Critical evaluation of a long-term, locally-based wildlife monitoring program in West Africa

Effective monitoring programs are required to understand and mitigate biodiversity declines, particularly in tropical ecosystems where conservation conflicts are severe yet ecological data are scarce. Locally-based monitoring has been advanced as an approach to improve biodiversity monitoring in developing countries, but the accuracy of data from many such programs has not been adequately assessed. I evaluated a long-term, patrol-based wildlife monitoring system in Mole National Park, Ghana, through comparison with camera trapping and an assessment of sampling error. I found that patrol observations underrepresented the park's mammal community, recording only two-thirds as many species as camera traps over a common sampling period (2006-2008). Agreement between methods was reasonable for larger, diurnal and social species (e.g., larger ungulates and primates), but camera traps were more effective at detecting smaller, solitary and nocturnal species (particularly carnivores). Data from patrols and cameras corresponded for some spatial patterns of management interest (e.g., community turnover, edge effect on abundance) but differed for others (e.g., richness, edge effect on diversity). Long-term patrol observations were influenced by uneven sampling effort and considerable variation in replicate counts. Despite potential benefits of locally-based monitoring, these results suggest that data from this and similar programs may be subject to biases that complicate interpretation of wildlife population and community dynamics. Careful attention to monitoring objectives, methodological design and robust analysis is required if locally-based approaches are to satisfy an aim of reliable biodiversity monitoring, and there is a need for greater international support in the creation and maintenance of local monitoring capacity.