Chlorophyll-a estimation from remote sensing data of sea surface temperature and aerosol optical thickness through a shallow neural network

The oceans cover most of the Earth surface, being therefore essential elements of the environmental balance of our planet. In this sense, the prediction of global change scenarios that may affect them is an issue of high scientific and social relevance. One of the elements that indicates the quality of the water is the concentration of Chlorophyll-a. It is well known that Chlorophyll-a is related to the sea surface temperature and other variables such as the presence of nutrients and wind. All of them have been monitored with remote sensing satellites for more than a decade ago. Thus, researchers have available temporary series of these data. In this work, the prediction of Chlorophyll-a concentration is addressed from data on sea surface temperature and the aerosol optical thickness. For this, a shallow neuronal network is designed and trained, whose performance is contrasted with other approaches. The results show that the tested methodology can be used to model predictors with the discussed climate variables.