Abstract : Knowledge of the regolith, i.e. surface geology, is increasingly demanded to serve societal needs. As a response to this demand, the reference information published worldwide in soil and regolith maps is the lithology. However, acquisition of this information in the field (and at the laboratory) is expensive and time consuming. Natural gamma-ray signals are influenced by lithological as well as physico-chemical properties of the first meter of the ground. In order to accelerate the mapping process, we investigate a predictive lithology mapping method at a regional to semi-local scale. In a siliciclastic-carbonated sedimentary environment, we combine airborne gamma-ray data with morphological information, as well as shallow lithology in boreholes. The proposed method allows 1/ " calibrating " airborne gamma-ray data in terms of dominant lithology, 2/ deriving geologically realistic cartographic polygons based on morphology and gamma-ray maps. The dominant lithology is attributed to each cartographic polygon resulting in a realistic predictive lithological map. A quick validation of this map is presented in comparison with an independent soil map and several other tests of the method in other parts of the Paris Basin, tend to accredit the robustness of the method.
