Fast Moment Tensor Inversion for Large Earthquakes using the Mexican Accelerographic Network

Abstract:

The moment tensor calculation that is computed immediately after the occurrence of a major earthquake is limited to the number of unsaturated records in stations near the epicenter and the number of stations that transmit their data in real-time. Accelerographic records, however, are not commonly saturated after major earthquakes. Taking advantage of the wide coverage of the Mexican Accelerographic Network, we use accelerograms observed in real time to compute moment tensor solutions after the occurrence of an earthquake. In our study, we compute the double-couple moment tensor inversion as a least squares problem by minimizing the misfit between synthetic waveforms in three components and observed waveforms. Synthetic Receiver Green’s Tensors for each station of the network within the model were previously calculated using a three-dimensional model of South-central Mexico. The database has horizontal spatial resolution of 20 km and a depth spatial resolution of 5 km. Our procedure fits windows containing the P and S waves to compute a fast first-solution. A revised solution is then calculated by fitting the full record. A first solution can be obtained seconds after the P-wave recorded in the station closest to the epicenter. Our results show that it is possible to obtain the moment tensor solution quickly and accurately. Furthermore, we show the resolution and range of uncertainty of the moment tensor solutions compared with that reported by specialized agencies for 30 selected strong earthquakes in Mexico from 2010 to 2014.