This product contains 4 parameters of water quality: the high permanganate index (COD_MN), total nitrogen (TN), total phosphorus (TP), and ammonia (NH₃), and each parameter unit is mg/L. The above 4 parameters are non-optically active. How to build satellite retrieval models of non-optically active parameters suitable for different water bodies is the main difficulty at present. From the data-driven perspective, this algorithm is based on Sentinel-2 satellite images and in-situ measurement data of the water quality parameter from the automatic water quality monitoring station in the Zhejiang Province. This dataset covers different optical water bodies and different seasons, as well as a wide range of concentrations. Based on this dataset, combined with the extreme gradient improvement tree (XGBOOST) machine learning algorithm, satellite retrieval models for the above 4 water quality parameters suitable for inland water bodies in the Zhejiang Province were developed.

The dataset provides the monthly satellite retrieval products of permanganate, total nitrogen, total phosphorus and ammonia nitrogen in the surface water of the major rivers entering the sea. The time span is from 2016 to 2023, with a spatial resolution of 10 meters and a monthly averaged temporal resolution.