Performance evaluation of physics-based ground-motion simulations in the New Madrid region: effects of embayment structure and stochastic velocity variability

Junfei Huang, Arben Pitarka, Rie Nakata, Tang Houjun, & David B. McCallen

Submitted August 30, 2026, SCEC Contribution #15502, 2026 SCEC Annual Meeting Poster #231

The New Madrid seismic zone produced the 1811-1812 earthquake sequence, among the largest earthquakes in the United States, yet ground-motion simulation and validation in this region remain challenging because both recordings and subsurface constraints are sparse. Physics-based three-dimensional (3D) ground-motion simulation is emerging as a promising approach to engineering applications, but performance evaluation and validation are essential before it can be used confidently. Compared with the western United States, simulations in the central and eastern United States, including the New Madrid region, carry greater uncertainty because of limited recorded data and less constrained regional structure and source characterization.

In this study, we evaluate broadband (0-5Hz) 3D ground-motion simulations for five recent small earthquakes in the New Madrid region using the USGS Central U.S. velocity model as a baseline. We compare the original model with modifications that include layered sedimentary basin structure and spatial stochastic velocity perturbations to better account for inherent impedance contrasts and wave scattering. In addition, shallow-basin velocities are increased based on recent findings on deep shear-wave velocity structure measured in the region. These modifications are motivated by the deficiency of the smooth baseline model to reproduce the complex and long-duration waveforms recorded in the New Madrid region.

The results demonstrate that the baseline velocity model generally produces overly simple and short-duration waveforms relative to the recordings. In contrast, the modified model improves agreement in the waveforms, main-phase arrival time and amplitude, duration, and Fourier amplitude spectrum. The local resonance-periods, inferred from horizontal-to-vertical spectral ratios are in better agreement with the recordings after the model modifications. Long-period and long-duration motions are also observed in the basin near Memphis even for small earthquakes. Overall, the results highlight the importance of 3D basin structure and small-scale velocity heterogeneity in simulated ground motion amplification pattern in the New Madrid region. The study provides a model validation strategy for improving existing 3D velocity models and simulation methodology. It also lays the groundwork for simulating strong ground motion with engineering applications for scenario earthquakes in the region.

Key Words
New Madrid seismic zone, physics-based ground-motion simulation, validation, velocity model, basin structure, stochastic velocity variability

Citation
Huang, J., Pitarka, A., Nakata, R., Houjun, T., & McCallen, D. B. (2026, 08). Performance evaluation of physics-based ground-motion simulations in the New Madrid region: effects of embayment structure and stochastic velocity variability. Poster Presentation at 2026 SCEC Annual Meeting.

Related Projects & Working Groups
Ground Motions (GM)