Abstract

Pseudotachylytes (frictional melts formed during seismic slip) in the metamorphosed anorthosites from Nusfjord (Lofoten, northern Norway) preserve a record of seismic rupture in the dry lower crust at 650–750 °C, 0.8 GPa. Field observations indicate that the Nusfjord pseudotachylytes represent single-earthquake events associated with large stress drops, on the order of hundreds of megapascal (MPa) to 1–2 gigapascal. Such large stress drops are interpreted to reflect the high strength of the intact anorthosite at the high confinement conditions of the lower crust. One important question is whether evidence of the high stresses necessary to initiate seismic rupture in the lower crust is preserved in the microstructure of the Nusfjord pseudotachylytes and of their damage zone. Pyroxene deformation microstructures associated with preseismic loading and coseismic fragmentation reveal strongly localized transient stresses that presumably reached GPa-level magnitude. Here we use high-angular resolution electron backscatter diffraction (HR-EBSD) on diopside grains to obtain spatial datasets of residual stresses that are retained in the crystal lattice of diopside. We apply this method combined with microstructural analysis on diopside in a sample from a pseudotachylyte from Nusfjord to reconstruct the spatial heterogeneities of stress and link them to the earthquake cycle and associated coseismic thermal effects. Diopside contains micro- to nanoscale deformation twins within 3 mm from the fault and in clasts in the pseudotachylyte. Strong lattice undulations are locally present in survivor clasts, indicating low-T plasticity at high stress. Residual stresses from the wall rock and in a survivor clast vary between ~600 and ~200 MPa and form a gradient of decreasing residual stress away from the pseudotachylyte, only elevated within 200 µm from the pseudotachylyte margin and with the highest values occurring within the clast. Microfaults crosscut the deformation twins, lattice undulations, and residual stress spatial heterogeneities within the clast. The latter appear strongly similar to the lattice undulations, in distribution and orientation. The obtained stresses are lower than estimated stress drops for the locality and than stresses expected during rupture propagation (both >1 GPa). As alternative stress source, we investigated thermal stress introduced by coseismic frictional heating. Calculations demonstrate that this process is only significant over a distance of less than 100 µm in the wall rock for a stress drop of 100 MPa, and less than 10 µm for a stress drop of 1 GPa. Instead, because coseismic microfaults crosscut twinning, lattice undulations, and the spatial heterogeneities of residual stress, we interpret that these features correspond to the progressive build-up of stress during preseismic loading. An explanation for the discrepancy between the residual stresses and suggested stress drop is that the stress build-up in diopside was partially dissipated by the formation of twins. Additionally, the stress drop is estimated at the scale of the bulk fault, whereas the residual stresses are measured at the single grain scale and as such are likely to vary locally depending on the microstructure and on the different ability of different phases to dissipate the stress build-up via e.g. twinning and recovery of dislocations.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call