Thermal control on the updip and downdip extents of megathrust earthquake rupture: revisiting Dr. Roy Hyndman's seminal contributions 30 years later

How far earthquake ruptures extend updip and downdip along subduction megathrusts has important science and hazard implications. In the Hyndman hypothesis proposed in the 1990s, the updip limit was set by a change in clay mineralogy at 100–150 °C, and the downdip limit by a temperature of ∼350 °C or the serpentinized mantle wedge corner (MWC), whichever was shallower. These limits were based on reported changes in fault friction behaviour with temperature and petrology. Many recent megathrust ruptures extended much beyond these limits, but the core idea of the hypothesis is still of great scientific value. Here, I highlight advances in earthquake science over the past three decades relevant to this subject and explain how the Hyndman hypothesis should be accordingly revised. The key point is that the clayey shallow segment of the megathrust and the MWC segment that is rich in lizardite serpentinite can act as “soft barriers” to seismic slip instead of hard limits. They impede seismic slip but may still participate in the slip while undergoing rate-strengthening, or they may even facilitate the slip by exhibiting dynamic weakening at high enough slip rates. In most subduction zones, antigorite serpentinites exhibit seismic behaviour to allow rupture to occur 10–20 km deeper than the MWC while radiating high-frequency seismic energy. In very warm subduction zones, seismic rupture is limited to be shallower than the MWC by thermally activated creep, and the ∼350 °C limit approximately holds. In this situation, the thermal and petrologic conditions at the MWC harbour Episodic Tremor and Slip.