Abstract
Water salinity is one of the important environmental factors of the sea and plays a significant role in the study and prediction of the oceanic surface currents, location analysis of the fish aggregation, density determination and studying its changes, and also in ecological properties. This parameter changes greatly with time and location, and proper recognition of it requires measurements at short time intervals (monthly) of multiple points in the study area.
In traditional ways, the assessment and evaluation of one or several specific factors of water quality is often costly and time-consuming, and cannot be a good indication for the entire area of a vast region. But in recent years, satellite and remote sensing technology have been considered as an appropriate tool for evaluating some water quality parameters because, given the digitality of these data, their wide availability, regular measurements, their repetition in short periods of time, Less cost and time, a wide range of projects can be achieved. The purpose of this study is mapping sea surface salinity of the Persian Gulf in Iran and the Gulf of St. Lawrence in Canada using MODIS satellite imagery. In this regard, a software has been produced in Iran for the first time that can prepare salinity, temperature and density maps of the sea surface in three different models with proper accuracy by entering the MODIS satellite imagery and CTD field data. High capability and flexibility of the Artificial Neural Network in approximation of nonlinear and linear continuous functions in hybrid space, led this study to provide a new method based on using this network in which salinity map is determined by a multilayer perceptron network.