have limited space in the final layout and therefore the reduction of font size to a minimum of 8 pt during typesetting should be taken into account when preparing tables). • Headings and other texts in table cells should to be concise. Tables including captions, legends and footnotes should be written in Microsoft Word (in exceptional cases Microsoft Excel can be used following agreement of the Editor). • Tables should be saved in a separate file (not as an integral part of manuscript) and the manuscript text should contain the reference to the position of the table (or figure) in brackets. NUMBERS • Numbers greater than 999 should be written with appropriate spaces and without commas, e.g 10 000 is correct, 10,000 and 10000 are incorrect
Peter Wood Independent, Oxford, UK
FOOTNOTES • References to footnotes in the main text should be marked with arabic numerals in superscript form.
The Editor, International Forestry Review, The Crib, Dinchope, Shropshire SY7 9JJ, UK Telephone: +44 (0)1588 672868 Email:
[email protected], Web: www.cfa-international.org Cover photo: Zebra in Acacia woodland, Hluhluwe–Imfolozi Park, South Africa (Credit: John Innes)
SCIENTIFIC NAMES • The complete scientific name (genus, species, authority and, where appropriate, cultivar) should be cited at the first occasion of its mention and written in italics (authority in normal type). If vernacular names are used, they must be accompanied by the correct scientific name at first use. CITATIONS • For text citations, papers should be referred to as (Smith 1998) and papers by the same author in the same year should be distinguished by lettering in sequence (1998a, 1998b, etc.). Where papers are written by a single author or two authors their names should be cited. If three or more authors are involved the first name should be listed followed by ‘et al.’, e.g. (Smith et al. 2002). • Citations should be separated by a comma, not a semi-colon, i.e. (Johnstone 2003, Smith 2002, Smith et al. 2002). • Multiple citations by different authors should be listed alphabetically, e.g. (Brown 2001, Jones 2003 and Smith 2002). • Multiple citations by the same author should conform to the following format (Brown 2001, 2005, 2009). QUOTATIONS • Direct quotations from papers or books should be referenced in the format (Smith 1998: 23-24). REFERENCES • At the end of the paper, the list of references must be arranged in alphabetical ordering without serial numbering. • References should be formatted with a ‚hanging‘ indent. • There should be no additional line spacing between individual references. • The following standard forms of citation must be used: Author’s name, all authors’ initials to follow surname, journal and book titles in italics. Volume number in bold. Second and subsequent lines should not be indented. For example: Journal paper LÄHDE, E., LAIHO, O., NOROKORPI, Y. and SAKSA, T. 1999. Stand structure as the basis of diversity index. Forest Ecology and Management 115 (2/3): 213-220.
TITLE • The title and author’s details should be in the following format Mutually beneficial company-community partnership in ensuring its long-term viability: emerging lessons from Indonesia A.A. NAWIR and L. SANTOSO Center for International Forestry Research, Jl. CIFOR, Situ Gede, Sindang Barang, Bogor 16680, Indonesia Email:
[email protected] and
[email protected]
Michael J. Wingfield Forest and Agricultural Biotechnology Institute (FABI), South Africa
Contact
•
GRAPHS AND FIGURES • Graphs, diagrams and other figures should be prepared in Microsoft Excel or in Microsoft PowerPoint and saved in separate files. Graphs and diagrams should be drawn in 2-D form (not in 3-D spatial form) and single columns or circle sectors (in case of pie diagrams) should be filled with graycscale colours (not with colour filling or by using of automatic hatching). The line weight of axes and other lines and also the size of used letters or numerals should be appropriate to the final size reduction of diagrams during layout of final magazine page (maximum width of finally reduced diagram is either 8 or 18 cm depending on whether it fits one or two columns). • Complex images such as maps (drawn in line draw or grayscale format) should be prepared in one of the following: CorelDraw, Adobe Illustrator, Adobe Photoshop, Macromedia Freehand or similar in which he finished images can be saved as *.eps (Encapsulated PostScript) file format. B&W photographs should be submitted as *.tif image file format with a resolution of at least 300 dpi. Colour
photographs may be submitted only after agreement with the Editor. Colour images should be saved in CMYK colour format, as *tif file format and with a resolution of at least 300 dpi. Titles should be in the format: TABLE X Title of table
Paper or chapter in proceedings SMITH, W.J. 2001. Selection of tree species for arid environments. In: BLACKBURN, J.W. (ed.) Multipurpose trees and shrubs for fuelwood and agroforestry. CNRD Monograph No4. 366 pp. Book PHILIP, M.S. 1994. Measuring trees and forests. 2nd edition, CAB International, Wallingford, England. 310 pp. •
•
Unnecessary use of capitals should be avoided. For example HOLMGREN, J., JOYCE, S., NILSSON, M. and OLSSON. H. 2000. Estimating Stem Volume and Basal Area in Forest Compartments by Combining Satellite Image Data with Field Data. Scandinavian Journal of Forest Research 15: 103–111. Is incorrect. HOLMGREN, J., JOYCE, S., NILSSON, M. and OLSSON. H. 2000. Estimating stem volume and basal area in forest compartments by combining satellite image data with field data. Scandinavian Journal of Forest Research 15: 103–111. Is correct. Websites should only be quoted in isolation where hard copies are not available.
ACKNOWELDGEMENT • It is necessary for authors to acknowledge suggestions made by referees with a simple statement such as ‘The valuable suggestions made by anonymous referees is gratefully acknowledged’. SUBMISSION • Manuscripts offered for publication should be submitted by email to
[email protected] REFEREEING • Contributions will be referred to at least two expert referees. Authors will be consulted if the paper is considered suitable but alterations are thought desirable. After alterations have been agreed and incorporated, the paper will be considered final. ASSISTANCE WITH PUBLICATION • For assistance with preparing manuscripts visit our Online guide to scientific writing at www.cfa-internationl,.org/OGSWintro.html * Original means that the paper, or any close derivative of it, has not been published previously in any form, including on the internet.
International Forestry Review Vol.18(1), 2016 i
Contents PAPERS What drives consumption of wood energy in the residential sector of small cities in Europe and how that can affect forest resources locally? The case of Bragança, Portugal J.C. AZEVEDO, M.C. FERREIRA, L.F. NUNES and M. FELICIANO
1
Analysis of the effective stakeholders’ involvement in the development of National Forest Programmes in Europe J. BALEST, M. HRIB, Z. DOBṠINSKÁ and A. PALETTO
13
Potentials of REDD+ in supporting the transition to a Green Economy in the Congo Basin K.E. ENONGENE and K. FOBISSIE
29
Evaluating the impacts of plantations and associated forestry operations in Africa—methods and indicators V. INGRAM, E. VAN DER WERF, E. KIKULWE and J.H.H. WESSELER
44
Higher forestry education in Kenya: bridging the gap between educational training and job market competencies S. RAMCILOVIC-SUOMINEN, Y. PUENTES RODRIGUEZ, B. KIRONGO and S. PITKÄNEN
56
The potential of rubber and acacia plantations for forest carbon stocks in Malaysia J. RATNASINGAM, K. THIRUSELVAM and F. IORAS
68
Decentralisation policy as recentralisation strategy: forest management units and community forestry in Indonesia M.A.K. SAHIDE, S. SUPRATMAN, Ahmad MARYUDI, Y.-S. KIM and L. GIESSEN
78
Financing sustainable forest management in developing countries: the case for a holistic approach B. SINGER
96
Exploring the impact of social norms and perceptions on women’s participation in customary forest and land governance in the Democratic Republic of Congo—implications for REDD+ L. STIEM and T. KRAUSE
110
Roles of forests in food security based on case studies in Yunnan, China S. TAKAHASHI and L. LIANG
123
BOOK REVIEWS 133
International Forestry Review Vol.18(1), 2016 1
What drives consumption of wood energy in the residential sector of small cities in Europe and how that can affect forest resources locally? The case of Bragança, Portugal J.C. AZEVEDO1, M.C. FERREIRA2, L.F. NUNES1 and M. FELICIANO1 1 2
Centro de Investigação de Montanha (CIMO), Instituto Politécnico de Bragança, Campus de Santa Apolónia, 5300-253, Bragança, Portugal Associação Florestal do Vale do Douro Norte, Casa Florestal de Mascanho/Carvas, 5090-077 Murça, Portugal
Email:
[email protected],
[email protected],
[email protected] and
[email protected]
SUMMARY A study in a small city in Portugal was conducted to improve the understanding of the role of wood energy in the residential energy services in small cities in Europe, the factors affecting the use of wood consumption in households, and how changes in consumption drivers affect forest resources. The hypothesis that small cities in Europe have wood energy consumption much larger than expected was analysed based on survey data collection. Drivers were analysed through statistical modelling. Wood, used in 42% of households, represented 43% of the city’s final energy consumption. The probability of wood energy use depended of resident’s age, construction type, area and year. The amount of wood used was explained by resident’s education, construction type and age, and energy function. Changes in drivers suggest a decrease in wood demand in the near future although new energy products and changes in energy use can balance this trend. Keywords: firewood, energy survey, wood energy use, wood energy consumption modelling, forest sustainability
Quels sont les moteurs de la consommation d’énergie du bois dans le secteur résidentiel de petites villes en Europe et comment cela peut affecter les ressources forestières au niveau local? Le cas de Bragança, Portugal J.C. AZEVEDO, M.C. FERREIRA, L.F. NUNES et M. FELICIANO Une étude dans une petite ville au Portugal a été menée pour améliorer la compréhension du rôle de l’énergie du bois dans les services énergétiques résidentiels dans les petites villes en Europe, les facteurs qui affectent l’utilisation de la consommation de bois dans les foyers, et comment les changements dans les moteurs de consommation affectent les ressources forestières. L’hypothèse que les petites villes en Europe ont une consommation d’énergie du bois beaucoup plus grande que prévue a été analysée sur la base de la collecte de données d’enquête. Les moteurs ont été analysés grâce à la modélisation statistique. Le bois est utilisé dans 42% des familles, représentant 43% de la consommation finale d’énergie de la ville. La probabilité d’utilisation d’énergie du bois dépend de l’âge des résidents, du type de construction, de l’espace et de l’année. La quantité de bois utilisé a été expliquée par l’éducation des résidents, le type de construction, l’âge, et la fonction de l’énergie. Les changements dans les forces motrices suggèrent une baisse de la demande de bois dans un avenir proche, bien que de nouveaux produits énergétiques et des changements dans la consommation d’énergie peuvent équilibrer cette tendance.
¿En qué se fundamenta el consumo de leña en el sector residencial de las pequeñas ciudades europeas y de qué manera afecta a los recursos forestales locales? El caso de Bragança, Portugal J.C. AZEVEDO, M.C. FERREIRA, L.F. NUNES y M. FELICIANO Este estudio, realizado en una pequeña ciudad de Portugal, se llevó a cabo para mejorar la comprensión del rol de la leña como energía doméstica en pequeñas ciudades europeas, los factores que repercuten en su uso y cómo los cambios en los hábitos de consumo afectan a los recursos forestales. La hipótesis de que en las pequeñas ciudades de Europa hay más consumo del esperado fue analizada a partir de encuestas. Los fundamentos de consumo fueron analizados mediante modelos estadísticos. La leña, con un uso doméstico del 42%, representa el 43% del consumo de energía final de la ciudad. La probabilidad de su uso depende la de edad de los residentes, tipo, superficie y año de la construcción. Y cantidad de leña utilizada depende del nivel educativo, tipo y edad de la construcción y tipo de quemador. Los cambios en los fundamentos sugieren un descenso de la demanda en un futuro próximo, aunque los nuevos productos energéticos y los cambios en el uso de los mismos pueden equilibrar esta tendencia.
2 J.C. Azevedo et al.
INTRODUCTION Wood is an important source of energy in numerous regions of the world. Globally, it corresponds to nearly 14% of the primary energy consumption (Balat and Ayar 2005) showing a tendency to increase (FAO 2010a, Lindroos 2011, Solberg et al. 2014). The significance of wood energy differs regionally. In developing countries, biomass (mainly wood) represents 35% of primary energy consumption (Balat and Ayar 2005). In Sri Lanka, however, wood energy sums up to 45% of the consumption (Rajmohan and Weerahewa 2007) and in Nepal it can be as high as 76% (Link et al. 2012). Inversely, biomass in industrialized countries represents only 3% of primary energy used (Balat and Ayar 2005). In the European Union, wood accounts for 4.8% of the gross inland energy consumption (Eurostat 2012) and in the USA it accounts for just 4% of it (Song et al. 2012b). In countries such as Latvia, Finland, and Sweden, wood energy represents, however, 27, 21, and 19% of the total energy consumption, respectively (Eurostat 2012). Most of the wood energy is used in the residential sector, either in developed or developing countries (Akther et al. 2010, Brito 1997, DeFries and Pandey 2010, Song et al. 2012a, Song et al. 2012b). In Nepal, for example, the residential sector is responsible for 99% of the total fuelwood consumption (Link et al. 2012). In the USA, households consume 23% of the energy produced from wood (Song et al. 2012b) but in Europe consumption in households is more than 60% of all wood energy produced, in some cases more than 80%, such as in Greece, the Netherlands, and France (FAO 1997). Although most of the consumption of wood energy takes place in the residential sector, the importance of wood in the total energy budget of this sector in Europe seems to be moderate. National level surveys indicate that wood represents less than 10% of the final energy consumption in households at the EU-25 level (Ó Broin 2007) even if it is high in particular countries like Portugal (25%) (INE 2010), Austria (25%), and Denmark (22%) (SEAI 2013). Trends for the transition of firewood and other energy products to cleaner and more efficient forms of energy have been described as “energy ladder” or “energy stacking” (Kroon et al. 2013). According to these hypotheses, countries or households tend to replace inefficient and more polluting energy conversion technologies by others in a sequential or simultaneous process, dependent on the level of income or status but also on the urbanization level and the price of energy sources (Masera et al. 2000, Rajmohan and Weerahewa 2007, Pachauri and Jiang 2008, Kroon et al. 2013). Available national level data for wood energy consumption in Europe (Eurostat 2009, INE 2010, SEAI 2013) indicates that the importance of wood energy in consumption is in general low, which fits in the transition theories mentioned above. In fact, in general there has been a definitive replacement of wood fuel by advanced and clean electricity or natural gas. A trend for further reduction in wood energy use in the residential sector has also been identified at the international level but at the same time it has also been assumed that the use of biomass for energy in households is often underestimated (FAO 2010b).
Wood energy consumption data available for European countries is usually aggregated at the national level referring simultaneously to large cities, small and medium sized cities, villages, and houses isolated in the countryside. Since detailed energy surveys at the level of cities of smaller units are seldom conducted, the information provided from national and European statistics can be misleading in what concerns the role of wood energy in the domestic sector at the regional level, in particular at the level of small cities. In the EU, small cities, here defined broadly as urban centres with population ranging from 5 000 to 50 000 inhabitants, represent near 60% of the resident population, i.e., near 217.5 million inhabitants (ESPON 2006). Consumption of wood energy in this city size class can be much higher than expected from general transitions theory and national level data. Detailed data to test this hypothesis is not, unfortunately, available. Drivers and preferences of wood energy use in the residential sector in Europe are practically unknown (Couture et al. 2009, Lindroos 2011). Research on these topics has been mostly conducted in developing countries in Asia and Africa, providing important information on sources, uses, and effects, particularly at the village scale (e.g., Zhu et al. 1983, Alabe 1996, Campbell et al. 2003, Pachauri and Jiang 2008). There is also considerable work done in the USA, (e.g., Hardi and Hassan 1986, Force 1989, Song et al. 2012a, Song et al. 2012b). Identified drivers and preferences of wood energy, usually in rural areas across the globe, include a wide range of factors such as elevation, size of a household’s private garden, per capita firewood forest area energy required, education level and access to a reliable supply of electricity, coal price, electricity price, and distance to coal purchasing place, among others (Force 1989, Mislimshoeva et al. 2014, Wei et al. 2012, San 2012, Cai 2008). The identification of drivers of consumption in small cities in Europe is important to understand the current use of wood energy in these centres, but also to assess the role of these cities in wood energy markets at the national and international levels and to evaluate the effects of changes in wood consumption on related domains such as land-use planning, forest management, forest conservation, and air quality and public health (Liu et al. 2003, Naeher et al. 2007, DeFries et al. 2010, Bouriaud et al. 2014). The goal of this work was to assess major drivers of wood energy consumption in Bragança, a small city in the North of Portugal, as a case study representing a range of small European cities to which our results can apply. The understanding of preferences of wood energy consumers as well as the potential effects of wood energy in existing forest resources, was also a purpose of this study. The specific objectives were to examine in detail (i) what is the role of wood in the residential energy services, (ii) what factors affect the use and the volume of wood consumption in households, (iii) what are the preferences of consumers, and (iv) how changes in drivers and preferences can affect forest resources locally. These objectives were tackled through a detailed energy survey at the city level and statistical modelling of wood energy consumption.
What drives consumption of wood energy in small cities in Europe? 3
MATERIAL AND METHODS Study area Our study was conducted in the city of Bragança, northeastern Portugal. Bragança is the largest city and the capital of the district and municipality with the same name. Resident population is 23 099 inhabitants distributed among 8851 households in three parishes: Sé, Santa Maria, and Samil (INE 2012). Population has grown over the last decades due to migration from parishes in the countryside (INE 2012). It is located in a predominantly rural region where agriculture/ forestry activities prevail. The economy of the city is centred in the services sector, particularly education and administration. In the municipality of Bragança land cover is dominated by forests (35.4%) and shrublands (32.3%) followed by agricultural land (30.5%) (AFN 2010). Forests are dominated by deciduous oaks, mainly Quercus pyrenaica (35.2%), maritime pine, Pinus pinaster, (29%), and chestnut, Castanea sativa (19.2%) (AFN 2010). The climate is temperate, with continental and Atlantic influences. Average annual temperature is 12.2°C; average minimum temperature is 6.8°C; and average annual precipitation is near 750 mm. Winters concentrate most of the precipitation. Elevation in the city is around 670 meters. Energy survey Wood and energy consumption in the residential sector in the city of Bragança was surveyed based on 201 structured interviews carried out in March and April of 2012, following a questionnaire consisting mostly of closed questions. The household sample was randomly selected and verified for representativeness of location (parish), housing typology, and construction age. Interviews were conducted with the member of the household responsible for the payment of utilities here considered as “head”. The sample size corresponds to a 6.8% confidence interval for a 95% confidence level. The interviewees were mostly females (57% vs. 43% males), averaged 41 years of age, completed high school (39%) or had a university degree (39%). The questionnaire was divided into five groups of questions: (I) personal data; (II) housing characterization; (III) biomass consumption and residential heating systems characterization; (IV) other energy sources characterization; and (V) additional questions. The first group related to the respondent’s gender, age and education. The housing units characteristics (group II) included items such as type, area, number of floors and rooms, number of residents, year of construction, isolation measures, efficiency class and existence or absence of a central heating system. Group III included questions on heating system type, function, and months of operation. When interviewees used firewood, additional questions were asked concerning the type, quantity, provenance and other characteristics of firewood. Quantity of firewood was expressed in Mg since firewood is usually sold locally in trailer loads of approximately 4 Mg. Questions in group IV
regarded the identification of energy sources other than biomass used in households, their main uses, periods, quantities, costs, and the use of other renewable energy production systems (solar, wind). These sources of energy were reported in terms of actual consumptions obtained mainly from electricity and gas bills. The final group of questions (V) included complementary aspects of interest related with energy use in the residential sector such as awareness of air quality and health related issues. Results on wood and energy consumption were directly scaled up to the city level based on the proportion of households. This procedure was followed since the proportion of building types in the sample was approximately the same observed from statistics data (INE 2012). Energy was expressed in Joule, according to the following lower heat content values: natural gas and bottled gas – 46 GJ/Mg; diesel – 42.6 GJ/Mg; wood biomass - 14.65 GJ/Mg; Pellets and briquettes – 16.8 GJ/Mg. Wood biomass lower heat content is an average value for firewood considering moisture in the 25–30% range. Electricity was also converted to primary energy using a factor of 2.5, corresponding to an efficiency of 0.4 (Directive 2006/32/EC, European Parliament and Council, April 5, 2006, and Despacho n.º 17313/2008 of the Portuguese Republic). Modelling biomass consumption An exploratory analysis based on logistic and multiple regression models was conducted to detect the most relevant factors that affect wood energy consumption. In the dataset, consumption of wood obtained from the surveys, was recorded as a 0/1 dichotomous variable corresponding to “no consumption” and “consumption”, respectively. For the households consuming wood the quantity of biomass consumed (Mg/yr) were registered. A two-step approach to data modelling was followed. Firstly, all the data (n=201) to model the probability of wood energy consumption in the city applying logistic regression was used. Then, multiple regression analysis to model the amount of biomass consumed using only data from households that use biomass (n=84) was applied. In the logistic regression procedure an initial exploratory analysis to fit models for each candidate to explanatory variable independently (Table 1) was conducted using SAS/STAT PROC LOGISTIC (SAS Institute Inc. 2008). This step enabled to examine the statistical significance of the possible regressors which included continuous and categorical variables as well as discretized versions of continuous variables (Table 1). In this early stage, one observation was eliminated after checking for the existence of outliers. Then, the stepwise method of PROC LOGISTIC was used for variable selection with the logit transformation of the dependent variable and avoiding colinearity among explanatory variables. The maximum likelihood (ML) method was used for parameter estimation. The hypothesized probability logistic model is of the form:
{
}
pˆ = 1 + exp − ( b0 + b1 X 1 + b2 X 2 +... + bm X m ) (1) −1
4 J.C. Azevedo et al.
TABLE 1 Variables used in logistic regression and multiple linear regression models Variable
Description
explanatory variables. The Akaike Information Criterion (AIC) for each model was also considered. The multiple linear regression model has the form: Y i = b0 + b1 X 1 + b2 X 2 + ... + bm X m (2)
NRooms
Number of rooms (numeric variable; range: 1 to 8)
NFloors
Number of floors (numeric variable; range: 1 to 4)
where Yi is the amount of biomass used per household, β0 is the intercept and the βi are the parameters of the Xi independent variables. Parameter estimates were obtained by the ordinary least squares (OLS) method. Quality of fit was evaluated by 2 the adjusted coefficient of determination Radj and the root mean square error (RMSE):
Age50
Age of household head over 50 years
EDPr
Education level of household head: Illiterate or primary
EDSe
Education level of household head: Secondary (9th grade)
EDHs
( )
n
2 Radj = 1−
Education level of household head: High School (12th grade)
EDGrad
Education level of household head: University
HTApt
Housing type Apartment
HTSdh
Housing type Semi-detached house
HTDh
Housing type Detached house
HTRh
Housing type Row house
CY1
Housing unit with more than 1 kitchen
BDACF
Burning device Air Closed Fireplace
BDOF
Burning device Open Fireplace
BDS
Burning device Stove
BDWCF
Burning device Water Closed Fireplace
All variables entered in the models as dummy variables (coded 0/1) with the exception of the numerical variables NRooms and NFloors.
(
)
Mpr =
MA pr =
∑ (y EF = 1 − ∑ (y
2
K1
1 n ∑ y i − yˆ i* , (5) n i =1
1 n ∑ y i − yˆ i* , (6) n i =1 − yˆ *i
i
)
2
−y)
2
i
, (7)
where yi is a real observation and yˆ *i is the corresponding predicted value calculated with the model fitted with the observation i deleted from the original data set. Both in the logistic regression and in the linear regression the inclusion of variables in the models with logical signs of their parameters was also considered. RESULTS
where pˆ is the probability of biomass consumption, β0 is the intercept, and the βi are the parameters associated with the Xi independent variables. For the regression model, the best subset regression was performed using SAS/STAT PROC REG (SAS Institute Inc. 2008) with the R2 selection method with the option BEST=3 to obtain the three best models with a different number of
Energy survey Housing characteristics Surveyed households were distributed mainly in apartments, which represented 69% of all accommodation units surveyed (Table 2). Buildings were constructed mostly after 1981 (88% of surveyed housing) but in a large proportion (56%) after
What drives consumption of wood energy in small cities in Europe? 5
TABLE 2 Types and sizes of sampled housing units in Bragança, Portugal Interviews (No.)
(%)
Average total surface (m2)
Apartment
139
69
131.2
8.2
2.2
Semi-detached house
24
12
262.5
12.0
2.9
Detached house
24
12
261.3
10.8
3.6
Row house
14
7
171.7
8.6
1.8
Housing type
1996 (Table 3). Dwellings were relatively large in size, as compared to European or Portuguese national standards (see, e.g., Eurostat database), both in area and number of rooms in all housing types (Table 2). The adoption of energy efficiency measures at the household level was relatively common in the sample as suggested by frequency of double walls (66% of housing), double glazing (76%), and insulation (61%) in all types of housing (Table 4). Energy sources Households in Bragança used, in general, more than one source of energy simultaneously. All households had electricity. Moreover, natural, and bottled gas and diesel were available in 64.7, 29.9, and 15.9% of the households, respectively. Wood was used as a source of energy in 42% of the
Average number of divisions (No.)
Average number of facades (No.)
households surveyed. Overall, natural gas was the most frequently used energy source and it was preferred for water heating and cooking (Table 5). Electricity was the second most frequently used source although it was not the first choice in any of the individual uses considered. Wood energy was used mostly for air heating representing the most frequent energy source providing this energy service. The use of wood energy for water heating and cooking was occasional (Table 5). Wood energy Wood alone provided an estimated 245.6 TJ/yr to the residential sector in Bragança which accounted for 42.8% of final energy (Table 6) and 31.3% of primary energy consumption (Table 7). Air heating was the predominant energy service for
TABLE 3 Absolute frequency of sampled housing units per construction year in Bragança, Portugal Housing type Construction year
Row house
Semi-detached house
Detached house
Apartment
Total
(No.)
(No.)
(No.)
(No.)
(No.)
2
1
2
1
6
300m2 is the construction area above 300 m2, and CY80 and CY90 are construction year in the eighties and nineties, respectively. Parameter estimates are presented in Table 9. The logistic model was statistically significant judging 2 from the results of the likelihood ratio test ( G (5) = 89.13 ; p
