Retrieving aerosol characteristics and sea-surface chlorophyll from satellite ocean color multi-spectral sensors using a neural-variational method

Abstract

We developed a two-step algorithm for retrieving and then monitoring the concentration of Saharan dusts and of the sea-surface chlorophyll from satellite ocean-color multi-spectral observations. The first step consisted in classifying the top of the atmosphere (TOA) spectra using a neuronal classifier, which provided the aerosol type and a first-guess value of the aerosol parameters that was used to initialize the variational method. The variational method was the second step, which retrieved accurate measurements of the aerosol and chlorophyll-a concentrations. The algorithm was conditioned to take into account the absorbing aerosols, such as the Saharan dusts. We used this algorithm to analyze 13years of SeaWiFS images (September 1997–December 2009) over an area of the Atlantic Ocean off the coast of West Africa. Since our method allowed us to take Saharan dusts into account, the number of pixels processed for retrieving the chlorophyll-a concentration was an order of magnitude higher than that processed by the standard SeaWiFS algorithm. The analysis of the SeaWiFS images showed that the Saharan dust concentration was maximal in summer during the rainy season and minimal in autumn, which could be explained by the seasonal variability of dust emission triggered by mesoscale atmospheric processes (low-level jet and convection) and soil characteristics (humidity and vegetation).