Crustal Stress State in Taiwan: Moderately Strong Crust Supporting Gravitational and Flexural Loading

Abstract

In Taiwan and other contractional tectonic wedges, there is evidence that the basal detachment is very weak and the overlying crust is relatively strong. In this study, we compute the absolute deviatoric 3-D stress field in Taiwan using focal mechanism data and coseismic stress changes for the 1999 M7.6 Chi-Chi earthquake. We estimate a minimum coefficient of friction of 0.3 in the upper 15 km crust, assuming hydrostatic pore pressure, consistent with a moderately strong crust. The estimated stress state is consistent with primary tectonic structures of Taiwan. The maximum compressive stress in the fold and thrust belt and the Coastal Range is approximately aligned with relative motion of the Eurasian and the Philippine Sea Plates, but the minimum and intermediate principal stresses vary laterally and with depth. A reverse-faulting stress regime wraps around the Peikang high in western Taiwan suggesting inherited basement structure influences crustal stress state. Vertical maximum compressive stress is inferred under the high topography of the Central Range mountains. We present a simple stress model for Taiwan assuming spatial variations in gravitational potential energy are supported elastically. The model stress state is broadly consistent with the stress inversion result. The model shows that vertical maximum compressive stress in the Central Range is a result of high topography and thick crustal root. Consistent with a moderately strong crust, the model deviatoric stresses in areas of high seismicity in Taiwan are of the order 50-100 MPa. Lower deviatoric stresses under the Central Range are consistent with low rates of observed seismicity.

Earth's Free Surface Complicates Inference of Absolute Stress From Earthquake‐Induced Stress Rotations

“…Additionally, the HH01 model simplifies the background stress amplitude and the stress change to single values, which may not be applicable for earthquakes with substantial spatial variability in slip and stress drop (e.g., Lay et al., 2011). In these cases, stress‐estimation methods based on numerical modeling (Wesson & Boyd, 2007; Yang & Johnson, 2020; Yang et al., 2013; Yoshida et al., 2014, 2015) could be used.…”

“…Additionally, the HH01 model simplifies the background stress amplitude and the stress change to single values, which may not be applicable for earthquakes with substantial spatial variability in slip and stress drop (e.g., Lay et al, 2011). In these cases, stress-estimation methods based on numerical modeling (Wesson & Boyd, 2007;Yang & Johnson, 2020;Yang et al, 2013;Yoshida et al, 2014Yoshida et al, , 2015 could be used. Yin and Rogers (1995) and Hardebeck and Hauksson (2001) developed analytical methods for estimating the differential stress from an earthquake-induced rotation of the principal stress axes.…”

“…Both methods relate the rotation of the principal stresses to the ratio between the earthquake stress change and the background stress. More sophisticated methods for estimating absolute stress from stress rotations have been introduced, but require numerical modeling of the earthquake stress changes (Wesson & Boyd, 2007;Yang & Johnson, 2020;Yang et al, 2013;Yoshida et al, 2014Yoshida et al, , 2015.…”