| الملخص: |
Globally, hundreds of thousands of wildlife rangers patrol wide areas within protected areas every day, observing biodiversity and illegal activities. Data collection by rangers therefore has enormous potential to track changes in biodiversity and threats to it, at scale and with little additional cost. However, ranger patrols are biased in space and time and detections are imperfect, so what rangers observe may not capture underlying reality well. Furthermore, even when monitoring results are reliable, they might not be used effectively to inform conservation management. Effective ranger-based monitoring also requires active engagement by the people collecting and using data (rangers and managers). In this Thesis I investigated factors affecting (a) the reliability of ranger-based monitoring data and, (b) the effective use of these data within conservation management. I used the monitoring and management of elephant poaching in the Mana-Chewore World Heritage Site, Zimbabwe, as a case study and combined statistical, mathematical, and qualitative methods. I began with a participatory modelling approach in which rangers and managers helped me to build and evaluate models of the spatial distribution of elephant poaching in Mana-Chewore, with statistical methods to account for patrol bias. Combining quantitative models and interview responses allowed for more robust inference in the face of uncertainty, with proximity to water emerging as the strongest driver of poaching (reflecting both poacher and elephant behaviour). Next, I developed mathematical simulations to quantify how patrol characteristics (effort, spatial coverage, etc.) interacted with poaching dynamics to affect the power of ranger-collected data to detect underlying spatial and temporal trends in poaching. Power to detect trends was low in many scenarios, with some non-intuitive results (such as spatially targeted patrols achieving power similar to spatially random patrols). Strategies required to achieve robust results depended heavily on monitoring objectives (the magnitude of change in poaching that managers wish to detect, for example). To complement these quantitative insights, I interviewed 23 rangers working in Mana-Chewore to investigate their perceptions of patrol-based data collection. I found that their occupational culture (including a strong sense of duty and deference to authority), as well as their awareness of how their data were used, shaped their engagement with monitoring. In a second qualitative analysis, I interviewed nine park managers and 17 senior staff of the national wildlife authority to investigate the extent to which ranger-collected data were used to inform anti-poaching. Managers valued and made basic use of ranger-collected poaching data but did not systematically analyse trends in these data to inform their anti-poaching strategies. Managers felt that management based on intuition, experience and more reactive data-use was more familiar and dependable, and therefore did not embrace data-based adaptive management. For ranger-based monitoring to contribute effectively to biodiversity conservation, practitioners and scientists must acknowledge, understand, and account for uncertainty in both monitoring data and the behaviour of those collecting and using it. Clearly defining monitoring and trend detection goals and critically evaluating the likelihood of achieving these goals is essential, as is meaningfully engaging the perspectives of rangers and managers. More generally, this research demonstrates the importance of interdisciplinarity in the study of socio-ecological systems, and the power of models for both understanding and dealing with the uncertainty inherent in these systems. |