next up previous contents index
Next: 7.1.12 Running HYP Up: 7.1 The hypocenter program, Previous: 7.1.10 Alternative model   Contents   Index

7.1.11 Using HYP to determine crustal structure

HYP has an option to locate a data set for a large number of different models and then determined which model gives the lowest average RMS for the data set. This might be a useful option, particularly when a sparse data set is available. In order to use this option, an additional input parameter file h_models.par is given. When this file is in the working directory, HYP will switch to multiple model mode SO ONLY HAVE THIS FILE IN WORKING DIRECTORY IF MULTIPLE MODEL MODE IS INTENDED. When using this option, all events must use the same STATIONx.HYP file, otherwise the program fails. The input MUST be from a single file, NOT from the data base. THE PROGRAM MUST RUN IN NON INTERACTIVE MODE. Below is an example of an input file.


layer #  start vp  delta vp   # delta   start h   delta h   # delta
      1      4.55       0.1         5       0.0       1.0         1
      2       6.3       0.1         5       4.0       1.0         1
      3       6.8       0.1         5      22.5       1.0         1
      4      7.90      0.05         3      32.5       1.0         1       
      4      8.05      0.05         4      40.0       1.0         1

The first line is info only. Layer # is also only for information. For each layer, there is a start P-velocity (start vp), increment in velocity (delta vp) and number of increments (# delta). The following inputs are then the same for layer depths. There must be an entry for each layer even if no variation is used. In the above example, no variation in layer thickness is tested for. An example input file is given in DAT. The parameters for location not set in h_model.par like Vp/Vs, Lg velocity etc remain unchanged. When HYP starts up, it will print out how many permutations are required. If more than a few thousand, reduce the number of models. In any case it is an advantage to first try with just a few models to get a feeling for how sensitive the data is for model changes.

An output file h_models.out is generated, see example below. For each model tested, one output line is given with the RMS and the model. In the example below only the last 5 models are shown. Since many models can have very similar average RMS, the best 10 models are printed at the end.


  0.946    4.95  0.00  6.70  4.00  7.20 24.50  7.90 32.00  8.20 40.00
  2.607    4.95  0.00  6.70  4.00  7.20 24.50  7.90 32.00  8.30 40.00
  0.934    4.95  0.00  6.70  4.00  7.20 24.50  8.00 32.00  8.10 40.00
  0.994    4.95  0.00  6.70  4.00  7.20 24.50  8.00 32.00  8.20 40.00
  2.677    4.95  0.00  6.70  4.00  7.20 24.50  8.00 32.00  8.30 40.00

  Minimum rms   0.764057

  The best models

  0.771    4.95  0.00  6.50  4.00  7.10 24.50  7.80 32.00  8.10 40.00
  0.766    4.85  0.00  6.50  4.00  7.00 23.50  7.80 32.00  8.10 40.00
  0.767    4.85  0.00  6.50  4.00  7.00 24.50  7.80 32.00  8.10 40.00
  0.769    4.85  0.00  6.50  4.00  7.10 23.50  7.80 32.00  8.10 40.00
  0.766    4.85  0.00  6.50  4.00  7.10 24.50  7.80 32.00  8.10 40.00
  0.772    4.85  0.00  6.50  4.00  7.20 24.50  7.80 32.00  8.10 40.00
  0.771    4.95  0.00  6.50  4.00  6.90 22.50  7.80 32.00  8.10 40.00
  0.771    4.95  0.00  6.50  4.00  7.00 22.50  7.80 32.00  8.10 40.00
  0.770    4.95  0.00  6.50  4.00  7.00 23.50  7.80 32.00  8.10 40.00
  0.771    4.95  0.00  6.50  4.00  7.00 24.50  7.80 32.00  8.10 40.00


next up previous contents index
Next: 7.1.12 Running HYP Up: 7.1 The hypocenter program, Previous: 7.1.10 Alternative model   Contents   Index
Peter Voss : Mon Feb 27 10:16:12 UTC 2017