Best practices for reporting individual identification using camera trap photographs

Abstract

Camera traps have become a popular sampling tool in ecological studies. This is especially true for studies that estimate population densities through spatial capture-recapture models of species with physical traits allowing individual identification. However, misidentification is a potential problem when trying to identify individuals from photographs. Researchers need rigorous methods for checking and rectifying errors, handling unclassifiable photographs and resolving inter-observer discrepancies to prevent or reduce misidentification. Through a systematic review of both traditional and spatial capture-recapture studies, we found that few studies described the methods used to avoid individual misidentification; moreover, most studies did not provide photographic evidence that individuals could be reliably differentiated. This lack of transparency and accountability prevents reviewers and readers from assessing the reliability and robustness of the studies, and falls short of the standards of reporting compared to other fields of research such as biomedical or genomic research. To address this problem, we present an Individual Identification Reporting Checklist (IIRC) listing minimum standards for reporting methods and photographic data for individual identification, which are necessary evidence in publications that rely on accurate individual identification from camera trap photographs. By adopting the IIRC, researchers will enhance the credibility of capture-recapture studies, facilitate information transfer among researchers, and improve the progress of downstream research and conservation efforts reliant on accurate photographic identification. We acknowledge that photographic data may be associated with heightened poaching risks. However, in the context of capture-recapture studies, we argue that frequently published information such as study site locations and abundance estimates carry much higher risks of poaching, while the additional threats incurred from the inclusion of photographic information are likely to be marginal and can be mitigated. @ 2020 The Authors