Decadal Erosion Rates and Sediment Buffering Identified Through Enhanced DEM Differencing Using Underutilized Global Satellite DEMs

Abstract

Quantifying decadal-scale erosion rates in tectonically active regions is essential for assessing landscape hazards and constraining sediment budgets. A key question in Earth surface processes is how contemporary erosion measurements influenced by recent climatic extremes relate to long-term geological rates. This study investigates this relationship by evaluating two decades of topographic change (2000–2020) in Taiwan's Laonong River Basin (1,373 km2), a mountainous landscape characterized by active landslides, steep relief, and intense rainfall, including Typhoon Morakot in 2009, the highest-rainfall typhoon on record in Taiwan. Several satellite-derived digital elevation models (DEMs) were benchmarked against high-resolution LiDAR to identify the most suitable data set for long-term change detection. The reprocessed SRTM NASADEM was identified as the optimal DEM due to minimal artifacts and limited void filling, with further improvement achieved through canopy removal and frequency-domain bias correction. DEM differencing across sub-catchments revealed that erosion rates averaged ∼15.3 mm/yr, reaching ∼551 mm/yr in landslide-affected tributaries. However, approximately 87.6% of the eroded material was stored within the basin, predominantly in valley-fill deposits, reducing effective net lowering across the basin to 5.06 ± 1.60 mm/yr. This rate is consistent with long-term (103–106 yr) cosmogenic and thermochronological estimates (∼3–6 mm/yr), indicating that sediment buffering spatially averages extreme localized erosion over decadal timescales. These findings highlight the role of internal sediment storage in shaping basin-scale erosion patterns and emphasize the importance of high-resolution erosion assessments for capturing geomorphic heterogeneity. The results also demonstrate the potential of underutilized global DEMs, once corrected, to resolve meaningful geomorphic signals in mountainous terrain.