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 southern arm of the New Madrid seismic zone of the central United States coincides with the buried, ~110 km by ~20 km Blytheville Arch antiform within the Cambrian–Ordovician Reelfoot rift graben. The Blytheville Arch has been interpreted at various times as a compressive structure, an igneous intrusion, or a sediment diapir. Reprocessed industry seismic-reflection profiles presented here show a strong similarity between the Blytheville Arch and pop-up structures, or flower structures, within strike-slip fault systems. The Blytheville Arch formed in the Paleozoic, but post–Mid-Cretaceous to Quaternary strata show displacement or folding indicative of faulting. Faults within the graben structure but outside of the Blytheville Arch also appear to displace Upper Cretaceous and perhaps younger strata, indicating that past faulting was not restricted to the Blytheville Arch and New Madrid seismic zone. As much as 10–12.5 km of strike slip can be estimated from apparent shearing of the Reelfoot arm of the New Madrid seismic zone. There also appears to be ~5–5.5 km of shearing of the Reelfoot topographic scarp at the north end of the southern arm of the New Madrid seismic zone and of the southern portion of Crowley's Ridge, which is a north-trending topographic ridge just south of the seismic zone. These observations suggest that there has been substantial strike-slip displacement along the Blytheville Arch and southern arm of the New Madrid seismic zone, that strike-slip extended north and south of the modern seismic zone, and that post–Mid-Cretaceous (post-Eocene?) faulting was not restricted to the Blytheville Arch or to currently active faults within the New Madrid seismic zone.

Our models show patterns reflecting local fault control on both shoreline regression and river deflections along the Atlantic Coastal Plain. In these models, maximum displacement is assumed to be at the center of a fault, and both uplifts and downwarps are assumed to be of sufficient magnitude to influence surface processes. Models show regional shoreline regression: (1A) without localized uplifts; (1B) with different rates of regional uplift at either end; (1C) without any localized uplifts but with a large river-dominated delta; (2A) with a fault parallel to the shoreline with seaward side down or (2B) with seaward side up; and (3) with a fault perpendicular to the shoreline. Model 1A has consistently spaced parallel shorelines and an absence of river deflections, such as characterizes most of the late Pleistocene coastal plain across Georgia. Model 1B has divergence of shorelines toward and deflection of rivers away from the end with greater uplift. Model 1C has seaward deflections of shorelines with spacing dependent upon rates of sediment influx and removal by coastal processes. Models 2A and 2B represent interruptions of model 1 patterns. Both produce a seaward deflection and wider spacing of younger shorelines on the uplifted side of the fault with associated river deflections toward the margins of the uplift. Both also produce a landward deflection and closer spacing of younger shorelines coupled with convergence of rivers toward the downdropped basin. Model 3 produces a seaward deflection and wider spacing of older shorelines across the uplift associated with river deflections toward the margins of the uplift on one side of the fault. On the other side, there is a landward deflection and narrower spacing of younger shorelines on the downdropped side of the fault where river deflections merge toward the lowest area. In model 3, shorelines are discontinuous and may be difficult to correlate across the fault, and fault length is constrained by resumption of model 1 shorelines seaward of the fault. Model 3 matches patterns in the vicinity of the 1886 Charleston earthquake, South Carolina, with a NW-trending fault of ~50 km length with the NE side up and uplift continuing since the early Pleistocene. Very similar patterns occur in the vicinity of Beaufort, South Carolina, and Wilmington, North Carolina, which suggest other NW-trending faults of comparable or greater length may be present near these localities. Model 2A matches patterns near the Okefenokee Swamp, which suggests that a 100-km-long, N-trending fault may border the east side of Trail Ridge near the Georgia-Florida state boundary. Model 2B was used by previous workers to explain zones of river anomalies in the Carolinas, but those anomalies do not match this model.

Prior to the early 1990s, nearly all surface faults recognized in south Louisiana were faults of the Baton Rouge system. Since then, the number of surface fault traces interpreted in the region has increased dramatically, owing to a combination of (1) application of traditional analysis of cues on topographic maps and aerial- photographic imagery over increasingly large areas, particularly in southwest Louisiana, (2) the employment of geophysical surveying techniques in the Holocene delta plain where surface scarp relief is negligible, and (3) the advent of light detection and ranging (LiDAR) digital elevation models (DEMs). Like faults of the Baton Rouge system, newly recognized surface faults of the Tepetate system show distinctive depth-displacement relations in Quaternary and pre-Quaternary strata indicating that they are active and are the surface expressions of deep-subsurface older Cenozoic growth faults that have been reactivated following extended periods of quiescence. The differential displacement of older relative to younger Pleistocene terrace surfaces characteristic of individual Tepetate–Baton Rouge system faults also characterizes surface faults of the other systems, suggesting they may share similar movement histories. Commonplace recognition of active surface faults throughout south Louisiana now suggests that many of the known deep-subsurface growth-fault systems have surface expression reflecting their reactivation in the late Cenozoic. The recently amplified picture of surface faulting in this region highlights an important aspect of coastal tectonics in the northern Gulf of Mexico setting, and can provide useful constraints for modeling tectonics in this and other coastal settings characterized by reactivation of growth faults following lengthy intervals of quiescence.

The recent availability of high-resolution light detection and ranging (LiDAR) data in Connecticut has led to the discovery of a 125-km-long, northeast-trending geomorphic lineament, herein named the Eastford lineament, in eastern Connecticut and south-central Massachusetts. It appears to represent a surface expression of the 50-km-long Eastford fault that continues to the northeast and southwest. The proposed fault zone appears to be post-Triassic in age, since the Eastford fault and other shorter mapped faults along the lineament offset the ~201-m.y.-old Higganum dike system and Paleozoic lithotectonic terranes and structures. An integration of borehole data from the Moodus Deep Well, the 1987 cluster of seismicity north of Moodus, and faults identified from recently acquired seismic surveys suggest that displacements on the proposed continuation of the Eastford fault north of Moodus are the source of the Moodus earthquakes. In south-central Connecticut, the Eastford lineament is truncated by a 14-km-long, northwest-trending fault zone, herein named the Bunker Hill fault zone, which is associated with an en echelon zone of LiDAR lineaments. Like the Eastford lineament, the Bunker Hill fault zone appears to dextrally offset the Higganum dike system and Paleozoic formations on the western flank of the Killingworth Dome. An abrupt change in strike of the Higganum dike system and LiDAR lineaments across the Bunker Hill fault zone suggests, however, that this offset may be, at least in part, from dip-slip motion on the Bunker Hill fault zone. To the northwest, the Bunker Hill fault zone continues across the Eastern Border fault and part of the Jurassic Portland Formation of the Hartford Basin, which suggests that it postdates the latest phase of rifting in the basin. It also coincides with a linear, northwest-trending aeromagnetic low that sinistrally offsets a linear, northeast-trending positive aeromagnetic anomaly by ~1.4 km.

Numerous faults in the Mississippi Embayment area of far southern Illinois have been active since Late Cretaceous time. They have been documented via geologic mapping, drilling and trenching studies, and high-resolution seismic-reflection surveys. The majority lie within the Fluorspar area fault complex, directly in line with the New Madrid seismic zone. Other active elements include the Commerce fault zone and southern end of the Ste. Genevieve fault zone. Most Cenozoic faults strike northeast; a few strike north to north-northwest. Dips are generally 60° to vertical. Narrow grabens, often deeper than they are wide, are typical. Some grabens contain sedimentary units that are elsewhere eroded, or restricted to the grabens. Faults have undergone multiple episodes of movement since the Late Cretaceous. Offsets of tens of meters are common in the late Miocene to early Pleistocene Mounds Gravel. At one site, the Mounds Gravel is downthrown by 150 m within a graben. Younger Pleistocene sediments overlie the Mounds Gravel and are confined to the graben. Illinoian and younger displacements are less common and amount to a few meters, at the most. Two sites show possible Holocene faulting. The overall fault pattern implies transtensional, right-lateral wrenching that produced pull-apart grabens. Such a pattern is consistent with the contemporary stress regime, both in southernmost Illinois and in the New Madrid seismic zone.

The East Tennessee seismic zone in the southern Appalachians is an ~75-km-wide, 350-km-long region of seismicity that extends from NE Alabama and NW Georgia to NE of Knoxville, Tennessee. It is the second most active seismic zone east of the U.S. Rocky Mountains. Although the East Tennessee seismic zone has not recorded historical earthquakes of M > 5, researchers have used hypothetical and theoretical relationships to suggest that it may be capable of generating an “infrequent” M ~7.5 quake. To help clarify the late Pleistocene earthquake history and the earthquake potential of the East Tennessee seismic zone, we conducted an 18-mo pilot study to seek evidence of paleoseismic activity and have made important discoveries. ENE of Knoxville, Tennessee, in late Pleistocene French Broad River alluvium, we discovered: (1) strike-slip, thrust, and normal faults involving bedrock and alluvium at three sites, and widespread bleached or clay-filled fractures; (2) paleoliquefaction; and (3) anomalous fractured and disrupted features at three sites attributable to liquefaction and forceful groundwater expulsion and fluidization during or immediately after two or more major late Quaternary earthquakes. All of these features were produced by seismic events with a probable minimum M ~6.5. Optically stimulated luminescence dates at four sites provide maximum ages of 73–112 ka for at least two events. Upward penetration of at least two generations of fractures, clastic-sediment intrusions, and faults into the Bt horizons of Ultisols at several sites implies that two strong shocks occurred sometime after ~73 ka, and possibly much later than 73 ka. Two exposures in terrace alluvium E and W of the Tennessee Highway 92 bridge S of Dandridge, Tennessee, were graded and geologically mapped at 1 in. = 5 ft. The site W of the bridge revealed at least three sets of crosscutting fractures that terminate upslope against the base of an overlying late Pleistocene colluvium. The E site revealed numerous fractures and a fault with ~20 cm of sinistral displacement. Moreover, several “fluidization boils” containing shale clasts from below are cut by younger, red, clay-filled fractures. Few of these fracture sets in the Quaternary sediments parallel those in bedrock of the Tennessee Valley and Ridge and Blue Ridge geologic provinces that host the East Tennessee seismic zone, and these fractures are poorly aligned with the present-day N70E maximum principal stress orientation. A third site, 5 km SW of Dandridge on the NW side of Douglas Reservoir, contains at least two NW-vergent thrust faults that transported weathered bedrock 25–50 cm over late Pleistocene alluvium. At the same site, a 12-m-long mode 1 branching fracture in Sevier Shale is filled with Quaternary sediment, and is truncated by the largest thrust fault at 1–2 m depth. This structure, including the Quaternary sediment it contains, is also displaced 10 cm along a NW-trending sinistral fault. The discovery of faults at the ground surface that displace both bedrock and terrace alluvium contrasts with the modern seismicity, which occurs at 5–26 km depth in rocks below the basal décollement of major Paleozoic thrust sheets. Collectively, these initial findings imply that the East Tennessee seismic zone has produced coseismic surface faulting and generated at least two strong (M > 6.5) earthquakes during the late Quaternary.

A fundamental goal of intraplate tectonics research is to understand the role of crustal discontinuities in the distribution of Quaternary surface ruptures. Geophysical studies of southern North America on the Gulf of Mexico Coastal Plain reveal a buried Cambrian craton margin (Alabama-Oklahoma transform) that strikes southeast beneath Mesozoic and Cenozoic passive-margin sediments and Paleozoic thrust sheets. Seismic-reflection profiles show a graben system (Saline River fault zone) related to an episode of Triassic rifting above this transform margin during initial opening of the Gulf of Mexico. Post-Triassic reactivation of the Saline River fault zone produced normal and reverse faulting and strike-slip flower structures that can be linked to Quaternary surface deformation. We investigated surface and shallow Quaternary faulting along the Saline River fault system in south-central North America. Our field sites show late Pleistocene to late Holocene surface and near-surface deformation along lineaments and scarps of the Saline River fault zone. Age constraints from three sites are consistent with a surface rupture as long as 70 km ca. 6–5 ka. Mid-Holocene sand blows in the region may record strong shaking from such a large earthquake about the same time, but available ages provide only broad constraints on the timing of large earthquakes in the Saline River fault zone. The Saline River fault zone significantly expands the known area of paleoseismicity in midcontinent North America. This study builds upon our understanding of the Saline River fault zone's relationship to a transform craton margin and thus advances our understanding of active intraplate deformation.

Intraplate earthquakes within the eastern United States represent brittle faulting in the upper crystalline crust at shallow to moderate depths. The premise—that intraplate seismicity in crystalline crust occurs along postorogenic brittle faults formed during extensional events—permits us to distinguish three large intraplate-earthquake domains in the eastern United States. These domains are: the Appalachian accreted terranes; the Grenvillian accreted terranes; and the midcontinent accreted terranes. These crystalline domains are separated by terrane boundaries, where metamorphic zones, formed during terrane accretion, seal off older brittle faults in an older terrane from any direct connection with newer brittle faults in a newer accreted terrane. Strain within each domain accumulates on preexisting fault zones formed during previous postorogenic extensional events. Although older faults in adjacent regions may be sealed off within independent terranes, extensional events that postdate accretion may generate younger brittle faults that cross the sealed boundaries. Thus, breaches created by younger migrating hotspots, failed rifts, and/or impact structures may provide local connecting zones across sealed boundaries. Detailed fracture studies within the Southern Appalachian accreted terranes, Triassic rift basins therein, and overlapping Cretaceous–Cenozoic passive-margin strata of the southern Atlantic Coastal Plain show that recurrent Mesozoic–Cenozoic brittle faulting: (1) is postorogenic and occurs along fracture sets formed during failed-to-successful Mesozoic rifting during the breakup of Pangea and opening of the Atlantic Ocean; (2) is not related to brittle reactivation of orogenic ductile shear zones or metamorphic fabrics; (3) is largely confined to accreted crystalline terranes in the Appalachians separated by metamorphosed zones, referred to previously, which are herein called sealed boundaries; (4) has younger, shorter fracture sets that are confined by older fault zones; and (5) formed in a temporal sequence of fracture sets that have a hierarchical ordering and scaling of recurrently active fault zones, with older, linked sets forming longer, more through-going fault zones, which bound large polygonal crustal blocks within accreted terranes. The second intraplate-earthquake domain is the Mesoproterozoic Grenville basement. Crystalline terranes of the Grenville orogen (ca. 1.2–0.9 Ga) were accreted during formation of the Rodinian supercontinent. The Rigolet phase (ca. 1.02–0.9 Ga) of the Grenville orogeny was characterized by a shift from contraction to NW-SE extension with development of core complexes (e.g., Adirondacks). Protracted cooling, continued NW-SE-extension, and passage of a hotspot (ca. 750–600 Ma) preceded the breakup of Rodinia and opening of the Iapetus Ocean at ca. 565 Ma. Thus, the Grenville orogen also contains a sequence of postorogenic brittle fracture sets, which subsequently formed recurrently active brittle fault zones. The third intraplate-earthquake domain contains the Archean–Mesoproterozoic orogenic belts in the midcontinent region, which are cut by brittle fracture sets related to: failed rifting events (e.g., ca. 1.1 Ga Midcontinent rift; Cambrian Southern Oklahoma aulacogen, Reelfoot rift, and Rough Creek graben); successful Triassic rifting and opening of the Gulf of Mexico; and Mid-Cretaceous passage of the Bermuda hotspot. Regional fracture sets formed during each of these postorogenic events and the temporal sequence of fracture sets determine the hierarchical ordering and scaling of recurrently active fault zones. Fault-plane solutions in major seismic zones within these three domains (Charleston, South Carolina, and central Virginia; east Tennessee and Giles County, Virginia; New Madrid and Wabash—Arkansas, Illinois, Kentucky, Missouri, Tennessee) are consistent with their occurrence along reactivated older fault zones.

The study of paleoliquefaction grew out of (1) the recognition that earthquakes left their imprint on soft sediments as deformational structures primarily through liquefaction, and (2) the need for applying paleoseismology to settings in which active faults were not readily recognizable, accessible, or did not reach the surface. Earthquake-induced liquefaction features are distinctive, and their formation is a result of strong ground shaking that may or may not result in lateral spreading. Paleoliquefaction features include sand blows, and intrusive dikes and sills, as well as less prominent, but equally informative features such as load casts, pseudonodules, ball-and-pillow structures, and convolute lamination. Fluvial depositional environments, with generally easy access and relatively abundant natural outcrops, have been the primary choice for conducting paleoseismic studies. In general, when relying on sand dikes and sills, sand blows, and their related structures common in fluvial sediments, one is restricted to river valleys. Depending on the density of the drainage network and the size of the streams, one may not obtain as much data as would be desired. Therefore, other environments that may contain earthquake-induced liquefaction structures may need to be sought out. Lacustrine and paleolacustrine deposits also have a distinctive suite of liquefaction-induced sedimentary structures, most commonly pseudonodules and load casts, and less commonly convolute lamination. However, these structures are not limited to lacustrine deposits, because they have been observed in paleoliquefaction source beds in the New Madrid seismic zone, liquefaction accompanying the 26 January 2001, Bhuj, India, earthquake, and liquefaction associated with the Charlevoix seismic zone. The earliest earthquake-induced paleoliquefaction features were described and correlated to specific earthquakes using soft-sediment deformational structures in lake sediments and a series of modern earthquakes in California, and deformational structures in prehistoric lake sediments. Ancient lake deposits have been used for paleoseismological studies with some success in the United States in California, Oregon, Washington, and Alaska. Such deposits have also been successfully used in Europe and the Middle East. The most promising of these studies have been in varved glaciolacustrine deposits. Varves are small-scale (centimeter to millimeter) sedimentary units. They form in a variety of marine and lacustrine depositional environments from seasonal variation in clastic, biological, and chemical sedimentary processes. The most common seismically induced structures that occur in varved glaciolacustrine deposits are pseudonodules.

Paleoseismic investigations in fluvial deposits frequently use large-scale (many centimeters to decimeters wide) ground-failure features of liquefaction origin as indicators of larger earthquakes (i.e., exceeding M ~6). Such large features are not the only signature of seismicity, however. Seismic shaking often produces an abundance of small-scale features (millimeter to centimeter in size) such as sills, small clastic dikes, and ground fractures, which can vary widely in height and range from paper-thin to a few centimeters wide. These small-scale seismic signatures commonly form in field settings where large liquefaction features are absent, such as regions with a reduced susceptibility for liquefaction or sites far from earthquake meizoseismal regions where shaking levels were lower. Thus, these small signatures have the potential to significantly expand the geographic area useful for paleoseismic studies, yet they are not typically sought in most paleoseismic field studies because many can develop nonseismically, and interpreting their formative origin can be challenging. We examined small-scale features that occur in association with large liquefaction features at a variety of field sites across the United States. We present new criteria, with many photographic examples, to evaluate whether small-scale features and ground fractures were seismically generated. Although this research was done primarily in fluvial settings in the United States, these criteria should be applicable worldwide in many field settings with clastic sediments, potentially giving the study of small-scale seismic features and fractures a significant role in future paleoseismic investigations.

In New England, earthquakes pose a risk to the built environment. Emergency preparedness and mitigation planning are prudent in this region as older unreinforced masonry buildings and numerous critical facilities are common. New England state geological surveys cooperate with the Northeast States Emergency Consortium (NESEC) to improve risk communication with emergency managers. To that end, Connecticut, Maine, Massachusetts, and Vermont employed surficial geologic maps, deglaciation history, knowledge of the glacial stratigraphy, and professional judgment to reclassify surficial geologic material units into one of the five National Earthquake Hazards Reduction Program (NEHRP) site classifications (A, B, C, D, and E). These new classifications were used as a substitute for the HAZards U.S. Multi-Hazard (HAZUS-MH) site class value of “D,” which is used throughout New England as a default value. In addition, coding of surficial geologic materials for the five NEHRP site classifications was compared with classifications using the Wald methodology, a method that uses a slope analysis as a proxy for shear-wave velocity estimates. Comparisons show that coding to site classes using the Wald methodology underestimates categories A (high-velocity shear-wave materials, least relative hazard) and E (lowest-velocity shear-wave materials, greatest relative hazard) when evaluated side by side with coding done with the aid of surficial geologic maps. North of the glacial limit, derangement of drainage resulted in extensive ponding of meltwaters and the subsequent deposition of thick sequences of lacustrine mud. Inundation by the sea immediately following deglaciation in New England resulted in the deposition of spatially extensive and locally thick sequences of glacial marine mud. Surficial geologic maps better capture this circumstance when compared with the Wald topographic slope analysis. Without the use of surficial geologic maps, significant areas of New England will be incorrectly classified as being more stable than the site conditions that actually exist. By employing surficial geologic information, we project an improved accuracy for HAZUS-MH earthquake loss estimations, providing local and regional emergency managers with more accurate information for locating and prioritizing earthquake planning, preparedness, and mitigation projects to reduce future losses.

In this paper, we describe the scientific basis for the source and ground-motion models applied in the 2008 National Seismic Hazard Maps, the development of new products that are used for building design and risk analyses, relationships between the hazard maps and design maps used in building codes, and potential future improvements to the hazard maps.