We simultaneously invert the strong-motion velocity records and the long- and intermediate-period teleseismic P waveforms of the 19 September 1985 Michoacan, Mexico, earthquake to recover the distribution of slip on the fault using a point-by-point constrained and stabilized, least-squares inversion method. A fault plane with strike fixed at 300° and dip fixed at 14° is placed in the region of the earthquake hypocenter and divided into 120 subfaults. Rupture is assumed to propagate at a velocity of 2.6 km/sec away from the hypocenter. Synthetic near-source ground motions and teleseismic P waveforms for pure strike-slip and dipslip dislocations are calculated for each subfault. The observed data are then inverted to obtain the amount of strike-slip and dip-slip displacement required of each subfault. We also invert the data sets using a time-window procedure where the subfaults are allowed to slip up to three times. This approach relaxes the constraint of fixed subfault rupture time imposed by a constant rupture velocity.
Inversion of the strong-motion data alone yields a slip model similar to the solution previously obtained using only teleseismic waveforms. This result supports the use of teleseismic waveform data for the derivation of fault dislocation models in the absence of strong-motion recordings. Our simultaneous inversion of both data sets suggests that rupture during the Michoacan earthquake was controlled largely by the failure of three major asperities located along the length and down the dip of a 150-km segment of the Cocos-North America plate boundary. The solution contains three major source regions including an 80 km by 55 km source near the hypocenter with a peak slip of 6.5 meters. Two additional sources are present on the southeast portion of the fault about 70 km away from the hypocenter. One of these sources, with a peak slip of 5 meters, covers a 45 km by 60 km area and extends downdip from a depth of about 10 km to 24 km. The third source region is somewhat smaller (30 km by 60 km area, 3.1-meters peak slip) and extends further downdip at depths between 27 km and 39 km. Aftershock activity following the earthquake was associated mainly with the two shallow sources. These two sources are separated by the aftershock zone of the 1981 Playa Azul earthquake.