Studies of Instrumental and Historical Seismicity
Fourier amplitudes and response spectra of more than 3000 waveforms from ∼300 Puerto Rico earthquakes of magnitude 3–5.5 have been analyzed. Due to a paucity of data at small distances and large magnitudes, the ground-motion data cannot be used directly to obtain ground-motion relations for magnitudes and distances of most engineering interest. Instead, the data were used to determine key attenuation parameters, such as regional anelastic attenuation, duration behavior and generic site amplification that are input to seismological ground-motion models. The data were also used to validate ground-motion model predictions.
To overcome the incompleteness of the data set, we applied the stochastic method to simulate waveforms for different magnitudes and distances. The stochastic method has been applied for other regions such as California, Cascadia, and eastern North America, and on average reproduces empirical attenuation relationships that can be obtained by direct use of enough data. The input parameters for the simulations are based on the attenuation parameters obtained from the recorded waveforms.
We simulated 1950 acceleration time series for magnitudes from M3.0 to M8.0 and distances from 2 km to 500 km. Simulation was performed in magnitude steps of 0.2 units. In order to provide a good database, we simulated data for both backward and forward directivity cases, as well as data for azimuths with minimal directivity effects. The response spectra of the simulated time series have been calculated. The maximum likelihood method has been applied in order to derive ground-motion relations for a generic soft rock site condition for frequencies from 0.1 to 20Hz. The stochastic-model ground-motion relations for Puerto Rico are validated using available seismographic data, and compared to ground-motion relations for other regions. These are the first region-specific ground-motion relations developed for seismic hazard analysis of Puerto Rico.