Rapid Geodetic Analysis of Subduction Zone Earthquakes Leveraging a 3‐D Elastic Green's Function Library

The 2018 M7.2 Pinotepa earthquake ruptured a shallow slab section along the Middle America subduction zone. We demonstrate how a geodetic Green's function (GF) library and efficient modeling algorithm can rapidly resolve earthquake slip and contribute to early warning systems. The source is characterized with InSAR data and a finite‐element model mimicking realistic slab geometry (Slab1.0) and velocity structures (CRUST2.0) that are not considered in the conventional homogeneous (HOM)‐ or layered(1‐D)‐crust solutions. The rupture is imaged 18 min after geodetic data are downloaded to a 16‐CPU‐thread workstation. Nearby Mw8.6‐megathrust scenarios are also studied with synthetic GPS data in the Guerrero/Oaxaca area. Our results show that earthquake slip solutions are sensitive to the materials of elastic‐modeling domains. Attaining smaller misfits or sums of squared errors, models by 3‐D GFs significantly better recover coseismic displacements than those estimated with 1‐D GFs or HOM GFs at 95% confidence. Plain Language Summary A Mw7.2 earthquake took place near the city of Pinotepa Nacinoal in Oaxaca, Mexico, on 16 February 2018 and became the third most significant event striking southern Mexico since mid‐2017, after the Mw8.2 Tehuantepec and Mw7.1 Puebla‐Morelos earthquakes. A new algorithm of deformation modeling is designed to use a fault library to efficiently characterize the earthquake source. This library accounts for fault slip within both rock and structural complexities documented within the Middle America subduction zone. We further study 100 Mw8.6 megathrust scenarios and discover that an elastic model of these complexities is necessary to better recover the slip and surface displacements than the customary layered or uniform‐crust solutions. These Mw8.6 scenarios are the largest expected events based on historical records. Our algorithm images the Pinotepa earthquake rupture 18 min after the ground displacement data are downlinked from the satellite and processed in local machines, providing an alternative of evaluating the earthquake mechanism in addition to other seismological methods. The majority of fault slip (up to 1.2 m) is located over a gently dipping slab segment at a depth of ~21 km. These results illuminate the effectiveness and applicability of the proposed algorithm for real‐time analyses, without compromising model accuracy to analytical solutions.