Decoupling of Basin Soil Erosion and Riverine TSM under Reservoir Regulation: A 40-Year Satellite Analysis

This dataset contains annual reach-scale mean Total Suspended Matter (TSM) concentrations (mg L⁻¹) for the main stem of the Songhua River from 1984 to 2023.

TSM concentrations were derived from Landsat TM, ETM+, and OLI top-of-atmosphere (TOA) reflectance imagery during the ice-free season (May–October). A Random Forest regression model was developed and calibrated using 542 in situ riverine samples collected during 2021–2022. The calibrated model was then applied consistently across the Landsat archive to generate annual TSM estimates.

Following pixel-level retrieval, the main stem was divided into four consecutive reaches (LR–SRH, SRH–FRH, FRH–JMR, and JMR–SRJ). Annual mean TSM concentrations were calculated for each reach.

The Excel file includes the following fields:
(1) Year (1984–2023)
(2) Reach (river segment identifier)
(3) Mean_TSM (annual mean TSM concentration, mg L⁻¹)

This dataset is suitable for analyzing long-term trends in riverine suspended matter concentration and for examining statistical relationships between basin soil erosion and in-stream TSM dynamics.

It should be noted that the dataset represents concentration-based estimates only and does not include sediment flux or sediment load information.