Analysing the proximal gamma radiometry in contrasting Mediterranean landscapes: Towards a regional prediction of clay content

Abstract The analysis of γ-radiometric data from proximal soil sensing measurements can be used to predict key soil properties for environmental and agricultural issues. These predictions require prior calibrations that use a large sampling of field data that could be significantly reduced if regional predictions can be established. However, such predictions may be challenging for Mediterranean regions because of the highly contrasted soils and specific geopedological characteristics (sedimentary rocks, low weathering, etc.). Our study focuses on the regional prediction of a key soil property (clay content), within highly contrasting geopedological groups (GPG) from the Languedoc region. The aims of the research were (i) to understand the relationship between γ-radiometric data and the clay content within the different complex GPGs and (ii) to predict the clay content from γ radiometry with a regional calibration based on the GPG as an auxiliary variable . A proximal soil survey using the γ radiometer, called “The Mole”, was conducted in 8 areas of the Languedoc region. The plots were selected over different GPGs that are representative of the soil diversity and materials according to their weathering intensity and igneous components. In total, 242 sites were sampled based on a random grid cell for soil laboratory measurements. The soil texture and stoniness were analysed, and the particle size fractions were separated for mineral identification via X-ray diffraction. The activity of 232 Th [1 to 71 bq/kg] and 40 K [40 to 771 bq/kg] exhibits the largest range between the different GPGs, which spans the range of values measured in previous European studies. The relationship between the clay content and 232 Th is significant at a regional level [R 2  = 0.60; RMSE = 42 g/kg]. The residuals are significantly distributed according to the different GPGs. A new linear model based on the 232 Th activity and the GPGs improves the prediction of the clay content [R 2  = 0.72; RMSE = 35 g/kg]. Our study showed that ground γ radiometry can be used to predict the clay content within a complex Mediterranean sedimentary context. The 232 Th radionuclide is the best predictor. The disturbance of the relationship between the 232 Th and the clay content corresponds to the fine earth mineralogy type and the proportion of igneous pebbles.