Abstract

As in most of the islands of the Lesser Antilles, the small-scale large pelagic fishery in Martinique (FWI) is based on the exploitation of fishes concentrated around three types of "aggregators": flotsams, seamounts and Fish Aggregating Devices (F ADs). The technical and human means implemented by the small scale fishermen to catch the fishes are the same whatever aggregator they target.\Due to the lack of consistent data series, CPUEs are commonly expressed as "catches by trip". The influence of several other fishing efIort components on the catches per trip data issued from the small scale large pelagic fishery in Martinique was modelled using General Linear Models. The aim of the analysis is not to get a predictive model but to identify the variables explaining most of the catch per trip variability so as to evaluate the fishing efIort more accmately.\At the scale of the globallarge pelagic fishery, the most influent variables were the type of aggregator exploited during the trip ("type of fishing") and the depth layerprospected with the fishing gears. The same analysis was conducted within the strata defined by the types of fishing. Most of the catch per trip variability of the moored F AD fishery was explained by the use or not of drifting longlines allowing to catch large adult pelagics (essentially Thunnus albacares and Makaira nigricans.\The fishing efIort components available seemed to be inefficient to explain the catch per trip variability of the very random ofIshore trolling fishery. The catch per trip variability of the coastal trolling fishery depended mainly on the trip length.\Generalized Linear Models appeared to be efficient to process the data issued from the small skewed data sets often encountered in the insular context. The results suggest several fisheries should be defined within the large pelagic fishery so as to better assess the catch per trip variability but also the age structure of the catches. The fisheries should be classically defined by the fishing area, the species targeted, the fishing gears but also by the type of aggregator prospected. The moored F AD fishing seems to substitute efIective fishing time (local fishing efficiency) to costIy travelling time (local searching efficiency) compared to the traditional coastal trolling fishery.

PDF Preview