Coupled Aerosol and Water Constituent Retrieval Algorithm for Hyperspectral Remote Sensing Sensors over Case II Water Lake Constance

Kurzfassung

A physically based method for the inversion of hyperspectral remote sensing data over case II waters is presented and applied to data of the satellite sensor CHRIS-PROBA. Field campaigns were carried out in Lake Constance simultaneously to the satellite overflights. Measurements of the atmospheric optical depth, of hyperspectral subsurface reflectance spectra and water constituent concentrations are used for validation and optical closure experiments. The data simulation and image processing is performed by use of the Modular Inversion Program MIP. The radiative transfer in MIP is simulated in a coupled, plane-parallel atmosphere-ocean system based on the FEM-method. Results of simulation are stored independent of sensor and recording conditions in extensive databases. The reduction to specific recording conditions and the adjustment of algorithms to the sensor characteristics are supported automatically in MIP. The libraries and analytical parameterizations of the subsurface irradiance reflectance form the search space for the retrieval algorithms. Here, 6 parameters are determined iteratively for each pixel by searching the set of media parameters for which the calculated radiance spectrum is closest to the measured spectrum. In this way, optical depths of 3 aerosol types, concentrations of suspended matter, chlorophyll, and Gelbstoff are obtained. These results are compared with the in-situ measurements, and an assessment of error sources is given