Reconciling viscoelastic models of postseismic and interseismic deformation: Effects of viscous shear zones and finite length ruptures

Abstract

We have developed a suite of earthquake cycle models for strike‐slip faults to investigate how finite ruptures and lithosphere‐scale viscous shear zones affect interseismic deformation. In particular, we assess whether localized and stationary interseismic deformation and large‐scale, rapidly decaying postseismic transients may be explained with models incorporating either or both of these features. Models incorporating viscous shear zones give more stationary interseismic deformation than layered half‐space, Maxwell viscoelastic models producing similar early postseismic deformation in the near field. This tendency is accentuated when the effective viscosity per unit width of the shear zone increases either with depth or with interseismic time. Models with finite (200 km long) ruptures produce time‐dependent deformation similar to that from models with infinitely long ruptures, though with smaller‐magnitude and somewhat more localized surface velocity fluctuations. Models incorporating a stiff lithosphere, a low‐viscosity mantle asthenosphere, and a crust‐ to lithosphere‐scale viscous shear zone can replicate postseismic and interseismic deformation typical of large, strike‐slip earthquakes. Such models require depth‐dependent, power law, or transient rheologies for the viscous shear zone material in the crust, as long as the effective viscosity per unit shear zone width is approximately 1015 Pa s/m at crustal depths during the postseismic interval. Below the Moho, the shear zone effective viscosity per unit width must be higher throughout the seismic cycle, at least over a short depth interval. For example, if shear zone viscosity per unit width is 5×1016 Pa s/m in the mantle lithosphere, a model with a 50 km thick lithosphere and asthenosphere effective viscosities of 1 to 5×1018 Pa s can reproduce reference velocity profiles for both postseismic and later interseismic deformation.

Kinematically consistent models of viscoelastic stress evolution

“…Many classes of viscoelastic models have been used to successfully explain postseismic observations after large earthquakes (vertically layered viscoelastic models with Maxwell [ Hearn et al , ; Ryder et al , ], standard linear solid [ Ryder et al ., ], Burgers (transient) [ Ryder et al ., ], and power law rheologies [ Freed and Bürgmann , ] and models incorporating viscous shear zones [ Hearn et al ., , ], among others). To our knowledge, across all three‐dimensional models, only those incorporating Burgers rheologies [ Hetland , ; Meade et al ., ; P. R. DeVries et al, Viscoelastic block models of the North Anatolian Fault: A unified earthquake cycle representation of pre‐ and post‐seismic geodetic observations, submitted to Bulletin of the Seismological Society of America , 2016] and weak shear zones [ Yamasaki et al ., ; Hearn and Thatcher , ] have so far been used to explain both preearthquake and postearthquake data. Here we focus on the former class of models (Figure , inset).…”

“…To our knowledge, across all three-dimensional models, only those incorporating Burgers rheologies [Hetland, 2006;Meade et al, 2013;P. R. DeVries et al, Viscoelastic block models of the North Anatolian Fault: A unified earthquake cycle representation of pre-and postseismic geodetic observations, submitted to Bulletin of the Seismological Society of America, 2016] and weak shear zones [Yamasaki et al, 2014;Hearn and Thatcher, 2015] have so far been used to explain both preearthquake and postearthquake data. Here we focus on the former class of models (Figure 1, inset).…”

Statistical tests of simple earthquake cycle models

“…More recently, a viscoelastic block model was used to explain both pre-and post-Izmit earthquake observations with a Burgers rheology (log 10 η M ≈ 18.6 À 19.0 Pa s and log 10 η K ≈ 18.0 À 19.0 Pa s) [DeVries et al, 2016]. A number of studies have examined the viscoelastic effects of past earthquakes across both individual faults and entire fault systems [e.g., Dixon et al, 2003;Pollitz, 2003;Johnson and Segall, 2004;Smith and Sandwell, 2006;Johnson et al, 2009;Hilley et al, 2009;Chuang and Johnson, 2011;Johnson, 2013]; here we focus on studies that have examined data from throughout the earthquake cycle [e.g., Hetland, 2006;Meade et al, 2013;Yamasaki et al, 2014;Hearn and Thatcher, 2015].…”

“…Unified models that simultaneously explain interseismic and postseismic observations could lead to a more comprehensive understanding of time‐dependent deformation through the earthquake cycle [ Hetland , ]. Recent studies have focused on explaining geodetic observations from both before and after large ( M w > 7) earthquakes on the North Anatolian fault in Turkey [ Yamasaki et al ., ; Hearn and Thatcher , ] and the Kunlun fault in Tibet [ DeVries and Meade , ]. This approach is limited geographically to a few strike‐slip faults worldwide, because it requires geodetic data across individual faults at different stages of the earthquake cycle.…”

Geodetically constrained models of viscoelastic stress transfer and earthquake triggering along the North Anatolian fault

“…Two sets of models that may be able to explain both pre‐earthquake and post‐earthquake geodetic observations incorporate either viscous shear zones [e.g., Yamasaki et al ., ; Hearn and Thatcher , ] or Burgers rheologies [e.g., Hetland , ]. In particular, models that incorporate Burgers rheologies can explain pre‐earthquake and post‐earthquake data from both the NAF [ Hetland , ] and 15 strike slip faults worldwide in two dimensions [ Meade et al ., ], and may also be consistent with time‐dependent spatial patterns of far‐field seismic activity after large earthquakes [ Marsan and Bean , ].…”