Optimizing nocturnal thermal-drone surveys for black-shanked douc langurs: calibration to ground distance sampling, gimbal-angle effects, and survey effort

Drone-mounted thermal infrared (TIR) imagery could improve monitoring of arboreal primates in closed-canopy forests, but inference remains constrained by imperfect detection. We evaluated nocturnal TIR drone transect surveys as a scalable method for monitoring the Critically Endangered black-shanked douc langur. We tested how gimbal angle affects detectability, calibrated drone encounter rates against ground distance-sampling (DS) density indices using ground DS as a reference, and used Monte Carlo simulations to identify efficient survey designs. We conducted diurnal DS and nocturnal TIR surveys along the same 22 fixed transects in Cát Tiên National Park, Vietnam, from January to May 2023, with repeated drone flights at 45° and 90° gimbal angles. Across 131 45° and 129 90° flight replicates, the 45° configuration yielded higher detection probability (0.80 vs 0.66) and higher encounter rates, consistently outperforming the 90° configuration. Drone encounter rates were positively associated with DS density indices, meaning that drone data reflect underlying density once calibrated to ground surveys, but cannot on their own provide direct estimates of absolute density. For our study site, where the mean encounter rate was 1.76 groups/km, simulations based on the preferred 45° configuration showed that three repeat flights per transect were enough to achieve a mean detection probability of 0.97, supporting a 10-transect screening design (CV = 0.23) and a 15-transect monitoring design (CV = 0.18). Beyond douc langur monitoring, this study presents a practical workflow for planning thermal-drone surveys and interpreting drone encounter rates for other arboreal mammals across sites and survey designs.