Quantifying decadal stability of lake reflectance and chlorophyll-a from medium-resolution ocean color sensors

Multi-decadal time-series of Lake Water-Leaving Reflectance (LWLR), part of the Lakes Essential Climate Variable, have typically been interrupted for the 2012–2016 period due to lack of an ocean color sensor with capabilities equivalent to MERIS (2002−2012) and OLCI (2016 - present). Here we assess, for the first time, the suitability of MODIS/Aqua to estimate LWLR and the derived concentration of chlorophyll-a (Chla) at the global scale across optically complex water types, in an effort to fill these information gaps for climate studies. We first compare the normalized water-leaving reflectance (Rw) derived from two atmospheric correction algorithms (POLYMER and L2gen) against in situ observations. POLYMER shows superior performance, considering the agreement with in situ measurements and the number of valid outputs. An extensive assessment of nine Chla algorithms is then performed on POLYMER-corrected Rw from MODIS observations. The algorithms are tested both in original parameterizations and following calibration against in situ measurements of Chla. We find that the performance of algorithms parameterized per Optical Water Type (OWT) allows considerable improvement of the global Chla retrieval capability. Using 3 years of overlapping observations between MODIS/Aqua and MERIS (2009–2011) and OLCI (2017–2019), respectively, MODIS-derived reflectance and Chla products showed a reasonable degree of long-term stability in 48 inland water bodies. These water bodies, therefore, mark the candidates to study long-term environmental change.