A probabilistic seismic hazard assessment of the Trans-Mexican Volcanic Belt, Mexico based on historical and instrumentally recorded seismicity

The Trans-Mexican Volcanic Belt (TMVB) is an active volcanic chain being deformed by an intra-arc extensional fault network. Although several crustal earthquakes with magnitude>7 have originated in the TMVB since the 16th century, the background seismicity of this geological structure is very low and the region is usually considered of low seismic hazard. In this study, we present an updated probabilistic seismic hazard model of the TMVB. The seismicity catalog used here includes forty-three historically and instrumentally recorded earthquakes, from 1858 to 2014; five of these are large earthquakes that occurred in the TMVB during the XIXth century. Due to the lack of a statically representative sample, we propose, in a qualitative manner, the seismicity catalog is complete for M >= 4 since 1964 and for M >= 6 since 1858. Moreover, we introduce three different earthquake frequency-magnitude relations. The first one is a. conventional Gutenberg Richter fit of the distribution of the instrumentally recorded earthquakes data. The other two are non-conventional, semi-parametric approaches that integrate the historical and the instrumental data to determine seismicity rates in the region. Our preferred model (seismicity model B) fits separately the instrumental and the historical data and merge the two fits into one curve. A uniform seismic hazard (USH) of the TMVB for a return period of 500 years was calculated considering three major sources of earthquakes: 1) Subduction thrust-faulting events in the Middle American Trench (MAT); 2) Earthquakes within the subducted Cocos plate and, 3) Shallow crustal earthquakes in the TMVB. According to the seismicity model B, the average recurrence time of a M >= 7 earthquake on the TMVB is approximately 150 years. In contrast, the recurrence time estimated from the instrumental catalog is 12,000 years. The results of this seismicity model, which is based on historical and instrumental data, agrees also with the return periods of prehistoric earthquakes, estimated for short segments of the fault system in the TMVB in paleoseismological studies. When comparing the results of our preferred seismicity model, the PGA estimated using only the instrumental seismicity are 18 to 56% smaller than those predicted by the model using the historical catalog.