PILOT : SMALL PELAGIC FISHERIES PLANNING

The goal of this pilot is to provide the crew and ship owners with information which benefits fisheries planning. It will run in two phases. The first phase will focus on providing an interactive end user tool for browsing compilations of past history, while the second level will provide the end user with a tool which will do predictions of where and when fish can be caught in an energy efficient manner.

This pilot will collect and combine as many data sets as possible which may be of interest for predicting/planning of future fisheries effort. Extensive data sets within fisheries activity and catch statistics will be combined with information from that time and history of the same, such as meteorological and oceanographic data, moon phase, time of day, time of year. Machine learning will be used to form a predictive model, which will then be used for predictions of fish availability.

This information will provide insight of how the fisheries best can be performed, and will give valuable input for the planning process.

Large and varied data sets are needed for this pilot. Information about all pelagic catches that have been landed in Norway the last decades will be provided by Norges Sildesalgslag, as well as current data on the same. Meteorological and oceanographic hindcasts and forecasts will be provided by e.g. the Norwegian Meteorological Institute. Operational data will be provided by the participating fishing vessels.

The analyses and data aggregation will be performed at the SINTEF Marine Data Center, using Big Data tools such as Apache Hadoop, Apache SPARK and GlusterFS for storage and analysis.

This pilot will benefit the small pelagic fishing fleet, typically covering the North Atlantic Ocean, with the Norwegian pelagic fishing fleet as the main stakeholders. In this project they are represented by the ship owning companies Ervik & Saevik, Eros, Kings Bay and Liegruppen.

The main research partner will be SINTEF Ocean. The developed methods will in general be easily transferable to other fisheries.

The main motivation for this pilot is:

• improve catch revenue through improved fisheries planning, that is improving catch efficiency through reducing the time spent looking for fish and the fuel consumption.

The main KPIs of the pilot are:

• The fuel/time baseline varies significantly based on other factors, and the available sensor data makes it difficult to establish a “ground truth” about fish distribution.