Are ranger patrols effective in reducing poaching‐related threats within protected areas?

Poachingis one of the greatest threatsto wildlife conservationworld-wide. However, the spatial and temporal patterns of poachingactivities within protected areas, and the effectiveness of ranger patrolsand ranger posts in mitigating these threats, are relatively unknown. We used 10 years (2006–2015) of ranger-based monitoring data and dynamic multi-season occupancy models to quantify poaching-related threats, to examine factors influencing the spatio-temporal dynamics of these threatsand to test the efficiency of management actions to combat poachingin Nyungwe National Park (NNP), Rwanda. The probability of occurrence of poaching-related threatswas highest at lower elevations (1,801–2,200 m), especially in areas that were close to roads and tourist trails; conversely, occurrence probability was lowest at high elevation sites (2,601–3,000 m), and near the park boundary and ranger posts. The number of ranger patrolssubstantially increased the probability that poaching-related threatsdisappear at a site if threatswere originally present (i.e. probability of extinction of threats). Without ranger visits, the annual probability of extinction of poaching-related threatswas an estimated 7%; this probability would increase to 20% and 57% with 20 and 50 ranger visits per year, respectively. Our results suggest that poaching-related threatscan be effectively reduced in NNP by adding ranger posts in areas where they do not currently exist, and by increasing the number of patrolsto sites where the probability of poachingactivities is high. Synthesis and applications. Our application of dynamic occupancy models to predict the probability of presence of poaching-related threatsis novel, and explicitly considers imperfect detection of illegal activities. Based on the modelled relationships, we identify areas that are most vulnerable to poaching, and offer insights regarding how ranger patrolscan be optimally deployed to reduce poaching-related threatsand other illegal activites, while taking into account potential sampling biases. We show that poachingcan be effectively reduced by increasing ranger patrolsto areas under high risk of poachingactivities, and by adding ranger patrolsnear these sites. These findings are broadly applicable to national parks and protected areas experiencing a high degree of poachingand other illegal activities.