Summary
POLARIS represents a significant advance in high-resolution soil property characterisation, assembling harmonised pedological data from national soil survey databases into probabilistic maps at 30-metre resolution across the contiguous United States. By depth harmonising and aggregating over 21,000 soil series whilst integrating conventional soil maps to improve accuracy, the authors provide a resource intended to support land surface and hydrologic modelling at previously unavailable spatial detail. Although validation metrics indicate moderate predictive performance (R² ~0.41), the probabilistic approach and comprehensive variable coverage address a long-standing gap in high-quality, spatially-detailed soil parameter estimates for environmental modelling.
UK applicability
Whilst POLARIS is specific to United States soils and geology, the methodological framework for depth harmonisation, soil series aggregation and probabilistic mapping could inform development of equivalent high-resolution soil property databases for the United Kingdom, potentially supporting improved hydrological and land surface modelling within British farming and environmental management contexts. UK practitioners would need to work with comparable national soil survey resources (such as the National Soil Resources Institute datasets) to replicate this approach.
Key measures
Soil texture, organic matter content, pH, saturated hydraulic conductivity, Brooks–Corey and Van Genuchten water retention curve parameters, bulk density, saturated water content; spatial resolution at 30, 300 and 3,000 metres; R², RMSE and MAE statistics; 100-bin histograms per grid cell
Outcomes reported
The study presents POLARIS, a database of 30-metre probabilistic soil property maps covering the contiguous United States, integrating data from 21,481 soil series with measured variables including texture, organic matter, pH, hydraulic conductivity, water retention parameters, bulk density and saturated water content. Validation against in situ measurements yielded an average R² of 0.41, normalised root-mean-square error of 12%, and normalised mean absolute error of 8.8%.
Topic tags
Dig deeper with Pulse AI.
Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.