LEA ZUCCALI

This PhD thesis is proposed in response to the increasing and regular availability of remote sensing data and the requirement for operational prediction of natural hazards. The main objective is to improve the near real-time integration of remote sensing data and dynamic geophysical models for natural hazard mitigation. This thesis is partially funded (50%) by the National Action Plan on Artificial Intelligence. The relevance of the methodology developed in this thesis 1) compared to the current one lies in the integration of geophysical knowledge (which allows to improve the interpretation of the results for operational purposes) and its implementation in quasi real time; 2) compared to previous attempts to improve the near real-time interpretation of InSAR data using the Kalman filter (e.g., Bato et al, 2017, Dalaison & Jolivet, 2020), lies in the ability to take into account non-Gaussian error statistics (which allow a better representation of reality). The first application will be volcanology, using InSAR and GNSS data, but the methodology can easily be applied to other natural hazards (e.g., landslides, slow slides...), as well as anthropogenic hazards such as forest fires.

This Ph.D thesis is proposed along with the increasing and regular availability of the amount of remote sensing data and the response to the requirement of operational prediction of natural hazards. The main objective is to improve the near-real time integration of remote sensing data and dynamical geophysical models for the mitigation of natural hazards. This thesis is partly (50%) funded by the national action plan in Artificial Intelligence. The relevance of the methodology developed in this thesis 1) compared to the current emerging data-driven methods, lies in the incorporation of geophysical knowledge (which helps increasing the interpretability of the results for operational purposes) and its near real-time implementation; 2) compared to previous attempts to improve the near real-time interpretation of InSAR data base on the Kalman Filter (e.g. Bato et al, 2017, Dalaison & Jolivet, 2020), lies in the capability in taking non-Gaussian error statistics (which can better represent the reality) into account. First application will be in volcanology, using InSAR andGNSS data, but the methodology can easily be applied to other natural hazards (e.g. landslides, slow slip..), as well as to anthropogenic hazards like forest fires.

__________________________________________________________________________________________________________

Encadrantes / Supervisors: Yajing Yan (LISTIC) and Virginie Pinel (ISTerre)