How can geological datasets help us to choose between rheological behaviours suggested by experimental measurements? Examples from the Alpine Fault and the Japan Trench.

Abstract:

Laboratory measurements of fault rock rheology are commonly performed on individual components of complex natural systems. Textural and structural inferences from outcrop observations provide one means to constrain the co-operative behaviour of multi-component natural fault systems. For example, the principal slip zone (PSZ) of New Zealand’s central Alpine Fault comprises a 1-10 cm thick sandwich of impermeable smectite-bearing ultracataclasite/gouge layers between higher permeability hanging wall cataclasite and footwall gravel (Boulton et al., 2012, doi: 10.1029/2011GC003872). Based solely on measured mechanical properties we expect earthquake ruptures to nucleate and propagate in the cataclasites rather than the PSZ. However, the PSZ gouge was preferentially comminuted so it must localise slip, and injection veins penetrate from it into the surrounding formation. This suggests that the gouge experiences coseismic pressurization and weakening, possible if slip is confined to one layer within the impermeable (thus undrained) gouge. On the southern Alpine Fault a clear PSZ is not well-developed; instead a wide shear zone crops out (Barth et al., 2013, doi: 10.1002/tect.20041). Mechanical data again demonstrate frictionally weak, velocity strengthening, low permeability materials, compatible with a creeping shear zone, but PSZ materials display velocity weakening behaviour at high slip rates if undrained, from which we infer seismic slip is possible in nature. Extensive paleoseismic records suggest the structure has accommodated regularly repeating earthquakes for the last 17 kyr (Berryman et al., 2013, doi: 10.1126/science.1218959). Our newly gathered geodetic datasets may resolve this apparent slip rate paradox. In situ measurements from active fault systems can also help interpret experimental data. For example, in material recovered from around the active slip zone of the 2011 Tohoku-oki earthquake, experiments suggest lower frictional strength for undrained than drained shear (Ujiie et al., 2013, doi: 10.1126/1243485). Perturbation of in situ thermal structure is consistent with the experimentally-measured friction in a partially undrained case, from which we infer that thermal pressurization also affects this fault in nature (Fulton et al., 2013, doi: 10.1126/1243641).