Fine-tuning established morphometric models through citizen science data

Abstract

For globally threatened species, patchy data pose challenges in both establishing historical baselines and assessing current distributions. In the case of sawfishes, much previous research has relied on amputated trophy rostra or historical photographs to fill data gaps in distribution and population estimates. However, the established morphometric models that allow the estimation of total length from rostra are restricted by the data available to generate them, with many constructed solely from juvenile morphometric data, and robust error testing rarely performed. This study generates an independent dataset from citizen science photographs and field measurements, incorporating sawfishes of all life history stages, to test the robustness of previously proposed models for the four Indo-Pacific sawfish species. Rostrum-to-length morphometric models (RLMs) were confirmed for Pristis clavata and Pristis pristis, and a new model was proposed for Pristis zijsron. A key source of error in previous RLMs for Anoxypristis cuspidata was identified as changing rostrum morphology with body size, but there were insufficient measurements in the present dataset to propose a new robust model for this species. This study highlights the problems associated with extrapolating total lengths of animals beyond the size bounds used to produce the original model. For historically exploited species like sawfishes, which are often represented in collections by trophies alone, these methods outline a critical step prior to use in further ecological study.