How can satellite observations help us understand and model the behaviour of active tectonic systems, from pre-earthquake deformation along the East Anatolian Fault Zone to recurring slow slip cycles in the Cascadia subduction zone?

Jan is a computational physicist working at the Faculty of Mathematics and Physics at Charles University. He studied general physics and geophysics at Charles University and completed his PhD in seismology in 2023. Subsequently, he worked as a postdoctoral researcher at the Géoazur Laboratory at the Université Côte d’Azur in France, where he explored the application of machine learning to the modelling of large earthquakes (2023-2025). His research focuses on the physics of earthquakes and slow slip, utilizing seismic and satellite observations and solving related large-scale inverse problems.

Research abstract and objectives

Seismic (earthquakes) or aseismic (creep, slow slip, tremors) slip releases stress accumulated at tectonic faults due to the movement of tectonic plates. These slip events occur at substantial depths of tens of kilometres, making it challenging to uncover their underlying physical mechanisms or to constrain their governing parameters through inverse modelling of geodetic (GNSS, InSAR) and seismic observations recorded at the Earth’s surface. The goal of the project is to develop dynamic rupture model inversions by utilizing modern machine learning approaches to enhance the computational efficiency of the method and pave the way for modelling longer-term fault movements. Two showcase applications will be performed: Bayesian inverse modelling of the interseismic slip in the decade preceding the catastrophic 2023 Kahramanmaraş, Türkiye, earthquake doublet, as well as cycles of slow slip events on the Cascadia subduction zone. This will enable to constrain the physical properties of both causative faults, explain their behaviour through a physics-based model, further informing the seismic hazard assessments. The main scientific results will be dynamic rupture models and spatial distributions of their parameters along the fault, which will explain the recorded InSAR and GNSS data and uncover possible differences in the physics-based mechanisms governing the tectonic slip in the two areas.

Project objectives:

• Develop physics-based models of longer-term development of tectonic faults to explain available geodetic and satellite (InSAR) data.
• Implement machine learning approach (Physics Informed Neural Nets) to constrain the model parameters.
• Investigate the interseismic slip on the East Anatolian fault and the cycles of slow slip events on the Cascadia Fault.

Scientific questions addressed by the project:

• What can we learn about the dynamic properties of the East Anatolian Fault Zone from satellite observations acquired prior to the 2023 earthquakes?
• Can the slow slip cycle in the Cascadia subduction zone be reproduced using a single dynamic model with stable frictional properties over multiple cycles?