Identifying fisheries operations in tropical multispecies fisheries: A comparative analysis of multivariate approaches and neural networks

Abstract

Efforts to assess and manage mixed fisheries started a shift from monospecific to multispecies approaches in European countries; this transition, however, is still far from being applied in other regions. Given the operational heterogeneity and diversity of species targeted, disentangling the complexity of those fisheries' complexities is the first step to understanding, monitoring, and managing them. Multivariate analytical tools supported by the métiers approach offer alternatives to untangle the fishing operations and identify management units. Among the analytical techniques used for this purpose, self-organizing maps (SOM) and different multivariate methods have been reported. The usage of the former has been less common, and there is limited information regarding its performance compared with the latter. In this study, we aimed to evaluate the performance of the SOM and multivariate techniques to identify the operational units of a mixed fishery from the southeast coast of Mexico. Five métiers or management units were identified using both analytical tools matching the groups; however, some differences were observed. We discuss the strengths and caveats of those techniques and examine the management implications of the findings in the context of the analyzed fishery.