Citation
The social and environmental context related to the utilization of improved cooking technology in rural Uganda

Material Information

Title:
The social and environmental context related to the utilization of improved cooking technology in rural Uganda
Creator:
Brunner, Nicole Michelle ( author )
Place of Publication:
Denver, Colo.
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (75 pages) : ;

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Geography and Environmental Sciences, CU Denver
Degree Disciplines:
Environmental sciences

Subjects

Subjects / Keywords:
Stoves ( lcsh )
Social conditions ( fast )
Stoves ( fast )
Social conditions -- Uganda ( lcsh )
Uganda ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Human dependence on natural resources, especially from forests, is most pronounced in developing countries such as Uganda, where many people live in poverty and rely on wood fuel for cooking. These demands often compete with conservation efforts aimed at protecting forests and biodiversity. A better understanding of the relationship between social and environmental conditions and the utilization of improved stove technology is essential to ensure the sustainability of forest-related socioeconomic activities and the conservation of biodiversity in Uganda's forests. Statistical modeling was used to explain the relationship between utilization of improved stove technology and socio-economic characteristics acquired from household surveys and environmental characteristics derived from remote sensed imagery. The findings show that socio-economic and environmental conditions are important determinants of improved stove adoption. Charcoal stoves were found to be related to higher levels of welfare, less forest cover and further distance from protected areas. Whereas, efficient wood stoves were associated with closer proximity to protected areas representing restricted access to resources. The results of this study could inform policies by providing context on the characteristics of households and communities utilizing improved stove technology.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Nicole Michelle Brunner.

Record Information

Source Institution:
University of Colorado Denver Collections
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
985411050 ( OCLC )
ocn985411050
Classification:
LD1193.L547 2016m B79 ( lcc )

Downloads

This item has the following downloads:


Full Text
THE SOCIAL AND ENVIRONMENTAL CONTEXT RELATED TO THE
UTILIZATION OF IMPROVED COOKING TECHNOLOGY IN RURAL UGANDA
By
NICOLE MICHELLE BRUNNER B.S., Metropolitan State University Denver 2012
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Environmental Science Program
2016


This thesis for the Master of Science degree by
Nicole Michelle Brunner has been approved for the Environmental Science Program by
Gregory L. Simon, Chair Peter Anthamatten Deborah Thomas
July 22, 2016
11


Brunner, Nicole Michelle (MS, Environmental Sciences)
The Social and Environmental Context Related to the Utilization of Improved Cooking Technology in Rural Uganda
Thesis directed by Associate Professor Gregory L. Simon.
ABSTRACT
Human dependence on natural resources, especially from forests, is most pronounced in developing countries such as Uganda, where many people live in poverty and rely on wood fuel for cooking. These demands often compete with conservation efforts aimed at protecting forests and biodiversity. A better understanding of the relationship between social and environmental conditions and the utilization of improved stove technology is essential to ensure the sustainability of forest-related socio-economic activities and the conservation of biodiversity in Ugandas forests. Statistical modeling was used to explain the relationship between utilization of improved stove technology and socio-economic characteristics acquired from household surveys and environmental characteristics derived from remote sensed imagery. The findings show that socioeconomic and environmental conditions are important determinants of improved stove adoption. Charcoal stoves were found to be related to higher levels of welfare, less forest cover and further distance from protected areas. Whereas, efficient wood stoves were associated with closer proximity to protected areas representing restricted access to resources. The results of this study could inform policies by providing context on the characteristics of households and communities utilizing improved stove technology.
The form and content of this abstract are approved. I recommend its publication.
Approved: Gregory Simon
m


ACKNOWLEDGEMENTS
There are many people who have helped me and provided guidance and feedback for this project. Among those I would like to acknowledge the researchers with whom I work at the Geoscience and Environmental Change Science Center, Dr. Darius Semmens, Dr. Kenneth Bagstad, and Dr. Todd Hawbaker and my committee members from the department of Geography and Environmental Sciences Dr. Deborah Thomas and Dr. Peter Anathamatten, who had to endure long meetings about methods and reading my thesis serval times, your comments were essential to the success of this project. I would also like to acknowledge Michael Verdone from IUCN and Dr. Bonnie Keeler from Natural Capital Project who directed me to Uganda, shared data and knowledge, and were there to discuss ideas at the beginning of this project. A special thank you to my advisor Dr. Gregory Simon who has supported me throughout my thesis and played a crucial role overseeing my progress through the program. He has not only given me guidance and encouragement over the last 3 years, but has broadened my knowledge and changed the way I think about issues. This experience would not has been the same without him. This research has been reviewed and was determined non-human subject research as defined by the policies of Colorado Multiple Institutional Review Board (COMIRB) in accordance with OHRP and FDA regulations. The protocol number is 16-
0281.
IV


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION..........................................................4
II. LITERATURE REVIEW.....................................................7
Traditional stove and energy use........................................7
Improved biomass stove technologies....................................12
Factors influencing improved fuel and stove use........................15
III. STUDY AREA..........................................................19
Background information.................................................19
Uganda energy use......................................................25
IV. DM A AM) METHODS.....................................................28
Data...................................................................28
Social survey data...................................................28
Environmental data...................................................30
Methods................................................................31
Conceptual framework.................................................32
Research questions...................................................33
Temporal analysis When?............................................33
Spatial analysis Where?............................................34
Explanatory analysis Why?..........................................35
ii


Analysis of relationships What?
36
Description of variables chosen for analysis.........................37
V. RESULTS.....................................................................43
The temporal trends of stove use..............................................43
The spatial analysis of stove use.............................................46
The potential benefits of improved stoves.....................................47
The relationship between variables driving the utilization of improves stoves.54
Analysis of efficient wood stoves...........................................54
Analysis of improved charcoal stoves........................................55
VI. DISCUSSION..................................................................58
VII. CONCLUSION.................................................................65
REFERENCES......................................................................68
iii


CHAPTERI
INTRODUCTION
Close to half of the global population uses traditional wood fuels for cooking and heating, which typically include firewood and charcoal. These fuels are predominantly used in poor and rural areas of sub-Saharan Africa. Poor ventilation of open fire stoves coupled with incomplete combustion of these wood fuels results in high levels of indoor air pollutants which impacts health (Hutton, et al., 2007; Adkins, et al., 2010; Bailis, et al., 2015). In addition to releasing health-damaging pollutants, traditional charcoal and wood burning stoves have been identified as a contributor to greenhouse gas (GHG) emissions. Also, in regions where resources are limited the amount of time spent collecting wood for fuel can be a considerable burden on impoverished rural households, especially women and children (Adkins, et al., 2010). These stoves are also often tied to unsustainable tree harvesting which can also impact climate change and biodiversity loss. However, the extent of these effects remains highly uncertain (Bailis, et al., 2015). Due to the negative impacts associated with the use of traditional fuels, improved more efficient stoves are being funded in order to promote sustainable development.
Ugandas population is highly reliant on traditional wood fuel (Khundi, et al., 2011). Firewood is the primary fuel in Uganda, however there has been a steady increase in charcoal use. Like many developing countries the majority of Ugandas population is rural, and their livelihoods and wellbeing are heavily reliant on natural resources. In addition, the population is growing rapidly, coupled with increasing production and resource extraction, which has raised concerns about energy and resource shortages.
Uganda is one of the most biodiverse as well as poorest countries in the world. It was ranked 163th out of 187 on the 2015 human development index (UNDP, 2015). Poverty is often
4


linked to environmental degradation. Traditional stove and fuel use, poverty, and environmental degradation are often coincide in developing countries. Therefore, poverty is often cited as causing environmental degradation (Aggrey, et al., 2010). Likewise traditional stove and fuel use are linked to deforestation.
There are concerns that increasing demand for natural resources in Uganda will lead to destruction of protected areas that are of particular conservation importance due to the countrys rare biodiversity, including highly endangered species such as chimpanzees and mountain gorillas, which also bring in significant revenue through eco-tourism. The primary activates cited as threats to conservation goals are agricultural clearing for cultivation and grazing, illegal hunting (poaching), logging, pit-sawing and charcoal production (Plumptre, et al., 2010).
While firewood collection does not drive large-scale deforestation, many local or regional assessments have cited wood fuel demand as a primary driver of forest loss (May-Tobin, 2011). This discrepancy suggests that impacts of wood fuel demand may vary by location and fuel type (Bailis, et al., 2015). For example, there are some contextual disparities between developing countries in Africa, where wood fuel has been cited as having a greater influence on land use change than in other regions of the world (Naughton-Treves, et al., 2007). Also, charcoal production also has a greater environmental impact than firewood collection. Unlike firewood, which typically consists of twigs and smaller branches, charcoal is usually made by burning larger tree trunks and limbs which leads to more forest loss than firewood collection (May-Tobin, 2011). Naughtons-Treves et al. (2007) determined that producing 50 tons of charcoal from hardwood would results in 1 km2 of deforestation. The per capita consumption for firewood in Uganda has been estimated to 680 kg per year for rural residents and 240 kg per year for urban residents. Charcoal consumption is about 4 kg for rural and 120 kg for urban dwellers (Mwaura,
5


et al., 2014). Therefore, about every 400 urban households requires 1km2 of forest to supply their energy requirements per year. Regardless of whether wood fuel use contributes to forest loss, there are concerns that forest loss in developing countries will lead to fuel shortages and increased vulnerability of already impoverished populations.
Based on traditional policy solutions that view environmental degradation as a local problem of over-consumption caused by poverty, global development initiatives and government organizations have formed policy aimed at distributing improved more efficient stoves. This improved stove technology claims to help by reducing fuel demand, potentially alleviating deforestation pressure in select locations, enhancing public health and alleviating poverty on a local scale, as well as mitigating climate change at the global level (Global Alliance for Clean Cookstoves, 2011). Therefore traditional wood fuel and stove use presents society with two important links between local and global impacts. Merging social science and remote sensing data can offer new insights into how issues such as poverty, gender, demography, or development relate to resource use and availability (Geoghegan, et al., 1998).
The objectives of this study are to explore (i) the geographic and temporal differences in fuel types and cooking technology in rural households in Uganda (ii) the claimed benefits of improved stove use, and (iii) the social and biophysical characteristics related to the adoption of improved cooking technology. Understanding how the household choice to adopt improved cooking technology is related to social and biophysical characteristics (such as demand and access to resources) will provide information that will hopefully contribute to policies aimed at addressing peoples basic needs as well as conservation of the environment.
6


CHAPTER II
LITERATURE REVIEW
This chapter reviews the literature addressing three topics relevant to this project. I begin by reviewing the contribution of traditional stove and fuel use to negative impacts associated with health, poverty and environment. Followed by an overview of improved cookstove technology. Finally is an overview of the methods other studies have used and factors found that influence the choice to utilize improved stove technology.
Traditional stove and energy use
Wood fuel and traditional stoves are used predominantly in developing countries. Globally, at least 2.6 billion people are estimated to use traditional wood fuel, and about 2.2 billion people use inefficient traditional stoves to bum those fuels. Traditional wood fuel primarily consists of firewood and charcoal, which is typically collected from nearby forests. It estimated that 55% of all wood harvested is used for fuel, which supplies 9% of the worlds primary energy (Bailis, et al., 2015). The use of traditional fuels and stoves is exceptionally high in sub-Sahara Africa, where 85% of the rural population uses traditional wood fuel, and up to 94% use traditional stoves (Okello, 2014). Traditional stoves are typically a 3-stone stove or open fire. The 3-stone stoves burn primarily firewood. The three stones support the cooking vessel and acts as the hearth, where wood is burned by pushing it through the openings in between the stones (Adkins, et al., 2010). Traditional charcoal stoves also referred to as traditional metal stoves, are typically made from found metal scraps such as roof materials, or reclaimed oil drums. Traditional stoves have no insulation which makes them very inefficient (Okello, 2014).
7


Traditional stoves produce massive amounts of smoke. The smoke, which is from incomplete combustion, is linked to global climate change due to the release of greenhouse gas emissions including Carbon Dioxide (C02), methane, black carbon, and short-lived climate forcers (SLCFs) (Panwar, et al., 2009; Simon, 2014; Bailis, et al., 2015). This smoke is also toxic and harmful to human health. Traditional stoves release carbon monoxide, particulate matter, and non-methane volatile organic compounds which are linked to eye irritation, cataracts, respiratory illnesses, including asthma, low birth weight and still birth, tuberculosis and cancer (Simon,
2014; Adkins, et al., 2010). Traditional stoves are most often used in main indoor living spaces that are poorly ventilated and where many people are exposed, which increases the negative health impacts. It is estimated that approximately 1.5 million people die annually from health issues induced by exposure to smoke from traditional stoves (WHO, 2006; Okello, 2014).
Another major social development challenge associated with traditional wood fuel use in developing countries is social and gender dimensions. Women and children primarily participate in the collection and processing of wood fuel. The collection process, in particular, is associated with a considerable amount of human labor, risk, and time which could have been spent pursuing more economically productive activities (Foell, et al., 2011). Women and children are also the most commonly exposed to indoor pollutants and are the most vulnerable to the harmful effects (WHO, 2006).
The health impacts caused by burning wood fuels in traditional stoves is increasingly well characterized, whereas the contribution to global climate change from the release of GHG emissions and fuel driven deforestation is still highly uncertain. It is estimated that 18% of the global atmospheric black carbon concentration originates from incomplete combustion (Foell, et al., 2011). Black carbon contributes to climate change, and is also linked to increasing glacial ice
8


melt as well as altering regional rainfall patterns (Okello, 2014). Ballis et al. (2015) estimated that in 2009, GHGs and SLCFs from wood harvested unsustainably and partial burning of fuel accounted for between 1.9 2.3% of global emissions, these emission estimates were divided equally between C02, black carbon, and SLCFs. The estimates for fractional non-renewable biomass (fNRB), which is wood harvested unsustainably, were found to be much less than carbon-financed improved stove and wood fuel projects estimate. Therefore, it is likely that projects are exaggerating the potential for improved stove technology to reduce emissions through reducing deforestation.
Most of the literature is divided on the extent to which wood fuel demand drives deforestation. While traditional wood fuel is a renewable resource, there are concerns about its sustainability with increasing energy demand, especially in areas with high rates of deforestation. These concerns are especially relevant in sub-Sahara Africa, where there are high population growth rates which often coincide rapid forest loss. The use of traditional and inefficient stove technologies only exacerbates the problem of fuel insecurity. Since the 1980s, the scientific community has, for the most part, accepted that fuelwood use may not necessarily lead to deforestation on a large scale and that fuelwood crisis may not become a reality (Mwampamba, 2007). However, studies at local scales have found that location is a key factor determining whether fuel harvesting is sustainable due to disparities between wood demand and resource availability (Arnold, et al., 2006). In a special issue published by the International Journal on Energy Policy in 1993, the most current research which dismissed the wood fuel crisis, emphasized the inaccuracy of attributing deforestation exclusively to wood fuel extraction (Mwampamba, 2007).
9


An assessment by Kaimowitz & Angelsen, (1998) on tropical deforestation found that several factors lead to forest loss. This study also and notes that there is not substantial evidence linking wood fuel demand and deforestation, although it is occasionally a driver of deforestation in parts of Africa and Asia. Similarly, another study examining multiple cases of tropical deforestation found numerous drivers, however tree harvesting for wood fuel was a more significant cause in some parts of Africa (Geist & Lambin, 2002). However, there have been very few systematic studies of wood fuel sustainability (Bailis, et al., 2015), and many studies that attribute the demand for wood fuel as the primary cause of deforestation and forest degradation dont include any empirical evidence.
However, studies that distinguish between the impacts of firewood versus charcoal find that while firewood collection rarely threatens forests, charcoal use has very different implications. Due to inefficient charcoal production and consumption, households use on average 4 to 6 times more wood than households using firewood (Mwampamba, 2007). In addition, charcoal production involves cutting larger branches or tree trunks which often results in tree removal, and the preferential wood species to produce charcoal are typically slow-growing, and therefore is less sustainable than firewood.
Kituyi, et al. (2001) found that fuel security in Kenya is determined by the presence of biodiversity of wood types which allows households to switch fuel types with relative ease. Similarly, households often respond to fuel scarcity by harvesting other available tree species that are growing on farms and in protected forests. The decreasing of major tree species used for wood fuel in Kenya is not entirely due to fuel demand, but due to competing demands for timber, agriculture and settlement. These increasing demands raise concerns about the long-term sustainability of forests to supply sufficient fuel demand in Kenya. A study done by Luoga, et al.
10


(2000) in eastern Tanzania shows that while the commercialization of wood resources, particularly charcoal production, provides significant income to rural households, it also contributes to significant forest resource loss.
A study in West Africa conducted by World Bank found that the concentration of charcoal production can be a primary driver of deforestation (Arnold, et al., 2006). A study done by Ekeh et al. (2014) examining the level of GHG emitted from charcoal production, transport and use in Uganda, found that 15% of GHG are emitted from the transportation phase. The production and use of charcoal accounted for the vast majority of GHG emissions. Therefore shifting to improved technology to produce charcoal would result in significant GHG reduction, however using sustainably sourced biomass would reduce GHG emissions by almost half, and therefore the best solution is sustainable harvesting practices of wood coupled with improved production methods.
Mwampamba, (2007) evaluated the sustainabiliy of current charcoal consumption rates, with the current efficiency of local charcoal production and current forest management policies to determine weather forests can contiune to meet current and future demands in Tanzania. 24 scenarios were developed to capture the numerous uncertainties of forested area required to meet charcoal demand. Their findings suggest that scenarios based on moderate levels of consumption paired with low production efficiency and very little reforestation efforts could substantially reduce total forest area on public lands by 2028. This study determines that charcoal consumption activities are a threat to the future of forests in Tanzania.
Charcoal is the primary fuel used for cooking in most urban centers sub-Saharan countries (Mwampamba, 2007). Increasing demand for charcoal from growing urban populations, coupled with poor resource management, and regulation present a problem for the
11


future for forests and fuel shortages in Africa (Mwampamba, 2007; May-Tobin, 2011; Foell, et al., 2011).
Traditional stoves are a primitive technology with several associated negative impacts. With diminishing resources and increasing demand to supply fuel, alternatives must be provided to address basic human welfare. Charcoal can be damaging not only to human health but also the environment.
Improved biomass stove technologies
Due to the increasing interconnectedness between environments, societies, and economies worldwide, there are a diversity of actors (i.e. NGOs, private sector, international organizations, civil society) shaping environmental governance. Carbon financing or carbon trading is a mechanism aimed at reducing greenhouse gas emissions (GHG) and mitigate climate change. Carbon credits (equivalent to 1 ton of carbon dioxide) are often purchased by developed countries and businesses through emission trading markets in order to meet emissions reductions obligations made under the Kyoto Protocol.
Due to the many recent reports and studies noting the contribution of traditional stoves and wood fuel use to GHG emissions, coupled with health concerns and deforestation, has led the promotion of improved stoves (Okello, 2014). Carbon financing has been providing funds to implement clean cookstove projects in developing countries (Simon, et al., 2012). These projects are often financed under umbrella initiatives such as the Clean Development Mechanism (CDM), and implemented under the Kyoto Protocol (Peskett, et al., 2011). Along with aiming to reduce GHG emissions, these programs often promote win-win scenarios in which local communities, forests, biodiversity and food security are all positively affecting by the program..
12


Improved stove designs range from locally made to manufactured models. Locally made stoves are constructed from materials found in or near the home and typically consist of clay soils or mud and dried grasses. Manufactured stoves are finished products produced entirely in factories, either domestically or internationally, then distributed to local villages. These manufactured stoves are typically based on the rocket stove design principle (Adkins, et al., 2010). Improved stoves have several claimed co-benefits besides burning fuel more efficiently than the widely used traditional stove. These additional benefits include reduced cooking time, improving health, lessening deforestation pressures, alleviating the burden placed on women and children in fuel collection, reducing greenhouse gas emissions, and mitigating poverty (Kees & Feldmann, 2011; Okello, 2014). The extent to which improved stove use benefits households, lessens deforestation and forest degradation, as well as mitigates climate change, remains highly uncertain (Kees & Feldmann, 2011; Bailis, et al., 2015). A study by Bailis, et al. (2015) found that the adoption of 100 million state-of-the-art cookstoves would decrease traditional wood fuel emissions by 98-161 million tons of CO2 per year, valued at $1.1-1.8 billion USD, an amount that far exceeds the current investments in improved cookstove projects.
Improved stove designs funded through carbon finance are typically manufactured in order to standardized technology and accurately account for emission reductions (Adkins, et al., 2010). However standardizing technology introduces a question of whether the technology practical for household use. If a household does not perceive benefits of improved stoves they will be less likely to use the stove. Studies have found that households with improved stoves are commonly used along with traditional stoves, commonly referred to as stove stacking. A household survey and efficiency test done by Adkins et al. (2010) found that while improved stoves reduced the amount of fuel used compared to the traditional counterpart, some models
13


showed significant increases in time spent cooking. For example, cooking matooke, a traditional staple food in Uganda, using a traditional 3-stone fire requires approximately 17 min, whereas the improved stove models tested both showed increases in time needed to cook matooke. The Ugastove showed a significant increase 22 min (27% longer), the StoveTec stove showed only a slight increase 18 min (5% additional time) compared to the 3-stone traditional model. The survey portion of the study found that despite the reduction in smoke and reduced fuel use, about 40% of households preferred traditional stoves over the improved model. The major complaints were difficulty lighting, easy to tip over, longer time spent cooking, and too much heat causing burning.
Masera, et al. (2000) found that very few rural households completely abandon traditional fuels and stoves. Cutlural and technical factors were found to be the primary reason for stove stacking in rural villages in Mexico. This practice occurred among households of varying socioeconomic status. Even though improved and modem stoves carry status as symbols of prosperity, they are still used along with tradtional stoves and fuels. This is the case because improved stove design that does not meet all of the of cultural and customary cooking requirements.
Access to clean and affordable energy has been identified as critical to achieving the millennium development goals due to the claimed co-benefits, such as improved maternal health, reduction in premature death, reducing environmental impacts, and saving time in reduced fuelwood collection. Overall, improving access to clean stoves is expected to improve human welfare and contribute to sustainable development in developing countries (Okello, 2014). However, there is debate about whether these win-win scenarios are realistic and feasible (Simon, et al., 2012)
14


Factors influencing improved fuel and stove use
Besides technical limitations of improved stove designs, household stove choice and the relative fuel consumption are influenced by factors at various structural levels ranging from micro- or household-level factors to biophysical features related to fuel availability, which together make up the social and environmental context that determines household preferences regarding cooking technology (Masera, et al., 2000). At the household or micro-level, the distinction in fuel and stove use in rural communities may be due to demographic, economic conditions, cultural values, health concerns or a combination of factors. At the village or community level, resource scarcity and the influence from proximity to urban centers, could be motivating households to switch to more efficient cooking options (Masera, et al., 2000). A study by (Mahapatra & Mitchell, 1999) in rural India found that while wood fuel use is influenced by socio-economic factors, limited forests resources does not reduce the demand for wood fuels. In other words, socio-economic factors influenced fuel use more than resource availability.
Masera et al. (2000) claim that fuel and stove switching is the result of simultaneous interactions of factors that push households away from traditional practices, and pull them back. Push factors include convenience, cleanliness, health and economic status. Pull factors usually result from defects or inadequacies of technology to meet the needs of traditional cooking practices.
The literature on household adoption of improved fuel and stoves in developing countries is for the most part inconsistent and predominantly uses qualitative methods. While there have been numerous attempts to assess improved stoves or fuel choice, few studies have used rigorous statistical analysis to confirm findings (Lewis & Pattanayak, 2012). A literature review by Lewis and Pattanayak (2012) found that the most common factors with a significant positive correlation
15


to improved fuel and stove adoption were education, income, and urban location. The roles of factors such as fuel availability, price, household size and composition are still largely unclear. Demographic characteristics showed varying results, household size and average age of head of household was inconclusive. Most of the studies that include education find a positive and statistically significant relationship with improved stove adoption. Most studies (56%) found that female heads of households were more likely to adopt improved fuels. However improved stove studies seldom considered female head of household in their analysis. The review also found that the majority of studies found no significant associations between improved stove adoption and livelihood or occupation, number of children, and the price of modem fuel. Some variables that were understudied include proximity to markets, roads and populated centers (Adkins et al.
2010). Therefore, there is greater need to explore more of these understudied variables as well as variables with known significant correlations to ensure consistency and verifiable results.
Wallmo & Jacobson (1998) found that protected areas are likely a factor associated with the use of improved stoves. The main reason households tended to implement improved stoves in 3 pilot parishes in western Uganda was to reduce the amount of fuel required for domestic energy use to lessen the need illegally harvest wood from the adjacent National Parks. People living outsides parks and protected areas are often excluded from accessing resources, settlement, and land use within the park boundaries (Hartter, 2010) which can foster attitudes of resentment towards the government authority, the protected forests, and the wildlife residing within. A study by Bush et al. (2004) found that the further residents were from four protected national forests in Uganda, the fewer resources were extracted and less total household income was derived from forests. This relationship was due to the increased levels of deforestation due to high density of agriculture and over-exploitation of resources (Hartter, 2010; Nabanoga 2005).
16


Therefore, distance from forests (especially protected forests) is likely an important factor in driving resource use and availability.
Community-level economic factors and resource availability are often closely interlinked. Access to forests has a critical effect on peoples survival and well-being, and plays an important part in livelihood strategies (Nabanoga, et al., 2010). Forests provide wood and non-wood forest products and services to the people who produce and consume forest products, and local people play a major role in the management of forest resources. Government authority is weak and property rights are often times ambiguous in developing countries, turning dynamics within the community closer to the ground into critical determinants of actual resource use and access.
There is an indisputable correlation between environmental degradation and poverty; as natural resources become depleted, local economies suffer and livelihoods become more vulnerable (Emerton, 2001). While there are circumstances in which local livelihood activities contribute to environmental degradation, there are also cases where the impacts of local production activities are relatively benign. However, local communities are commonly blamed for environmental degradation and unsustainable harvesting which is a reflection of exercising control over resources by governments and authorities (Robbins, 2005). Therefore, cultural politics play a major role in conservation struggles due to resource claims based on traditional or customary land rights. Having been alienated from their resources, many local people have been re-claiming those resources and as populations increase around protected areas, this problem has only intensified. These tactics often include trespassing, a steady encroachment into areas of restricted access, poaching the protected wildlife, illegal grazing, tree harvesting, fuelwood collection, threatening government personnel, and setting fires to gain control and access (McCarthy, 2002). These restrictive boundaries of government lands which elevate ecological
17


goals above basic human needs, contradict contemporary conservation trends, which have emphasized the need for porous boundaries to protected areas, extractive reserves, and community support for and participation in conservation (McCarthy, 2002).
While linking exploring proximate drivers or causes of environmental degradation are important, such as proximate causes of deforestation (slash-and-bum cultivation or wood fuel collection). It is also important to explore relationships between factors such as socio-economic status and biophysical characteristics in order to understand the underlying processes and structures that influence household level decisions that in turn give rise to environmental outcomes (Geoghegan, et al., 1998).
18


CHAPTER III
STUDY AREA
The study area selected for this study is Uganda, mainly due to the high levels of dependence on traditional fuels and diminishing forest resources. In addition, there are several global development and carbon financing initiatives aimed at promoting improved stoves in the country. This chapter presents an overview of Uganda, the energy sector, environmental concerns with regards to deforestation and traditional wood fuel use, as well as an overview of the global organizations and efforts aimed at promoting improved cookstove and forest conservation projects in Uganda.
Background information
Uganda is a landlocked country located along both sides of the equator, between 1 30' S and 4 N latitude and between 29 30' East and 35 West longitude, in Eastern Africa. Sudan borders Uganda on the north, Kenya is on the eastern border, Rwanda, and Tanzania are to the south and Democratic Republic of Congo is to the west (Figure 1) (NEMA, 2009).
19


0 1,250 2,500 5,000 Kilometers
1 _I___I__I___I__I___I__I__I
| Waterbodies Uganda Regions
270'0"E 300'0"E 3300"E 3600"E
I______________________________i_______________________________i________________________________i_______________________________I_______________________________i_______________________________i______________________________i_______________________________I
Figure 1: Map of Africa, location of Uganda
The boundaries that delimit Uganda were created by Britain during colonization in the
late 19th century, which grouped together many different ethnic groups, political systems and cultures. There were difficulties in establishing a united political front after Ugandas independence from colonial rule in 1962, ultimately resulting in the dictatorships of military leader Idi Amin from 1971-1979 followed by Milton Obote from 1980-1985, who engaged in guerrilla warfare and human rights abuses. During this time the country plunged into political and economic instability, and there was total breakdown of law and order. Conflict claimed the lives of at least 600,000 Ugandans between 1971 and 1986 (Hamilton, et al., 2000). Civil strife coupled with increasing demand on resources led to diminished forests and wildlife in protected forests (Turyahabwe & Banana, 2008).
The current president of Uganda is Yoweri Museveni, who has been in power since 1986 and has brought relative stability with a developing economy. There is currently an armed
20


rebellion in the Northern region that has displaced hundreds of thousands. In addition, there can be spill over from the violence occurring in the DRC. Due to recent conflicts between government forces and the Lord's Resistance Army, there are currently an estimated 32,447 displaced residents in northern Uganda. This region remains politically and economically insecure, due to fighting, which resulted in an estimated 1.8 million people residing in internally displaced persons (IDP) camps (CIA, 2013) during the height of the conflict in 2011.
The country is split up into 4 regions which represent very diverse ecoregions. The Western region is characterized by Terrestrial Ecoregions of the World (TEOW) as being primarily being tropical and subtropical moist broadleaf forests. The North and Central regions are tropical and subtropical grasslands, savannas and shrublands, also referred to as Victoria Basin Forest-Savanna Mosaic. The North region which is more dry and arid than the rest of the country is comprised of East Sudanian Savanna and thickets as well as Northern Acacia-Commiphora bushlands. The Central region is more populated and houses the capital Kampala. The Eastern region is diverse with tropical and subtropical moist broadleaf forests near Mt. Elgon and tropical and subtropical grasslands, savannas and shrublands (Figure 2; Olson, et al., 2001).
21


Ecoregions by Olson(2001)
| Albertine Rift Montane Fcrests | East African Montane Fo'ests | East African Montane Moorlands East Stdaman Savanna | Lake Afrctropic
__Northern Acacia-Commiphora Bushlands And Thickets
Northern Corgolian Forest Savanna Mosaic Ruv/enzon-Vircnga Montane Moorlands Victoria Basin Forest-Savanna Mosaic
Figure 2: Olson 2001 Ecoregions with Uganda political region boundaries
The total surface area of Uganda is 241,550 km2. The current estimated population of Uganda is 31.8 million (Table 1). The capital is Kampala, which has a population of approximately 1,863,000. The average annual population growth rate from 2010-2015 was 3.3%. Agriculture accounts for about 25% of the gross domestic product (GDP), 50% of total exports and employs almost 75% of the labor force (IMF, 2010). Approximately 70% of agriculture produced is for subsistence purposes (Ruhanga & Manyindo, 2010). Most smaller farms mix commercial crops such as coffee and tea with subsistence staples include mostly starches such as maize, rice, millet, plantains, and cassava (flour) (Benson, et al., 2008; Chiputwa, et al., 2015). Coffee is the largest earner of foreign exchange, followed only by the tourism industry. Consequently, there is growing interest and investment in ensuring the protection of the resources supporting tourism, particularly wildlife conservation.
22


Table 1: Demographic and economic information for Uganda: Source: (UNDP, 2015; CIA, 2013; UNSD, 2016)
Parameter Value Year
Population, total (millions) 38.8 2015
Population density (persons per square km) 160.7 2015
Urban population growth rate (average annual 5.4 2010-2015
Rural population growth rate (average annual %) 3.0 2010-2015
Kampala population 1.936 2015
Urban population (%) 16.1 2015 est.
Gross domestic product (GDP) (million current US$) 26444 2013
GNI: Gross national income per capita (current US$) 681.5 2013
Forested area (%) 14.1 2012
Tourist arrivals at national borders (000) 1206 2013
Tourism (%) increase 17.5 2005-2013
Most of the population is rural (87%) and impoverished; about 10 million people fall under the national rural poverty level and rely heavily on the health and resilience of the ecosystems which support their livelihoods (Nabanoga, et al., 2010; IFAD, 2012). Forests and woodlands, in particular, are essential for employment, economic growth, and basic household uses (Obua, et al., 2010; Waiswa, et al., 2011; Brickwell, et al., 2012). Therefore, continued access to forest resources will be necessary for securing rural livelihoods (Bush, et al., 2005).
Forests have high agricultural potential which makes them particularly vulnerable to encroachment for agricultural production (Nakakaawa, et al., 2011). Between 2000 and 2005, the annual rate of forest loss in Uganda was reported to be 2.2%, which is among the highest globally (Okello, 2014; FAO, 2006). If the current rate of forest loss continues, Uganda will lose all of its remaining forests by 2040 (NEMA, 2014). The drivers of deforestation vary across Uganda, but the largest contributors have been identified as agricultural expansion, followed by wood fuel collection, and timber harvesting (IMF, 2010). According to the government of
23


Uganda, wood fuel demand drives deforestation in the northern and eastern regions, while timber and agricultural demand drives forest loss in the Central and West regions (Brickell et al. 2012). These various threats, in addition to unsustainable harvesting of resources, pose serious challenges to forest conservation, resource degradation, and livelihoods (Nabanoga, et al., 2010).
In Uganda, forest conservation is undertaken through an established protected area system (Figure 2). During colonialiation authorities seized ownership of forests and wildlife, prohibiting hunting and use of park resources, and imposed national park boundaries over many traditional hunting and grazing lands. Wildlife today, particularly large species, survive mainly in isolated national parks (Naughton-Treves, 1999). These parks and wildlife belonging to the government are a remnant of the colonial history of Uganda. While these areas tend to bring in significant revenue through tourism there is still resentment towards these boundaries by local residents who reside on the edge of parks, and frequently complain that the government's animals raid and destroy their crops for which they are never compensated and cannot defend against because they are protected by the law (Naughton-Treves, 1997).
About 35% of Ugandas forest is government regulated. 18-20% of the country's forests are National Parks or wildlife refuges which are designated as strict nature reserves. Central Forest Reserves (CFR), these are low-intensity zones allowing for some sustainable harvesting and non-consumptive use, account for the remaining 15- 17% of state-controlled forests (Peskett, et al., 2011). This situation leaves the majority of forest on private or communal land (Brickwell, et al., 2012). In many cases of established conservation in Uganda, access to and control of resources and landscapes has been taken away from local people.
In Uganda, four of the major national parks, Kidepo Valley National Park, Murchison Falls National Park, Lake Mburo National Park and Queen Elizabeth National Park are primarily
24


grasslands. Mt. Elgon National Park, Bwindi Impenetrable National Park, Mgahinga Gorilla National Park, Kibale National Park and Semuliki National Park are forested. Two national parks are mountainous: Mt. Rwenzori National Park and Mt. Elgon National Park.
Uganda energy use
In Uganda traditional biomass, predominantly firewood, charcoal, and agricultural residues account for over 90% of the national domestic consumption (Khundi, et al., 2011; Okello, 2014; Egeru, 2014). Modern fuels, such as electricity and other petroleum-based fuels contribute less than 10% of energy use. Charcoal is more common in urban households, while firewood and agricultural residues are more common in rural households. Wood fuel is also commonly used for small-scale industrial production activities (e.g., brick making, agroprocessing, and charcoal production). Most of the firewood used is harvested or collected from natural forests on communal, government or private land is largely unregulated (Tabuti, et al., 2003).
The largest consumer of wood energy in Uganda is the residential sector, with firewood being the most prominent used fuel type (Okello, 2014). Wood fuels are typically burned in inefficient traditional stoves open-fire stoves. Traditional stoves are used by about 87.5% of households (Egeru, 2014; Okello, 2014). 72.7% of the population use the 3-stone stoves, and 14.8% of households use traditional charcoal stoves (Byakola and Mukheibir, 2009).
Ugandans consumes an estimated 18 million tons of firewood and 500,000 tons of charcoal annually (Egeru, 2014). Reduced forest resources in Uganda may lead to increases in poverty due to higher wood fuel costs, and time spent collecting wood (Kazoora, et al., 2008). Current wood fuel collection practices are nearly always impacted by high rates of forest loss,
25


usually not driven by wood fuel collection. The demand for traditional wood fuel is reported to be growing at a rate ranging from about 2.5% to 6% annually (Tabuti, et al., 2003).
Because there are high levels of deforestation in Uganda, there are concerns about the sustainability of wood fuel supplies, especially with a rapidly growing population. Additionally, the increasing harvesting demand may lead to additional deforestation, resulting in overall environmental degradation (Okello, 2014; Naughton-Treves, et al., 2007; Egeru, 2014; Tabuti, et al., 2003). Given the heavy reliance on wood fuel, the anticipated depletion of forest stocks could prove to be a threat not only to the economic welfare of rural households but also to ecosystem health and biodiversity (Bailis, et al., 2015).
The use of efficient stove technology is becoming more common in Uganda due to the promotion of renewable energy and energy efficiency programme supported by the Uganda government and backed by the German Agency for International Cooperation. Since 2005, an estimated 500,000 stoves have been distributed in Uganda., with a goal of providing 250,000 more improved charcoal stoves and 4 million more improved wood stoves to households by 2017 (Kees & Feldmann, 2011; Okello, 2014).
The UGASTOVE project produces improved charcoal stoves and wood stoves for a variety of purposes in Uganda. Adkins et al. (2010) give a detailed description of the UGASTOVE design. The UGASTOVE started in 2004 and received carbon financing for emission reductions in 2009. The goal of the project was to distribute 180,000 stoves within 7 years which would decrease carbon dioxide released into the atmosphere by approximately 600,000 tons. The project far appears to have successfully distributed household stoves in areas experiencing high levels of deforestation (Simon, et al., 2012). Even with these projects and goals in place the current use of improved stove technology in households in Uganda is still quite
26


low (Okello, 2014). Byakola and Mukheibir (2009) estimate that 8.7% of Ugandan households use improved stoves (Byakola & Mukheibir, 2009).
27


CHAPTER IV
DATA AND METHODS
Data
The objectives of this study were to determine differences in household fuel and stove use, analyze the potential benefits of improved stoves, and explore the social and environmental context related to improved stove adoption. To address these objectives, household and community survey data were obtained from the 2005/06 Uganda National Household Survey (UNHS) and 2010/11 Uganda National Panel Surveys (UNPS). SERVIR (RCMRD, 2015) land cover maps (2000-2014), as well as protected area boundaries (NFA, 2007) which are data derived from remote sensing, were used to explore community-level resource availability. Most studies on rural fuel and stove use in developing countries are very localized or done at regional scales with small survey sizes (Adkins, et al., 2010; Cooke, et al., 2008; Egeru, 2014; Foell, et al., 2011; Lewis & Pattanayak, 2012; Mahapatra & Mitchell, 1999; Okello, 2014; Otieno & Buyinza, 2010; Pouliot, et al., 2012; Wallmo & Jacobson, 1998). This study broadens the scale of analysis to the national level by utilizing national panel surveys and data derived from remote sensing.
Social survey data
Survey data are necessary to address questions related to the social context associated with adoption of improved stove technology. These surveys can provide information on what influences household level decisions, including socio-economic status. The 2005/06 Uganda National Household Survey (UNHS) and 2010/11 Uganda National Panel Surveys (UNPS) are multi-purpose household and community surveys designed and funded by World Bankss Living Standard Measurement Study (LSMS) project and administered by the Uganda Government.
28


Both UNHS 2005/06 and UNPS 2010/11 were carried out annually, over a twelve-month period for the purpose of accommodating the seasonality associated with agriculture and consumption. The surveys are designed to interview each household twice each year, in visits six months apart.
The two surveys were designed using a two-level stratified random sampling design applied to select the enumeration area at the first level and households at the second level. Both survey samples cover the entire country and were selected in such a way that they are representative of the entire population of Uganda, for urban and rural populations, and the four regions: Central, Eastern, Northern and Western. The 2010/11 UNPS sample includes approximately 3,200 households distributed over 322 enumeration areas. All of the enumeration areas included in the 2010/11 survey and most of the households were re-interviewed from the 2005/2006 UNHS. The 2010/11 sample also includes additional randomly-selected households allowing for a representative sample at the national, urban/rural, and regional levels. The UNPS 2010/11 tracked and re-interviewed a sample of 2,716 households that were previously interviewed for the UNHS 2005/06.
The data used in this study was derived from the household and community modules. The household questionnaire gathered information on socio-economic characteristics. This includes information on the primary stove type used for cooking in 2010, 2005 and retrospectively in 2001. Types of stoves differ between the 2005/06 and 2010/11 surveys. Therefore information reported on stove use in 2005/06 and 2001 was only used descriptively due to its limitation for detailed analysis. The analytical part of the study utilizes respondent information on stove use during the survey in 2010/2011. The stoves were grouped into three categories including efficient wood, charcoal and traditional. The community survey aimed at collecting information related to the characteristics of communities residing in the sampled enumeration areas.
29


The 2010/11 survey also provides modified GPS coordinates. This allows for inclusion of spatial variables while preserving the confidentiality of the surveyed households and communities. The locations have been generated following the method used by the Demographic and Health Surveys (DHS) program. The coordinates included represent a random offset of the center-point of clustered household GPS locations within the enumeration area and within a certain distance depending on weather that community is designated urban or rural. An offset if 0-5km was used for most rural locations due to households being more dispersed. Therefore taking into account the 0-5km offset the GPS coordinates are representative at the cluster level and can be used for spatial analysis and merging data acquired from remote sensing.
Figure 3: Rural community GPS survey points
Environmental data
From a social science perspective, remote sensing can be useful to provide data on the on the biophysical or environmental context that shapes social events. In other words, remotely
30


sensed data can offer an additional layer for multilevel analysis by providing the biophysical context within which people live and relate to the environment (Liverman, et al., 1998). The environmental data used for this study include forest cover, percent forest change (2000-2014) and protected areas (National Parks and Wildlife Refuges).
Forest cover and forest change information are necessary to determine how resource availability is related to the adoption of improved stove technology. Land cover maps developed by SERVIR for the Green House Gases Inventories project provided information on forest cover (2014) and forest change (2000-2014) (RCMRD, 2015). These maps were commissioned to provide baseline land use / land change data for the forestry sector. The maps were created using LANDSAT 5 thematic mapper (30 by 30-meter resolution) imagery for Eastern and Southern Africa. The land cover maps include years 2000 and 2014. The 2000 land cover map includes imagery collected between 1999-2001 and the 2014 map includes data from 2013-2015. The overall accuracy of the data are 86.33% (436/505) and the KAPPA coefficient is 0.8444 (RCMRD, 2015).
Methods
This chapter introduces the conceptual framework used to address this study, followed by the specific research questions derived from this framework. The last section introduces the methods employed to address the questions as well as the variables incorporated into the analysis. This study focuses on traditional and improved stoves which utilize firewood and charcoal fuels, due to these being the predominant fuels used in rural households. Improved stoves were considered to include efficient wood stoves and improved charcoal stoves. Modern stoves werent included as improved stoves, due to the small sample size of households who use
31


these stoves in rural areas. Statistical analysis used in this study includes descriptive statistics and logistic regression analysis.
Conceptual framework
The relationship between access to resources, biophysical characteristics and the decision to use an improved primary stove was examined with the assumption that communities with a higher demand for fuel would be the most likely incorporate efficient technology. Traditional wood fuel use is associated with a variety of development challenges, including health, economic and environmental concerns. This research applies a human-environment framework aimed at improving the understanding of how socio-economic status, resource availability and biophysical characteristics influence household choice to incorporate improved cooking technology. To understand how the physical environment relates to household-level choices, we must first understand the likely costs and benefits of switching to improved stoves. Applying the utility theory, the cost and benefit of switching to improved stoves can be analyzed simultaneously, if the benefits outweigh costs the household will switch stoves.
I hypothesize that households will be more inclined to switch to improved stoves when demand for fuel is higher (which reflects resource availability and higher prices of all fuel types) and less inclined to switch if households are more remote and there are plentiful resources. The households that have fewer available resources may view improved stoves as a benefit because they may save time in collecting firewood and save money because of greater efficiency in burning wood or charcoal. Households that have abundant resources, on the other hand, may not recognize the time savings as a valuable benefit. Potential costs preventing the adoption of stoves will likely be poverty (poor welfare) as less well-off households will likely to be less inclined (or able) to switch stoves.
32


Research questions
In order to better understand how social and environmental context relates to the utilization of improved stove technology this study addresses the following questions:
1. Are stoves being adopted and utilized over-time? (temporal analysis)
2. What are the spatial differences in improved stove utilization? (spatial analysis)
3. Why are households utilizing or not utilizing improved stoves? What are some of the potential benefits of improved stoves? (explanatory analysis)
4. What are the main ecological and socio-economic variables which drive the utilization of improved stoves? (analysis of relationships)
Temporal analysis When?
In order to determine the temporal trends in improved stove adoption the two surveys, which include information on primary stove type used in 2000/01, 2005/06, 2010/11, were merged used. The UNITS 2005/06 data that are provided could be linked with the 2010/11 data at the household levels through the unique household identifier (HffiD). A variable for household status was included in the UNPS 2010/11 data which identifies (i) households from the 2005/06 survey that were interviewed at the same location, (ii) households from the 2005/06 survey that had moved and were interviewed at an alternative location. After the rural household survey data from 2005/06 and 2009/10 were merged, a total of 1,743 households in 233 enumeration areas were available for the present analysis.
In the UNITS 2005/06, households were asked what type of stove was primarily used for cooking in 2000/2001. Therefore, all households sampled both in the 2005/06 and 2009/10 surveys were used to assess changes in cooking technology and fuel use between 2001 and 2010. The 2005/2006 survey responses for primary stove options did not relate directly to the
33


2010/2011 survey data (Table 2). Therefore, some assumptions were made to permit comparison of these data sets and determine differences over time. The most important of these was that improved charcoal stove in the 2006 survey is the same as the charcoal stove in the 2010 survey. Among the households with primary charcoal stoves in 2010; 17% reported having improved charcoal stove in the 2006 survey, 35% reported having a traditional metal stove, and 37% had traditional 3-stone. This is concerning because traditional metal may have been considered charcoal stove by some respondents in the 2010/11 survey.
Table 2: Modifications made to relate household survey data from 2005/06 an 2010/11
Response Measure
Primary Stove
Which is the 2010 Wood / Sawdust Burning, Open fire l=Traditional
stove used most often by this Efficient Wood Burning 2= Efficient wood
Charcoal 3=Improved charcoal
nousenoiu: 2005 Traditional metal stove Traditional 3-stone stove l=Traditional
Improved firewood stove 2= Efficient wood
Improved charcoal stove 3=Improved charcoal
Spatial analysis Where?
In order to determine spatial relationships of improved stove utilization the hotspot analysis too in ARCGIS was used to calculate the Getis-Ord Gi statistic for every community point location in the 2010/11 survey dataset. The output provides z-scores and p-values which provide information on significant hot spots (high values) and cold spots (low values) based on spatial clusters. The results show how point location relates within the context of neighboring points. For example, in order to be considered a statistically significant hot spot, a location with
34


improved stove use will be surrounded by other point locations with improved stoves (Getis & Ord, 1995).
Explanatory analysis Why?
To understand how stove choice is related to environmental and social factors it is necessary to understand the likely costs and benefits of switching to improved stoves. Why are households utilizing or not utilizing improved stoves? What are some of the potential benefits of improved stoves? What are the regional differenced in fuel and stove use? Examining regional differences in fuel and stove use can provide information on how location characteristics may influence improved stove use. The hypothesis was addressed that regions with less resources (wood for fuel) would be more likely to adopt and utilize improved stoves. Data used to examine regional differences in fuel and stove use and potential benefits of improved stoves include variables selected based on providing a narrative about factors that may be driving improved stove use. The 2010/11 survey was used for these questions which includes 1,970 rural households and 233 enumeration areas. The variables selected to represent resource availability include firewood collection time, fuel types used for cooking, number of fuels used for cooking, source of fuel, time spent cooking, fuel quantity used, household stove location and presence of a chimney. Regional stove use should correspond to more diverse fuel types used for cooking and more sources, under the assumption that thus represents less resources availability. The potential benefits of improved stoves and how this may be influencing stove utilization efficiency of improved stoves, was examined under the hypothesis that if benefits improved stoves outweigh the costs the technology would be more likely to be utilized and maintained over time.
35


Analysis of relationships What?
A logistic regression model was used to investigate what ecological and socio-economic variables drive the utilization of improved stoves (efficient wood and charcoal stove). Each household is labeled yi =1 if they responded as utilizing an improved stove as their primary stove used for cooking, 0 if they primarily utilize a traditional stove model. I excluded respondents that used other models or had no stove, leaving a sample size n of 1,822 for efficient wood stoves, and 1,904 for charcoal stoves. I model the probability that y = 1 using,
Pr(yt = 1) = logit-1 (At/?),
Where X included 17 household and community level predictors. Due to nonlinearity of the logistic regression model, coefficients can be difficult to interpret. The curve of the logistic regression required a choice where to evaluate changes, whether to interpret the results on a probability scale. Probabilities are calculated from the equation,
1. a + (3x = 0
2. logit-1(a + (3x) = 0.5
3. /?e/(l + e0)2 = /?/4
Accordingly p/4 represents the maximum difference in Pr(y =1) corresponding to a single unit difference in x. Therefore dividing the coefficients by 4 gives an upper limit of the predictive difference corresponding to one unit of difference in x (Gelman & Hill, 2007).
Incorporating interactions of ecosystems and human activities at different scales, from regional, community and to household, is important in remote sensing and social studies (Semeels, et al., 2007). The processes operating at these various scales are mostly nested and interdependent. For example, ecosystem functioning might influence micro-level household
36


processes, and so these factors should thus be analyzed simultaneously. I have included variables at 3 different spatial scales (1) the household nested within (2) a community, which is nested within (3) a region. The results will determine the relative importance of factors driving the household decision to adopt improved stove technology at the household level. These types of models that incorporate both ground-based social data and biophysical characteristics from remote sensing are able to determine relationships on a much larger scale. Because the data are at different scales, each household within the enumeration area (community) was given identical community and biophysical characteristics, due to not having the exact location of the households. The logistic regression gives a single outcome, the probability that a respondent adopts an improved primary cookstove. The statistical analyses were run using R statistical software package (R Core Team, 2013).
Description of variables chosen for analysis Household characteristics
The household level variables included in the analysis represent demographic and socioeconomic characteristics. The indicators selected for analysis were found to be significant in other similar studies or have not been studied extensively. The data selected from the surveys include: (i) household information (number of people in a household, tenure security, female head of household, number of fuel sources); (ii) welfare indicators (measures taken to make water safer, whether every member of the household has at least one pair of shoes and two sets of clothes, and whether every child in the household has a blanket. Welfare indicators represent poverty in place of income; income can be subjective in rural areas where there are many subsistence farms and welfare may not be best reflected as a monetary value. Choice of an improved stove was predicted for households with improved welfare, and those with larger
37


family size, larger parcel size, and greater number of fuel sources. Since investment in improved cooking technology typically requires having the means to purchase the stove households with improved welfare were assumed to be more likely to use modem technology. Households with many members were also assumed to be likely to adopt improved stoves due to the benefits of long term cost savings due to fuel efficiency. Number of fuel sources was chosen as a variable to potentially represent resource availability.
Table 3: Household socio-economic characteristics and indicators
Variable Type Description Count- Count-No
Name Yes
Charcoal stove binary Yes/No 134 1821
Efficient wood stove binary Yes/No 52 1903
Tenure security binary Yes/No 189 1754
Female head of household binary Yes/No 587 1368
Min Max Mean
Safe water continuous More Safe/Less safe 0 3 0.63
Number of fuel continuous Purchase/Collect village/Collect 0 3 1.18
sources own land
Welfare measure continuous 0=High poverty l=Moderate 2=Low /not 3=Not impoverished 0 3 1.49
Household size continuous Number of people 1 30 6
Parcel Size continuous Acres 0 200 2.37
Some of the variables from the survey were aggregated for the purpose of simplifying the analysis. Several welfare indicators were provided in the 2010/11 household survey (Table 4). Individually these welfare indicators may not provide meaningful information about the poverty level of the household, and many of them may be correlated within a household (if one indicator is present in an impoverished household more than one may be as well). Therefore, a welfare measure was obtained by adding the number of indicators present within a household. Only the yes or no indicators were used that were obviously associated with poor welfare.
38


Table 4: Variables used to measure welfare from survey welfare indicators
Response Measure
Welfare Measure
Does every member of the Yes=l 0= very low
household have at least two No=0 l=low
sets of clothes? 2=moderate
Does every child in this Yes=l 3 or (2+NA)= high
household (all those under No=0
18 years old) have a blanket? NA
Does every member of the Yes=l
household have at least one pair of shoes? No=0
Safe Water
What is the main source of Private connection to More Safe =1
water for drinking for your household? pipeline Public taps Protected well/spring Vendor/Tanker truck Private connection to pipeline Public taps Protected well/spring
Bore-hole Vendor/Tanker truck
Unprotected well/spring River, stream, lake, pond -Or- Boil and filter
Gravity flow scheme Boil only
Rain water Filter only
What do you do to the water Nothing is done
to make it safer for Boil and filter Less Safe = 0
drinking? Boil only Filter only Nothing is done -And- Bore-hole Unprotected well/spring River, stream, lake, pond Gravity flow scheme Rain water
Community characteristics
Community characteristics considered as important in influencing choice of stove for cooking by a household were specific region, distance to protected areas (National Parks and Wildlife Refuges), distance to infrastructure (major roads, markets and population centers
39


+20,000), forest cover (2014) and forest cover difference (2000-2014) and number of forest crop within the community (Table 5). Access to alternative energy sources have a location dimension e.g. charcoal is prominent closer in proximity to urban areas. Distance to a population center and markets may inadvertently represent higher population densities. It is likely that communities located closer to population centers would contain more people, due to urban sprawl. Rural communities located closer to population centers would likely either have less available resources due to the greater demand for fuel from the larger population.
Also included in the explanatory variables, was distance to protected areas, forest cover, and forest crops, all chosen to represent resource availability. There are different types of protected areas with differing management practices and restrictions. Certain kinds of protected areas allow subsistence resource extraction. National Parks and Wildlife Refuges, however, are
Figure 4: Map of Uganda showing the major protected areas; Data Source: Uganda Wildlife Authority 2007
40


restricted access areas and allow no resource extraction; these boundaries most likely represent limited resource availability (Figure 4).
Table 5: Community and environmental variables representing resource demand and availability
Variable measure min max mean
FC Area 2014 meters2 0 10,521,085 2,576,945
Dist. to NPAVLR meters 0 188,155 58,238
Pet. FC Diff % -1 98.76 2.14
FC Diff meters2 -8160513 5840830 -368707
Avg. time spent collecting firewood Hrs./7days 0 30 3.08
Number Tree Crops count 0 4 1.67
Distance to population center (+20000) km 0 101.28 25.47
Distance to main road km 0 40.43 8.75
Distance to market km 0 116.18 33.30
The forest cover was clipped to a 5 km buffer surrounding the rural community survey point location. The buffer distance of 5 km was chosen based on the cluster methodology which derived the point locations (see social survey data) (Figure 5). In addition, the total forest cover was calculated for within the 5km buffer area for 2000 and 2014, and the difference was then determined for the 14-year period.
A few studies found that households generally do not travel far distances to collect fuel unless resources are scarce (Tabuti, et al., 2003; GoU, 2002; Mahapatra & Mitchell, 1999). Tabuti et al. (2003) found that in Bulamogi County Uganda that most people (residing in rural areas) traveled short distances of 2 km or less, to collect firewood, and that local people believe there is an abundance of firewood. Supporting this finding, the Uganda National Forest Plan (2002) found that Ugandans traveled less than two km to collect firewood. However, from 1992-2000 there was a dramatic increase in average time spent collecting, which may have resulted from the loss of forest resources (Figure 6) (GoU, 2002). The Mahapatra & Mitchell, (1999)
41


study conducted in rural India reported that people frequently traveled greater than 8 km to collect fuel where there is firewood scarcity.
Figure 5: 2000 and 2014 forest cover within a 5km buffer of a community survey point
1.4
1.2
_ 1
* 0.8 zj
I 0.6
X
- 0.4
0.2 0
Figure 6: Average distance travelled to collect household fuelwood; Source: (GoU, 2002)
1992
1999 Ki C c c





Central Eastern Northern Western Uganda Region
42


CHAPTER V
RESULTS
This study investigates the social and ecological context surrounding improved stove adoption. In order to achieve this it is necessary to explore the temporal trends, geographic differences, and potential benefits of improved stoves. This information can shed insight on whether improved stoves are being utilized as primary stoves and being sustained over time. Additionally, the potential benefits of adopting improved stove technology were investigated to determine the reasons for improved stove adoption as well as to explain why households are not maintaining improved stoves as the primary stove over time. The benefits considered align with the Global Alliance for Clean Cookstoves (2011) statements, which include less time spent cooking, less fuel used due to higher efficiency, less time spent collecting firewood due to less fuel used, and the potential health benefits due a reduced amount of indoor air pollution.
The temporal trends of stove use
Descriptive statistics were used to determine stove change from 2001 to 2010 using the 2005/06 and 2009/10 survey responses. Exploring the stove use trends over time can give insight to the social factors influencing stove adoption and long-term use. Overall, the number of improved and modem stove users is increasing over time, and traditional stove users are decreasing (Figure 7). However, households adopting the improved stove technology are not maintaining the technology as their primary stove over time (Figures 8 and 9).
43


140
120
100
80
60
40
20
0
Sum of Modem
Sum of Efficient Wood
Sum of Charcoal
Figure 7: Count of stove use 2001, 2006 and 2010, n = 1852
Figure 8 shows that out of the 26 households who claimed utilizing efficient wood stoves as their primary stove in 2001, 22 still used them as their primary stove in 2006. By 2010 none of those households still listed them as their primary stove. Instead, those 21 households switched back to using a traditional 3-stone stove as their primary stove and 1 had a charcoal stove. Of the additional 20 households who adopted efficient wood stoves between 2001 and 2006, only 3 still listed them as their primary stove in 2010; one household reported having a charcoal stove and the rest switched back to traditional stoves. Between the 2005/06 survey and the 2010/11 survey an additional 41 household adopted efficient wood stoves as their primary stove.
44


o
JS
0>
s
o
n
4-
o
-
a>
.a
S
s
45
40
35
30
25
20
15
10
5
<< 0
2001 2005 2010
Sum of 2001 Stove Adoption 26 22 0
Sum of 2005/06 Stove Adoption 0 20 3
Sum of 2010/11 Stove Adoption 0 0 41
Figure 8: Adoption of efficient wood stoves 2001-2010 n=1852
There were more improved charcoal stoves users in 2001 who maintained long-term use as their primary stove through 2010 (Figure 9). Out of the 70 households who adopted improved charcoal stoves in or prior to 2001, 64 still stated them as their primary stove in 2005, 13 retained the technology in 2010. Of those who didnt retain use of the technology 50 switched back to traditional stoves. An additional 25 households adopted improved charcoal stoves in 2005 who didnt have them in 2001, 9 still had them as their primary stove in 2010. In 2010, 104 additional households adopted charcoal stoves as their primary stove. Both improved stove types spiked in primary use in 2010. This spike may be related to the Ugastove project receiving carbon financing in 2009 and upping distribution goals.
45


o
.s
1)
s
o
n
4-
o
-
0>
.a
S
s
120
100
80
60
40
20
2001 2005 2010
Sum of 2001 Stove Adoption 70 64 13
Sum of 2006 Stove Adoption 0 25 9
Sum of 2009 Stove Adoption 0 0 104
Figure 9: Changes in primary improved charcoal stove adoption 2001-2010
The spatial analysis of stove use
The results of exploring regional differences in stove use (Table 6) revealed that the Central region has the highest percentage of primary charcoal stove users (63% of all primary charcoal stoves) and modern stove users (57% of all modem stoves). The East has the highest percentage of primary efficient wood stove users (42% of all efficient wood stoves). Traditional stoves are for the most part evenly distributed throughout the country.
Table 6: Percent stove use by region
2010 stove use by region Charcoal Efficient Wood Burning Traditional Modern
Central 63.43% 25.00% 21.42% 57.14%
East 6.72% 42.31% 27.47% 0.00%
North 18.66% 21.15% 27.98% 21.43%
West 11.19% 11.54% 23.12% 21.43%
Total 100.00% 100.00% 100.00% 100.00%
Figure 9 shows the point survey locations where there are statistically significant clusters or hotspots and coldspots of improved charcoal and efficient wood stoves. The hotspot for charcoal stove adoption is more predominant in the Central region, especially concentrated
46


around the capital city Kampala. The hotspots for efficient wood stoves are more concentrated of around the protected areas, paticularly Mt. Elgon National Park in the East region, and in the western region, and near Bwindi Impenetrable Forest in the south western region, which is an important area for tourism revenue due to the Mountain gorilla (Figure 10).
a) b)
Hotspots
* Cold Spot 99% Confidence
* Cold Spot 95% Confidence o Cold Spot 90% Confidence o Not Significant
c Hot Spot 90% Confidence
* Hot Spot 95% Confidence
* Hot Spot 99% Confidence
H Protect edAreas (NPs & WLRs) Waterbodi es
0 80 160 320 Kilometers
1 _i_i_i__I_i__i_i_I
Figure 7: Results of a hotspot analysis showing statistical significance of clusters of improved a) charcoal stoves b) wood stoves
The potential benefits of improved stoves
The potential benefits of improved stove use include potential fuel saving of improved stoves, time spent cooking and collecting wood, resource availability (examined by fuels used and sources of fuel), and cooking practices (where the stove is located in the household and
47


whether there is a chimney. Table 7 shows a comparison of the quantity of firewood and charcoal used for different stove types as well as the average time spent per day using the stove. On average, households who primarily use improved stoves (efficient wood burning or charcoal) use much less fuel than traditional open fire or wood/sawdust burning stoves. Efficient wood stoves use approximately 32% less fuel that open fire stoves. This finding is also reflected in time spend collecting firewood compared with stove use. Charcoal stove users, who generally purchase fuel, spend less average time collecting firewood than wood stove users (0.18 hours/week). Not surprisingly, households with an efficient wood stove spend less average time collecting firewood (1.25 hours/week) than households with traditional stoves (1.52 hours/week), which is most likely related to using less fuel. The charcoal stove is less efficient than the wood burning stove but still more efficient than the open fire pit and wood/sawdust burning stoves.
The amount of firewood used by the wood/sawdust burning stove was similar to the open fire pit, much less charcoal was used. However, this type of stove is not common and charcoal may not be used often with this kind of stove, which could be why the numbers are less. Therefore, this analysis further justifies grouping wood/sawdust burning stove into the traditional stove category and classifying charcoal as an improved stove.
However, the efficient wood burning stoves were used more hours per day than the less efficient traditional stoves for both fuel types (18-32% more time than open fire). This finding is consistent with what Adkins et al.s (2010) report. This study also found that in western Uganda households spend on average 3.8 hours cooking per day, which is less than what the survey data shows for the national average.
48


Table 7: Average hours per day used and average quantity of firewood and charcoal used per month (30 days) based on primary stove type
2010 Firewood Charcoal
Stove Type Average Hrs. / Day Used Average Qty. Used Average Hrs. / Day Used Average Qty. Used
Wood/Sawdust burning 5.79 1350 7.16 920.45
Efficient wood burning 6.39 915.38 6.76 348.86
Charcoal 4.89 864.10 5.07 7855.92
Open Fire 4.84 1261.40 5.75 9996.16
Time spent collecting firewood is another indirect indicator of resource availability at the community level. Table 8 presents the average total time spent collecting firewood by region, including the source from where households reported getting their firewood. Therefore, only the averages were assessed for households who responded as collecting firewood either from their land or the village, this was to reduce potential errors due to people reporting time spent collecting firewood who purchase their fuel. Overall, households spend about 2 hours collecting fuel. This finding is consistent with Tabuti et al. (2003), who reported that in Bulamogi County in Uganda, most people spent less than 2 hours collecting firewood, commonly collecting daily or once a week. The time spent collecting by those who collect from their land is less than those who primarily get their fuel from the village. There is a lot of variance in fuel collection time in the Central region. The Western has high variance in fuel collection time from those collecting from the village.
49


Table 8: Average time spent collecting firewood of surveys who reported gathering or collecting firewood as their main source, by region.
Firewood Source 2010 Average Firewood Collection Time Standard Deviation Variance
Central rural
Collect from own land 1.54 1.79 3.22
Collect from village 2.40 3.78 14.26
East rural
Collect from own land 1.23 1.34 1.80
Collect from village 1.45 1.32 1.73
North rural
Collect from own land 1.53 2.02 4.09
Collect from village 1.91 2.16 4.68
West rural
Collect from own land 1.42 1.32 1.73
Collect from village 1.78 2.37 5.62
Understanding fuel characteristics by region can relate to how resources may influence household stove choices. Table 9 shows differences in rural fuel use from the 2010/11 survey by region. The results indicate that fuel characteristics differ by region. Reflecting the patterns reported from regional characteristics of stove use the Western region has a greater dependency on firewood use (88.06%). The Central region uses more charcoal (31% compared to the second highest the Northern region at about 12%). The Eastern region has a higher percentage of crop residues (17.73%) used as fuel. The Northern regions has the second highest percentage of firewood, charcoal and crop residue use.
50


Table 9: Rural fuel use by region
Rural fuels used for cooking 2010 Charcoal Crop residue Firewood Total
Central 31.16% 5.45% 63.39% 100%
East 8.51% 17.73% 73.76% 100%
North 11.54% 10.07% 78.39% 100%
West 5.33% 6.61% 88.06% 100%
Uganda (16.24%) (9.73%) (74.03%) (100%)
Table 10 reveals that most of Uganda households use no more than 2 fuels for cooking (99%). The Eastern region has the highest percent of households using 2-3 fuels for cooking (about 28%). The Western region has the least amount of fuel diversity with 91% of households relying on firewood.
Table 10: Multiple Fuels used for cooking by region
Rural Fuel Use 2010 1 fuel for cooking 2 Fuels for cooking 3 fuels for cooking Total
Central 434 79.34% 112 20.48% 1 0.18% 100.00%
East 358 71.89% 128 25.70% 12 2.41% 100.00%
North 395 77.91% 106 20.91% 6 1.18% 100.00%
West 432 90.76% 43 9.03% 1 0.21% 100.00%
Uganda 1619 79.83% 389 19.18% 20 0.99% 100.00%
Fuel source can also be an important variable to represent resource availability, in that households that purchase fuel are likely doing so due to a lack of access to natural resources. Fuel source may also be a factor influencing household stove choice. For example, if a household has to spend money to purchase fuel, they may be more likely to adopt a more efficient stove to save money. Six primary sources of fuel were recorded in the survey: (1) Purchase from a shop; (2) Purchase from the market; (3) Public utility; (4) The black market; (5) Collect or gather from your own land; and (6) Collect or gather from the village (Table 11). The
51


majority of firewood for all regions is collected from the village (69.16%) and 25.34% of households collected from their private land. The east rural and west rural households lean more towards collecting firewood on their property. Table 10 shows the percent of the total fuel used by the source of the fuel. Charcoal is primarily purchased (88.8%) from either a marketplace (73.5%) or a shop (15.3%). The black market (illegal trade) is mainly used for charcoal (5%). Firewood is mostly gathered or collected (91.3%), primarily collected from the village (67.2%), but also from the households own land (24.1%). Crop residue is primarily gathered from the households land but also collected from the village.
Table 11: Source of fuel normalized by 100% of each fuel-type category.
Source 2010 Charcoal Crop residue Firewood
Gather / collect own land 3.82% 72.25% 24.79%
Gather / collect village 5.10% 27.75% 68.70%
Marketplace 67.20% 0.00% 3.97%
Public Utility 2.23% 0.00% 0.21%
Shop 18.47% 0.00% 1.48%
Black market 3.18% 0.00% 0.85%
Total 100% 100% 100%
Overall, the Central region has more charcoal use, and the majority of charcoal is purchased. Its therefore not surprising that the Central region has the highest percent of households who purchase fuel (22%). The Western region, more reliant on firewood, has a greater percent of households who gather wood from the village (63%). The East region has the most fuel collection from private land (36%), probably due to higher use of crop residues. However, all regions still have a high percentage of households who are collecting fuel from the village (44% 63%).
The potential health impacts of traditional stove and fuel use was assessed by looking at the cooking characteristics of households. Houses chimneys or where the residents cook
52


outdoors may be less exposed to smoke from traditional stoves, and consequently may be less inclined to switch to improved stoves. The 2009/10 survey asked whether stoves had chimneys and where in the house stoves were located. The results in Table 12 shows that the vast majority of all stoves types do not have chimneys in Uganda (94% of charcoal stoves, 93% of efficient wood burning stoves, 96% of open fire pits and 100% of wood/sawdust burning).
Table 12: Percent of Stove types that have chimneys
Chimney Stoves rural No Yes Total
Charcoal 176 94.12% 11 5.88% 187 100%
Efficient Wood Burning 50 92.59% 4 7.41% 54 100%
Open fire 1704 96.71% 58 3.29% 1762 100%
Wood / Sawdust Burning 43 100% - - 43 100%
Stoves are primarily located in a separate kitchen (74%), although a surprising percentage of stoves are located outside (20%) which would not require a chimney for ventilation (Table 13). Just 6.45% of stoves were located in a main living space. Households that use traditional stoves without chimneys and are located indoors likely represent high levels of indoor air pollution; when that stove is in a main room of the house, it is probable that the entire family is being exposed.
Table 13: Where stoves are located in the household by region
Rural In a room in In a separate In an outdoor Total
Cooking the dwelling not kitchen space
practices just devoted to
cooking
Central 32 6.19% 337 65.18% 148 28.63% 517 100%
East 30 5.23% 466 81.18% 78 13.59% 574 100%
North 60 10.93% 341 62.11% 148 26.96% 549 100%
West 12 2.75% 381 87.19% 44 10.07% 437 100%
Uganda 134 6.45% 1525 73.42% 418 20.13% 2077 100%
53


The relationship between variables driving the utilization of improves stoves
The model was ran using a classical logistic regression in which stove use is binary variable (1= improved stove, O=traditional stove). Improved charcoal stoves and efficient wood stoves were separated for analysis since the use of charcoal and wood likely represent very different characteristics.
Analysis of efficient wood stoves
Table 14 contains results from a logistic model that was implemented using a binary indicator of efficient wood stove use as the dependent variable. Socio-economic variables as well as variables representing proximity to resource demand (roads, markets, and population centers) were not associated with the choice to adopt this type of cookstove. The most significant factor that helps to explain patterns of adoption of efficient wood stove technology was the distance to National Parks (NP) and Wildlife Refuges (WLR) (p < 0.001) which likely represent resource scarcity, in that those are restricted-access areas, or active stove distribution areas. Households located closer to National parks and Wildlife refuges have a higher probability of adoption efficient wood stoves as their primary stove type, every meter further distance away from a protected area corresponds to approximately 50% lower probability of utilizing the efficient wood stove as the primary stove. The Northern region (at p < 0.05) and Western region (at p < 0.001) were significant and negatively correlated to efficient wood stove adoption, indicating that households in these regions are less likely to adopt this stove model as their primary stove. Overall the Area Under the Curve (AUC) value, which ranges from 0.5 to 1.0 (0.5 indicates no better than a random prediction and 1.0 is a perfect fit) the model performance was moderate (AUC=0.69), indicating that the variables included in this model are about 70% of the variation in wood stove adoption.
54


Table 14: Logistic regression results for efficient wood stoves (n=1955)
Fixed Effects Coef.est Coef.se Pr(>|z|)
(Intercept) -2.110 0.890 0.018 *
c.HH Size -0.070 0.060 0.193
Female Head HH 0.030 0.310 0.936
Welfare Measure 0.180 0.140 0.205
Safe water Measure 0.300 0.250 0.229
Parcel Size -5.390 8.620 0.532
Secure Tenure -1.170 0.740 0.116
# Fuel Sources -0.420 0.330 0.198
Tree Crops Community -0.470 0.370 0.206
FC Difference (%) 0.000 0.020 0.835
n.Dist NP/WLR (m) -2.130 0.800 0.008 **
FC 2014 Area (m2) -0.290 0.890 0.745
Dist. road (km) 0.000 0.020 0.952
Dist. market (km) 0.000 0.010 0.799
Dist. pop. center +20000 (km) 0.000 0.010 0.812
Region: Eastern -0.150 0.460 0.740
Region: Northern -0.930 0.530 0.078 .
Region: Western -1.970 0.610 0.001 **
AUC 0.692
Significance codes: 0***0.001 0 01 0.05 V
Analysis of improved charcoal stoves
The results from the improved charcoal model suggest that household socio-economic characteristics are very much related to the utilization of improved charcoal stoves (Table 15). Factors identified as being positively associated with the adoption of charcoal stoves include older average household age, higher welfare, safer water and number of fuel sources, and distance to National Parks and Wildlife Refuges. The coefficient for household welfare is positive and significant (p < 0) implying that higher welfare is associated with households that are more likely to have a charcoal stove. This result corresponds to 15% greater probability of using improved charcoal stove with every unit of increasing welfare measure. Safe water is also an important variable, suggests that households taking extra measures to make their water safer are 17.5% more likely to incorporate improved stoves. This variable may also be viewed as an additional welfare indicator. The greater number of fuel sources corresponds to approximately
55


9% greater probability of charcoal stoves. Households larger that the centered value (5) are 3% more likely to utilize charcoal stoves. National Parks and Wildlife Refuges was highly significant and positively associated with improved charcoal stove use, this result can be interpreted as households that are further from these restricted use areas have a higher probably (70%) of incorporating charcoal stoves with every meter of distance; this relationship is opposite from the efficient wood stove results. This relationship is likely due improved charcoal stoves being more central, while most of the NP and WLR are more remote and located closer to the countys borders. There was also a significant negative relationship to forest cover, which means that areas adopting improved charcoal stove are less forested. Every meter2 less of forested area makes the households 50% more likely to primarily utilize a charcoal stove. Overall the AUC value, was quite high for the charcoal stove model (AUC= 0.87).
There was no relationship between having a female head of household and adoption of improved stoves in this study, however, other studies on clean fuel adoption have found the gender of the household head to be a relevant factor (Lewis & Pattanayak, 2012). The indicators of resource demand (distance to a population center, distance to a road, and distance to a market) were not significant predictors in the models. In other words, proximity to areas of high fuel demand does not have a statistically significant influence household's decision to adopt improved stove.
56


Table 15: Logistic regression results for charcoal stoves (n=1955, groups= region, 4)
Fixed Effects Coef.est Coef.se Pr(>|z|)
(Intercept) -4.990 0.710 0. .000 ***
c.HH Size -0.120 0.040 0. .007 **
Female Head HH -0.230 0.230 0. .320
Welfare Measure 0.610 0.110 0. .000 ***
Safe water Measure 0.700 0.180 0. .000 ***
Parcel Size -20.610 10.670 0. .053
Secure Tenure -0.450 0.410 0. .272
# Fuel Sources 0.360 0.200 0. .068
Tree Crops Community -0.210 0.250 0. .400
FC Difference (%) -0.010 0.020 0. .500
n.Dist NP/WLR (m) 2.820 0.600 0. .000 ***
FC 2014 Area (m2) -2.010 0.740 0. .007 **
Dist. road (km) -0.790 0.590 0. .183
Dist. market (km) -0.280 0.830 0. .740
Dist. pop. center +20000 (km) 0.160 0.850 0. .851
Region: Eastern -0.870 0.440 0. .048 *
Region: Northern 0.560 0.390 0. .148
Region: Western -0.460 0.480 0. .342 ***
AUC 0.873
Significance codes: 0***0.001 0 01 0.05 V
When comparing the averages of these variables, you can clearly see that efficient wood stove users have a much lower average welfare measure (Table 16). Parcel size was larger for efficient wood stoves users than charcoal stove users, which is likely due to households being more remote and rural. Households utilizing charcoal stoves have more fuel sources and higher safe water measure.
Table 16: Average of variables used in the regression analysis by stove type
Average of Variables Efficient Wood Stove Charcoal Stove Traditional Stove
Welfare Measure (0-3) 1.58 2.35 1.42
Parcel Size (acres) 1.78 1.05 2.49
Safe water measure (0-3) 0.65 1.13 0.59
Household age 21.5 19.7 21.8
Number of fuel sources 1.15 1.35 1.19
Dist. NP/WLR (meters) 47,340 113,042 54,407
FC Area 2014 (m2) 2,794,264 1,817,464 2,628,087
57


CHAPTER VI
DISCUSSION
While the use of improved stoves has increased since 2001 and traditional stove use has decreased, it appears that there may be a problem with households maintaining the technology as their primary stove over time. The majority of households that adopted an improved stove in Uganda ended up switching back to a traditional primary stove within 5 years. With all of the claimed co-benefits of improved stove technology, it appears that theyre falling short and therefore structures funded through the carbon trade may imposing unpractical and inappropriate technology on households.
One of the claimed co-benefits of improved stove technology verified by this study is fuel savings. On average efficient wood stove models use less fuel than the traditional counterpart (32%) and spend less average time collecting firewood. This finding is consistent with Adkins et al. (2010) who evaluated fuel efficiency of the two types of improved stoves in Uganda. The results of the study showed that both stoves were substantially more efficient burning firewood compared to the 3-stone stove. The Ugastove stove model burned 46% less wood, and the Stove Tec model showed 38% fuel savings (Adkins, et al., 2010). In regions with limited resources, fuel savings of 32% can have a significant impact on household security and reducing vulnerability. Fuel savings means less time collecting firewood or less cost of purchasing fuel, which places less of a burden on impoverished households.
However, another finding suggests that households spent on average more time cooking with efficient wood stoves (18-32% more time than open fire). Longer time spent cooking contradicts one of the co-benefits claimed by development projects and non-profits (Global Alliance for Clean Cookstoves, 2011). This finding is important because if households are
58


spending more time cooking, they may not perceive time savings collecting wood as a significant enough benefit not to switch back to the traditional stove. In addition, longer time spent cooking may lead to longer exposure to smoke making the claimed health benefits negligible. Based on the literature longer cooking time is not observed in all improved stove models, however, Adkins et al. (2010) found this to be the case, especially with the Ugastove model in Uganda, where improved stove models took anywhere from 5% to 27% more time to cook food than the traditional 3-stone stove. 80% of the survey respondents complained about the increased cooking time of the improved Ugastove as well as other reported inconveniences (i.e. too hot to touch, easy to knock over, difficult to light), and many preferred the traditional stove over the improved stove (Adkins, et al., 2010).
There are some potential limitations of these results including discrepancies terminology used for stove types between the two surveys (improved charcoal and charcoal). There are also likely technical as well as cultural limitations of improved stove models causing households to switch back to traditional primary stoves, which are outside the realm of this study. Switching back to traditional stoves could also be evidence of people selling/bartering their stoves to cover other expenses, especially if these stoves were received in kind from carbon or development projects. In addition, primary stove use as a binary variable may not be the best reflection of household stove choice, in that multiple stove types are likely being used simultaneously.
Stove stacking is also common practice in Uganda and there may be times when improved stoves are used more frequently than traditional (such as rainy seasons). There also could be technical malfunctions that limit the lifespan of the stoves or make them unpractical, with households reverting to back to the reliable traditional stoves. Masera et al. (2000) found that the main technical reason for households utilizing multiple stoves in rural Mexico is that
59


improved and modem stoves were not suitable for some cultural cooking practices. A popular local food for Ugandans is matooke, in which the process of steaming plantains is very traditional and involves filling the bottom of the cooking vessel with banana stems to create a space separation between the boiling water and the peeled plantains wrapped in leaves. The plantains are steamed until they reach a desired softness, which is tested by hand (Adkins, et al., 2010). A study by Atkins et al. (2010) found that cooking this dish with an improved stove took significantly more time than cooking it with a traditional stove. Stove stacking in rural areas can also be described as tension between maintaining a cultural identity and simultaneously incorporating western values. This has described this relationship as the autonomous culture where indigenous people maintain identifying elements, such as possessions, traditions, or ideologies, in order to feel a sense of control over their cultural spaces (Masera, et al., 2000).
A high percentage of Ugandans cook outdoors which makes reducing smoke from traditional stoves a lessened benefit. Fuel use characteristics and consumption of fuels within households may be due to regional differences, poverty, fuel scarcity or potentially cultural differences in cooking practices. The results from this study indicate that fuel and stove characteristics differ by region, which is likely due to climatic differences, land cover features, development and resource availability.
The more heavily forested, tropical Western region has a greater dependency on firewood use from collecting and gathering (88.06%). This region has the lowest percent of primary efficient wood stoves used and the highest percent of households who cook indoors. Which likely indicates that resources are plentiful in the western regions, however households are likely being the most adversely affected by indoor air pollution. Charcoal stoves are used more commonly as primary stove in the Central region. The Central region which uses the most
60


charcoal fuel (31% compared to the second highest the Northern region at about 12%), is primarily a mosaic of grasslands, savannas, and shrublands. The rural areas in the Central region may also have more charcoal use due to resource depletion (less forest cover) due to higher rates of population and development. The Eastern region is classified as tropical and subtropical moist broadleaf forests near Mt. Elgon and tropical and subtropical grasslands, savannas and shrublands; it has a higher percentage of crop residues (17.73%) used as fuel. The Eastern region also has the highest percent of efficient wood stoves and household who cook outdoors. This region also has the highest percent of households using multiple fuels for cooking (crop residue, charcoal and firewood) which may indicate resource scarcity. The Northern region has the highest percentage of stoves located in a main area of the house. This likely means that these families are the most exposed to indoor air pollution and this could also be an indicator of extreme poverty. The Northern region is also more vulnerable due to recent conflict which has displace a large portion of the population.
The results of the regression analysis to determine the relationship between various factors driving the utilization of improved stove technology showed that socio-economic and demographic characteristics were important determinants of households who use a primary charcoal stove but not for efficient wood stoves. Safe water and higher welfare measure were found to have a strong positive relationship to utilizing an improved charcoal stove. Safe water is likely related to welfare, therefore more affluent households are more likely to have access to safe water and subsequently improved stove technology.
This relationship may indicate that more impoverished households are not getting access to improved stove technology, potentially due to cost. This relationship could pose a problem where poorer households who are likely more vulnerable to adverse effects of traditional open
61


fire stoves, continue to worsen their situation. However, it is difficult to make an assumption about the causal relationship between household adoption of improved stoves and welfare. In other words, it is unclear whether improved welfare was followed by improved stove adoption or improved stoves were followed by reduced poverty. In addition, impoverished households without access to safe water are more likely affected by indoor air pollution, as well as being disproportionately more vulnerable to environmental disturbances such as impacts of climate change. Also, households without access to safe water sources may need to spend more time collecting and boiling the water, exposing the inhabitants to more smoke. Therefore, time savings and reduced smoke from improved models could be a more significant benefit for these households. Programs to promote improved stoves in areas without access to safe water could make progress in alleviating poverty and increasing the security of rural households that could benefit from improved stove technology.
The regression results also showed community level indictors of resource availability were significant in predicting the household choice to utilize charcoal stoves. Communities with less forest cover were more likely to use charcoal stoves are their primary cooking apparatus.
The distance to National Parks and Wildlife Refuges which represent protected areas with restricted access, were important predicators for both stove types. However with opposite relationships. The charcoal stoves were more likely to be utilized in communities further from National Parks and Wildlife refuges. Households in closer proximity to protected areas are more likely to adopt an efficient wood stove. This finding is consistent with a study by Wallmo & Jacobson (1998) who found that households adjacent to national parks in western Uganda tend to implement improved stoves due to a lack of access to resources.
62


There are some limitations with using a general logistic regression model does not account for spatial autocorrelation due to similarities of households within communities all having the same community and biophysical characteristics. Individual-level analysis while straight forward ignores group level group level variation beyond that explained community level variables such as forest cover or proximity to roads (Gelman & Hill, 2007).
This relationship between efficient wood stove use and proximity to National Parks and Wildlife refuges could also be due to the preferential distribution of stoves from carbon and development organizations coupling conservation goals or forest restoration projects. These organizations aimed at distributing improved stove technology may be influencing patterns of stove use by making the technology available to rural communities located close to protected areas which have perceived conservation importance such as Mt. Elgon and Bwindi National Forests. However, regardless of how whether or not the stove is tactfully distributed the choice to use the technology as the primary stove is still a household decision and therefore there must be a perceived benefit, whether that be fuel savings or health benefits.
Parks and protected areas can be controversial due to the inequitable social impacts. The zones around the park boundary tend to become high priority areas for environmental groups who tend to bring in international funding, to pursue conservation goals, often leading to contention and conflict. Protected areas that are set aside for strictly ecological and financial purposes may fail to consider the social impacts of their establishment and management. Although inclusionary management has become increasingly more common the past decade, in which communities are involved and have power over the use and management of resources. Exclusionary conservation is by far still more prominent in conservation efforts despite the
63


negative social impacts and the uncertain effectiveness of protected areas in protecting forests (Bates & Rudel, 2000).
Conservation and sustainability are difficult to achieve in times of immediate uncertainty whether related to extreme poverty, civil unrest or both. Uganda has had a long history or both. Beginning with British colonization where natural resources and property rights were suddenly confiscated and National parks and protected areas were established restricting access to resources and evicting local people from their ancestral land. To a series of militant dictatorships resulting in death and free for all resource grab, which occurred only in the recent past. There are certain conservation priorities such as the Bwindi National Forest and World Heritage Site, which is considered the most important forest in Uganda to westerners aiming to protect biodiversity including the endangered mountain gorilla, as well as by the government for economic incentives due to high amounts of tourism revenue. Like many of the naturally remaining forests in Uganda, Bwindi is surrounded by dense agriculture and very little forest remains outside the parks boundaries (Hamilton, et al., 2000). Due to the high reliance on wood fuel for energy, there are concerns regarding pressure to illegally harvest wood from inside parks and protected forests.
64


CHAPTER VII
CONCLUSION
The objectives of this study were to examine the social and environmental context related to the utilization of improved stove use in rural Uganda. The results of this study suggest that location, resource availability (i.e. proximity to protected areas, forest cover) and welfare play and important role in determining improved stove utilization. This study has observed a high dependency on traditional fuels and stoves used for cooking in rural Ugandan communities. While there has been an increase in improved stove use, there are definite spatial trends related to the utilization of these stoves which differ based on fuel type (firewood or charcoal). While variables representing welfare were found to have an influence on charcoal stove use, it appears as though environmental context is also an important factor influencing improved stove utilization in Uganda. Therefore, it is important to consider not only how formal structures influence household level decision-making, but also the influence of informal structures, such as resource availability.
Although improved stove use is increasing there may be issues as to whether the technology is being maintained as the households primary stove overtime. This raises questions as to whether the technology is practical and whether improved stoves technology under carbon financing are able to achieve both climate and development goals. Some of the findings presented in this study refute the potential for win-win scenarios or co-benefits promoted by organizations and development programs funded to implement improved stoves. Stove development projects may have a difficult time fitting technology solutions onto a diverse social-ecological landscape and as a result, there are mixed outcomes. While improved stove technology can offer various advantages, those technologies need to be adaptable and practical
65


within the household. Unfortunately, stove development projects tend to follow certain pathways of convenience, efficiency, and have particular prerogatives such as standardizing technology and distribution preferences and it may be difficult to overcome those tendencies.
As population continues to increase and fuel becomes scarcer due to increasing rates of deforestation, communities become more vulnerable and environmental degradation exasperates poverty. High levels of poverty and constant conflicts arising within the country and in the surrounding areas makes exclusionary and top down management of forest resources in order to protect the countries rare biodiversity and promote tourism difficult. Therefore any conservation efforts must be approached with consideration of the context of Ugandas historical and current political and economic instability. These efforts must be compatible with supporting local communities and alleviating poverty in order to be successful.
Being that traditional wood fuels are still predominantly used, with increasing demand for fuel and little forest remaining outside of protected areas there is a need to provide alternative fuel sources and efficient stove technology. Co-benefits and win-win policy objectives are possible, however there must be more consideration for location and demand for technologies to be utilized. Rural areas surrounding protected forests may be vulnerable due to restricted access to resources and alternative fuels. However, conservation efforts should also keep in mind the likely impacts of displaced environmental degradation from growing demand for charcoal from urban centers. Providing efficient stoves to urban populations may have more success in reducing forests loss. These households may also be more impacted by indoor air pollution due to dense and confined living quarters.
Access to clean and more fuel efficient stoves has the potential to be a critical element in reducing vulnerability of impoverished people, as well as prolonging resources and protecting
66


the countries biodiversity. An important lesson to be drawn from this study is that informal structures outside direct policies promoting conservation and development are likely influencing the success of improved stove projects, which depends solely on households utilizing the technology. Household decisions likely depend on a number of internal and external factors, such as culture, resource availability, or welfare. It follows that future consideration should be given to the context within which communities are embedded, which will likely influence the successful utilization of improved stoves. When lofty polices are implemented aiming to promote development or progress in the global south, while also reducing global emissions and deforestation this often results in grouping populations of people which over-simplifies solutions into a one-model fits all approach. Which raises questions such as; what is the ultimate goal of improved stoves? For whom are these improved stoves manufactured? At whose cost?
67


REFERENCES
Adams, W. M. & Hutton, J., 2007. People, Parks and Poverty: Political Ecology and Biodiversity Conservation. Conservation and Society, 5(2), pp. 147-183.
Adkins, E. et al., 2010. Field testing and survey evaluation of household biomass cookstoves in rural sub-Saharan Africa. Energy for Sustainable Development, Volume 14, pp. 172-185.
Aggrey, N., Wambugu, S., Karugia, J. & Wanga, E., 2010. An Investigation of the Poverty-Environmental Degredation Nexus: A Case Study of Katonga Basin in Uganda. Research Journal of Environmental and Earth Sciences, 2(2), pp. 82-88.
Arnold, J. E. M., Kohlin, G. & Persson, R., 2006. Woodfuels, livelihoods, and policy interventions: changing perspectives. World Development, 34(3), pp. 596-611.
Bailis, R., Drigo, R., Ghilardi, A. & Masera, O., 2015. The carbon footprint of traditional woodfuels. Nature Climate Change, Volume 5, pp. 266-272.
Bates, D. & Rudel, T. K., 2000. The Political Ecology of Conserving Tropical Rain Forests: A Cross-National Analysis. Society & Natural Resources: An International Journal, pp. 13:17, 619-634.
Benson, T., Mugarura, S. & Wanda, K., 2008. Impacts in Uganda of rising global food prices: the role of diversified staples and limited price transmission. Agricultural Economics, Issue 39, pp.513-524.
Brickwell, E., McFarland, W. & Mwayafu, D. M., 2012. Unlocking progress on REDD+: sector coordination in Uganda, London: Overseas Development Institute.
Bush, G. K., Nampindo, S., Aguti, C. & Plumptre, A. J., 2004. Valuing Ugandas forests: A livelihood and ecosystems approach, Kampala, Uganda: Wildlife Conservation Society.
Bush, G., Nampindo, S., Aguti, C. & et al, 2005. The Value ofUnganda'sforests: A livlihoods and ecosystems approach, Natonal Forest Authority Kampala, Uganda: Consultancy Report to Wildlife conservation Society, Albertine Rift Programme.
Byakola, T. & Mukheibir, P., 2009. Energy systems: vulnerability-adaptation-resilience (VAR) regional focus: sub-Saharan Africa, Uganda, Paris: InHellio International.
Chiputwa, B., Spielman, D. J. & Qaim, M., 2015. Food standards, certification, and poverty among coffee farmers in Uganda. World Development, Volume 66, pp. 400-412.
CIA, 2013. The WorldFactbook, Washington, DC: Central Intelligence Agency.
Cooke, P., Kohlin, G. & Hyde, W., 2008. Fuelwood, forests and community management: evidence from household studies. Environment and Development Economics, 13(01), pp. 103-135.
Egeru, A. S., 2014. Rural Households' Fuelwood Demand Determinants In Dryland Areas of Eastern Uganda. Economics, Planning, and Policy, 9(1), pp. 39-45.
68


Ekeh, O., Fangmeier, A. & Muller, J., 2014. Quantifying greenhouse gases from the production, transportation and utilization of charcoal in developing countries: a case study of Kampala, Uganda. Int J Life Cycle Assess, Volume 19, pp. 1643-1652.
Emerton, L., 2001. Community-Based Incentives for Nature Conservation, s.l.: IUCN-The World Conservation Union.
FAO, 2006. Global forest resources assessment 2005: progress towards sustainable forest management, Rome,: Food and Agriculture Organisation of the United Nations,.
Foell, W., Pachauri, S., Spreng, D. & Zerriffi, FL, 2011. Household cooking fuels and technologies in developing economies. Energy Policy, Volume 39, pp. 7487-7496.
Geist, H. J. & Lambin, E. L., 2002. Proximate causes and underlying driving forces of tropical deforestation. BioScience, 52(2), pp. 143-150.
Gelman, A. & Hill, J., 2007. Data analysis using regression and multilevel hierarchical models. 1st ed. New York: Cambridge Unitversity Press.
Geoghegan, J. et al., 1998. "Socializing the pixel" and "pixelizing the social" in land-use and land-cover change. In: D. Liverman, E. F. Moran, R. R. Rindfuss & C. P. Stern, eds. People and pixels: linking remote sensing and social science. Washington, D.C.: National Academy Press, pp. 51-69.
Getis, A. and J.K. Ord. 1992. The Analysis of Spatial Association by Use of Distance Statistics. Geographical Analysis, 24(3).
Global Alliance for Clean Cookstoves, 2011. Igniting change: a stategy for universal adoption of clean cook stoves andfuels, Washington DC: United Nations Foundations .
GoU, 2002. The National Forest Plan, Kampala: Ministry of Water, Lands and Environment, Republic of Uganda.
Hamilton, A., Cunningham, A., Byarugaba, D. & Kayanja, F., 2000. Conservation in a region of political instability: Bwindi Impenetrable Forest, Uganda. Conservation Biology, 14(6), pp. 1722-1725.
Hartter, J., 2010. Resource use and ecosystem services in a forest park landscape. Society & Natural Resources, 23(3), pp. 207-223.
Hutton, G., Rehfuess, E. & Tediosi, F., 2007. Evaluation of the costs and benefits of interventions to reduce indoor air pollution. Energy for Sustainable Development, 11(4), pp. 34-43.
IFAD, 2012. Enabling poor rural people to overcome poverty, Rome, Italy: International Fund for Agricultural Development.
IMF, 2010. Uganda: Poverty Reduction Strategy Paper, s.l.: IMF Country Report No. 10/141.
69


Kaimowitz, D. & Angelsen, A., 1998. Economic models of tropical deforestation: A review, Bogor, Indonesia: Center for International Forestry Research (CIFOR).
Kazoora, C., Birungi, Z. S. & Dranzoa, L., 2008. Poverty forest linkages in Uganda, Washington DC: PROFOR.
Kees, M. & Feldmann, L., 2011. The role of donor organisations in promoting energy effcient cook stoves. Energy Policy, Volume 39, pp. 7595-7599.
Khundi, F., Jagger, P., Shively, G. & Sserunkuuma, D., 2011. Income, poverty and charcoal production in Uganda. Forest Policy and Economics, 13(3), pp. 199-205.
Kituyi, E. et al., 2001. Biofuel availability and domestic use patterns in Kenya. Biomass and Bioenergy, Volume 20, pp. 71-82.
Lewis, J. J. & Pattanayak, S. K., 2012. Who Adopts Improved Fuels and Cookstoves? A Systematic Review. Environmental Health Perspectives, 120(5), pp. 637-645.
Liverman, D., Moran, E. F., Rindfuss, R. R. & Stern, P. C., 1998. People and Pixels: Linking Remote Sensing and Social Science. 1 ed. Washington DC: National Academy Press.
Lopez, R., 1992. Environmental degredation and economic openness in LDCs: the poverty linkage. Amercian Journal of Agriculture and Economy, pp. 1183-1243.
Luoga, E. J., Witkowski, E. F. & Balkwill, K., 2000. Economics of charcoal production in miombo woodlands of eastern Tanzania: some hidden costs associated with commercialization of the resources. Ecological Economics, Volume 35, p. 243-257.
Mahapatra, A. K. & Mitchell, C. P., 1999. Biofuel consumption, deforestation, and farm level tree growing in rural India. Biomass and Bioenergy, Volume 17, pp. 291-303.
Martinez-Alier, J., 2002. The Environmentalism of the Poor: A Study of Ecological Conflicts and Valuation. Cheltenham, UK: Edward Elgar Publishing Limited.
Masera, O. R., Saatkamp, B. D. & Kammen, D. M., 2000. From Linear Fuel Switching to Multiple Cooking Strategies: A Critique and Alternative to the Energy Ladder Model. World Development, 28(12), pp. 2083-2103.
May-Tobin, C., 2011. Chapter 8: Woodfor Fuel, s.l.: Union of Concerned Scientists: The Root of the Problem.
McCarthy, J., 2002. First World political ecology: lessons from the Wise Use movement. Environment and Planning, pp. 34:1281-1302.
Mwampamba, T. H., 2007. Has the woodfuel crisis returned? Urban charcoal consumption in Tanzania and its implications to present and future forest availability. Energy Policy, Volume 35, p.4221-4234.
Mwaura, F., Okoboi, G. & Ahaibwe, G., 2014. Determinants of households choice of cooking energy in Uganda, Kampala, Uganda: Economic Policy Research Centre (EPRC).
70


Nabanoga, G., Namaalwa, J. & Ssenyonjo, E., 2010. REDD Working Papers: REDD and sustainable development perspective from Uganda, London: IIED.
Nakakaawa, C. A., Vedeld, P. O. & Aune, J. B., 2011. Spatial and temporal land use and carbon stock changes in Uganda: implications for a future REDD strategy. Mitig Adapt Strateg Glob Change, p. 16:25-62.
Naughton-Treves, L., 1999. Whose animals? A history of property rights to wildlife in Toro, Western Uganda. Land Degradation and Development, Volume 10, pp. 311-328.
Naughton-Treves, L., Kammen, D. & Chapman, C., 2007. Burning biodiversity: Woody biomass use by commercial and subsistence groups in western Ugandas forests. Biological Conservation, Volume 134, pp. 232-241.
NEMA, 2009. Uganda Atlas of Our Changing Environment, Nairobi : National Environment Management Authority United Nations Environment Programme.
NEMA, 2014. Fifth national report to the converntion on biological diversity, Kampala, Uganda: National Environment Management Authority.
NFA, 2007. Gazetted Areas GIS database, Kampala, Uganda: Government of Uganda, National Forest Authority.
Obua, J., Agea, J. G. & Ogwal, J. J., 2010. Status of Forests in Uganda. African Journal of Ecology, 48(4), pp. 853-859.
Okello, C., 2014. Assessing sustainability of bioenergy systems in developing countries: methodological development and application, University of Naples Federico II: Department of Agriculture .
Olson, D. M. et al., 2001. Terrestrial ecoregions of the world: a new map of life on Earth. Bioscience, 51(11), pp. 933-938.
Otieno A. C. & Buyinza, M., 2010. Collaborative Forest Management in Uganda: A Strategy for Controlling Deforestation in West Bugwe Forest Reserve, Busia District. Earth Sciences,
4(2), pp. 95-102.
Panwar, N. L., Kurchania, A. K. & Rathore, N. S., 2009. Mitigation of greenhouse gases by adoption of improved biomass cookstoves. Mitigation and Adaption Strategies for Global Change, Volume 14, pp. 569-578.
Peskett, L., Schreckenberg, K. & Brown, J., 2011. Institutional approaches for carbon financing in the forest sector: learning lessons for REDD+ from forest carbon projects in Uganda. Environmental Science & Policy, Volume 14, pp. 216-229.
Plumptre, A. et al., 2010. Impacts offire and large mammals on the ecology of Queen Elizabeth Park, Uganda, s.l.: Wildlife Conservation Society and Woods Hole Research Centre.
71


Pouliot, M., Treue, T., Darko Obiri, B. & Ouedraogo, B., 2012. Deforestation and the Limited Contribution of Forests to Rural Livelihoods in West Africa: Evidence from Burkina Faso and Ghana. Royal Swedish Academy of Sciences, Volume 41, pp. 738-750.
R Core Team, 2013. R: A language and environment for statistical computing, Vienna, Austria: R Foundation for Statistical Computing.
RCMRD, 2015. Uganda Land Cover 2000 and2014 Scheme II. s.l.:s.n.
Robbins, P., 2005. Political Ecology. In: s.URoutledge.
Ruhanga, R. & Manyindo, J., 2010. Ugandas environment and natural resources: enhancing Parliaments oversight, s.l.: UWS and UNEP/GRID-Arendal 2010.
Serneels, S., Linderman, M. & Lambin, E. F., 2007. A multilevel analysis of the impact of land use on interannual land-cover change in East Africa. Ecosystems, 10(3), pp. 402-418.
Simon, G., 2014. If you can't stand the heat, get into the kitchen: obligatory passage points and mutually supported impediments and the climate-development interface. Royal Geographical Society, 46(3), pp. 268-277.
Simon, G. L., 2010. The 100th Meridian, Ecological Boundaries, and the Problem of Reification. Society and Natural Resources, Volume 0, p. 1-7.
Simon, G. L., Bumpus, A. G. & Mann, P., 2012. Win-win scenarios at the climate-development interface: Challenges and. Global Environmental Change, pp. 22:275-287.
Tabuti, J. S., Dhillion, S. S. & Lye, K. A., 2003. Firewood use in Bulamogi County, Uganda: Species selection harveesting and consuption patterns. Biomass andBioEnergy, Volume 25, pp. 581-596.
Turyahabwe, N. & Banana, A. Y., 2008. An overview of history and development of forest policy and legislation in Uganda. International Forestry Review, 10(4), pp. 641-656.
UNDP, 2015. Human Development Report: work for human development, New York, NY: United Nations Developement Programme.
UNSD, 2016. UNdata. [Online]
Available at: http://data.un,org/CountryProfile.aspx?crName=Uganda [Accessed 5 May 2016],
Waiswa, D. et al., 2011. Decentralisedforest management in Uganda: local communities' participation andforest sustainability, Hyberabad India: Paper to the 13th meeting of the International Association for the study of the Commons (IASC).
Wallmo, K. & Jacobson, S. K., 1998. A social and environmental evaluation of fuel-efficeinet cook-stoves and conservation in Uganda. Environmental Conservation, 25(2), pp. 99-108.
WHO, 2006. Fuel for life: household energy and health, Geneva, Switzerland: World Health Organization.
72


Full Text

PAGE 1

THE SOCIAL AND ENVIRONMENTAL CONTEXT RELATED TO THE UTILIZATION OF IMPROVED COOKING TECHNOLOGY IN RURAL UGANDA By NICOLE M ICHELLE BRUNNER B.S., Metrop olitan State University Denver 2012 A thesis submitted to the F aculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Environmental Science Program 2016

PAGE 2

ii This thesis for the Master of Science degree by Nicole Michelle Brunner has been approved for the Environmental Science Program by Gregory L. Simon, Chair Peter A n thamatten Deborah Thomas July 22, 2016

PAGE 3

iii Brunner, Nicole Michelle (MS, Environmental Sciences) The Social and Environmental Context Related to the Utilization of Improved Cooking Technology in Rural Uganda Thesis directed by Associate Professor Gregory L. Simon ABSTRACT Human dependence on natural resources, especially from forests, is most pronounced in developing countries such as Uganda, where many people live in poverty and rely on wood fuel for cooking. These demands often compete with conservation efforts aimed at protec ting forests and biodiversity. A better understanding of the r elationship between social and environmental conditions and the utilization of improved stove technology is essential to ensure the sus tainability of forest related socio economic a Statist ical modeling was used to explain the relationship between utilization of improved stove technology and socio economic characteristics acquired from household surveys and environmental characteristics derived from remote sensed imagery. The findings show t hat socio economic and environmental conditions are important determinants of imp roved stove adoption. Charcoal stoves were found to be related to higher levels of welfare, less forest cover and further distance from protected areas. Whereas, efficient woo d stoves were associated with closer proximity to protected areas representing restricted access to resources. The results of this study could inform policies by providing context on the characteristics of households and communities utilizing improved stov e technology The form and content of this abstract are approved. I recommend its publication. Approved: Gregory Simon

PAGE 4

iv ACKNOWLEDGEMENTS There are many people who have helped me and provided guidance and feedback for this project. A mong those I would like to acknowledge the researchers with whom I work at the Geoscience and Environmental Change Science Center, Dr. Darius Semmens, Dr. Kenneth Bagstad, and Dr. Todd Hawbaker and m y committee members from the department of Geography and Environmental Sciences Dr. Deborah Thomas and Dr. Peter Anathamat ten, who had to endure long meetings about methods and reading my thesis serval times, your comments were essential to the success of this project. I would also like to acknowledge Michael Ve r done from IUCN and Dr. Bonnie Keeler from Natural Capital Project who directed me to Uganda shared data and knowledge and were there to discuss ideas at the beginning of this project. A special thank you to my advisor Dr. Gregory Simon who has supported me throughout my thesis and played a crucial role overseeing my progress through the program. He has not only given me guidance and encouragement over the last 3 years but has broadened my knowledge and changed the way I think about issues. This experie nce would not has been the same without him. This research has been reviewed and was determined non human subject research as defined by the policies of Colorado Multiple Institutional Review Board (COMIRB) in accordance with OHRP and FDA regulations. The protocol number is 16 0281.

PAGE 5

ii TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ................................ ......... 4 II. LITERATURE REVIEW ................................ ................................ .............................. 7 Traditional stove and energy use ................................ ................................ ..................... 7 Improved biomass stove technologies ................................ ................................ ........... 12 Factors influencing improved fuel and stove use ................................ .......................... 15 III. STUDY AREA ................................ ................................ ................................ .......... 19 Background information ................................ ................................ ............................... 19 Uganda energy use ................................ ................................ ................................ ........ 25 IV. DATA AND METHODS ................................ ................................ .......................... 28 Data ................................ ................................ ................................ ............................... 28 Social survey data ................................ ................................ ................................ ...... 28 Environmental data ................................ ................................ ................................ .... 30 Methods ................................ ................................ ................................ ......................... 31 Conceptual framework ................................ ................................ .............................. 32 Research questions ................................ ................................ ................................ .... 33 Temporal analysis When? ................................ ................................ ....................... 33 Spatial analysis Where? ................................ ................................ .......................... 34 Explanatory analysis Why? ................................ ................................ ..................... 35

PAGE 6

iii Analysis of relationships What? ................................ ................................ ............. 36 Description of variables chosen for analysis ................................ ............................. 37 V. RESULTS ................................ ................................ ................................ ................... 43 The temporal trends of stove use ................................ ................................ ................... 43 The spatial analysis of stove use ................................ ................................ ................... 46 The potential benefits of improved stoves ................................ ................................ .... 47 The relationship between variables driving the utilization of improves stoves ............ 54 Analysis of effici ent wood stoves ................................ ................................ .............. 54 Analysis of improved charcoal stoves ................................ ................................ ....... 55 VI. DISCUSSION ................................ ................................ ................................ ............ 58 VII. CONCLUSION ................................ ................................ ................................ ........ 6 5 REFERENCES ................................ ................................ ................................ ................. 68

PAGE 7

4 CHAPTER I INTRODUCTION Close to half of the global population uses traditional wood fuels for cooking and heating which typically include firewood and charcoal. These fuels are predominantly used in poor and rura l areas of sub Saharan Africa. Poor ventilation of open fire stoves coupled with incomplete combustion of these wood fuels results in high levels of indoor air pol lutants which impacts health (Hutton, et al., 2007; Adkins, et al., 2010; Bailis, et al., 2015) In addition to releasing health damaging pollutants, traditional charcoal and wood burning stoves have been identified as a contributor to greenhouse gas (GHG) emissions. Also, in regions where resources are limited the amount of time spent collecting wood for fuel can be a considerable burden on impoverished rural households, especially women and children (Adki ns, et al., 2010). These stoves are also often tied to unsustainable tree harvesting which can also impact climate change and biodiversity loss However, the extent of these effects remains highly uncertain (Bailis, et al., 2015). Due to the negative impac ts associated with the use of tra ditional fuels, improved more efficient stoves are being funded in order to promote sustainable development. (Khundi, et al., 2011) Firewood is the primary fuel in Uganda, however there has been a steady increase in charcoal use. Like many developing countries the majority of Uganda is rural, and their livelihoods and wellbeing are heavily reliant on natural resources. I n addition, the population is growing rapidly, coupled with increasing production and resource extraction, which has raised concerns about energy and resource shortages. Uganda is one of the most biodiverse as well as poorest countries in the world. It wa s ranked 163th out of 187 on the 2015 human development index (UNDP, 2015) Poverty is often

PAGE 8

5 linked to environmental degradation. Traditional stove and fuel use, poverty, and environmental degradation are often coincide in deve loping countries. Therefore, poverty is often cited as causing environmental degradation (Aggrey, et al., 2010). Likewise traditional stove and fuel use are linked to deforestation. There are concerns that increasing demand for natural resources in Ugan da will lead to destruction of p rotected a reas that are of particular conservation importance due to rare biodiversity, including highly endangered species such as chimpanzees and mountain gorillas, which also bring in significant revenue thr ough eco tourism. The primary activates cited as threats to conservation goals are agricultural clearing for cultivation and grazing, illegal hunting (poaching), logging, pit sawing and charcoal production (Plumptre, et al., 2010). While firewood collection does not drive large scale deforestation, many local or regional assessments have cited wood fuel demand as a primary driver of forest loss (May Tobin, 2011). This discrepancy suggests that impacts of wood fuel demand may vary by location and fuel type (Ba ilis, et al., 2015). For example, there are some contextual disparities between developing countries in Africa, where wood fuel has been cited as having a greater influence on land use change than in other regions of the world (Na ughton Treves, et al., 2007). Also, charcoal production also has a greater environmental impact than firewood collection. Unlike firewood which typically consists of twigs and smaller branches, charcoal is usually made by burning larger tree trunks and li mbs which leads to more forest loss than firewood collection (May Tobin, 2011) Naughtons Treves et al. (2007) determined that producing 50 tons of charcoal from hardwood would results in 1 km 2 of deforestation. The per capita consumption for firewood in Uganda has been estimated to 680 kg per year for rural residents and 240 kg per year for urban residents. Charcoal consumption is about 4 kg for rural and 1 20 kg for urban dwellers (Mwaura,

PAGE 9

6 et al., 2014) Therefore about every 400 urban households requires 1km 2 of forest to supply their energy requirements per year. Regardless of whether wood fuel use contributes to forest loss, there are concerns that forest loss in developing countries will lead to fuel shortages and increased vulnerability of already impoverished populations. Based on traditional policy solutions that view environmental degradation as a local problem of ove r consumption caused by poverty, g lobal development initiatives and government organizations have formed policy aimed at distributi ng improved more efficient stoves. This improved stove technology claims to help by reducing fuel de mand potentially alleviating deforestation pressure in select locations, enhanc ing public health and alleviat ing poverty on a local scale, as well as mitigating climate change at the global level (Global Alliance for Clean Cookstoves, 2011). Therefore tr aditional wood fuel and stove use present s society with two important links bet ween local and global impacts. Merging social science and remote sensing data can offer new insights into how issues such as poverty, gender, demography, or development relate t o resource use and availability (Geoghegan, et al., 1998) The objectives of this study are to explore (i) the geographic and temporal differences in fuel types and cooking technology in rural households in Uganda (ii) the cl aimed benefits of improved stove use, and (iii) the social and biophysical characteristics related to the adoption o f improved cooking technology. Understanding how the household choice to adopt improved cooking technology is related to social and biophysi cal characteristics (such as demand and access to resources) will provide information that will hopefully contribute to policies aimed at

PAGE 10

7 CHAPTER II LITERATURE REVIEW This chapter review s the literature addressing three topics relevant to this project I begin by reviewing the contribution of traditional stove and fuel use to negative impacts associated with health, poverty and environment. Followed by an overview of improved cooks tove technology. Finally is an overview of the methods other studies have used and factors found that influence the choice to utilize improved stove technology. Traditional stove and energy use Wood fuel and traditional stoves are used predominantly in de veloping countries. Globally at least 2.6 billion people are estimated to use traditional wood fuel, and about 2.2 billion people use inefficient traditional stoves to burn those fuels. Traditional wood fuel primarily consists of firewood and charcoal, w hich is typically collected from nearby forests. It estimated that 55% of all wood harvested is used for fuel, which supplies 9% of the primary energy (Bai lis, et al., 2015). The use of traditional fuels and stoves is exceptionally high in sub Saha ra Africa where 85% of the rural population uses traditional wood fuel, and up to 94% use traditional stoves (Okello, 2014) Traditio nal stoves are typically a 3 ston e stove or stone stoves burn primarily firewood. The three stones support the cooking vessel a nd act s as the hearth where wood is burned by pushing it through the openings in between the stones (Adkins, et al., 2010). Traditional charcoal stoves also referred to as traditional metal stoves, ar e typically made from found metal scraps such as roof materials, or reclaimed oil drums. Traditional stoves have no insulation which makes them very inefficient (Okello, 2014).

PAGE 11

8 Traditional stoves pro duce massive amounts of smoke. The smoke which is from incomplete combustion is linked to global climate change due to the release of greenhouse gas emissions including Carbon Dioxide (CO2), methane, black carbon and short lived climate forcers (SLCFs) (Panwar, et al ., 2009; Simon, 2014; Bailis, et al., 2015) This smoke is also toxic and harmful to human health. Traditional stoves release carbon monoxide, particulate matter, and non methane volatile organic compounds which are linked to eye irritation, cataracts respiratory illnesses, including asthma, low birth weight and still birth, tuber culosis and cancer (Simon, 2014; Adkins, et al., 2010). Traditional stoves are most often used in main indoor living spaces that are poorly ventilated and where many people a re exposed, which increases the negative health impacts. It is estimated that approximately 1.5 million people die annually from health issues induced by exposure to smoke from traditional stoves (WHO, 2006; Okello, 2014) Another major social development challenge associated with traditional wood fuel use in developing countries is soc ial and gender dimensions. Women and children primarily participate in the collectio n and processing of wood fuel. The collection process in particular, is associated with a considerable amount of human labor, risk, and time which could have been spent pursuing more economically productive activities (Foell, et al., 2011) Women and children are also the most c ommonly exposed to indoor pollutants and are the most vulnerable to the harmful effects (WHO, 2006) The health impacts caused by burning wood fuels in traditional stoves is increasingly well charac terized, whereas the contrib ution to global climate change from the release of GHG emissions and fuel driven deforestation is still highly uncertain. It is estimated that 18% of the global atmospheric black carbon concentration originates from incomplete combustion (Foell, et al., 20 11). Black carbon contributes to climate change, and is also linked to increasing glacial ice

PAGE 12

9 melt as well as altering regional rainfall patterns (Okello, 2014). Ballis et al. (2015) estimated that in 2009 GHGs and SLCFs from wood harvested unsustainably and partial burning of fuel accounted for between 1.9 2.3% of global emissions, t hese emission estimates were divided equally between CO2, black carbon, and SLCFs. The estimates for fractional non renewable biomass (fNRB), which is wood harvested unsusta inably, were found to be much less than carbon financed improved stove and wood fuel projects estimate. Therefore, it is likely that projects are exaggerating the potential for improved stove technology to reduce emissions through reducing deforestation. M ost of the literature is divided on the extent to which wood fuel demand drives deforestation. While traditional wood fuel is a renewable resource, there are concerns about its sustainability with increasing energy demand, especially in areas wit h high rat es of deforestation. These concerns are especially relevant in sub Sahara Africa where there are high population growth rates which often coincide rapid forest loss. The use of traditional and inefficient stove technologies only exacerbates the problem of fuel insecurity. Since the 1980s, the scientific community has, for the most part, accepted that fuelwood use may not necessarily lead to deforestation on a large scale and that fuelwood crisis may not become a reality (Mwampamba, 2007) However, studies at local scale s have found that location is a key factor determining whether fuel harvesting is sustainable due to disparities between wood demand and resource availability (Arnold, et al., 2006). In a special issue published by the International Journal on Energy Policy in 1993, emphasized the inaccuracy of attributing deforestation exclusively to wood fuel extraction (Mwampamba 2007)

PAGE 13

10 An assessment by Kaimowitz & Angelsen, (1998) on tropical deforestation found that several factors lead to forest loss. This study also and notes that there is not substantial evidence linking wood fuel demand and deforestation, although it is occasionally a driver of deforestation in parts of Africa and Asia. Similarly, another study examining multiple cases of tropical deforestation found numerous drivers, however tree harvesting for wood fuel was a more significant cause in some parts of A frica (Geist & Lambin, 2002). However, there have been very few systematic studies of wood fuel sustainability (Bailis, et al., 2015), and many studies that attribute the demand for wood fuel as the primary cause of deforestation and forest degradation don However, studies that distinguish between the impacts of firewood versus charcoal find that while firewood collection rarely threatens forests, charcoal use has very different implications. Due to inefficient charcoal production and consumption households use on average 4 to 6 times more wood than households using firewood (Mwampamba, 2007) In addition, c harcoal production involves cutting larger branches or tree trunks which often results in tree removal, and the preferential wood species to produce charcoal are typically slow growing and therefore is less sustainable than firewood Kituyi, et al. (2001) found that fuel security in Kenya is determined by the presence of bio diversity of wood types which allows households to switch fuel types with relative ease. Similarly, households often respond to fuel scarcity by harvesting other available tree species that are growing on f arms and in protected forests. The decreasing of m ajor tree species used for wood fuel in Kenya is not entirely due to fuel demand, but due to competing demands for timber, agriculture and settlement. These increasing demands raise concerns about the long term sustainability of forest s to supply sufficien t fuel demand in Kenya A study done by Luoga, et al.

PAGE 14

11 ( 2000) in eastern Tanzania shows that while the commercialization of wood resources particularly charcoal production provides significant income to rural households it also contributes to significan t forest resource loss. A study in West Africa conducted by World Bank found that the concentration of charcoal production can be a primary driver of deforest ation (Arnold, et al., 2006). A study done by Ekeh et al. (2014) examining the level of GHG emitt ed from charcoal production, transport and use in Uganda, found that 15% of GHG are emitted from the transportation phase. The production and use of charcoal accounted for the vast majority of GHG emissions. Therefore shifting to improved technology to pro duce charcoal would result in significant GHG reduction, however using sustainably sourced biomass would reduce GHG emissions by almost half, and therefore the best solution is sustainable harvesting practices of wood coupled with improved production metho ds. Mwampamba, (2007) evaluated the sustainabiliy of current charcoal consumption rates, with the current efficiency of local charcoal production and current forest management policies to determine weather forests can contiune to meet current and future demands in Tanzania 24 scenarios were developed to capture the numerous uncertainties of forested area required to meet charcoal demand. Their findings suggest that scenarios based on m oderate levels of consumption paired with low production efficiency and very little reforestation efforts could substantially reduce total forest area on public lands by 2028. This study determines that charcoal consumption activities are a threat to the f uture of forests in Tanzania C harcoal is the primary fuel used for cooking in most urban centers sub Saharan countries (Mwampamba, 2007) Increasing demand for charcoal from growing urban populations, coupled with poor resourc e management, and regulation present a problem for the

PAGE 15

12 future for forests and fuel shortages in Africa (Mwampamba, 2007; May Tobin, 2011; Foell, et al., 2011) Traditional stoves are a primitive technology wi th several associated negative impacts. With diminishing resources and increasing demand to supply fuel, alternatives must be provided to address basic human welfare. Charcoal can be damaging not only to human health but also the environment. Improved bio mass stove technologies Due to the increasing interconnectedness between environments, societies, and economies worldwide, there are a diversity of actors (i.e. NGOs, private sector, international organizations, civil society) sh aping environmental govern ance. Carbon financing or carbon trading is a mechanism aimed at reducing greenhouse gas emissions (GHG ) and mitigate climate change. Carbon credits (equivalent to 1 ton of carbon dioxide) are often purchased by developed countries and businesses through e mission trading markets in order to meet emissions reductions obligations made under the Kyoto Protocol. Due to the many recent reports and studies noting the contribution of traditional stoves and wood fuel use to GHG emissions, coupled with health concerns and deforestation, has led th e promotion of improved stoves (Okello, 2014). C arbon financing has been providing funds to projects are often financed under um brella initiatives such as the Clean Development Mechanism (CDM), and implemented under the Kyoto Pr otocol (Peskett, et al., 2011). Along with aiming to reduce GHG emissions, t hese programs often promote communities, fore sts, biodiversity and food security are all positively affecting by the program.

PAGE 16

13 Improved stove designs range from locally made to manufactured models Locally made stoves are constructed from materials found in or near the home and typically consist of c lay s oils or mud and dried grasses. Manufactured stoves are finished products produced entirely in factories, either domestically or internationally, then distributed to local villages. These manufactured stoves are typically based on the rocket stove desi gn principle (Adkins, et al., 2010). Improved stoves have several claimed co benefits besides burning fuel more efficiently than the widely used traditional stove. These additional benefits include reduced cooking time improving health lessening deforest ation pressures, alleviating the burden placed on women and children in fuel collection, reducing greenhouse gas emissions and mitigating poverty (Kees & Feldmann, 2011; Okello, 2014) The extent to which improved stove use benefits households, lessens deforestation and forest degradation, as well as mitigates climate change, remains highly uncertain (Kees & Feldmann, 2011; Bailis, et al., 2015) A study by Bailis, et al. (2015) found of the would decrease traditional wood fuel emissions by 98 161 million tons of CO 2 per year, valued at $1.1 1.8 billion USD an amount that far exceeds the curr ent investments in improved cookstove projects. I mproved stove designs funded through carbon finance are typically manufactured in order to standardized technology and accurately account for emission reductions (Adkins, et al., 2010). However standardizing technology introduces a question of whether the technology practical for household use. If a household does not perceive benefits of improved stoves they will be less likely to use the stove. Studies have found that households with improved s toves are commonly used along with traditional stoves, commonly A household survey and efficiency test done by Adkins et al. (2010) found that while improved stoves reduced the amount of fuel used compared to the t raditiona l counterpart some models

PAGE 17

14 showed significant increases in time spent cooking. For example, cooking matooke, a traditional staple food in Uganda, using a traditional 3 stone fire requires approximately 17 min, whereas t he improved stove models tested both showed increases in time needed to cook matooke. The Ugastove showed a significant increase 22 min (27% longer), the StoveTec stove showed only a slight increase 18 min (5% additional time) compared to the 3 stone tradi tional model The survey portion of the study found that despite the reduction in smoke and reduced fuel use about 40% of households preferred traditional stoves over the improved model. The major complaints were difficulty lighting, easy to tip over, lon ger time spent cooking, and too much heat causing burning. Masera, et al. (2000) found that very few rural households completely abandon traditional fuels and stoves C utlural and technical factors were found to be the primary reason for stove stacking in rural villages in Mexico. This practice occurred among households of varying socio economic status. E ven though improved and modern stoves carry status as symbols of prosperity, they are still used along with tradtional stoves and fuels. This is the case because improved stove design that does not meet all of the of cultural and customary cooking requirements. A ccess to clean and affordable energy has been identified as critical to achieving the millennium development goals due to the claimed co benefits such as improved maternal health, reduction in premature death, reducing environmental impacts, and saving time in reduced fuelwood collection. Overall, improving access to clean stoves is expected to improve human welfare and contribute to sustainable d evelopment in developing countries (Okello, 2014) However, there is debate about whether these win win scenarios are realistic and feasible (Simon, et al., 2012)

PAGE 18

15 Factors influencing improved fuel and stove use Besides technical limitations of improved stove designs, h ousehold stove choice and the relative fuel consumption are influenced by factors at various structural levels ranging from micro or household level factors to biophysical features related to fuel availabi lity, which together make up the social and environmental context that determin es household preferences regarding cooking technology (Masera, et al., 2000) At the household or micro level th e distinction in fuel and stove use in rural communities may be due to demographic, economic conditions, cultural values, health concerns or a combination of factors. At the village or community level, resource scarcity and the influence from proximity to urban centers, could be motivating households to switch to more efficient cooking o ptions (Masera, et al., 2000). A study by (Mahapatra & Mitchell, 1999) in rural India found that while wood fuel use is influenced by socio economic factors, limited forests resources does not reduce the demand for wood fuels I n other words, socio economic factors influenced fuel use more than resource availability. Masera et al. (2000) claim that fuel and stove switching is the result of simultaneous interaction s of factors that push households away from traditional practices and pull them back. include convenience, cleanlines s, health and economic status. usually result from defects or inadequacies of technology to meet the needs of traditional cooking practices The literature on household adoption of improved fuel and stoves in devel oping countries is for the most part inconsistent and pre dominantly uses qualitative methods While there have been numerous attempts to assess improved stoves or fuel choice, few studies have used rigorous statis tical analysis to confirm findings (Lewis & Pattanayak, 2012) A literature review by Lewis and Pattanayak (2012) found that t he most common factors with a significant positive correlation

PAGE 19

16 to improved fuel and stove adoption were education, income, and urban location. The roles of f actors such as fuel availability, pric e, household size and compo sition are still largely unclear. Demographic characteristics showed varying results, household size and average age of head of household was inconclusive. Most of the studies that include education find a positive and statisti cally signifi cant relationship with improved stove adoption. Most studies (56%) found that f emale head s of households were more likely to adopt improved fuel s H owever improved stove studie s seldom considered female head of household in their analysis The review also found that the majority of studies found no significant associations between improved stove adoption and livelihood or occupation, number of children, and the price of modern fuel Some variables that were understudied include proximity to markets roads and populated centers (Adkins et al. 2010). Therefore, there is greater need to explore more of these understudied variables as well as variables with known significant correlations to ensure consistency and verifiable results. Wallm o & Jaco bson (1998) found that protected areas are likely a factor associated with the use of improved stoves. The main reason households tend ed to implement improved stoves in 3 pilot parishes in western Uganda was to reduce the amount of fuel required for domest ic energy use to lessen the need illegally harvest wood from the adjacent National Parks. People living outsides parks and protected area s are often excluded from accessing resources, settlement, and land use within the park boundarie s (Hartter, 2010) which can foster attitudes of resentment t owards the government authority, the protected forests and the wildlife residing within A study by Bush et al. (2004) found that the further residents were from four protected national forests in Uganda the fewer resources were extracted and less total household income was derived from forests. This relationship was due to the increased levels of deforestation due to hi gh density of agriculture and over exploitation of resources (Hartter, 2010; Nabanoga 2005).

PAGE 20

17 Therefore, distance from forests (especially protected forests) is likely an important factor in driving resource use and availability. Community level economic factors and resource availability are often closely interlinked. being, and plays an important part in livelihood strat egies (Nabanoga, et al., 2010) Forest s provide wood and non wood forest products and services to the people who produce and consume forest products, and local people play a major role in the management of forest resources. Government authority is weak and property rights are often times ambig uous in developing countries, turning dynamics within the community closer to the ground into critical determinants of actual resource use and access. There is an indisputable correlation between environmental degradation and poverty ; as natural resources become depleted, local economies suffer and livelihoods become more vulnerable (Emerton, 2001) While there are circumstances in which local livelihood activities contribute to environmental degradation, there are also cases where the impacts of local production act ivities are relatively benign. However, local communities are commonly blamed for envi ronmental degradation and unsustainable harvesting which is a reflection of exercising control over resources by governments and authorities (Robbins, 2005) Therefore, cultural politics play a major role in conservation struggles due to resource claims based on tradition al or custom ary land rights. Having been alienated from their resources, many local people have been re claiming those resources and as populations increase around protec ted areas this problem has only intensified. These tactics often include trespassing, a steady encroachment into areas of restricted access, poaching the prote cted wildlife, il legal grazing, tree harv esting, fuelwood collection, threatening government per sonnel, and setting fires to gain control and access (McCarthy, 2002) These restrictive boundaries of governme nt lands which elevate ecological

PAGE 21

18 goals above basic human needs, contradict contemporary conservation trends, which have emphasize d the n eed for porous boundaries to protected areas extractive reserves, and community support for and participation in conservation (McCarthy, 2002) While linking exploring proximate drivers or causes of environmental degradation are important such as proximate causes of deforestation (slash and burn cultiva tion or wood fuel collection). It is also important to explore relationships between factors such as socio economic status and b iophysical characteristics in order to understand the underlying processes and structures that influence household level decisions that in turn give rise to environmental outcomes (Geoghegan, et al., 1998).

PAGE 22

19 CHAPTER III STUDY AREA The study area selected for this study is Uganda, mainly due to the high levels of dependence on traditional fuels and diminishing forest resources. In addition, there are several global development and carbon financing initiatives aimed at promoting improved stoves in the count ry This chapter presents an overview of Uganda, the energy sector, environmental concerns with regards to deforestation and traditional wood fuel use, as well as an overview of the global organizations and efforts aimed at promoting improved cookstove and forest conservation projects in Uganda. Background information Uganda is a landlocked country located along both sides of the equator, between 1 30' S and 4 N latitude and between 29 30' East and 35 West longitude in Eastern Africa. Sudan borders Uganda on the n orth, Kenya is on the e ast ern border Rwanda, and Tanzania are to the s outh and Democratic Republic of Congo is to the w est (Figure 1) (NEMA, 2009)

PAGE 23

20 Figure 1 : Map of Africa, location of Uganda The boundaries that delimit Uganda were created by Britain during colonization in the late 19 th century which grouped together many different ethnic groups, political systems and cultures. T her independence from colonial rule in 1962 ultimately result ing in the dictatorships of military leader Idi Amin from 1971 1979 followed by Milton Obote from 1980 1985, who engaged i n guerrilla warfare and human rights abuses During this time t he country plunged into political and economic instability, and there was total breakdown of law and order. Conflict claimed the lives of at least 600,000 Ugandans between 1971 and 1986 (Hamilton, et al., 2000). Civil strife coupled with increasing demand on resources led to diminished forests and wildlife in protected forests (Turyahabwe & Banana, 2008). The c urrent president of Uganda is Yoweri Museveni who has been in power since 1986 and has brought relative stability with a developing economy. There is currently an armed

PAGE 24

21 rebellion in the Northern region that has displaced hundreds of thousands. In addition there can be spill over from the violence occurring in the DRC. Due to recent conflicts between government forces and the Lord's Resistance Army, there are currently an estimated 32,447 displaced residents in northern Uganda. This region remains politic ally and economically in secure, due to fighting, which resulted in an estimated 1.8 million people residing in internally displaced persons (IDP) camps (CIA, 2013) during the height of the conflict in 2011. The country is split up into 4 regions which repr esent very diverse ecoregions. The Western region is characterized by Terrestrial Ecoregions of the World (TEOW) as being primarily being tropical and subtropical moist broadleaf forests The North and Central regions are tropical and subtropical grass lands, savannas and shrublands also referred to as Victoria Basin Forest Savanna Mosaic. The North region which is more dry and arid than the rest of the country is comprised of East Sudanian Savanna and thickets as well as Northern Acacia Commiphora bus hlands. The Central region is more populated and houses the capital Kampala. The Eastern region is diverse with tropical and subtropical moist broadleaf forests near Mt. Elgon and tropical and subtropical grasslands, savannas and shrublands ( Figure 2; Olson, et al., 2001)

PAGE 25

22 Figure 2 : Olson 2001 Ecoregions with Uganda political region boundaries The total surface area of Uganda is 241,550 km 2 The current estimated population of Uganda is 31.8 million (Table 1) The capital is Kampala, which has a population of approximately 1,863,000. The average annual population growth rate fro m 2010 2015 was 3.3%. Agriculture accounts for about 25% of the gross domestic product (GDP), 50% of total exports and employs almost 75% of the labor force (IMF, 2010) Approximately 70% of agriculture produced is for subsistence purposes (Ruhanga & Manyindo, 2010) Most smaller farms mix commercial crops such as coffee and tea with subsistence staples include mostly starches such as maize, rice, millet, plantains, and cassava (flour) (Benson, et al., 2008; Chiputwa, et al., 2015). Coffee is the largest earner of for eign exchange followed only by the tourism industry. Consequently there is growing interest and investment in ensur ing the protection of the resources supporting tourism, particularly wildlife conservation.

PAGE 26

23 Table 1 : Demographic an d economic information for Uganda: Source: (UNDP, 2015; CIA, 2013; UNSD, 2016) Parameter Value Year Population, total (millions) 38.8 2015 Population density (persons per square km) 160.7 2015 Urban population growth rate (average annual %) 5.4 2010 2015 Rural population growth rate (average annual %) 3.0 2010 2015 Kampala population 1.936 2015 Urban population (%) 16.1 2015 est. Gross domestic product (GDP) (million current US$) 26444 2013 GNI: Gross national income per capita (current US$) 681.5 2013 Forested area (%) 14.1 2012 Tourist arrivals at national borders (000) 1206 2013 Tourism (%) increase 17.5 2005 2013 Most of the population is rural (87%) and impoverished ; about 10 million people fall under the national rural poverty level and rely heavily on the health and resilience of the ecosystems which support their livelihoods (Nabanoga, et al., 2010; IFAD, 2012) Forests and woo dlands, in particular, are essential for employment, economic growth, and basic household uses (Obua, et al., 2010; Waiswa, et al., 2011; Brickwell, et al., 2012) Therefore continued access to forest resourc es will be necessary for securing rural livelihoods (Bush, et al., 2005) Forests have high agricultural potential which makes them particularly vulnerable to encroachment for agricultural production (Nakakaawa, et al., 2011) Between 2000 and 2005, the annual rate of forest loss in Uganda was reported to be 2.2%, which is among the highest globally (Okello, 2014; FAO, 2006) If the current rate of forest loss continues Uganda will lose all of its remaining forests by 2040 (NEMA, 2014) The drivers of deforestation vary across Uganda, but the largest contributors have been identified as agricultural expansion, followed by wood fuel collection, and timber harvesting (IMF, 2010) According to the government of

PAGE 27

24 Uganda, wood fuel demand drives deforestation in the northern and eastern regions, while timber and agricultural demand drives forest loss in the Ce ntral and West regions (Brickell et al. 2012). These various threats, in addition to unsustainable harvesting of resources, pose serious challenges to forest conservation, resource degradation, and livelihoods (Nabanoga, et al., 201 0) In Uganda, forest conservation is undertaken through an established p rotected a re a system (Figure 2). During c olonial iation authorities seized ownership of forests and wildlife, prohibiting hunting and use of park resources, and imposed national park boundaries over many traditional hunting and grazing lands. Wildlife today, particularly large species, survive mainly in isolated national parks (Naughton Treves, 1999). These parks and wildlife belonging to the government are a remnant of the colon ial history of Uganda. While these areas tend to bring in significant revenue through tourism there is still resentment towards these boundaries by local d and destroy their crops for which they are never compensated and cannot defend against because they are protected by the law (Naughton Treves, 1997). 20% of the country's forests are National Parks or wildlife refuges which are designated as strict nature reserves. Central Forest Reserves (CFR) these are low intensity zones allow ing for some sustainable harvesting and non consumptive use account for the remaining 15 17% of state controlled forest s (Peskett, et al., 2011) This situation leaves the majority of forest on private or communal land (Brickwell, et al., 2012) In many cases of established conservation in Uganda access to and control of resources and landscapes has been taken away from local people. In Uganda, four of the major national parks, Kidepo Valley National Park, Murchison Falls National Park, Lake Mburo National Park and Queen Elizabeth National Park a re primarily

PAGE 28

25 grasslands. Mt. Elgon National Park, Bwindi Impenetrable National Park, Mgahinga Gorilla National Park, Kibale National Park and Semuliki National Park are forested. Two national parks are mountain ous: Mt. Rwenzori National Park and Mt. Elgon National Park. Uganda energy use In Uganda traditional biomass, predominantly firewood, charcoal, and agricultural residues account for over 90% of th e national domestic consumption (Khundi, et al., 2011; Okello, 2014; Egeru, 2014) Modern fuels, such as electricity and other petroleum based fuels contribute less than 10% of energy use. Charcoal is more common in urban households while firewood and agricultural residue s are more common in rural households Wood fuel is also commonly used for small scale industrial production activities (e.g. brick making, agro processing, and charcoal production). Most of the firewood used is harvested or collected from natural forests on communal, government or private land is largely unregulated (Tabuti, et al., 2003) The largest consumer of wood energy in Uganda is the residential sector, with firewood being the most prominent used fuel type (Okello, 2014) Wood fuels are typically burned in inefficient traditional stoves open fire stoves. Traditional stoves are used by about 87.5% of households (Egeru, 2014; Okello, 2014) 72 .7% of the population use the 3 stone stoves, and 14.8% of households use traditional charcoal stoves (Byakola and Mukheibir, 2009). Ugandans consumes an estimated 18 million tons of firewood and 500,000 tons of charcoal annually (Egeru, 2014) Reduced forest resources in Uganda may lead to increases in poverty due to higher wood fuel costs, and time spent collecting wood (Kazoora, et al., 2008) Current wood fuel collection practices are nearly alw ays impacted by high rates of forest loss,

PAGE 29

26 usually not driven by wood fuel collection The demand for traditional wood fuel is reported to be growing at a rate ranging from about 2.5% to 6% annually (Tabuti, et al., 2003) Bec ause there are high levels of deforestation in Uganda there are concerns about the sustainability of wood fuel supplies, especially with a rapid ly growing population. Additionally, the increasing harvesting demand may lead to additional deforestation res ulting in overall environmental degradation (Okello, 2014; Naughton Treves, et al., 2007; Egeru, 2014; Tabuti, et al., 2003) Given the heavy reliance on wood fuel, the anticipated depletion of forest stocks could prove to be a threat not only to the economic welfare of rural households but also to ecosystem health and biodiversity (Bailis, et al., 2015) The use of e fficient stove technology is becoming more common in Ugan da due to the government and backed by the German Agency for International Cooperation. Since 2005, an estimated 500,000 stoves have been distributed in Uganda., w ith a goal of providing 250,000 more improved charcoal stoves and 4 million more improved wood stoves to households by 2017 (Kees & Feldmann, 2011; Okello, 2014) d wood stoves for a variety of purposes in Uganda. Adkins et al. (2010) give a detailed description of the UGASTOVE design. The UGASTOVE started in 2004 and received carbon financing for emission reductions in 2009. The goal of the project was to distribu te 180,000 stoves within 7 years which would decrease carbon dioxide released into the atmosphere by approximately 600,000 tons. The project far appears to have successfully distributed household stoves in areas experiencing high levels of deforestation (Simon, et al., 2012) Even with these projects and goals in place the current use of improved stove technology in households in Uganda is still quite

PAGE 30

27 low (Okello, 2014) Byakola and Mukheibir (2009) estimate that 8.7% of Ugandan households use improved stoves (Byakola & Mukheibir, 2009)

PAGE 31

28 CHAPTER IV DATA AND METHODS Data The objectives of this study were to determine differences in household fuel and stove use, analyze the potential benefits of improved stoves, and explore the social and environmental context related to improved stove adoption. To address these objectives, h ousehold and community survey data were obtained from the 2005/06 Uganda National Household Survey (UNHS) and 2010/11 Uganda National Panel Surveys (UNPS) SERVIR (RCMRD, 2015) land cover maps (2000 2014), as well as protected area boundaries (NFA, 2007) which are data derived from remote sensing, were used to explo re community level resource availability Most studies on rural fuel and stove use in developing countries are very localized or done at regional scales with small survey sizes (Adkins, et al., 2010; Cooke, et al., 2008; Egeru, 2014; Foell, et al., 2011; Lewis & Pattanayak, 2012; Mahapatra & Mitchell, 1999; Okello 2014; Otieno & Buyinza, 2010; Pouliot, et al., 2012; Wallmo & Jacobson, 1998) This study broadens the scale of analysis to the national level b y utilizing national panel surveys and data derived from remote sensing. Social survey data Survey data are necessary to address questions related to the social context associated with adoption of improved stove technology. The se surveys can provide information on what influences household level decisions, including socio economic status. The 2005/ 06 Uganda National Household Survey (UNHS) and 2010/11 Uganda National Panel Surveys (UNPS) are multi Standard Measurement Study (LSMS) project and administered by the Uganda Governm ent.

PAGE 32

29 Both UNHS 2005/06 and UNPS 2010/11 were carried out annually, over a twelve month period for the purpose of accommodating the seasonality associated with agriculture and consumption. The surveys are designed to interview each household twice each year, in visits six months apart. The two surveys were designed using a two level strati fied random sampling design applied to select the enumeration area at the first level and h ouseholds at the second level. Both survey sample s cover the entire country and were selected in such a way that they are representative of the entire population of Uganda, for urban and rura l populations, and the four regions: Central, Eastern, Northern and Western. The 2010/11 UNPS sample includes approximately 3,200 households distribu ted over 322 enumeration areas. All of the enumeration areas included in the 2010/11 survey and most of the households were re interviewed from the 2005/2006 UNHS. The 2010/11 sample also includes additional randomly selected households allowing fo r a representative sample at the national, urban/rural, and regional levels The UNPS 2010/11 tracked and re interviewed a sample of 2,716 households that were previously interviewed for the UNHS 2005/06. The data used in this study was derived from the h ousehold and community modules. The household questionnaire gathered information on socio economic characteristics. This includes information on the primary stove type used for cooking in 2010, 2005 and retrospectively in 2001. Types of stoves differ betwe en the 2005/06 and 2010/11 surveys. Therefore information reported on stove use in 2005/06 and 2001 was only used descriptively due to its limitation for detailed analysis The analytical part of the study utilizes respondent information on stove use durin g the survey in 2010/2011. The stoves were grouped into three categories including efficient wood, charcoal and traditional. The community survey aimed at collecting information related to the characteristics of communities residing in the sampled enumeration area s.

PAGE 33

30 The 2010/11 survey also provides modified GPS coordinates. This allows for inclusion of spatial variables while preserving the confidentiality of the surveyed households and communities. The locations have been generated following the method used by the Demographic and Health Surveys (DHS) program. The coordinates included represent a random offset of the center point of clustered household GPS locations within the enumeration area and within a certain distance depending on weather tha t community is designated urban or rural. An offset if 0 5km was used for most rural locations due to households being more dispersed. Therefore taking into account the 0 5km offset the GPS coordinates are representative at the cluster level and can be use d for spatial analysis and merging data acquired from remote sensing. Figure 3 : Rural community GPS survey points Environmental data From a social science perspective, remote sensing can be useful to provide data on the on the biophysical or environmental context that shapes social events In other words, remote ly

PAGE 34

31 sensed data can offer an additional layer for multilevel analysis by providing the biophysical context within which people live and relate to the environment (Liverman, et al., 1998) The environmental data used for this study include forest cover, percent forest change (2000 2014) and protected areas (National Parks and Wildlife Refuges) Forest cover and forest change information are necessary to determine how resource availability is related to the adoption of improved stove technology. Land cover maps developed by SERVIR for the Green House Gases Inventories project provided information on forest cover (2014) and forest change ( 2000 2014) (RCMRD, 2015). These maps were commissioned to provide baseline land use / land change data for the forestry sector. The maps were created using LANDSAT 5 thematic mapper (30 by 30 meter resolution) imagery for Eastern and Southern Africa. The l and cover maps include years 2000 and 2014. The 2000 land cover map includes imagery collected between 1999 2001 and the 2014 map includes data from 2013 2015. The overall accuracy of the data are 86.33% (436/505) and the KAPPA coefficient is 0.8444 (RCMRD 2015). Methods This chapter introduces the conceptual framework used to address this study, followed by the specific research questions derived from this framework. T he last section introduce s the methods employed to address the questions as well as the variables incorporated into the analysis. This study focuses on traditional and improved stoves which utilize firewood and charcoal fuels, due to these being the predominant fuels used in rural households. Improved stoves were considered to include efficie nt wood stoves and improved charcoal stoves. Modern

PAGE 35

32 these stoves in rural areas. Statistical analysis used in this study includes descriptive statistics and logi stic regression analysis. Conceptual framework The relationship between access to resources, biophysical characteristics and the decision to use an improved primary stove was examined with the assumption that communities with a higher demand for fuel would be the most likely incorporate efficient technology. Traditional wood fuel use i s associated with a variety of development challenges, including health econo mic and environmental concerns This research applies a human environment framework aimed at improving the understanding of how socio economic status, resource availability and biophysical characteristics influence household choice to incorporate improved cooking technology. T o unders tand how the physical environment relates to household level choices, we must first understand the likely cost s and benefits of switching to improved stoves Applying the utility theory, the cost and benef it of switching to improved stoves can be analyzed simultaneously, if the benefits outweigh costs the household will switch stoves. I hypothesize that households will be more inclined to switch to improved stoves when demand for fuel is higher (w hich reflects resource availability and higher prices of all fuel types) and less inclined to switch if households are more remote and there are plentiful resources. The households that have fewer available resources may view improved stoves as a benefit because they may save time in collecting firewood and save mo ney because of greater efficiency in burning wood or charcoal H ouseholds that have abundant resources on the other hand, may not recognize the time savings as a valuable benefit. Potential costs preventing the adoption of stoves will likely be poverty ( poor welfare) as less well off households will likely to be less inclined (or able) to switch stoves.

PAGE 36

33 Research question s In order to better understand how social and environmental context relates to the utilization of improved stove technology t his study address es the following questions: 1. Are stoves being adopted and utilized over time? (temporal analysis) 2. What are the spatial difference s in improved stove utilization? (spatial analysis) 3. Why a re households utilizing or not utilizing improved stoves? What are some of the potential benefits of improved stoves? (explanatory analysis ) 4. What are the main ecological and socio economic variables which drive the utilization of improved stoves? (analysis of relationships) T emporal a nalysis When? In order to determine the temporal trends in improved stove adoption the two surveys which include information on primary stove type used in 2000/01, 2005/06, 2010/11 were merged used. The UNHS 2005/06 data that are provided could be linked with th e 2010 /11 data at the household levels through the unique household ident ifier (HHID). A variable for household status was included in the UNPS 2010/11 data which identifies (i) households from the 2005/06 survey that were interviewed at the same location (ii) households from the 2005/06 survey that had moved and were interviewed at an alternative location After the rural household survey data from 2005/06 and 2009/10 were merged a total of 1,743 households in 233 enumeration areas were available for the pre sent analysis. In the UNHS 2005/06, households were asked what type of stove was primarily used for cooking in 2000/2001. Therefore, all households sampled both in the 2005/06 and 2009/10 surveys were used to assess changes in cooking technology and fuel use between 2001 and 2010. The 2005/2006 survey responses for primary stove options did not relate directly to the

PAGE 37

34 2010/2011 survey data (Table 2 ). Therefore, some assumptions were made to permit comparison of these data sets and d etermine differences ove r time. The most important of these was that survey. Among the households with primary charcoal stove s in 2010; 17% reported having improved charcoal stove in the 2006 survey 35% reported having a traditional metal stove, and 37% had traditional 3 stone This is concerning because traditional metal may have been Table 2 : Modifications made to relate household survey data from 2005/06 an 2010 /1 1 Response Measure Primary Stove Which is the stove used most often by this household? 20 10 Wood / Sawdust Burning, Open fire 1=Traditional Efficient Wood Burning 2= Efficient wood Charcoal 3=Improved charcoal 2005 Traditional metal stove Traditional 3 stone stove 1=Traditional I mproved firewood stove 2= Efficient wood Improved charcoal stove 3=Improved charcoal S patial a nalysis Where? In order to determine spatial relationships of improved stove utilization the hotspot analysis too in ARCGIS was use d to calculate the Getis Ord Gi statistic for every community point location in the 2010/11 survey dataset. The output provides z scores and p values which provide information on significant hot spots (high values) and cold spots (low values) based on spatial clusters. The results show how point location relates within the context of neighboring points For example, in order to be considered a statistically signif icant hot spot, a location with

PAGE 38

35 improved stove use will be surrounded by other point locations with improved stoves (Getis & Ord, 1995) Explanatory analysis Why? To understand how stove choice is related to environmental and social factors it is necessary to understand the likely costs and benefits of switching to improved stoves. Why are households utilizing or not utilizing improved stoves? What are some of the potential benefits of improved stov es? What are the regional differenced in fuel and stove use? Examining regional differences in fuel and stove use can provide information on how location characteristics may influence improved stove use. The hypothesis was addressed that regions with less resources (wood for fuel) would be more likely to adopt and utilize improved stoves. Data used to examine regional differences in fuel and stove use and potential benefits of improved stoves include variables selected based on providing a narrative about f actors that may be driving improved stove use. The 2010/11 survey was used for these questions which includes 1,970 rural households and 233 enumeration areas. The variables selected to represent resource availability include firewood collection time, fuel types used for cooking, number of fuels used for cooking, source of fuel, time spent cooking, fuel quantity used, household stove location and presence of a chimney Regional stove use should correspond to more diverse fuel types used for cooking and more sources, under the assumption that thus represents less resources availability T he potential benefits of improved stoves and how this may be influencing stove utilization efficiency of improved stoves, was examined under t he hypothesis that if benefits i mproved stoves outweigh the costs the technology would be more likely to be utilized and maintained over time.

PAGE 39

36 Analysis of r elationships What? A logistic regression model was used to investigate what ecological and socio economic variables drive the uti lization of improved stoves (efficient wood and charcoal stove). Each household is labeled yi =1 if they responded as utilizing an improved stove as their primary stove used for cooking, 0 if they primarily utilize a traditional stove model. I excluded respondents that used other models or had no stove, leaving a sample size n of 1,822 for efficient wood stoves, and 1,904 for charcoal stoves. I mo del the probability that y = 1 using, Where X included 17 household and community level predictors. D ue to nonlinearity of the logistic regression model, coefficients can be difficul t to interpret. The curve of the logistic regression required a choice where to evaluate changes, whether to interpret the results on a probability scale. Probabilities are calculated from the equation, 1. 2. 3. y = 1) corresponding to a single unit difference in x Therefore dividing the coefficients by 4 gives an upper limit of the predictive difference corresponding to one unit of difference in x (Gelman & Hill, 2007) Incorporating interactions of ecosystems and human activities at different scales, from regional, community and to household, is important in remote sensing and social studies (Serneels, et al., 2007). T he processes operating at these various scales are mostly nested and interdependent. For example, ecosystem functioning might influence micro level household

PAGE 40

37 processes, and so these factors should thus be analyzed simultaneously. I have included variables at 3 different spatial scales (1) the household nested within (2) a community which is nested within (3) a region. The results will determine the relative importance of factors driving the household decision to adopt improved stove te chnology at the house hold level. These types of models that incorporate both ground based social data and biophysical characteristics from remote sensing are able to determine relati onships on a much larger scale. Because the data are at different scales e ach household within the enumeration area (community) was given identical community and biophysical characteristics, due to not having the exact location of the househ olds. The logistic regression gives a single outcome, the probability that a respondent adopts an improved pr imary cook stove. The statistical analyses were run using R statistical software package (R Core Team, 2013) Description of variables chosen for analysis Household characteristics The household level variables included in the analysis represent demographic and socio economic characteristics. The indicators selected for analysis were found to be significant in other similar studies or have not been studied extensively. The data selected from the survey s include: (i) household information (n umber of people in a household, tenure security, female head of household number of fuel sources ); (ii) welfare indicators (measure s taken to make water safer, whether every member of the household has at least one pair of shoe s and two sets of clothes, and whether every child in the household has a blanket Welfare indicators represent poverty in place of income ; income can be subjective in rural areas where there are many subsistence farms and welfare may not be best reflected as a monetary value. Choice of an improved stove was predicted for households with improved welfare, and those with larger

PAGE 41

38 family size, larger parcel size, and greater number of fuel sources. Since investment in improved cooking technology typically requi res having the means to purchase the stove household s with improved welfare were assumed to be more likely to use modern technology. Ho use holds with many members were also assumed to be likely to adopt improved stoves due to the benefits of long term cost saving s due to fuel efficiency Number of fuel sources was chosen as a variable to potentially represent resource availability. Table 3 : Household socio economic characteristics and indicators Variable Name Type Description Count Yes Count No Charcoal stove binary Yes/No 134 1821 Efficient wood stove binary Yes/No 52 1903 Tenure security binary Yes/No 189 1754 Female head of household binary Yes/No 587 1368 Min Max Mean Safe water continuous More Safe/Less safe 0 3 0.63 Number of fuel sources continuous Purchase/Collect village/Collect own land 0 3 1.18 Welfare measure continuous 0=High poverty 1=Moderate 2=Low /not 3=Not impoverished 0 3 1.49 Household size continuous Number of people 1 30 6 Parcel Size continuous Acres 0 200 2.37 Some of the variables from the survey were aggregated for the purpose of simplifying the analysis. Several welfare indicators were provided in the 2010/11 household survey (Table 4 ). Individually these welfare indicators may not provide meaningful information about the poverty level of the household, and many of them may be correlated within a household (if one indicator is present in an impoverished household more than one may be as well). Therefore, a welfare measure was obt ained by adding the number of indicator s present within a household. Only the yes or no indicators were used that were obviously associated with poor welfare.

PAGE 42

39 Table 4 : V ariables used to measure welfare from survey welfare indicators Response Measure Welfare Measure Does every member of the household have at least two sets of clothes? Yes=1 No=0 0= very low 1=low 2=moderate 3 or (2+NA)= high Does every child in this household (all those under 18 years old) have a blanket? Yes=1 No=0 NA Does every member of the household have at least one pair of shoes? Yes=1 No=0 Safe Water What is the main source of water for drinking for your household? Private connection to pipeline Public taps Protected well/spring Vendor/Tanker truck Bore hole Unprotected well/spring River, stream, lake, pond Gravity flow scheme Rain water More Safe =1 Private connection to pipeline Public taps Protected well/spring Vendor/Tanker truck Or Boil and filter Boil only Filter only Less Safe = 0 Nothing is done And Bore hole Unprotected well/spring River, stream, lake, pond Gravity flow scheme Rain water What do you do to the water to make it safer for drinking? Nothing is done Boil and filter Boil only Filter only Community characteristics Community characteristics considered as important in influencing choice of stove for cooking by a household were specific region, distance to protected areas (National Parks and Wildlife Refuges), distance to infrastructure (major roads, markets and popula tion centers

PAGE 43

40 +20,000), forest cover (2014) and forest cover difference (2000 2014) and number of forest crop within the community (Table 5) Access to alternative energy sources have a location dimension e.g. charcoal is prominent closer in proximity to ur ban areas. D istance to a population center and markets may inadvertently represent higher population densities. It is likely that communities located closer to population centers would contain more people, due to urban sprawl. Rural communities located clo ser to population centers would likely either have less available resources due to the greater demand for fuel from the larger population. Also included in the explanatory variables, was distance to protected areas, forest cover, and forest crops, all chosen to represent resource availability. There are different types of protected areas with differing management practices and restrictions. Certain kinds of protected areas allow subsistence resource extraction. National Parks and Wildlife Refuges, however, are Figure 4 : Map of Uganda showing the major protected areas; Data Source: Uganda Wildlife A uthority 2007

PAGE 44

41 restricted access areas and allow no resource extraction; these boundaries most likely represent limited resource availa bility (Figure 4) Table 5 : Community and environmental variables representing resource demand and availability Variable measure min max mean FC Area 2014 m eters 2 0 10,521,085 2,576,945 Dist. to NP/WLR meters 0 188,155 58,238 Pct. FC Diff % 1 98.76 2.14 FC Diff m eter s 2 8160513 5840830 368707 Avg. time spent collecting firewood Hrs./7days 0 30 3.08 Number Tree Crops count 0 4 1.67 Distance to population center (+20000) km 0 101.28 25.47 Distance to main road km 0 40.43 8.75 Distance to market km 0 116.18 33.30 The forest cover was clipped to a 5 km buffer surrounding the rural co mmunity survey point location. The buffer dista nce of 5 km was chosen based on the cluster methodology which derived the point locations (see social survey data) (Figure 5 ). In addition, the total forest cover was calculated for within the 5km buffer area for 2000 and 2014, and the difference was then determined for the 14 year period A few studies found that households generally do not travel far distances to collect fuel unless resources are scarce (Tabuti, et al., 2003; GoU, 2002; Mahapatra & Mitchell, 1999) Tabuti et al. (2003) found that in Bulamogi County Uganda that most people (residing in rural areas) traveled short distances of 2 km or less to collect firewood, and that local people believe there is an abundance of firewood. Supporting this finding, the Uganda National Forest Plan (2002) found that Ugandans traveled less tha n two k m to c ollect firewood. However, from 1992 2000 there was a dramatic increase in average time spent collecting, which may have resulted from t he loss of forest resources (Figure 6) (GoU, 2002) The Mahapatra & Mitchell, ( 1999)

PAGE 45

42 study c onducted in rural India reported that people frequently traveled greater than 8 km to collect fuel where there is firewood scarcity Figure 5 : 2000 and 2014 forest cover within a 5km buffer of a community survey point Figure 6 : Average distance travelled to collect household fuelwood; Source: (GoU, 2002)

PAGE 46

43 CHAPTER V RESULTS This study investigates the social and ecological context surroun ding improved stove adoption. In order to achieve this it is necessary to explore the temporal trends, geographic differences, and potential benefits of improved stoves. This information can shed insight on whether improved stoves are being utilized as primary stoves and being sustained over time Additionally, the potential benefits of adopting improved stove technology were investigated to determine the reasons for improved stove adoption as well as to explain why households are not maintaining improved stoves as the primary stove over time. The benefits considered align with the Globa l Alliance for Clean Cookstoves ( 2011) statements, which include less time spent cooking, less fuel used due to higher efficiency, less time spent collecting firewood due to less fuel used, and the potentia l health benefits due a reduced amount of indoor air pollution. The t emporal trends of stove use Descriptive statistics were used to determine stove change from 2001 to 2010 using the 2005/06 and 2009/10 survey responses. Exploring the stove use trends over time can give insight to the social factors influencing stove adoption and long term use. Overall, the number of improved and modern stove users is increasing over time, and traditional stove users are decreasing (Figure 7). However, households adopti ng the improved stove technology are not maintaining the technology as their primary stove over time (Figures 8 and 9)

PAGE 47

44 Figure 7 : Count of stove use 2001, 2006 and 2010 n = 1852 Figure 8 shows that out of the 2 6 households who claimed utilizing effi cient wood stoves as their primary stove in 2001, 22 still used them as their primary stove in 2006. By 2010 none of those households still lis ted them as their primary stove. I nstead those 21 households switched back to using a traditional 3 stone stove as their pri mary stove and 1 had a charcoal stove. Of the additional 20 households who adopted efficient wood stoves between 2001 and 2006, only 3 still listed them as their primary stove in 2010; one household reported having a charcoal stove and the rest switched back to traditional stoves. Between the 2005/06 survey and the 2010/11 survey an additional 41 household adopted efficient wood stoves as their primary stove. 2001 2006 2010 Sum of Modern 2 6 13 Sum of Efficient Wood 26 42 44 Sum of Charcoal 70 89 127 0 20 40 60 80 100 120 140 2001 2006 2010 Total 1619 1686 1644 1580 1600 1620 1640 1660 1680 1700 Traditional Stoves

PAGE 48

45 Figure 8 : Adoption of efficient wood stoves 2001 2010 n=1852 There were mor e improved c harcoal stoves users in 2001 who maintained long term use as their primary stove through 2010 (Figure 9). Out of the 70 households who adopted improved charcoal stoves in or prior to 2001, 64 still stated them as their primary stove in 2005, 13 retained switched back to traditional stoves. An additional 25 households adopted improved charcoal stoves in 2005 who stove in 2010. In 2010 104 additional households adopted charcoal stoves as their primary stove Both improved stove types spiked in primary use in 2010. This spike may be related to the Ugastove project receiving carbon financing in 2009 and upping dis tribution goals. 2001 2005 2010 Sum of 2001 Stove Adoption 26 22 0 Sum of 2005/06 Stove Adoption 0 20 3 Sum of 2010/11 Stove Adoption 0 0 41 0 5 10 15 20 25 30 35 40 45 Number of Households

PAGE 49

46 Figure 9 : Changes in primary improved charcoal stove adoption 2001 2010 The s patial analysis of stove use The results of exploring regional differences in stove use (Table 6) revealed that the Central region has the highest percentage of primary charcoal stove users (63% of all primary charcoal stoves) and modern stove users (57% of all modern stoves). The Eas t has the highest percentage of primary efficient wood stove users (42% of all efficient wood stove s). Traditional stoves are for the most part evenly distributed throughout the country. Table 6 : Percent stove use by region 2010 stove use by region Charcoal Efficient Wood Burning Traditional Modern Central 63.43% 25.00% 21.42% 57.14% East 6.72% 42.31% 27.47% 0.00% North 18.66% 21.15% 27.98% 21.43% West 11.19% 11.54% 23.12% 21.43% Total 100.00% 100.00% 100.00% 100.00% Figure 9 shows the point survey locations where there are statistically significant clusters or hotspots and coldspots of improved ch arcoal and efficient wood stoves The hotspot for charcoal stove adoption is more predominant in the Central region, esp ecially concentrated 2001 2005 2010 Sum of 2001 Stove Adoption 70 64 13 Sum of 2006 Stove Adoption 0 25 9 Sum of 2009 Stove Adoption 0 0 104 0 20 40 60 80 100 120 Number of Households

PAGE 50

47 around the capital city Kampala. The hotspots for efficient wood stoves are more concentrated of around the protected areas, paticularly Mt. Elgon National Park in the East region, and in the western region, and near Bwindi Imp enetrable Forest in the south w estern region, which is an important area for tourism revenue due to the Mountain gorilla (Figure 10). Figure 7 : Results of a hotspot analysis showing statistical significance of clusters of improve d a) charcoal stoves b) wood stoves The potential benefits of improved stoves The potential benefits of improved stove use include potential fuel saving of improved stoves, time spent cooking and collecting wood, resource availability (examined by fuels used and sources of fuel), and cooking practices (where the stove is located in the household and

PAGE 51

48 whether there is a chimney. Table 7 shows a comparison of the quantity of firewood and charcoal used for different stove types as well as the average time spent per day using the stove. On average households who primarily use improved s toves (efficient wood burning or charcoal) use much less fuel than traditional open fire or wood/sawdust burning stoves. Efficient wood stoves use approximately 32% less fuel that open fire stoves. This finding is also reflected in time spend collecting fi rewood compared with stove use. Charcoal stove users, who generally purchase fuel, spend less average time collecting firewood than wood stove users (0.18 hours/week). Not surprisingly, households with an efficient wood stove spend less average time collec ting firewood (1.25 hours/week) than households with traditional stoves (1.52 hours/week), which is most likely related to using less fuel. The c har coal stove is less efficient tha n the wood burning stove but still more efficient than the open fire pit and w ood/ s awdust burning stoves The amount of firewood used by the w ood/ s awdust burning stove was similar to the open fire pit, much less charcoal was use d. H owever this type of stove is not common and charcoal may not be used often with this kind of stove, which could be why the numbers are less. Therefore this analysis further justifies grouping wood/sawdust burning stove into the traditional stove category and classifying charcoal as an improved stove However, the efficient wood burning stoves were used more hours per day than the less efficient traditional stoves for both fuel types (18 32% more time than open fire) This finding is consistent with what ( 2010) report. This study also found that in western Uganda households spend on a verage 3.8 hours cooking per day which is less than what the survey data shows for the national average.

PAGE 52

49 Table 7 : Average hours per day used and average quantity of firewood and charcoal used per month (30 days) based on primary stove type 20 10 Firewood Charcoal Stove Type Average Hrs / Day Used Average Qty Used Average Hrs / Day Used Average Qty Used Wood/Sawdust burning 5.79 1350 7.16 920.45 Efficient wood burning 6.39 915.38 6.76 348.86 Charcoal 4.89 864.10 5.07 7855.92 Open Fire 4.84 1261.40 5.75 9996.16 Time spent collecting firewood is another indirect indicator of resource availability at the community level. Table 8 presents the average total time spent collecting firewood by region, including the source from where households reported getting their firewood Therefore only the averages were assessed for households who responded as collecting firewood either from their land or the village, this was to reduce potential errors due to people reporting time spent collecting firewood who purchase their fuel. Overall, households spend about 2 hours collecting fuel. This finding is consistent with Tabuti et al. (2003) who reported that in Bulamogi County in Uganda most people spent less than 2 hours colle cting firewood, commonly collecting daily or once a week. T he time spent collecting by those who collect from their land is less than those who primarily get their fuel from the village. T he re is a lot of variance in fuel collection time in the Central reg ion The Western has high variance in fuel collection time from those collecting from the village.

PAGE 53

50 Table 8 : Average time spent collecting firewood of surveys who reported gathering or collecting firewood as their main source, by region. Firewood Source 2010 Average Firewood Collection Time Standard Deviation Variance Central rural Collect from own land 1.54 1.79 3.22 Collect from village 2.40 3.78 14.26 East rural Collect from own land 1.23 1.34 1.80 Collect from village 1.45 1.32 1.73 North rural Collect from own land 1.53 2.02 4.09 Collect from village 1.91 2.16 4.68 West rural Collect from own land 1.42 1.32 1.73 Collect from village 1.78 2.37 5.62 Understanding fuel characteristics by region can relate to how resources may influence household stove choices. Table 9 shows differences in rural fuel use from the 2010/11 survey by region. The results indicate that fuel characteristic s differ by region Reflecting the patterns reported from region al ch aracteristics of stove use the Western region has a greater dependency on firewood use (88.06%) The C entral region uses more charcoal (31% compared to the second highest the Northern region at about 12%). The E astern region has a higher percentage of crop residues (17.73%) used as fuel. The Northern regions has the second highest percentage of firewood, charcoal and crop residue use.

PAGE 54

51 Table 9 : Rural fuel use by region Rural fuels used for cooking 2010 Charcoal Crop residue Firewood Total Central 31.16% 5.45% 63.39% 100% East 8.51% 17.73% 73.76% 100% North 11.54% 10.07% 78.39% 100% West 5.33% 6.61% 88.06% 100% Uganda (16.24%) (9.73%) (74.03%) (100%) Table 10 reveals that most of Uganda households use no more than 2 fuels for cooking (99%). The Eastern region has the highest percent of households using 2 3 fuels for cooking (about 28%). The Western region has the least amount of fuel diversity with 91% of hous eholds relying on firewood. Table 10 : Multiple Fuels used for cooking by region Rural Fuel Use 2010 1 fuel for cooking 2 Fuels for cooking 3 fuels for cooking Total Central 434 79.34% 112 20.48% 1 0.18% 100.00% East 358 71.89% 128 25.70% 12 2.41% 100.00% North 395 77.91% 106 20.91% 6 1.18% 100.00% West 432 90.76% 43 9.03% 1 0.21% 100.00% Uganda 1619 79.83% 389 19.18% 20 0.99% 100.00% Fuel source can also be an important variable to represent r esource availability in that h ouseholds that purchase fuel are likely doing so due to a lack of access to natural resources. Fuel source may also be a factor influencing household stove choice. For example, if a household has to spend money to purchase fuel they may be more likely to adopt a more efficient stove to save money. Six primary sources of fuel were recorded in the survey: (1) Purchase from a shop; (2) Purchase from the market; (3) Public utility; (4) The black market; (5) Collect or gather from your own land; and (6) Collect or gather from the village (Table 11). The

PAGE 55

52 majority of firewood for all regions is collected from the village (69.16%) and 25.34% of households collected from their private land. The east rural and west rural households lean more towards collecting firewo od on their property. Table 10 shows the percent of the total fuel used by the source of the fuel. Charcoal is primarily purchased ( 88.8%) from either a marketplace (73.5%) or a shop (15.3%). The black market (illegal trade) is mainly used for charcoal (5 %). Firewood is mostly gathered or collected (91.3%), primarily collected from the village (67.2%), but also from the households own land (24.1%). Crop residue is primarily gathered from the Table 11 : Source of fuel normalized by 100% of each fuel type category. Source 2010 Charcoal Crop residue Firewood Gather / collect own land 3.82% 72.25% 24.79% Gather / collect village 5.10% 27.75% 68.70% Marketplace 67.20% 0.00% 3.97% Public Utility 2.23% 0.00% 0.21% Shop 18.47% 0.00% 1.48% Black market 3.18% 0.00% 0.85% Total 100% 100% 100% Overall, the Central region has more charcoal use, and the majority of charcoal is households who purchase fuel (22%). The Western region more reliant o n firewood has a greater percent of households who gat her wood from the village (63%). The East region has the most fuel collection from private land (36%), probably due to higher use of crop residues. However, all regions still have a high percentage of households who are collecting fuel from the village (44% 63%). The potential health impacts of traditional stove and fuel use was assessed by looking at the cooking characteristics of households. Houses chimneys or where the residents cook

PAGE 56

53 outdoors may be less exp osed to smoke from traditional stoves, and consequently may be less inclined to switch to improved stoves. The 2009/10 survey asked whether stoves had chimneys and where in the house stoves were located. The results in Table 1 2 show s that the vast majority of all stoves types do not have chimneys in Uganda (94% of charcoal stoves, 93% of efficient wood burning stoves, 96% of open fire pits and 100% of wood/sawdust burning). Table 12 : Percent of Stove types that have chimneys Chimney Stoves rural No Yes Total Charcoal 176 94.12% 11 5.88% 187 100% Efficient Wood Burning 50 92.59% 4 7.41% 54 100% Open fire 1704 96.71% 58 3.29% 1762 100% Wood / Sawdust Burning 43 100% 43 100% Stoves are primarily located in a separate kitchen (74%), although a surprising percentage of stoves are located outside (20%) which would not require a chimney for ventilation (Table 13). Just 6.45% of stoves were located in a main living space. Households that use traditional stoves without chimneys and are located indoors likely r epresent high levels of indoor air pollution ; when that stove is in a main room of the house, it is probable that the entire family is being exposed. Table 13 : Where stoves are located in the household by region Rural Cooking practices In a room in the dwelling not just devoted to cooking In a separate kitchen In an outdoor space Total Central 32 6.19% 337 65.18% 148 28.63% 517 100% East 30 5.23% 466 81.18% 78 13.59% 574 100% North 60 10.93% 341 62.11% 148 26.96% 549 100% West 12 2.75% 381 87.19% 44 10.07% 437 100% Uganda 134 6.45% 1525 73.42% 418 20.13% 2077 100%

PAGE 57

54 The relationship between variables driving the utilization of improves stoves The model was run using a classical logistic regression in which stove use is binary variable (1= improved stove, 0=traditional stove). Improved charcoal stoves and efficient wood stoves were separated for analysis since the use of charcoal and wood likely represent very different characteristics. An alysis of efficient wood stoves Table 14 contains results from a logistic model that was implemented using a binary indicator of efficient wood stove use as the dependent variable. Socio economic variables as well as variables representing proximity to re source demand (roads, markets, and population centers) were not associated with the choice to adopt this type of cookstove. The most significant factor that help s to explain patterns of adoption of efficient wood stove technology was the distance to National Parks (NP) and Wildlife Refuges (WLR ) (p < 0.001 ) which likely represent resource scarcity, in that those are restricted access areas, or active stove distribution areas. Households located closer to Natio nal parks and Wildlife refuges have a higher probability of adoption efficient wood stoves as their primary stove type, every meter further distance away from a protected area corresponds to approximately 50% lower probability of utilizing the efficient wo od stove as the primary stove. The Northern region (at p < 0. 05) and Western region (at p < 0.001) were significant and negatively correlated to efficient wood stove adoption, indicating that households in these regions are less likely to adopt this stove model as their primary stove. Overall the Area Under the Curve ( AUC ) value, which ranges from 0.5 to 1.0 (0.5 indicates no better than a random prediction and 1.0 is a perfect fit) the model performance was moderate (AUC=0.69), indicating that the variable s included in this model are about 70% of the variation in wood stove adoption.

PAGE 58

55 Table 14 : Logistic regression results for efficient wood stoves (n=1955) Fixed Effects Coef.est Coef.se Pr(>|z|) (Intercept) 2.110 0.890 0.018 c.HH Size 0.070 0.060 0.193 Female Head HH 0.030 0.310 0.936 Welfare Measure 0.180 0.140 0.205 Safe water Measure 0.300 0.250 0.229 Parcel Size 5.390 8.620 0.532 Secure Tenure 1.170 0.740 0.116 # Fuel Sources 0.420 0.330 0.198 Tree Crops Community 0.470 0.370 0.206 FC Difference (%) 0.000 0.020 0.835 n.Dist NP/WLR (m) 2.130 0.800 0.008 ** FC 2014 Area (m2) 0.290 0.890 0.745 Dist. road (km) 0.000 0.020 0.952 Dist. market (km) 0.000 0.010 0.799 Dist. pop. center +20000 (km) 0.000 0.010 0.812 Region : Eastern 0.150 0.460 0.740 Region : Northern 0.930 0.530 0.078 Region : Western 1.970 0.610 0.001 ** AUC 0.692 Significance Analysis of improved charcoal stoves The results from the improved charcoal model suggest that household socio economic characteristics are very much related to the utilization of improved charcoal stoves (Table 15). Factors identified as being positively associated with the adoption of charcoal stoves include older average household age, higher welfare safer water and number of fuel sources, and distance to National Parks and Wildlife Refuges. The coefficient for household welfare is positive and significant (p < 0) implying that higher welfare is associated with households that are more likely to have a charcoal stove. This result corresponds to 15% greater probability of using improved charcoal stove with every unit of increasing welfare measure. Safe water is also an important variabl e, suggests that households taking extra measures to make their water safer are 17.5% more likely to incorporate improved stoves. This variable may also be viewed as an additional welfare indicator. The greater number of fuel sources corresponds to approxi mately

PAGE 59

56 9% greater probability of charcoal stoves. Households larger that the centered value (5) are 3% more likely to utilize charcoal stoves. National Parks and Wildlife Refuges was highly significant and positively associated with improved charcoal stov e use, this result can be interpreted as households that are further from these restricted use areas have a higher probably (70%) of incorporating charcoal stoves with every meter of distance; this relationship is opposite from the efficient wood stove res ults. This relationship is likely due improved charcoal stoves being more central, while most of the NP and WLR are more remote and located closer to the tha t areas adopting improved charcoal stove are less forested. Every meter 2 less of forested area makes the households 50% more likely to primarily utilize a charcoal stove. Overall the AUC value, was quite high for the charcoal stove model (AUC= 0.87). Ther e was no relationship between having a female head of household and adoption of improved stoves in this study, however other studies on clean fuel adoption have found the gender of the household head to be a relevant facto r (Lewis & Pattanayak, 2012). The indicators of resource demand (distance to a population center, distance to a road, and distance to a market) were not significant predictors in the models. In other words, p roximity to areas of high fuel demand does n o t have a statistically significant i nfluence household's deci sion to adopt improved stove.

PAGE 60

57 Table 15 : Logistic regression results for charcoal stoves (n=1955, groups= region, 4) Fixed Effects Coef.est Coef.se Pr(>|z|) (Intercept) 4.990 0.710 0.000 *** c.HH Size 0.120 0.040 0.007 ** Female Head HH 0.230 0.230 0.320 Welfare Measure 0.610 0.110 0.000 *** Safe water Measure 0.700 0.180 0.000 *** Parcel Size 20.610 10.670 0.053 Secure Tenure 0.450 0.410 0.272 # Fuel Sources 0.360 0.200 0.068 Tree Crops Community 0.210 0.250 0.400 FC Difference (%) 0.010 0.020 0.500 n.Dist NP/WLR (m) 2.820 0.600 0.000 *** FC 2014 Area (m2) 2.010 0.740 0.007 ** Dist. road (km) 0.790 0.590 0.183 Dist. market (km) 0.280 0.830 0.740 Dist. pop. center +20000 (km) 0.160 0.850 0.851 Region : Eastern 0.870 0.440 0.048 Region : Northern 0.560 0.390 0.148 Region : Western 0.460 0.480 0.342 *** AUC 0.873 W hen comparing the averages of these variables, you can clearly see that efficient wood stove users have a much lower average welfare measure (Table 16 ). Parcel size was larger for efficient wood stoves users than charcoal stove users which is likely due to households being more remote and rural. H ouseholds utilizing charcoal stoves have more fuel sources and higher safe water measure Table 16 : Average of variables used in the regression analysis by stove type Average of Variable s Efficient Wood Stove Charcoal Stove Traditional Stove Welfare Measure (0 3) 1.58 2.35 1.42 Parcel Size (acres) 1.78 1.05 2.49 Safe water measure (0 3) 0.65 1.13 0.59 Household age 21.5 19.7 21.8 Number of fuel sources 1.15 1.35 1.19 Dist. NP/WLR (meters) 47,340 113,042 54,407 FC Area 2014 (m2) 2,794,264 1,817,464 2,628,087

PAGE 61

58 CHAPTER VI DISCUSSION While the use of improved stoves has increased since 2001 and traditional stove use has decreased, it appears that there may be a problem with households maintaining the technology as their primary stove over time. The majority of households that adopted a n improved stove in Uganda ended up switching back to a traditional primary stove within 5 years. With all of the claimed co therefore structures funded through the carbon tra de may imposing unpractical and inappropriate technology on households. One of the claimed co benefits of improved stove technology verified by this study is fuel savings. On average efficient wood stove models use less fuel than the traditional counterpart (32%) and spend less average time collecting firewood. This finding is consistent with Adkins et al. ( 2010) who evaluated fuel efficiency of the two types of improved stoves in Uganda The r esults of the study showed that both stoves were substantially more efficient burning firewood compared to the 3 stone stove. The Ugastove stove model burned 46% less wood and the Stove Tec model showed 38% fuel savings (Adkins, et al., 2010) In region s with limited resources fuel savings of 32% can have a significant impact on household security and reducing vulnerability Fuel savings means less time collecting firewood or less cost of purchasing fuel, which places less of a burden on impoverished households. However, another finding suggests that h ouseholds spent on average more time cooking with efficient wood stoves ( 18 32% more time than open fire ) Longer time spent cooking contradicts one of the co benefits claimed by development projects and non profits (Global Alliance for Clean Cookstoves, 2011) This finding is important because if households are

PAGE 62

59 spending more time cooking, they may not perceive time savings collecting wood as a significant enough benefit not to switch back to the traditional stove. In addition, longer time spent cooking may lea d to longer exposure to smoke making the claimed health benefits negligible. Based on the literature longer cooking time is not observed in all imp roved stove models, however, Adkins et al. (2010) found this to be the ca se, especially wit h the Ugastove mod el in Uganda, where i mproved stove models took anywhere from 5% to 27% more time to cook food than the traditional 3 stone stove 80% of the survey respondents complained about the increased cooking time of the improved Ugastove as well as other reported i nconveniences (i.e. too hot to touch, easy to knock over, difficult to light) and many preferred the traditional stove over the improved stove (Adkins, et al., 2010) There are some potential limitations of these results incl uding discrepancies terminology used for stove types between the two surveys (improved charcoal and charcoal) T here are also likely technical as well as cultural limitation s of improved stove models causing households to switch back to traditional primary stoves, which are outside the realm of this study Switching back to traditional stoves could also be evidence of people selling/bartering their stoves to cover other expenses especially if these st oves were received in kind from carbon or development projects In addition, primary stove use as a binary variable may not be the best reflection of household stove choice, in that multiple stove types are likely being used simultaneously. Stove stacking is also common practice in Uganda and there may be times when improved stoves are used more frequently than traditional (such as rainy seasons). There also could be technical malfunctions that limit the lifespan of the stoves or make them unpractical with households reverting to back to the reliable traditional stoves. Masera et al. (2000) found that the main technical reason for households utilizing multiple stoves in rural Mexico is that

PAGE 63

60 improved and modern stoves were not suitable for some cultural cook ing practices. A popular in which the process of steaming plantains is very traditional and involves filling the bottom of the cooking vessel with banana stems to create a space separation between the boiling water and the peeled plantains wrapped in leaves. The plantains a re steamed until they reach a desired softness, which is tested by hand (Adkins, et al., 2010). A study by Atkins et al. (2010) found that cooking this dish with an improved stove took significantly more time than cooking it with a traditional stove. Stove stacking in rural areas can also be described as tension between maintaining a cultural ident ity and simultaneously where indigenous people maintain identifying elements, such as possessions, traditions, or ideologies, in order to feel a sense of control over their cultural spaces (Masera, et al., 2000). traditional stoves a lessened benefit. Fuel use characteristics and c onsumption of fuels within households may be due to region al differences, poverty, fuel scarcity or potentially cultural differences in cooking practices. The results from this study indicate that fuel and stove characteristic s differ by region which is likely due to climatic differences, land cover features, de velopment and resource availability. The more heavily forested, tropical Western region has a greater dependency on firewood use from collect ing and gathering (88.06%) This region has the lowest percent of primary efficient wood stoves used and t he highe st percent of households who cook indoors. Which likely indicates that resources are plentiful in the western regions, however households are likely being the most adversely affected by indoor air pollution. C harcoal stoves are used more commonly as prima ry stove in the Central region. The C entral region which uses the most

PAGE 64

61 charcoal fuel (31% compared to the second highest the Northern region at about 12%), is primarily a mosaic of grasslands, savannas, and shrublands The rural areas in the Central region may also have more charcoal use due to resource depletion (less forest cover) due to higher rates of population and development. The E astern region is classified as tropical and subtropical moist broadleaf forests near Mt Elgon and tropical and sub tropical grasslands, savannas and shrublands ; it has a higher percentage of crop residues (17.73%) used as fuel. The Eastern region also has the highest percent of efficient wood stoves and household who cook outdoors. This region also has the highest perc ent of households using multiple fuels for cooking (crop residue, charcoal and firewood) which may indicate resource scarcity. The Northern region has the highest percentage of stoves located in a main area of the house. This likely means that these famili es are the most exposed to indoor air pollution and this could also be an indicator of extreme poverty. The Northern region is also more vulnerable due to recent conflict which has displace a large portion of the population. The results of the regression analysis to determine the relationship between various factors driving the utilization of improved stove technology showed that s ocio economic and demographic characteristics were important determinants of households who use a primary charcoal stove but no t for efficient wood stoves Safe w ater and higher welfare measure were found to have a strong positive relationship to utilizing an improved charcoal stove. Safe water is likely related to welfare, therefore more affluent household s are more likely to hav e access to safe water and subsequently improved stove technology. This relationship may indicate that more impoverished households are not getting access to improved stove technology, potentially due to cost This relationship could pose a problem where poorer households who are likely more vulnerable to adverse effects of traditional open

PAGE 65

62 fire stoves, continue to worsen their situation. However, it is difficult to make an assumption about the causal relationship between household adoption of improved stoves and welfare. In other words, it is unclear whether improved welfare was followed by improved stove adoption or improved stoves wer e followed by reduced poverty. In addition, i mpoverished households without access to safe water are more l ikely affected by indoor air pollution, as well as being disproportionately more vulnerable to environmental disturbances such as impacts of climate change. Also, households without access to safe water sources may need to spend more time collecting and bo iling the water, exposing the inhabitants to more smoke. Therefore, time savings and reduced smoke from improved models could be a more significant benefit for these households. Programs to promote improved stoves in areas without access to safe water coul d make progress in alleviating poverty and increasing the security of rural households that could benefit from improved stove technology. The regression results also showed community level indictors of resource availability were significant in predicting the household choice to utilize charcoal stoves. Communities with less forest cover were more likely to use charcoal stoves are their primary cooking apparatus. The distance to National Parks and Wildlife Refuges which represent protected areas with restri cted access, were important predicators for both stove types. However with opposite relationships. The charcoal stoves were more likely to be utilized in communities further from Natio nal Parks and Wildlife refuges. Households in closer proximity to protec ted areas are more likely to adopt an efficient wood stove. This finding is consistent with a study by Wallmo & Jacobson ( 1998) who found that households adjacent to national parks in western Uganda tend to implement improved stoves due to a lack of access to resources

PAGE 66

63 There are some limitations with using a general logistic regression model does not account for spatial autocorrelation due to similarities of households within communities all having the same community and biophysical characteristics. Indiv idual level analysis while straight forward ignores group level group level variation beyond that explained community level variables such as forest cover or proximity to roads (Gelman & Hill, 2007) This relationship between efficient wood stove use and proximity to National Parks and Wildlife refuges could also be due to the preferential distribution of stoves from carbon and development organizations coupling conservation goals or forest restoration projects. These o rganizations aimed at distributing improved stove technology may be influencing patterns of stove use by making the technology available to rural communities located close to protected areas which have perceived conservation importance such as Mt. Elgon a nd Bwindi National Forests However regardless of how whether or not the stove is tactfully distributed the choice to use the technology as the primary stove is still a household decision and therefore there must be a perceived benefit, whether that be fu el savings or health benefits. Parks and protected areas can be controversial due t o the inequitable social impacts. The zones around the park boundary tend to become high priority areas for environmental group s who tend to bring in international funding to pursue conservation goals, often leading to contenti on and conflict. Protected areas that are set aside for strictly ecological and financial purposes may fail to consider the social impacts of the ir establishment and management. Although inclusionary management has become increasingly more common the past decade in which communities are involved and have power over the use and management of resources. E xclusionary conservation is by far still more prominent in conservation efforts despite the

PAGE 67

64 negative social impacts and the uncertain effectiveness of protected areas in protecting forests (Bates & Rudel, 2000). Conservation and sustainability are difficult to achieve in times of immediate uncertainty whether related to extreme poverty, civil unrest or both. Uganda has had a long history or both. Beginning with British colonization where natural resources and property rights were suddenly confiscated and National parks and protected areas were established restricting access to resources and evicting loca l people from their ancestral land. To a series of militant dictatorships resulting in death and free for all resource grab, which oc curred only in the recent past. There are certain conservation priorities such as the Bwindi National Forest and World Heritage Site, which is considered the most important forest in Uganda to westerners aiming to protect biodiversity including the endangered mountain gorilla, as well as by the government for economic incentives due to high amounts of tourism revenue. Like many of the naturally remaining forests in Uganda Bwindi is surrounded by dense agriculture and very little forest remains outside the parks boundaries (Hamilton, et al., 2000). Due to the high reliance on wood fuel for energy, there are concerns regardi ng pressure to illegally harvest wood from inside park s and protected forests

PAGE 68

65 CHAPTER VII CONCLUSION The objectives of this study were to examine the social and environmental context related to the utilization of improved stove use in rural Uganda. Th e results of this study suggest that location, resource availability (i.e. proximity to protected areas, forest cover) and welfare play and important role in determin ing improved stove utilization. T his study has observed a high dependency on traditional fue ls and stoves u sed for cooking in rural Uganda n communities While there has been an increase in improved stove use, there are definite spatial trends related to the utilization of these stoves which differ based on fuel type (firewood or charcoal). While variables representing welfare were found to have an influence on charcoal stove use, i t appears as th ough environmental context is also an important factor inf luencing improved stove utilization in Uganda. Therefore it is important to consider not only h ow formal structures influence household level decision making, but also the influence of informal structures, such as resource availability. Although improved stove use is increasing there may be issues as to whether the technology is being maintained as the households primary stove overtime This raises questions as to whether the technology is practical and whether improved stoves technology under carbon financing are able to achieve both climate and development goals. Some of the finding s presented in this study refute the potential for win win scenarios or co benefit s promoted by organizations and development programs funded to implement improved stoves. Stove development projects may have a difficult time fitting technology solutions onto a diverse social ecological landscape and as a result, there are mixed outcomes. While improved stove technology can offer various advantages, those technologies need to be adaptable and practical

PAGE 69

66 within the household. Unfortunately, stove development projects tend to follow certain pathways of convenience, efficiency, and have particular prerogatives such as standardizing technology and distribution preferences and it may be difficult to overcome those tendencies. As population continues to increase and fuel become s scarcer due to increasing rates of deforestation, communities become more vulnerable and environmental degradation exasperates poverty H igh levels of poverty and constant conflicts arising within the country and in the surrounding areas makes exclusiona ry and top down management of forest resources in order to protect the countries rare biodiversity and promote tourism difficult. Therefore a ny conservation pol itical and economic instability. These efforts must be compatible with supporting local communities and alleviating poverty in order to be successful. Being that traditional wood fuels are still predominantly used, with increasing demand for fuel and little forest remaining outside of protected areas there is a need to provide alternative fuel sources and efficient stove technology Co benefits and win win policy objectives are possible, however there must be more consideration for location and demand for technologies to be utilized. R ural areas surrounding protected forests may be vulnerable due to restricted access to resources and alternati ve fuels However, conservation efforts should also keep in mind the likely impacts of displaced environmental degradation from growing demand for charcoal from urban centers. Providing efficient stoves to urban populations may have more success in reducing forests loss. The s e households may also b e more impacted by indoor air pollution due to dense and confined l iving quarters. A ccess to clean and more fuel efficient stoves has the potential to be a critical element in reducing vulnerability of impoverished people, as well as prolonging resources and protecti ng

PAGE 70

67 the countries biodiversity. An important lesson to b e drawn from this study is that informal structures outside direct policies promoting conservation and development are lik ely influencing the success of improved stove projects which depends solely on house holds utili zing the technology. Household decisions likely depend on a number of internal and external factors, such as culture resource availability, or welfare It follows that future consideration should be given to the context within which communities are embedded, which will likely influenc e the successful utilization of improved stoves. When lofty polices are implemented aiming to promote development or progress in the global south, while also reducing global emissions and deforestation this often results in grouping populations of people w hich over simplifies solutions into a one model fits all approach Which raises questions such as; what is the ultimate goal of improved stoves? F or whom are these improved stoves manufactured ? At whose cost?

PAGE 71

68 REFERENCE S Adams, W. M. & Hutton, J., 2007. People, Parks and Poverty: Political Ecology and Biodiversity Conservation. Conservation and Society, 5(2), pp. 147 183. Adkins, E. et al., 2010. Field testing and survey evaluation of household biomass cookstoves in rural sub Saharan Af rica. Energy for Sustainable Development, Volume 14, pp. 172 185. Aggrey, N., Wambugu, S., Karugia, J. & Wanga, E., 2010. An Investigation of the Poverty Environmental Degredation Nexus: A Case Study of Katonga Basin in Uganda. Research Journal of Environm ental and Earth Sciences, 2(2), pp. 82 88. Arnold, J. E. M., Kohlin, G. & Persson, R., 2006. Woodfuels, livelihoods, and policy interventions: changing perspectives. World Development, 34(3), pp. 596 611. Bailis, R., Drigo, R., Ghilardi, A. & Masera, O., 2 015. The carbon footprint of traditional woodfuels. Nature Climate Change, Volume 5, pp. 266 272. Bates, D. & Rudel, T. K., 2000. The Political Ecology of Conserving Tropical Rain Forests: A Cross National Analysis. Society & Natural Resources: An Internat ional Journal, pp. 13:17, 619 634. Benson, T., Mugarura, S. & Wanda, K., 2008. Impacts in Uganda of rising global food prices: the role of diversified staples and limited price transmission. Agricultural Economics, Issue 39, pp. 513 524. Brickwell, E., McF arland, W. & Mwayafu, D. M., 2012. Unlocking progress on REDD+: sector coordination in Uganda, London: Overseas Development Institute. Bush, G. K., Nampindo, S., Aguti, C. & Plumptre, A. J., 2004. livelihood and ecosystems appro ach, Kampala, Uganda: Wildlife Conservation Society. Bush, G., Nampindo, S., Aguti, C. & et al, 2005. The Value of Unganda's forests: A livlihoods and ecosystems approach, Natonal Forest Authority Kampala, Uganda: Consultancy Report to Wildlife conservatio n Society, Albertine Rift Programme. Byakola, T. & Mukheibir, P., 2009. Energy systems: vulnerability adaptation resilience (VAR) regional focus: sub Saharan Africa, Uganda, Paris: In Hellio International. Chiputwa, B., Spielman, D. J. & Qaim, M., 2015. Fo od standards, certification, and poverty among coffee farmers in Uganda. World Development Volume 66, pp. 400 412. CIA, 2013. The World Factbook Washington, DC: Central Intelligence Agency. Cooke, P., Kohlin, G. & Hyde, W., 2008. Fuelwood, forests and community management: evidence from household studies. Environment and Development Economics, 13(01), pp. 103 135. Egeru, A. S., 2014. Rural Households' Fuelwood Demand Determinants In Dryland Areas of Eastern Uganda. Economics, Planning, and Policy, 9(1), pp. 39 45.

PAGE 72

69 Ekeh, O., Fangmeier, A. & Mller, J., 2014. Quantifying greenhouse gases from the production, transportation and utilization of charcoal in developing countries: a case study of Kampala, Uganda. Int J Life Cycle Assess, Volume 19, pp. 1643 1652 Emerton, L., 2001. Community Based Incentives for Nature Conservation, s.l.: IUCN The World Conservation Union. FAO, 2006. Global forest resources assessment 2005: progress towards sustainable forest management, Rome,: Food and Agriculture Organisation o f the United Nations,. Foell, W., Pachauri, S., Spreng, D. & Zerriffi, H., 2011. Household cooking fuels and technologies in developing economies. Energy Policy, Volume 39, pp. 7487 7496. Geist, H. J. & Lambin, E. L., 2002. Proximate causes and underlying driving forces of tropical deforestation. BioScience, 52(2), pp. 143 150. Gelman, A. & Hill, J., 2007. Data analysis using regression and multilevel hierarchical models. 1st ed. New York: Cambridge Unitversity Press. Geoghegan, J. et al., 1998. "Socializin g the pixel" and "pixelizing the social" in land use and land cover change. In: D. Liverman, E. F. Moran, R. R. Rindfuss & C. P. Stern, eds. People and pixels: linking remote sensing and social science. Washington, D.C.: National Academy Press, pp. 51 69. Getis, A. and J.K. Ord. 1992. The Analysis of Spatial Association by Use of Distance Statistics. Geographical Analysis, 24(3). Global Alliance for Clean Cookstoves, 2011. Igniting change: a stategy for universal adoption of clean cook stoves and fuels Wa shington DC: United Nations Foundations GoU, 2002. The National Forest Plan, Kampala: Ministry of Water, Lands and Environment, Republic of Uganda. Hamilton, A., Cunningham, A., Byarugaba, D. & Kayanja, F., 2000. Conservation in a region of political ins tability: Bwindi Impenetrable Forest, Uganda. Conservation Biology, 14(6), pp. 1722 1725. Hartter, J., 2010. Resource use and ecosystem services in a forest park landscape. Society & Natural Resources, 23(3), pp. 207 223. Hutton, G., Rehfuess, E. & Tediosi F., 2007. Evaluation of the costs and benefits of interventions to reduce indoor air pollution. Energy for Sustainable Development, 11(4), pp. 34 43. IFAD, 2012. Enabling poor rural people to overcome poverty, Rome, Italy: International Fund for Agricultural Development. IMF, 2010. Uganda: Poverty Reduction Strategy Paper, s.l.: IMF Country Report No. 10/141.

PAGE 73

70 Kaimowitz, D. & Angelsen, A., 1998. Economic models of tropical deforestation: A review, Bogor, Indonesia: Center for International Forestry Research (CIFOR). Kazoora, C., Birungi, Z. S. & Dranzoa, L., 2008. Poverty forest linkages in Uganda, Washington DC: PROFOR. Kees, M. & Feldmann, L., 2011. The role of donor organisations in promoting energy effcient cook stoves. Energy Policy, Volume 39, pp. 7595 7599. Khundi, F., Jagger, P., Shively, G. & Sserunkuuma, D., 2011. Income, poverty and charcoal production in Uganda. Forest Policy and Economics, 13(3), pp. 199 205. Kituyi, E. et al., 2001. Biofuel availability and domestic use patterns in Keny a. Biomass and Bioenergy, Volume 20, pp. 71 82. Lewis, J. J. & Pattanayak, S. K., 2012. Who Adopts Improved Fuels and Cookstoves? A Systematic Review. Environmental Health Perspectives, 120(5), pp. 637 645. Liverman, D., Moran, E. F., Rindfuss, R. R. & Ste rn, P. C., 1998. People and Pixels: Linking Remote Sensing and Social Science. 1 ed. Washington DC: National Academy Press. Lopez, R., 1992. Environ m e ntal degredation and economic openness in LDCs: the poverty linkage. Amercian Journal of Agriculture and E conomy, pp. 1183 1243. Luoga, E. J., Witkowski, E. F. & Balkwill, K., 2000. Economics of charcoal production in miombo woodlands of eastern Tanzania: some hidden costs associated with commercialization of the resources. Ecological Economics, Volume 35, p. 243 257. Mahapatra, A. K. & Mitchell, C. P., 1999. Biofuel consumption, deforestation, and farm level tree growing in rural India. Biomass and Bioenergy, Volume 17, pp. 291 303. Martinez Alier, J., 2002. The Environmentalism of the Poor: A Study of Ecologi cal Conflicts and Valuation. Cheltenham, UK: Edward Elgar Publishing Limited. Masera, O. R., Saatkamp, B. D. & Kammen, D. M., 2000. From Linear Fuel Switching to Multiple Cooking Strategies: A Critique and Alternative to the Energy Ladder Model. World Deve lopment, 28(12), pp. 2083 2103. May Tobin, C., 2011. Chapter 8: Wood for Fuel, s.l.: Union of Concerned Scientists: The Root of the Problem. McCarthy, J., 2002. First World political ecology: lessons from the Wise Use movement. Environment and Planning p p. 34:1281 1302. Mwampamba, T. H., 2007. Has the woodfuel crisis returned? Urban charcoal consumption in Tanzania and its implications to present and future forest availability. Energy Policy, Volume 35, p. 4221 4234. Mwaura, F., Okoboi, G. & Ahaibwe, G., 2014. energy in Uganda, Kampala, Uganda: Economic Policy Research Centre (EPRC).

PAGE 74

71 Nabanoga, G., Namaalwa, J. & Ssenyonjo, E., 2010. REDD Working Papers: REDD and sustainable development perspective from Uganda, London: IIED. Nakakaawa, C. A., Vedeld, P. O. & Aune, J. B., 2011. Spatial and temporal land use and carbon stock changes in Uganda: implications for a future REDD strategy. M itig Adapt Strateg Glob Change, p. 16:25 62. Naughton Treves, L., 1999. Whose animals? A history of property rights to wildlife in Toro, Western Uganda. Land Degradation and Development, Volume 10, pp. 311 328. Naughton Treves, L., Kammen, D. & Chapman, C. 2007. Burning biodiversity: Woody biomass Biological Conservation, Volume 134, pp. 232 241. NEMA, 2009. Uganda Atlas of Our Changing Environment, Nairobi : National Environment Managem ent Authority United Nations Environment Programme. NEMA, 2014. Fifth national report to the converntion on biological diversity, Kampala, Uganda: National Environment Management Authority. NFA, 2007. Gazetted Areas GIS database, Kampala, Uganda: Governm ent of Uganda, National Forest Authority. Obua, J., Agea, J. G. & Ogwal, J. J., 2010. Status of Forests in Uganda. African Journal of Ecology, 48(4), pp. 853 859. Okello, C., 2014. Assessing sustainability of bioenergy systems in developing countries: meth odological development and application, University of Naples Federico II: Department of Agriculture Olson, D. M. et al., 2001. Terrestrial ecoregions of the world: a new map of life on Earth. Bioscience, 51(11), pp. 933 938. Otieno A. C. & Buyinza, M., 2010. Collaborative Forest Management in Uganda: A Strategy for Controlling Deforestation in West Bugwe Forest Reserve, Busia District. Earth Sciences, 4(2), pp. 95 102. Panwar, N. L., Kurchania, A. K. & Rathore, N. S., 2009. Mitigation of greenhouse gases by adoption of improved biomass cookstoves. Mitigation and Adaption Strategies for Global Change, Volume 14, pp. 569 578. Peskett, L., Schreckenberg, K. & Brown, J., 2011. Institutional approaches for c arbon financing in the forest sector: learning lessons for REDD+ from forest carbon projects in Uganda. Environmental Science & Policy, Volume 14, pp. 216 229. Plumptre, A. et al., 2010. Impacts of fire and large mammals on the ecology of Queen Elizabeth P ark, Uganda, s.l.: Wildlife Conservation Society and Woods Hole Research Centre.

PAGE 75

72 Pouliot, M., Treue, T., Darko Obiri, B. & Ouedraogo, B., 2012. Deforestation and the Limited Contribution of Forests to Rural Livelihoods in West Africa: Evidence from Burkina Faso and Ghana. Royal Swedish Academy of Sciences, Volume 41, pp. 738 750. R Core Team, 2013. R: A language and environment for statistical computing, Vienna, Austria: R Foundation for Statistical Computing. RCMRD, 2015. Uganda Land Cover 2000 and 2014 Sc heme II. s.l.:s.n. Robbins, P., 2005. Political Ecology. In: s.l.:Routledge. Ruhanga, R. & Manyindo, J., 2010. s.l.: UWS and UNEP/GRID Arendal 2010. Serneels, S., Linderman, M. & Lambin, E. F., 2007. A multilevel analysis of the impact of land use on interannual land cover change in East Africa. Ecosystems 10(3), pp. 402 418. Simon, G., 2014. If you can't stand the heat, get into the kitchen: obligatory passage points and mutual ly supported impediments and the climate development interface. Royal Geographical Society, 46(3), pp. 268 277. Simon, G. L., 2010. The 100th Meridian, Ecological Boundaries, and the Problem of Reification. Society and Natural Resources, Volume 0, p. 1 7. Simon, G. L., Bumpus, A. G. & Mann, P., 2012. Win win scenarios at the climate development interface: Challenges and. Global Environmental Change, pp. 22:275 287. Tabuti, J. S., Dhillion, S. S. & Lye, K. A., 2003. Firewood use in Bulamogi County, Uganda: S pecies selection harveesting and consuption patterns. Biomass and BioEnergy, Volume 25, pp. 581 596. Turyahabwe, N. & Banana, A. Y., 2008. An overview of history and development of forest policy and legislation in Uganda. International Forestry Review, 10( 4), pp. 641 656. UNDP, 2015. Human Development Report: work for human development New York, NY: United Nations Developement Programme. UNSD, 2016. UNdata. [Online] Available at: http://data.un.org/CountryProfile.aspx?crName=Uganda [Accessed 5 May 2016]. Waiswa, D. et al., 2011. Decentralised forest management in Uganda: local communities' participation and forest sustainability, Hyberabad India: Paper to the 13th meeting of the International Association for the study of the Commons (IASC). Wallmo, K. & J acobson, S. K., 1998. A social and environmental evaluation of fuel efficeinet cook stoves and conservation in Uganda. Environmental Conservation, 25(2), pp. 99 108. WHO, 2006. Fuel for life: household energy and health, Geneva, Switzerland: World Health O rganization.