Bayesian occupancy monitoring for Annamite endemic biodiversity in central Vietnam

Abstract

Given the crisis facing South-east Asian biodiversity evidence led conservation, including assessing the impact of innovative protected area management models, is urgently needed. Bayesian statistics provide an intuitive way to interpret biodiversity monitoring data but are largely unused, or poorly understood, by field biologists and protected area managers. We built Bayesian occupancy models for two threatened endemics of the Annamite mountains: northern yellow-cheeked gibbon Nomascus (gabriellae) annamensis and crested argus Rheinardia ocellata ocellata based on auditory surveys in three protected areas in central Vietnam. Occupancy of 2 × 2-km grid cells across the landscape was 0.76 ± SE 0.03 for northern yellow-cheeked gibbon and 0.68 ± SE 0.05 for crested argus. Models predicted higher probability of gibbon occurrence at lower elevations and higher probability of crested argus presence with increasing dense forest cover. Bayesian modeling is a useful tool for assessing the effectiveness of conservation interventions and for measuring progress against conservation goals. The wider application of Bayesian statistics in conservation monitoring should allow more intuitive and user-friendly representation of sampling uncertainty, including visual representation of probability distributions and more rigorous testing for changes in the status of conservation targets.