For the source characterization of Mexican subduction earthquakes we used records from the Guerrero Accelerograph Array for earthquakes occurring during the last 15 years with seismic moments ranging from 6.37E+23 to 1.15E+28 dyne cm. After removing the path and site effects, we obtained the flat level from the source acceleration spectra versus seismic moment and compared it with the relationship proposed by Dan et al. (2001). A new alternative method for characterizing size and stress drop of asperities in source modeling is to use the relationship between acceleration source spectral level and seismic moment. We found that the relationship proposed by Dan et al. works appropriately for predicting the flat level of the acceleration spectra for the Mexican subduction earthquakes.
To contrast these results we analyzed the aftershocks of the Tokachi-oki, Japan, earthquake of 26 September 2003 and aftershocks. After removing the path and site effects, the flat level of the acceleration spectra in high frequencies clearly and systematically is underestimated by Dan’s relation. Even Dan’s relation has been obtained based on data from different subduction zones in the world; our results show that it is not always applicable.
We also found that the fault areas estimated from Brune’s formula (Brune, 1970) and the corner frequencies obtained from data of this regional array gave a good estimation of the size of the biggest asperity. These sizes are well predicted by the relationship for subduction earthquakes from the source inversion results given by Somerville et al. (2002).
The relationships presented in this work are useful for the source characterization of Mexican subduction earthquakes for the prediction of strong ground motions. That is made from the relationship between acceleration source spectral level and seismic moment. The asperities with area and stress drop (effective stress) are characterized from the acceleration source spectral levels, total source area, and seismic moment.