USE OF GENERALIZED LINEAR-MODELS TO INVES TIGATE THE FISHING STRATEGIES OF TUNA PURSE SEINERS, IN PRESENCE OF FLOATING OBJECTS IN THE WESTERN ATLANTIC-OCEAN

Citation
D. Gaertner et al., USE OF GENERALIZED LINEAR-MODELS TO INVES TIGATE THE FISHING STRATEGIES OF TUNA PURSE SEINERS, IN PRESENCE OF FLOATING OBJECTS IN THE WESTERN ATLANTIC-OCEAN, Aquatic living resources, 9(4), 1996, pp. 305-323
Citations number
55
Categorie Soggetti
Fisheries,"Marine & Freshwater Biology
Journal title
ISSN journal
09907440
Volume
9
Issue
4
Year of publication
1996
Pages
305 - 323
Database
ISI
SICI code
0990-7440(1996)9:4<305:UOGLTI>2.0.ZU;2-B
Abstract
The fishing strategies used by Venezuelan purse-seine fishermen in the Caribbean Sea are based on sighting indices, which mark the presence of tuna schools at the surface of the ocean. Once detection has been m ade, the purse seiner can seek the help of a baitboat to increase the percentage of successful sets. In order to consider the variability of factors, such as season, floating objects associated with schools, he lp of a baitboat, tuna species and mean weight of fish in the school, the probabilities of some events have been estimated by legit modeling , drawn from log-linear models. The probabilities of capturing a schoo l of large tunas are high when yellowfin tuna is the main species in t he school and when the set is made on non-associated school or on whal e-sharks; this effect is greater during the dry season. In contrast, i f skipjack tuna is the dominant species in the school, the probability of catching large tunas is low, especially for schools associated wit h whales. The results of these models have been included in a tree dec ision analysis. If we use the criterion of the mathematical expectatio n of outcomes, the tactic which consists of using a baitboat does not lead to the best choice. However, the effect of this tactic is signifi cant, since the rate of positive sets increases from 58%, without the baitboat, to 85% with the baitboat. The efficiency of this help is lin ked with the nature of the sighting index. This efficiency is greater for sets associated with whales and especially with non-associated sch ools than for sets associated with whale-sharks. A Monte Carlo simulat ion shows that some important outcomes can be expected, without the ba itboat, for sets associated with whales. Hence, the captains of the sm all and median purse seiners, who use baitboats more frequently than t he captains of the large ones, could be characterized as risk-averse, without neglecting some limiting characteristics of their boats such a s the speed. Compared to other tuna fishing grounds, the Caribbean Sea is characterized by a high proportion of sets made in presence of wha le sharks (37% of the sets) and whales (24%). In contrast, the frequen cy of log sets is very low. This fact could be explained by the nature of ocean currents which carry away these logs out off the Venezuelan purse-seiner fishing grounds. To conclude, the similarities of adaptiv e attitudes developed against risk in different fields, such as agricu lture and fisheries, are briefly discussed.