A preliminary habitat suitability model for devil rays in the western Indian Ocean

Citation
Guirhem G, Arrizabalaga H, Lopetegui L, et al (2021) A preliminary habitat suitability model for devil rays in the western Indian Ocean. In: IOTC- 17th Working Party on Ecosystems & Bycatch (Assessment). IOTC-2021-WPEB17(AS)-27_rev1, Online
Abstract

The European tropical tuna purse seine fishery incidentally captures three highly migratory and endangered species of devil rays, spinetail devil ray Mobula mobular, sicklefin devil ray M. tarapacana, and bentfin devil ray M. thurstoni in the Indian Ocean. Due to their global decreasing populations, understanding the factors of their spatial and temporal distributions and the associated environmental conditions are fundamental for their management and conservation. Yet, the spatial and temporal distribution of devil rays in the Indian Ocean is poorly understood. Here we developed a habitat suitability model for devil rays in the Western Indian Ocean depicting the seasonal and interannual changes in their spatial distributions and underlying environmental conditions. We used bycatch data collected between the period 2010-2020 by the EU tropical tuna purse seine observer programs to determine which environmental variables influence the occurrence of devil rays using generalized additive models. A separate modelling was done for the spinetail devil ray, and for the three species of devil rays combined, since many individuals are only recorded at the genus level. The environmental variables associated with the presence of devil rays were chlorophyll, sea surface height and sea surface temperature fronts. When modelling the habitat suitability for spinetail devil ray, the most influential environmental variables were net primary production of phytoplankton and sea surface temperature fronts. Both the interannual and seasonal variability in habitat suitability of devil rays were explained by these environmental variables. We also showed that devil rays are associated to permanent hotspots in the Mascarene Plateau and Central Indian Ridge, and to seasonal hotspots in the Western Arabian Sea and Equatorial regions where there is a high occurrence of devil rays during winter monsoon. We found that setting on large tuna schools decreases the chances of devil ray bycatch. Both models predicted a higher probability of incidental catch of devil rays in fishing sets on free swimming schools of tunas than in sets on fish aggregating devices. The identified hotspots and associated environmental characteristics provide information about the habitat use and ecology of the devil rays in the Western Indian Ocean. Furthermore, the habitat suitability models, and biological hotspots identified in this study could also to be used to inform the development of future spatial management measures, including time-area closures, to minimize the interaction of pelagic fisheries with these vulnerable species.