The power of mixed survey methodologies for detecting decline of the Bornean orangutan

Abstract

For many threatened species, it is difficult to assess precisely for large areas the change in their abundances over time and the relative impacts of climate and anthropogenic land use. This is because surveys of such species are typically restricted to small geographic areas, are conducted during short time periods, and use different survey protocols. We assessed the change in the abundance of Bornean orangutan Pongo pygmaeus morio in Sabah, Malaysia, and to identify environmental drivers affecting the change by integrating different types of survey data. We used nest count data obtained from aerial and ground transect surveys and occurrence data obtained from reconnaissance walks and interview survey over the past decade. We built a spatially-explicit dynamic population model within the Bayesian framework allowing these varying survey data to be analyzed jointly by explicitly accounting for each surveys sampling rate. We found that sampling rates vary across survey types, reflecting each surveys associated effort. Orangutan survival rates were strongly determined by natural forest extent and moderately by temperature. Orangutan migration rates across more than 1 km distance between forest patches were low, which underlines the importance of maintaining ecological connectivity. The paucity of species abundance data collected in a consistent manner over many years across broad extents often hinders the assessment of species population trend and their persistence across regional scales. We demonstrate that this can be addressed by integrating multiple survey data across different localities, provided that sampling rate inherent to each survey is accounted for.