Research areas

A brief overview of several research topics is provided here. Click on the navigation tabs directly below this text to see more detailed information on selected topics.

 
    Research Topics: General | Epsilon | Pulse Classification | Attenuation scripts |

Attenuation Scripts

by Jack Baker, Nirmal Jayaram and Ting Lin, April 2008

Some of the research presented on this web site relies on ground motion prediction ("attenuation") models to estimate the probability distributions of observed ground motion intensity associated with a given earthquake magnitude, source-to-site distance, etc. As a courtesy, Matlab scripts are provided here for a variety of popular models published in the scientific literature. You are welcome to use these scripts as long as you acknowledge this web site as their source (and of course cite the original model publications).

The following models for spectral acceleration (Sa) were published in Volume 68, Issue 1 of Seismological Research Letters, in 1997.

The following models were developed as part of the PEER Next Generation Attenuation project. Reference information will be updated once final publications are produced. These models were all coded by Nirmal Jayaram, a PhD student at Stanford University.

Correlations between spectral acceleration values are also required for some of the calculations in the research presented here. The following script provides a calculation of predicted correlations between spectral acceleration values at two periods, from a single ground motion. The equation was developed by Jack Baker and Nirmal Jayaram, and a paper documenting its development is in review.

Note: there is a potential for ambiguity in the definition of "horizontal spectral acceleration," and the precise definition has some impact on the predicted probability distribution. More explanation is provided here and here. The 1997 models generally provided the distribution of the geometric mean of spectral acceleration from two horizontal ground motion components. The new NGA models will be more explicit as to which definition they use.