Department of Civil Engineering,Sharif University of Technology
Department of Civil Engineering,Carleton University
Using stochastic point source and nite fault modeling, the stochastic stress drop is
estimated for 52 large shallow earthquakes listed in the 'Pacic Earthquake Engineering Research Center
(PEER) Next Generation Attenuation of Ground Motions (NGA)' database. The Pseudo Spectral
Acceleration (PSA) of 541 accelerograms, recorded at National Earthquake Hazards Reduction Program
(NEHRP) C-class sites from 52 earthquakes are simulated and compared with the PSA listed in the PEER
NGA database. The magnitude of the analyzed earthquakes ranged from M4:4 to M7:6. The stress drop
is calibrated by trial and error and based on the analysis of residuals where the residual is dened as
the log of the observed PSA minus the log of the predicted PSA by stochastic methods. The symmetric
distribution of residuals around the zero line is considered as an indicator of good agreement between the
simulated and observed PSAs. The calculated stress drops based on stochastic point source and nite fault
modeling are dierent from the static stress drops that are currently listed in the PEER NGA database.
It seems that there is no clear relation between stress drop and earthquake magnitude, but there is a good
linear relation between estimated stress drops based on stochastic point source and nite fault modeling.
The sensitivity of the estimated stress drop values on the Kappa factor and geometric speeding factor
(b value) is also investigated.