Spatiotemporal variations in Pearl River plume dispersion over the last decade based on VIIRS-derived sea surface salinity

A river plume indicates the dispersion and transport path of pollutants from runoff, monitoring the spatiotemporal variation of river plume distribution from space is crucial for marine environmental governance. This study focuses on the Pearl River Plume (PRP), and takes the Pearl River Delta coastal waters, China as study area. We developed a machine learning-based sea surface salinity (SSS) estimation algorithm for the Visible Infrared Imaging Radiometer Suite (VIIRS) on-board the Suomi National Polar-orbiting Partnership satellite (SNPP), leveraging extensive field-measured SSS data from the study area. Independent validation of the algorithm produced an R² of 0.89 and a mean relative percentage error of -1.29 %. By applying the algorithm to long-term VIIRS/SNPP imagery (2012-2022), we generated seasonal, monthly, and annual SSS maps. Using these SSS data, we conducted a detailed analysis of PRP's spatiotemporal distribution and occurrence frequency. Furthermore, we examined the impacts of the El Niño-Southern Oscillation (ENSO), Pearl River discharge, wind forcing, and precipitation on SSS and PRP variability. Our findings demonstrate that the machine learning-derived SSS estimates effectively capture river plume dynamics, providing valuable insights into freshwater transport processes. These estimates contribute to a better understanding of coastal hydrodynamic processes, supporting marine environmental management and sustainable coastal development.