A Spatiotemporally Seamless Chlorophyll-a Concentration Dataset for Typical Lakes in the Yangtze–Huai River Basin

After SeaDAS OCSSW radiometric and atmospheric correction, lake cropping, cloud and land masking, outlier pixel removal, and Chl-a inversion, missing observations were filled in using DINEOF. Three lakes are independently reconstructed and processed within a one-year window; Values less than 0 μ g/L or greater than 100 μ g/L are referred to as missing values, and the missing positions are initialized with 0 without standardization or mean centering.

Select 30% of the valid observations from each lake annual matrix for cross validation, and choose the mode with the smallest cross validation error among 1-25 EOF modes to complete the final reconstruction. The temporal resolution of the dataset is 1 d, the spatial resolution is 250 m, and the unit of Chl-a is μ g/L. The total data volume is about 230 GB. The data is divided into three sub datasets according to the Taihu Lake Lake, Chaohu Lake and Hongze Lake.