ABSTRACT
VMS (Vessel Monitoring System) data are increasingly being viewed as essential in commercial stock research and management. However, the usefulness of these data for the estimation of the fishing effort is highly dependent on the correct identification of the VMS registers associated to fishing operations. In this study, a VMS sample of unprocessed data with an average resolution of 10 minutes, belonging to four vessels licensed for purse seining, was used to analyze the fishing operations, starting with the identification of fishing trips and then the identification of hauls within trips. The methodological approach was based on the in-depth knowledge of the typical fishing operations carried out by the purseseine fleet, thus allowing for the correct interpretation of the information on position, course and speed contained in the VMS data. An algorithm was developed incorporating that knowledge, resulting in a first approach to the automatic delimitation of the different phases of a single trip, basically by differentiating between fishing and non-fishing VMS points. The data time resolution was not enough to enable the identification of the gear deployment phase, which is an extremely fast operation. Haul back, carried out at very low speeds and extending over time, was found to be the best candidate to be a marker of the fishing activity.
