The development of reliable and scalable biological indicators is a major challenge for the implementation of the EU Soil Strategy and Mission Soil. Among soil mesofauna, Collembola (springtails) are key regulators of decomposition processes, microbial dynamics, and nutrient cycling, yet their potential as indicators of soil health remains insufficiently explored at continental scale.
Within the Horizon Europe SOB4ES project, we analysed Collembola communities across more than 430 sites distributed among nine European pedoclimatic regions and contrasting land-use systems. Using a harmonized dataset of sampling plots integrating biological, soil, climatic, management, and Earth Observation data, we identified the key environmental drivers and anthropogenic pressures shaping Collembola abundance and diversity.
Biological observations were integrated with soil physico-chemical properties, climatic variables, topographic descriptors, management information, and Earth Observation products. Machine-learning models and explainable artificial intelligence (SHAP) were applied to identify the factors governing Collembola abundance and diversity.
The analyses revealed a striking contrast between density and species-richness responses. Collembola density showed relatively low predictability, reflecting strong local-scale variability and dependence on soil structure, moisture conditions, and climate. In contrast, species richness was predicted with very high accuracy by the XGBoost model (R² ≈ 0.92), indicating a strong and consistent response to broad environmental gradients. Soil pH emerged as the dominant predictor, followed by precipitation, soil moisture, nitrogen availability, temperature, and pedoclimatic context. Species richness increased under alkaline, nutrient-rich, and moist conditions, while specific soil and landscape characteristics further modulated diversity patterns.
These findings demonstrate that Collembola species richness is a robust and sensitive indicator of soil ecological condition across Europe. By combining standardized biodiversity monitoring with Earth Observation data and explainable artificial intelligence, this study provides a scientifically grounded and operational framework for integrating soil biodiversity into future European soil health monitoring systems.
Authors: Cristina Fiera¹, Minodora Manu1, Ioana Vicol1, Monica Mitoi1, Constantin-Tiberiu Sahlean1, Dariusz Skarżyński², Jörg-Alfred Salamon3, Marjetka Suhadolc4, Karen Vancampenhout5, Wim H. van der Putten6,7, George Zalidis8, María Jesús Iglesias Briones9, and SOB4ES consortium