Sensitivity study for aquatic ecosystem monitoring with the DESIS hyperspectral sensor

Abstract

The aquatic ecosystems of coastal and inland waters are more variable than the open ocean as water constituents and bottom substrates differ considerably in type, concentration and optical properties. Sensors with high spatial, spectral and radiometric resolution are therefore required to provide enough detail for mapping these highly complex environments. A new Earth observation instrument, designed for monitoring spectrally complex areas, is the DLR Earth Sensing Imaging Spectrometer (DESIS). This hyperspectral sensor, which covers a spectral range from 450 nm to 950 nm and has a spatial resolution of 30 meters, will be launched to the ISS in summer 2018. The goal of the present study is to specifically evaluate the performance of DESIS over aquatic ecosystems based on its sensor specifications. For this purpose, a sensitivity study was conducted in order to explore the spectral and radiometric properties of DESIS in the context of optically deep and shallow water mapping. The optical water properties and benthic cover types were chosen to represent typical inland and coastal waters. Forward simulations of hyperspectral measurements were made using the Water Color Simulator software WASI to study the expected DESIS radiances and their signal-to-noise ratios (SNR) for different atmospheric conditions. The impact of sensor noise on the retrieval of water constituents, benthos types and water depth was analyzed by applying inverse modelling to these reflectance spectra. The study assesses the mapping potential and limitations of a DESIS type sensor for these complex environments, which are usually much darker than land surfaces.