Improving conservation and fishery assessments with local knowledge: future directions

While there is great interest in the potential of local knowledge to inform quantitative conservation, there are no formal guidelines for its interpretation or analysis. Our work to quantify retrospective bias in fisher interviews and explore effects on IUCN Red List assessments shows that assumptions made when analyzing local knowledge affect conservation assessments (O'Donnell, Pajaro & Vincent, 2010). Aswani (2010), Daw (2010), and Patterson (2010) suggest several important ways to improve our understanding of local knowledge and its incorporation into conservation and management. We concur with Patterson (2010) who notes that the existence of bias does not render local knowledge useless. As part of ‘re-inventing fisheries management’ (Pitcher, Hart & Pauly, 1998) scientists must confront uncertainty by quantifying and including it in management recommendations (Ludwig, Hilborn & Walters, 1993). Fisher knowledge should be probed with the same analytical rigor as scientifically collected data, and to do so we need methods to quantify and correct bias in local knowledge. Daw suggests research to identify and quantify memory and perception bias that will improve our ability to accurately incorporate local knowledge into quantitative conservation assessments (2010). Psychologists have identified biases that can occur when we recall events from the past (e.g. Bradburn, Rips & Shevell, 1987), but there is currently no approach for quantifying those biases. As Daw (2010) suggests, there is a need for collaboration with psychologists to understand the nature and existence of retrospective bias along with a host of other memory and perception biases (table 1; Daw, 2010). Developing methods of quantifying these biases and/or procedures to correct for them, as we have explored here, would allow researchers to interpret interview responses appropriately, improving conservation assessments. A range of existing techniques should be utilized to help us collect and apply local knowledge as accurately as possible. There is a growing literature about how to design studies, phrase questions, and choose appropriate respondents (e.g. Neis et al., 1999a; Davis & Wagner, 2003; Anadón et al., 2009). Quantitative assessment protocols are being developed that explicitly incorporate local knowledge (e.g. Walmsley, Medley & Howard, 2005). It is necessary to acknowledge memory biases that may be present (Table 1; Daw, 2010) and consider the motivations of respondents to over- or under-report (reviewed by Yasué, Kaufman & Vincent, 2010). Comparing multiple sources of local knowledge (‘local’ data-collection methods reviewed by Danielsen, Burgess & Balmford, 2005) with scientific observations can help determine the strengths and weaknesses of both approaches. As well, assumptions about biases should be made explicit in research papers and researchers should experiment with metrics to account for biases and evaluate effects on results. As both Aswani (2010) and Daw (2010) point out, the importance of understanding the perceptions of resource users extends beyond filling data gaps in single-species monitoring. Aswani reminds us that the appropriate application of local knowledge requires a broader understanding of community, beliefs, and economic context of that knowledge (Neis et al., 1999a). For example, our study (O'Donnell et al., 2010) is part of a larger research project to evaluate options for seahorse conservation and management action. We have consulted with a range of stakeholders and experts from fishers to scientists about preferred management options (Martin-Smith et al., 2004). We are currently investigating the spatial and behavioral dynamics of fishing, noted by Aswani (2010) as being necessary for moving beyond managing for single species to a more comprehensive social-ecological system. We also agree with Daw's (2010) suggestion that a better understanding of how humans perceive and recall environmental change can help us understand conflicts in conservation and resource governance. Conflict can arise when there is a gap between fishers' perceptions and scientific knowledge used to inform management of fishing. From another study conducted in our focal community, Handumon, we know that locals perceive their Marine Protected Area (MPA) to be more successful than scientific data suggest (Yasuéet al., 2010). Since local support is a key determinant of ecological success in community-based MPAs, it will be important to acknowledge and then explore this discrepancy in order to ensure long term viability of the MPA. It has been more than a decade since the first illustrations of the utility and applicability of local knowledge to quantitative fishery management were published (e.g. Neis et al., 1999b), but few conservation or fishery assessments are yet based on local knowledge (Brook & McLachlan, 2008). With the number of threatened species on the rise and the vast majority (>95%) of described species yet to be assessed for extinction risk (Vié, Hilton-Taylor & Stuart, 2009), the need for cost-effective, reliable information on species abundance and exploitation has never been greater. To close this gap we encourage others also to: (1) acknowledge and explore the strengths and weaknesses of local knowledge; (2) recognize biases associated with documenting local knowledge; (3) clearly state assumptions about local knowledge bias and evaluate the effect of assumptions on assessments; (4) improve methods for quantifying bias in local knowledge, allowing researchers to appropriately interpret and apply it to conservation assessments and management recommendations; (5) place local knowledge within the context of broader ecological and societal systems. Determining how best to integrate contextual or qualitative information with quantitative information for conservation and management gain is an area ripe for further research.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.