Exploring Earthquake Processes Through Models and Experiments

Patrick Bianchi addresses one of the most complex challenges in geophysics: improving our ability of forecasting natural and induced earthquakes. In his thesis he successfully combined measurements retrieved at the laboratory scale within the Rock Physics and Mechanics Laboratory (RPMLab) (cm-scale) with numerical simulations to investigate the rock behaviour before an earthquake. Therewith, he aimed to improve the understanding of earthquake triggers and, eventually, set the basis for future upscale experiments at the BedrettoLab (100 m-scale).

In a first step, Patrick conducted experiments in the RPMLab, in which rock samples equipped with acoustic sensors and fiber optics where put under high pressure (2500 kN) to produce tiny earthquakes and, ultimately, to fracture the rock. Thanks to the implemented monitoring technologies, he was able to observe premonitory signals (localisation of seismic and aseismic deformation) shortly before the sample failure. With the available data, he validated numerical simulations performed with a physics-based computational tool that allowed to effectively depict the complexity linked to such preparatory processes.
Next, he focused on how different types of cracking affect stress levels, seismic statistics, and energy release in the rock. He found that certain types of cracks consume more energy, which could help explaining why some earthquakes are bigger than others. By performing additional simulations, he proved that small cracks may merge at the final stages of the experiments, creating larger fractures, which, eventually, lead to a macrofractures in the sample and its complete failure.
Finally, he investigated the effects of fault pre-conditioning in a sample of Rotondo granite retrieved from the BedrettoLab. Before the test, the rock sample was drilled, creating tiny boreholes serving as an artificial fault, which was then saturated with water. His tests showed that the injection of water with pre-conditioning caused the fault to reactivate in specific ways, creating a system of channels that allow the fluid to flow. This process can lead to small, local and sometimes silent movements in the rock that could eventually lead to an earthquake.
By combining lab experiments on different rocks and performing numerical simulations, his research helps to better understand how earthquakes nucleate. This knowledge could eventually lead to more accurate earthquake predictions, benefiting both geothermal energy projects and improving society's ability to prepare for earthquakes.