Effects of atmospheric correction over inland waters on water quality parameters retrieval

The exploitation of remote sensing techniques for the detection of water quality parameters requires an accurate estimation of water reflectance, key physical quantity for parameters retrieval. The water-leaving signal generally only accounts for around 10% of the total signal recorded at sensor, while atmospheric contribution, due to gases and aerosol presence, is predominant. Thus, the atmospheric contribution removal (i.e. the atmospheric correction) is a crucial phase in image processing for water quality parameter estimation. For clear waters, where phytoplankton (chlorophyll) is the predominant optical component, atmospheric correction can be performed assuming water is almost dark in Near Infrared (NIR) bands, assumption on the basis of several algorithms. The phase becomes more tricky considering inland and coastal waters, where water leaving signal in the NIR is not null, and for which a more accurate characterization of atmospheric conditions could be required. The aim of this work is the evaluation of the effect of different data and techniques for atmospheric correction over inland or coastal waters, on the estimation of water quality parameters (chlorophyll and suspended matter concentration, transparency). Almost three different codes will be used for new generation satellites products correction: ATCOR, 6SV, and Seadas, all fed using alternatively ground-based or satellite-based measurements of atmospheric parameters, such as water vapor, pressure, ozone concentration, aerosol optical thickness and its optical parameters. Results (water reflectance values) will be inter-compared and validated through in situ measurements. Finally water reflectance will be used to feed a bio-optical model for the retrieval of water quality parameters, to evaluate possible improvement with increasing characterization of atmospheric conditions