Soil indicators for ecosystem services: a focus on water regulation

Abstract. Soil health assessments increasingly rely on indicators to infer soil functions and ecosystem services; however, the extent to which these indicators accurately represent water-related soil processes remains uncertain. This study investigates the relationships between soil properties and provision of water regulation ecosystem services across three contrasting pedo-climatic regions in Austria, Italy, and Tunisia. Using 315 soil profiles, we applied a process-based soil–water model to quantify infiltration, runoff triggering, groundwater recharge, and crop water stress index under representative climatic conditions. We evaluated commonly used soil indicators, including saturated hydraulic conductivity, available water content, bulk density, organic matter content, clay content, saturated soil water content, soil depth, and macroporosity. Pairwise correlation and multiple linear regression analyses were employed to assess interactions between soil properties and soil water balance components. Results show that indicator–process relationships vary considerably across sites and are often non-linear, with specific correlations reflecting local combinations of soil texture, structure, profile development, and climate. For example, in the Marchfeld region (Austria), infiltration exhibited a strong positive correlation with bulk density (r=0.74, p<0.001), while the crop water stress index showed a significant negative correlation with soil depth (r= −0.35, p<0.001). In the Bologna area (Italy), the study also indicated that groundwater recharge was positively correlated with soil macro-porosity (r=0.45, p<0.001), whereas macro-porosity exhibited a strong negative correlation with flux-to-runoff (r= −0.66, p<0.001), underscoring the key role of soil structural characteristics in controlling infiltration–recharge–runoff dynamics. In addition, multiple linear regression models were developed to assess the relevance of the individual soil properties and their interactions in controlling soil water balance components. For instance, the infiltration model for Marchfeld (r=0.79, p<0.001) was highly predictive and incorporated clay content, organic matter, and soil depth. Several widely used indicators exhibited weaker or inconsistent relationships with water-related processes than commonly assumed. For instance, saturated hydraulic conductivity alone was not a robust predictor of infiltration and recharge across sites, whereas soil depth and clay content emerged as recurrent controls, especially when considered jointly. Overall, this study highlights the value of process-based modelling for disentangling soil–climate interactions and cautions against the generalized use of static soil indicators in hydrological and soil health assessments.