Feasibility study for an aquatic ecosystem earth observing system

Abstract

This study presents the work developed as a CEOS action for which CSIRO and DLR taken the lead on a feasibility assessment to determine the benefits and technological difficulties of designing an Earth observing satellite mission focused on the aquatic (non-oceanic) ecosystems. I n fact, many Earth observing sensors have been designed, built and launched with primary objectives of either terrestrial or ocean remote sensing applications. Often the data from these sensors are also used for biogeochemistry in inland, estuarine, deltaic and near coastal waters as well as for mapping macrophytes, macro-algae, sea grasses and coral reefs. However, such land and ocean specific sensors are not designed for these complex aquatic environments and consequently are not likely to perform as well as a dedicated sensor would. The results indicate that a dedicated sensor of aquatic ecosystems could be a multispectral sensor with about 26 bands in the 380-780 nm wavelength range for their ability to estimate: algal pigment concentrations of chlorophyll-a, accessory pigments, cyanobacteria pigments (cyanophycoerythrin and cyanophycocyanin especially) as well as other wavelengths relevant for phytoplankton functional types research; sun-induced algal fluorescence; suspended matter, possibly split up into organic and mineral matter; colored dissolved organic matter (CDOM) and discriminate terrestrial from marine sources; spectral light absorption and backscattering of the optically active components; transparency of water such as Secchi disk transparency, turbidity and spectral vertical attenuation of light. Then for optically shallow waters also the ability to measure and map: the water column depth with substratum type and cover; plants floating at or just above the water surface. Extending the range to shortwave helps the retrieval of suspend matter in very turbid waters and support research activities in mapping functional traits in floating/emerging vegetation. By including a specific band at 810 nm allows to explore lakes with high CDOM, since in these lakes the signal detectable by remote sensing sensors occurs as two peaks near 710 nm and 810 nm. Moreover, about 15 bands between 360-380 nm and 780-1400 nm are needed for removing atmospheric and air-water interface effects. These requirements are very close to defining an imaging spectrometer with spectral bands between 360 and 1000 nm (suitable for Si based detectors), possibly augmented by a SWIR imaging spectrometer. In that case the spectral bands would ideally have 5 nm spacing, although it may be necessary to go to 8 nm wide spectral bands to obtain enough signal to noise. The spatial resolution of such a global mapping mission would be between 17 and 33 m enabling imaging of the vast majority of water bodies (lakes, reservoirs, estuaries etc.) larger than 0.2 ha and 25% of river reaches globally (at 17 m resolution) whilst maintaining sufficient radiometric resolution. As spectral and spatial resolution are the core sensor priorities, the temporal resolution needs to be as high as is technologically and financially possible. In fact, although high revisit frequency is probably not critical for applications focused on benthic mapping and inventory, it is critical for tracking processes such as algal blooming.