Periodic Reporting for period 1 - FEAT (The role of Fluid pressure in EArthquake Triggering)
Periodo di rendicontazione: 2015-12-01 al 2017-11-30
In recent years, human induced seismicity associated with underground wastewater disposal and fluid injection has become a matter of societal concern. During oil/gas extraction and hydraulic fracturing large amounts of contaminated water (i.e. wastewater) are produced and generally re-injected at high pressure within subsurface permeable formations. These operations modify the subsurface stress field causing seismicity rates to increase dramatically in regions far from active tectonic margins, and stable continental regions like the Western Canada Sedimentary basin and the central United States have seen sharp increases of moderate to large earthquakes, with Mw > 5 events becoming common. In Europe, induced earthquakes during fluid pressure stimulation of subsurface reservoirs have been documented in several notable cases including Switzerland, southern Italy and the Netherlands. In this context, understand how fluid pressure interacts with faults is of primary importance to produce comprehensive models to evaluate the seismic hazard related to injection of fluids at depth. However, the role of fluid pressure in fault stability represent a conundrum in earthquake physics. Fluid overpressure has been proposed as one of the primary mechanisms that facilitate earthquake slip along tectonic faults. However, elastic dislocation theory combined with Rate- and State Friction (RSF) laws suggests that fluid overpressure may inhibit the dynamic instabilities that result in earthquakes. This controversy poses a serious problem in our understanding of earthquake physics, with severe implications for both natural and human-induced seismic hazard. The overall objective of FEAT is to produce a comprehensive experimental study on the role of fluid pressure on fault frictional stability to shed light on the physical processes at the origin of induced seismicity.
I designed state-of-the-art laboratory experiments to shed light on the effect of fluid pressure on fault rheology and frictional stability. I performed experiments using a word class apparatus (BRAVA) at the HP-HT laboratory installed at INGV in Rome, with fluid pressure ranging from sub hydrostatic to near lithostatic, and showed that the friction stability parameter (a-b) evolves with increasing fluid pressure from velocity strengthening (indicative of aseismic creep) to velocity neutral behaviour. Furthermore, the critical slip distance, Dc, dramatically decreases as the pore fluid pressure increases. These observations indicate that fluid overpressure can facilitate earthquake nucleation since it controls the evolution of fault zone rheology (Scuderi and Collettini, 2016). However, while this study was the first to show the effect of fluid pressure on the friction stability parameters, these results did not show clear evidence of velocity weakening behavior, which is required to nucleate a seismic instability. To dig further into the physical mechanism at the origin of fluid driven fault slip, I have since developed a new experimental approach that consists in deforming simulated fault gouge under constant applied stress field (i.e. creep experiments), which is inferred for intraplate faults, and monitored the evolution of fault slip during fluid pressure build up. Remarkably, and for the first time, my results showed that, when the Coulomb-Mohr criterion for fault reactivation is satisfied, a frictional instability can spontaneously nucleate on a velocity strengthening fault gouge due to fluid pressurization (Scuderi et al., 2017). I proposed that under the stress conditions of induced seismicity the second order frictional variations, as evaluated by the RSF approach, are overcome by the weakening induced by the fluid pressure. Based on mechanical data and microstructural analysis, I proposed a microphysical mechanism that relates fault zone deformation with the evolution of the stress state through an energy balance to explain the observed nucleation of slip instability.
Another aspect that was developed within FEAT was to understand the physical processes during fault zone deformation using non-destructive techniques, such as ultrasonic wave propagation, which may prove effective in the future to implement systems of early warning. By matching the fault rheological properties with the elastic surrounding I was able to reproduce, for the first time, the spectrum of fault slip behaviors in the laboratory and investigate the underlying physical mechanisms. By analyzing ultrasonic wave propagation during stick-slip frictional sliding, and developing a new technique that uses coda waves, I have shown that P-wave velocity changes before the earthquake stress drop can be clearly detected within laboratory fault zones. The results, that have been published in Nature Geoscience and Nature Communication, have built the ground for a better understanding of the physical mechanisms at the origin of fault slip. For each experiment, we have collected the resulting fault zone and analyzed to characterize the evolution of shear fabric in granular fault gouge from stable sliding to stick slip and understand the implications for fault slip mode. This is a long-standing problem in fault mechanics with implication in identifying the record of past co-seismic slip and evaluate the seismic hazard. I have shown that once shear localizes along narrow slip zones the slip behavior is controlled by the elastic interaction between the fault zone and the surrounding. This suggest that a single fault segment can experience a spectrum of slip behaviors, from aseismic creep, slow earthquakes to dynamic rupture, depending on the evolution of fault rock frictional properties and elastic conditions of the loading system.
Another aspect that was developed within FEAT was to understand the physical processes during fault zone deformation using non-destructive techniques, such as ultrasonic wave propagation, which may prove effective in the future to implement systems of early warning. By matching the fault rheological properties with the elastic surrounding I was able to reproduce, for the first time, the spectrum of fault slip behaviors in the laboratory and investigate the underlying physical mechanisms. By analyzing ultrasonic wave propagation during stick-slip frictional sliding, and developing a new technique that uses coda waves, I have shown that P-wave velocity changes before the earthquake stress drop can be clearly detected within laboratory fault zones. The results, that have been published in Nature Geoscience and Nature Communication, have built the ground for a better understanding of the physical mechanisms at the origin of fault slip. For each experiment, we have collected the resulting fault zone and analyzed to characterize the evolution of shear fabric in granular fault gouge from stable sliding to stick slip and understand the implications for fault slip mode. This is a long-standing problem in fault mechanics with implication in identifying the record of past co-seismic slip and evaluate the seismic hazard. I have shown that once shear localizes along narrow slip zones the slip behavior is controlled by the elastic interaction between the fault zone and the surrounding. This suggest that a single fault segment can experience a spectrum of slip behaviors, from aseismic creep, slow earthquakes to dynamic rupture, depending on the evolution of fault rock frictional properties and elastic conditions of the loading system.
The results collected during FEAT represent the first comprehensive laboratory investigation of the role of fluid pressure on fault stability with direct implication for the evaluation of the seismic risk during fluid injection operations. Form the results that I have produced with FEAT emerge that an increase in fluid pressure beyond a critical stress state causes accelerated fault slip that evolves in dynamic slip inducing earthquakes regardless of the original frictional rheology of the fault. We propose that during injection the fluid pressure and volume of fluids should be carefully monitored and never exceed this critical value. However, the inaccessibility of faults at depth associated with the difficulty in imaging the fault zone structure and the complications in understanding the state of stress close to faults due to heterogeneous distribution of fluids and fluid pressure raise major challenges. In this context, more work should be done, from a theoretical and technological point of view, to better understand what are the limit conditions that should not be overcome and how to monitor the true state of stress at depth to reduce the risk of induced seismicity.