A novel tool for adapting land atmospheric correction to water

Kurzfassung

A novel tool has been developed and implemented in the latest version 7 of the open-source software WASI (available at https://ioccg.org/resources/software/) which can be applied directly to Level-2 data to convert atmospherically corrected image data to Rrs. It is based on well-established physical models of water reflectance and specular reflections at the water surface. Two methods, inverse modelling and image-specific neural networks, are combined to process all water pixels. Errors caused by spectral ambiguities are quantified by comparing the results of the two methods. The application of the Rrs tool is illustrated using examples from different satellites (Sentinel-2, Landsat-9, EnMAP, DESIS) for different water types and varying sun glint conditions. Validation results are presented using multispectral in-situ measurements from AERONET-OC stations and hyperspectral in-situ measurements from field spectrometers. The algorithms used in the Rrs tool will be implemented in ESA’s atmospheric correction software for CHIME, which is currently under development at DLR.