Michel Dreyfus-León

Individual-based modelling of fishermen search behaviour with neural networks and reinforcement learning

A model to mimic the search behaviour of fishermen is built with two neural networks to cope with two separate decision-making processes in fishing activities. One neural network deals with decisions to stay or move to new fishing grounds and the other is constructed for the purpose of finding prey within the fishing areas. Some similarities with the behaviour of real fishermen are found: concentrated local search once a prey has been located to increase the probability of remaining near a prey patch and the straightforward movement to other fishing grounds. The artificial fisherman prefers areas near the port when conditions in different fishing grounds are similar or when there is high uncertainty in its world. In the latter case a reluctance to navigate to other areas is observed. The artificial fisherman selects areas with higher concentration of prey, even if they are far from the port of departure, unless a high uncertainty is related to the fishing ground. Connected areas are preferred and followed in orderly fashion if a higher catch is expected. The observed behaviour of the artificial fisherman in uncertain scenarios can be described as a risk-averse attitude. The approach seems appropriate for an individual-based modelling of fishery systems, focusing on the learning and adaptive characteristics of fishermen and on interactions that take place at a fine scale.

A spatial individual behaviour-based model approach of the yellowfin tuna fishery in the eastern Pacific Ocean

A spatial individual-based model of the yellowfin tuna fishery in the eastern Pacific Ocean is presented. Schools of fish and individual fishing vessels are represented with artificial neural networks. These representations are intended to model behaviour decisions of movement in space. Schools of fish search continually for comfort and tuna vessels search for the tuna schools during a fishing trip. Two scenarios are considered: one with no fishing regulation and another with area closure during the last quarter of the year. This model is focused on spatial dynamics of fishing effort. Effort redistribution when regulations are implemented is not well understood and this modelling approach can help fishery managers to envisage some regulation effects in the fishery.

Modelling Cooperation between fishermen with a Cellular automaton. A framework for fishing effort spatial dynamics

Abstract In many fisheries sharing information between vessels is an important characteristic of fishermens behaviour rarely modelled or analyzed. A Cellular Automaton is designed in an attempt to understand circumstances that favour group formation. The simulated world is toroidal with a static fishing resource distributed in patches. Movement decisions are random in the case of fishermen in a local scale. After a certain time interval, sharing information is possible between fishermen in a dynamic Moores neighbourhood of cells at a broader scale and movement to adjacent areas may occur according to a set of rules. The dynamic neighbourhood is a novel concept defined in this work within the framework of Cellular Automata. Decision making by each fisherman is a function of the influence other fishermen (neighbours) exert on them as well as on personal knowledge, to form an opinion of the areas (cells) quality, and take action consequently.

Recruitment prediction for Pacific herring (Clupea pallasi) on the west coast of Vancouver Island, Canada

The accurate prediction of recruitment to the fishery is a very important tool within the management structure of any fish stock being exploited. In the case of the Pacific herring, Clupea pallasi, fishery in Canada, a forecast of the abundance of each herring stock is particularly important for formulating an annual catch quota. The sustainable management of the fishery and the resource is based in part on accurate recruitment forecasting because Pacific herring are short-lived and so the recruitment contributes a significant part of the total spawning run targeted by the fishery each year. Several factors are believed be important in determining the success of recruitment besides spawners biomass. Since herrings are “r” strategists, conditions related to the egg, the planktonic, or even the juvenile stage might determine the future level of recruitment. Recently a formula that defines conditions for a semi-quantitative level of recruitment forecast was elaborated using genetic algorithms and current study attempts to improve on this model. Using salinity in two quarterly periods during the planktonic and pre-recruit stages, temperature and spawning biomass for the west coast of Vancouver Island stock, classification rules that define recruitment in 3 different levels (low, medium and high) were developed with a genetic algorithm, setting low and high boundaries for each condition. A 75% success in classifying recruitment was obtained. The model was shown to be particularly effective at predicting when the recruitment would be low, which could be important from the perspective of the Precautionary Approach and the sustainable management of this stock.

Adaptive responses of tropical tuna purse-seiners under temporal regulations.

Abstract The failure to achieve fisheries management objectives has been broadly discussed in international meetings. Measuring the effects of fishery regulations is difficult due to the lack of detailed information. The yellowfin tuna fishery in the eastern Pacific Ocean offers an opportunity to evaluate the fishers’ responses to temporal regulations. We used data from observers on-board Mexican purse-seine fleet, which is the main fleet fishing on dolphin-associated tuna schools. In 2002, the Inter-American Tropical Tuna Commission implemented a closed season to reduce fishing effort for this fishery. For the period 1992–2008, we analysed three fishery indicators using generalized estimating equations to evaluate the fishers’ response to the closure. We found that purse-seiners decreased their time spent in port, increased their fishing sets, and maintained their proportion of successful fishing sets. Our results highlight the relevance of accounting for the fisher behaviour to understand fisheries dynamics when establishing management regulations.