Regional prediction of soil organic carbon content over temperate croplands using different multiscale measurements by hierarchical modelling

Spatial monitoring of soil quality especially in terms of soil organic carbon (SOC) content for the purpose of evaluating the effects of C-stocking practices at different spatial scales, is critical for both stakeholders and public authorities. Various studies have recently demonstrated the usefulness of hyperspectral satellite images for faster and cheaper prediction of SOC compared to standard soil chemical analyses. In this study we combine both sources of information by means of hierarchical modelling accounting for different spatial scales to provide an accurate map of SOC over a broad regional scale in a peri-urban region.