Using the SCEC Version 4 southern California reference seismic velocity model 9/27/05 H. Magistrale First read: Magistrale, H., S. Day, R. W. Clayton, and R. Graves, 2000. The SCEC southern California reference 3D seismic velocity model Version 2, Bull. Seismol. Soc. Am., v. 90, no. 6B, p. S65-S76 This has pictures of the basin boundaries, tomographic coverage, etc. A pdf file of a preprint of the manuscript is in the anonymous ftp site: moho.sdsu.edu in the directory: pub/Version2 with the file name 'manuscript.pdf'. Version 3 differs from Version 2 in including tomographically determined Upper mantle seismic velocities. See: Kohler, M., H. Magistrale, and R. Clayton, 2003, Mantle heterogeneities and the SCEC three-dimensional seismic velocity model version 3, Bulletin Seismological Society of America 93, 757-774. Version 4 - the current version - differs from Version 3 in the following ways: San Bernardino Valley: A new San Bernardino Valley basement is based on recent USGS inversion of gravity data confirmed by comparison to a seismic reflection line. The new model features a deep trough in the central valley, in contrast to the previous flat-bottomed valley model. The new basement joins smoothly to the relatively shallow Chino basin to the west. Salton Trough: A new model is motivated by the needs of TeraShake simulations of southern San Andreas fault events. Depth to basement is defined by a combination of seismic refraction surveys, inversion of gravity observations, surface geology, and boreholes. Sediment velocity-depth gradients depend on the nature of the basement, smoothing merging into deep metasedimentary basement, and having a discontinuity above shallow crystalline basement. The model includes the portion of the Trough south of the international border. Vp-density: The new Vp-density relation is based on density measurements from oil well samples in the Los Angeles basin and the San Gabriel Valley, geotechnical boreholes throughout southern California, and 12 oil wells along the LARSE lines. The newly determined Vp-density ratio is constant, in contrast to the old relation. This is true even for low Vp, as defined by the geotechnical data. The new densities are higher, for a given Vp, than the old. This will tend to lower the Poisson ratio, which will lower Vp/Vs; that is, changing the Vp-density relation produces a new Vs model. The model exists as a fortran code and associated files. The code reads a file of points specified by latitude, longitude and depth, and writes out Vp, Vs, and density at those points. The compressed tar file Version4.0.tar.Z contains all the files. First uncompress (assuming a Unix machine): uncompress Version4.0.tar.Z or zcat Version4.0.tar.Z > Version4.0.tar Next extract the files: tar xvf Version4.0.tar Then compile: f77 version4.0.f -o version3.0 The file of points should be named "btestin" (this can be changed in the subroutine "readpts"). Here is a sample file: 8 33.50000 -118.50000 30.0 33.50000 -118.50833 30.0 33.50000 -118.51667 30.0 33.50000 -118.52500 30.0 33.97200 -118.08800 3000.0 33.97200 -118.08800 3001.0 33.97200 -118.08800 3002.0 33.97200 -118.08800 3003.0 The first line contains the number of points in the file (8 in this example). The code has array dimensions allowing up to 750,000 points. To handle more points, the arrays can be redimensioned by changing the parameter "ibig" in the include file "in.h", or by dividing the total point set into chunks of less than 750,000. The remaining lines are the latitude, longitude (both in decimal degrees, with the longitude negative, because this is the western hemisphere), and depth (in meters. Note in the version 1 model the depth was given in feet). These are read with free format. It is a common error to forget the negative longitudes. The model consists of rule and object parameterized basins embedded in a tomography background. The objects are reference surfaces outlined by polygons, so the code must figure out which polygon an input point is within. You can speed up the code by putting all of the points at the same latitude and longitude sequentially in the input file (as in the last 4 points in the sample). See note 2 below. The output will be a file name "btestout" (this name can be changed in the subroutine "writepts"). Here is a sample (for the last 4 points of the input sample): 33.97200 -118.08800 3000.00 3509.8 1817.4 2304.0 33.97200 -118.08800 3001.00 3510.8 1818.1 2304.2 33.97200 -118.08800 3002.00 3511.8 1818.9 2304.3 33.97200 -118.08800 3003.00 3512.8 1819.6 2304.5 These echo the latitude, longitude and depth, and then give Vp, Vs (both in m/s), and density (in kg/m^3). Notes and warnings: 1. This is a research code, and bugs and glitches may be revealed as it is subjected to new applications. 2. Input points can be in any order, although the code will run faster if the points are given with the same lat-longs sequentially. If you want to recover the geotechnical borehole shear wave you must give the lat-long to within 50 m of the borehole's lat-long. 3. The code assigns a minimum density of 1500 kg/m^3 to points with a Vp of 1586 m/s or less. This is done at line 825 of the main program. 4. The boundary between the Los Angeles area basins (see note 6) and the background model is relatively smooth except along the south and west edges (mostly in the ocean) of the basins model. This produces an artificial abrupt transition from the seismically slow basin sediments to the faster background model. This may produce unrealistic artifacts in, for example, waveform modeling. Watch out. 5. Points outside the area of the tomographic background (see note 6) will be assigned velocities from a smooth Hadley-Kanamori 1D model.