Validation of Landsat-8 and Sentinel-2 algorithms for atmospheric correction on Lake Geneva

Remote-sensing of water surfaces such as large lakes is a useful tool for monitoring and understanding large-scale processes. Suspended particles in lakes can have various sources such as autochthonously produced organic matter, urban or agricultural runoff or rivers inflows. Understanding their sources and impacts on natural waters is highly important as they are the mirror of the ecological well being. As suspended solids can be retrieved by remote-sensing based on their reflectances, satellite images are an important tool to study the lake and detecting important processes and events. Before using the reflectances measured by the sensor for monitoring lakes, atmospheric corrections have to be applied on the image for retrieving accurate data. This project compares three different algorithms for atmospheric corrections, namely Acolite developed in Belgium, OPERA developed at Vito and C2RCC designed by Brockmann Consulting, applied on the case study, the Lake of Geneva. The resulting reflectances from the different algorithm settings are compared to the in-situ water-leaving reflectance measurements. Additionally, the total suspended matter concentrations derived from the output reflectances were validated with in-situ concentrations. C2RCC with elevation indication performs a good atmospheric correction and yield good particulate concentrations by non-ideal atmospheric conditions. Acolite with a linear parameter of 9.5 gives good results by perfect weather conditions. This projects depicts guidelines and recommendations for performing atmospheric corrections over the Lake Geneva with these algorithms.