The 2015 Gorkha (Nepal) earthquake sequence: I. Source modeling and deterministic 3D ground shaking

Geometric Controls on Pulse‐Like Rupture in a Dynamic Model of the 2015 Gorkha Earthquake

“…These findings suggest that the extent of the Gorkha earthquake rupture was limited by the geometry of the MHT fault surface, although that geometrical effect in our model was realized indirectly, by adopting a frictional parameterization compatible with fault topography. The along‐dip and along‐strike extents of the simulated rupture, and its oval‐shaped coseismic slip distribution, in addition to being consistent with most published source models (e.g., Avouac et al, ; Galetzka et al, ; Wei et al, ), also closely correlate with the model of a flat decollement bounded by two ramps along the dip and two pinch points along the strike. The result is that only a small portion of the MHT ruptured during this event, unlike the nearby, devastating 1934 M w 8.4 Nepal‐Hihar earthquake that occurred nearby to the east and that ruptured all the way up to the ground surface.…”

“…The peak slip of 8 m is located on the decollement segment, to the south of the asperity. This peak slip is slightly larger than most observationally inferred peak coseismic slips (5–7 m; Avouac et al, ; Galetzka et al, ; Grandin et al, ; Hayes et al, ; Lay et al, ; Wang & Fialko, ; Wei et al, ; Yagi & Okuwaki, ; Yue et al, ), because our velocity‐weakening zone (motivated by the assumed shape and spatial extent of the decollement segment) restricted the rupture area in the simulation, requiring higher slip to conform with seismic moment estimates. This difference may also arise in part from smoothing in the finite‐fault inversion.…”

“…Numerous kinematic studies have inferred a pulse‐like rupture for the Gorkha earthquake (e.g., Galetzka et al, ; Yue et al, ), as also shown in the Gorkha dynamic rupture model. The kinematically inverted source models (e.g., Galetzka et al, ; Wei et al, ; Yue et al, ) vary somewhat in their estimates of the slip‐pulse duration but typically give estimates near 6 s, close to the slip‐weighted simulation mean (Figure ). In simple dynamic models where neither heterogeneities (in, e.g., stress state and near‐fault rock stiffness) nor dynamic weakening effects strongly affect rupture duration, pulse duration is controlled principally by the ratio of fault width to rupture velocity, as suggested by dimensional arguments and confirmed by numerical simulations (Day, ).…”

“…The other frictional parameters are variable and distributed as shown in Figures c–f. We introduce a velocity‐weakening portion of the plate interface consisting of the decollement segment, a narrow strip along the top of the downdip ramp down to 14.5‐km depth, and a more extensive part of the downdip ramp corresponding to an asperity that appears in many source imaging studies (Avouac et al, ; Grandin et al, ; Hayes et al, ; McNamara et al, ; Qiu et al, ; Wei et al, ; Yue et al, ). Apart from that asperity, the lower velocity‐weakening boundary follows the 14.5‐km depth contour (the white contour band in Figure c) and does not coincide exactly with the kink joining the decollement and downdip ramp (dashed red curve in each of Figures c–f).…”

Constraining Frictional Properties on Fault by Dynamic Rupture Simulations and Near‐Field Observations

“…The rupture starts at a hypocentral depth of ~15 km and propagates eastwards along the lower edge of the locked portion of the Main Himalayan Thrust (Avouac et al, ), where the Indian and the Eurasian plates collide at a rate of ~18 mm/yr (Lavé & Avouac, ). Moment tensor solutions show that this earthquake was a nearly pure double‐couple reverse faulting event, with the fault plane estimated to have a strike of 293°, a dip angle of 7–10°, and a rake of 95–100° (Avouac et al, ; Galetzka et al, ; Wei et al, ). The final slip distribution and rupture speed obtained from kinematic inversions and backprojection methods consistently show that this earthquake had a relative simple rupture pattern with an average speed of 2.8–3.2 km/s (Fan & Shearer, ; Grandin et al, ; Lay et al, ; Liu et al, ; Wei et al, ; Yagi & Okuwaki, ; Yin et al, ; Yue et al, ) (Figure a and Table ).…”