HMM Imaging of the Aquatic Ecosystem

Abstract

The oceans are monitored since decades using multispectral sensors on satellite, but very little information is available for the majority of inland waters since they are too small for ocean colour satellites, optically too complex for most multispectral sensors and too numerous (around 120 million lakes > 15 m) for traditional sampling. Since inland waters cover less than 4 % of Earth's surface, their importance for global processes has long been overlooked, but new data indicate that they may be more important than the oceans in some aspects, e.g. they bury twice as much carbon from the atmosphere by sedimentation. A number of hyperspectral space sensors will be launched in the next years whose resolution of 30 m is suited to monitor water quality of nearly 90 % of the lake areas. I present the principles of the models that are used to analyse hyperspectral data over water and discuss the potential and challenges of hyperspectral imaging for optically complex water types.