In May and December 1994, two medium-size, intermediate-depth-focus earthquakes occurred in Guerrero, Mexico, eastward of the rupture area of the great Michoacan earthquake of September 19, 1985. Even though these are not major earthquakes (~6.4 Mw), they were widely felt through central and southern Mexico, with minor damage at Zihuatanejo and Acapulco, located along the Pacific coast, and Mexico City.
Both earthquakes, separated by ~100 km, have similar focal depths and magnitudes, however, their focal mechanisms, based upon the polarities of first arrivals, show some differences. The May earthquake shows a clear normal faulting mechanism ( = 307o, δ = 55o, = -108o), whereas the December earthquake mechanism solution suggests an initial thrust faulting ( = 313o, δ = 62o, = 98o) process. Although previous analysis, including local and teleseismic stations, reported a normal faulting for the December earthquake, we find that modeling using the CMT focal mechanism solution fails to reproduce the first 5 sec of the observed P-wave signal at the nearest broadband station ( = 168 km) and the S-wave polarity at two strong ground-motion local stations ( = 32, 53 km); in fact, the best fit for these stations is obtained using the thrust focal mechanism calculated from the first-motion method.
Seismic moment value and rupture duration time deduced from the teleseismic spectral analysis are: 2.0 x 1018 N-m and 6.9 sec for the May event; 2.8 x 1018 N-m and 7.1 sec for the December earthquake. From the inferred seismic moment, an average of ~15 bars for both earthquakes is obtained.
Inversion of teleseismic P-wave data indicates a better fit using the CMT focal mechanism solution (normal faulting) than the first-motion mechanism for both earthquakes, although the adjustment's differences are small for the May event; for this earthquake, the rupture consisted of two sources separated by ~7 sec, starting at a depth of ~40 km and then propagating downdip, reaching a depth of ~60 km. The December earthquake however, released, all its energy at a depth of 50 km in two main sources separated by ~10 sec.
The non-double-couple components values are -0.004 and -0.01 for the May and December events, respectively, indicating that the December shock has a small contribution of non-double-couple radiation that could be the result of a changing mechanism. This result agrees with the hypothesis that a slab subducting at a shallower angle (our case) is associated with the existence of random subfaults with different fault orientations.
From a tectonic point of view, the complexity of the December earthquake could be the result of the observed complexity of the stress distribution around 101o W and the existence of compressional events beneath the normal faulting earthquakes near the coastline. This feature permits the flexural stresses associated to the slab bending upward to become subhorizontal at the Guerrero region. We conclude that the May earthquake corresponds to a pure normal faulting, whereas the December shock is a complex event with a variable fault geometry.