Nets and Networks: Vessel Associations and Fishing Success in Commercial Fisheries

Fishing activities provide the link between fish harvesters and the populations that they exploit.  Within the context of foraging theory, such predatory relationships have been studied using a variety of optimization models.  This has resulted in relatively simple, successful predictive models of foraging decisions and provided insights into factors that can influence foraging success.  In the context of fishing, these established principles from behavioural ecology can provide insights into factors that influence fishing mortality through choices made by fish harvesters.  However, many of the models of foraging theory start from the simplifying assumption that foragers have perfect knowledge of their environment.  This assumption is seldom testable in either non-human foragers or fish harvesters, but is likely of varying validity among different contexts and individuals.  Within fishing fleets, the most likely source of this knowledge is personal experience and the experience of other members of the fleet. Information is often shared within fishing fleets, but not equally and completely among participants.  Instead, some individuals (and their vessels) can be expected to share information on conditions and success more freely than others, creating information networks within fishing fleets. Network theory provides a quantitative basis for examining associations within fleets and their relationship to fishing success.  Unfortunately, we generally have little direct data on information exchange in commercial fisheries. However, vessel monitoring systems (VMS) can provide regular records of the geographical position for entire fleets throughout a fishing season. They are usually employed to monitor regulatory compliance, but the data generated also allows the detailed spatial examination of fishing activities.

In this study we developed a procedure using VMS data to define a network structure among Dutch beam trawlers and then examined the relevance of this network to individual vessel performance. The vessel activity associated with any single VMS record was inferred from a Gaussian mixture model.  After identifying VMS records where vessels were fishing a network was constructed based upon vessel proximity. Vessel performance was estimated by the value of the annual catch in the catch records standardized by days fishing, vessel horsepower, and non-fishing time at sea. Our spatially inferred network demonstrated behavioural differences among individual vessels that were relevant to their realized fishing power and their contribution to fishing mortality. This has implications for the interpretation of catch rate time series and regulatory measures based upon targeted regulation or removal of specific vessels from an active fleet.