Earthquake induced residual stresses preserved in fault rocks exhumed from the lower crust
- 1University of Oslo, The Njord Centre, Department of Geosciences, Oslo, Norway (luca.menegon@geo.uio.no)
- 2Department of Geosciences, University of Padua, Padua, Italy
- 3Department of Earth Sciences, University of Cambridge, Cambridge, UK
Field studies established that seismicity in the lower crust is linked to brittle failure of dry, strong rocks. Failure of these strong rocks implies build-up of differential stresses to gigapascal (GPa) levels, but this requirement contrasts with current models of continental lithosphere deformation, which favour distributed flow of weak viscous lower crust. Although several mechanisms have been proposed to generate transiently high stresses, direct measurements are lacking. Recent advancements in microanalytical techniques (i.e., high-angular resolution electron backscatter diffraction, HR-EBSD) have proven successful at measuring the residual stress resulting from elastic strain retained in mineral grains.
We investigated with HR-EBSD the residual stresses in garnet and diopside from exhumed faults containing pseudotachylytes (quenched frictional melts produced during seismic slip). The samples come from Lofoten (Norway), Holsnøy (Bergen Arcs, Norway), and Musgrave Ranges (Central Australia). Pseudotachylytes from all three localities represent single earthquake events and formed at lower-crustal conditions (T = 500–720 °C, P = 0.5–1.0 GPa). Pseudotachylytes from Holsnøy show an asymmetric damage distribution, where host-rock garnet is pulverized nearby the fault on the side subjected to predominantly tensional stresses during rupture propagation, while garnet is intact on the other side. This asymmetry provides an opportunity to compare the residual stresses on both sides of the fault.
All samples preserve intragrain residual stress heterogeneities reaching 100s of MPa to GPa levels due to local high density of unrecovered lattice defects (dislocations). However, the timing of formation of lattice defects with respect to the seismic event differs. In samples from the Musgrave Ranges, the absence of any later deformation along with the sluggish mobility of dislocations in garnet at the ambient deformation conditions (500 °C, 0.5 GPa) allowed preservation of the high dislocation density produced during the earthquake rupture propagation, recording stress heterogeneities of as much as 6 GPa. In Holsnøy, residual stress heterogeneities of up to 1 GPa are only measured in pulverized grains and are also associated with unrelaxed dislocations generated during the earthquake rupture propagation. Intact garnet grains from the less damaged side of the fault show a limited range of intragrain stress heterogeneities, generally within 100 MPa, and a low density of dislocations. Residual stresses in diopside from Lofoten are only elevated (600 MPa) within 200 µm of the pseudotachylyte. Diopside recorded the progressive build-up of stress during interseismic loading, as suggested by the presence of coseismic cracks crosscutting lattice undulations that preserve the greatest stress heterogeneities. However, the ability of diopside to build up stress is limited, as stress is efficiently dissipated by the development of deformation twins.
In conclusion, great stress heterogeneities can be preserved in mineral grains that experienced the earthquake cycle in the lower crust. Different mineral phases can preserve stress heterogeneities to different extents, depending on the mobility of dislocations after their formation and on other relaxation mechanisms (e.g., twinning). Information on residual stress have important implications for the energy budget of an earthquake, the earthquake cycle deformation, and crustal rheology.
How to cite: Menegon, L., Toffol, G., van Schrojenstein Lantman, H. W., Wallis, D., Pennacchioni, G., and Jamtveit, B.: Earthquake induced residual stresses preserved in fault rocks exhumed from the lower crust, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9228, https://doi.org/10.5194/egusphere-egu24-9228, 2024.