Characterising retrieval uncertainty of chlorophyll-a algorithms in oligotrophic and mesotrophic lakes and reservoirs

Remote sensing product uncertainties for phytoplankton chlorophyll-a (chla) concentration in oligotrophic and mesotrophic lakes reservoirs were characterised across 13 existing algorithms using an situ dataset of water constituent concentrations, inherent optical properties (IOPs) remote-sensing reflectance spectra Rrsλ collected from 53 (346 observations; chla < 10 mg m-3, median 2.5 m-3). Substantial shortcomings retrieval accuracy evident with absolute percentage differences (MAPD) > 37% mean (MAD) 1.82 m-3. Using the Hyperspectral Imager Coastal Ocean (HICO) band configuration improved accuracies by 10–20% compared to Land Colour Instrument (OLCI) configuration. Retrieval attributed biogeochemical machine learning models through SHapley Additive exPlanations (SHAP). The uncertainty most semi-analytical was primarily determined absorption composition. Machine showed relatively high sensitivity light coloured dissolved organic matter (CDOM) non-algal pigment particulates (NAP). In contrast, red/near-infrared algorithms, which aim lower presence CDOM NAP, explained total at 673 nm (aϕ(673)) variables related backscatter. Based on these characterisations we discuss suitability evaluated algorithm formulations, make recommendations estimation improvements oligo- reservoirs.