The predictability of earthquakes has been debated since many decades, particularly their long-term recurrence (when, where and which magnitude). The nonlinearlity of the governing physics behind the process, the complexity of the Earth's interior and the uncertainty of the observable parameters are all factors which make earthquake prediction almost impossible. However, we also know some systematic behaviours of earthquake occurrence both from field data analysis and laboratory experiments. 1) bring the synergy to understand better physically and statistically the variation of seismicity and 2) tackle the forecast of seismicity evolution due to small stress perturbations. We bring the field and experimental data analysis to more precisely detect systematic or abnormal behaviours. We then put our efforts on numerical simulations (mathematical and fully-elastodynamics), considering the randomness of heterogeneity and its evolution and sensibility over various scales in both space and time, for discussing the predictability. For these purposes, we establish a multi-disciplinary between geoscience laboratories, a physical and a mathematical department, including international collaborations.
The reserach project PREMs (PRedictability of Earthquakes and Mathematical Models) has been proposed to Agence National de Recherche in the generic call 2024, in particular in the axis H16: Interfaces: Mathematics, compuitng science - Earth Science and Environmenet. The project scientifically starts on the 1st January 2025 for the period of 48 months. The project benefits the financal support of 661 k euros from ANR.