Active fault slip-rates and earthquake recurrence controlled by stress transfer and viscous flow

Seismic hazard assessment and understanding of continental deformation are hindered by unexplained slip-rate fluctuations on faults, associated with (a) temporal clusters of damaging earthquakes lasting 100s to 1000s of years, and (b) longer-term fault quiescence lasting tens to hundreds of millennia. We propose a new unified hypothesis explaining both (a) and (b), involving stress interactions between fault/shear-zones and neighbouring fault/shear-zones; however key data to test this are lacking. We propose measurements and modelling to test our hypothesis, which have the potential to quantify the processes that control continental faulting and fluctuations in the rates of expected earthquake occurrence, with high societal impact. Our aspiration is that cities and critical facilities worldwide will gain additional protection from seismic hazard through use of the calculations we pioneer herein.

The background is that slip-rate fluctuations hinder understanding because they introduce uncertainty about whether specific faults are active or not. For example, a review in Japan of earthquake risk to critical facilities, such as the Tsuruga nuclear power plant (NPP), revealed a geological fault under a nuclear reactor (Chapman et al. 2014). The question that arose was whether the fault was active or not. Japan's Nuclear Regulatory Authority (NRA) has guidelines defining fault activity, and considered the fault under the reactor to be active, evidenced by faulting in sediments <~125,000 years in age. The Japan Atomic Energy Power Company (JPAC) disagreed, following study by an independent team of geoscientists. In 2014, the Tsuruga NPP remained closed due to ongoing debate between the NRA and JPAC, with similar debates ongoing for other NPPs. We suggest that defining fault activity as simply "active" or "inactive" is unsatisfactory because it is debatable even amongst experts. In fact a fault that has not slipped in many millennia may, in reality, not be inactive, but instead may simply have a low slip-rate, with the capability to host a damaging earthquake after a long recurrence interval.

Our breakthrough is we think slip-rate fluctuations over both timescales (a and b) are a continuum, sharing a common cause involving interaction between fault/shear-zones. For the first time, we provide calculations that describe this interaction, quantifying slip-rate fluctuations and seismic hazard in terms of probabilities. We show that slip during an earthquake cluster on a brittle fault in the upper crust occurs in tandem with high strain-rate on the viscous shear-zone underlying the fault. This deformation of the crust produces changes in differential stress on neighbouring fault/shear-zones. Viscous strain-rate is known to be proportional to differential stress, so, given data on slip-rate fluctuations one can calculate changes in differential stress, and then calculate implied changes to viscous strain-rates on receiver shear zones and slip-rates on their overlying brittle faults. We provide a quantified example covering several millennia, but lack data allowing a test over tens to hundreds of millennia. If we can verify our hypothesis over both timescales, through successful replication of measurements via modelling, we will have identified and quantified a hitherto unknown fundamental geological process. We will study the Athens region, Greece, where a special set of geological attributes allows us to measure and model slip-rate fluctuation over both time scales (a and b), the key data combination never achieved to date. We know of no other quantified explanation that links slip-rate fluctuations over the two timescales; the significance and impact of accomplishing this is that it has the potential to change the way we mitigate hazard for cities and critical facilities.

Chapman et al. 2014, Active faults and nuclear power plants, EOS, 95, 4