23.1.5 Program operation

The program makes a grid-search and finds how many polarities and amplitude ratios fit each possible solution. All solutions with less than a given number of wrong polarities and/or amplitude ratios within given error limits, are then written out and can be plotted. With a cursor, the user can then select the preferred solution, which can be stored in the input file or the database. The program is intended to work from within EEV (option F), however it can also work independently (see below). The program uses an input file called focmec.inp (automatically generated). This is a Nordic format file. Direct waves have angle \bgroup\color{black}$>90$\egroup and refracted arrivals angle \bgroup\color{black}$<90$\egroup degrees. If the angle is \bgroup\color{black}$>90$\egroup, the polarity is plotted at an azimuth+180. If the user wants to use FOCMEC as a freestanding program, the angle of incidence information may have to be put in manually in a standard CAT-file, which is then renamed focmec.inp. This can be done automatically by FOCMEC if a hyp.out and corresponding print.out file is available. FOCMEC can also be used to convert angles, like dip, strike and rake to T and P-axis, simply say 'focmec a', where argument a stands for angles and you will be prompted for input.

NOTE that the number of polarity errros is an integer number if no relative weighting of polarities is used and a real number if relative weighting is used, see intro for FOCMEC above and/or FOCMEC manual in DAT.

When the program runs, all amplitude information and corresponding corrections are listed. First there is a question of which type of amplitudes to be used. In the example below, automatic amplitudes are chosen.

Number of polarities:            7
Amplitude types:   Manual:       0   Automatic:   10   Spectral:    10

Amplitude to use:  Manual(1), Automatic(2), Spectral(3) ?

 Q: Local: Qp= 142.0**0.71  Qs= 142.0** 0.7   Global: t*(P)=1.10  t*(S)=4.20

 FOO   Z PG      2883   0.28   13.4    1.5    119     40  1.5   83    77
 FOO   T SG     23712   0.24   23.4    2.2    119     40  2.0   83    77
 SUE   Z PG      1448   0.28   13.8    1.6    118     41  1.4  129    81
 SUE   T SG     24591   0.32   24.0    2.1    118     41  2.0  129    81
 OSG   Z PG       734   0.29   18.2    1.8    106     46  1.3  199   120
 OSG   T SG     11610   0.25   31.6    2.8    106     46  2.0  199   120
 HYA   Z PG      1090   0.28   21.0    2.0    102     47  1.3  105   144
 HYA   T SG     12161   0.28   36.6    3.2    102     47  2.0  105   144
 MOL   Z PG       505   0.24   32.2    2.9     94     48  1.3   59   238
 MOL   T SG      2866   0.28   56.1    6.0     94     48  2.0   59   238

 STAT  Ratio type  T     Amp 1    Amp 2  Fcor LogRat
 FOO   SH(T)/P(Z)  H     23712     2883   0.7   0.93
 SUE   SH(T)/P(Z)  H     24591     1448   0.7   1.22
 OSG   SH(T)/P(Z)  H     11610      734   0.7   1.23
 HYA   SH(T)/P(Z)  H     12161     1090   0.7   1.08
 MOL   SH(T)/P(Z)  H      2866      505   0.6   0.87

 total obs =  12  gap in az = 220.0  gap in ain =  60.0

The abbreviations are STAT: Station code, C: Component, PH: Phase, AMP: Amplitude in count, PER: Period in sec, TRTIME: Travel time in sec, QCOR: Log Q-correction, ANGINC: Angle of incidence at the source, ANGEMG: Angle of emergence at the station, Fcorr: Free surface correction for this amplitude, Az: Azimuth from the event to the station, DIST: Epicentral distance in km., Ratio type (see text), T: indicator of ratio type, Amp1 and Amp2: The two amplitudes (count) in the ratio, Fcor is the free surface correction in the amplitude ratio (to be multiplied with ratio) and LogRat is the logarithm of the corrected amplitude ratio used. Total obs: Total number of ratios and polarities, gap in az: the gap of no observations (amplitude and polarity), should be less then 180 deg, gap in ain (amplitude and polarity): gap in angle of incidence, should be as small as possible.

Following, the user get the choices:

Stop                        (0)
Plot saved solution(s)      (1)
Plot new solutions          (2)
Plot selected solution      (3)
Find new solutions          (4)
-1, -2, -3 also plot station

  1. This is the solution(s) already stored in the data base (S-file). See secetion "Storing and selecting fault plane solutions" below.
  2. Plotting new solution after having used option 4
  3. Plotting the selected solution after using option 4
    Using e.g. -1 instead of 1, also plots the stations to help identify them on the plot, see Figure 23.1
  4. Starting a search for new solutions

    Option 4 gives the following information and questions:

     There are   7 polarity readings
     Use realtive weight, y/n=default
     Maximum number of allowed polarity errors, enter for 1

    Depending on number of data values, 0-5 is a good answer.

    If relative weighting is used, the question about number of polarity errors is:

    Realtive weight, number of polarity errors, can be a fractional number, enter for 1.0

    Relative weighting will give more solutions since the numbers summed are smaller than 1.0. So for a good solution, 0.5 miught be a good number to use.

    Equivalent for ratios to 'Maximum number of polarity errors', however, error is defined by amplitude ratio error. Number of errors depends on number of observations. For 5 observations 0-1 errors is reasonable.

    Maximum amplitude ratio error,  enter for default of .2

    Give maximum allowed difference between observed and computed log amplitude ratio, default is 0.2, which often is a good value.

    Degree increment in search, enter for default 2

The program will now start the searching and write out on the screen (and in a file) the solutions which fit the requirement of number of misfits. The maximum number of solutions is limited to 100 as a default, or to the value defined by `FOCMEC MAXSOL' in SEISAN.DEF. At the end, the number of acceptable solutions is written out as well as the minimum number of bad fits. This can then be used for the next search. Now option 0 to 4 can be used again.

When plotting the solution with option 2, the cursor comes up. Also, the solutions will be printed in text form to the screen, see Figure 23.1.

The abbreviations are Pol: Number of polarity errors for P note it is a real number, SV(not used) and SH(not used), AccR/TotR: Number of accepted to total number of ratio errors, RMS RErr: The RMS error for the ratios used, AbsMaxDiff: The maximum error. See also FOCMEC manual.

The polarities and amplitude ratios can be plotted on the focal sphere using the same convention as the original FOCMEC program, which is:

o = compression
\bgroup\color{black}$\Delta$\egroup = dilatation
V = amplitude ratio SV/P
S = amplitude ratio SV/SH
H = amplitude ratio SH/P

The user can select a preferred solution by moving the cursor near one of the letters T or P (T and P axis). By pressing T, the program will find the nearest T axis (same for P and nearest P-axis) and corresponding fault plane solution, which can be stored in the database and/or plotted with option 3. If no solution is to be selected, press q for quit. If a solution has been selected, the user will be asked if it is to be saved or not after selecting option 0. The saved solution goes into the focmec.out and from there into the S-file (type F-line) in the database if FOCMEC is operated from EEV and the solutiosn will also be written to fps.out.

When working from EEV, the event will always be located before the FOCMEC program starts up. In the Nordic format the solution is stored simply as strike, dip, rake and number of bad polarities (3f10.1,I5). Aki and Richards convention is used. In addition, the name FOCMEC will be written near the end of the line to indicate that the fault plane solution was made by FOCMEC. The line type is F.

The following files are created:

focmec.dat: Input parameters to FOCMEC_EXE.
focmec.log: Log of the FOCMEC_EXE run.
focmec.lst: More details on solutions
focmec.out: Gives input parameters and solutions
focmec.eps: A Postscript plot file of LATEST plot
focmec.run: Run parameters for FOCMEC_EXE, you can re-run FOCMEC by `focmec_exe < focmec.run'