Author(s): Sergio Martinez-Aranda; Jose Segovia-Burillo; Mario Morales-Hernandez; Daniel Caviedes-Voullieme

Linked Author(s):

Keywords: Soil erosion; Sediment transport; GPU-based computation; HPC; Water-driven erosion

Abstract: Hydraulic soil erosion is an increasingly critical issue worldwide driven by intensive farming, wildfires, and extreme hydrometeorological events. Many approaches have been put forward to model soil erosion, and thus provide valuable information to researchers, stakeholders and managers. However, these models are often heavily parameterized, demand careful calibration, and have limited applicability into highly resolved and spatially distributed problems. In contrast, physically-based approaches leveraging on shallow water and sediment transport equations can potentially resolve local erosion phenomena, strongly coerced by topographic features and human structures, and effectively redistribute and transport sediment through the system (Fernandez-Pato et al., 2020). The cost is of course the associated computational effort. Nevertheless, the spatially resolved information is arguably of higher value, and the physical generality of the model relaxes the constraints imposed by calibration on specific events or conditions. Thus, this type of model may become a central tool to assess future soil loss under climate change.

DOI:

Year: 2025