The prediction of distribution pattern of Saurida tumbil in the northern Persian Gulf using GIS and ANN

In conformity to study and predict the distribution pattern of lizardfish (Saurida tumbil) in the northern Persian Gulf (Hormuzgan province waters), the catch data of 2008 were collected and analyzed. The primary statistical analysis were done using Excel and then distribution maps were provided using GIS–software based on Catch Per Unit of Area index (CPUA). Then to forecast the distribution pattern, the maps of different physical and chemical parameters of sea water consist of temperature, turbidity, salinity, sigma T, oxygen, pH, chlorophyll a, electrical conductivity, sound speed, depth, distance from coast, geographical position and time of towing were prepared and were used as independent value to be assessed with CPUA of lizardfish as dependent value. These maps were used as input of Artificial Neural Networks (ANNs) software of which 60% of data for training, 20% for testing and 20% for validation were applied to prepare the best ANN model. With applying this model on catch and CPUA data, the distribution pattern with emphasize on fishing ground can be predicted for further fisheries management and leading the fishing activities to be concentrated in fishing grounds.