Stage/Thèse en modélisation statistique des sols : système d’inférence spatiale des sols avec incertitude quantifiée

In the framework of the EU-project DeepHorizon (https://cordis.europa.eu/project/id/101156701), we are looking for an excellent PhD candidate to develop statistical methods supporting the development of a spatial soil inference system for European soils. A soil inference system uses known measurements, each with a certain level of uncertainty, to predict related soil properties with minimal error, by applying a series of logically connected (pedo)transfer functions (PTFs). The PhD candidate will start with an inventory of existing soil pedotransfer functions relevant to European soils and to calibrate usual mechanistic biogeochemical models. A large part of the work involves the exploration, development and application of new statistical approaches relevant for the inference system. The approaches should handle missing data along with uncertainty quantification of the input soil properties and propagation of the uncertainty throughout the inference engine. The candidate is expected to collaborate closely with other PhD candidates of the project consortium and with a project partner in Belgium, for which temporary stay could be envisioned.