Search

An inslab earthquake occurred below Atoyac, Guerrero, Mexico, on 13 April 2007 ( Mw 5.9) at a depth of 37 km. It was strongly felt in Acapulco (Δ ∼65 km) and Mexico City (Δ ∼270 km); the peak ground accelerations (PGAs) at soft sites in these two cities were as high as 200 cm/s2 and 12 cm/s2, respectively. This was the first significant earthquake in Mexico since the recent strengthening of strong-motion and broadband networks in the country. This was also the first moderate event for which ground-motion maps for Mexico City were produced and distributed without human intervention in near real time. In this paper, we report on the source characteristics and the tectonic implications of the earthquake and the ground motions that it produced. We also discuss the performance of the seismic alert system (SAS) and near real-time generation and distribution of expected ground-motion maps for Mexico City during the earthquake. Our analysis of the Atoyac earthquake is based on broadband seismograms and accelerograms from stations operated by the National Seismological Service (SSN) of the Instituto de Geofisica (IG) at the Universidad Nacional Autonoma de Mexico (UNAM), and the Instituto de Ingenieria (II), UNAM. We also used data from an accelerograph operated by the Centro Nacional de Prevencion de Desastres (CENAPRED) in Acapulco. About 35 aftershocks with M ≥ 3.4 were reported by SSN in the first four days of the earthquake, the largest being one of magnitude 5.3. The SSN network is sparse. However, the accelerograph network operated by the II is relatively dense in Guerrero (Anderson et al. 1994) where a mature seismic gap has been identified (Singh et al. 1981). Merged data from both networks permit reliable location of earthquakes in the region. As the accelerograms from the II network are not …

SUMMARY We compare ground motions recorded at the surface and in boreholes at five different locations of the lake-bed zone of the Valley of Mexico with theoretical dispersion curves and eigenfunctions calculated for the first two modes of Rayleigh and Love waves. We find that (1) the maximum in the horizontal-to-vertical displacement ratio, which occurs at the dominant frequency of the site (0.4 Hz), corresponds to the higher mode rather than to the fundamental mode of the Rayleigh waves, (2) borehole records at depths from 0 to 100 m show that the normalized vertical displacement does not decrease rapidly below the superficial clay layer, as should be the case for the fundamental mode, but remains #0.8, and (3) the measured phase velocity at a period of about 2.5 s (2.0t0.5 km s x1 ) is too fast for the fundamental mode predicted for the known crustal velocity structure. These observations lead us to conclude that the wavefield in the lakebed zone in Mexico City is dominated by higher-mode surface waves. This provides a plausible explanation for the long duration of the coda in the lake-bed zone. Although shear wave Q is very small (10‐20) in the clay layer, the higher modes of surface waves do not propagate in the superficial clay layer but in the underlying structure where Q-values are likely to be relatively high. Thus, while the clay layer plays the passive role of amplifying the ground motion, its contribution in damping out the motion is insignificant. The results have two important practical implications. (1) The strain estimate from recorded ground velocity differs significantly for the fundamental mode as compared to the higher-mode surface waves. (2) If the ground motion is dominated by the fundamental mode, then knowledge of the superficial layer and the velocity contrast with the underlying structure is sufficient for understanding and modelling of the ground motion. If, however, the higher-mode surface waves dominate, then a detailed knowledge of the deeper structure is required.

To understand the cause of observed amplification of seismic waves even at hill-zone sites in the Valley of Mexico, digital accelerographs have been installed at three especially chosen sites. Two of these sites, MADI and TEXC, located on hard Pleistocene lava (a few meters in thickness) overlying Oligocene andesites, were expected to be free of site effects. Analysis of the data recorded by these and other accelerographs during three moderate, shallow subduction zone events, however, shows significant amplification at MADI and TEXC between 0.2 and 0.6 Hz with respect to established attenuation relations. The cause of the amplification at hill-zone sites in the Valley of Mexico, including MADI and TEXC, may be pervasive low S-wave velocity in, and complex structure of, the upper layers of the volcanic rocks. If so, then there may not be a truly “hard” rock site in the valley.

Aceelerograms recorded on the Guerrero, Mexico, strong motion accelerograph network illustrate the dependence of strong ground motion on the magnitude and the hypocenter distance. These data suppport the hipothesis that complexity in accelerograms at short distances arises from complexity at the source. The duration of strong shaking is controlled by the source dimension at short distances, and extended by wave porpagation effects as distance increases. Peak amplitudes (peak acceleration, velocity) saturate at different magnitudes at different distances. This change in shape of attenuation curves as magnitude increases can be explained by the trnasition of Green's functions from simple, short pulses at short distances to longer duration wave trains at large distances. Spectral amplitudes demonstrate scalin relations in which loe frequency amplitudes are proportional to seismic moment but high frecuencies increase much less rapidly. The beginnings of large earthquakes look like small earthquakes, consistent with a cascade model for the growth of large events. The mostrecent large earthquake, on sept. 14,995 (Mw=7.4) shows surprisingly weak peak accelerations.