Operational Near Real Time Global Riverine Sediment Flux Estimates From Space

Abstract We present global estimates of riverine sediment flux derived from an open‐source suite of algorithms using the Harmonized Landsat Sentinel (HLS) and discharge products. We estimated Sediment Concentration (suspended sediment concentration [SSC]) via neural networks, trained and tested on matchups of in situ and HLS data, obtaining a median error of 13 mg/L and R 2 of 0.53. We paired SSC estimates with Surface Water and Ocean Topography (SWOT)‐derived discharge and GRADES‐hydroDL separately to provide 30 million estimates of suspended sediment flux between April 2023 and May 2025 as a proof‐of‐concept. Median difference on river reaches where HLS from Landsat and Sentinel sources are available is 24%. The SWOT Confluence SSC Module data product is designed to estimate SSC in every river in the SWOT river database every 2–3 days and are stored alongside SWOT discharge at the Physical Oceanography Distributed Active Archive Center (PO.DAAC) to enable sediment flux estimates through Earth's large rivers.