Efficiency is doing things right: high-throughput, automated, 3D methods in the modern era of otolith morphometrics

Abstract

The morphometrics of fish otoliths have been commonly used to investigate population structures and the environmental impacts on ontogeny. These studies can require hundreds if not thousands of otoliths to be collected and processed. Processing these otoliths takes up valuable time, money, and resources that can be saved by automation. These structures also contain relevant information in three dimensions that is lost with 2D morphometric methods from photographic analysis. In this study, the otoliths of three populations of coho salmon (Oncorhynchus kisutch) were examined with manual 2D, automated 2D, and automated 3D otolith measurement methods. The automated 3D method was able to detect an 8% difference in average otolith density, while 2D methods could not. Due to the loss of information in the z axis, and the longer processing time, 2D methods can take up to 100 times longer to reach the same statistical power as automated 3D methods. Automated 3D methods are faster, can answer a wider range of questions, and allow fisheries scientists to automate rather monotonous tasks.