Automatic detection of Moroccan coastal upwelling from chlorophyll-a concentration images using the Expectation-Maximization method

The inshore water along the Atlantic coast of Morocco is part of one of the four major upwelling systems in the world. This region is characterized by the presence of upwelling activity almost all years around. In this work, we propose an unsupervised algorithm for identifying and extracting automatically this vital zone using the satellite observation of chlorophyll-a concentration (CHL-a). In fact, the approach is started with the application of Expectation-Maximisation algorithm for detecting the nutrient-rich region over CHL-a image. Then, a region-growing algorithm is applied to remove the noisy structures appeared in the offshore waters not belonging to the upwelling region. The proposed methodology is tested over a database of 230 weekly CHL-a images spanning the period from the years 2016 to 2020. The results show that the proposed method seems to be applicable for any other ecosystem.