Deciphering riverine CODMn sources: driver identification and carbon cycle impacts

Abstract

The large-scale monitoring of riverine organic pollution is of significant practical importance for river basin ecological management and sustainable development to replace the traditional monitoring stations. The development for remote sensing of riverine chemical oxygen demand (COD Mn ) has pressing demand as a low cost and high effective tool for addressing water quality degradation and dissolved organic carbon (DOC) estimation. In this study, we develop remote sensing monitoring models of COD Mn in individual rivers across China using Sentinel-2 images and national cross-section datasets. The performance of R 2 for main rivers in China with eXtreme Gradient Boosting (XGBoost) algorithms could reach over 0.80. The retrieved average COD Mn from 2021 to 2023 were 3.70 mg/L, 3.53 mg/L and 3.33 mg/L respectively, and the COD Mn trends to be decreased significantly ( p < 0.05). The accumulation influences of precipitation to COD Mn in southern rivers was observed while the dilution effect to COD Mn was shown in northern rivers. Areas where crop land and impervious cover are the dominant land types have significantly higher COD Mn for all rivers. Also, the increase in the number of industrial parks did not lead to an increase in riverine COD Mn . With the relationship of COD Mn and DOC, the total DOC flux from river to sea was ranged from 3.71 to 6.30 Tg C. By bridging remote sensing innovation with water quality parameter of COD Mn in rivers, this research advances sustainable river management and carbon cycle strategies.