Validation of remote sensing reflectance derived with WASI from EnMAP L2A land data using AERONET-OC data

Kurzfassung

To retrieve remote sensing reflectance, Rrs, from satellite images of water bodies, it is necessary to subtract the surface reflections of the sun and sky (glint), Rrssurf. A newly introduced Rrs tool from the software WASI allows to simulate Rrssurf for bottom of atmosphere reflectance images, which are affected by glint and minor path radiance errors from atmospheric correction. This work presents examples of Rrs spectra obtained with WASI for eight EnMAP scenes, which were atmospherically corrected using EnMAP's L2A processor for land, PACO. For each EnMAP scene, atmospheric parameters (ozone column, AOT, water vapor) are input into WASI and the user optimizes the processing parameters individually. Then, a pixel-wise glint correction estimates the sky radiance reflected at the water surface in sensor direction as fractions of three components originating from direct solar radiation (sun glint) and diffuse components caused by molecular and aerosol scattering (sky glint). The generated glint image is subtracted from the input L2A image, resulting in a glint corrected Rrs image. Two scenes each are chosen at four AERONET-OC sites: Venice Lagoon (AAOT), Bahia Blanca, Chesapeake Bay and Lucinda. The in-situ data that is timely closest to each EnMAP acquisition is chosen to validate the Rrs results, which are also compared with EnMAP's standard water L2A product. A strong correspondence between the WASI derived Rrs and AERONET-OC measurements is achieved for all scenes and sites. The aquatic Rrs spectra obtained from the EnMAP land product closely resemble those from the standard water product. In the case of Bahia Blanca, the high concentration of total suspended matter leads to a glint overestimation below 500 nm. Further investigation is needed to determine if additional fine-tuning of the WASI processing parameters might compensate this effect. Overall, the results validate the glint correction method implemented in the new Rrs tool.