Tiger abundance and density estimates from non-invasive sources: Camera-trap and genetic capture-recapture analyses

We present tiger density and abundance estimates derived independently from two non-invasive sampling techniques: camera-trapping and fecal genetic sampling in same geographic areas. Camera-trap data collected in 2009 covered 359.2 km 2 of lowland core area of Chitwan National Park using 187 camera-trap locations and 2,821 trap-nights obtaining 264 identifiable photographs of 41 tigers with 103 recaptures. Fecal genetic sampling in 2011 surveyed 325 km 2 along three spatial routes covering 1,173 km and collecting 257 tiger fecal samples. An eight multi-locus genotyping assay identified 31 tigers with 42 recaptures. We analyzed each dataset separately using spatially explicit capture-recapture models under Maximum likelihood and Bayesian framework. Maximum Likelihood-based density estimates from DENSITY were 2.55 (SE 0.6) tigers per 100 km² for camera traps and 2.57 (SE 0.9) for genetic data, while Bayesian estimates from SPACECAP were slightly lower at 2.4 (SE 0.3) and 2.2 (SE 0.5), respectively. Maximum Likelihood abundance estimates were 66 (SE 7.9) for camera data and 88 (SE 19.3) for genetic data, whereas Bayesian estimates were 64 (SE 7.9) and 93 (SE 19), respectively. Two independent data sources broadly comparable density estimates, though uncertainty was greater for the genetic dataset, which included fewer identified individuals and spatial recaptures. Because the data were surveyed in different years and differed in sampling effort, differences in precision cannot be attributed solely to sampling method. Non-invasive genetic sampling nonetheless produced density and abundance estimates broadly consistent with an established camera trap protocol, suggesting that integrating the data sources can enhance robust tiger monitoring and inform effective conservation planning.