Water surface temperature and salinity estimation from EO satellites for estuarine dynamics assessment in the Mediterranean and Black Sea

Estuaries are known as transition zones which modulate the freshwater inputs into the sea, with ocean salt water entering the river mouth and merging with the zero-salinity river streamflow. Understanding their dynamics is important for several purposes including the estimate of the salinization of inland waters and the effects in the thermohaline variability of the shelf to the open sea. The Copernicus Service Evolution Project EstuarIO proposes a low-to-high complexity modeling of the estuaries by  merging 1D box and 3D unstructured modeling approaches. The final aim is to better represent the river release (in terms of runoff, temperature and salinity) within the Copernicus forecasting Centres over the Southern European Seas. A source of uncertainty is that most estuaries are poorly monitored, river discharge measurements are taken far from river outlets, and salinity and temperature at the river mouths are mostly unknown. One of the EstuarIO objectives is to strengthen the calibration and validation of the estuarine models applied to target sites (Rhone, Po, Ebre and Danube deltas), using water temperature and salinity data derived from EO satellites. Landsat 8 and 9, along with other data sources such as MODIS are used as preliminary data sources for the riverine and coastal surface temperature (ST). The Landsat scenes used in the study were the L1TP (calibrated top-of-atmosphere reflectance and brightness temperature) data, with a combined repeat coverage of 8 days and spatial resolution of 30 m for the Operational Land Imager (OLI) multispectral bands and 100 m resampled to 30 m for the Thermal Infrared Sensors (TIRS) bands. Atmospheric correction and cloud masking were applied before retrieving the ST values. Current results suggest that the Landsat 8 and 9 imageries can be utilized to obtain high-resolution riverine and coastal ST data. A multilayer perceptron neural network based (MPNN) model is under testing in the target estuaries to estimate SSS values with in situ observations as benchmark to judge this innovative approach. Preliminary results on SSS extraction will be presented as well.