SCEC Award Number 18146 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title SDSU BBP Module Extension: Temporal-Spatial Ground Motion Correlation
Investigator(s)
Name Organization
Kim Olsen San Diego State University
Other Participants Rumi Takedatsu, Research Assistant, SDSU
SCEC Priorities 4c, 4b, 5b SCEC Groups GM, CME, EEII
Report Due Date 03/15/2019 Date Report Submitted 05/02/2019
Project Abstract
 Ground motion simulations can be viable alternatives to empirical relations for seismic hazard analysis when data are sparse. However, in many cases, simulated ground motion time series, in particular those originating from stochastic methods, lack inter-frequency correlation revealed in recorded seismic data, which has implications for seismic risk. We develop a post-processing method to rectify simulation techniques that otherwise produce synthetic time histories deficient in inter-frequency correlation structure. We show that the addition of correlation to the synthetic time histories results in insignificant changes in the bias. The method generates correlated synthetic time series that are very similar to the original results from current SDSU BBP Module. Finally, we show that our method also works well to incorporate realistic inter-frequency correlation into PSA.
Intellectual Merit Ground motion simulations can be viable alternatives to empirical relations for seismic hazard analysis when data are sparse. However, in many cases, simulated ground motion time series, in particular those originating from stochastic methods, lack inter-frequency correlation revealed in recorded seismic data, which has implications for seismic risk. We develop a post-processing method to rectify simulation techniques that otherwise produce synthetic time histories deficient in inter-frequency correlation structure.
Broader Impacts The use of the refined ground motion generator with inter-frequency correlation will improve the loss estimation based on the synthetic ground motions.
Exemplary Figure Figure 1. The inter-frequency correlation coefficients of epsilon at reference frequencies 0.2 Hz, 0.5 Hz, 1 Hz, 2 Hz, 5 Hz and 10 Hz from the empirical correlation coefficients (dashed lines) and the SDSU SCEC BBP Module after applying our method. (a) Using independent random variables at two horizontal components (solid lines), and (b) using correlated random variables (with correlation coefficient equals 0.7) at two horizontal components (solid lines) for the Loma Prieta event.

Credit: Nan Wang.