Heather M. Savage

Potentially induced earthquakes in Oklahoma, USA: Links between wastewater injection and the 2011 Mw 5.7 earthquake sequence

Significant earthquakes are increasingly occurring within the continental interior of the United States, including five of moment magnitude (Mw) ≥ 5.0 in 2011 alone. Concurrently, the volume of fluid injected into the subsurface related to the production of unconventional resources continues to rise. Here we identify the largest earthquake potentially related to injection, an Mw 5.7 earthquake in November 2011 in Oklahoma. The earthquake was felt in at least 17 states and caused damage in the epicentral region. It occurred in a sequence, with 2 earthquakes of Mw 5.0 and a prolific sequence of aftershocks. We use the aftershocks to illuminate the faults that ruptured in the sequence, and show that the tip of the initial rupture plane is within ∼200 m of active injection wells and within ∼1 km of the surface; 30% of early aftershocks occur within the sedimentary section. Subsurface data indicate that fluid was injected into effectively sealed compartments, and we interpret that a net fluid volume increase after 18 yr of injection lowered effective stress on reservoir-bounding faults. Significantly, this case indicates that decades-long lags between the commencement of fluid injection and the onset of induced earthquakes are possible, and modifies our common criteria for fluid-induced events. The progressive rupture of three fault planes in this sequence suggests that stress changes from the initial rupture triggered the successive earthquakes, including one larger than the first.

Enhanced remote earthquake triggering at fluid-injection sites in the midwestern United States

Movers and Shakers We tend to view earthquakes as unpredictable phenomena caused by naturally shifting stresses in Earths crust. In reality, however, a range of human activity can also induce earthquakes. Ellsworth (p. 10.1126/science.1225942) reviews the current understanding of the causes and mechanics of earthquakes caused by human activity and the means to decrease their associated risk. Notable examples include injection of wastewater into deep formations and emerging technologies related to oil and gas recovery, including hydraulic fracturing. In addition to directly causing increased local seismic activity, activities such as deep fluid injection may have other ramifications related to earthquake occurrence. Van der Elst et al. (p. 164; see the news story by Kerr) demonstrate that in the midwestern United States, some areas with increased human-induced seismicity are also more prone to further earthquakes triggered by the seismic waves from large, remote earthquakes. Improved seismic monitoring and injection data near deep disposal sites will help to identify regions prone to remote triggering and, more broadly, suggest times when activities should, at least temporarily, be put on hold. Wastewater injected deep underground can make some faults more susceptible to triggering by large remote earthquakes. A recent dramatic increase in seismicity in the midwestern United States may be related to increases in deep wastewater injection. Here, we demonstrate that areas with suspected anthropogenic earthquakes are also more susceptible to earthquake-triggering from natural transient stresses generated by the seismic waves of large remote earthquakes. Enhanced triggering susceptibility suggests the presence of critically loaded faults and potentially high fluid pressures. Sensitivity to remote triggering is most clearly seen in sites with a long delay between the start of injection and the onset of seismicity and in regions that went on to host moderate magnitude earthquakes within 6 to 20 months. Triggering in induced seismic zones could therefore be an indicator that fluid injection has brought the fault system to a critical state.

Effects of acoustic waves on stick–slip in granular media and implications for earthquakes

It remains unknown how the small strains induced by seismic waves can trigger earthquakes at large distances, in some cases thousands of kilometres from the triggering earthquake, with failure often occurring long after the waves have passed. Earthquake nucleation is usually observed to take place at depths of 10–20 km, and so static overburden should be large enough to inhibit triggering by seismic-wave stress perturbations. To understand the physics of dynamic triggering better, as well as the influence of dynamic stressing on earthquake recurrence, we have conducted laboratory studies of stick–slip in granular media with and without applied acoustic vibration. Glass beads were used to simulate granular fault zone material, sheared under constant normal stress, and subject to transient or continuous perturbation by acoustic waves. Here we show that small-magnitude failure events, corresponding to triggered aftershocks, occur when applied sound-wave amplitudes exceed several microstrain. These events are frequently delayed or occur as part of a cascade of small events. Vibrations also cause large slip events to be disrupted in time relative to those without wave perturbation. The effects are observed for many large-event cycles after vibrations cease, indicating a strain memory in the granular material. Dynamic stressing of tectonic faults may play a similar role in determining the complexity of earthquake recurrence.

Biomarkers heat up during earthquakes: New evidence of seismic slip in the rock record

During earthquakes, faults heat up due to frictional work. However, evidence of heating from paleoearthquakes along exhumed faults remains scarce. Here we describe a method using thermal maturation of organic molecules in sedimentary rock to determine whether a fault has experienced differential heating compared to surrounding rocks. We demonstrate the utility of this method on an ancient, pseudotachylyte-hosting megathrust at Pasagshak Point, Alaska. Measurements of the ratio of thermally stable to thermally unstable compounds (diamondoids/n-alkanes) show that the melt-bearing rocks have higher thermal maturity than surrounding rocks. Furthermore, the mineralogy of the survivor grains and the presence of any organic molecules allow us to constrain the temperature rise during the ancient earthquakes to 840–1170 °C above ambient temperatures of ~260 °C. From this temperature rise, we estimate that the frictional work of the earthquake was ~105–228 MJ/m 2 . Using experimental friction measurements as a constraint, we estimate that the minimum slip necessary for heating was ~1–8 m. This paper demonstrates that biomarkers will be a useful tool to identify seismic slip along faults without frictional melt.

Critical evaluation of state evolution laws in rate and state friction: Fitting large velocity steps in simulated fault gouge with time‐, slip‐, and stress‐dependent constitutive laws

The variations in the response of different state evolution laws to large velocity increases can dramatically alter the style of earthquake nucleation in numerical simulations. But most velocity step friction experiments do not drive the sliding surface far enough above steady state to probe this relevant portion of the parameter space. We try to address this by fitting 1–3 orders of magnitude velocity step data on simulated gouge using the most widely used state evolution laws. We consider the Dieterich (Aging) and Ruina (Slip) formulations along with a stress dependent state evolution law recently proposed by Nagata et al. (2012). Our inversions confirm the results from smaller velocity step tests that the Aging law cannot explain the observed response and that the Slip law produces much better fits to the data. The stress dependent Nagata law can produce fits identical to, and sometimes slightly better than, those produced by the Slip law using a sufficiently large value of an additional free parameter c that controls the stress dependence of state evolution. A Monte Carlo search of the parameter space confirms analytical results that velocity step data that are well represented by the Slip law can only impose a lower bound on acceptable values of c, and that this lower bound increases with the size of the velocity step being fit. We find that our 1–3 orders of magnitude velocity steps on synthetic gouge impose this lower bound on c to be 10–100, significantly larger than the value of 2 obtained by Nagata et al. (2012) based on experiments on initially bare rock surfaces with generally smaller departures from steady state.

Can flat-ramp-flat fault geometry be inferred from fold shape?: A comparison of kinematic and mechanical folds

The inference of fault geometry from suprajacent fold shape relies on consistent and verified forward models of fault-cored folds, e.g. suites of models with differing fault boundary conditions demonstrate the range of possible folding. Results of kinematic (fault-parallel flow) and mechanical (boundary element method) models are compared to ascertain differences in the way the two methods simulate flexure associated with slip along flat-ramp-flat geometry. These differences are assessed by systematically altering fault parameters in each model and observing subsequent changes in the suprajacent fold shapes. Differences between the kinematic and mechanical fault-fold relationships highlight the differences between the methods. Additionally, a laboratory fold is simulated to determine which method might best predict fault parameters from fold shape. Although kinematic folds do not fully capture the three-dimensional nature of geologic folds, mechanical models have non-unique fold-fault relationships. Predicting fault geometry from fold shape is best accomplished by a combination of the two methods.

Frequency dependence of delayed and instantaneous triggering on laboratory and simulated faults governed by rate‐state friction

Earthquake triggering by transient stresses is commonly observed; however, some aspects remain unexplained. The first is the often-observed delay between the triggered earthquakes and the triggering waves, and the second is the unexpected effectiveness of transient stressing in the seismic frequency band. Previous theoretical and laboratory studies have suggested that seismic transients should have little impact on faults if the duration of the transient is smaller than the timescale for nucleation of slip. We reexamine the dynamics of stress triggering during stick-slip sliding on a laboratory fault and make three important observations that pertain to earthquake triggering. (1) Delayed triggering (clock advance) occurs for both bare granite surfaces and granular gouge prior to the onset of instantaneous triggering. (2) Triggering occurs much earlier in the stick-slip cycle than expected for a simple Coulomb stress threshold. (3) Shorter-period (higher stressing rate) pulses are more effective at triggering than longer-period pulses of the same stress amplitude. We use numerical simulations to show that rate-state friction can explain each of the observed features but not all three simultaneously. Only the Ruina slip law for state evolution, in which faults must slip to heal, can reproduce early-onset and stressing rate-dependent triggering. The laboratory and numerical experiments show that faults can remain relatively weak over much of the seismic cycle and that the triggered response depends on a competition between healing and weakening during triggered slip. Transient stressing at seismic frequencies may be more effective at triggering earthquakes than previously recognized.

Sediment provenance and controls on slip propagation: Lessons learned from the 2011 Tohoku and other great earthquakes of the subducting northwest Pacific plate

The ∼50 m slip of the Tohoku earthquake occurred along a very fine grained red-brown smectitic clay horizon subducting in the Japan Trench. This clay, cored in the plate boundary fault at Integrated Ocean Drilling Program Expedition 345, Site C0019, correlates with similar pelagic clay recovered seaward of the trench at Deep Sea Drilling Project Sites 436 and 1149. Comparable clays occur throughout the northwest Pacific Basin. Backtracking of ocean drilling Sites 436, C0019, and 1149 indicates that they formed during the Early Cretaceous at the Kula-Pacific Ridge. These sites traveled northwestward through the equatorial zone, accumulating siliceous and calcareous oozes until ca. 100–85 Ma. Sites 436, C0019, and 1149 then entered the realm of pelagic clay deposition where they remained until ca. 15 Ma. From ca. 15 Ma to the present, Sites 436, C0019, and 1149 accumulated clays and silty clays with variable amounts of siliceous microfossils and volcanic ash, representing the transition from deep-sea conditions to a continental margin sedimentary environment. The predicted backtracked vertical sequence of sediments fits well with the cores at Sites 436, 1149, and C0019, after accounting for structural complications in the latter. Pelagic clay occurs in numerous boreholes penetrating the relatively smooth ocean floor of the Pacific plate north and northeast of the Tohoku earthquake. Here the widespread pelagic clay apparently fosters tsunami and tsunamigenic earthquakes. Seamounts rising above the normal oceanic crust accumulated sequences of calcareous sediments as their crests remained above the calcite compensation depth for most of their history. A seafloor including pelagic clay and carbonate-covered seamounts occurs south and southeast of the southern extent of the Tohoku earthquake rupture zone. This area has no historic tsunami or tsunamigenic earthquakes along the Japan and Izu-Bonin Trenches with the possible exception of the poorly located Enpo earthquake of A.D. 1677. We believe that the seamounts incoming on the oceanic plate to the south and southeast of the Tohoku rupture zone interfere with long-distance propagation of slip in the pelagic clay, limiting earthquake magnitude, shallow slip, and tsunami generation.

Possible precursory signals in damage zone foreshocks

Foreshocks may provide a precursory signal of an impending earthquake, but their role in nucleation of the mainshock is unclear. One way to further our understanding of foreshock failure mechanisms is to determine where they occur in the fault zone. However, earthquake locations commonly include uncertainties large enough to allow rupture on either the main fault interface or on subsidiary fractures within a surrounding damage zone. Here we obtain precise earthquake locations, with ~10 m uncertainty, for foreshocks and aftershocks of an Mw5.0 near Prague, OK, USA. Repeating earthquakes imply that some precursory slow slip occurred before the mainshock. In addition, we show that foreshocks initially rupture faults and fractures throughout the 300-meter-thick fault damage zone, and later localize onto a narrower zone (<100 m thick) nearer the mainshock hypocenter. Focal mechanisms corroborate that foreshocks occur in the surrounding damage zone as well as on the mainshock rupture interface. These results highlight that earthquake nucleation is most likely a complex feedback between frictional failure processes on the fault interface and deformation in the surrounding damaged rock, not just nucleation on a single surface.

Dynamically triggered slip leading to sustained fault gouge weakening under laboratory shear conditions

We investigate dynamic wave-triggered slip under laboratory shear conditions. The experiment is composed of a three-block system containing two gouge layers composed of glass beads and held in place by a fixed load in a biaxial configuration. When the system is sheared under steady state conditions at a normal load of 4 MPa, we find that shear failure may be instantaneously triggered by a dynamic wave, corresponding to material weakening and softening if the system is in a critical shear stress state (near failure). Following triggering, the gouge material remains in a perturbed state over multiple slip cycles as evidenced by the recovery of the material strength, shear modulus, and slip recurrence time. This work suggests that faults must be critically stressed to trigger under dynamic conditions and that the recovery process following a dynamically triggered event differs from the recovery following a spontaneous event.