Remote sensing of brine shrimp cysts in salt lakes

Using laboratory and field experiments as well as spectral analysis of satellite images, we demonstrate that surface aggregations of brine shrimp ( Artemia ) cysts (BSC) in salt lakes can be identified unambiguously in satellite imagery. This is because of the unique reflectance spectral shapes of the BSC image slicks, where a sharp and monotonic increase in reflectance is found at wavelengths >550 nm and two inflection points are found around ~550 nm and ~ 650 nm. Such spectral characteristics differentiate BSC slicks from other floating matters. Based on this principle, a deep learning model is developed to extract BSC features from MERIS (Medium Resolution Imaging Spectrometer, 2002–2012) and OLCI (Ocean and Land Color Instrument, 2016 - present) satellite images to quantify BSC abundance, spatial distribution patterns, and their temporal changes in the Great Salt Lake (GSL), the world's largest contributor of BSC commercial products. A clear seasonality is found in BSC abundance, with the primary peak in April – May and secondary peak in October – November. The two peaks may be explained by food availability to brine shrimp. The inter-annual variability and the recent increasing trend in BSC abundance, on the other hand, are difficult to explain by fluctuations in wind, temperature, or salinity, while recent increase in commercial harvest does not appear to be associated with the variability in BSC abundance estimated by satellites. Because many salt lakes around the world, for example the Aral Sea, Lake Urmia, and the Dead Sea, also show BSC slicks in satellite imagery, this study suggests that it is possible to perform a systematic evaluation of BSC abundance and possibly brine shrimp populations in all major salt lakes, especially under a changing climate and increased human activities. Surface aggregations of brine shrimp ( Artemia ) cysts (BSC) in salt lakes show unique reflectance spectral shapes, therefore can be identified from passive optical remote sensing imagery. Based on the spectral and spatial characteristics of surface aggregations of BSC, a deep learning model is developed to detect and quantify surface BSC in the southern arm of the Great Salt Lake from satellite imagery of 2002–2012 and 2016–2020. Results show clear seasonality, inter-annual variability, and a recent increase in BSC amount. While such information may complement those obtained from shipborne surveys to help understand the lake's ecology and fisheries, the study represents the first of its kind in satellite remote sensing of shrimp eggs, which may be extended to other salt lakes on Earth to have a systematic evaluation. • Reflectance of brine shrimp ( Artemia ) cysts (BSC) show unique spectral shapes. • A deep-learning model is developed to detect BSC slicks in MERIS and OLCI imagery. • Linear unmixing is applied to quantify BSC abundance in the Great Salt Lake (GSL). • BSC in the GSL shows seasonality, inter-annual variability, and recent increases. • BSC image slicks are also found in the Aral Sea, Lake Urmia, and the Dead Sea.