Soil hot spots: why global models fail in complex areas – A scientific study

Can we rely on global datasets to tell us what is happening locally in our farms? Our latest study, co-funded by the EU under the Path4Med project, suggests that in geomorphologically and geologically complex areas, we need to dig deeper.

Our study provides a rigorous accuracy assessment of widely used global soil datasets (ISRIC and ESDAC) against over 10,000 local field ground-truth data in Greece. We identified significant discrepancies, including low classification accuracy (19–21%) and a failure to capture extreme soil values. Crucially, we found that these errors are not random but are clustered in specific “hot spots” linked to complex geomorphological and geological features.

This publication represents foundational work for the Path4Med project. Our findings confirm that we cannot rely solely on global models for regional decision-making. This research is the first step in our strategy to generate improved, locally calibrated datasets. By understanding exactly where current models fail, we can better target our efforts to enhance soil health monitoring across our demonstration sites, ensuring our solutions are robust and applicable to real-world conditions.

Please check the full open-access paper here: https://www.mdpi.com/3605342

Gerontidis, S.; Soulis, K.X.; Stavropoulos, A.; Nikitakis, E.; Kalivas, D.P.; Kairis, O.; Kopanelis, D.; Soulis, X.K.; Palli-Gravani, S. Assessment of the Accuracy of ISRIC and ESDAC Soil Texture Data Compared to the Soil Map of Greece: A Statistical and Spatial Approach to Identify Sources of Differences. Soil Syst. 2025, 9, 133. https://doi.org/10.3390/soilsystems9040133