EPSS Colloquium fall-2015

The 2015 Mw 7.8 Nepal-Gorkha earthquake with casualties of over 9,000 people is the most devastating disaster to strike Nepal since the 1934 Nepal–Bihar earthquake. Its rupture process is imaged by teleseismic back-projections (BP) of seismograms recorded by three large regional networks in Australia, North America and Europe. The source images of all three arrays reveal a unilateral eastward rupture. But the propagation directions and speeds differ significantly between different arrays. To understand the spatial uncertainties of the BP analysis, we image four moderate-size aftershocks recorded by all three arrays exactly as done for the mainshock. We find that the apparent source locations inferred from BPs are systematically biased from the catalog location, which can be explained by the uncertainty of the seismic velocity derived from the 1D reference model (e.g. PREM). We introduced a slowness error term in travel time that successfully mitigates the source location discrepancies of different arrays. The calibrated BPs are mutually consistent and reveal a unilateral rupture propagating eastward at a speed of 2.7 km/s over ~150km along the down-dip edge of the locked Himalaya thrust zone, predominantly located in a relatively narrow and deep swath. In that sense, the 2015 Nepal earthquake is an incomplete rupture that failed to rupture the entire Himalayan décollement to the surface, which can be regarded as an intermediate event during the interseismic period of larger and complete Himalyan ruptures. The high-frequency ground shaking in the near-field region rarely exceeds the European Macroseismic Intensity (EMS) 8, lower than anticipated by the Ground motion prediction Equations (GMPE) given the magnitude and proximity to the rupture. Such modest level of ground motions can be explained in part by the deep and thus further away high-frequency source radiations from the Kathmandu basin.