[S-SS08] Fault Rheology and Earthquake Physics

[SSS08-04] Estimation of absolute shear stress in the source region of large earthquakes based on estimates of fault fracture energy

*Nelson Pulido¹ (1.National Research Institute for Earth Science and Disaster Resilience)

Keywords:fracture energy, large earthquake, shear stress, apparent stress, rupture velocity, fault

Fault fracture energy (G), a fundamental physical property of earthquake rupture has been widely estimated from the energy budget during earthquake fault rupture. To that purpose a simple relationship between fault slip and shear stress (linear slip weakening friction law without under or overshoot), as well as a constant fault rupture velocity are usually assumed. Under these assumptions, an estimate of fracture energy often referred to as G’, can be obtained from the difference of the average stress drop and two times the apparent stress. However fault rupture propagation and shear stress release during large earthquakes can be highly heterogeneous, and therefore the estimation of G from seismic radiation and the aforementioned assumptions, may lead to wrong estimates. To clarify this point, in a previous study I calculated fracture energy of large earthquakes (G) based on a heterogeneous fault rupture process, using the finite width slip pulse dynamic rupture model of Rice et al (2005). I also estimated G’ values based on fault average stress drops and apparent stress (earthquake radiated energy normalized by seismic moment times rigidity). To calculate G and G’ I used the extensive database of finite fault rupture models of large earthquakes (M>7) by NEIC. My results indicated that G’ systematically underestimates G. To provide a physical explanation for this difference I compared the energy budget from earthquakes obtained from simple linear slip weakening friction law, to the one obtained from a more general relationship where dynamic shear stress values (σ_d ) can drop below fault residual shear stress (σ_r ) during fault rupture. It can be shown that G can be expressed as follows; G = G’ + (σ_r - σ_d) δ , where δ is fault slip. Assuming a nearly complete strength drop during fault rupture (namely, σ_d ~ 0) for large earthquakes (as suggested by recent megathrust earthquakes), residual shear stress can be obtained as the difference between G and G’ divided by fault slip. Based on this idea I calculated lower bound estimates of absolute residual fault shear stress (as well as absolute initial shear stress , σ_o , by adding stress drops), using my dataset of G and G’ values. My estimations of average fault shear stress ( [σ_o+σ_r]/2 ) range from 0.3 MPa to 40 MPa. These results also enable the estimation of seismic efficiency (the ratio of radiated energy to total available potential energy), which has remained unresolved for earthquakes to date. My estimation of average seismic efficiency for the entire dataset is around 12%. I also obtained that seismic efficiency increases with increasing average fault rupture velocity.
My results also indicate that average shear stress is highly correlated with apparent stress. These results render possible the first order estimation of absolute shear stress in source regions of earthquakes globally, based on available measurements of apparent stress (radiated energy). My results also show that average shear stress and earthquake slip, have a weak negative correlation with seismic efficiency. These results may indicate to some extent that large earthquakes within high background shear stress regions might be less seismically efficient compared to earthquakes in regions with lower background shear stress.