Assessing pan-European soil resilience by combining modelled soil erosion with LUCAS topsoil indicators

Abstract

Background Soil erosion and topsoil chemical condition are usually assessed separately, although both shape how soils respond to land use pressure. This study tested whether the two are related across Europe and how robust composite indicators of topsoil condition are between surveys. Methods National mean soil loss for 2001 and 2012 was derived from the Global Soil Erosion Modelling platform (GloSEM) for 28 countries; in this model, differences between the two years reflect land use and vegetation-cover change. European maps of rainfall erosivity, soil erodibility and support practices were sampled at LUCAS points. Soil pH, electrical conductivity and soil organic matter from the LUCAS Topsoil surveys of 2009, 2015 and 2018 (18,744–21,859 points each) were combined into a threshold-based Soil Resilience Index, tested against a linear formulation and on revisited points. Results National soil loss changed little between 2001 and 2012 (13 countries increased, 15 decreased; p = 0.15), with largest decreases in Poland (−13.7%) and Italy (−12.4%). It followed rainfall erosivity closely (Spearman ρ = 0.90). Topsoil pH was stable: 81.0% of 12,898 revisited points kept their pH class between 2009 and 2018. The index was highest under shrubland and grassland and lowest under cropland and bare land. Soil loss was not associated with the index across 27 countries (ρ = −0.04, p = 0.83), and rainfall erosivity was unrelated to it at 40,501 points (ρ = 0.02). The change in the index between 2015 and 2018 depended on its formulation and coincided with systematic inter-survey differences in pH and conductivity. Conclusions Erosion pressure and topsoil chemical condition are largely independent at the European scale, so neither can stand in for the other in soil health assessment. Composite indices of topsoil condition are sensitive to formulation and survey comparability, and temporal assessments should use revisited points and report sensitivity analyses.

The authors' abstract, as published at the source. Open Research Europe, 2026 · DOI ↗