Fully Dynamic Spontaneous Rupture Due to Quasi‐Static Pore Pressure and Poroelastic Effects: An Implicit Nonlinear Computational Model of Fluid‐Induced Seismic Events. Issue 11 (9th November 2018)

• Record Type: Journal Article
• Authors: Jin, Lei; Zoback, Mark D.
• Is Part Of: Journal of geophysical research. Volume 123:Issue 11(2018)
• Page Start: 9430
• Page End: 9468
• Publication Date: 2018-11-09
• Subjects: induced seismicity -- fault fluid pressure -- poroelasticity -- fully dynamic spontaneous rupture -- slip weakening -- modeling and computation; Geomagnetism -- Periodicals; Geochemistry -- Periodicals; Geophysics -- Periodicals; Earth sciences -- Periodicals; 551.1
• DOI: 10.1029/2018JB015669
• Languages: English
• ISSNs: 2169-9313

Abstract

Abstract: Fluid perturbations play a pivotal role in triggering earthquakes. However, the role of fluid in the coseismic rupture process remains largely unknown. To this end, we develop a 2‐D fully dynamic spontaneous rupture model for fluid‐induced earthquakes. The effect of fluid in the preseismic quasi‐static regime is modeled as either pore pressure diffusion or fully coupled poroelasticity, using our Jin and Zoback (2017, https://doi.org/10.1002/2017JB014892 ) computational model. The two approaches lead to radically different predictions on the time of earthquake nucleation. Correspondingly, the evolved fluid pressure or poroelastic stress on the fault, together with the spatially altered density of the fluid‐saturated hosting rock, is passed to the dynamic regime. Under the assumption of an undrained coseismic fluid‐solid system, we discretize the fully dynamic Cauchy equation of motion subjected to an exact fault contact constraint using a split‐node finite element method in space and an implicit Newmark family finite difference method in time. Within each time step, a fully implicit Newton‐Raphson scheme is implemented iteratively for linearizing the fully discrete equations. Within each Newton iteration, a physics‐based and nonstationary preconditioner is designed to accelerate the convergence of the selected generalized minimal residual method iterative linear solver. The effect of fluid is highlighted throughout the discretization and computational procedures. Finally, by conducting a numerical experiment, we illustrate that a fully coupled poroelastic model can lead to significantly different predictions on coseismic rupture behaviors and wavefields compared to a decoupled model. Our computational model can also serve as one of the earliest full‐physics modeling tool for fluid‐induced earthquakes. Key Points: A new full‐physics model of fluid‐induced fully dynamic and spontaneously rupturing earthquake is introduced A new numerical modeling framework including discretization, linearization, and physics‐based nonstationary preconditioning is presented Fully coupled poroelastic models can lead to notably different predictions on earthquake nucleation, dynamic rupture, and wavefields than decoupled models