SSM perspective
Background and objective
It is a well-known phenomenon that artificial increased in pore pressure at depth, due to subsurface fluid injection, can trigger earthquakes (Walsh and Zoback 2015, and references therein). Increased pore pressures at depth can also occur naturally, for instance due to the pressure changes imposed by a waxing and waning ice sheet.
A geological repository for spent fuel is to be safe for timescales involving several future glacial cycles. The advance or retreat of an ice sheet increases stress field anisotropy, leading to greater fault instability and the potential for glacially induced earthquakes. Specifically, during the deglaciation of a warm-based ice sheet, high hydraulic gradients can elevate pore fluid pressure to the point where fault stability is compromised. Thus, regarding a geological repository for spent fuel in Forsmark, it's important to examine the potential for fault instability by fluid pressure increase associated with future glacial cycles.
Empirical data from the fluid injection-induced fault slip experiment at Mont Terri provides an opportunity to evaluate numerical model’s ability to reproduce observed results. This process can build confidence in the methodology used to assess potential future fault slip caused by elevated pore fluid pressure.
Results and conclusions
Both the model and the experiment demonstrated a sharp increase in pressure at the monitoring point on the Main fault around 800 seconds, marking the onset of fluid migration from the Injection fault and activation of the Main fault. However, the experiment results showed a more abrupt and faster dissipation of pressure, while the model exhibited a more gradual decline, indicating slower fluid migration and pressure dissipation in the fault model. This suggests that the model could be refined to better capture the fluid retention and faster dynamics observed in natural fault systems, especially the physical involved in the post-activation.
The shear displacement curves at the monitoring point in the Main fault show a similar pattern of rapid increase during the fluid migration event, but the model results stabilize quickly after the peak displacement, while the experiment shows a gradual decline in shear displacement. The experiment highlights more sustained mechanical activity in the Main fault after fluid injection, indicating that in-situ fault systems may continue to experience post-activation slip even after the peak pressure has dissipated. The model captures the initial slip well but under predicts the ongoing post-activation slip mechanical processes.
The experimental results demonstrate a clear pressure-controlled slip behaviour at the injection point, with shear displacement closely tracking the stepwise injection pressure and exhibiting largely reversible elastic–frictional response. In contrast, the numerical model does not reproduce this fundamental pressure–slip coupling, instead predicting smoother, largely irreversible displacement that is weakly linked to the injection protocol and inconsistent with the expected slip direction. This mismatch indicates that the current modelling approach does not adequately represent pressure-driven fault mechanics at the Injection fault and requires refinement to capture the dominant controlling role of fluid pressure and fault heterogeneity.
Need for further research
The findings of this study suggest that while the numerical model is capable of fault dynamics such as fluid-induced fault activation, refinements are needed to improve the simulation of displacement evolution at the Injection fault. Specifically, incorporating more detailed fault characteristics, such as heterogeneity and variability in material properties, could lead to a better match with experimental data.