Utilizing a 3D Global P-Wave Tomography Model to Improve Backprojection Imaging: A Case Study of the 2015 Nepal Earthquake

We carry out a backprojection analysis of high-frequency P-wave signals to image the rupture process of the 8 September 2017 Mw 8.2 Chiapas earthquake based on the 3D velocity model LLNL-G3Dv3. Results show that the entire ∼56 s unilateral rupture process clearly involves two stages. The first stage lasted for ∼37 s, with a northwest (∼322° azi- muth) propagation over a length of ∼89 km at a speed of ∼2:8 km=s. Then, the rupture made a right turn within a short distance and propagated to ∼348° azimuth for ∼74 km at a higher speed of ∼4:6 km=s. The two-stage rupture was further confirmed by the power release, data coherency, beamforming patterns, and aftershock distribution. The overall rupture can be well explained by a two-fault model. Rupture on the second fault was positively triggered by that on the first one. The Tehuantepec Fracture Zone, which transversely intersects the Middle American trench as well as the coseismic rupture zone, likely functioned as a barrier to the first stage. Upon reaching the barrier, the rupture was temporarily blocked and deceler- ated, but eventually detoured to cross the barrier with the aid of the accumulated stress. The abrupt release of stress led to a much higher rupture speed in the second stage.