Climate Variability : an entry point to assess vulnerability of small-holder farmers in Samia District, in Western Kenya

(1)

Climate Variability: An entry point to assess vulnerability of

small-holder farmers in Samia District, in Western Kenya

University of Amsterdam

International Development Studies

Research Master Thesis

Student: Georgia Papoutsi

Student number: 10463933

Email: jeopap@hotmail.com

(2)

2 | P a g e

Supervisor

Dr. Nicky R.M. Pouw

Assistant Professor, Research Master International Development Studies Graduate School of Social Sciences

University of Amsterdam Nieuwe Achtergracht 166 1018 WV Amsterdam The Netherlands Phone: +31(0)20-5254105 E-mail: n.r.m.pouw@uva.nl http://home.medewerker.uva.nl/n.r.m.pouw/

Second Reader

Dr. Hebe J.L.M. Verrest

Assistant Professor International Development Studies, Director Research Master Urban Studies Graduate School of Social Sciences

University of Amsterdam

Nieuwe Achtergracht 166 | room B 4.06 1018 WV Amsterdam The Netherlands Phone: +31(0) 205254180 E-mail: h.j.l.m.Verrest@uva.nl http://home.medewerker.uva.nl/h.j.l.m.verrest/

Local Supervisor

Dr. Hannington Odame

Centre for African Bio-Entrepreneurship Lavington

P.O. Box 25535-00603, Nairobi Kenya

(3)

3 | P a g e

List of Tables

Table 1 . Operationalization of vulnerability index and the functional relationship of variables to vulnerability ... 27

Table 2. Sampling of households for villages within Luchululo and Bujwang’a sub-locations ... 36

Table 3. Sample Population Characteristics for Bujwang’a and Luchululo Sub-locations ... 49

Table 4. Interaction of remittances with indicators other than financial ones ... 56

Table 5. Ranking of information sources ... 76

Table 6. Statistical significance of participation in social groups to the perceptions of farmers with regard to climate variability ... 78

(6)

6 | P a g e

List of Figures

Figure 1: Integrated vulnerability context associated with climate variability ... 28

Figure 2: Vulnerability index of villages within the two sub-locations of Luchululo and Bujwang’a ... 52

Figure 3: Contribution of Sustainable Rural Livelihoods Approach capitals to the vulnerability of villages ... 53

Figure 4: Coping strategies of households to temperature and rainfall fluctuations ... 62

Figure 5: The coping strategies of vulnerability groups found in Samia area ... 63

Figure 6: Strategies of Samian households for coping with food shortage ... 66

Figure 7: Crops cultivated by the majority of Samian households within the area of Bujwang’a and Luchululo ... 67

Figure 8: Total amount of rainfall for the period 2008-2012 ... 73

(7)

7 | P a g e

Abbreviations

ACCI Adaptation to Climate Change and Insurance

AI Average Index

ASAL Arid and Semi-Arid Lands

CABE Center for African Bio-Entrepreneurship CIA Central Intelligence Agency

DFID British Department for International Development ENSO El Niño Southern Oscillation

FAO Food and Agriculture Organization FGD Focus Group Discussions

GDP Gross Domestic Product

GIS Geographical Information Systems

GoK Government of Kenya

HIV Human Immunodeficiency Virus IDS Institute for Development Studies

IFAD International Fund for Agricultural Development IPCC Intergovernmental Panel on Climate Change

KACCAL Kenya Adaptation to Climate Change in Arid and Semi-Arid Lands KAPAP Kenya Agricultural Productivity and Agribusiness Project

KMI Kenyan Meteorological Institute KNBS Kenyan National Bureau of Statistics MoA Ministry of Agriculture

NCCAP National Climate Change Action Plan NCCRS National Climate Change Strategy NGO Non-Governmental Organization

NPBMF National Performance and Benefit Measurement Framework

PALWECO Program for Agriculture and Livelihoods in Western Communities in Kenya SL Sustainable Livelihoods

SLA Sustainable Livelihoods Approach SLF Sustainable Livelihoods Framework SPSS Statistical Package for Social Sciences SRLA Sustainable Rural Livelihoods Approach UNDP United Nations Development Programme UNEP United Nations Environment Programme

UNICEF United Nations International Children's Emergency Fund

WB World Bank

WCED World Commission on Environmentand Development WEDO Women's Environment and Development Organization

(8)

8 | P a g e

Acknowledgements

First and foremost, I would like to express my cordial gratitude to all people of Samia District that participated in my research, opened their houses, narrated their stories and shared their fears and expectations on rural development with me. My thesis would have never been accomplished without their help and input. I am always grateful to the sub-chief of Bujwang’a, Mr. Leonard that was always available to assist me, inviting me to all local events and gatherings, facilitating me to meet the local community of Bujwang’a. I would like to express as well my gratitude to Mr. Charles and Mr. David from Mumbao, in Luchululo, who supported my research in the area from the first day, even without the absolute support of most of the village elders. Without them my fieldwork in Luchululo would have been impossible. Many thanks also have to be given to my skillful and dedicated research assistants Erina and Tandi who were my colleagues and supporters in all fieldwork difficulties.

My parents deserve my deepest gratitude for prioritizing always my education. They have generously supported me both financially and psychologically to go through this two-year educational challenge. Being farmers in another rural context, they constituted inspiration to get involved in agriculture and study its multiple stressors through a social sciences perspective. I would like to extend my gratitude to my promoter Nicky Pouw, whose expertise, understanding and patience, added considerably to my graduate experience.

I am always grateful to Mr. H. Odame, Mr. L. Haggai, Mrs. E. Kangai and Mr. O. Okumu for their expertise and support throughout fieldwork. Their warm welcoming meant a lot for me and this why they are my CABE family. Thanks must also be sent to Mr. E.Ogombe who shared his knowledge in the area with me during the first difficult weeks of fieldwork. A special thank you to Mr. H. Odame and his Samian family (Mrs. Lucy and Mr. Sammy) for their hospitality, facilitation and friendship.

Last, I want to thank all my friends that supported me these two years to accomplish this study. Special thanks to my friends V.Palleki, D. Petalios and S. Hammer for proofreading and making useful suggestions.

(9)

9 | P a g e

Abstract

Climate variability is predominantly challenging in rainfed agricultural systems, such as those found in the countries of Sub-Saharan Africa and more specifically in Kenya. One main cause of variability is the El Niño Southern Oscillation (Stige et al. 2006; Giannini et al. 2008). The ENSO phenomenon causes high season-to-season variability and adds to small-holder farmers’ difficulty in deciding how to plan and manage their seasonal activities (Morton 2007), increasing their livelihood insecurity. Several studies assess livelihoods through the construction of vulnerability indexes (Hahn et al., 2009; Fraser et al., 2011). Following this line of thought, this study uses the bottom-up approach of sustainable livelihoods in combination with a vulnerability assessment index. The research question that is addressed is how climate variability has affected the livelihood strategies of vulnerable small-holder farmers over a recent five-year period (2008-2012) in Samia in Western Kenya. Using meteorological data, household survey data and multiple participatory tools, this study explores and assesses the characteristics that make Samians vulnerable to climate variability, and delves into local perceptions with regards to climate variability, focusing on the social capital (participation in social groups, access to and use of information) of small-holder farmers. Results show that local communities experience different levels of vulnerability to climate variability based on their access to the capital assets of sustainable livelihoods approach, (social, human, and physical, financial and natural capital). Moreover, the analysis indicates that the official meteorological data do not coincide with farmers’ perceptions. Social capital is found to be catalytic on farmers’ perceptions. Overall the findings suggest that policy makers, development practitioners and meteorological services need to proceed on interventions that target assets, mainstream climate variability adaptation strategies, strengthen social capital and enhance channels of dissemination of weather information (other than social networks) to the vulnerable farmers at the local level.

Keywords: climate variability, smallholder farmers, Kenya, vulnerability assessment, information, social assets

(10)

10 | P a g e

CHAPTER 1- Introduction

The Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (IPCC 2007) has revealed that any change in climate will have detrimental socioeconomic impacts all over the world, but especially in Africa because of the vulnerability of society and the sensitivity of the environment there. Fluctuations in temperature and rainfall have a great impact on the natural resources that determine the livelihoods of rural communities. Therefore, climate variability is predominantly challenging in rainfed agricultural systems, such as those found in the countries of Sub-Saharan Africa and more specifically in Kenya. One main cause of weather anomalies is the El Niño Southern Oscillation (ENSO) (Stige et al., 2006; Giannini et al., 2008), which produces high season-to-season variability, thereby increasing the vulnerability of those who are highly reliant on farming and livestock, two climate-sensitive economic activities (Anyah & Qui, 2012). Generally, agricultural production as well as food security is threatened by climate change and variability in the majority of African regions (Nelson et al., 2009), as they disrupt farming activities.

Interactions between human and biological environments have been assessed through the construction of vulnerability indexes (Hahn et al., 2009; Fraser et al., 2011). These two components should not be viewed separately because understanding how vulnerability is manifested in a context-specific environment by investigating a vulnerable group of small-holder farmers could contribute to policy and decision-making with regard to resource allocation in a specific area (Eakin and Bojorquez-Tapia, 2008). Moreover, Hulme (2010) notes that climate is a constructed idea having complex physical and cultural connotations. Given the fact that ecosystem dynamics and climate variability in the region of Africa have not been thoroughly studied (Shongwe et al., 2009), it seems important to understand the interaction between social perceptions with reference to information dissemination regarding environmental changes, as this can stimulate more focused adaptation actions within the context of farm-level decision making (Hulme 2010; Reid & Vogel 2006; O’Brien et al., 2009; Boissière et al., 2013).

This study aims to illustrate how climate variability has affected the livelihood strategies of vulnerable small-holder farmers over a recent five-year period (2008-2012) in Samia in Western

(11)

11 | P a g e

Kenya. Adopting a mixed-methods approach, the study explores and assesses the characteristics that make Samians vulnerable to climate variability, and delves into local perceptions with regards to climate variability, focusing on the social capital (participation in social groups, access to and use of information) of small-holder farmers. The two locations that were selected for this study, the sub-locations of Bujwang’a and Luchululo, are in Samia District. The area is rather underserviced and relatively neglected, with a large population depending on rainfed agriculture, thus providing much of interest for this specific research topic. Applying an integrated conceptualization of vulnerability that blends the sustainable livelihoods approach with vulnerability assessment index to this specific research locale enabled me to describe the multiple vulnerabilities and social processes that hinder adaptation to climate variability.

The thesis consists of seven chapters. Following Chapter 1, the Introduction, Chapter 2 gives an in-depth overview of the concepts of vulnerability and sustainable livelihoods. Chapter 3 elucidates the research design that guided this study and Chapter 4 describes the research context, from national to district level, and the research locations. Chapter 5 explores how vulnerability to climate variability is manifested among small-holder farmers and how they cope with climate variability, focusing on stressors that prevent adaptation. Chapter 6 investigates how small-holder farmers have perceived climate variability over a recent five-year period (2008-2012). The two analytical chapters (Chapter 5 and 6) intend to show that adaptation to climate variability depends on multiple stressors that increase the vulnerability of the rural poor. Strategies, perceptions and capacities are interconnected, thus influencing adaptation to climate variability. The last chapter (Chapter 7) responds to the main research question, reflects on the methodological and theoretical approach that was employed in this study, referring to the limitations of the study, and suggests several policy and research recommendations to address livelihood security.

(12)

12 | P a g e

CHAPTER 2-Theoretical Framework

The purpose of this chapter is to provide an in-depth overview of the concepts that will be presented in this thesis and to introduce previous research conducted on vulnerability assessment topics. The theoretical dimensions of the concepts will be presented to make explicit the following research analysis. This chapter is divided into two sections, one discussing the concept of sustainable livelihoods and the other the concept of vulnerability. Many points from the previous research work on these subjects/concepts, which are referenced in this chapter, will be reconsidered in the analysis chapters of the thesis.

2.1 Background of the Sustainable Rural Livelihoods Approach

The concept of sustainable livelihoods, and its relationship to development, has been widely analyzed by scholars throughout the last decades. It is a concept that has given researchers the opportunity to go one step further in their analysis of poverty eradication. The conservative definitions and approaches of the past focused only on the material dimensions of poverty without taking into account the context within which poverty has its roots. Then, in the early 1980s, vulnerability and participation became two of the central concepts in the analysis of poverty, while the multiple institutional factors and processes that influence the ability of the poor to make a living in a sustainable way were also included.

The sustainable livelihoods concept finds its early definitions in the Brundtland report (WCED, 1987), which first raised awareness of resource, environment and development issues (Chambers, 1987). The concept was expanded further in 1992, when Chambers and Conway gave the first definition of “sustainable livelihood.” They stated that:

“A livelihood comprises the capabilities, assets (stores, resources, claims and access) and activities required for a means of living; a livelihood is sustainable which can cope with and recover from stress and shocks, maintain or enhance its capabilities and assets, and provide

(13)

13 | P a g e

sustainable livelihood opportunities for the next generation; and which contributes net benefits

to other livelihoods at the local and global levels and in the short and long term.” (p. 7) This definition embodies three fundamental attributes: the possession of human capabilities,

access to tangible assets (e.g., stores and material resources) and intangible assets (claims and access), and the existence of economic activities (Krantz, 2001). In their definition, Chambers and Conway (1992) support the idea that sustainable livelihoods in accordance with capabilities are the means that can lead to well-being while sustainability and equity constitute essential principles of the livelihoods approach. Moreover, Chambers (1997, p. 9-12) and Chambers and Conway (1992, p. 35-36) recognize five elements of the definition that are considered critical not only in the evolution of the livelihoods approach but also in the domain of development per se: well-being, livelihood security, capability, equity and sustainability. The livelihoods approach places the poor and the marginalized in the center of development (Chambers, 1995) and aims at their empowerment.

In 1999, Carney provided a simpler definition of “livelihood” which was related to the definition of Chambers and Conway. She stated that:

“A livelihood comprises the capabilities, assets (including both material and social resources) and activities required for a means of living.” (p. 4)

To incorporate the notion of sustainability, Ian Scoones (1998) provided a more comprehensive definition for sustainable livelihoods on behalf of the Institute for Development Studies (IDS) and the British Department for International Development (DFID). The definition of Sustainable Livelihoods (SL) was then given as:

“A livelihood comprises the capabilities, assets (including both material and social resources) and activities required for a means of living. A livelihood is sustainable when it can cope with and recover from stresses and shocks and maintain or enhance its capabilities and assets, while not undermining the natural resource base.”(p. 5)

The difference between the definition of Chambers and Conway (1992) and the later definition by Scoones (1998) is that Scoones does not include “net benefits to other livelihoods” as a prerequisite for a livelihood to be considered as sustainable. The Scoones definition was the precursor of a framework for the analysis of sustainable rural livelihoods. This framework had

(14)

14 | P a g e

three basic elements: livelihood resources, livelihood strategies, and institutional processes and organizational studies.

Livelihood resources are conceptualized as different types of “capital” which constitute a resource base “from which different productive streams are derived from which livelihoods are constructed” (Scoones 1998, p. 7). The following five forms of capital are vital for the analysis of livelihood security and embrace the people-centered component that characterizes the approach (Scoones, 1998, p. 7-8; Carney, 1999, p. 3, 7; FAO, 2005):

 Human capital, which consists of education, knowledge, skills and labor that enable households to pursue successful livelihood strategies. Differentiations of human capital affect the other capitals tremendously. For example, poor education can affect the adoption of agricultural innovations because education constitutes an essential contributor to the diffusion of technology and the efficiency of production (see also Welch, 1970; Jamison & Lau, 1982; Feder et al., 1985; Strauss et al., 1991; Binam et al., 2004). At the household level, household size plays a determining role on the livelihood strategies that it adopts.

 Physical capital, which consists of livestock, equipment, vehicles, houses, irrigation pumps, communications and equipment useful in production. Lack of irrigation pumps either for household or farm use means that people must spend a lot of their daily time fetching water for multiple uses instead of engaging in agricultural activities, thereby diminishing labor available for livelihood activities. The improvement of physical capital can be of great importance for rural households.

 Natural capital, which consists of land, forests, water, grazing, air, biodiversity. Because natural capital constitutes the base for sustaining rural livelihoods, lack of productive land constrains tremendously small-holder farmers. Depletion of natural resources is a common characteristic of rural poor areas because farming methods are limited while lack of electricity increases the use of charcoal, which results in deforestation.

 Financial capital, which consists of savings or debt, income, credit, insurance and remittances and any other financial means. Being well-off gives a household the opportunity to educate its children, to access healthcare in times of disease, and to buy food. Financial means refers not only to monetary capital but also to assets that can be

(15)

15 | P a g e

sold and converted to money. Therefore, possessing land and livestock that can be converted to liquidity enhances successful livelihood strategies.

 Social capital, which describes the social networks, group membership, neighborhood/religious associations, social support deriving either from friends or family, and access to information that gives people the opportunity to act collectively.

Changes in the components of “livelihood”– capabilities, assets and activities, play a key role in the adaptation of people to the stress and shocks that they face (Chambers and Conway, 1992, p. 7). Scoones (1998) states that “identifying what livelihood resources (or combinations of “capitals”) are required for different livelihood strategy combinations is a key step in the process of analysis” (p. 9). At the household level, the combination of activities and assets that is adopted to deal with these internal or external shocks is often mentioned as “livelihood strategy”.

According to Ellis (2000), a livelihood strategy incorporates not only activities that generate income but also other kinds of elements such as cultural and social choices. The “cultural asset” could be incorporated in the framework and include for instance “beliefs, traditions, language, identity, festivals and sacred rites’’ (Adato & Meinzen-Dick, 2002, p. 26). Borrowing from Ellis, he defines livelihood diversification as the process by which rural families construct a diverse portfolio of activities and social support capabilities in their struggle for survival and in order to improve their standards of living” (Ellis, 1998, p. 4).

The combination of assets (social, human, physical, financial and natural) among individuals and households varies through the years and therefore livelihood trajectories diversify. Many scholars have explored how the vulnerability of livelihoods could be reduced through combinations of capitals (Adger, 2003; Yohe and Tol, 2002; Ziervogel et al., 2006). For instance, changes in weather patterns are identified throughout long periods of time and they can be accompanied by major shifts in the combinations of assets that people adopt. Seasonality is an intrinsic feature of livelihoods diversification. That is the reason why a historical approach must be taken into account whenever the analysis is based on strategies that people adopt to cope with such shocks. Up to this point, the sustainable rural livelihoods approach explained how livelihood security can be achieved through access to a range of livelihood assets. These could be natural, social, financial, physical, and human capitals, which in turn enable individuals and households to pursue successful strategies.

(16)

16 | P a g e

To understand the complexity of the processes through which the livelihoods are constructed and diversified, institutions and processes that influence the access to and control over assets must be taken into account. To upscale interventions between the local and national levels, further analysis of institutions, social processes and structures is necessary. According to Eyhorn (2007), this stage constitutes the meso level which “determines access to assets, livelihood strategies and decision-making bodies, terms of exchange among different types of capital, and returns to any given livelihood strategy” (p. 45). The institutional landscape in which transformations take place can be comprised of both formal and informal institutions and organizations while the level at which they operate ranges from the household to the community and from the national to the global level.

2.2 Social Capital, a Lens for Institutional Landscape

To analyze the institutional landscape, this study emphasizes the analysis of social capital, limited this time to participation in social groups, and access to and use of information. The opportunity to access information is more likely to happen within environments that can facilitate the exchange of information (Thuo et al., 2013). The dissemination of information depends on social relations among members (Sorenson et al., 2006) or the social integration of the members (Fritsch & Kauffeld-Monz, 2010). Social connections and social networks constitute part of the social capital of poor farming communities in the absence of formal institutions, assisting farmers with identifying and locating opportunities and resources (Kasarjyan et al., 2007; Thuo et al., 2013). There is data on the importance of weather information for different groups of farmers in Sub-Saharan Africa. In Zimbabwe, the majority of surveyed communal farmers who had information on the 1997/98 seasonal forecast had indicated plans to alter the planting date and the cultivated crop as well as the cultivated area (Phillips et al., 2001). In the Machakos District of Kenya, most of the farmers participating in the research acknowledged that they had adopted management recommendations regarding their cultivations based on weather forecasts (Ngugi, 2002). Therefore, understanding how small-holder farmers perceive weather anomalies and how they obtain and use weather information would inform the weather community and policy makers on how to convey more effectively the uncertainty of the

(17)

17 | P a g e

weather forecasts (Morss et al., 2008; Lazo et al., 2009) to meet the needs of those who are most dependent on weather.

The aspect of institutional processes and organizational issues in the approach of sustainable rural livelihoods is interlinked with the concept of freedoms that Sen (2000) introduced in his book “Development as Freedom.” In his book, he explains that people should have the right and the opportunity to evaluate and improve the choices they have. He argues that those freedoms (rights and opportunities) and therefore social justice are essential in order for people to reinforce their capabilities. The notion of capability is explained as “the various combinations of functionings (beings and doings) that a person can achieve” (Sen, 2000, p. 151). A poor set of initial endowments and entitlements combined with a lack of capabilities means that individuals and households are unable to secure their well-being. Such entitlements are state resources that can be accessed by those who have sufficient capability; it is the notion of capabilities that can therefore clarify the uneven access to natural resources.

People cannot become better off if even one of the elements necessary to deal with climate variability is lacking and therefore their well-being seems impossible. Well-being and capabilities are two interrelated elements initially incorporated in the first part of the definition of “sustainable livelihoods” (Chambers and Conway, 1992; Scoones, 1998).

2.3 The Concept of Vulnerability and Vulnerability Assessment

The concept of vulnerability is incorporated in the analysis of multiple research contexts such as sustainability science, climate impacts and adaptation, development and poverty, famine and secure livelihoods (Fussel, 2007). For the Sustainable Livelihoods Approach (SLA), diversifications in the different types of capital introduce the concept of vulnerability (Morse, McNamara, & Acholo, 2009). This is the context within which the sustainable livelihoods (Chambers and Conway, 1992; Scoones, 1998) are framed. Issues of “seasonality,” “trends,” and “shocks” describe the vulnerability of livelihoods. The ability “to cope with and recover from stresses and shocks,” adds the sustainability dimension to the definition of resilience.

Many scholars have analyzed the relationship between livelihood adaptation, vulnerability and resilience. Evidence can be found in Liverman (1990), who noted that vulnerability has been

(18)

18 | P a g e

related or equated to concepts such as resilience, marginality, susceptibility, adaptability, fragility, and risk. In relation to vulnerability, Walker et al. (2006) defined resilience as “the capacity of a system to absorb disturbance and reorganize while undergoing change so as to still retain essentially the same function, structure, identity, and feedbacks” (p. 4). Marschke and Berkes (2006) went a step further by stating that “resilience offers a lens with which to explore stresses and shocks and to understand livelihood dynamics” while “it is used to characterize a system’s ability to deal with change” (p. 2). The latter view sees resilience as part of the social realm and depicts the “social resilience” of Adger (2000), who defined resilience as the ability of groups or communities to cope with external stresses and disturbances as a result of social, political, and environmental change. Thus, the concept of resilience is important to understanding the capacity to adapt to climate variability.

The conceptualization of vulnerability that best describes all these aforementioned explanations is the one of the IPCC. According to McCarthy et al. (2001), vulnerability is defined as “the degree to which a system is susceptible to, or unable to cope with, adverse effects of climate change, including climate variability and extremes. Vulnerability is a function of the character, magnitude, and rate of climate variation to which a system is exposed, its sensitivity, and its adaptive capacity”. The dimensions of vulnerability are accordingly defined by McCarthy et al. (2001) as:

 Exposure, referring to the character, the magnitude and the rates of future climate change a system is or will be facing.

 Sensitivity, referring to the degree to which a system is affected by climate stimuli.  Adaptive capacity, which is the potential, capability or ability of a system to adapt to

climate change stimuli or their effects or impacts.

In a livelihood system the adaptive capacity can be conceptualized as the five capitals of sustainable livelihoods (natural, physical, human, social, financial) that people must have access to in order to pursue successful livelihood strategies. For instance, phenomena such as droughts influence water and soil stocks (resulting in, for example, failed crop harvests and loss of livestock) and therefore affect the financial capital of the local populations who depend on agricultural activities. Floods also affect natural resources but their impacts on physical capital are more obvious. Extensive damage to buildings and infrastructure such as road washouts,

(19)

19 | P a g e

damage to and destruction of irrigation, telecommunication and electricity systems are some of the most distinctive. These examples show that different types of climate shocks affect different types of assets. These result in different responses and strategies that enable people to avoid, cope with and resist such shocks (Payne and Lipton, 1994; Schubert, 2005). It is important to mention that a historical perspective regarding the responses and strategies should be incorporated in the assessment of vulnerability. The knowledge of the past can constitute a basis for coping with (short term) or adapting to (long term) climate variability on behalf of rural poor. According to the literature, vulnerability is conceptualized in different ways by different scholars, based always on their scientific field, multiple methodologies and goals of research. Therefore, there are many methodological and conceptual approaches to assessing this concept and guiding the analysis of vulnerability within a specific research area (Deressa et al., 2008). Analysis techniques range from GIS and mapping (Wilhelmi & Wilhite, 2002) and complicated mathematical models (Brooks et al., 2005; Deressa et al., 2008; Fontaine and Steinemann, 2009; Slejko et al., 2009) to cluster analysis (Antwi-Agyei et al, 2012; Poff et al., 2010) and vulnerability indexes (Adger, 1999; Patnaik and Narayanan, 2009; Hahn et al., 2009). The latter is the approach employed in this study.

Indicator-based vulnerability assessments have been conducted at various levels, both national and district (Ericksen et al., 2011; Hahn et al., 2009). The indicators that have been used in many of these studies are subjective, deriving from the literature (Ericksen et al., 2011) without taking into account the knowledge and perceptions of local communities regarding climate variability (Thomas, 2007). According to Eakin and Bojorquez-Tapia (2008), vulnerability assessments based on census data at a national level may not identify the level of vulnerability of rural areas at the local level in terms of access to assets.

This study aims to analyze the concepts of sustainable rural livelihoods and vulnerability at the village level and furthermore to identify the factors that cause vulnerability, using information about social dynamics that usually are absent in national and regional assessments. To address this lack, this study aims to expand scientists’ perception about coping strategies and adaptation of local communities to climate variability and change, by providing valuable insights into the drivers and structure of vulnerability (see e.g., Eakin and Bojorquez-Tapia, 2008; Antwi-Agyei et al, 2012).

(20)

20 | P a g e 2.3 Conclusion

Despite the various definitions and conceptualizations of vulnerability, this study blends a theoretical approach (sustainable rural livelihoods) with a practical approach (vulnerability assessment using indicators) to elucidate the dimensions of vulnerability at the village level in the District of Samia in Western Kenya. Moreover, it highlights the need to account for variability in social processes influencing vulnerability to climate change through the analysis of social capital in the small-holder farmers’ perceptions. Interventions based on grounding policies that take into account the social processes and outcomes at the local level are needed (Ellis, 2000; Scoones et al., 2005) to prevent livelihood insecurity in rural areas. The empirical data that will be analyzed through the lens of sustainable rural livelihoods could inform policy makers and practitioners to prioritize limited livelihood resources in designing interventions aimed at reducing vulnerability and increasing the adaptive capacity of rural poor.

(21)

21 | P a g e

CHAPTER 3 - Research Design

This chapter will describe the epistemological position that underlies the methodological planning, operationalization and analysis of this research. Following from the tenets that support the epistemological positioning, the chapter will continue with the presentation of the main research question and its subsidiary research questions, the conceptual scheme and the concept operationalization. The methods that were used to accomplish the inductive livelihood research will be further discussed while the tools used for analysis are mentioned in the following section. The chapter ends with a section detailing the sampling process, continuing with the process of gaining access to the study area, which is then followed by the ethical challenges met during fieldwork in both sub-locations of Bujwang’a and Luchululo within Samia District in western Kenya.

3.1 Epistemological Positioning and Research Methodology

The epistemological positioning that guided the planning of this study is subjectivism along with several characteristics of critical realism. The study aimed to highlight the variability and complexity of factors that make small-holder farmers vulnerable to climate variability and hinder their capacity to adapt. Therefore, it focuses on the empirical to capture the experience of the agents and their perceptions regarding climate variability, on the actual to assess the factors that enhance agents’ vulnerability and record the strategies that were used to deal with the stressor of climate, and the real to get an overview of the social dynamics that influence how the small-holder farmers’ adapt to climate variability. A bottom-up approach guides this explorative study. The framework of sustainable livelihoods integrated within the IPCC definition of vulnerability provides the basis for understanding how adaptive capacity can be strengthened through livelihood strategies to enable people to cope with climate variability and change, and increase resilience. Given that the Sustainable Livelihoods Framework (SLF) is formulated around a households’ five capitals (human, natural, social, physical and financial), it can be used to

(22)

22 | P a g e

identify ways in which these capitals can be exchanged or used either to enhance the adaptive capacity of local households under stress or to reduce their sensitivity. Therefore integrating current empirical data on the five capitals that make small-holder farmers vulnerable to climate variability may allow policy makers and development practitioners to come up with a combination of available capitals that will increase context-specific adaptation strategies.

The subjectivist approach that guided the research enabled me to acquire a bottom-up and comprehensive understanding of the different dimensions that underlie small-holder farmers’ vulnerability, especially in terms of sensitivity to risks and shocks and adaptive capacity. Moreover, the critical realist approach focuses mainly on two factors, that of agency and structure (Giddens, 1984), enabling the researcher to “identify the structures at work that generate those events and discourses ... These structures are not spontaneously apparent in the observable pattern of events; they can only be identified through the practical and theoretical work of the social sciences.” (Bhaskar,1989). Through critical realism, I was able to unveil the underlying social aspects that influence the perceptions and behavior of these farmers (May, 2001), using both practical and theoretical approaches. Multiple methods were employed to conduct the data collection and analysis.

3.2 Research Questions

Following from the tenets of subjectivism, the research questions that were formed in order to accomplish this study were adapted throughout the fieldwork. At first, the target of the study was to explore how climate variability affects rural livelihoods of small-holder farmers with regard to water management in Samia District. Once in the study area, it became evident that data needed to support the study was not available and other preliminary vulnerabilities needed to be primarily assessed. In addition, the concept of water management was not widespread among the locals. It became also apparent that in order to inform downscaled adaptation policies, the study had to capture the social dynamics that either hinder or encourage adaptation to climate variability.

(23)

23 | P a g e

3.2.1 Main research question

The preliminary research question that guided this research is the following:

 How has climate variability affected the livelihood strategies of vulnerable small-holder farmers over the recent five-year period in Samia in western Kenya?

3.2.2 Research sub-questions

To answer the main research question, three subsidiary research questions were deployed. These questions are:

 How is vulnerability to climate variability manifested among small-holder farmers in Samia?

 How do small-holder farmers perceive climate variability and how has the weather changed over a recent five year period?

 How do small-holder farmers cope with climate variability in Samia district?

3.3 Conceptual Framework

To measure vulnerability to climate variability, the study begins with the definition of the concept of vulnerability that was given by IPCC (2001). Therefore, vulnerability to climate variability and change is expressed as “a function of the character, magnitude and rate of climate variation to which a system is exposed, its sensitivity and its adaptive capacity” (McCarthy et al., 2001, p. 995). Following from the tenets of the Sustainable Rural Livelihoods Framework, during focus group discussions and from a household survey, five capital assets were ascertained using various indicators (social, economic, human, physical and natural). In the focus group discussions (FGDs), participants were asked to identify indicators linked to each form of capital asset-livelihood resource (human, financial, natural, physical and social capitals). This preliminary identification gave input to the household questionnaire, which collected complementary information on assets such as access to remittances and size of land owned.

(24)

24 | P a g e

3.3.1 Vulnerability index

A vulnerability index was constructed to quantify the strengths and weaknesses of the existing livelihood system and the capacity of Samian communities to adopt coping strategies that will allow them to deal with climate variability impacts. This conceptual approach allows the study to embrace the socio-economic factors that contribute to climate vulnerability at the community level, thus providing information to policy makers and development practitioners useful for formulating geographically targeted interventions in the domain of agriculture.

The indicators that were selected for the construction of the vulnerability index are based on the five assets of the SLA integrated with the IPCC vulnerability definition (exposure, sensitivity, adaptive capacity). As it was not possible to assess the dimension of exposure because of lack of downscaled rainfall measurements at the village level, exposure is considered to be the same for all villages. To deal with this limitation of rainfall data, the study subscribes to the argument of Moss et al. (2002) that “the inclusion of ‘exposure’ in the definition of vulnerability appears to be problematic, since exposure by itself does not necessarily contribute to negative outcomes” (p. 2). Consequently, the livelihood vulnerability index depends on two major components: sensitivity and adaptive capacity.

The variables that were chosen in order to operationalize the concepts of adaptive capacity and sensitivity are presented within the categories of SLF capitals.

Adaptive capacity: Refers to the entire capabilities, resources and institutions of a country or region to implement effective adaptation measures. It is measured through the following dimensions:

 Financial Capital. Financial assets such as remittances, crop yield and ownership of livestock are considered vital for reducing vulnerability. Within the context of Samia, remittances are the result of the migration of household members to urban centers such as Kisumu and Nairobi. This financial aid is of great importance for family members living in rural areas, contradicting the finding of Vincent (2004) who argues that the high vulnerability of rural areas is a result of increasing urbanization. Crop yields and livestock ownership are included in the adaptive capacity of households as well. For example, high crop yields reduce food insecurity in times of food shortage and therefore

(25)

25 | P a g e

reduces vulnerability, while the ownership of livestock gives the household the opportunity to sell its property when crops fail.

 Human Capital. Human capital is constituted of two variables: completed education at a secondary level and household members working in the family’s farm. A well-educated population has the potential to reduce vulnerability through the diversification of its livelihood. Moreover, education provides the opportunity to increase the sources of information (from paper media to Internet) regarding agrometeorological practices, increasing the adaptive capacity of communities. As for the second aspect, human capital, availability of family labor is the bedrock of farming in Samia as it reduces the cost of production.

 Social capital. Social capital comprises one component within the vulnerability index, that of participation in social groups, expressing institutional importance and strength at the community level. A community with stronger institutional structure can adapt more easily to climate variability impacts. Moreover, access to and use of information is of importance to the adaptive capacity of households. All households were found to have access to weather information; therefore this variable was neutral for the index. This aspect of social capital is analyzed in Chapter 6, explaining the factors that influence the perceptions of small-holder farmers about climate variability.

Sensitivity: Refers to the degree to which a system is affected by climate stimuli. Therefore, the capitals that were identified to express this dimension are:

 Natural capital. Natural capital consists of two indicators: size of cultivated land owned and access to natural water available for agriculture. It is assumed that the larger the size of a farm holding, the larger the yield or the variety of crops that are cultivated, and therefore the less its vulnerability to climate variability. The survey also considered size of farms owned from a psychological perspective; ownership of land has indirect effects on environmental sustainability and crop productivity, affecting the level of a household’s vulnerability (Antwi-Agyei et al., 2012), and adding to its sensitivity. Access to a source of natural water for agriculture is assumed to reduce the sensitivity of the household as this access can increase the productivity of the farm.

(26)

26 | P a g e

 Physical capital. Physical capital comprises two indicators as well: access to irrigation facilities and ownership of media access. Irrigation facilities have the same effect on sensitivity as access to a natural water source. Farms accessing water through irrigation facilities show an increase in their yields when using regulated irrigation systems. As for ownership of media access, this also reduces sensitivity, and hence vulnerability, as small-holder farmers then have the opportunity to be informed about weather forecasts that may affect how they plan their agricultural activities.

Along with the fact that the concept of exposure is not integrated in the vulnerability index because of fieldwork limitations, it should be noted that figures used for rainfall were based on rainfall data at the district level over a period of five years, 2008-2012.

3.3.2 Index methodology

To calculate the vulnerability index, we used the method of equal weights. That is, all variables have equal weight despite the different number of components that constitute them (see Table 1), while the formula uses the average index of each source of vulnerability. Having selected variables that are measured on different scales, all indicators were first standardized according to the UNDP (2007) procedure of normalizing indicators for the life expectancy index. The formula used is:

Standardized index value = (Observed value – minimum value) / (Maximum value – minimum value)

The standardized indicators took values ranging from “0 to 1 (Appendix 1).” Having standardized the indicators and having expressed the functional relationship of each and every variable to vulnerability (Table 1), the vulnerability index is calculated according to the following formulation of Patnaik and Narayanan, (2005):

(27)

27 | P a g e

where n is the number of sources of vulnerability and AI is the average index for each of the sources of vulnerability (in this case n=α=2). The scale of the index ranges between 0 (less vulnerable) to 1 (more vulnerable).

Table 1: Operationalization of vulnerability index and the functional relationship of variables to vulnerability.

Concept Dimensions Variables Indicators

Functional relationship with vulnerability

V

u

ln

era

b

il

it

y

Sensitivity Natural Capital

Size of farm holding expressed in acres of land

Percentage of households that have no access to natural water resource

Physical Capital

Percentage of households that have access to irrigation facilities

Percentage of households that own any type of media (radio, newspaper, etc.) Adaptive capacity Financial Capital

Percentage of households that receive remittances

Average crop yield in bags (measured for both raining seasons)

Percentage of households that own livestock

Social Capital

Percentage of households that have at least one member participating in a social group (either formal or informal)

Human Capital

Percentage of households that have at least one member who has completed secondary education

Average of household members working on the family farm

(28)

28 | P a g e

3.3.3 Conceptual scheme

Figure 1 illustrates the connection of the concepts already presented. It shows that vulnerability to climate variability is expressed through two basic types of factors, climatic and non-climatic. On one hand, the climatic factors, which in our study take the form of climate variability, focused on how rainfall exposes the community to certain impacts that increase their vulnerability. On the other hand, non-climatic factors consisting of indicators within the five capitals of the sustainable livelihoods approach contribute to the sensitivity and adaptive capacity of local communities. Adaptation depends not only on the adaptive capacity of the community but also on the perceptions of the agents, in our case small-holder farmers. The effect of adaptive capacity is expressed by a solid arrow while perceptions are visualized by the dashed one.

N=Natural capital, P=Physical capital, F=Financial capital, H=Human capital, S= Social capital

Figure 1. Integrated vulnerability context associated with climate variability (adapted from Fussel & Klein, 2006)

(29)

29 | P a g e

3.4 Mixed Methods

Several methods were used both in alternation and in combination during fieldwork in order to collect, complement, triangulate and analyze data. These consist of observations and informal conversations, five participatory events, semi-structured interviews with small-holder farmers and key informants, a household survey and a vulnerability index. When dealing with a complex phenomenon like climate variability and change, using a mixed–methods approach enables the researcher to gain a better understanding of the various dimensions of the problem (Adger et al., 2009).

3.4.1 Household survey

A household survey tool was employed to obtain descriptive information about the sample population. This tool enabled the study to acquire a quantified breakdown of rural livelihoods with reference to agriculture, access to meteorological information, perceptions and coping strategies regarding climate variability, and value positions in relation to social dynamics, agrometeorological practices and meteorological information. The household survey was conducted in all six villages within the sub-locations of Luchululo and Bujwang’a and for this purpose a team of five research assistants was engaged in the research. Initially, a team of nine assistants was trained in a full-time three-day workshop where all research targets were thoroughly explained in English and Kisamian. During the training workshop, each participant was asked to give a presentation of both participatory tools and questionnaires. Individual follow-up of these presentations on the third day of the workshop led to the elimination of 4 participants, resulting in the team of five. In 9 pilot surveys, a total of 130 sub-surveys were conducted of which 10 were invalid and 120 (n=120) were used in the analysis of the data.

3.4.2 Observations and informal conversations

To better understand the context within which the research would take place and gain insight into the accuracy of my interpretation of the collected data, it was important that I observe the local communities. Becoming familiar with the locals provided me the opportunity to be involved in

(30)

30 | P a g e

events and activities that ordinarily I would not be invited to attend. Because I was present during the long rains period, I was able to observe first hand when and how such agricultural activities as preparing land and planting crops took place, which gave me insight into the labor division within those agricultural activities, and acquainted me with the daily routine of the locals. Moreover, the method of observing as a participant allowed me to become well known to the community opinion leaders, who are highly important nodes of the community structure as they are engaged in the flow of agrometeorological information at the community level. Lastly, having a better overview of where villages are located, how much time it takes to commute and when and where it is better to have an interview with each gender enabled me to develop more targeted methods.

3.4.3 Participatory research methods

Participatory research methods were employed in order to actively engage community members and multiple stakeholders in discussions regarding vulnerability, climate variability, coping strategies and the social dynamics that influence the aforementioned concepts. The aim of these methods was not only to extract information but also to give participants the opportunity to learn from each other and share knowledge (especially about strategies for coping with climate variability), to analyze for themselves the relation between livelihood resources and weather fluctuations, and to provide their validation of the preliminary research findings. The participatory methods were employed at both the beginning and the end of fieldwork, allowing vulnerabilities to be identified and also providing an opportunity to cross-check the data collected. The methods used involved: Two (2) community mappings, six (6) focus group discussions, six (6) seasonal calendars, six (6) vulnerability matrixes, and six (6) community timelines.

 Community Mapping.

At the beginning of the fieldwork, community mapping was a useful tool for identifying the boundaries of each sub-location, for orienting myself to the area and for obtaining a geographical overview of the villages. In the process of this mapping, which was conducted with the

(31)

31 | P a g e

assistance of village elders, a sub-chief, an interim-sub-chief and two opinion leaders, further discussion revealed opportunities to reduce exposure to climate variability. The greatest opportunity for increased agricultural productivity, the reduction of poverty and food insecurity proved to be irrigation. Therefore, the main criterion for selecting villages for the research was their access to water for agriculture. During mapping, there was opportunity to ask about existing community groups (formal and informal), community services that support the most vulnerable groups and recommendations for further improvement of rural livelihoods.

 Six (6) Focus Group Discussions (FGDs)

Focus group discussions were one of the basic participatory research tools used to generate information about the perceptions of farmers regarding climate variability. Six (6) FGDs were conducted, one in each village, with an average of 15 participants, male and female, in each group. No separate FGDs were held with only men or only women as an FGD was an opportunity to emphasize interactions within the groups. Short interviews were held after the focus groups with different genders to identify the social restrictions of participating in a mixed-gender focus group. Age diversification (18-78) was intentionally employed, as one of the complementary research tools to the focus group discussions was asking the participants to recall historical trends regarding climate, agricultural development and also all events that influenced their livelihoods and their well-being as a community. Themes that were discussed include: climate variability and gender specific impacts, coping strategies and obstacles to adopting efficient coping strategies, gender division of labor and its cultural implications, and food insecurity with emphasis on weather and cultural influence.

 Six (6) Vulnerability Matrixes

Vulnerability matrixes were employed to make it easier for all participants to measure the impact of shock arising from climate variability. On a scale from 0-3, (0→no impact, 1→low impact, 2→medium impact, 3→significant impact), the participants had to evaluate the effect of the shock to their livelihood resources, addressing the five capitals (natural, physical, financial,

(32)

32 | P a g e

social and human capital) of the SRL approach. The relevant measures were determined with the help of my local supervisor before the focus group discussions were held. For example, impact on education was measured by lost days of schooling. Whenever an expected livelihood resource was mentioned by the group, further explanation was encouraged, trying to avoid any imposition on their answers and thoughts. When coming to agreement on the score to be given a particular item, participants were always asked to elaborate on their answers, triggering debate among them. The conversation continued on to the coping strategies that they tend to adopt to deal with these climate variability impacts, emphasizing the importance of weather information. Lastly, participants were asked to rank their coping strategies in terms of efficiency, which would facilitate later analysis of these strategies. The overview of vulnerability matrixes was discussed and complemented by a teacher, a Palweco community facilitator and an e-agricultural extension officer.

 Six (6) Seasonal Calendars

Seasonal calendars were used to obtain a schedule of activities that are carried out throughout the year and to visualize the weather patterns affecting the villages. This tool helped to stimulate discussion on changes in livelihood activities over a recent five-year period, changes in coping strategies regarding climate variability, gender division of labor and changes within seasons, always in relation to weather. For example, participants were asked to answer questions such as: Have seasonal patterns shifted in the last five years? What are the coping strategies of locals during stressful times in terms of drought, erratic rainfall, floods, and food shortage? Lastly, they were asked to assess how climate information affects and/or informs the planning of their socio-economic activities. The data collected were triangulated with multiple questions to all agricultural extension officers who participated not only as key informants to the study but also as extension informants through short interviews. Weather patterns were analyzed with a meteorologist based in Busia County.

(33)

33 | P a g e

 Six (6) community timelines: The River of Life

The River of Life tool was employed to trace the history of each community and create an intergenerational exchange of knowledge regarding negative weather experiences, challenges that those experiences brought to the community, key interventions in agricultural development that took place in the area, and actors that were involved in the development of the sector by introducing new projects, crops and facilities. In general, this tool enabled somehow the bonding of the community, as there was an extensive exchange of valuable information as youth learned from their elders while the latter were informed about new community initiatives and coping strategies. The River of Life enabled the study to target key informants with prominent roles in the development of the agricultural sector and to be more precise about questions regarding the efficiency of the projects that they have launched to diminish the vulnerability of local communities.

3.4.4 Semi-structured interviews with small-holder farmers and key informants

Semi-structured interviews with key informants and small-holder farmers allowed me to obtain in-depth information and details on how agrometeorological information is downscaled to farmers, how locals perceive climate variability, and what strategies small-holder farmers in Samia have been using to cope with climate variability.

 Fifteen (15) interviews with small-holder farmers were conducted, mostly after the household survey, in order to saturate data regarding weather changes, coping strategies and factors that hinder the adoption of several coping strategies influencing the agricultural planning of local farmers. The climate incidents that make farmers vulnerable were valued according to their impact on livelihood resources and ranked accordingly. Further questions were posed with regard to the sources of agrometeorological information that they obtain, such as the level of trust and the efficiency of these sources in meeting their needs. Reasoning on how participation in social groups makes them prosper was provided, and information was elicited regarding the importance of participating in social groups in responding to climate variability. The interviewees were purposively selected to obtain information regarding the importance of social capital

(34)

34 | P a g e

(eight (8) were participating in social groups and seven (7) were not). Comparison and triangulation was needed to analyze subjective positioning on perceptions.

 Fourteen (14) key informant interviews were conducted with purposively selected interviewees who had a role in the setting and possessed knowledge on the topic under research. All originated from different sectors, allowing the exploration of agrometeorological issues from varying perspectives, thereby avoiding biased results and findings. Nine (9) government officials (an irrigation officer, three agricultural extension officers from the Ministry of Agriculture (MoA), one of them involved in the Adaptation to Climate Change and Insurance (ACCI) project, the Program for Agriculture and Livelihoods in Western Communities in Kenya (PALWECO) community facilitator, a Kenya Agricultural Productivity and Agribusiness Project (KAPAP) service provider, a meteorologist at County Meteorological Services, the Busia County Director of the Environment, and the National ACCI project coordinator), as well as two community leaders, the chairman of the Samia Irrigation Fruits self-help group, and the coordinator of an agroforestry nursery project, contributed to the accomplishment of in-depth interviews. The key informant interviews provided information on the dissemination of weather information and the role of specific institutions, the multiple stressors that enhance the vulnerability of small-holder farmers, the coping strategies that are proposed from the top down and also information about the criticism of locals about the reliability and efficiency of weather channels.

3.4.5 Sampling and research population

The unit of analysis for this study was first the household and then the community in order to capture the vulnerabilities of small-holder farmers in Samia District and to identify the multiple stressors that create those vulnerabilities, with an entry point that is given through the stressor of climate variability as the entry point. The respondents participating in the research were methodically chosen based on a baseline survey (Odame et al, 2010). Quota sampling based on the division of the overall population by wealth categories was employed to identify the individuals and households (n= 120) that participated in the focus group discussions, interviews and household survey. Snowball sampling was used to approach several key informants while all

Climate Variability : an entry point to assess vulnerability of small-holder farmers in Samia District, in Western Kenya