Regional-Scale Controls on Earthquake Ground Motions and their Implications for Assessing Historical Earthquake Magnitudes

Abstract

Historical earthquake magnitudes estimated from building damage distributions are essential for constraining the potential magnitudes of future earthquakes on fault systems, and their accumulated slip deficits, which underpin seismic hazard assessments. However, earthquake-induced damage patterns reflect complex trade-offs between earthquake magnitude and location, wave propagation effects, site conditions, and building vulnerability. The relative influence of these factors remain poorly quantified in continental foreland basin settings, where thrust earthquakes generate destructive ground shaking that propagates through laterally-varying geological structures. This study addresses this issue using seismic wavefield simulations with varying earthquake magnitudes, thrust fault locations, and basin structures. Combining these simulated ground motions with established PGV-intensity-magnitude relationships, we quantify how source and basin characteristics can contribute to over- or under-estimation of historical earthquake magnitudes that are based upon intensity observations. The simulation results show that proximity of the fault rupture to the foreland basin, and earthquake magnitude, are the dominant controls on the resultant ground motions. Increasing the fault's distance from the range front by ~50 km produces ground-motion reductions equivalent to decreasing earthquake magnitude by 1--2 Mw units. Basin structure superimposes an additional control on these ground motions, increasing peak ground accelerations by factors of 1.5--2 near the basin margin but becoming negligible for more distant events. In range-front settings, this study quantifies which locations and magnitudes of earthquakes are likely to be unrecorded in the historical record. Additionally, we show that whether or not the presence of foreland basin structure is accounted for when analyzing macroseismic intensity distributions can lead to mis-estimation of historical earthquake magnitudes by up to 0.5 Mw, with total source-path-site uncertainties reaching ~1 Mw. These systematic biases have significant implications for assessing historical moment release in continental collision zones.