An advanced global soil erodibility (K) assessment including the effects of saturated hydraulic conductivity

Surya Gupta; Pasquale Borrelli; Panos Panagos; Christine Alewell

doi:10.1016/j.scitotenv.2023.168249

An advanced global soil erodibility (K) assessment including the effects of saturated hydraulic conductivity

Sci Total Environ. 2024 Jan 15:908:168249. doi: 10.1016/j.scitotenv.2023.168249. Epub 2023 Oct 31.

Authors

Surya Gupta¹, Pasquale Borrelli², Panos Panagos³, Christine Alewell⁴

Affiliations

¹ Department of Environmental Sciences, University of Basel, Basel 4056, Switzerland. Electronic address: surya.gupta@unibas.ch.
² Department of Environmental Sciences, University of Basel, Basel 4056, Switzerland; Department of Science, Roma Tre University, Rome, Italy.
³ European Commission, Joint Research Centre (JRC), Ispra, Italy.
⁴ Department of Environmental Sciences, University of Basel, Basel 4056, Switzerland.

PMID: 37918722
DOI: 10.1016/j.scitotenv.2023.168249

Abstract

USLE-type models are widely used to estimate average annual soil loss at large scales, with the erodibility factor (K) being the sole component that accounts for soil's susceptibility to erosion. The factor includes the information on permeability in the equation, however, most definitions of the K factor consider the soil hydrological influence only very crudely and indirectly. Thus, the direct impact of surface runoff infiltration and drainage on soil erosion is largely neglected. The objective of this study is to incorporate soil hydraulic properties in the K factor map by merging available global-scale measured saturated hydraulic conductivity (Ksat) data with soil texture and organic carbon information into a modified K factor. To achieve this, the Wischmeier and Smith (1978) soil texture- and permeability-based equation (K_Wischmeier factor) was modified to include Ksat, called K_ksat factor. Using the Random Forest machine learning algorithm, the K_Wischmeier factor and the K_ksat factor were each correlated with soil and remote sensing covariates for spatial extrapolation of two independent K factor maps at 1 km spatial resolution. We noted a clear decrease in the mean value of the K_ksat factor (0.023 t ha h ha^-1 MJ^-1 mm^-1) compared to the mean value of the K_Wischmeier factor (0.027 t ha h ha^-1 MJ^-1 mm^-1). The reduction in K_ksat factor values was most pronounced in tropical regions reflecting the difference in soil properties (e.g., clay and iron), whereas other climate regions showed relatively minor changes in comparison to the K_Wischmeier factor as well as to the recent global modeling of Borrelli et al. (2017) (K_GloSEM factor maps). As many studies discussed an overall overestimation of (R)USLE based erosion rates compared to measurements, this reduction in the K factor might improve modeled erosion rates in the right direction. The K_ksat marks an important initial step in integrating hydraulic properties into the K factor of USLE-type models and can prove their significance in future studies.

Keywords: K factor; Random Forest; Soil hydraulic properties; Soil texture; Tropical regions; USLE.