Global declines in biological diversity are increasingly well documented and threaten the welfare and resilience of ecological and human communities. Despite international commitments to better assess and protect biodiversity, current monitoring effort is insufficient and conservation targets are not being met (e.g., Convention on Biological Diversity 2010 Target). Protected areas are a cornerstone of attempts to shield wildlife from anthropogenic impact, yet their effectiveness is uncertain. In this dissertation, I investigated the monitoring and conservation of wildlife (specifically carnivores and other larger mammals) within the context of a poorly studied savanna reserve in a tropical developing region: Mole National Park (MNP) in the West African nation of Ghana.

I first evaluated the efficacy of the park's long-term, patrol-based wildlife monitoring system through comparison with a camera-trap survey and an assessment of sampling error. I found that park patrol observations underrepresented MNP's mammal community, recording only two-thirds as many species as camera traps over a common sampling period. Agreement between methods was reasonable for larger, diurnal and social species (such as many larger ungulates and primates), but camera traps were more effective at detecting smaller, solitary and nocturnal species (particularly carnivores). Long-term patrol data were also subject to considerable sampling variation that could make interpretation of wildlife trends unreliable, and I suggest ways in which this locally based monitoring program may be improved.

Given the ecological and cultural importance of carnivore species, their propensity for human conflict, and the difficulty with which they are monitored, I assessed their status and vulnerability to extinction in MNP. Only 9 of 16 historically occurring carnivore species were detected in the camera-trap survey (covering 253 stations deployed for 5,469 trap days between October 2006 and January 2009). A hierarchical multi-species occupancy model applied to camera-trap data indicated a low overall likelihood of the presence of undetected species. Results from concurrent sign, call-in, and village surveys, as well as patrol records, provided more equivocal evidence of carnivore occurrence but supported the conclusion that many carnivores have declined and are likely functionally or fully extirpated from the park, including the top predator, lion (Panthera leo). Evidence of local human-carnivore conflict was also documented, including hunting of carnivores for traditional use and in retaliation for livestock depredation. Contrary to expectation, variation in carnivore persistence was not explained by ecological or life-history traits such as body size, home range size or fecundity, thus raising doubt as to the predictability of carnivore community disassembly.

I extended the multi-species occupancy model to test hypotheses about extrinsic influences on carnivore community dynamics in MNP. I derived spatially explicit GIS descriptors of heterogeneity in illegal hunting pressure, law enforcement patrol effort, prey biomass, and habitat productivity, and used a Bayesian modeling framework to assess support for their effects on carnivore occurrence. The framework explicitly accounted for spatial autocorrelation and variation in species- and site-specific detection probabilities. Contrary to my expectation, there was no indication of a consistent, negative effect of illegal hunting activity on spatial patterns of carnivore occurrence. By contrast, occurrence patterns of most species were positively associated with prey biomass, and several species had either positive or negative associations with riverine forest (but not with other indicators of habitat heterogeneity).

I conclude that pressure from hunting and other anthropogenic impacts remains high for West African wildlife, even within protected areas, but that human-wildlife relations are complex and their consequences inadequately predicted by simple models of extinction risk. Existing monitoring programs may generate data unsuitable for strong inference on wildlife community dynamics, and careful attention to objectives and methodology is needed. More attention to the protection and recovery of carnivore populations is also needed, as are further focused and interdisciplinary efforts to inform and improve wildlife conservation in West Africa.