The 23 August 2011 M w (moment magnitude) 5.7 ± 0.1, Mineral, Virginia, earthquake was the largest and most damaging in the central and eastern United States since the 1886 M w 6.8–7.0, Charleston, South Carolina, earthquake. Seismic data indicate that the earthquake rupture occurred on a southeast-dipping reverse fault and consisted of three subevents that progressed northeastward and updip. U.S. Geological Survey (USGS) “Did You Feel It?” intensity reports from across the eastern United States and southeastern Canada, rockfalls triggered at distances to 245 km, and regional groundwater-level changes are all consistent with efficient propagation of high-frequency seismic waves (~1 Hz and higher) in eastern North America due to low attenuation. Reported damage included cracked or shifted foundations and broken walls or chimneys, notably in unreinforced masonry, and indicated intensities up to VIII in the epicentral area based on USGS “Did You Feel It?” reports. The earthquake triggered the first automatic shutdown of a U.S. nuclear power plant, located ~23 km northeast of the main shock epicenter. Although shaking exceeded the plant’s design basis earthquake, the actual damage to safety-related structures, systems, and components was superficial. Damage to relatively tall masonry structures 130 km to the northeast in Washington, D.C., was consistent with source directivity, soft-soil ground-motion amplification, and anisotropic wave propagation with lower attenuation parallel to the northeast-trending Appalachian tectonic fabric. The earthquake and aftershocks occurred in crystalline rocks within Paleozoic thrust sheets of the Chopawamsic terrane. The main shock and majority of aftershocks delineated the newly named Quail fault zone in the subsurface, and shallow aftershocks defined outlying faults. The earthquake induced minor liquefaction sand boils, but notably there was no evidence of a surface fault rupture. Recurrence intervals, and evidence for larger earthquakes in the Quaternary in this area, remain important unknowns. This event, along with similar events during historical time, is a reminder that earthquakes of similar or larger magnitude pose a real hazard in eastern North America.

The Mineral, Virginia (USA), earthquake occurred at 17:51:3.9 UTC (Coordinated Universal Time) on 23 August 2011; the hypocenter was at 37.905°N, 77.975°W and depth was 8 km. The widely reported moment magnitude (M w ) was 5.7 ± 0.1. The m b (teleseismic short-period) magnitude estimated here using 26 global network stations is 5.77 ± 0.23, in agreement with m b values reported by national and international data centers, and not significantly different from M w . However, the m bLg magnitude of the earthquake determined here using 386 stations in eastern North America is 6.28 ± 0.26. The m bLg magnitude is a short-period magnitude correlated with m b that is based on the amplitude of the Lg phase at regional distances. Lg is a crust-guided phase that represents the largest amplitudes observed on short-period seismograms at regional distances in eastern North America. The m bLg magnitude was the primary magnitude appearing in catalogs of eastern United States earthquakes until superseded recently by M w . The catalog of previous earthquakes in central Virginia is keyed to m bLg , rather than M w . The Mineral shock reveals large regional variations in the Lg phase attenuation in the eastern United States. The expected value for the return period of m bLg 6.3 and larger earthquakes in the Central Virginia seismic zone is 752 yr, with a 95% confidence interval of 385–1471 yr. The Mineral earthquake caused Modified Mercalli Intensity (MMI) VIII damage in the epicentral area, with several instances of partial and total collapse of masonry chimneys and walls. A finite-fault, full wavefield simulation of the motions within 30 km of the epicenter fits the velocity recordings and Fourier spectral amplitudes in the 1–10 Hz frequency band, at the only strong-motion station in that distance range. The strongest motions are predicted to have occurred in two areas offset to the northwest and southeast of the epicenter, within which peak ground accelerations may have approached 2 g, and peak velocities were probably well in excess of 20 cm/s. The only factor mitigating damage in this earthquake was the brief (<3 s) duration of strong shaking.

We characterize shear-wave velocity versus depth (Vs profile) at 16 portable seismograph sites through the epicentral region of the 2011 M w 5.8 Mineral (Virginia, USA) earthquake to investigate ground-motion site effects in the area. We used a multimethod acquisition and analysis approach, where active-source horizontal shear (SH) wave reflection and refraction as well as active-source multichannel analysis of surface waves (MASW) and passive-source refraction microtremor (ReMi) Rayleigh wave dispersion were interpreted separately. The time-averaged shear-wave velocity to a depth of 30 m (Vs30), interpreted bedrock depth, and site resonant frequency were estimated from the best-fit Vs profile of each method at each location for analysis. Using the median Vs30 value (270–715 m/s) as representative of a given site, we estimate that all 16 sites are National Earthquake Hazards Reduction Program (NEHRP) site class C or D. Based on a comparison of simplified mapped surface geology to median Vs30 at our sites, we do not see clear evidence for using surface geologic units as a proxy for Vs30 in the epicentral region, although this may primarily be because the units are similar in age (Paleozoic) and may have similar bulk seismic properties. We compare resonant frequencies calculated from ambient noise horizontal:vertical spectral ratios (HVSR) at available sites to predicted site frequencies (generally between 1.9 and 7.6 Hz) derived from the median bedrock depth and average Vs to bedrock. Robust linear regression of HVSR to both site frequency and Vs30 demonstrate moderate correlation to each, and thus both appear to be generally representative of site response in this region. Based on Kendall tau rank correlation testing, we find that Vs30 and the site frequency calculated from average Vs to median interpreted bedrock depth can both be considered reliable predictors of weak-motion site effects in the epicentral region.

Earthquake damage is often increased due to local ground-motion amplification caused by soft soils, thick basin sediments, topographic effects, and liquefaction. A critical factor contributing to the assessment of seismic hazard is detailed information on local site response. In order to address and quantify the site response at seismograph stations in the eastern United States, we investigate the regional spatial variation of horizontal:vertical spectral ratios (HVSR) using ambient noise recorded at permanent regional and national network stations as well as temporary seismic stations deployed in order to record aftershocks of the 2011 Mineral, Virginia, earthquake. We compare the HVSR peak frequency to surface measurements of the shear-wave seismic velocity to 30 m depth (Vs30) at 21 seismograph stations in the eastern United States and find that HVSR peak frequency increases with increasing Vs30. We use this relationship to estimate the National Earthquake Hazards Reduction Program soil class at 218 ANSS (Advanced National Seismic System), GSN (Global Seismographic Network), and RSN (Regional Seismograph Networks) locations in the eastern United States, and suggest that this seismic station–based HVSR proxy could potentially be used to calibrate other site response characterization methods commonly used to estimate shaking hazard.

A dense seismic array was deployed at a 2 km spacing to record the aftershocks of the M w (moment magnitude) 5.8 Mineral, Virginia (USA), earthquake in 2011. The three-component seismometers, installed on a 60-km-long profile, recorded 40 aftershocks over 9 days of deployment. Based on manual picking of P-wave (primary, compressional) and S-wave (secondary, shear) arrival times of 15 aftershocks, we find that the P-wave propagates with a velocity of 6.15 km/s through the upper crust, and the direct S-wave travels with a velocity of 3.66 km/s within the first 20 km (Vs <20km ) and decreases slightly to 3.54 km/s (Vs >20km ) for distances >20 km. Hence, the aftershock data show a Vp/Vs ratio of 1.68 within the first 20 km of hypocentral distance, and a ratio of 1.73 for distances >20 km. We attribute the small decrease in Vs with increased distance to the complex geologic setting: the recording array was deployed across the geologic boundary between the Quantico Formation and the Ta River Metamorphic Suite. Near-source attenuation of S-waves (amplitude decay with hypocentral distance R) was measured using ~1200 digital seismograms (north-south and east-west components) from 40 aftershocks. The decay of amplitude was extracted using a nonlinear least-squares regression for different frequency bands: 1–2, 2–4, 4–8, and 8–16 Hz. For 1–2 Hz the decay can be described as a function of distance (R) as R −0.8 , for 2–4 Hz as R −0.9 , for 4–8 Hz as R −1.05 , and for 8–16 Hz as R −1.15 . The decay exponents, or b values, increase ~9%–15% from a lower to the next higher analyzed frequency band. These values are valid to a distance of as much as ~45 km from the aftershocks.

The M5.8 23 August 2011 Mineral, Virginia, earthquake was felt over nearly the entire eastern United States and was recorded by a wide array of seismic broadband instruments. The earthquake occurred ~200 km southeast of the boundary between two distinct geologic belts, the Piedmont and Blue Ridge terranes to the southeast and the Valley and Ridge Province to the northwest. At a dominant period of 3 s, coherent postcritical P-wave (i.e., direct longitudinal waves trapped in the crustal waveguide) arrivals persist to a much greater distance for propagation paths toward the northwest quadrant than toward other directions; this is probably related to the relatively high crustal thickness beneath and west of the Appalachian Mountains. The seismic surface-wave arrivals comprise two distinct classes: those with weakly dispersed Rayleigh waves and those with strongly dispersed Rayleigh waves. We attribute the character of Rayleigh wave arrivals in the first class to wave propagation through a predominantly crystalline crust (Blue Ridge Mountains and Piedmont terranes) with a relatively thin veneer of sedimentary rock, whereas the temporal extent of the Rayleigh wave arrivals in the second class are well explained as the effect of the thick sedimentary cover of the Valley and Ridge Province and adjacent Appalachian Plateau province to its northwest. Broadband surface-wave ground velocity is amplified along both north-northwest and northeast azimuths from the Mineral, Virginia, source. The former may arise from lateral focusing effects arising from locally thick sedimentary cover in the Appalachian Basin, and the latter may result from directivity effects due to a northeast rupture propagation along the finite fault plane.

Groundwater levels were offset in bedrock observation wells, measured by the U.S. Geological Survey or others, as far as 553 km from the M w 5.8 Mineral, Virginia (USA), earthquake on 23 August 2011. Water levels dropped as much as 0.47 m in 34 wells and rose as much as 0.15 m in 12 others. In some wells, which are as much as 213 m deep, the water levels recovered from these deviations in hours to days, but in others the water-level offset may have persisted. The groundwater-level offsets occurred in locations where the earthquake was at least weakly felt, and the maximum water-level excursion increased with felt intensity, independent of epicentral distance. Coseismic static strain from the earthquake was too small and localized to have contributed significantly to the groundwater-level offsets. The relation with intensity is consistent with ground motion from seismic waves leading to the water-level offsets. Examination of the hydrographs indicates that short-period ground motion most likely affected the permeability of the bedrock aquifers monitored by the wells.

The 23 August 2011 M 5.7 intraplate earthquake occurred in the Central Virginia seismic zone near Mineral, Virginia (USA), far from the nearest plate boundaries. I suggest here that this earthquake, as well as others occurring in this region since 1774, was triggered by pore-fluid pressure diffusion associated with groundwater recharge. Using finite element modeling (FEM) estimates are made of the magnitude and timing of pressure diffusion with respect to the time of the earthquake. Two scenarios are considered: (1) the diffusion took place along vertical to near-vertical diffusion paths, or (2) the diffusion was restricted to a hydraulically transmissive fracture zone that was later illuminated by thousands of aftershocks. Both scenarios have merit. The fracture zone may not have been entirely generated by the main shock. For either model a fractured crust is assumed to be recharged at the surface of the Earth by groundwater recharge with subsequent pore-fluid pressure diffusion propagating to the hypocenter. The transient behavior of pore-fluid pressure is examined at the focal depth of 8 km. These results are compared with the duration and timing of base flow (groundwater recharge) as estimated from a hydrograph separation. The delays of the peaks in pore-fluid pressure diffusion as computed by the FEM simulations are found to be consistent with the start and duration of groundwater recharge with respect to the timing of the Virginia earthquake.

A reconnaissance team documented the geotechnical and geological aspects in the epicentral region of the M w (moment magnitude) 5.8 Mineral, Virginia (USA), earthquake of 23 August 2011. Tectonically and seismically induced ground deformations, evidence of liquefaction, rock slides, river bank slumps, ground subsidence, performance of earthen dams, damage to public infrastructure and lifelines, and other effects of the earthquake were documented. This moderate earthquake provided the rare opportunity to collect data to help assess current geoengineering practices in the region, as well as to assess seismic performance of the aging infrastructure in the region. Ground failures included two marginal liquefaction sites, a river bank slump, four minor rockfalls, and a ~4-m-wide, ~12-m-long, ~0.3-m-deep subsidence on a residential property. Damage to lifelines included subsidence of the approaches for a bridge and a water main break to a heavily corroded, 5-cm-diameter valve in Mineral, Virginia. Observed damage to dams, landfills, and public-use properties included a small, shallow slide in the temporary (“working”) clay cap of the county landfill, damage to two earthen dams (one in the epicentral region and one further away near Bedford, Virginia), and substantial structural damage to two public school buildings.

The Mineral, Virginia (USA), earthquake of 23 August 2011 was an unusually strong seismic event in the eastern United States. It caused widespread structural damage to residential property near the epicenter. An analysis of residential property damage reports, in conjunction with visits to some damaged residences, reveals a 40 km 2 area of concentrated damage centered 11 km south of the town of Louisa. This area is west of the earthquake’s epicenter and may be in the immediate hanging wall of a northeast-striking, moderately southeast dipping causative fault suggested by seismic data. The degree of damage in this area is consistent with a maximum Modified Mercalli Intensity (MMI) of VIII. A surrounding area of ~550 km 2 reported damage that is consistent with an MMI intensity of VII. A statistical analysis of dwelling characteristics confirms that home age and condition were factors that influenced the frequency and severity of reported property damage. The median damage to homes constructed between 1900 and 1973, relative to assessed value, was approximately twice that of homes constructed after 1973 in Louisa County, and three times greater within areas of MMI intensity VI, VII, and VIII.

The 2011 Mineral, Virginia, earthquake is one of the larger recorded seismic events occurring east of the Rocky Mountains since seismic instrumentation was first deployed. The operation of the North Anna nuclear power station (NANPS), located ~22 km northeast of the epicenter, was affected by the earthquake vibration. This moderate event caused the first incident in which a commercial U.S. nuclear power plant experienced a safe shutdown as a result of earthquake ground motion. Post-earthquake investigations confirmed that important safety-related structures, systems, and components (SSCs) at the NANPS did not have any detectable damage. Damage at the NANPS consisted of cracking and spalling of some of the non-safety-related ancillary structures, and the plant was restarted after three months of intensive inspections and reviews. Response spectra developed from the recorded ground motion at the NANPS showed a modest exceedance of the plant seismic design levels for safety-related SSCs, but these SSCs were not damaged and maintained their functionality. The NANPS performance, in combination with other global examples, shows that nuclear power plants have been able to function safely even when earthquake ground motions exceeded the design levels of the SSCs. In this paper we describe the observed earthquake effects at the NANPS and discuss the original geologic and seismic characterization of the plant site. We also discuss the impacts of other earthquakes on the performance of various nuclear power plants, and previous and current seismic hazard and risk evaluations for U.S. nuclear power plants.