We recently spent three and a half weeks in the Soviet Union as part of the Soviet/American Working Group on earthquake prediction. We visited the seismological field station at Garm as well as several institutes in Moscow, Tashkent, and Dushanbe. We examined and discussed at some length some of the ongoing experimental research pertaining to earthquake mechanics and prediction. In the following note we mention the laboratories and individuals currently engaged in experimental work, for the benefit of our American colleagues who may wish to communicate with their Soviet counterparts.
Abstract On 15 June 2010, a M w 5.7 earthquake occurred near Ocotillo, California, in the Yuha Desert. This event was the largest aftershock of the 4 April 2010 M w 7.2 El Mayor‐Cucapah (EMC) earthquake in this region. The EMC mainshock and subsequent Ocotillo aftershock provide an opportunity to test the Coulomb failure hypothesis (CFS). We explore the spatiotemporal correlation between seismicity rate changes and regions of positive and negative CFS change imparted by the Ocotillo event. Based on simple CFS calculations we divide the Yuha Desert into three subregions, one triggering zone and two stress shadow zones. We find the nominal triggering zone displays immediate triggering, one stress shadowed region experiences immediate quiescence, and the other nominal stress shadow undergoes an immediate rate increase followed by a delayed shutdown. We quantitatively model the spatiotemporal variation of earthquake rates by combining calculations of CFS change with the rate‐state earthquake rate formulation of Dieterich (1994), assuming that each subregion contains a mixture of nucleation sources that experienced a CFS change of differing signs. Our modeling reproduces the observations, including the observed delay in the stress shadow effect in the third region following the Ocotillo aftershock. The delayed shadow effect occurs because of intrinsic differences in the amplitude of the rate response to positive and negative stress changes and the time constants for return to background rates for the two populations. We find that rate‐state models of time‐dependent earthquake rates are in good agreement with the observed rates and thus explain the complex spatiotemporal patterns of seismicity.
Similar precursory phenomena have been observed before earthquakes in the United States, the Soviet Union, Japan, and China. Two quite different physical models are used to explain these phenomena. According to a model developed by US seismologists, the so-called dilatancy diffusion model, the earthquake occurs near maximum stress, following a period of dilatant crack expansion. Diffusion of water in and out of the dilatant volume is required to explain the recovery of seismic velocity before the earthquake. According to a model developed by Soviet scientists growth of cracks is also involved but diffusion of water in and out of the focal region is not required. With this model, the earthquake is assumed to occur during a period of falling stress and recovery of velocity here is due to crack closure as stress relaxes. In general, the dilatancy diffusion model gives a peaked precursor form, whereas the dry model gives a bay form, in which recovery is well under way before the earthquake. A number of field observations should help to distinguish between the two models: study of post-earthquake recovery, time variation of stress and pore pressure in the focal region, the occurrence of pre-existing faults, and any changes in direction of precursory phenomena during the anomalous period.
Abstract Seismicity induced by fluid injection and withdrawal has emerged as a central element of the scientific discussion around subsurface technologies that tap into water and energy resources. Here we present the application of coupled flow‐geomechanics simulation technology to the post mortem analysis of a sequence of damaging earthquakes ( M w = 6.0 and 5.8) in May 2012 near the Cavone oil field, in northern Italy. This sequence raised the question of whether these earthquakes might have been triggered by activities due to oil and gas production. Our analysis strongly suggests that the combined effects of fluid production and injection from the Cavone field were not a driver for the observed seismicity. More generally, our study illustrates that computational modeling of coupled flow and geomechanics permits the integration of geologic, seismotectonic, well log, fluid pressure and flow rate, and geodetic data and provides a promising approach for assessing and managing hazards associated with induced seismicity.
An experimental method designed to measure ultrasonic velocities in simulated fault gouge subjected to normal and direct shearing stress yields these results. For a well‐compacted, unsorted, fine‐grained, dry granite gouge under constant normal stress, reversible changes in V p with shear stress are observed prior to stable sliding. V p decreases by as much as 9% with increasing shear stress for the conditions of these experiments. For layers of intact granite slabs subjected to constant normal stress, and whose interfaces are relatively free of gouge material, V p increases with increasing shear stress. The observed velocity changes in the granite gouge experiments are believed to be produced by reversible opening and closing of small cracks. For a gouge material consisting of moderately compacted, sorted, coarser‐grained, dry sand, irreversible changes in V p were observed. The predominant changes in V p for the sand gouge are believed to result from grain fracturing and from readjustment of grains to a denser state of aggregation with changes in the general stress.
Abstract This paper presents a phenomenological stochastic model for earthquake recurrence processes involving physical interaction among fault segments. Slip on one segment may reduce (or increase) the time to the next event on another segment or possibly induce an immediate slip on that segment as well. The gross behavior of this model is first observed through simulations; temporal and spatial disorder are observed even when the stochastic aspects are minimized. To estimate the strength of these interactions, we derive factors from the output of three-dimensional elastic dislocation analyses, relating induced stress changes to temporal changes in next-event dates. In a final section, we derive approximate analytical expressions and numerical results for future probabilistic earthquake risk and site hazard, conditional on the elapsed times since events on all relevant fault segments and on the number of events since that may have caused stress changes (interactions).
Abstract. Modelling the seismic potential of active faults and the associated epistemic uncertainty is a fundamental step of probabilistic seismic hazard assessment (PSHA). We use SHERIFS (Seismic Hazard and Earthquake Rate In Fault Systems), an open-source code allowing to build hazard models including earthquake ruptures involving several faults, to model the seismicity rates on the North Anatolian Fault (NAF) system in the Marmara region. Through an iterative approach, SHERIFS converts the slip-rate on the faults into earthquake rates that follow a Magnitude Frequency Distribution (MFD) defined at the fault system level, allowing to model complex multi-fault ruptures and off-fault seismicity while exploring the underlying epistemic uncertainties. In a logic tree, we explore uncertainties concerning the locking state of the NAF in the Marmara Sea, the maximum possible rupture in the system, the shape of the MFD and the ratio of off-fault seismicity. The branches of the logic tree are weighted according to the match between the modelled earthquake rate and the earthquake rates calculated from the local data, earthquake catalogue and paleoseismicity. In addition, we use the result of the physics-based earthquake simulator RSQSim to inform the logic tree and increase the weight on the hypotheses that are compatible with the result of the simulator. Using both the local data and the simulator to weight the logic tree branches, we are able to reduce the uncertainties affecting the earthquake rates in the Marmara region. The weighted logic tree of models built in this study is used in a companion article to calculate the probability of collapse of a building in Istanbul.
The concept of apparent fracture energy for the shear failure process is employed by many authors in modeling earthquake sources as dynamically extending shear cracks. Using records of shear strain and relative displacement from stick‐slip events generated along a simulated, prepared fault surface in a large (1.5m × 1.5m × 0.4m) granite block and a slip‐weakening model for the fault, direct estimates of the apparent shear fracture energy of the stick‐slip events have been obtained. For events generated on a finely ground fault surface, apparent fracture energy ranges from 0.06 J/m² at a normal stress of 1.1 MPa to 0.8 J/m² at a normal stress of 4.6 MPa. In contrast to estimates for tensile crack formation, we find that the apparent fracture energy of stick‐slip events increases linearly with normal stress. The results for the slip‐weakening model for the stick‐slip events are generally consistent with constitutive fault models suggested by observations of stable sliding in smaller scale experiments.
