Data collected as part of the SEMS project, funded by the U.S. Minerals Management Service, are available; click here to obtain a zip file containing the data. The data are in smc format (see the BAP documentation from http://nsmp.wr.usgs.gov ; software for working with smc files are available from my online software link. Information about SEMS and the data can be found in Boore, D.M. and C.E. Smith (1999). Analysis of earthquake recordings obtained from the Seafloor Earthquake Measuring System (SEMS) instruments deployed off the coast of southern California, Bull. Seism. Soc. Am. 89, 260--274. As the name of the paper implies, most of these data were obtained from instruments on the sea floor. Other, onshore, data from some of these events are available; see http://db.cosmos-eq.org/.
These data are provided as a service to users. No warranty, expressed or implied, is made by me or the U.S. Geological Survey as to the accuracy of the data. In addition, answers to questions regarding the data should be sought from the before-mentioned paper rather from me.
UPDATE: 07/25/06 -- Weight "sig" added to travel-time files
As discussed in Open-File report 03-191 (A Compendium of P- and S- Velocities from Surface-to-Borehole Logging: Summary and Reanalysis of Previously Published Data and Analysis of Unpublished Data, by David M. Boore, with assistance in some data analysis by James F. Gibbs and Magda Rodriguez), data are available for the borehole logging performed by the USGS, starting in 1975. Before 1992 layered models were not fit to the travel times; only interval velocities were determined. Between 1992 and 1999 layered models were used, but the observed times were corrected to vertical. Starting in 1999, the slant times were used and the fitting program accounted for refraction at layer boundaries. In the results reported on in OFR 03-191 and available from this web site, all pre-1999 borehole measurement were reinterpreted using the modern analysis program. The data have been put into several zip files, as described below.
As corrections and additions are inevitable, I wrestled with how best to indicate when changes have been made. I decided that the only reasonable way to do so was to reconstruct the zip files, using the "update and new" switch so that only changed or new files are added to the zip file. The same switch can be used in unzipping the files (I use PKZip). In this way, the file creation dates can be used to tell if files have been changed. The Fortran programs used to fit models to the travel times and to produce the files included here are available upon request from me; I will eventually put them onto the online_software page of my web site.
The data are available from the following files:
BH_CSV_XLS (approx. 320Kb): Contains files with the basic borehole information, the velocity models, and the travel times used to derive the velocity models in both comma separated text files (*.csv files) for convenience in importing into spreadsheet or database programs and in Excel spreadsheet format (*.xls files). The files in spreadsheet format are particularly useful because they can be sorted in various ways. Separate sets of files are given for the P-waves and the S-waves, and the filenames indicate whether the file contains borehole and measurement information, travel times, or models. The files and their contents are as follows:
BH_VEL_OUT.ZIP (approx. 1.7Mb): Contains text files of the velocity models (*.vel files) and the Poisson's ratios (*.out files) in the format used in recent open-file reports. The files for each open-file report have been put into separate folders rather than being collected together; the relative folder information is carried along in the zip files. This was originally done because in some cases the 3- or 4-letter code used to name a station is the same for stations in different open-file reports. For the convenience of the user, however, duplicated station codes names were renamed and all velocity files were collected into one directory (VELFILES). The zip file therefore contains two copies of each velocity file.
BH_PS.ZIP (approx. 4.4Mb): Contains Postscript plot files showing the fit of the model to the travel times, the residuals to the fit, and the velocity model. In the plots of travel time versus depth the times are the slant times, uncorrected for distance, and the arrival times, calculated only at depths for which measurements were made, are joined by straight lines. The actual travel times would be curved for shallow depths, and therefore the display can seem odd at shallow depths (for example, 84-562\EC6.PS).
BH_GEOLOGS.ZIP (approx. 400Kb): Contains scanned versions of the summary geologic logs, from the open-file reports. The files have names such as GEOLOGS_XX-XXX.PDF, where XX-XXX is the open-file year and number. The geologic logs are in the process of being scanned and will be added as time goes on; I did not want to hold up putting the models on the web until all the logs had been scanned. As there will be in excess of 13Mb of scanned files, I will probably break BH_GEOLOGS.ZIP into several parts. The summary logs often have plots of the velocity models on the same page, but the user is warned that these models might differ somewhat from the models contained in the files available from the web site, because the analysis methods have changed through the years (as discussed earlier). This can result in small differences in the velocities from 1992 to 1999, and larger differences for the earlier models, compared to the reinterpreted models contained here.
STA_BH.ZIP (approx. 33Kb) Contains the correspondence between strong-motion recording stations and boreholes. The fields of information include: hole_code; borehole site name; borehole latitude; borehole longitude; borehole reference; depth to bottom of borehole; distance needed to extrapolate to 30m; average velocity to 30m, including any needed extrapolation; sta_code; agency code 1; agency code 2; station name; station location; station latitude; station longitude. hole_code and sta_code are unique numbers assigned in my database to each borehole and station. The station database was started by importing the contents of some USGS open-file reports containing station information; this database was then built up using the various data reports for the bigger California earthquakes (Loma Prieta, Cape Mendocino, Landers, and Northridge). Because of the evolutionary nature of the database, the two "agency code" fields may not be accurate in all cases (stations were sometimes transferred between the USGS and CDMG [California Division of Mines, now the California Geological Survey]), but they can be useful in identifying stations, because the agencies providing data for the stations know nothing about my database code numbers. Note that more than one station can be associated with a borehole (usually it does not work the other way around), and therefore a borehole may occur in a number of rows in the file, each with a different strong-motion station. OFR 03-191 contains a discussion of how the correspondence strong-motion stations and boreholes was established.
kntclss.xls (approx. 250Kb) contains the spreadsheet I used to estimate site classes at K-Net sites. Because no K-Net shear-wave velocities were determined for depths below 20 m, it is necessary to extrapolate the velocity models in order to estimate the site classes. The method for doing so was discussed in Appendix A of Atkinson and Boore (2003) and expanded on in Boore (2004). It is important to realize that the estimates are based on a statistical analysis and no claim is made that the estimated site class is correct at each station. See the aforementioned papers for more details.