Temperature insensitive viscous deformation limits megathrust seismogenesis

Three models have been proposed to explain the downdip limit of the subduction seismogenic zone. The first is a temperature-controlled transition in rate-and-state frictional properties between 350-510°C, which inhibits earthquake nucleation. The second places the limit at the frictional and viscous failure envelope intersection. The third combines thermal and lithological controls, where ‘warm’ subduction zones are controlled by a 350°C frictional transition and ‘cold’ subduction zones are limited by the overriding plate Moho. To evaluate these hypotheses, we integrate thermal models with seismicity catalogs from 17 subduction zones. Observed depth limits remain remarkably consistent (~50 km) across a temperature range exceeding 250°C, indicating that the temperature-controlled rate-and-state friction model cannot fully explain observed depths. While warm subduction zones can be reasonably explained as a rate-and-state stability transition, the overriding plate Moho in cold subduction zones is too shallow, challenging the combined thermal-lithological model. To test the frictional-viscous model, we analyze power law creep and low-temperature plasticity for quartz, feldspar, olivine, antigorite, and talc. We find that power law creep in any tested mineral is overly temperature sensitive. In contrast, wet olivine, antigorite, and talc low-temperature plasticity fits observed depth limits to a ~6 km misfit. However, only talc is consistent with the weak megathrust paradigm of effective friction coefficients <0.1 and shear strengths of tens of MPa. We conclude that a frictional-viscous transition with a weak and temperature-insensitive viscous mechanism, such as talc low-temperature plasticity, is most consistent with the downdip seismicity limit and constraints on megathrust strength.