The Impact of Atmospheric Correction Processors in Spatio-Temporal Fusion for Monitoring Chlorophyll-A Concentration in Inland Lakes

Abstract

Remote sensing technology has great potential in monitoring chlorophyll a (Chl-a), which is an important indicator of eutrophication in water bodies. However, the spatial and temporal continuity of remote sensing data are inevitably influenced by the limitation of sensor resolution and cloud contamination, which prevent the highly dynamic monitoring of water quality in inland median and small water bodies. Spatial and temporal fusion (STF) provides an effective way to address this issue. However, the errors might be introduced into the remote sensing reflectance () in the pre-processing and fusion process, which might bring large uncertainties in the derived Chla datasets. In this paper, an analytical study was designed to understand the influence of using different atmospheric correction processors for generating the images in STF, and the accuracy of the estimated Chla using the corresponding fusion images was validated with the in-situ samples. The experimental results show that ACOLITE DSF processor achieved the best performance for processing Multi-spectral Instrument (MSI) and Ocean and Land Color Instrument (OLCI) images in the atmospheric correction tests. Moreover, the machine-learning based Chla inversion accuracy of fusion images was comparable with that of real MSI images.