Stress field constraints on intraplate seismicity in eastern North America

Abstract

Focal mechanisms of 32 North American midplate earthquakes (mb = 3.8–6.5) were evaluated to determine if slip is compatible with a broad‐scale regional stress field derived from plate‐driving forces and, if so, under what conditions (stress regime, pore pressure, and frictional coefficient). Using independent information on in situ stress orientations from well bore breakout and hydraulic fracturing data and assuming that the regional principal stresses are in approximately horizontal and vertical planes (±10°), the constraint that the slip vector represents the direction of maximum resolved shear stress on the fault plane was used to calculate relative stress magnitudes defined by the parameter ϕ = (S2 ‐ S3)/(S1 ‐ S3) from the fault/stress geometry. As long as the focal mechanism has a component of oblique slip (i.e., the B axis does not coincide with the intermediate principal stress direction), this calculation identifies which of the two nodal planes is a geometrically possible slip plane (Gephart, 1985). Slip in a majority of the earthquakes (25 of 32) was found to be geometrically compatible with reactivation of favorably oriented preexisting fault planes in response to the broad‐scale uniform regional stress field. Slip in five events was clearly inconsistent with the regional stress field and appears to require a localized stress anomaly to explain the seismicity. Significantly, all five of these events occurred prior to 1970 (when many regional networks were installed), and their focal mechanisms are inconsistent with more recent solutions of nearby smaller events. The frictional likelihood of the geometrically possible slip on the selected fault planes was evaluated in the context of conventional frictional faulting theory. The ratio of shear to normal stress on the fault planes at hypocentral depth was calculated relative to an assumed regional stress field. Regional stress magnitudes were determined from (1) S1/S3 ratios based on the frictional strength of optimally oriented faults (the basis for the linear brittle portion of lithospheric strength profiles), (2) the computed relative stress magnitude (ϕ) values, and (3) a vertical principal stress assumed equal to the lithostat. Two end‐member possibilities were examined to explain the observed slip in these less than optimally oriented fault planes. First, the frictional coefficient was held constant on all faults, hydrostatic pore pressure was assumed regionally, and the fault zone pore pressure was determined. Since pore pressure is a measurable quantity with real limits in the crust (P0 < S3), this end‐member case was used to determine which of the geometrically possible slip planes were frictionally likely slip planes. Alternately, pore pressure was fixed at hydrostatic everywhere, and the required relative lowered frictional coefficient of the fault zone was computed. Slip in 23 of the 25 geometrically compatible earthquakes was determined to also be frictionally likely in response to an approximately horizontal and vertical regional stress field derived from plate‐driving forces whose magnitudes are constrained by the frictional strength of optimally oriented faults (assuming hydrostatic pore pressure regionally). The conditions for slip on these 23 relatively “well‐oriented” earthquake faults were determined relative to this regional crustal strength model and require only moderate increases in pore pressure (between about 0.4–0.8 of lithostatic, hydrostatic is about 0.37 of lithostatic) or, alternately, moderate lowering (<50%) of the frictional coefficient on the faults which slipped. Superlithostatic pore pressures are not required. Focal mechanisms for the two other earthquakes with slip vectors geometrically consistent with the regional stress field, however, did require pore pressures far exceeding the least principal stress (or extremely low coefficients of friction). These events may reflect either local stress rotations undetected with current sampling or poorly constrained focal mechanisms. The analysis also confirmed a roughly north to south contrast in stress regime between the central eastern United States and southeastern Canada previously inferred from a contrast in focal mechanisms between the two areas: most central eastern United States earthquakes occur in response to a strike‐slip stress regime, whereas the southeastern Canadian events require a thrust faulting stress regime. This contrast in stress regime, with a constant maximum horizontal stress orientation determined by far‐field plate‐driving forces, requires a systematic lateral variation in relative stress magnitudes. Superposition of stresses due to simple flexural models of glacial rebound stresses are of the correct sense to explain the observed lateral variation, but maximum computed rebound‐related stress magnitude changes are quite small (about 10 MPa) and do not appear large enough to account for the stress regime change if commonly assumed stress magnitudes determined from frictional strength apply to the crust at seismogenic depths.