Comparison of phycocyanin concentrations in Chaohu Lake, China, retrieved using MODIS and OLCI images

Phycocyanin (PC) concentration is used as an indicator to characterize cyanobacteria biomass while monitoring eutrophication in inland water. Remote sensing provides useful methods for quantifying PC concentration; however, there a shortage of datasets the long-term when only single remote data used. Therefore, concentrations obtained from multisource images should be compared before integrating them monitoring. In this study, machine learning (ML) regression algorithms are develop retrieval models suitable Moderate Resolution Imaging Spectroradiometer (MODIS) and Sentinel-3 Ocean Land Colour Instrument (OLCI) images, their accuracies compared. The two optimal applied satellite acquired on same days compare spatial consistency results. results show that sensitive spectral range 560–680 nm. Among ML algorithms, gradient boosted tree (GBT) exhibits highest accuracy both MODIS ( R 2 = 0.82, RMSE 61.9 μg/L) OLCI 0.86, 45.44 μg/L). retrieved bloom no-bloom periods have high most areas Chaohu Lake. Their correlation coefficient also exceeds 0.7, average relative error reaches 0.293 μg/L. However, large difference exists with concentrations, which may cause by poor applicability atmospheric correction these areas. proposed developed using GBT paper can expend idea quantitative other water quality parameters water, conclusions enable effective integration time series reservoirs lakes.