Improving Sediment Source Tracing Accuracy by Coupling Particle-Size Information

The particle size sorting effect during erosion significantly undermines sediment source tracing accuracy. This study proposes an innovative strategy integrating particle size information to enhance traditional methods. Laboratory experiments using sediments with fine-particle enrichment ratios showed that under strong sorting, traditional methods incurred high errors (MAE: 29.3%), while the new strategy reduced MAE to 6.8% by optimizing fingerprint matching. Validation via simulated rainfall across slopes (5°–25°) and intensities (60–120 mm/h) confirmed broad applicability. MAEs for geochemical and mid-infrared spectral tracing remained below 9.7% and 8.3%, respectively, significantly outperforming conventional methods. In high-sorting scenarios (e.g., 5°/120 mm/h), accuracy gains ranged from 31.3% to 64.4%. Mechanistically, the strategy isolates sorting interference, preserving source-like fingerprint characteristics. It also reveals spectral tracing's superior sensitivity to mineralogical differences in fine fractions, supporting multi-dimensional tracing systems. This research provides a breakthrough in overcoming the long-standing constraint of particle size sorting, with practical implications for understanding erosion and enabling precision source management.