Comparison of mark-resight methods to estimate abundance and rabies vaccination coverage of free-roaming dogs in two urban areas of south Bhutan

In Bhutan, Capture-Neuter-Vaccinate-Release (CNVR) programs have been implemented to manage the dog population and control rabies, but no detailed evaluation has been done to assess their coverage and impact. We compared estimates of the dog population using three analytical methods: Lincoln-Petersen index, the Chapman estimate, and the logit-normal mixed effects model, and a varying number of count periods at different times of the day to recommend a protocol for applying the mark-resight framework to estimate free-roaming dog population abundance. We assessed the coverage of the CNVR program by estimating the proportion of dogs that were ear-notched and visually scored the health and skin condition of free-roaming dogs in Gelephu and Phuentsholing towns in south Bhutan, bordering India, in September October 2012. The estimated free-roaming dog population in Gelephu using the Lincoln-Petersen index and Chapman estimates ranged from 612 to 672 and 614 to 671, respectively, while the logit-normal mixed effects model estimate based on the combined two count events was 641 (95% CI: 603-682). In Phuentsholing the Lincoln-Petersen index and Chapman estimates ranged from 525 to 583 and 524 to 582, respectively, while the logit-normal mixed effects model estimate based on the combined four count events was 555(95% Cl: 526-587). The total number of dogs counted was significantly associated with the time of day (AM versus PM; P= 0.007), with a 17% improvement in dog sightings during the morning counting events. We recommend to conduct a morning marking followed by one count event the next morning and estimate population size by applying the Lincoln-Peterson corrected Chapman method or conduct two morning count events and apply the logit-normal mixed model to estimate population size.