Participatory Indicator Development for Sustainable Natural Resource Management
Andrew Dougill and Mark Reed
School of the Environment, University of Leeds, Leeds, LS2 9JT
Contact:
Abstract
There have been few attempts to develop, test or apply sustainability indicators for natural resource management at a farm-level, and even fewer have fully consulted land users from the initiation of this process. While environmental sustainability indicators need to be accurate and reliable, land users also need them to be rapid, cost-effective and easy to use, in addition to being linked to management objectives such that they can be targeted at community empowerment and sustainable natural resource management. In order to develop farm-level sustainability indicators that meet these criteria, a framework for participatory indicator development has been constructed that integrates land user knowledge with biogeographical information. Integration of qualitative and quantitative approaches incorporating biogeographical evaluation of land user indicators in the context of initial livelihoods analyses may lead to a more concise and useable list of the most effective indicators. Case study material shows that Kalahari pastoralists in South West Botswana use over eighty sustainability indicators for natural resource management. However, indicator knowledge is sparsely dispersed and varies between different social groups in the community, and consequently there is little consensus over which are most effective. Dissemination of sustainability indicator information is proposed through the distribution of rangeland evaluation guidebooks through the Ministry of Agriculture that integrate participatory rangeland condition assessment with management strategy advice. In this way, it may be possible to empower communities to effectively monitor and respond to rangeland degradation to enhance the sustainability of natural resource management.
Key Words: sustainability indicators, natural resource use, participatory approaches, environmental monitoring, Kalahari pastoralists.
Paper presented at Conference on Combining Qualitative and Quantitative Methods in Development Research, Centre for Development Studies, Swansea July 1-2, 2002.
Introduction: Sustainability Indicators and Participatory Research
There has been much recent debate in development and environmental literature about the ability of participatory research to provide both qualitative insights into livelihood options and decision-making controls together with quantitative assessments of environmental sustainability (e.g. Abbot and Guijt, 1998; Rigby et al., 2000; Morse et al., 2001). One way in which this can be achieved is to extend the largely qualitative, livelihoods analytical framework (Scoones, 1998; Carney, 1998), to identify quantifiable indicators of environmental sustainability with relevant stakeholders. Thus far farmer-led approaches have focused largely on deriving indicators of land degradation, rather than of sustainability, and have concentrated on mixed and arable farming systems (e.g. Rigby et al., 2000, Scoones, 2001; Stocking and Murnaghan, 2001). As yet, no farmer-led studies focus on the complexity of issues and environmental changes that typify the highly dynamic semi-arid rangelands (Behnke et al., 1993) that support the livelihoods of over 25 million African pastoralists (Lane, 1998).
If sustainability indicators are to empower communities, they must be simple, rapid and inexpensive, in addition to being credible, transferable, dependable and confirmable (Pretty, 2001). Further, to ensure acceptance within natural science debates and to maximise the impact on policy makers, it is vital that indicators can be viewed in a positivist manner as being accurate and reliable. As such, the majority of sustainability indicators for natural resource management have been highly quantitative. They have been identified, selected and applied by researchers and they often carry little meaning for local communities who require specialist training and equipment to use them (e.g. Tongway, 1994; NRC, 2000). As such, there is no accepted framework for participatory identification, evaluation and selection of sustainability indicators that can feed into rangeland management decisions, both on individual, community and institutional levels. A small number of recent studies have taken a more qualitative approach to indicator development, eliciting indicators from local communities (e.g. Bellows, 1995; Kipuri, 1996; Woodhouse et al., 2000), but have been unable to ensure accuracy and reliability. By integrating qualitative and quantitative participatory methods it is possible to combine the strengths of both approaches.
This paper reports on research with pastoralists in Kgalagadi District, South West Botswana. It aims to develop a methodological framework capable of extending qualitative livelihoods analysis to obtain indigenous knowledge on rangeland environmental changes and provide quantitative assessments of rangeland condition. It is proposed that the outputs of integrated assessments can then be used to enable pastoralists to monitor environmental changes themselves independent of Government extension workers. This participatory environmental monitoring could then empower them to discuss pastoral management decisions within their more informal social networks. The ultimate research aim therefore remains the empowerment of pastoral communities to manage their rangelands in a more environmentally sustainable manner. Consequently, this paper concludes with a discussion of how involvement of the key advisory and policy-makers in the research process can contribute towards such empowerment, which has rightly been identified as the key participatory research frontier to explore (Chambers, this conference). It is argued that involvement of farmers, extension workers and policy-makers in integrated qualitative and quantitative monitoring of environmental change will improve the opportunities for research leading to such community empowerment.
Background: Frameworks for Participatory Indicator Development
From Degradation to Sustainability Indicators
Despite the ubiquity of the sustainability definition in the Brundtland Report (WCED, 1987) and the popularization of the concept, there is little consensus over a precise operational definition and meaning. The meaning of sustainability differs in time and space, and between individuals. Without a precise operational definition, eliciting sustainability indicators from local communities becomes problematic. This may explain the absence of local participation in the development of farm-level sustainability indicators to date (e.g. Gomez et al., 1996; Rigby et al., 2001).
Operational definitions of land degradation are however well established (UNEP, 1997), and eliciting degradation indicators from communities is relatively straightforward (Stocking and Murnaghan, 2001). As the antithesis of sustainability, degradation indicators may be reversed to derive sustainability indicators. This research advocates the integration of quantitative approaches into participatory (largely qualitative) livelihood analyses to validate and short-list the most effective community indicators.
Existing Sustainability Indicator Frameworks
A number of frameworks have been developed to classify sustainability indicators with much emphasis on how to involve local communities fully without the opposite neglect of scientific knowledge. The most widely used frameworks employed by development agencies are the Framework for the Evaluation of Sustainable Land Management (FESLM) (Smyth and Dumanski, 1995) used by the World Bank, and the Pressure-State-Response (PSR) (OECD, 1993) and Driving Force-State-Response (DSR) (UNCSD, 1996) frameworks.
Although none of these frameworks are explicitly participatory, classification frameworks like these go further than the positivist scientifically driven attempts to define indicator evaluation criteria that have guided the development of many previous sustainability indicators (Bell and Morse, 1999). For example, NRC (2000) selected indicators according to predefined criteria and a list of factors that most strongly influence ecosystem functioning. Breckenridge et al. (1995) developed a framework in which indicators were identified by interdisciplinary teams of researchers, who subsequently evaluated them with reference to a list of predefined evaluation criteria and tested them with empirical research and then re-evaluated them. Rennie and Singh (1996) developed a similar framework, in which indicators were identified by researchers; however, communities short-listed indicators which became “community-derived”.
The development of pre-defined, externally generated evaluation criteria for indicators does not acknowledge the diversity of stakeholders with wide-ranging perceptions of relevant criteria. Participation of stakeholders in the development of evaluation criteria is therefore essential to select appropriate indicators. Evaluation criteria directly influence indicator selection, and are themselves influenced by the objectives for which users wish to develop indicators (Krugmann, 1996). Determining user objectives and evaluation criteria prior to the identification and selection of indicators is therefore a key step, but one that has been rarely addressed in indicator development frameworks. In the Kalahari research case study provided here, this is undertaken by differentiating analysis on the basis of pastoralist land tenure status. In such pastoralist societies, land tenure can be viewed a key element of an individual's environmental entitlement (Leach et al., 1997).
A few recent studies have taken the identification of sustainability indicators a step further towards empowering communities, by incorporating them into indicator-based management tools to assess land degradation (e.g. Savory, 1988; Tongway, 1994; Milton et al., 1998; Stocking and Murnaghan, 2001). However, community involvement in their development and testing has been limited and therefore empowerment has been negligible. They tend to have been developed with little reference to land user objectives, strengths or constraints, and to encompass a limited range of indicators. Milton et al. (1998) for example, provide one of the most comprehensive suites of indicators for farm-level assessment. However, their guide was restricted to vegetation and soil-based indicators, with limited reference to socio-economic indicators that are of greater importance to individual pastoralists. Similarly, Tongway's (1994) user manual for Australian farmers focused exclusively on soils and required considerable technical expertise. Methods based on chemical soil testing (e.g. Stocking and Murnaghan, 2001) are inappropriate to the nature of environmental changes in the Kalahari (Dougill et al., 1999). Savory's (1988) biological monitoring method is more appropriate in that it used a combination of plant, animal and soil factors as indicators of range condition, but the suggested data collection methods were too complicated and time-consuming to achieve widespread uptake by farmers. Consideration of socio-economic variables in assessments of environmental degradation (or sustainability), together with more applicable indicators of rangeland ecological changes, are both therefore required if more applicable indictors of environmental sustainability are to be provided for African rangelands.
Link from Livelihoods Analysis Approaches and Frameworks
The recognised need for holistic discussion on the impact of changes in the natural resources upon which pastoralists depend, matches the evolving focus of interdisciplinary livelihoods analytical approaches (Scoones, 1998). This livelihoods framework designed for systematic analysis of poverty has also gained widespread acceptance and adoption in natural resource related development projects (Ashely and Carney, 1999; Ellis, 2000). Consequently, the research framework proposed here for sustainability indicator identification evolves from more general sustainable livelihoods analyses that have been employed on DFID-funded projects elsewhere in this region (Ashley, 2000; Twyman et al., 2001; in prep.).
Livelihoods approaches can be extended using a range of participatory approaches designed to elicit both qualitative and quantitative information on the nature of changes in ecological communities and indicators. This information can then be used to develop indicators that can be employed to monitor changes in their natural resource base. Participatory approaches designed to elicit information on environmental change that can guide participatory environmental monitoring have been outlined in a number of previous studies (e.g. Abbot and Guijt, 1998; Scoones, 2001). Key methods include participatory mapping of different ecological resource zones around villages (e.g. Twyman et al., 2001). Mapping is often aided by aerial photographs and transect walks or drives with key informants (Chambers, 1997). Another vital method in such dynamic rangeland systems is the use of oral testimonies or histories (Slim and Thompson, 1993) with key stakeholders who can outline key transitions between ecological states (e.g. Kepe and Scoones, 1999). These can often be related to rainfall histories (Conway, 1988; Twyman et al., 2001). It is thus essential that a research framework for sustainability indicator identification integrates qualitative livelihoods analysis with appropriate quantitative elements of participatory environmental monitoring. It is to this end, that the methodological framework proposed here was applied in Kalahari rangelands of South West Botswana.
Research Design and Methods: Proposed Framework
Building on existing frameworks, the model proposed here stresses the need for farmer involvement at every stage in the research process (Figure 1 - subsequent numbers in parentheses refer to the research stage outlined in the Figure). Although the simplified diagrammatic presentation implies a series of distinct stages the key element throughout is the simultaneous use of the different approaches in order to integrate the qualitative and quantitative, and the environmental with the socio-economic.
The research framework starts from semi-structured interviews (1 -3) conducted with pastoralists and extension workers to identify the key livelihood constraints and the causes of poverty at a household level, together with discussion of the objectives that different groups wish indicators to meet. It is critical that local communities define the indicator objectives and the evaluation criteria against which indicators may be evaluated (1). Much of this can be achieved using a livelihoods approach to structure the initial phases of pastoralist interviews (2). In addition, interviews were extended to include discussion of what interviewees used as rangeland degradation indicators (3). The 67 pastoralists interviewed at this initial research stage were asked to identify objectives they would like indicators of rangeland degradation to meet. Where this was problematic, they were asked to identify their main objectives for farm management, followed by an assessment of the extent indicators could help meet these objectives. Consistent with accepted degradation definitions (Abel and Blaikie, 1989; UNEP, 1997), rangeland degradation was defined as areas of formerly productive rangeland that have declined in their agricultural productivity independent of rainfall variability. Pastoralists were also asked at this stage to identify which of these indicators they would expect to change first before a degraded state was reached. These "early warning" indicators were predominantly process-based and are vital to improving agricultural extension advice and enabling individual pastoralists to respond in time to prevent rangeland degradation.
By differentiating analysis of responses from the different groups within societies (based on a range of factors) it may be possible to better target outputs. The resulting list of parameters were also evaluated in community focus groups to assess their accuracy and ease of use as indicators (4). A number of criteria for evaluating indicators were elicited from farmers, though it is noticeable that "accuracy" and "ease of use" summarise the majority of them well (Table 1).
The next stage of the proposed methodological framework involves investigation of the degradation indicator short-list (5), using quantitative biogeographical and economic techniques. To achieve this, study sites are identified on a continuum from degraded to non-degraded (6), using economic criteria, notably herd size trends, consistent with the UNEP (1997) definition of land degradation. Trends are normalised for rainfall, and other factors influencing trends are explored in semi-structured interviews with farmers at each borehole. Degradation is only inferred where declining herd size is involuntary, and linked to rangeland condition. Changes in rangeland condition are elicited from farmers using oral histories. Herd size trends are derived both from livestock census data and from pastoralists' records. Indicators on the short-list are then measured at each study site using biogeographcial techniques, to determine which indicators best characterise the identified degraded sites (7). The community in further focus groups (8) may then evaluate this statistically ranked list before it is disseminated to other farmers, extension workers and policy-makers.
Dissemination should facilitate farm level application and thus aid decision-making by individual pastoralists. In this case, we propose to disseminate findings through rangeland assessment manuals, similar to those produced for the Karroo by Milton et al. (1998). Draft manuals need to be developed, incorporating information about management strategies that may prevent or ameliorate degradation. This information has been gained through all the stages of the research framework, notably pastoralist interviews (1-3), community focus groups (4,8) and key stakeholder interviews (8) with extension workers and Government officials. These guidebooks will be trialed with a representative sample of community members in order to optimise their adoption potential (9), prior to wider production and distribution through the Ministry of Agriculture (10). The process is iterative throughout, with even after initial dissemination, periodic reviewing of indicator performance in collaboration with local communities (11,12), to ensure that guidebooks remain relevant to dynamic user objectives and assets. This will lead to rejection, adaptation, and/or adoption of indicators for future manual editions.
Table 1: Indicator evaluation criteria identified in semi-structured interviews with Kalahari pastoralists
Indicators should:Ease of use criteria
Be easily measured
Be rapid to measure
Be timely
Make use of existing skills and knowledge
Accuracy Criteria
Be reliable and robust
Be representative of system variability over space and time
Be scientifically credible
Be relevant to the system/region being evaluated
Be diverse, encompassing a wide range of variables
Figure 1: Methodological framework for participatory indicator analysis
Case Study Findings
The application of the proposed methodological framework to pastoral livelihoods research in Kgalagadi District, South West Botswana remains ongoing. The preliminary findings and discussion provided here extends a previous review of the indicator identification stages (1-3) (Reed and Dougill, in press). This paper also discusses the findings by incorporating preliminary discussion of the primary indicator evaluation stages (4,5). Research is ongoing to conduct the next evaluative stages (6-8) (in July and August 2002) and therefore are simply discussed on a methodological basis at this stage to outline the methodological issues currently faced for discussion.