Spatial Multiscale Agent-Based Fish Behavioural Model

The common sole (Solea solea) is a fish species of high commercial interest, subject to an increasing fishing activity. In order to assess and predict the availability of the fish stock under different fishing efforts, some of the co-authors have developed FIShPASs (FIshing Stock Probabilistic Agent-Based Simulator) [1], an agent-based model for the behavior of the common sole considering probabilistic aspects. In this work we generalize the model in order to support any fish species. The new model accounts for feeding information, taken from a food-web model [2], and for recruitment mechanisms, based on the SURBA (SURvey-Based Assessments) method [3]. The model is also extended with 3D spatial information, giving rise to two different spatial scales: a micro-scale describing in an individual-based approach the behavior of fish in a square kilometer of sea, according to the above-mentioned features; and a macro-scale that considers the Adriatic Sea as a discrete grid where each single cell represents a model at the micro-scale. This multiscale model can help to understand (at the micro-scale) the spatial dynamics of fish stocks over large areas, considering the migration through the fish life-cycle. The information on individual fish at the micro level feeds the macro model with detailed data in a multiscale simulation. The impact of fishing activity can be evaluated at both scales, in an “ecosystem approach to fisheries”.