Rainwater harvesting

A suitable poverty reduction strategy for small-scale farmers in developing countries?

Lisa Bunclark

A dissertation submitted to the School of International Development of the University of East Anglia in part-fulfilment of the requirements for the Degree of Master of Arts

November 2010

Table of Contents
Page 1 1 1 3 5 5 6 7 9 9 10 10 12 13 15 15 18 18 18 19 23 24 27 32 33

1. Introduction and Methods...... 1.1. Introduction.. 1.1.1. Background. 1.1.2. The case study: Botswana.. 1.2. Methods.... 1.2.1. Approach.... 1.2.2. Data collection and analysis.. 1.2.3. Limitations. 2. The problematisation of rainwater harvesting.... 2.1. The need for a new framework..... 2.2. Critical issues in RWH for small-scale agriculture.. 2.2.1. The challenges to crop production in marginal regions. 2.2.2. Understanding the priorities of small-scale farmers.. 2.2.3. The role of governance and institutions. 2.3. Beyond rainwater harvesting: Dynamic and complex systems 2.4. The situation in Botswana.... 3. Results and discussion 3.1. Results.. 3.1.1. Rainwater harvesting practices in Botswana. 3.1.2. Factors affecting adoption. 3.2. Discussion.... 3.2.1. The ability of rainwater harvesting to increase crop production and reduce poverty in Botswana.... 3.2.2. The suitability of rainwater harvesting: Towards a matrix for assessment 3.3. Summary.. 4. Conclusion

Appendix A Appendix B

Abstract

The first objective of this research is to determine the ability of rainwater harvesting (RWH) to increase rain-fed crop productivity and reduce poverty in Botswana. The second objective is to examine the factors that determine the suitability of RWH use in small-scale agriculture in developing countries as a whole. The overall aim is to produce a decision-making matrix that may be used in developing regions to assess the suitability of the technology for increasing crop production and reducing poverty in any given context. Primary data collected through the course of several interviews with key individuals are analysed via a process of coding, which leads to the categorisation and contextualisation of the range of factors that affect both the initial adoption and sustainable use of RWH systems by smallscale farmers. Together with findings gathered from secondary data, results from the primary data analysis are used to develop a decision-making matrix.

The results of this study indicate that current potential for increases in crop production through the use of RWH in both Botswana and developing countries as a whole is uncertain; this is primarily due to impacts of climate change and alterations to rural livelihood strategies as a result of economic development. It is shown that the suitability of RWH for increasing crop production and reducing poverty in developing countries depends on factors related to climate and ecology, farming practice, availability of assets, livelihood strategy, governance and institutions. With appropriate adaptation of systems and the development of community level institutions to provide in-depth training to farmers potential may improve, but it is possible that the prevalence of pastoral farming may prove too great a barrier to be overcome in some areas. It is recommended that further research is conducted to refine and expand the proposed decision-making framework into a more comprehensive RWH implementation framework.

Acknowledgements

First and foremost I would like to thank my supervisor, Dr Bruce Lankford, for his guidance throughout the duration of this research and without whom my fieldwork in Botswana would not have been possible. Secondly, many thanks to the research participants for their time and agreeing to share their experiences with me. I would also like to thank the staff of the German Development Cooperation (GTZ) in Gaborone and all individuals who provided valuable assistance in various stages of this research project, both within the UK and Botswana. Thank you to family and friends for unwavering support and encouragement, as well as their help with proof-reading and editing. Finally, a special thanks to both the Jack Wright Memorial Trust and Engineers Without Borders UK, who provided much appreciated financial assistance that enabled the undertaking of this research.

1. Introduction and Methods

1.1.

Introduction

1.1.1.

Background

Agriculture represents the primary livelihood strategy for the majority of the rural population in the developing world (Baguma and Loiskandl, 2010) and plays a crucial role in economic development and poverty reduction (Balke, 2008). Dependence on small-scale farming is greatest in the semi-arid and dry sub-humid climatic regions (CA, 2007) where rainfall is generally low and may not be sufficient for crop basic needs (Oweis and Hachum, 2006). However, in many cases the main challenge to crop production is not the availability of an adequate volume of water in absolute terms, but the inappropriate distribution of this water (Nigi, 2009; Reij et al., 1988) and the short-term agricultural drought associated with this (Falkenmark and Rockstrm, 2008). It is suggested that the introduction of an approach to directly bridge these intra-seasonal dry spells may be the most appropriate way to improve agricultural systems in these regions by reducing the moisture deficit in the soil and thereby reducing the risk of crop failure (Rockstrm, 2003)

Debate concerning the most suitable strategy for achieving adequate crop water availability centres around increasing the efficiency of green water (water stored as soil moisture), rather than blue water (water stored in rivers and aquifers), through the use of rainwater harvesting (RWH) (see Falkenmark, 2007; Rockstrm, 2003; Hatibu and Mahoo, 1999). RWH is a collective term used to describe the process of rainfall runoff collection and storage for subsequent beneficial use (Barron, 2009; Khaka et al., 2005; Mati et al., 2006; Oweis and Hachum, 2006). The categorisation of different RWH methods varies greatly, but in general systems are grouped into two categories: in-situ (or micro catchment) and ex-situ (or macro catchment) approaches. In-situ RWH approaches encompass any system in which runoff is collected in close proximity to crop growing area and used to replenish soil moisture directly. Ex-situ methods involve the collection of rainfall runoff from large areas, which may not be in close proximity to the crop land, and storage in ponds, containers or 1

underground reservoirs, for use as supplemental irrigation when necessary (Khaka et al., 2005; Critchley and Siegert, 1991; Botha et al., 2008)

In the past two decades interest in RWH has grown steadily and several global and regional declarations have pronounced it to be a solution to the developing worlds growing water needs (Nijhof et al., 2010). However, evidence indicates that predicted improvements in agricultural production and reductions in levels of poverty as a result of RWH use have not been achieved (Reij et al., 1988; Hatibu et al., 2006). Countries are increasingly urged to integrate RWH into their water resources management strategies (UN, 2006) and the governments of many developing countries within both Asia and Africa now endorse the use of the technology (Nijhof et al., 2010). Within the small-scale agricultural sector it is claimed that with the use of RWH there is a potential to double current crop yields (NWP, 2007) and lift farmers out of the poverty trap (Barron, 2009; Vohland and Barry, 2009). Nonetheless, techniques have failed to be adopted on a wide-scale despite the undertaking of a significant number of pilot projects and research (Oweis and Hachum, 2006; Reij et al., 1988). Furthermore, in regions where RWH is a traditional method of agricultural water management it has been observed that systems have fallen into disrepair and been abandoned (Kumar et al., 2008).

RWH technologies should not be viewed as a panacea for small-scale agriculture and although some systems have been successfully used by small-holder farmers in parts of the developing world for thousands of years (Critchley and Sieigert, 1991; Mamdouh, 1999), the same technology may not prove effective in other areas. The suitability of any technique for agricultural use depends on a wide range of factors (AfDB 2007) and a technology must be accessible, affordable and appropriate for the target community (Coupe, 2001; Coventry, 2003) if it is to be successfully adopted and sustainably used. Existing research largely ignores the complex environment that RWH systems must fit into (Cullis and Pacey, 1992; Scoones et al., 2007; Vohland and Barry, 2009) and despite an acknowledgement of the importance of contextual factors on the suitability of RWH systems in recent years, implementation frameworks remain focused on the technical aspects of RWH only (see Hatibu and Mahoo, 1999; AfDB, 2007; Young et al., 2002).

Although researchers have successfully identified factors that influence the adoption and sustainable use of RWH (Botha et al., 2008; Kundhlande, et al., 2004; Baiphethi, et al., 2

2009), little is known about how these factors interact (Andersson et al., 2009). Key issues identified in previous research include climatic characteristics, availability of resources, livelihood strategy, policies and institutions, but what specific factors affect the adoption of RWH by small-scale farmers in the case study country of Botswana and how these can be categorised? Furthermore, what contextual issues shape these influential factors? With the knowledge of these issues, what conclusions can be drawn regarding the ability of RWH to reduce poverty in Botswana and the suitability of the technology for small-holder farmers in developing countries in general? The overall aim of this research is to produce a decisionmaking matrix that may be used by those considering the implementation of RWH schemes in developing regions.

This paper argues that the current approach of researchers and developmental organisations to the use of RWH in agriculture does not place enough emphasis on the context within which the systems are placed. In some cases RWH may not be able to increase crop production or reduce poverty and steps need to be taken to incorporate non-technical factors into implementation frameworks to ensure that the technology is only advocated where suitable. The remainder of this section explains the use of Botswana as a case study for this research and outlines the methodological approach used in the collection and analysis of both primary and secondary data. The second section summarises the findings from current literature regarding factors that affect the successful use of RWH in developing countries to reduce poverty and highlights the gaps present in past research. The third section presents and analyses the results from the primary data collected and proposes a decision-making matrix that can be used to aid those considering the implementation of RWH schemes. Tentative conclusions regarding the suitability of RWH for reducing the poverty of smallscale farmers in developing countries in general are drawn in the final section.

1.1.2.

The case study: Botswana

As shown in Figure 1.1, Botswana is a land locked country in southern Africa encompassing approximately 582,000 km2 (FAO, 2005). The nation has a population of about 1,950,000 (World Bank, 2010), over 80 per cent of which is concentrated to the east of the country (FAO, 2005). Since gaining Independence in 1966 Botswana has developed rapidly, but

despite the large increases in per capita income the country still faces serious challenges of poverty (GoB, 2009).

The need for productivity increases in smallscale farming to help improve incomes in rural areas is recognised as a key concern in efforts to alleviate poverty in Botswana (Acquah, 2003) and this is one of the reasons for the selection of the country as a case study for this research. Traditional crop production is characterised by low-input, low-management rain-fed systems and

Figure 1.1: Location of Botswana

primarily involves the cultivation of sorghum and millet (GoB, 2009) Despite attempts by the government to assist small-scale farmers through extension services traditional farms continue to perform poorly (GoB, 2006a) and productivity remains low at between 200-300 kilograms per hectare (CAR, 2007; FAO, 2005; Whiteside, 1998), which is only 30 per cent of potential yields (Rockstrm et al., 2010). High incidences of both short- and long-term drought represent significant barriers to crop production (CAR, 2007). RWH systems have been traditionally used in Botswana for many years for domestic purposes (Pacey and Cullis, 1991), but views regarding the suitability of the technology for agricultural production vary (c.f. FAO, 2003; Mati et al., 2006); the aim of this research is to reach a conclusion on this issue.
(Source: Batisani and Yarnal, 2009)

The concept of RWH in agriculture is not new to the Government of Botswana. Since 1970 the Water Development Section of the Ministry of Agriculture (MoA) has provided syndicates of farmers with small earth dams to harvest rainwater for cattle watering and a limited amount of supplemental irrigation (FAO, 2005), although it is reported that many of these dams have fallen into disrepair and are no longer operational (FAO, 2005). The Government of Botswana recognises that low productivity in the rain-fed agricultural sector is primarily due to poor management of soil moisture (GoB, 2006a) and has identified the benefits that could be obtained through the use of RWH in rain-fed agriculture (GoB, 1999). There have been several governmental schemes aimed at traditional arable farmers including 4

Accelerated Rain-fed Arable Programme (ARAP), Arable Lands Development Project (ALDEP), National Master Plan for Arable Agriculture and Dairy Development (NAMPAADD) and Integrated Support for Arable Agricultural Development (ISPAAD) (AfDB, 2008, CAR, 2007), of which ALDEP was the largest scheme responsible for rainwater tank construction (GoB/UNDP, 2004). However, the impact of these schemes on increasing production appears to be relatively modest and the uptake of the technology has been low (Acquah, 2003). Despite heavy subsidies and assistance from the government water harvesting packages adopted by farmers through ALDEP comprised less than 2 per cent of all packages distributed (GoB, 2006a). Little examination of these schemes has been conducted and one of the objectives of this research is to determine the range of factors that led to such low levels of adoption of RWH systems by small-scale farmers in Botswana in the presence of what appears to be an enabling environment. The analysis of these results will provide key additional insights into factors that affect the suitability of RWH for agriculture in other developing regions.

1.2.

Methods

1.2.1.

Approach

The findings of this research are drawn from both primary and secondary qualitative data. The secondary data provide a comprehensive summary of past research and experience in the field of RWH within Botswana, the southern African region and the developing world as a whole. The primary data collected from a series of interviews conducted during fieldwork, provides in-depth information on the context of RWH within Botswana specifically. Interviews were chosen as the principal method of primary data collection as they provide rich qualitative data related to experiences, opinions and values (May, 2001) and this is the type of data needed to answer the research questions posed. A semi-structured interview in particular was used as this type of interview allows for a certain degree of flexibility and the ability to modify the structure of the discussion based upon what the researcher perceives as suitable at the time (Robson, 2002). It is via this combination of generalised secondary research and more focussed primary research that key concepts have been identified and analysed.

Grounded theory, as described by Strauss and Corbin (1998), was used as the basis for this research and data collection and analysis was driven by concepts identified in existing theories. Following the process of theoretical sampling, as explained by Bryman and Burgess (1994), the preliminary collection of secondary data was used to generate categories that shaped the interview guide and the collection of primary data. These interviews led to the identification of further themes via a process of coding and additional secondary data was collected where necessary until all themes were fully explored and considered saturated (Strauss and Corbin, 1998). This particular approach is appropriate to the aim of this research due to the ability it has to generate concepts and categories (Bryman, 2004), which is needed for the formulation of the decision-making matrix.

1.2.2.

Data collection and analysis

Secondary data were gathered from a range of reliable sources, including books, peerreviewed journals and material published by relevant organisations and research institutions, such as the Government of Botswana (GoB) and the Food and Agriculture Organisation of the United Nations (FAO). The material was located through searches of library catalogues, academic databases and general internet search engines; some material was also acquired through personal contacts. This data contributed to the formulation of an interview guide (see Appendix B) that increased the level of consistency between each interview, which acted to maintain a certain degree of objectivity and quality of data obtained (Robson, 2002).

Since the identification of those with knowledge or experience of the use of RWH in smallscale farming was problematic participants were identified through a process of snowball sampling. According to Bryman (2004) snowball sampling comprises the identification of an initial set of participants relevant to the research topic, who are then used to establish contact with other relevant groups or individuals. This sampling method is particularly useful for situations where no specific sampling frame exists (Bryman, 2004; Faugier and Sargeant, 1997) and takes advantage of social networks of participants to expand the sample size. For the purposes of this study, an initial list of potential participants was compiled with the use secondary data and these individuals were then used to establish contacts with other relevant organisations and individuals with knowledge and experience of RWH in small-scale agriculture, whether through research and policy making, or via first-hand experience in the field. A total of twelve individuals were interviewed, representing ministries, non6

governmental organisations (NGOs) and institutions working at national and local levels in Botswana; the traditional Batswana farming community; and two individuals from academic institutions in South Africa.

In order to ensure the interviews met with ethical approval, each potential participant was read and given a copy of the pre-prepared Research Participation Information Form. Informed consent was obtained from all participants and a list of those interviewed was recorded through the course of the fieldwork. To ensure informed consent was provided by each participant, interviewees were made aware that they may refuse to answer any questions they choose and can withdraw from the research project at any time. Finally, participant confidentially and anonymity was maintained throughout the course of the research in order to prevent any adverse effects that may occur due to their involvement in the project.

Although recordings of the interviews were not carried out, extensive notes were taken and subsequently word processed to increase the ease of analysis. Copies of the notes taken during the interviews can be found in Appendix B; as mentioned previously, the interviewees have been anonymised in order to protect their identity. Initially all notes were read several times to increase familiarisation with the data and allow preliminary coding, the software package Nvivo was later used to conduct more extensive and systematic coding using word-based techniques. As mentioned above, themes were partially pre-determined by the concepts identified in the secondary data, although additional categories were determined through the course of the coding process. Word repetition and the analysis of the context within which key words were used, as outlined by Ryan and Bernard (2003), led to the identification of additional themes as text was sorted into groups with similar meaning.

1.2.3.

Limitations

Although efforts have been made to ensure the validity of this research, some limitations are still identifiable. Firstly, it is possible that insufficient triangulation of data obtained from the NGOs, government ministries and research institutes with that of traditional farmers may reduce the credibility of the findings (Bryman, 2004). Furthermore, the small size of the total sample may not be representative of the population and may limit the transferability of the conclusion drawn from this research, although it is anticipated that the depth of data obtained 7

will compensate for this to a certain extent (Bryman, ibid). Lastly, problems with reliability and accuracy of data may also be increased due to the use of interviews as the method of primary data collection. Although efforts were made to ensure the positionality of the interviewer were minimised, it is never possible to guarantee neutrality in an interview (Rapley 2004). Data may also have been affected in the cases where more than one participant was interviewed simultaneously, as the presence of other individuals may have affected the participants response to questions posed (Chacko, 2004).

2.

The problematisation of rainwater harvesting

In this section the research topic is explored through existing literature and main theories are presented and examined. The importance of the context within which rainwater harvesting (RWH) systems must operate is discussed and the need for an updated framework for the implementation of the technology is argued. Gaps in knowledge regarding the suitability of rainwater harvesting (RWH) for improving production in small-scale agriculture and reducing poverty among farmers are highlighted. Key variables and factors that affect the appropriateness of RWH are presented and the complex nature of the problem of low crop yields is examined. Evidence is presented from the case study of Botswana and questions regarding the suitability of the technology for reducing poverty within the country are posed.

2.1

The need for a new framework

We need to offer the poor real technology choice over affordable, appropriate and accessible options. It is not hi-tech or low-tech but right tech. (Coventry 2003:1) RWH technologies have been traditionally used by farmers in many marginal regions of the developing world and so are undeniably relevant, but the introduction of new and inappropriate methods must be avoided (Pacey and Cullis, 1991). Within the field of development, there is a tendency to assume that successful technologies in one region can be transferred easily to another (Scoones et al., 2007) and that RWH systems produced by research are likely to show equally promising results in the field (Rling, 2009). The dynamic context within which these technologies must fit is often ignored (Cullis and Pacey, 1992; Scoones et al., 2007; Vohland and Barry, 2009) and fundamental factors which contribute to the success or failure of a scheme are addressed inadequately (Parr and Shaw, 1999). Although researchers have successfully identified factors that affect RWH (Botha et al., 2008; Kundhlande et al., 2004; Baiphethi et al., 2009), little is known about the interactions between these factors (Andersson et al., 2009), or how they affect adoption trends at household level (Baguma and Loiskandl, 2010). Issues regarding the sustainability of the technology are also often not addressed (Kundhlande et al., 2004). 9

Despite the acknowledgement of the importance of contextual factors on the suitability of RWH schemes, there is no evidence of their incorporation into models or frameworks that assess the use of the technology in agriculture (see Young et al., 2002; AfDB, 2007; Hatibu and Mahoo, 1999). Existing models and frameworks, as shown in Appendix A, focus on the technical aspects of RWH only and this has led to problems with many projects implemented (Hatibu and Mahoo, 1999). It is argued that there is a need for the compilation of a much wider range of factors into a decision-making matrix that can be used by those working with RWH in agriculture. RWH is not a silver bullet (Barron, 2009) and a combination of interconnected approaches will be necessary to solve a problem as complex as low crop production and high poverty (Pauli and Bjerregaard, 1999).

2.2

Critical issues in RWH for small-scale agriculture

2.2.1 The challenges to crop production in marginal regions

RWH is thought to be particularly suited to the application of supplemental irrigation in arid and semi-arid areas where the rainfall level is low, highly variable, often dispersed over a small number of high intensity events and yield losses high due to moisture stress (Barron, 2009; Nigi et al., 2007; Reij et al., 1988). The optimum rainfall level for the successful use of RWH in agriculture falls within the range of 200mm-1200mm (Pacey and Cullis 1991; Mati et al., 2006), although it is primarily the intensity of rainfall, rather than volume, that determines the degree of runoff and quantity of water harvested (Critchley and Siegert, 1991; Pachpute, et al., 2009). In these regions variations in rainfall may range between 33 per cent to 200 per cent of the long term average (Stewart 1988 in Rockstrm et al 2002) and in the most arid areas it is not unlikely that no rain will fall for several consecutive years, with an upper variation limit as high as 350 per cent of the long term average (Critchley and Siegert, 1991). This variability in rainfall presents the greatest challenge to crop production (Barron, 2009; Nigi, 2009) and there is an opinion that in some areas rainfall is simply too erratic for RWH to sustain crop yields successfully (Reij et al., 1988). Furthermore, rainfall is often not distributed in line with the crop-growing seasons (Balke,, 2008; Oweis and Hachum, 2006) and RWH can only prove effective if sufficient amounts of runoff can be harvested and stored outside of the growing period (Rost et al., 2009). In consideration of this, it is possible 10

that the use of RWH to bridge intra-seasonal dry spells may be limited in areas where assistance is needed most as ultimately crop water supply remains governed by the reliability of rainfall (Kumar et al., 2008).

The majority of small scale farmers in developing countries are located in areas with less than ideal conditions for growing crops, where low and erratic rainfall is coupled with poor soils (Kundhlande et al,. 2004) and high evaporation rates (Kumar et al., 2008). Soil fertility is generally the second most limiting factor on crop production after water scarcity (Critchley and Siegert, 1991) and a certain amount of organic matter needs to be present in soils to produce satisfactory crop yields (Balke, 2008). The implementation of RWH may be fruitless in regions where small holder farmers work with infertile soils and either lack the financial resources to use fertilisation methods (Hatibu et al., 2006), or do not perceive it worth investing in such measures due to high risk of crop failure (Rockstrm et al., 2002), yet this is a factor that is often overlooked (FAO, 2003). Even in areas where soil fertility is initially adequate for high yields, water harvesting schemes combined with a lack of nutrient replenishment may lead to depletion of fertility through nutrient mining, meaning initial increases in crop yields are unsustainable in the long-term (Critchley et al., 1992; Falkenmark and Rockstrm, 2004 in Falkenmark, 2007). Furthermore, soils with a high infiltration rate such as those with a high sand content can pose limitations on potential runoff (Boers, 1994), have a low water holding capacity resulting in low water availability in the root zone (Rockstrm et al., 2002), as well as lacking the structure necessary for the construction of RWH system components such as bunds (Critchley and Siegert, 1991). Finally, high potential evaporation may equate to higher soil moisture stress and a reduction in runoff harvested, but also a reduction in water volume available for irrigation for systems where water is stored in open reservoir (Kumar et al., 2008).

The suitability of RWH to provide adequate water to meet crop demand and reduce poverty is further contested when considered in the context of climate change. Since the beginning of the twentieth century, patterns of reduced rainfall levels and number of rainy days have been observed across southern Africa (Batisani and Yarnal, 2009; Hulme et al., 2001; Parida and Moalafhi, 2008) and are predicted to continue (Hulme et al., 2001; Makurira et al., 2009). Decreased and more stochastic rainfall may cause difficulties with the design of an effective RWH system (Boers, 1994; Critchley and Siegert, 1991; Reij et al., 1988) and may decrease

11

adoption levels as farmers view the degree of uncertainty and variability in rainfall as too great to invest in the technology (Botha et al., 2008; Boyd and Turton, 2000).

2.2.2

Understanding the priorities of resource poor small-scale farmers

The availability of resources such as finances, land and labour are widely cited as constraining factors to the adoption of RWH systems by farmers (Pachpute et al., 2009; UNESCO-IHE & IWMI, 2009). Many RWH systems demand a high initial labour input, which can present problems for some families, particularly those that are poorest, or headed by women (Cullis and Pacey, 1992), even where the willingness to replicate these systems is substantial (Nijhof et al., 2010). In some cases, the land demanded by RWH may leave the technology inaccessible to the poorest farmers who have little or no land suitable for crop production (Ellis, 2000; Kumar et al., 2008). Permanent RWH systems may also be unsuitable for nomadic farmers (Pacey and Cullis, 1991; Swatuk and Kgomotso, 2008), or those who do not formally own the land upon which they farm (Critchley and Siegert, 1991; Balke, 2008), due to the short-term potential benefits available from the installation of RWH on the land.

Rural livelihoods centre around the need to reduce the level of risk and uncertainty to ensure survival and well-being (Whitehead, 2002) and the suitability of RWH depends on the acceptance of the risk involved in the use of the technology by farmers (Andersson et al., 2009). Nigi et al. (2007) argues that the extent of unfavourable agricultural conditions with which farmers in developing countries are faced encourages them to adopt technologies such as RWH to lower risk levels. However, Toulin and Chambers (1990) are of the opinion that the harsh and widely varying conditions experienced in complex, diverse and risk prone areas prevent farmers from adopting RWH, as the technology fails to adequately reduce the risk levels involved in crop production (Boyd and Turton, 2000).

In marginal areas RWH cannot be considered as a stand-alone activity (Ramisch, 1999) and concentrating on improving water availability only will not solve all the problems connected to low agricultural productivity (Rockstrm et al., 2010), other inputs and cultural practices must also be optimised (Oweis and Hachum, 2006). In some cases the key resource lacking from farmers may the knowledge and skill required to manage their farmland effectively; therefore even with the provision of RWH to increase water availability it is possible that 12

production levels may remain low. Aside from aspects such as the application of fertilisers mentioned above, poor farm management practices that may limit yield increases include the selection of crops with a water demand that does not provide the best fit to rainfall patterns (Yuan et al., 2003), the untimely sowing of seeds (Kronen, 1994) and the inefficient application of water made available from RWH (Barron, 2009).

The suitability of RWH for agriculture and poverty reduction in developing countries also depends on the synchronisation of the technology with common farming livelihood strategies at both household and community level. At household levels livelihood strategies comprise extensive land use characterised by low levels of input (Dixon et al., 2001) and diversification to spread the level of risk and provide a higher level of buffering (Toulin and Chambers, 1990). On this basis it is argued that RWH may be unsuitable for small-scale farmers, as it is a practice that demands land use intensification (Jodha, 1990) and may limit the potential for diversification due to competition the technology may create between cattle and crops for land and resources may (Whitehead, 2002). At a community level the transformation of land that has historically been reserved for grazing into cropland for the purposes of RWH may intensify any conflict occurring between pastoral and arable farmers (Vohland, K. and Barry, B 2009). It is recommended that any RWH strategies consider methods for integrating both pastoral and arable farming (Botha et al., 2008; Pacey and Cullis, 1991), as schemes are likely to fail unless cattle grazing can be controlled (Critchley and Siegert, 1991; Vohland and Barry, 2009).

Possible future trends in livelihood strategies are also important to consider as these will relate to the appropriateness of RWH for small scale farmers in the long-term. If farming practices are intensified as predicted (Netting, 1993 in Ellis 2000; Ramisch,1999), RWH may become more in-line with and appropriate for farmers livelihood approaches. However, if crop production is regarded with increasingly low priority in the livelihood strategy as suggested by Ellis (2000), insufficient resources may be allocated to allow for the adoption of the technology (Borhang, 1992; Boyd and Turton, 2000; Hatibu and Mahoo, 1999).

2.2.3

The role of governance and institutions

Any government policy that has a significant influence on livelihood strategies of farmers, such as those regarding economic development, welfare, agriculture, investment and land 13

tenure, will affect the adoption and sustainability of RWH (Agarwal and Narain, 1999; Boyd and Turton 2000; Kumar et al., 2008) and is likely to long after it is ever withdrawn (Rethman and Muhangi, 2009). The involvement of the government is said to be a key influence on implementation of large-scale RWH strategies (Kahinda et al., 2007) and in areas where the government provides incentives for farmers, the adoption rate of RWH schemes appears to be higher (Baguma and Loiskandl, 2010; Tumbo et al., in press). Furthermore, government policies that lead to increased investments in infrastructure may allow for easier access to markets, which is thought to be an important requirement in encouraging farmers to adopt and sustain RWH (Rockstrm et al., 2010). However, Nigi (2009) found political intervention, at national or local level, to be detrimental to the success of projects and Jodha (1990) observed that public policies, such as drought relief schemes, disregard traditional coping strategies of farmers and replace the satisfaction of needs at the local level, which may reduce the adoption of RWH schemes as farmers rely more on external sources of food and income.

RWH schemes must be considered together with institutional and organisational environment (Cullis and Pacey, 1992) as a lack of emphasis on institutions may minimise the impact of any government policies or investments designed to encourage RWH (Baiphethi, et al., 2009). Institutional structures similar to those for full irrigation schemes are needed for the successful implementation of RWH schemes (Rockstrm et al., 2010) and catchment scale institutions in particular are important for ensuring the sustainability of crop production increases by both upstream and downstream farmers (Pachpute, et al. 2009). Nevertheless, the introduction of formal institutions and legal frameworks by governments has reduced the effectiveness of traditional and informal arrangements in RWH schemes in the past (Jodha, 1986; Boyd and Turton, 2000) and the resulting loss of traditional knowledge has led to a reduction in adoption and use of RWH (Boyd and Turton, 2000).

Institutions are a key factor to ensuring that RWH techniques move beyond a simple disseminated technology, to a sustainably adapted and managed technology (Botha et al., 2008) capable of reducing poverty. Institutions play a primary role in learning and knowledge exchange, development of best practices and experiences, farmer support and the management of RWH systems (Nijhof et al., 2010) and may help provide the poorest households with resources needed for the adoption of the technology (Fox et al., 2005 in Rockstrm et al 2010). However, RWH needs to be developed within the context of small14

scale farms (Whiteside, 1998) and although demonstrative field trials have proved a successful form of dissemination in some areas (Pachpute, et al., 2009), extensive training is likely to be a necessary part of any RWH scheme to provide farmers with the knowledge to not only apply the technology, but adapt it to their needs to ensure maximum benefit (UNESCO-IHE & IWMI, 2009).

2.3

Beyond Rainwater Harvesting: Dynamic and complex systems

Although RWH systems are generally viewed as benign (Batchelor et al., 2003; Kumar et al., 2008), benefits accrued by schemes implemented in one area may result in significant negative trade-offs in others (Batchelor et al., 2003). At a basin level RWH has the potential to drastically affect hydrology and ecosystems (Barron, 2009) as one third of rainfall is required to sustain the wider environment (UNEP, 2006). To maximise reductions in poverty and land degradation, societal and ecological demands need to be considered in combination with crop production (Vohland and Barry, 2009), as hydrology and ecosystems in semi-arid and arid areas are highly sensitive to changes in green water flow and small changes to rainfall and runoff levels may be large in a relative sense (Falkenmark, 2007). The analysis of implementation of RWH schemes, therefore, cannot be restricted to individual systems in isolation, particularly in the context of non-equilibrium environments that are characteristic of arid and semi-arid regions (Scoones et al., 2007).

2.4

The situation in Botswana

Botswana has a semi-arid to arid environment with a mean annual rainfall of 416mm (FAO, 2005), although in the east where the majority of small-scale crop production is carried out mean rainfall is between 200-550mm (see Figure 2.1). Rainfall is erratic and generally occurs in localised intense showers (Rethman and Muhangi, 2009); most regions of the country have a rainfall reliability of between 0.5 and 0.7 (Tsheko, 2003). Potential evaporation is about four times the average annual rainfall at 2000mm (Ganesan, 2001) and rates are highest in the summer, between October and April, when the vast majority of the rainfall occurs (FAO, 2005). The country is characterised by gently undulating topography populated by occasional rocky outcrops (FAO, 2005), and sandy infertile soils which are not 15

very suitable for crop production (FAO, 2003). However, the hardveld region in the east consists of more fertile loamy clay soils with higher water holding capacity and more suitable for crop growth (Rethman and Muhangi, 2009).

Economic growth has lead to rapid urbanisation with formal employment becoming a major source of income in some areas (Rethman and Muhangi, 2009), but arable farming is still said to provide an important contribution to livelihoods, with 65 per cent of households using it as a source of livelihood (BIDPA, 2001 in CAR, 2007). Figure 2.1 indicates the distribution of farming livelihood zones in Botswana and highlights the areas in which crop production is most important.

Figure 2.1: The farming livelihood zones in Botswana (shown with hydrology and rainfall levels included)

60 55 50 45 40

40 35 35

40 45 30 25 50 55

(Sources: Rethman and Muhangi, 2009 for livelihood zones map; GoB, 2009 for rainfall data)

16

Water scarcity in Botswana is high and this poses a significant restriction on crop production, but water availability is not the only factor affecting arable farming; poor farm management, inadequate arable land, an ageing farming population, inadequate market access, damage to crops by livestock and a lack of community level institutions are all aspects thought to limit crop production (AfDB, 2008; GoB, 2009; GoB, 2006a; Whiteside, 1998).

Evidence indicates that the government has a significant influence on crop production as the strong social support system is said to have led to the development of an enduring belief within the rural population that the government will provide come what may (Rahm et al., 2006). Drought relief programmes in particular are said to have reduced the uptake of yield enhancing technologies such as RWH (CAR, 2007). Furthermore, land allocation policies have provided farmers with plots that are unsuitable for crop production in some cases (Borhang, 1992) and the drive for rapid economic development by the Government has created more competition for labour and finances (Howard et al., 2007). However, little is known about how these issues may limit the adoption and sustainable us of RWH in Botswana.

Looking into the future, the Government of Botswana recognises a continued need to increase the productivity within the arable sector to reduce the growing level of food imports and the dependence on government in rural areas (GoB, 2009), yet there is no explicit evidence of plans to encourage the use of RWH in small-scale agriculture (GoB, 2006b). Meteorologically there is a trend towards a reduction in rainfall levels, decrease in rainy days and decrease in rainfall intensity in the future as a result of climate change (Batisai and Yarnal, 2009) and mean annual rainfall could decrease by between 8 per cent and 54 per cent by 2100 (Kenabatho et al., 2009); therefore water availability will continue to be a primary concern among farmers. It is possible that this need to increase small-scale agricultural productivity could be met by RWH, but research needs to be conducted to determine the ability of the technology to overcome the risks involved in crop production, including those associated with climate change.

17

3. Results and discussion

3.1

Results

Twelve interviews were carried out as part of the primary data collection in Botswana. During the interview process participants were asked to describe the types of rainwater harvesting (RWH) used by small-scale farmers in Botswana and outline the factors that affect the adoption of the technology. Comprehensive notes from the interviews can be found in Appendix B; as mentioned previously, the interviewees have been anonymised in order to protect their identity. A coding process was used to identify key themes in the participants responses and the findings are summarised in the following section.

3.1.1 Rainwater harvesting practices in Botswana

[RWH is] the most important technical innovation for farmers dependent on rain-fed agriculture. (Interviewee A)

The primary data indicates that RWH is used by small-scale farmers in Botswana, but that systems have not been widely adopted . Systems are built either independently by farmers, or through government-led schemes implemented by the Ministry of Agriculture (MoA). A range of different RWH techniques was mentioned by the participants, including both in-situ and ex-situ systems, the latter of which appear to be the most extensively used. The use of in-situ RWH methods was observed at one location only, no other evidence of the use of insitu systems in small-scale agriculture was found. Many participants referred to the use of pans to provide water for agriculture, particularly in the west of Botswana where it was reported they often provide the only source of water. Other ex-situ methods of water collection reportedly used are sand rivers, hafirs and small earth dams. However, there was mixed opinion as to how much these methods of water collection are continuing to be utilised by small-scale farmers.

18

3.1.2

Factors affecting adoption

Hydro-ecological factors

The volume of water is not enough to provide for irrigation due to low levels of rainfall, even if roof tops are utilised as well as land. (Interviewee L)

Insufficient rainfall was the most frequent factor cited for the non-adoption of RWH strategies by small-scale farmers in Botswana: more than half of the participants were of the opinion that the volume of water collected does not provide adequate irrigation to crops. Nonetheless, a number of participants were of the opinion that rainfall levels in Botswana are sufficient to provide water to meet crop demand for the entire duration of the crop growing season. Interviewee K stated that with the use of an in-situ RWH system, a good harvest was virtually guaranteed. A common agreement among the vast majority of participants was that traditional methods of RWH have successfully maintained agricultural water availability for crop production in the past, but no longer provide farmers with enough water for the entire growing season. According to Interviewee C rainfall cannot be depended upon these days and the increasingly stochastic nature of rainfall in Botswana has contributed to the inability of traditional methods to continue providing adequate water for crop growth. It was reported that traditional rainfall patterns are still essentially in place in some areas and rainfall is slightly delayed or heavier than usual. In others, rainfall no longer has an observable pattern and significant volumes of rain can occur in the traditionally dry winter months of June to August.

The high evaporation rate was reported to have a marked affect on the ability of RWH systems to provide adequate water to increase production, contributing to a lack of adoption of systems by some farmers. Participants commented that a high rainfall-evaporation ratio results in the loss of a significant proportion of rainfall harvested that would otherwise be available for crop consumption, particularly in traditional systems where water is lost through evaporation form surface storage in pans. In in-situ systems evaporation directly from the soil causes most problems and the application of mulch on crop land was said to be needed to reduce the loss of water stored in the soil. However, in some areas the majority if mulch is consumed by cattle and so unavailable for use in RWH schemes. 19

Availability of assets

Rainwater harvesting [for small-scale farmers in Botswana] is a good idea in theory, but not in practice. (Interviewee K)

An important factor affecting the adoption of RWH systems (even those which the Government has heavily subsidised) is the reluctance of farmers to dedicate their time and labour to implementing the systems. Two participants involved in the MoA Water Development Sections small earth dam construction scheme stated that in some cases farmers have only agreed to help with construction of dams upon receipt of payment, despite the fact that labour is expected to be given in-kind in return for the construction of the dam by the Government. According to the participants, problems regarding labour availability are primarily attributed to a lack of willingness to work on farmland among the rural population, which is partly due to competition for labour from the formal sector.

Availability of labour for RWH implementation is apparently restricted further by a lack of community cohesion within Botswana and a lack of willingness to work at group level. Many of the interviewees mentioned that people in Botswana are not keen to work together and any strategies involving the group cooperation of farmers are not likely to be successful. In some cases this was attributed to divides created by unevenly distributed development causing social tension, but in others this was said to be due to the co-existence of pastoral and arable farmers with conflicting priorities.

Interviewees gave mixed opinions regarding the influence that availability of financial capital has on RWH schemes, but the majority stated it to be a barrier to the initial adoption of the technology, even when heavily subsidised by the Government.

There was a government scheme in the past to provide a roof rainwater harvesting system., but farmers had to provide funds for this and so the scheme ultimately didnt work due to farmers lack of financial resources (Interviewee B)

20

Issues associated with a lack of financial resources regarding crop production in general also appear to have an indirect affect on the adoption of RWH by small-scale farmers. For example, an inability to purchase fencing reduces the likelihood of crop production, because without fencing the potential of damage to crops by livestock is high. Furthermore, insufficient finances for public transport prevents many farmers from accessing their fields, which are generally located several kilometres away from the homestead. Possibilities to obtain grants and loans are said to be available for farmers through the Citizen Entrepreneurial Development Agency (CEDA), but a lack of collateral appears to cause problems in gaining access to these.

An inability of some individuals to obtain a plot of land at all also poses problems. One reason for farmland scarcity is that elders are reportedly retaining family crop land, depriving younger generations from land that would have traditionally been passed through family generations. It is possible for these young farmers, or any household without farmland, to obtain a plot through the Land Board, but the process is said to be convoluted and applications to take many years to process. Moreover, it is reported that the Land Board has in some cases allocated land unsuitable for crop production.

The final asset that has been identified as limiting the uptake of RWH, is farming skill and knowledge. The training available to farmers is said to be inadequate and poor farm management at both an individual and group level restricts the performance of agriculture, which in turn makes farmers reluctant to invest in RWH technology.

Livelihood strategies Pastoral farming is a fundamental part of the traditional livelihood strategy for many Batswana and this appears to be a key factor limiting the adoption of RWH. It seems that farmers in Botswana are reluctant to invest in crop production technologies, such as RWH, when the level of competition with cattle is high due to the risk this imposes on crop yield levels. Although the erection of fencing may reduce the risk to crops posed by freely grazing cattle, in some areas this may not be possible due to limited land resources and the use of a plot of land purely for crop production may create conflict between arable and pastoral farmers. The competition for resources between crops and cattle was a re-occurring theme in the interviews and evidence indicates the needs of cattle are generally prioritised over those 21

of crops at both household and community level due to the higher value attributed to cattle through greater potential income.

Growing changes in livelihood strategies in Botswana also appear to limit the potential adoption rate of RWH in the future due the decreasing priority allocated to crop production. In recent years high economic development and a dramatic increase in the availability of formal employment has led to migration from rural areas, particularly among young Batswana, and farming no longer provides an attractive career choice. All participants spoke of a general trend of moving away from agrarian livelihoods in rural areas; traditional farming households appear to be abandoning their cropland and moving towards other livelihood strategies. The priority now appears to be on pursuing strategies that provide higher incomes that will enable the purchase of food, rather than continuing to grow it. Those households remaining in the farming sector are reported to be increasingly concentrating on pastoral farming over crop production, due to low prices and lack of accessible market for traditionally grown crops, such as maize, compared to high prices and large accessible market available for cattle.

Governance and institutions The findings from this research indicate that the implementation of policies and plans by the Government of Botswana related to RWH, have had little influence on increasing uptake of the technology by small-scale farmers. The participants confirmed that the impacts of incentivised government schemes aimed at increasing the adoption of RWH, such as the Arable Lands Development Plan (ALDEP), have had some success, but in most cases confirmed that dissemination of the technology has not extended past the pilot projects.

The increasing provision of water resources infrastructure, such as hand dug wells and boreholes, in rural areas under the current Integrated Support for Arable Agricultural Development (ISPAAD) scheme, is said to reduce the perceived need for RWH among farmers and increased dependence on the government for water supplies. In addition, extensive support given to rural populations through drought relief schemes has reduced not only the uptake of technologies such as RWH in crop production, but led to a reduction in arable farming in general:

22

Rainwater harvesting will not work in Botswana as there is no pressure to grow food. People receive too much help from the government and so have stopped farming. (Interviewee K)

Reductions in arable farming are also said to have occurred in recent years due to an emphasis by the Government on growth of the formal employment sector, commercialisation of agriculture, reduced support for small-scale farmers and a focus of support on pastoral farming. The increasing allocation of land by the Land Board of plots that are unsuitable for crop production was also stated to present problems, as mentioned previously, along with the allocation of land in the runoff path of traditional RWH pans.

Finally, insufficient institutional capacity in Botswana is reported to have limited adoption of RWH, due to both a lack of number present and the ineffectiveness of existing institutions. Several of the interviewees stated that the Agricultural Extension Workers from the MoA have insufficient time to attend to farmers needs in their designated area and inadequate transport to reach the most rural areas. Furthermore, many of the extension workers do not possess adequate knowledge to advise farmers suitably on RWH schemes. The number of institutions available to assist farmers in Botswana has declined significantly in the past few decades; those that remain appear not to work efficiently and group organisation has become a problem across the whole agricultural sector. NGOs, farming organisations and cooperatives that encouraged the adoption of RWH schemes among farmers in the past, but sharp falls in funding from donors since Botswanas achievement of middle-income status is said to have led to the collapse of many such organisations. This has made it more difficult for farmers to obtain funds for projects and gain knowledge to allow them to implement RWH on their land.

3.2

Discussion

The results indicate that RWH has been traditionally used by small-scale farmers in Botswana, but current use is not widespread. Factors identified as limiting the adoption of the technology include climate, availability of assets, rural livelihood strategy, government policy and institutions. In the first part of this section, the implications of these findings are 23

discussed and the ability of RWH to reduce poverty in Botswana is analysed. In the second part of this section, the findings from this analysis are used together with past research to propose factors that may affect the suitability of RWH for increasing crop production and reducing poverty in developing countries in general. Finally, these factors are summarised in a decision-making matrix that may potentially be used to aid the implementation of RWH projects.

3.2.1 The ability of rainwater harvesting to increase crop production and reduce poverty in Botswana.

Dynamic and interdependent environment Primary data indicates that in the past RWH has traditionally played an important part in small-scale farming in Botswana, but the abandonment of these schemes in recent years, as reported by the participants, is an indication that the benefits originally gained from using the technology may no longer be felt. Despite reductions in rainfall volume during the past few decades (Batisani and Yarnal, 2009) the mean annual rainfall for Botswana still falls within the optimum range for RWH (Pacey and Cullis 1991; Mati et al., 2006); therefore it is possible that rainfall has simply become too erratic to sustain rain-fed crop production, as found in other dryland areas (Reij et al., 1988). It is possible that the adaption of these traditional methods to include the storage of water in a closed container as opposed to an open reservoir may increase the ability of RWH to increase crop production in Botswana, as this would reduces losses due to evaporation and increase water availability for irrigation, but more research is required to determine if this is the case. In the future it is possible that even adapted traditional systems may prove unsuccessful at increasing crop yields as climate change predictions for the region indicate a continuing decrease in volume and increase in variability of rainfall (Hulme et al., 2001; Kenabatho et al, 2009).

Interviewee K provided evidence for the ability of in-situ methods to increase crop production, but this approach is relatively new to the country and there is no evidence as to whether in-situ RWH has the ability to consistently increase crop yields in the long-term, or whether it can be up-scaled across rural Botswana without impacting negatively on the wider environment. Again, further investigation is required before the full extent of the potential of in-situ RWH can be assessed.

24

Allocation of minimal resources to achieve maximum benefits The findings from this research confirm current ideas that the availability of labour and finances poses a significant barrier to the adoption of RWH (Pachpute, et al., 2009; UNESCO-IHE & IWMI, 2009; Pauli and Bjerregaard, 1999). The requirements for initial inputs on systems seem to cause problems for some farmers, as has been experienced in other areas (Cullis and Pacey, 1992; Nijhof et al, 2010). However, in Botswana even government-led schemes that provide a high degree of assistance towards labour and finances have not helped to encourage the widespread adoption of RWH, in contradiction to findings by others (Baguma and Loiskandl, 2010; Tumbo et al, in press). This suggests that other factors more influential than labour and finances may pose a greater barrier to the adoption of RWH and potential for increases in crop production.

In accordance with RWH literature (Barron, 2009; Kronen, 1994; Yuan et al., 2003), water is one of the many barriers to crop production and findings indicate that benefits from the adoption of RWH in Botswana may be limited unless improvements in farm management are also made. Farmers traditionally adopt a low input and low management approach to agriculture and have little knowledge of effective farm practice; as a result of this any potential benefits due to the adoption of RWH may either not materialise, or may be unsustainable in the long-term. Moreover, rural livelihood strategies focus on the need to reduce uncertainty (Whitehead 1997), so unless RWH can be shown to adequately reduce risk levels involved in crop production in Botswana, farmers will be reluctant to allocate any additional resources other than the minimum necessary (Boyd and Turton, 2000).

Regardless of whether RWH can be proven to decrease the risk associated with crop production, it is argued that adoption rates may remain low due to the conflict the use of the technology creates with other key sources of livelihood. Pastoral farming and the process of freely grazing cattle is an important part of livelihood for the majority of rural Batswana and it is argued that increases in crop production may only be achieved if a RWH strategy integrating both pastoral and arable farming is considered, as recommended by Botha, et al. (2008) and Pacey and Cullis (1991). As found by Pacey and Cullis (1991) in Kenya, the use of RWH in Botswana appears to intensify conflict between arable and pastoral farming at both community and household level, as land usually available for grazing is transformed into cropland; competition for water and vegetation also causes conflict (Whitehead 1997). Due to the higher potential cattle have to provide livelihood security to small-scale farmers 25

their needs are generally prioritised over those of crops; therefore as long as RWH schemes in Botswana do not allow for the effective production of livestock in conjunction with crops, it is predicted that the technology is not likely to be adopted by farmers.

In recent years evidence suggests that economic development and a rapid increase in formal sector employment has led to a reduction in importance of arable farming in rural livelihoods, in agreement with Ellis (2000). It is suggested that this has resulted in the insufficient allocation of resources (primarily labour) for RWH, as identified in previous research (Borhang, 1992; Boyd and Turton, 2000; Hatibu and Mahoo, 1999) and if the importance of arable farming continues to decrease into the future, the potential for widespread adoption of RWH may be low. It is proposed that due to the level of development that has occurred in Botswana and the low priority allocated to crop production by households, the potential to increase small-scale crop production through RWH may be limited.

Enabling environment and support systems The results from this research confirm findings in the literature that governments may have a detrimental effect on the ability of RWH to increase small-scale crop production (Agarwal and Narain, 1999; Boyd and Turton, 2000; Kumar et al., 2008). In agreement with findings by Jodha (1990), the implementation of extensive social security measures by the Government of Botswana has created a high level of dependence within the rural population and a disregard for coping strategies at the local level. It is suggested that high levels of dependence on the state have led to low levels of RWH adoption as farmers see no need to increase productivity because any requirements for food during times of need are met through government relief schemes. It is argued that unless attempts are made to increase the independence of the rural population on the state, the possibilities for increasing crop production remain low.

As highlighted in previous research (Jodha, 1986; Boyd and Turton, 2000), the involvement of the Government in RWH schemes and agriculture appears to have reduced the effectiveness of traditional and informal arrangements in rural areas. In particular, increased bureaucracy has reduced the adoption of traditional RWH methods, due to the need to apply for permission to excavate pans to collect runoff. Furthermore, a focus on the commercialisation of agriculture and an apparent lack of provision of access to local markets 26

may have discouraged farmers from increasing small-scale productivity, in accordance with findings by Rockstrm et al (2010). If Government policies were adjusted to reduce the degree of bureaucracy and provide more incentives for small-scale farmers to increase productivity, it is likely that the uptake of RWH may increase.

It is suggested that a lack of appropriate institution building has minimised the impact of RWH schemes in Botswana, as found in South Africa by Baiphethi, et al. (2009). If RWH is to have the potential to increase small-scale crop productivity on a widespread basis it is suggested that the capacity of existing institutions, such as the MoA extension scheme, need to be improved and new institutions need to be created at the community level to provide a platform for learning and knowledge exchange between farmers (Nijhof et al., 2010), along with more extensive training to equip them with the skills to adapt RWH systems to their particular needs (UNESCO-IHE & IWMI, 2009). However, unless the unwillingness of Batswana to work together at community level can be overcome, the sustainability of any institutions may be low.

3.2.2

The suitability of rainwater harvesting: Towards a matrix for assessment

The evidence from Botswana reiterates that dryland agricultural systems are inherently risky (Enfors and Gordon, 2007) and that the suitability of RWH for increasing crop production and reducing poverty ultimately depends on the potential the technology holds for reducing the risk involved in arable farming, without restricting benefits gained from other important livelihood sources. Drawing on both current literature and findings from this research, key requirements needed to ensure the suitability of RWH in any particular context have been proposed and divided into those affecting initial adoption and those affecting longer-term sustainability of RWH. These requirements are outlined in the following section and summarised in a potential decision-making matrix in Table 1.

Climate and ecology This research has confirmed that one of the greatest risks to rain-fed crop productivity is high rainfall variability and the ability to reduce the reliance on stochastic rainfall is the key to the suitability of any technology aimed at increasing productivity (Falkenmark and Rockstrm, 2008). In order to be suitable for increasing crop production the distribution of the rainfall must be timely as RWH can only decrease the negative impact of rainfall 27

variability on crop production to a certain extent (Kumar et al., 2008). In some semi-arid and arid areas the variability of the rain may be too great for certain types of RWH to provide sufficient benefits (Reij et al, 1988). Furthermore, RWH must enable crop water demand to be met both in the short- and longer-term future, maintaining a lowered level of risk in arable farming even in the context of climate change. The findings from this research indicate that the relationship between the climate and crop water demand is extremely complex and each case needs to be assessed on an individual basis to determine if the combination of climatic characteristics is compatible with RWH and if adequate water can be harvested to meet crop water demand for the duration of the growing season.

Climatic factors need to be considered closely with wider ecological issues, as these can have a marked influence on yields. For example, areas with unfavourable soil characteristics, such as low moisture holding capacity, or low fertility may not be suitable for RWH (Critchley and Siegert 1991). Data collection in Botswana reiterated the importance of combining RWH with soil conservation measures if crop production is to be most successful (AfDB, 2007, Rockstrm et al., 2002).

Farming practice Findings from this research also highlight the need for effective farming practice, as aspects such as unsuitable sowing time (Kronen, 1994) and ineffective use of water harvested (Barron, 2009) may restrict crop production despite increased water availability. Additionally, although data gathered from Botswana was unable to confirm that the presence of RWH in traditional farming increases the likelihood of the uptake of new RWH strategies, literature suggests that this has proved the case in other projects (AfDB, 2007). As a result it may be possible to conclude that RWH may be more suitable for use in crop production in areas where it has been used traditionally, as farmers are likely to be more familiar with the systems.

Availability of assets Although socio-economic factors are often neglected by those working on the implementation of RWH schemes, this research has highlighted that such factors have a crucial influence on the suitability of schemes. A lack of resources, including finances, labour and land, is a key constraint to the adoption of RWH by the poorest farmers (Pachpute et al., 2009; UNESCO-IHE & IWMI, 2009) and although government schemes in 28

Botswana were unsuccessful, in general the provision of grants and assistance from governments or NGOs has been shown to reduce the barrier to the use of the technology (Baguma and Loiskandl, 2010; Tumbo et al., in press).

Livelihood strategy Evidence suggests that in order sufficient resources to be allocated for the adoption and sustained use of RWH, crop production should be allocated with a relatively high priority level within the household livelihood strategy. The conflict between pastoral and arable farming poses perhaps the greatest barrier to the use of RWH in countries where livestock make a large contribution to livelihoods and competition for land, water and vegetation may lead to the failure of RWH systems unless an appropriate system that allows the co-existence of cattle and crops can be implemented (Pacey and Cullis, 1991). Lastly, in nations with high levels of economic development and formal employment where the importance of crop production in rural livelihood strategies is reducing (Ellis, 2000), RWH may not be suitable for increasing crop production as the investment in non-farm income is seen by farmers as having greater potential for reducing livelihood vulnerability than those in arable farming.

Governance The benefit of policies and schemes involving subsidies and grants specifically for the purpose of RWH adoption has already been discussed, but if the use of RWH is to be sustainable on a large-scale it is suggested that these incentives need to be accompanied by complimentary policies that encourage the growth of small-scale agriculture. This may include policies comprising improvements to rural infrastructure to allow market access for farmers (Rockstrm et al., 2010) and an appropriate reduction in drought relief to provide incentives to increase the efficiency of crop production.

Institutions It is suggested that suitability of RWH in small-scale agriculture will be greater if both informal and formal institutions exist at local and national level, as this will ensure the sustainable utilisation of RWH (Botha et al., 2008) where benefits are shared equally among all users and allow for the training of farmers and sharing of experiences in the field (Nijhof et al., 2010). This research has highlighted the difficulties that may be experienced with the involvement of political institutions in RWH schemes and small-scale agriculture in general and reiterated the need for community level institutions (Nigi, 2009). 29

Table 1: Decision-making matrix indicating requirements for the suitability of RWH systems in agriculture in terms of initial adoption and longer-term sustainability.

Factor Climate and ecology

Initial adoption Adequate data on rainfall, evaporation and soil properties to allow for effective design of systems Potential rainfall and runoff volume and distribution compatible with crop water demand Soil with good water holding capacity (and sufficient structure if required for any construction in association with RWH system) Soil nutrient level capable of sustaining crop growth in at least the short-term

Longer-term sustainability Sufficient availability of water to maintain wider ecosystems in region despite presence of RWH systems Equal benefits for both downstream and upstream users in basin Minimal affects of climate change on ability of RWH to provide adequate water Rainfall patterns offer opportunity for enhancement via RWH with little excessive drought High predictability of rainfall, or provision of weather forecasts, to allow for timely farming practice and efficient use of water harvested

Farming practice

Traditional use of RWH in crop production Growth of relatively high value cash crops Labour and equipment investment acceptable

Combined use of RWH with soil conservation methods and application of fertiliser Optimisation of farm management skills to decrease limitations on crop production caused by factors other than water availability (eg. seed sowing) Fits wider farming systems in location

Availability of assets

Availability of finances, materials and labour required for adoption through subsidies and assistance from appropriate institutions Adequate land availability and land tenure Knowledge and understanding of RWH Low input demand for adoption

Adequate availability of land suitable for long-term crop production close to homestead Low input demand for maintenance Availability of finances, materials and labour required for maintenance through subsidies and assistance from appropriate institutions Possession of skills to adapt RWH system to meet specific needs of farm/catchment

30

Livelihood strategy

Crop production high priority in livelihood strategy Significant reduction in risk of crop failure with implementation of scheme Rapid return on initial investment Lack of conflict with other current livelihood strategies (eg. pastoral farming)

No detrimental impact on wider livelihood strategy (eg. diversification) Provides consistent boost to household income and nutrition Sustained high priority of agriculture in livelihood strategy Low competition for resources from other livelihood strategies (eg. formal employment)

Governance

Incentivised policies and schemes, including grants and subsidies Provision of adequate institutions to provide training and assistance with adoption Policies encouraging independence of rural population from government Minimal government bureaucracy involved in adoption of RWH schemes

Complimentary policies encouraging the increased importance and growth of small scale agriculture and crop production. Legal framework defining rights and responsibilities of water users Provision of infrastructure to increase access to markets Provision of adequate institutions to provide training and assistance with maintenance and use

Institutions

Government with high capacity to implement relevant policies and schemes Presence of local level institutions to implement farmer centred research and extension Assistance of community/village leaders in adoption issues Presence of institutions to provide resources for initial investment (eg. micro credit organisations)

Catchment level institutional linkages between upstream and downstream users to monitor and manage water supply and demand within both agriculture and other sectors Community level institutions to allow for farmer participation in planning, cost sharing, continual evaluation and

improvement of systems

(Source: various)

31

3.3

Summary

The results indicate that RWH is being used by some farmers in Botswana, but that widespread adoption has not taken place. The factors affecting RWH in Botswana have been found to not only include issues related to the technical capability of the technology to increase crop production, but also issues related to the context within which the technology must fit, such as the availability of assets, livelihood strategy of farmers, government policy and institutions. The current potential for increases in crop production through the use of RWH in Botswana appear to be limited and potential for the future is uncertain, primarily due to impacts of climate change and alterations to rural livelihood strategies due to economic development. With appropriate adaptation of systems and the development of community level institutions to provide in-depth training to farmers potential may improve, but it is possible that the prevalence of pastoral farming may be too great a barrier to be overcome.

The factors that affect the suitability of RWH for improving crop yields and reducing poverty in any particular developing country are evidently significantly more complex than those considered in existing frameworks. This research indicates that important factors needing consideration relate to climate and ecology, farming practice, availability of assets, livelihood strategy, governance and institutions; all of which need to be taken into consideration when considering the implementation of RWH in small-scale agriculture. The suitability of RWH for use in developing countries in general is uncertain as a wide range of factors influence the successful adoption and sustained use of the technology. Each potential project needs to be assessed individually, with particular emphasis given to the context within which the system(s) must fit.

32

4. Conclusion

In the past two decades interest in rainwater harvesting (RWH) has grown steadily (Nijhof et al., 2010) and there is a general belief that the technology has the potential to lift farmers out of the poverty trap through the improvements it can provide to agricultural productivity (Barron, 2009, NWP, 2007; Vohland and Barry, 2009). However, in many areas empirical evidence does not support this hypothesis and improvements to livelihoods due to the use of RWH have been low (Reij et al., 1980; Hatibu et al., 2006). With the use of Botswana as a case study and source of primary research, one of the main objectives of this research has been to categorise and contextualise the range of factors that affect both the initial adoption and longer-term sustainability of the technology; the implications these factors have on the ability of the technology to increase crop production has also been analysed. The overall aim has been to examine the factors that determine the suitability of RWH use in small-scale agriculture for increasing crop production and reducing poverty in developing countries in general and produce a decision-making matrix that can be used to assess the level of suitability in any particular context.

The findings from this research indicate the primary factors that affect the adoption of RWH in Botswana are: hydro-ecological factors, availability of assets; rural livelihood strategy, governance and institutional capacity. An analysis of these factors and the context within which they occur indicates that the potential of the technology to increase small-scale crop production in the country depends on much more than the ability of the technology to provide sufficient water for supplemental irrigation. The factors have been shown to occur within the context of a dynamic and interdependent environment where farm management skills are poor and the incentive to increase crop production is low. The ability of the technology to reduce livelihood vulnerability and provide maximum benefit for minimal input is a key consideration for farmers. As in Botswana, the suitability of RWH for increasing crop production and poverty reduction in developing countries in general also depends on factors related to climate and ecology, farming practice, availability of assets, livelihood strategy, governance and institutions; all of which need to be taken into consideration when considering the implementation on RWH in small-scale agriculture. 33

The current potential for increases in crop production through the use of RWH in Botswana appear to be limited and potential for the future is uncertain; this is primarily due to impacts of climate change and alterations to rural livelihood strategies as a result of economic development. With appropriate adaptation of systems and the development of community level institutions to provide in-depth training to farmers potential may improve, but it is possible that the prevalence of pastoral farming may prove too great a barrier to be overcome. The suitability of RWH for use in developing countries in general is uncertain as a wide range of factors influence the successful adoption and sustained use of the technology; each potential project needs to be assessed individually, with particular emphasis given to the context within which the system(s) must fit. It is recommended that further research is conducted into the factors that determine the suitability of RWH for increasing crop productivity and poverty reduction in developing countries; this will allow for the refinement and expansion of the proposed decision-making framework presented into a more comprehensive RWH implementation framework.

34

References
Acquah B K (2003) Making technologies work for resource poor farmers in Botswana. Pula: Botswana Journal of African Studies, Vol. 17, No.2. African Development Bank (AfDB), (2007). Assessment of best practices and experiences in rainwater harvesting. Rainwater Harvesting Handbook. Tunis, Tunisia: AfDB African Development Bank (AfDB), (2008). Botswana Pandamatenga Agricultural Infrastructure Development Project Appraisal Report, Botswana: AfDB. Andersson, J., Zehnder, A., Jewitt, G., and Yang, H., (2009). Water availability, demand and reliability of in situ water harvesting in smallholder rain-fed agriculture in the Thukela River Basin, South Africa. Hydrology and. Earth System Sciences, Vol.13, pp.23292347. Agarwal A and Narain S (1999) Making Water Management Everybodys Business: Water Harvesting and Rural Development in India, Gatekeeper Series No.87, IIED, London . Baguma, D. and Loiskandl, W,. (2010). RWH technologies and practises in rural Uganda: a case study. Mitigation and Adaptation Strategies for Global Change, Vol.15, pp.355369. Baiphethi, M.N. et al., (2009). Reducing poverty and food insecurity by applying infield rainwater harvesting (IRWH): How rural institutions made a difference. African Journal of Agricultural Research, Vol. 4, No.12, pp.1358-1363. Balke, K-D., (2008) Agricultural technology for developing countries, Annex 1: Rainwater Harvesting. Science and Technology Options Assessment, Germany: Universitt Tbingen. Barron, J. (Ed.), (2009). Rainwater harvesting: A lifeline for human well-being, Sweden: UNEP/SEI. Batchelor, C. H., Rama Mohan Rao, M.S., and Manohar Rao, S., (2003). Watershed development: a solution to water shortages in semi-arid India or part of the problem?. Land Use and Water Resources Research, Vol.3, pp.1-10. Batisani, N., and B. Yarnal. (2010). Rainfall variability and trends in semi-arid Botswana: implications for climate change adaptation policy. Applied Geography. Vol.30, No.4, pp.483489. 35

Boers, Th.M., (1994). Rainwater Harvesting in Arid and Semi-Arid Zones, Publication 55, The Netherlands: ILRI. Borhang, K., (1992) Politics, Administration and Agricultural Development: The Case of Botswanas Accelerated Rainfed Arable Programme, Ph. D, Department of Social Science and Development, University of Bergen. Botha, J.J. et al., (2008). On-Farm Application of In-Field Rainwater Harvesting Techniques on Small Plots in the Central Region of South Africa, Volume 1 of 2: Main Report, TT313/07, South Africa: Water Research Commission. Boyd, C and Turton, C. (Eds.), (2000) The Contribution of Soil and Water Conservation to Sustainable Livelihoods in Semi-Arid Areas of Sub-Saharan Africa, Agricultural Research and Extension Network, Paper, No.102, London: Overseas Development Institute. Bryman, A., (2004). Social Research Methods, Second Edition, Oxford: Oxford University Press. Bryman, A. and Burgess, R.G., (1994). Analysing qualitative data, London; New York: Routledge. Centre for Applied Research (CAR) (2007) SADC MAPP Draft country assessment report: Botswana, [Online] Available at: http://www.sadc.int/fanr/agricresearch/mapp/reports/Botswana%20%20National%20situatio n%20analysis.pdf. [Accessed: 12th July 2010]. Chacko E (2004). Positionality and praxis: Fieldwork experiences in rural India. Singapore Journal of Tropical Geography, Vol.25, No.1, pp.51-63. Comprehensive Assessment of Water Management in Agriculture (CA), (2007). In: Molden,D. (Ed.), Water for Food, Water for Life. A Comprehensive Assessment of Water Management in Agriculture. International Water Management Institute (IWMI), London: Earthscan. Coupe, S., (2001). Poverty and Technology. http://www.practicalaction.org.uk/t4sl/povertyandtechnology, 2010]. [Online] Available at: nd [Accessed: 22 October

Coventry, C., (2003). The role of technology in poverty reduction, In, Practical Action South Asia Group/ITDG, South Asia Conference on Technologies for Poverty 36

Reduction: Working towards a Regional Strategy, 10th and 11th October 2003, New Delhi: Practical Action/ITDG. Critchley, W. and Siegert, K (1991) A Manual for the Design and Construction of Water Harvesting Schemes for Plant Production, Rome: FAO. Critchley, W.R.S., Reij, C.P. and Turner, S.D., (1992). Soil and Water Conservation in SubSaharan Africa: Towards sustainable production by the rural poor, Rome: IFAD. Cullis, A., and Pacey, A., (1992) A development dialogue: Rainwater Harvesting in Turkana, London: IT Publications. Dixon, J., Gulliver, A., and Gibbon, D., (2001) Farming Systems and Poverty: Improving farmers livelihoods in a changing world. Rome/Washington DC: FAO and World Bank. Ellis, F., (2000). Livelihoods, Diversification and Agrarian Change. In Rural Livelihoods and Diversity in Developing Countries, Oxford: Oxford University Press. Efors, E.I. and Gordon, L.J. (2007). Analysing resilience in dryland agro-ecosystems: A case study of the Makanya catchment in Tanzania over the past 50 years. Land Degradation and Development, Vol. 18, pp.680-696. Falkenmark, M., (2007). Shift in thinking to address the 21st century hunger gap: Moving focus from blue to green water management. Water Resources Management, Vol.21, pp.318. Falkenmark, M. and Rockstrom, J., (2008). Building resilience to drought in desertificationprone Sub-Saharan Africa: The water perspective. Natural Resources Forum, Vol.32, pp.93-102. Faugier, J. and Sargeant, M. (1997) Sampling hard to reach populations. Journal of. Advanced. Nursing, Vol. 26, pp.790-797. Food and Agriculture Organisation of the United Nations (FAO) (2003) A perspective on water control in southern Africa: Support to regional investment initiatives, Land and water discussion paper 1, Rome: FAO. Food and Agriculture Organisation of the United Nations (FAO), (2005). Botswana, AQUASTAT country profile. [Online] Available at: http://www.fao.org/nr/water/aquastat/countries/botswana/index.stm [Accessed: 24th April 2010]. 37

Ganesan, C.T., (2001). Water Resources Development and Management: A Challenging Task for Botswana. Water International, Vol.26, No.1, pp80-85. Government of Botswana (GoB), (1999). Water Conservation Strategy Framework, Department of Water Affairs, Gaborone: GoB. Government of Botswana (GoB), (2006a). National Water Master Plan review: Final Report, Volume 8 Agriculture, Ministry of Minerals, Energy and Water Resources, Department of Water Affairs, Gaborone: GoB. Government of Botswana (GoB), (2006b). National Water Master Plan review: Final Report, Volume 3 Surface Water Resources, Ministry of Minerals, Energy and Water Resources, Department of Water Affairs, Gaborone: GoB. Government of Botswana (GoB), (2009). National Development Plan 10, Ministry of Finance and Development Planning, Gaborone: GoB. The Government of Botswana and UNDP (2004) Botswana Technology Needs Assessment on Climate Change, Final Report, Botswana: Ministry of Environment, Wildlife and Tourism, Department of Meteorological Services and UNDP. Hatibu, N., Mutabazi, K., Senkondo, E.M. and Msangi, A.S.K., (2006). Economics of rainwater harvesting for crop enterprises in semi-arid areas of East Africa. Agricultural Water Management, Vol.80, pp.7486. Hatubu, N. and Mahoo, H., (1999). Rainwater Harvesting Techniques for Agricultural Production: A Case for Dodoma, Tanzania, Sokoine University of Agriculture, Tanzania. [Online] Available at: http://www.animaltraction.net/contil/contil-hatibu-waterharvestingTZ.pdf [Accessed on 27th August 2010]. Howard, K., Sigwele. K. and Malope, P., (2007). Report of the national stakeholder workshop on SADC multi-country agricultural productivity programme (SADC MAPP), 3-4 October 2007 at Mogoditshane, Botswana, Department of Agricultural Research, Ministry of Agriculture, Gaborone: GoB. Hulme M, et al., (2001) African climate change: 1900-2100. Climate Change Research, Vol.17, pp.145-168. Jodha, N.S., (1986). Common property resources and rural poor in dry regions of India. Economic and political Weekly, Vol.21, No.26, pp.1169-1181. 38

Jodha,N.S., (1990). Drought Management: The Farmers Strategies and Their Policy Implications, Issues Paper No.21, Dryland Networks Programme, International Institute for Environment and Development. Kahinda, J-M.M., Rockstrom, J., Taigbenu, A.E.and Dimes, J., (2007). Rainwater harvesting to enhance water productivity of rainfed agriculture in the semi-arid Zimbabwe. Physics and Chemistry of the Earth, Vol.32, pp.10681073. Kenabatho, P., McIntyre, N.R. and Wheater, H.S., (2009). Impacts of climate change on water resource planning in Botswana. [article] (Personal communication, 2nd June 2010). Khaka, E, Malesu, M., Oduor, A., and Mati, B., (2005). Rainwater harvesting and the millennium development goals. Nairobi, Kenya: RELMA. Kronen, M., (1994). Water harvesting and conservation techniques for smallholder crop production systems. Soil and Tillage Research, Vol.32, pp.71-86. Kumar, M.D., Patel, A. and Singh, O.P., (2008). Rainwater Harvesting in Water-scarce Regions of India: Potential and Pitfalls. In, Amarasinghe, U.A.; Sharma, B.R., (eds.). Strategic analyses of the National River Linking Project (NRLP) of India, series 2. Proceedings of the Workshop on Analyses of Hydrological, Social and Ecological Issues of the NRLP. Colombo, Sri Lanka: International Water Management Institute, pp289 314. Kundhlande G; et al., (2004). Socio-economic study on water conservation techniques in semi-arid areas, Research Report No.1267/1/04, South Africa: Water Research Commission Makurira, H., Savenije, H.H.G., Uhlenbrook, S., Rockstrom, J. and Senzanje, A., (2009). Investigating the water balance of on-farm techniques for improved crop productivity in rainfed systems: A case study of Makanya catchment, Tanzania. Physics and Chemistry of the Earth, Vol.34, pp.93-98. Mamdouh Nasr (1999) Assessing Desertification and Water Harvesting in the Middle East and North Africa: Policy Implication, ZEF Discussion Papers on Development Policy No.10, Centre for Development Research, Bonn, pp.59. Mati, B.et al., (2006). Mapping the potential of RWH in Africa, A GIS overview on development domains for the continent and ten selected countries, Technical Manual No.6, Nairobi, Kenya: RELMA.

39

May, T., (2001). Social Research: Issues, Methods and Process. Milton Keynes: Open University Press. Netherlands Water Partnership (NWP), (2007). Smart water harvesting solutions: Examples of innovative low-cost technologies for rain, fog, runoff water and groundwater. The Netherlands: Netherlands Water Partnership. Nigi, S.N., Savenije, H.H.G. and Gichuki, F.N., (2007). Land use changes and hydrological impacts related to up-scaling of rainwater harvesting and management in upper Ewaso Ngiro river basin, Kenya. Land Use Policy, Vol. 24, pp.129140. Nigi, S.N. (Ed.), (2009) Climate Change Adaptation Strategies Water Resources Management Options for Smallholder Farming Systems in Sub-Saharan Africa. New York: The MDG Centre for East and Southern Africa, The Earth Institute at Columbia University. Nijhof, S., Jantowski, B., Meerman, R. and Schoemaker, A., (2010). Rainwater Harvesting in Challenging Environments: Towards institutional frameworks for sustainable domestic supply. Waterlines, Vol.29, No.3, pp.209-219. Oweis, T. and Hachum, A., (2006). Water harvesting and supplemental irrigation for improved water productivity of dry farming systems in West Asia and North Africa. Agricultural Water Management, Vol.80, pp.5773. Pachpute, J. S., Tumbo, S. D., Sally, H. and Mul, M. L, (2009). Sustainability of Rainwater Harvesting System in Rural Catchment of Sub-Saharan Africa. Water Resources Management, Vol.23, pp.28152839. Pacey, A. and Cullis, A., (1991). Rainwater harvesting: The collection of rainfall and runoff in rural areas, London Intermediate Technology Publications. Parr, J., and Shaw, R. (1999). Choosing an appropriate technology, Technical briefing No.49, WEDC, Loughborough: University of Loughborough. Pauli, G and Bjerregaard, O (1999) Towards a Technology Strategy for Sustainable Livelihoods, Sustainable Livelihoods Programme, UNDP.

Rahm, D., Swatuk, L. and Matheny, E., (2006). Water resource management in Botswana: Balancing sustainability and economic development. Environment, Development and Sustainability, Vol.8, pp.157-183.

40

Ramisch, J., (1999). The long dry season: Crop-livestock linkages in Southern Mali, Issue Paper No.88, IIED, London: Drylands Programme. Rapley, T. (2004), Interviews, in Swale, C., Gobo, G., Gubrium, J. and Silverman, D, Qualitative Research Practice, London: SAGE. Reij, C., Mulder, P. and Begemann, L., (1988). Water Harvesting for Plant Production, Washington DC: The World Bank. Rethman C. and Muhangi P. (2009) The Botswana National Livelihood Zoning Draft Report, SADC Food Agriculture and Natural Resources Vulnerability Assessment Committee, Gaborone: SADC. Robson C. (2002), Real World Research, Oxford: Blackwell. Rockstrom, J., Barron, J. and Fox, P., (2002). Rainwater management for increased productivity among small-holder farmers in drought prone environments. Physics and Chemistry of the Earth, Vol.27, pp.949-959. Rockstrm, J., (2003). Water for food and nature in drought-prone tropics: vapour shift in rain-fed agriculture. Philosophical Transactions of the Royal Society, London B 385, 19972009. Rockstrm, J. et al., (2010). Managing water in rainfed agricultureThe need for a paradigm shift. Agricultural Water Management, Vol.97, pp.543550. Rling, N., (2009). Conceptual and Methodological Developments in Innovation, in Sanginga P.C., Waters-Bayer A., Kaarla S., Njuki J. and Wettasinha C., Innovation Africa: Enriching farmers livelihoods, Earthscan, London. Rost, S. et al., (2009). Global potential to increase crop production through water management in rainfed agriculture. Environmental Research Letters, Vol.4, pp.9. Ryan, G.W. and Bernard, H.R., (2003). Techniques to Identify Themes in Qualitative Data. Field Methods, Vol.15, No.1, pp.85-109. Scoones, I. et al., (2007). Dynamic systems and the challenge of sustainability, STEPS Working Paper 1, Sussex: Institute of Development Studies.

41

Strauss, A. and Corbin, J., (1998). Basics of qualitative research: Techniques and Procedures for Developing Grounded Theory, Thousand Oaks California, Sage. Swatuk, L.A. and Kgomotso, P.K., (2008). Communicating IWRM in Botswana. Physics and Chemistry of the Earth, Vol.33, pp.881-888. Toulin, C. and Chambers, R., (1990). Farmer-First: Achieving Sustainable Dryland Development in Africa, Issues Paper No.19, Dryland Networks Programme, International Institute for Environment and Development. Tsheko, R., (2003). Rainfall reliability, drought and flood vulnerability in Botswana. Water South Africa, Vol.29, No.4, pp389-392. Tumbo, S.D., Mutabazi, K.D., Byakugila, M.M. and Mahoo, H.F.M., (In press). An empirical framework for scaling-out of water system innovations: Lessons from diffusion of water system innovations in the Makanya catchment in Northern Tanzania. Agricultural Water Management, Accepted for publication April 2010). United Nations (UN), (2006). Coping with water scarcity: A strategic issue and priority for system-wide action, UN-Thematic Water Initiatives, UN-Water.

United Nations Environment Programme (UNEP), (2006). Harvesting Rainfall a Key Climate Adaption Opportunity for Africa. [press release],13th November 2006, Available at: http://www.unep.org/documents.multilingual/default.asp?DocumentID=485&ArticleID=542 0&l=en [Accessed: 2nd November 2010]. UNESCO-IHE & IWMI, (2009). Strategies of Water for Food and Environmental Security in Drought-prone Tropical and Subtropical Agro-ecosystems (Tanzania and South Africa), Smallholder System Innovations in Integrated Watershed Management, UNESCO-IHE & IWMI. Vohland, K. and Barry, B., (2009). A review of in situ rainwater harvesting (RWH) practices modifying landscape functions in African drylands. Agriculture, Ecosystems and Environment, Vol.131, pp119127. Whitehead, A., (2002). Tracking livelihood change: Theoretical, methodological and empirical perspectives from northeast Ghana. Journal of South African Studies, Vol.28, No.3, pp.575-598.

42

Whiteside, M., (1998). Encouraging Sustainable Smallholder Agriculture in Botswana, in Whiteside, M., Encouraging Sustainable Smallholder Agriculture in Southern Africa, London: Earthscan. World Bank (2010) World Development Indicators [Online} Available at:

http://data.worldbank.org/country/botswana [Accessed: 12th October 2010].

Young, M.D.B., et al., (2002). Parched-Thirst: development and the validation of a processbased model of rainwater harvesting,.Agricultural Water Management, 55, pp121-140. Yuan, T., Fengmin, L. and Puhai, L.,(2003). Economic analysis of rainwater harvesting and irrigation methods, with an example from China. Agricultural Water Management, Vol.60, pp.217226.

43

Appendix A

Rainwater harvesting implementation frameworks

Figure A1: Decision tree on rainwater harvesting in agriculture

(Source: African Development Bank 2007)

Figure A2: Decision tree on rainwater harvesting in agriculture

(Source: Hatibu and Mahoo 1999)

Appendix B

Interview guide and notes

Semi-structured interview for rainwater harvesting use in small-scale farming, Botswana. 1. Background Information a) What do you use your RWH system for? b) What size is your farm? c) What crops do you grow? d) Is your farm for subsistence or commercial use? e) What RWH system are you using? traditional or modern? f) When was your RWH system installed? g) Have there been any changes to your farm as a result of installing the system? -Crop yields, Income, Farm size, Crops grown h) Have there been any changes to your livelihood as a result of installing the system -Family employment, Education

2. The RWH System a) How did you originally find out about RWH? b) Why did you choose to install it on your farm? c) How did you acquire your RWH system? i. How was it designed? ii. How was it built? iii. How was it financed? (self, loan, NGO, gov) d) How much did your RWH system cost to install? e) How effective do you find your system at collecting rainwater? f) What volume of rainwater do you collect a month in the rainy season? g) What is the maximum volume of water that you can store? h) How do you store the rainwater?

i) What percentage of its capacity is it working at?

3. Individual a) Attitude -Have any changes occurred in rainfall in recent years? -What do you think will happen to water levels in the future? -What will you do if there is no water? b) Knowledge/ Skills - Traditional or modern style system? - What training was given to you on how to use the system? - Where you involved in the design and construction of the system? - How confident do you feel using and maintaining the system? c) Resources -How do you maintain your system? -Where do you get spare parts/materials to mend the RWH system? -How much do you spend on maintaining your system each year? -Who cares for the RWH system? - Support from government?

4. Institutions a) Are there any institutions that you have been in contact with about your RWH system? b) Do you receive any help or guidance from the local/national government with your RWH system? c) What has been your experience of interacting with the government about your RWH system?

a. Have they helped or hindered information flow? d) Are there any traditional systems in place to help with the construction and maintenance of your RWH system? e) Have the farmers created any organisations/ committees to help in the implementation and maintenance of RWH systems?

Interview A - 12/07/2010 Interviewee A has been working in the field of RWH in Australia for 20years, with the aim of conserving resources to cope with climate change. Now he is based in South Africa and is working on the production of the African Organic Manual. Interviewee As PhD was in the field of catchment management. The Water Research Commission have done a great deal of work regarding RWH, with around 20 different projects completed, providing good overviews of the technology. Dirk Versveld has written a book about RWH in Africa, which looks at over 100 different technologies. The key issue for small scale farmers in this region is the need to reduce vulnerability. The most fruitful areas of rainwater harvesting in the region are small dams as rainfall occurs with high intensity and infiltrates rapidly when stored behind the dams. However, these small dams structures can have a negative impact on ecosystems due to alterations in water flow and also lead to an artificial drought downstream of the structures. The ideal approach is to place wetlands on the fringe above the dam to act as a silt trap, reducing sedimentation behind the dam, whilst also providing a grazing area for cattle. The rainwater harvesting system developed by Interviewee A involves a system of swales, which are dead level contour bunds. Swales double water infiltration and mulch, which is used in conjunction with swales, reduced evaporation by 40%. This system is an effective form of rainwater harvesting in South Africa due to the steep slopes and high rainfall, but in other areas, such as Botswana, there is insufficient rainfall intensity and a lack of steep slopes to utilize this system. The public works programme in South Africa has been unsuccessful. The permaculture movement developed out of the agricultural problems caused in the United States in the 1930s that lead to the creation of a dust bowl and the consequent use of swales placed downstream of forested areas that proved successful. Mollison and Polinsky have written on this topic and look specifically at farm land use planning, where topography, climate and soil are all considered and controlled for different purposes. P A Yeomans developed the swale system into a new approach called the key line system. This was used in the Water for Every Farm Project in South Africa and used a system of patterned ploughing to help conserve water. Rainwater harvesting is regarded by Interviewee A as the most important technical innovation for farmers dependent on rain-fed agriculture. However, it must also be recognised that it is not rainwater harvesting alone that is important and although the technology can be effective when used independently, other factors can help to increase its positive impact. Rainwater harvesting is the first practical way in which problems experienced on small scale

farms in the region can be tackled and it is an extremely empowering for farmers as provides clearly visible results. A key point is that often policy makers see the construction of big dams as an answer to the problem of water scarcity in the region, but this is often actually the worst possible solution. Decentralised solutions, such as rainwater harvesting, is more appropriate as although it has less political clout, it has more social clout. These kind of large dam schemes create a culture of dependence on the government as communities are not involved in the project process. Where traditional communities and culture are still in-tact natural resources are managed in a much more effective way. For example in Zimbabwe a law came into force the specific areas of wetland must be conserved. The Shauna people didnt stick to these laws, but still conserved important areas of wetlands due to cultural knowledge and traditional stories that are used to encourage the preservation of important ecological areas. In South Africa, this traditional knowledge has been lost. Trying to manage natural resources without people understanding the new management system can lead to problems with dams, as people spontaneously try using such schemes to manage water resources without proper consideration of the risks to communities and the environment involved. The runoff coefficient for rainwater harvesting is highest on bare soil, therefore it is necessary to plant less/ remove some vegetation in order to collect the largest volume of water from rainfall. However, farmers also want to keep vegetation in order to prevent soil erosion and keep the soil in good condition. At Interviewee As farm there is an 11 hectare catchment area and 6 hectares of farmland. A large proportion of the soil is left bare to give high water runoff. Without implementing rainwater harvesting on his farm, he states that only 3 % of rainfall would infiltrate and recharge the groundwater, but with rainwater harvesting in place 26% now infiltrates and recharges. A study by the Institute of Hydrology in Wallingford, UK, conducted the Romwe Project in the mid-1990s and this work found that the larger the catchment used for rainwater harvesting, the less effective rainwater harvesting is in term of the improvements to yields observed. Small scale rainwater harvesting has a small footprint and can be implemented by normal people and has few negative impacts on others in the catchment. Rainwater harvesting is not a one size fits all approach and often doesnt work well when done on a very large scale. The ACRU model looks at the consequences of dam installation at the macro-scale. The implementation of rainwater harvesting is can be more problematic in dry areas due to lower levels of rainfall, but its use is more important as water is a precious resource and needs to be managed very effectively. Even the collection of very small volumes of runoff can produce significant results on farms. Some ecologists argue that some soil types are unable to support annual crops; therefore it may still be necessary for farmers to concentrate on traditional perennial crops despite

increased water availability when using rainwater harvesting as it is easy to destroy the delicate soil ecosystem. 86% of Africa is currently grassland and a large percentage of this should be left as such. It is important to choose the areas to implement rainwater harvesting systems for irrigation carefully as not all areas are appropriate. Pereira (1973) talks about the effect of road construction in Africa on soil erosion. Some of the most severe soil erosion in Africa is as the result of the development of roads and more intelligent design is needed, such as the incorporation of rainwater harvesting systems to control and utilise the runoff from the roads. The use of rainwater harvesting is all part of a process to encourage communities to use the resources available to them more effectively. Roling has developed a catchment education tool to build a platform for this to occur. J. Jiggers also works in the field of farming extension work.

Joint Interview with Interviewees B, C, D and E 22/07/2010 C Farmers attempt to make small earth dams for rainwater harvesting independently, but they do not provide adequate water for irrigation due to the changing rain patterns in Botswana. Traditionally there are two periods of rain, one in the beginning of September and one in mid-June, but this trend is now changing and the rains are becoming more unpredictable. After planting and harvesting people tend to move away from their farms and back to their villages as the small scale dams that provide water for irrigation are dry and so no new crops can be sewn. These dams provide water for an average of 2 months. B Some pans used to harvest rainwater are located far away from farmers. There was a government scheme in the past to provide a roof rainwater harvesting system, including guttering and storage tank, but farmers had to provide funds for this and so the scheme ultimately didnt work due to farmers lack of financial resources. Work to encourage farmers to use rainwater harvesting was also done by the Botswana Rainwater Harvesting Association in conjunction with other NGOs, but insufficient funding lead to the collapse of the NGO. B The government scheme to provide small scale dams to groups of farmers is not entirely appropriate as farmers in the syndicate cannot be from the same family, therefore it is not entirely effective as people do not work together outside of the family unit in Botswana. People do not even talk to their neighbours on a day to day basis, it is only politics and religion that bring people in Botswana together. C Immediately after independence there were cooperatives in place but these later collapsed and there is no longer anything to bring people together. Politics and uneven distribution of welfare have divided the population. B People in the rural areas do not have enough money to get to the field or build a shelter there. Things are difficult especially for the young people, although help is received from CEDA. There is a fund especially to help young farmers by supplying them with loans, but collateral is needed to obtain any credit. Aside from this, most young farmers also lack the necessary management skills to start up on their own. D Another problem young farmers face is a lack of land availability, as elders are continuing to hold on to land rather than pass it on through the family. C Land goes from generation to generation, but no one actually owns it. In Botswana, it is only possible to apply for the rights to a piece of land after the age of 21. This causes a big problem of unemployment and crime amongst young people in the rural areas. In addition, applications for land can take up to 10 years to process from their experience. B Citizens from the surrounding area come to the Permaculture Trust for help, but the NGO is collapsing due to a lack of funding and assistance from volunteers. The NGO used to depend on donors, but these have now pulled out and there are no longer any funds to build the capacity of the NGO. After independence donors became increasingly aware that funds were being channelled inappropriately by the government and so stopped providing funds. Botswana achieving middle income status also lead to an outflow of donors.

C The situation is the same for all Batswana. D No form of training is available to young people who wish to enter the field of agriculture for a living. Three to six weeks of training are provided through government funds and groups are taken to the rural training centre, but all training received is theoretical and no practical hands-on experience is included as part of this. C Borehole drilling is very expensive and in general there are no boreholes on farms or at cattle posts. To sink a borehole 1m it costs approximately 230P and in general a depth of 50m is needed before the water table is reached in this region. Problems are caused for the rural people as they must pay the driller their fee for drilling the borehole whether water is reached or not and so farmers are reluctant to get boreholes drilled due to the risk that they have to pay for a borehole they can not use. It is possible to have a survey to determine at what level the water is before the borehole is drilled, but this costs 10,000P on average. The Pantamatenga Plains are the area with the most water in the country and a great deal of money was spent by the government to improve the area and prevent regular flooding from destroying crops. However, the farmers benefiting from this scheme were mainly foreign and the local farmers gained few benefits. D There are many problems with infrastructure. Public transport is not reliable, farmers often have no money for public transport, there is no water source to a large proportion of agricultural land and some individuals are unable to gain access to a plot of land suitable for farming. B The government does provide extension workers in the rural areas, but they are not provided with transport and so find it difficult to do their job effectively. D Extension workers can take up to a month to give help and guidance once asked for assistance. B The seed saving scheme in the past attracted a great deal of farmers to the Permaculture Trust, as it gave farmers who lost their crops in the droughts a chance to grow crops again the next year. However, the fields were not protected from the animals and the amount of rain received reduced so farmers began to leave their fields. C When people look at improving farming methods they neglect old knowledge eg the introduction of using tractors over donkeys. B - Hybrid seeds are now encouraged but farmers must buy these each year. C The government buys seeds from abroad and many of these do not grow when planted by farmers here. Farming before independence was much easier. D Around the Cheche river the government sunk a large number of boreholes, but applications must be made in order to gain access to land for agriculture in this area. B There are no special government schemes for women headed households, although the is a government department dedicated to womens affairs.

D The main problem experienced by farmers in this region is with respect to water and water scarcity is the main issue driving people away from small scale agriculture. B people are now bringing cattle into villages where water sources are located and are letting their cattle drink from stand pipes or cutting pipes in order to divert the flow and gain access to water for their livestock. A lack of water is causing conflict and people would rather pay fines issued by the council from having their cattle in a village rather than leave them at the cattle post. People have to pay a fine to the council, but the victim of damage done by the cattle receives nothing. D The government suggests that farmers should form a syndicate in order to put in an application for a borehole or a dam, but this does not work well in areas where there are a mixture of arable and pastoral farmers. B The government used to provide food baskets to farmers, but then suddenly stopped ths programme and told farmers to go and plant. C- Ponds cannot be used as a source of water for crops as the rain cannot be depended upon these days. Land boards are now allocating land for planting on paths through which water used to flow into ponds and so the degree of runoff reaching the ponds is reduced. D It is not possible to build your own pond and permission has to be applied for. C The government will state that they are building a pond, but then many delays are experienced and it can take a long period of time before one is actually excavated. The government will eventually excavate a pond and fence it to prevent livestock from reaching the water, but the water within it will be designated to a certain area of land. And if it is found that the water is being taken and used on any additional land the farmer concerned will receive a fine. C Some people are benefiting from government pond construction twice, as they can use the water for both their crops and livestock. More help is provided for livestock farmers than arable farmers. D Prices for maize are very low (50kg for 80P) and so most farmers are moving towards cattle farming where prices are higher due to the BWC scheme. D Water needs to be transported from other areas to use for watering crops. E Some farmers have the idea to use rainwater harvesting for crop production, but they have no funds to hire the necessary equipment or to fence off the required area of land. Rainwater harvesting systems worked in the past as farmers cooperated, but this is no longer the case. D There is no longer adequate rainfall to make rainwater harvesting feasible and a great deal of water contamination occurs. C Health departments discourage people from using surface water as it is dirty and children are being taught that the traditional ways are bad. Lots of traditional stories exist to

encourage people and condition them for the environment in which they live, A big conflict now occurs between the traditions and more developed ways. Modern living is not designed to work in conjunction with traditional culture. Extension workers will provide information on new schemes being introduced, but there are just 'sweeteners' and have no real impact B Many programmes are implemented that come from other countries and are not adjusted to Botswana. For example, people were paid to destump their land in the past for cattle feed, but this lead to a great deal of soil erosion. Permaculture is now encouraging farmers to replant their land and provide contouring, which has stopped erosion and improved the land. C If one thing is encouraged in one area and proves successful, it will automatically be replicated in another area by the government, despite varying geographic conditions. B Farmers do not have a voice in Botswana; there are no farmers forums in existence or representatives in parliament. People in government do not care about the farmers and there is poor infrastructure for communication between farmers themselves. There used to be a cooperative for educating farmers, but this is no longer in existence. Bureaucracy has made it difficult to get funds to support such schemes and issues are never taken to parliament. C If the government was organised differently maybe this would help improve the situation for farmers.

Interview G - 26/07/2010 A lot of work has been done on macro systems, but not micro systems. Levels of evapotranspiration are also high in Botswana. In Botswana the main problem is that rainfall is erratic and unreliable. Irrigation is a demanding enterprise and rainwater harvesting may be able to provide significant water for supplemental irrigation, but not enough for total irrigation allowance. Sometimes intense storms are experienced and using water from this you may be able to increase hectaridge on a farm, but then no water is received for the next three months. There may not be enough water to irrigate crops using groundwater, so rainwater harvested may be used to top up water for irrigation. To depend entirely on rainwater harvesting to provide water to crops may be a constraint. A tank to hold 5,000l of water would cost approximately 3,000 Rand. The government has just been informed that they will be looking into the supply of rainwater harvesting tanks, along with the provision of a small vegetable patch. The cost of the whole package will be 9000 P.

Interview H and I 27/07/2010 The dam is 7-8 years old. The Agricultural Extension workers are responsible for inspections The fence meant to be preventing cattle from entering the reservoir area and watering at the side of the water is broken in several places; the current situation of cattle in the reservoir area is against the government policy. The dam construction include a valve chamber and pipes leading to feeding troughs downstream of the dam, but a large amount of seepage has occurred and cattle can now water from any location downstream of the dam. In one of the National Development Plans by the Government a focus was put on repairing the earth dams already constructed. Some structures built under this scheme are very old and need rehabilitation. In some locations the farmers had to be paid for their labour, which should have been given in kind, as otherwise the project would not have been able to be completed. Each farmers syndicate has to have its own constitution and organise annual subscriptions to pay for maintenance of the dam. Poor management by farmers causes the biggest problems in these projects. Farmers are unable to fix the seepage problems themselves and need to contact the government in order to get the work done.

Interview J 29/07/2010 There has been no formal review of the ISPAAD programme. Under past schemes rainwater harvesting was encouraged among farmers through a package to provide small scale earth dams to individuals, but these reservoirs often did not last for the whole season. Now, as an alternative under the ISPAAD scheme, boreholes are provided to syndicates of farmers for irrigation and this is more suitable as it provides a source of water for the entire year; it Is a more reliable source of water. This is also better than the dam scheme as time is saved by removing the need for farmers to commute between the field and water supply. Group projects are also better as they provide more water to more people. Group projects are more value for money for the government and this approach is based on money rather than anything else People used to have to transfer water from their village to their fields for watering their cattle, therefore the main reason to encourage the formation of additional water sources onfarm is to reduce this commute rather than due to climate change. Making people work together is against human nature, people want to do things individually, therefore group schemes are more challenging to implement successfully, but it doesnt prevent it all together. The ISPAAD scheme has been running since 2008 and is now in its second year. Finances are the most important thing to consider. Rainwater harvesting was incorporated into the ALDEP scheme to provide watering for donkeys and cattle and not for irrigation purposes. ISPAAD is encouraging mechanisation and so therefore rainwater harvesting is no longer considered necessary. Most people already have rainwater harvesting for cattle due to the ALDEP scheme. This ALDEP scheme consisted of several components. Farmers were encouraged to construct both under and over-ground storage tanks to collect water harvested for the roofs of their property. ALDEP 2 evaluation has been completed, but ALDEP 3 still in the process of being carried out.

Interview K 25/07/2010 Interviewee K has lived and worked in Botswana for a period of 35 years. He established WildFoods, a company that produces food stuffs from fruits and other natural foods grown locally. Water is currently from runoff on the hillside behind the farm and there are plans to install a system for harvesting runoff from the roof of the property. In one area a zig-zag formation is currently used to harvest rainwater for the irrigation of lemon trees. This is combined with the placement of mulch at the base of the trees. Shade is required from the leaves of the trees to provide shade and stop the microryzer in the soil from being killed. This combination of mulch and the creation of a micro-climate is needed to obtain the best yields when using rainwater harvesting. Another rainwater harvesting technique was used as part of an agro-forestry project, which was completed several years ago. This rainwater collection system consisted of troughs created around each trees, between which raised semi-circular bunds ran parallel to the contours At regular intervals, longitudinal semi-circular bunds were then placed to divide up the contour bund into several sections. This system was set up across the whole field in which the trees were grown. In the large gaps down the slope between the trees, more contour and longitudinal bunds were constructed; to collect additional rainfall runoff. In the spaces immediately upstream of the bunds running between the trees maize was planted. From this project Interviewee K observed that he could plant 20% of this field with maize, concentrating the plants in a 1m wide strip upstream of the contour bunds and could obtain the same yield as the average subsistence farmer planting the whole field with maize; this is as a result of the rainwater harvesting system used, which increased the volume of water reaching the crops substantially. Another benefit observed from using this rainwater harvesting system was that the absorbency of the soil was increased in areas where planting had taken place immediately upstream of the bunds. Modifications were made with this rainwater harvesting system to try and prevent microrisen in the soil from being killed by high temperatures by planting lemon trees to provide shade, but this reduced the amount of runoff significantly and so these were later removed. Interviewee K finds that the systems of rainwater harvesting used in his fields are reliable and provide adequate water for irrigation even with the high levels of rainfall variability experienced in Botswana. A significant degree of manual labour is needed to set up the system , but then labour is saved during planting as only 20% of the area needs to be cultivated; the systems is also easy to maintain with no significant costs involved. No storage tanks are used as part of the rainwater harvesting system. Other local farmers have been brought to Interviewee Ks farm to learn about the rainwater harvesting system, but farmers are conservative and are not keen to start new practices that are not traditionally used. The harvest from Mr Taylors fields using only rainwater harvesting for irrigation is virtually guaranteed from his experience, whereas the average failure rate for farmers growing rain-

fed crops in Botswana in 1 in 3. The Ministry of Agriculture do not do anything in practice, there is a great deal of reinventing the wheel going on, but this is not effective as farmers in Botswana are conservative and like to use tried and tested methods that their fore-fathers have been using. Rainwater harvesting is not a traditional practice in Botswana. Not all crops can be grown with the use of rainwater harvesting for irrigation, such as some forms of green leafed vegetables 35-40 years ago there was an organisation called Serowe Farmers Brigade. Which encouraged the use of haffirs to store water, but Interviewee K is not aware of any organisations in existence today that encourage rainwater harvesting in the country. Interviewee K does not see a future in the use of rainwater harvesting for irrigation in Botswana as there is no pressure for people to grow their own food; people receive too much help from the government and so have stopped farming. The average annual yield of grain per hectare in the region is stated by Interviewee K to be 168kg, whereas a commercial farm can yield 40,000kg per hectare per annum. It is for this reason that the government are more focussed on increasing commercial agriculture than supporting small scale farmers. Also, farming is no longer seen as a career choice in Botswana and it is now the culture to have a white collar job. Interviewee K does not have the view the lack of farming in Botswana is due to laziness, they are traditionally pastoralists that have been subjugated by other tribes and marginalised, and are no longer able to live in their traditional manner. Traditionally Batswana live season to season, they have a slow pace of life, but now they must live in a different dynamism and a consumer society has developed. The scheme by the Ministry of Agriculture to provide dams to farming syndicates works, but there are many conflicts that occur. Also, these dams are just for the provision of watering livestock. Problems also occur with soil degradation as a result of trampling by cattle and sedimentation rates are high in the dams. Interviewee K is of the opinion that rainwater harvesting cannot be used in areas where cattle grazing is occurring as flat land is needed to maximise rainfall runoff; if cattle are graving on land they will cause it to break up and significantly reduce runoff volumes. If rainwater harvesting was to occur on a large scale farmers would have to fence their farms to prevent cattle from wandering onto crop land and this will not be feasible in Botswana. Rainwater harvesting is a good idea in Botswana in theory, but not good in practice due to the high number of cattle farmers. Sand reservoirs and hafirs can provide good sources of harvested rainwater. Hafirs are used extensively in northern Africa and the middle-east, but Interviewee K is sceptical as to whether they would be successful in Botswana due to the high degree of manual labour involved to construct them. Vernon Gibberd is currently involved in the implementation of vegetable garden schemes in Botswana. Trials have shown that a plot of 40m3 can yield enough produce to feed an average Batswana family for a year. In practice Interviewee K feels that this kind of scheme would not work if up-scaled to the basin level; as not enough people would be interested in

carrying it out as will not create any money for the families, and this is the main driving force for the Batswana population. However, in northern Namibia, the situation is very different; there are many womens groups who farm a great deal of marula. With Interviewee Ks rainwater harvesting system, he observed improvements in the yields immediately, during the first growing season. The absorbency of the soil was found to take longer to develop and improvements were observed after the second season of planting. A large amount of practical work has been conducted in the Karoo region of South Africa. This is a very flat region, where bunds are placed in river beds to facilitate the overflow of water out of the river channel and into fields located on the flood plains. These fields are prepared to receive water during the dry seasons when several bunds are constructed to control the flow of water through the field; large volumes of wheat are grown in this way.

Interview L - 21/07/10 Many small dams have been built by the Ministry of Agricultures Water Development Section, primarily for the reason that farmers in rural areas do not fall under the jurisdiction of the Ministry of Water Affairs. The dam scheme is mainly to provide water for livestock farming, although some of the larger structures have an allowance for irrigation too. Most of the dams constructed are still in operation and are maintained by farmers, who have the responsibility of installing a fence around the dam and reservoir in order to prevent livestock from damaging it. Botswana does not have very strong farmers associations. Interviewee L is unsure if this is due to a absolute lack of financial resources or not, but suspects it is most likely due to a lack of organization that makes it difficult to organise the finances between the farmers; group organisation is a problem across the whole agricultural sector. Finances are also unlikely to be an issue as the vast majority of the costs for the dams are covered by the government and only 1% is contributed by the farmers. However, maintenance is thought to be an issue. A new scheme is to be introduced by the government which will help to ensure that all sectors can benefit from infrastructure projects. In Thune new dams are being built that will be large enough to provide water allocation for irrigation as well as livestock (unlike most other dams that have been built to date, which are for livestock watering only). There is currently a policy shift in water resources management by the government towards the concept of integrated water resources management and the aim is to provide water for all communities within Botswana, even in the rural areas where communities previously had to provide water for themselves. A new scheme for ISPAAD will assist farmers to drill boreholes for use in pastoral and arable farming. Previously farmers had to organise their own access to water, but now farmers will be given their own boreholes under this new scheme. Well development is also another scheme being introduced in areas with low head/high water table, particularly in areas close to rivers. Shallow dug wells are being installed together with a range of different lifting mechanisms depending on the site, all of which is funded by the government. Alternative sources of water in rural areas are sand rivers and this is where people have accessed water traditionally, but now problems with sand mining are occurring close to the larger cities as large amounts of sand are required for construction. In the Kalahari area pans provide a source of water as they collect rainfall. In the past runoff farming was encouraged by the government, but now there are issues regarding the quality of the water and also the volume of water collected is not enough to provide for irrigation due to low levels of rainfall, even if roof tops are utilized for collecting rainwater as well as land. An average of 4.8-5 mm of irrigation is needed per day for most crops and rainwater harvesting will not provide enough for this. In the past water for irrigation has been obtained from groundwater and mulit-purpose dams. A mixture of large and small scale interventions are needed for a good watershed

management strategy, with small dams used to augment groundwater supply. Claims were made in the past that the installation of small scale dams for farmers intercepted the flow into the larger dams, such as Gaborone dam, which provide water for domestic needs to large populations. However, these claims have been shown to be ill founded and the effect of these small dams on the level of the reservoir at Gaborone was found to be insignificant. The introduction of an integrated water resources management approach by the government will help reduce problems of water availability for irrigation as stakeholder participation will increase across all sectors and people will be informed of progress from a lower level through decentralisation. The key to solving the current problems is participation. The current schemes involving the formation of cooperatives among farmers are not successful because people in Botswana do not like working together. Farmers may pretend to be working together in order to receive funding or have a dam built, but then once this is achieved the syndicate falls apart and conflicts occur. Estimates of rainfall in Botswana in the future should be investigated to determine whether levels are high enough to provide a significant volume of water for irrigation if harvested. Central and local government, along with village level governance are involved in solving issues around water resources management.

Interview M 11/08/2010 In-field rainwater harvesting has been the most effective of the methods of rainwater harvesting trialled in South Africa as it is very simple and duplicable. The only problems experienced have been with respect to institutional arrangements, which involve support and extension to the community. This is one of the most important issues when working in developing countries as the politics, internal policies and social structure are more important than is often realised. Leadership within a community is very important and can have a huge influence on the success of a project. Demonstration of techniques to farmers was found to be very helpful in changing the attitudes they had towards rainwater harvesting. However, many people who were originally using integrated rainwater harvesting techniques implemented as part of Interviewee Ms most recent research project are now no longer using the technology. Interviewee M thinks this is because it is easy for individuals to gain access to grants and there is a large degree of dependence on the government. Farmers in these areas move from project to project, trying different strategies for a short time, rather than sticking to one strategy and making it work for them over time. Boronto is another rainwater harvesting research project that is being run in conjunction with WRC in South Africa. Lack of institutions has appeared to cause problems with implementation of projects Interviewee M has worked on. The focus of research in South Africa is currently on in-field rainwater harvesting methods as non-field methods are expensive to implement and maintain. In some cases, in-field and non-field techniques have been combined, but non-field techniques have not been used on their own. These schemes are useful as they allow larger areas to be irrigated, but using rainwater harvesting for full irrigation may be a problem in semi-arid areas. Pastoral and arable farming can work together well as crops can be grown for cattle, which then in turn provide manure for the cultivated land. However, difficulties can occur due to the need to fence off cultivated areas. In the Eastern Cape (South Africa) rainfall is high and in addition to planting food crops such as maize, farmers plant fodder crops. However, more than 700mm of rainfall is needed to do this and the fodder must be cut and carried to the cattle, they are not allowed to graze on the crop land. Animals are always part of the farming system and in some regions it is normal to let cattle onto the arable land to graze, but this is not possible if in-field rainwater harvesting measures have been implemented. In Saaidamme ex-field rainwater harvesting has been carried out and water is used for flood irrigation for lucern. Channeling water directly as it falls and flooding fields is another option and can work better for farming in some areas, such as Karoo where less than 250mm of rainfall is received annually on average. There are techniques that can be used in conjunction with livestock, but these are not really

implemented, as it is advised more often on planted pasture using a conservation technique Before the implementation of any rainwater harvesting scheme details of the natural resources in the area must be known, along with long-term rainfall. The appropriate rainwater harvesting technology then needs to be matched to the site. Observations must be carried out in the first year to determine the impact of implementing the system. The farmers must then be consulted and a decision should be made together with the researcher as to which technique works best. However, it is difficult to get farmers to change their way of farming. Traditional knowledge has led to the development of Interviewee Ms research as it was farmers that suggested the introduction of vegetable to the crops tested, as these are traditionally grown in the area. Normally croplands are not right next to a farmers house and so demonstrations were initially unsuccessful due to the distances that had to be travelled. However, when demonstration sites were moved to backyards, the project was much more successful. In South Africa, most individuals in rural areas are not active on their croplands. There are large bio-physical challenges in the form of poor soil, inadequate rain, lack of maintenance to land and lack of capital for investment. One community visited in Interviewee Ms trials had not cultivated their land for 25 years and a large amount of theft and damage is occurring in the fields where farmers are now choosing to plant crops.

Interview N 28/07/2010 The problem in Botswana is that it is a semi-arid climate, with low annual rainfall. 500mm of rain is received in a good year, which is a lot of water if managed correctly. A large proportion of the rainfall normally comes in the form of short intense storms and as a result a large amount of the water is lost as runoff. If the water is collected into reservoirs and is left open then in the high summer temperatures large amounts of evaporation takes place. It has been shown that the Gaborone dam loses more water from evaporation each year than is used. The methods of storing the water need to be considered carefully. Interviewee N has a 10,000l of water storage at his home and is able to water vegetables and fruit trees using this. Now it is more or less government policy that when a school is built a rainwater harvesting tank must be installed. Tanks are normally provided for each building within a school. Small dams are not a good method of storing water from runoff due to the high evaporation rates in Botswana. An irrigation put into place to use water harvested should also be carefully considered and should be as efficient as possible, such as a drip irrigation system. In Botswana people are not willing to work. Conservation farming is now being encouraged, which involves no till, crop rotation and permanent soil cover. In Botswana the challenge is not the no till option as little work is involved, as in Botswana farms are very large the land board give 10 hectares to farmers as yields per unit area are not large. With conservation farming leaving the land helps the soil and will increase yields. Digging reservoirs involves mechanisation and is therefore no feasible for farmers with just cattle for power. There is definitely lots of resistance from farmers for using rainwater harvesting practices. Rainwater harvesting is not a traditional practice and so farmers of today do not do it. People do not think that rainwater harvesting is necessary and to some extent the government can be blamed for this. Every year drought is often declared by the government and drought relief schemes are started. Instead people should be made to be more proactive. The government ALDEP scheme was aimed at assisting small scale farmers and increasing their productivity. Rainwater harvesting was a small component of this and resources were provided for the construction of an underground reservoir. This water was used during the dry season. However, these ideas were not extended past the pilot projects. The ALDEP scheme has been replaced by the ISPAAD scheme run by the Ministry of Agriculture, but the rainwater component is not included within the new scheme. In Botswana rainwater harvesting should be placed as a top priority, but in order to accept these new innovations the farmers need to be pushed and encouraged and the government is not currently doing this. ALDEP was not extended past demonstration farms and although farmers were taken to the sites in order to be shown how to implement the technology, they did not replicate the scheme on their own farms. Where Interviewee N was raised there was a well that collected runoff from the surrounding fields. He is unsure why this method of water collection is no longer used. New farmers are disconnected with old farmers and knowledge is not passed on.

In northern Botswana every farm would have a similar well, but when the government began distributing water people began to neglect their wells. The government has provided new infrastructure for accessing water and so people have abandoned their traditional ways. The traditional processes need to be restarted. The presence of a large number of cattle in Botswana is a disadvantage for the use of rainwater harvesting, but the government is trying to encourage fencing to be incorporated into any project they carry out. However, there are still many fields that are not fenced and animals move freely. Stub and mulch should be introduced to reduce evaporation from the soil, but most of this is used for cattle feed; arable and pastoral farming have competing needs which poses problems in Botswana. In general Botswana has a system of range grazing and cattle roam freely to each grass, but in years when rainfall levels are low and grass is sparse, cattle are forced to eat stumps and mulch. There is a shortage of grazing in most years in Botswana. The government scheme for dam construction is not very appropriate as the dams silt up rapidly and large evaporation losses take place from the reservoir. The capacity of the dam reduces annually due to siltation and makes the problem of evaporation worse. A more structured management scheme is needed, as the syndicate scheme creates problems where people do not take ownership of the dam. In a borehole syndicate that Interviewee N is part of there are 10 farmers, but only 3 actually actively manage the borehole. Farmers want to use the dam but do not want to, or are unable to, pay for the subscription fees; a sense of ownership is missing. These systems only work where there are a few dedicated farmers within the syndicate who manage the system, in all other cases it breaks down. Perhaps the government has allowed people to depend on them too much and people now expect them to do things for them.

Interview O 30/06/2010 The NGO was founded in 1991 by three lecturers at the University of Botswana. The NGO concentrates on environmental awareness, organic gardening and resource conservation. ST run many resource conservation projects and the idea is that people form the local community will visit their ecological park and replicate what has been done there. Schools have been involved in an organic gardening and tree planting scheme. Work has been carried out with water conservation and communities have been encouraged to harvest water from their roofs to use for irrigation in small back garden patches, as well as for domestic purposes, such as dish washing. Some people have had tanks installed in their back garden. Lobbying for change has not been fruitful due to the government systems. Using sand rivers to harvest water is not very feasible as people excavate the sand for their personal use and break the natural path of the water and so less runoff reaches the dam for storage. A backyard garden project was taken out to the communities by ST and tanks were provided to inidivuals for use to store water. People were keen inititally, but then in many cases when representatives from ST returned the gardens had been paved over. People are too lazy and are reluctant to take up initiatives introduced as they assume the government will provide for them. Funding for such schemes is also a constraint.

Interview P 27/07/2010 The structures built under the government scheme providing farming syndicates with earth rainwater harvesting dams are the responsibility of the farmers to maintain, however, initially no structure was put in place for monitoring the dams and farmers thought that these were the property of the government and so made no attempt to repair them. The government became aware of this situation and then started to train farmers in how to maintain the structures. Initially the dams were built in areas where the government decided there was a need, but since approximately the year 2000 the process is demand driven and dams are built in areas where farmers syndicates request for one. The dams initially built by the government were the property of the community rather than a group of farmers and in these situations no funds were gathered for maintenance. When the dam scheme was changed to being demand driven, conflict occurred between farmers over the required annual financial input to cover maintenance costs; not all members of the syndicate wanted to contribute funds or help with the construction. Problems also often occur due to siltation which takes place very quickly due to the fact that cattle are not drinking at the designated drinking trough, but at the edge of the reservoir, causing soil around the perimeter to enter the reservoir. Every group that have a dam constructed for them under the governments scheme receives 1 weeks training; this training occurs in several stages, before, during and after construction. Topics covered included group dynamics, maintenance and fencing. Dams built by the government before the scheme was demand driven experience the greatest difficulties in being managed and maintained by village groups. For these dams, a kotla, a traditional Botswana community meeting, is held to elect members of a dam committee, which operates on a rotational basis. Group owned dams are more successful than the community dams, as most conflict arises in relation to the financial contributions. In community owned dams people often refuse to pay the subscription fee as there is no proper control over the funds, so there are often mis-used. However, the more recent farming syndicates have a bank account for their funds and so members are more assured that their money will actually be used for its allocated purpose. It is difficult for people to run-away from the dams up in the north as no boreholes are in the area and the dams can be the only source of water. In these areas the Extension Workers liaise with local authorities, chiefs, local Councillors and Agricultural Demonstrators in order to develop the most appropriate management system. In these areas the villages are small and the dam is named after the village and hence chief of the village. This can create problems as it is difficult to dis-associate the dam from the community and ensure it is used for its intended purpose of providing water for cattle. In Botswana, the culture is that if something belongs to the community then the whole community should receive benefits, but community ownership also means that people are less willing to put work into maintaining the dam. Additionally, the community-dam relationship can lead to problems in management as personal relationships can lead to the most appropriate person for the committee positions. Group personal conflicts are drawn into dam management. The Botswana Government is

trying by all means to move away from community dams. Multi-purpose dams are government driven, these have a small irrigation field next to them on which crops can be grown to utilise the water allocation from the dam. Some dams provide water for the whole year, some not. In the western parts of the country pans (natural and man-made) are used as a source of irrigation as they collect a large volume of rain and they are capable of providing water for the whole year. The dams constructed by the Government are capable of holding between 1.3million and 200,000 cubic metres of water. The exact volume of water stored at each location depends on the type of terrain, which dictates the size of the dam. In some areas levels of seepage through these dams is very high, whereas other dams hold water very well and volumes are sustained to such a level that fish can be supported in the reservoir. In reservoirs large enough, the Government is introducing fish, which provides communities with an additional source of income. It is hoped that by providing training to local communities on fishing practice the communities will have more ownership over the dams, although only approximately 10% of the reservoirs are large enough for this to be carried out. In newer dams owned by a farmers syndicate and with fencing around them, water appears to hold in the reservoir more and water quality is greater. In recent years a different rain pattern has been experienced each year. Traditionally rains would arrive in September; seeds would be planted and then harvested in March/April; another short period of rain would then occur in August, but crops would not be planted again until the rains arrived in September. Now significant volumes of rain sometimes fall in the winter months (June-August); rainfall is more erratic with no observable pattern. In some areas the traditional rainfall pattern is still present, but is delayed or heavier than usual. Locations suitable for the placement of dams are not running out, the main problem is actually with regards to the machinery used. One of the small earth dams used to take approximately 3 months to build, but now it takes 6-7 months due to frequent machinery breakdowns. The machinery used for the construction works was acquired from the UN in the 1970s. 4 dams a year used to be built, but this has now reduced to about 2 per year today. The Government used to provide fences to farmers as part of the small earth dam scheme, but now farmers must provide these themselves, as it is thought by the Government that this increases the ownership felt by the farmers over the dam. The Government does, however, provide fencing to the embankment to prevent cattle from walking over this and causing structural damage. The Government is trying to enhance ownership of the dam by the farmers in several ways during both construction and maintenance. The current institutional structure is working, but is in need of a lot of improvement. The Agricultural Demonstrators live in the regions with the farmers they are responsible for, but they are agronomists and not engineers. Agricultural Extension Officers provide the Agricultural Demonstrators with the application form necessary for the earth dam construction. Before a dam can be constructed the farmers need to produce a constitution for their syndicate and possess land rights to the location in which the dam is to be built (farmers

need to apply for land rights separately). In the past the Land Board has provided the farmers with land rights and then subsequent inspection by an Agricultural Extension worker has shown the area of land concerned not to be a suitable site for dam construction. In order to improve the efficiency of the system the Ministry of Agriculture has advised the Land Board to communicate with the Misitry of Agriculture to see if permission for dam construction has already been provided before land rights are given. In some cases problems have occurred where not all parties owning land that will be flooded by the construction of a dam have been consulted before land rights are given to the farmers. In these instances the Government has had to provide compensation. Some farmers collect rainwater from their roofs and store it in tanks. These received the equipment to allow them to do this as part of the ALDEP scheme, which has since been phased out. Hand dug wells are now provided as part of the ISPAAD scheme. These are concrete lined and a push pump is provided as part of the package. The owner of the land who is receiving the well from the Government is required to pay in kind for the construction, but this requirement is not specific enough and the Government wants to change the policy to something more workable. The ALDEP scheme was phased out for political reasons. A lot still needs to be done to provide people in rural areas with water. The north western areas of the country have the highest level of water scarcity and most people in the areas depend of hafirs for additional water. These hafirs were mostly constructed as part of a drought relief project with the intention of harvesting sufficient rainfall to irrigate crops. Traditionally people in this country have been digging small hafirs on their own properties and drought relief schemes have provided assistance in making them larger. The local government now pays communities to make their hafirs larger, but this is a politically driven initiative and is not consistently implemented. The Government needs to target where rainwater harvesting should be carried out. In the east of the country dams can be constructed, but this approach is unsuitable for the central region, where boreholes are a more appropriate means of obtaining water. The digging of such boreholes is expensive for the poor. The only rivers in the country are located to the east; in the west the terrain is more sandy and the only source of water is pans that can hold water for approximately 2 months. These pans are very open and water collection and allocation here is not controlled