Can modelling the drift of bycaught dolphin stranded carcasses help estimate total by-catch and identify involved fisheries? A feasibility study
Peltier, Helene, Authier, Matthieu, Dabin, Willy, Dars, Cecile, Demaret, Fabien, Van Canneyt, Olivier, Daniel, Pierre, Ridoux, Vincent
International Whaling Commission
The aim of this work is to test an approach that would be able to identify which fisheries could be involved in the observed unusual stranding events of January-March 2017. This is in line with ToR #3 of the expert group established to develop an independent review of the bycatch estimates derived from strandings in the Bay of Biscay. To do this we examined how the likely distributions of mortality of bycaught dolphins inferred from carcass drift modelling coincide with fishing effort statistics in the same area at the same dates for different fleets. The likely mortality areas at sea of common dolphins stranded with bycatch evidences were predicted using the reverse drift modelling methodology. The fishing effort data were generated from the Vessel Monitoring System, that automatically collects positional data of fishing vessels, along with other official sources of data to inform fishing gears and target species.
The number of dead small cetaceans from the tip of Brittany to the Spanish border in February and March 2017 was estimated at 4 430 small cetaceans (IC95% [2 750; 8 530]), including 3 690 common dolphins (IC95% [2 230; 6 900]) of which about 3500 would have been bycaught in fisheries.
Generalised Additive Models (GAMs) were used to explore the spatial consistency of the different fishing gears with the distribution of bycaught common dolphins at sea as obtained by carcass drift back-calculation.
Among all considered fisheries, three were positively correlated with bycaught dolphin mortality areas: the French pelagic trawl fishery, the French Danish seine fishery, and the Spanish bottom trawl fishery.
This work described one of the most intense stranding event ever recorded along the French Atlantic coasts. VMS data have been used to analyse the spatiotemporal correlation of fishing effort and likely mortality areas of stranded animals. Fisheries identified as having spatial and temporal correlation with these events corresponded to several gear types and target species. However, beyond this diversity of gears, two characteristics appeared to be shared: targeting predatory fishes in winter (sea bass and hake) and using high vertical opening gears.
In order to refine the understanding of these interactions with fisheries, and to identify more accurately some segments of the fleets, landed species related to fishing effort should be incorporated into the GAM analyses, in particular for the three fisheries that were shown to be positively correlated with bycaught cetaceans.