The crustal viscosity gradient measured from post-seismic deformation: A case study of the 1997 Manyi (Tibet) earthquake

We here analyse an InSAR dataset which constrains the temporal and spatial distribution of post-seismic ground displacement following the 1997 Manyi (Tibet) earthquake to estimate the depth-variation of creep viscosity within the crust. We use a method which depends on the detection and measurement of a systematic variation in post-seismic relaxation rate with distance from the fault. At a given distance the displacement rates may be compared with the prediction of a uniform viscosity model with apparent viscosity ηu. The rate at which ηu decreases with distance from the fault is directly diagnostic of the parameter that controls the vertical gradient of viscosity in a stratified crust. Our best-fit viscosity profile for Manyi exponentially decreases from 1.2×1021 Pa s at the top of the visco-elastic layer (at 10 km) to 1.6×1018 Pa s at the Moho (at 60 km). The predicted temporal and spatial variations of post-seismic displacements generally fit the observed post day-165 displacements within error, if we assume that displacements pre day-165 are attributed to after-slip on the fault. Within error, the crustal viscosity at 10 km depth may range between ∼6.1×1020 Pa s and ∼1.4×1022 Pa s. Estimates of the Maxwell time constant consistent with an inter-seismic period of ∼500–1000 yr imply an effective elastic layer thickness of ∼20 km for best-fit and upper bound viscosity solutions. We also show that the estimated viscosity profile is broadly consistent with laboratory measurements of creep for crustal materials like quartzite and granite by deducing from our viscosity profile the ratio (Q/n) of activation energy to stress exponent at ∼45–81 kJ/mole, thus providing a new link between laboratory-scale measurements of creep deformation and viscosity measured from post-seismic relaxation.