2D finite-element modelling of the interaction between poroelastic effects and viscoelastic relaxation during the seismic cycle

<p>The analysis of the Coulomb stress changes has become an important tool for seismic hazard evaluation because such stress changes may trigger or delay next earthquakes. Processes that can cause significant Coulomb stress changes include coseismic slip, earthquake-induced poroelastic effects as well as transient postseismic processes such as viscoelastic relaxation. In this study, we investigate the spatial and temporal evolution of pore fluid pressure changes and fluid flow during the seismic cycle, their dependency on the permeability in the crust and the interaction with postseismic viscoelastic relaxation. To achieve this, we use 2D finite-element models for intra-continental normal and thrust faults, which include coseismic slip, poroelastic effects, postseismic viscoelastic relaxation and interseismic stress accumulation. In different experiments, we vary (1) the permeability of the upper and lower crust while keeping the viscosity structure constant and (2) the viscosity of the lower crust and lithospheric mantle, while we keep the permeabilities constant. (1) The modelling results show that the highest changes in pore fluid pressure during and after the earthquake occur within a distance of ~ 1 km around the lower fault tip at the transition between upper and lower crust. The evolution of pore pressure and fluid flow depends primarily on the permeability in the upper crust. With decreasing permeability, the possibility of the pore fluids to flow decreases and thus, in the postseismic phase, the duration of the poroelastic relaxation increases, from a few days to several years, until the pore pressure reaches the initial pressure of the preseismic phase. In contrast, the influence of variations of the permeability in the lower crust on the pore pressure changes is negligible. For high upper-crustal permeabilities, postseismic vertical velocities are high and decreases rapidly with time, from around 120 mm/a after the first year by two orders of magnitude after 10 years, whereas for low permeabilities they remain consistently low over the years after the earthquake. (2) Models with low viscosity of the lower crust show that the timescales of poroelastic effects and viscoelastic relaxation overlap and affect the postseismic velocity already in the early postseismic phase and that both processes decay within a few years after the earthquake. For higher viscosities, the velocity is initially dominated by pore pressure changes during the first few years, whereas viscoelastic relaxation lasts for decades. Both processes also show differences in their spatial scale. Poroelastic effects occur within a few kilometers around the fault, whereas viscoelastic relaxation acts on tens to hundreds of kilometers. As both processes can cause Coulomb stress changes on faults in the vicinity of the earthquake source fault, it is important to understand the spatial and temporal evolution, the effects on the individual faults and the interaction of both processes during the earthquake cycle. Future work will therefore include the calculation and examination of Coulomb stress changes on intra-continental normal and thrust faults using 3D models that include poroelastic effects and viscoelastic relaxation.</p>