This section describes the tomography results and the naming of the files.
This effort inverted Love and Rayleigh phase and group velocity observations on a rectangular grid. In addition a spike test was performed to give an idea of the resolution possible with each data set. The results of each of these inversions is placed into different directories, within which there are files with names vc_000_001.xyz, vc_002_001.xyz, ..., per.uniq and per.uniq.NUM.
In the distribution the results for a region are to be found in the following directories:
LOVEc [Directory for Love phase velocity tomography] LOVEU [Directory for Love group velocity tomography] RAYLc [Directory for Rayleigh phase velocity spike test tomography] RAYLU [Directory for Rayleigh group velocity spike test tomography] LOVEc.spk [Directory for Love phase velocity tomography] LOVEU.spk [Directory for Love phase velocity tomography] RAYLc.spk [Directory for Rayleigh phase velocity spike test tomography] RAYLU.spk [Directory for Rayleigh group velocity spike test tomography]
In each of these directories, there will be the inversion output for each period, e.g., vc_002_001.xyz, which is for period number 3 (C notation, 3 = 1+ 002), that consists of four columns:
24.112 -120.155 3.370 0.00 24.112 -119.909 3.370 0.00 24.112 -119.662 3.370 0.00 24.112 -119.416 3.370 0.00 ..................... 34.905 -90.466 3.392 39.98 34.905 -90.192 3.395 69.98 34.905 -89.917 3.392 328.20 34.905 -89.643 3.367 199.95
The first column and second columns are the latitude and longitudes of the center of the cell. The third column is the dispersion velocity from the tomography, and the last column is sum of all path lengths passing through that cell.
the use of the fourth column must be discussed. The inversions use a 25 km x 25 km grid. In addition the inversion is initialized with the average dispersion velocity of the entire data set. In addition the inversion is iterative and not formal inverse matrix is obtained from which to estimate errors In the tabulation above, the first four coordinates all have the same dispersion value, but no rays actually passed through the cells. Thus there was no information upon which to estimate the dispersion, and thus the value should not be used.
This distribution contains the source code doquad.f for
the program doquad which outputs the dispersion in the
Computer Programs in Seismology surf96 format for a
specified latitude - longitude pair by interpolating the values of
the surrounding cells. If, and this is hardwired into the code,
the fourth column is less than 50km, then it is assumed that there
were not sufficient paths though that cell to use the
dispersion value.
To assist in associating the particular output file to a
period, two files are created - per.uniq and per.uniq.NUM.
The per.uniq file is just a listing of the periods,
with the first associated with the index 000, the second
with 001, etc. This is perhaps too programming
oriented. So the file per.uniq.NUM is also created.
In that file the first column is the index, e.g., vc_index_001.xyz,
the second column is the period, and the last column is the name
of the tomography output file.
per.uniq
2.000000 2.100000 2.200000 2.300000per.uniq.NUM
000 2.00 vc_000_001.xyz 001 2.10 vc_001_001.xyz 002 2.20 vc_002_001.xyz 003 2.30 vc_003_001.xyz