Satellite-based mapping of sediment dynamics and planform mobility in large river basins

Suspended sediment concentration (SSC) is a key indicator of river morphodynamics and water quality, yet long-term spatial monitoring remains constrained by sparse in-situ measurements. This study develops a satellite-based framework to estimate SSC along the Loire River (France) over 2005–2025 by combining Landsat 5/7/8/9 imagery processed in Google Earth Engine (GEE) with field SSC observations. A feed-forward neural network (R² = 0.94, RMSE = 3.65 mg/l) was trained on five spectral bands across three turbidity regimes and approximated by a compact surrogate model (R² = 0.89) suitable for operational GEE deployment. SSC maps reveal contrasting morphodynamic behaviour along the reach, with secondary-channel connectivity upstream and channel deepening downstream. River discharge emerged as the dominant SSC control, while rainfall has secondary importance. The approach demonstrates the scalability of combining machine learning and satellite imagery for long-term sediment monitoring in large rivers.