Accuracy assessment of input parameters for water column correction approach for case 2 waters

Numerous approaches characterising the radiation field of a water column have been developed and correction attempts for remote sensing data have been applied successfully. Various algorithms describe the complex interaction of biophysical parameters with down- and upwelling radiation in a water body and form the basis for water column correction. Parameters such as varying bottom reflectances and bathymetry aggravate an accurate parameterization of water column correction models. Applying these models, special interest lies in their sensitivity to both quality and accuracy of model input parameters. In this paper we discuss the sensitivity of the water column correction model MIP2 to bio-physical parameters, i.e. suspended matters (SM) and chlorophyll (CHL), in case 2 waters. In August 2010, hyperspectral AISAeagle data have been acquired; in-situ measurements were conducted concurrently to the airborne campaign. The study was conducted at the rocky shores of the island Helgoland (North Sea, Germany). The study area is characterised by a heterogeneous water body resulting in varying and spatially uncorrelated concentrations of SM and CHL, which aggravate an accurate water column correction. During analysis, special focus is set on areas with varying water characteristics such as vegetated bedrock, shallow sandy spots and deep water areas. Water column correction is performed using a sub-module of MIP, i.e. WATCOR. Reflectance deviation results show that variations of SM concentrations have a stronger influence than variations of CHL within the water column correction. Whereas, the shallow sandy spots reveal the highest sensitivity at constituent concentration variation followed by the deep water and the vegetated bedrock areas.