Spatial variability and environmental drivers of the soil erodibility factor (K) at a regional scale in the Italian Alps

Abstract

Understanding and predicting soil erodibility in mountain environments is challenging due to complex interactions among environmental, pedological, and biological processes, which contribute to high spatial variability. This is particularly evident in the Aosta Valley Region (NW Italian Alps), where previous studies reported pronounced soil heterogeneity. Building on these findings, we estimated the topsoil erodibility factor (K factor of the USLE), assuming that, given the uniform texture, soil organic matter (SOM) would be the main driver of K variation. K was calculated using two equations—USLE and EPIC. We also tested, in a demonstrative way, SOM values beyond the conventional threshold of the USLE nomograph to explore its influence on K in highly organic alpine soils. A digital soil mapping (DSM) approach with machine learning was used to model the spatial distribution of K. Pedological field data were analyzed to evaluate their relationship with K, and USLE-based erosion values were calculated for observed profiles to assess K estimate reliability. Results show that: (i) the USLE K equation better captures mountain complexity; (ii) SOM significantly reduces K, with stone cover exerting additional influence; (iii) the model identified key regional drivers of K (carbon stock, elevation, pH), producing consistent spatial maps at 40 m resolution; and (iv) K values vary across soil horizons, humus systems, land uses, and soil types. Complementary analysis of erosional denudation supports the central role of SOM in enhancing alpine soil resistance. These findings provide insights for future soil monitoring, conservation, and restoration strategies in mountain ecosystems.