Pradeep Talwani

Pore pressure diffusion and the mechanism of reservoir-induced seismicity

The study of reservoir-induced seismicity offers a controlled setting to understand the physics of the earthquake process. Data from detailed investigations at reservoirs in South Carolina suggested that the mechanism of transmission of stress to hypocentral locations is by a process of diffusion of pore pressure (Pp). These results were compared with available worldwide data. The ‘seismic’ hydraulic diffusivity, αs, was estimated from various seismological observations, and was found to be a good estimate of the material hydraulic diffusivity, α. Application of these results to a dedicated experiment to understand RIS at Monticello Reservoir, S.C., suggested that the diffusing Pp front plays a dual role in the triggering of seismicity. The spatial and temporal pattern of RIS can be explained by the mechanical effect of diffusion of Pp with a characteristic hydraulic diffusivity within an order of magnitude of 5×104 cm2/s, corresponding to permeability values in the millidarcy range. The triggering of seismicity is due to the combined mechanical effect of Pp in reducing the strength and, possibly, the chemical effect in reducing the coefficient of friction between the clays in the pre-existing fractures and the rocks that enclose these fractures.

Recurrence rates of large earthquakes in the South Carolina Coastal Plain based on paleoliquefaction data

We present a reanalysis of results of 15 years of paleoliquefaction investigations in the South Carolina Coastal Plain. All earlier radiocarbon age data and locations of organic material collected by various investigators were reviewed and recalibrated to obtain a uniform data set. The calibrated dates and the spatial extent of the sandblows having similar dates were used to estimate ages and magnitudes of prehistoric earthquake episodes. The results of this analysis suggest seven episodes (episodes A-G) of prehistoric liquefaction in the past 6000 years and two possible scenarios for their occurrence. In the first scenario, three seismic sources exist within the Coastal Plain of South Carolina; at Charleston (A, B, E, and G) with magnitudes M 71, Georgetown (C and F), and Bluffton (D) with magnitudes M ;6. In the second scenario, episodes C and D are combined into one episode, episode C9. In this scenario all earthquakes occurred at Charleston and with M 71. Episodes A and B seem to be more representative of the earthquake cycle and suggest a recurrence time of 500 - 600 years for M 71 earthquakes at Charleston. The recurrence times and magnitudes for episodes C and D are estimated at

Some seismological and geometric features of intraplate earthquakes

2000 years and ;6.0, respectively. The older episodes are less frequent, a fact that may be attributable to times of low ground water table. Before ;6000 years B.P., the ground water table was too low to permit observable liquefaction features to develop at the surface.

Paleoseismic Evidence for Recurrence of Earthquakes near Charleston, South Carolina

To understand the nature and cause of intraplate earthquakes we carried out a systematic search for common features associated with them. Various geological, geophysical and seismic data were analyzed and geometric and seismic parameters were compared for 29 intraplate events (M ⩾ 4.5) occurring worldwide. The results suggest that intraplate earthquakes occur by the reactivation of pre-existing zones of weakness in a compressional stress regime which is generally oriented parallel to the absolute direction of plate motion. Two styles of faulting were observed. For type A earthquakes, observed in the eastern United States, eastern China, western Europe and West Africa, deformation occurs by strike-slip motion on steeply dipping faults. Type B events are usually associated with thrust or normal faulting and were observed in eastern Canada, Fennoscandia, Australia and peninsular India. Focal depths did not explain the differences between type A and type B earthquakes; rather, the differences are due to the perturbation of a regional stress field by a local stress field. The regional stress field for type A events is primarily due to ridge push forces. The method of stress perturbation varied from region to region. The earthquakes in the type B category were further divided into three subcategories: type B1 (occurring in eastern Canada and Fennoscandia), type B2 (occurring in Australia, peninsular India and central Brazil) and type B3 (occurring in the Gulf Coast of the U.S.). The inferred causes of the perturbing stresses are deglaciation, intracontinental resistive forces resulting from the collision of plates and the deposition of a large thickness of sediments, respectively. Type A events were found to be associated with intersecting faults and generally supported the intersection model of Talwani (1988) and had analogs in plate boundary events.

Evidence of possible tectonic upwarping along the South Carolina coastal plain from an examination of river morphology and elevation data

A destructive earthquake that occurred in 1886 near Charleston, South Carolina, was associated with widespread liquefaction of shallow sand structures and their extravasation to the surface. Several seismically induced paleoliquefaction structures preserved within the shallow sediments in the meizoseismal area of the 1886 event were identified. Field evidence and radiocarbon dates suggest that at least two earthquakes of magnitudes greater than 6.2 preceded the 1886 event in the past 3000 to 3700 years. The evidence yielded an initial estimate of about 1500 to 1800 years for the maximum recurrence of destructive, intraplate earthquakes in the Charleston region.

Eastern Piedmont fault system: Speculations on its extent

Although the fault(s) responsible for the 1886 Charleston, South Carolina, earthquake have not yet been identified (primarily because of the lack of surface rupture), evaluation of Landsat imagery, aerial photography, and topographic maps have revealed an ∼200-km-long, ∼15-km-wide, north- northeast-trending zone composed of subtle topographic highs and morphologic changes in rivers that may be associated with tectonic activity. River anomalies observed within this zone include river bends that are convex toward the north- northeast, incised channels, changes in river patterns, and convex-upward longitudinal profiles. Analyses of geologic and geophysical data further indicate that these surface features may be the result of ongoing tectonic uplift along a north-northeast-trending fault zone possibly associated with recent seismicity near Charleston.

Evidence for a buried fault system in the Coastal Plain of the Carolinas and Virginia—Implications for neotectonics in the southeastern United States

Geologic mapping, interpretation, and field checking of recent aeromagnetic data suggest the existence of a closely associated series of faults and splays extending from Alabama to Virginia, herein termed the Eastern Piedmont fault system. Characteristic magnetic anomalies were found to be associated with known faults and were used to trace them through covered intervals. The fault system extends northeastward from the Goat Rock fault of Alabama and west-central Georgia, crossing the lower Piedmont of South Carolina, passes beneath a segment of the Coastal Plains in the Carolinas, and then flanks the Raleigh belt in North Carolina and continues into Virginia. From east-central Georgia to Virginia, cataclastic rocks along the faults of the system are bounded to the northwest and southeast by rocks of the Carolina slate belt, forming perhaps the most extensive fault system in eastern North America. Its movement history is similar to that of the Brevard fault: an early ductile mylonitic phase, followed by periods of brittle deformation. We interpret the fault system to have been initiated during collapse of the late Precambrian–early Paleozoic Carolina slate belt island arc. The Paleozoic continental suture probably lies farther east, buried beneath the Coastal Plain.

Internally consistent pattern of seismicity near Charleston, South Carolina

Geomorphic, geologic, and geophysical data suggest the presence of an ∼600-km-long, north-northeast–trending buried fault system in the Coastal Plain of the Carolinas and Virginia, herein named the East Coast fault system. The East Coast fault system is expressed by anomalous changes in fluvial geomorphology that locally coincide with one or more of the following features: linear aeromagnetic anomalies, buried faults interpreted from seismic reflection data, surface faults offsetting Pliocene–Pleistocene surficial units, locally brecciated phyllites and argillites, gently upwarped sediments, topographic highs, and seismicity near Summerville, South Carolina. River anomalies where the rivers traverse unconsolidated upper Pleistocene–Holocene flood-plain sediments are evidence of deformation along the East Coast fault system during the past 130 to 10 k.y. which may be ongoing. The fault system traverses the epicentral area of the 1886 Charleston, South Carolina, earthquake and lies west of paleoliquefaction sites along the outer South Carolina Coastal Plain; thus, the fault system could be the source of the Charleston earthquake and other large prehistoric earthquakes. Therefore, confirmation and demarcation of the East Coast fault system and elucidation of its history will help better assess seismic hazards in the southeastern United States.

Mechanism of Initial Seismicity Following Impoundment of the Monticello Reservoir, South Carolina

An improved velocity model for the meizoseismal area of the 1886 Charleston earthquake was used to relocate current seismicity, which showed marked separation into clusters. The relocated hypocenters and composite focal plane solutions were compared with available geophysical data to interpret their tectonic significance and possible association with the 1886 earthquakes. There is a distinct velocity discontinuity at a depth of about 10 km, where Vp increases from 5.9 to 6.45 km/s. The relocated hypocenters and composite focal plane solutions delineate two main source zones lying at different depths. The shallower zone, at 4 to 8 km depth and collinear with the Ashley River, is herein named the Ashley River seismogenic zone. The composite focal plane solution suggests reverse faulting on a steeply dipping northwest-striking fault with the southwest side upthrown. This zone is also associated with aeromagnetic and gravity anomalies. The deeper zone, at 9 to 13 km, suggests a right slip on a fault extending N26°E from east of Ravanel to Jedburg, a distance of more than 25 km, and dipping steeply to the west-northwest. Its location and extent are similar to the so-called Woodstock fault. Examination of geomorphic data suggests that there may be some ongoing tectonic uplift and subsidence in the area. The inferred P axes from fault-plane solutions are oriented S60°W. Firsthand accounts of the 1886 earthquakes suggest that two source areas were active in 1886 and the months that followed; I postulate that the two zones of current seismicity are coincident with the 1886 source areas.

Characteristic Features of Intraplate Earthquakes and the Models Proposed to Explain Them

Induced seismicity has been observed near Monticello Reservoir, South Carolina, since December 1977. Deployment of a seismic network before impound- ment allowed for detection and accurate location of pursuant seismicity since its inception. Corroborative fault-plane solutions, together with geological and borehole data on fracture orientations, made it possible to determine the structures associated with the initial seismicity. Earlier descriptions attributed this seismicity to the un- drained elastic response to impoundment of the reservoir or to a coupled poroelastic response, where diffusion of pore pressure and subsequent weakening was the pre- dominant cause. Quantitative evaluation of strength changes at hypocentral locations of a subset of 53 well-located earthquakes that followed the initial impoundment led to the following results: (1) The rocks in the vicinity of Monticello Reservoir are critically stressed, and strength changes less than or equal to 0.1 MPa are adequate to trigger seismicity; (2) except at locations on the periphery of the reservoir, and at shallow depths within it (1 km), impoundment of the reservoir led to strengthening at hypocentral locations due to the undrained elastic effect; (3) diffusion of pore pressure is the dominant mechanism for the observed seismicity; and (4) the inferred permeability of the fractures associated with seismicity, 5 10 14 m 2 (50 mD), lies within the range of seismogenic permeability associated with induced seismicity.