Geology Reports⌕ Search

USGS · 70028671

Earthquake ground-motion prediction equations for eastern North America

Abstract

New earthquake ground-motion relations for hard-rock and soil sites in eastern North America (ENA), including estimates of their aleatory uncertainty (variability) have been developed based on a stochastic finite-fault model. The model incorporates new information obtained from ENA seismographic data gathered over the past 10 years, including three-component broadband data that provide new information on ENA source and path effects. Our new prediction equations are similar to the previous ground-motion prediction equations of Atkinson and Boore (1995), which were based on a stochastic point-source model. The main difference is that high-frequency amplitudes (f ??? 5 Hz) are less than previously predicted (by about a factor of 1.6 within 100 km), because of a slightly lower average stress parameter (140 bars versus 180 bars) and a steeper near-source attenuation. At frequencies less than 5 Hz, the predicted ground motions from the new equations are generally within 25% of those predicted by Atkinson and Boore (1995). The prediction equations agree well with available ENA ground-motion data as evidenced by near-zero average residuals (within a factor of 1.2) for all frequencies, and the lack of any significant residual trends with distance. However, there is a tendency to positive residuals for moderate events at high frequencies in the distance range from 30 to 100 km (by as much as a factor of 2). This indicates epistemic uncertainty in the prediction model. The positive residuals for moderate events at < 100 km could be eliminated by an increased stress parameter, at the cost of producing negative residuals in other magnitude-distance ranges; adjustment factors to the equations are provided that may be used to model this effect.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

G. M. Atkinson, D.M. Boore. 2006. Earthquake ground-motion prediction equations for eastern North America. https://doi.org/10.1785/0120050245

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

Re-evaluation of shaking, dynamic strain, and the source of the 1886 Charleston, South Carolina, earthquake

As the largest historical event in the eastern United States, the 1886 Charleston, South Carolina, earthquake, plays a prominent role in our nation’s seismological and human history and an outsized role in our probabilistic ground‐motion hazard estimates, affecting building codes, insurance rates, and geophysical monitoring. Thus, constraining the characteristics of the 1886 earthquake source, as well as the nature of its shaking distribution and impacts to the extent possible, is essential. Given its vintage, however, societal, geological, and geophysical datasets are limited compared with those used in modern seismological analyses. Because of these limitations, some characteristics of the rupture model are fundamentally nonunique. Still, by re‐evaluating all available data, I can constrain several key aspects of this earthquake. I show that several interpretations in recent analyses are inconsistent regarding the inferred source; shaking; ground failure; and, particularly, railway impacts. Importantly, I have shown that although dynamic strain caused the ground failure that led to track deformation, without the intermediate liquefaction step, dynamic strain alone was unlikely to be the source of rail deformation, contrary to earlier studies. Had earlier pronouncements that dynamic strain is a real concern for bending railways held true, there would be serious ramifications for railway engineering as well as for other above‐ and below‐ground linear infrastructure.

South Carolina↗

ETAS-positive: Extracting unbiased epidemic-type aftershock models from incomplete earthquake catalogs

Earthquake catalogs are inherently incomplete, and short-term fluctuations in catalog completeness can strongly bias estimates of seismicity parameters. Here, we extend the positive statistics framework recently introduced for estimating Gutenberg–Richter a and b values to the Epidemic-Type Aftershock Sequence (ETAS) model. The approach focuses on complete magnitude–time intervals defined by pairs of earthquakes in which the second event is larger than the first. Comprehensive synthetic tests demonstrate that the proposed ETAS + method retrieves nearly unbiased parameter estimates even from highly incomplete catalogs, whereas parameters obtained using classical estimation approaches are strongly biased. Application to Southern California reveals a similar trend, with the productivity scaling parameter α approaching the b value, consistent with self-similar earthquake triggering within estimation uncertainties. Owing to its robustness against catalog incompleteness and computational efficiency, the ETAS + method provides a powerful new tool for seismicity analysis and earthquake forecasting.

California↗