资源环境科技发展态势分析平台

The efficiency of thermal pressurization as a dynamic weakening mechanism relies on the thermal and hydraulic properties of the rocks forming the fault core. Here we assess the effectiveness of thermal pressurization by comparing predictions of temperature rise to field estimates based on pseudotachylyte-bearing rocks. We measure hydraulic and transport properties of a suite of fault rocks (a healed cataclasite, an unhealed breccia, and the intact parent rock) from the pseudotachylyte-bearing Gole Larghe fault in the Adamello batholith (Italy) and use them as inputs in numerical simulations of thermal pressurization. We find that the melting temperature can be reached only if damaged, unhealed rock properties are used. A tenfold increase in permeability or a fourfold increase in pore compressibility of the intact rock is required to achieve melting. Our results emphasize the importance of damage processes that strongly modify fault rock properties and dynamic weakening processes during earthquake propagation.

Plain Language Summary During earthquakes, faults slide rapidly past each other, typically at several meters per second. Such fast sliding rates at great depth in the crust are expected to generate large amounts of heat and should melt the rocks at the sliding interface. However, such melted rocks are not systematically observed in faults. A number of mechanisms have been suggested to explain this apparent discrepancy. One convincing explanation is that the presence of water within the porosity of the fault rocks buffers the fault temperature by being rapidly pressurized and reducing the fault friction. This effect has been predicted from physical models, but these predictions depend strongly on poorly constrained rock properties and have not yet been tested in nature. Here we measured key physical properties in rocks adjacent to a fault that underwent frictional melting, and we test whether model predictions for the effects of pressurized water are consistent with the presence of melt. We find that it is the case only if rock properties are altered to account for the effect of microfractures that are likely created during earthquake rupture. Our results provide constrains to improve earthquake simulations and highlight the key role of microfracture damage on fault sliding processes.