Modelling the Spatio-temporal Distribution of Devil Rays in Western Indian Ocean using Observer Data from Tropical Tuna Purse Seine Fisheries.
Supervisors: Maria José Juan-Jordá, Haritz Arrizabalaga (AZTI)
The tropical tuna purse seine fisheries capture three highly migratory and endangered species of devil rays, Mobula mobular, M. thurstoni, M. tarapacana as bycatch. Due to their global decreasing population, understanding the factors of their spatial and temporal distribution and the associated environmental conditions are fundamental for their management and conservation. Yet, the spatial and temporal distribution of devil rays in the Western Indian Ocean is still poorly understood. Thus, the objective of this study is to develop a habitat suitability model for devil rays in the Western Indian Ocean depicting the seasonal and interannual changes in their spatial distributions and the underlying environmental conditions. In this study, I used the bycatch data collected by the EU purse seine observer program with 26523 fishing sets between the period 2010-2020 to determine which environmental variables influence the occurrence of devil ray using generalize additive model and EU Copernicus Marine Environment Monitoring Service as source of environmental covariates. A separate modelling was done for 1) species of devil ray (Mobula mobular, M. thurstoni, M. tarapacana, Mobula sp.) and 2) spinetail devil rays (Mobula mobular). Results revealed that devil rays are associated with persistent areas (Mascarene Plateau and Central Indian Ridge) and seasonal areas (Western Arabian Sea and Equatorial regions). For seasonal hotspots, there is a high occurrence of devil rays during winter monsoon (week 1-12, 49-53). Both models are associated with sea surface temperature fronts (0.02-0.03 oC. km -1). However, model for devil rays is linked with chlorophyll (0.3-0.5 mg. m-3) and sea surface height (0.4 m) while model for spinetail devil ray Mobula mobular is linked to net primary production of phytoplankton (10-30 mg. m-3. day-1). 1) Chlorophyll and sea surface height explains the distribution of devil rays in both persistent and seasonal hotspot while 2) net primary production of phytoplankton explains the distribution of spinetail devil rays (Mobula mobular) and only identify persistent hotspots. Both models predict more presence in school sets than fish aggregating devices. The hotspots identified in this study can contribute to inform holistic spatial management strategies to reduce bycatch of vulnerable and threatened species caught by pelagic fisheries in the Western Indian Ocean.

Keywords: devil rays, Mobula, bycatch, species distribution models, tuna fisheries, Indian ocean