MFAST_PACKAGE


If you use this package, please let me know at this address so I can keep you up to date on any changes: Martha.Savage [ at ] vuw.ac.nz

Versions 2.0 and later include modifications to the fortran codes for grading and for calculating the error bars.
The error bar calculation uses corrections detailed in Walsh, Arnold and Savage, Silver and Chan Revisited, currently online in J. Geophysical Research, 2013.
Version 2.2 uses gmt 5, and also fixes a bug that used a wrong search window and therefore has fewer good measurements. If you want to use gmt 4 you might need to mix and match the routines from version 2.1 and 2.2. See the 00README file for changes made, or contact me.

Here is the package, containing everything including codes and manuals: mfast_package_v2.2.tar.gz
Here is the JGR submission describing the results (if you have access to JGR it is recommended that you access the final published version from the JGR website): savage_etal_autosplit_jgr_rev2.pdf
Here is the JGR final published version from the JGR website: http://onlinelibrary.wiley.com.helicon.vuw.ac.nz/doi/10.1029/2010JB007722/pdf
Here is a link to the paper describing the original Teanby et al. method: http://bssa.geoscienceworld.org
Here is the documenatation alone: mfast_manual_v2.2.pdf
Here is a description of how to set up all the programs that mfast depends on: how_to_setupmfast_fromscratch
Here is a copy of a motivational presentation at a workshop for MFAST : mfast_workshop_2013_am_mks_includeskyushu.pdf
Here is a copy of a presentation for explaining mfast at a workshop for MFAST : mfast_workshop_2013_pm_mks.pdf
Here is the website for TESSA, a packge that does spatial averaging for fast directions and 2-D tomography for delay times, using the output of MFAST: TESSA

####################################################################
Copyright (C):
All scripts and programs are property of the authors unless otherwise noted.
Scientific non-commercial use is ok. Email me if you want to use it for commercial use.

Andreas Wessel (awbochum@gmail.com), 2007, at Victoria University of Wellington, New Zealand.
And Martha Savage (Martha.Savage@vuw.ac.nz) 2015 at Victoria University of Wellington, New Zealand
####################################################################

Installation:
- put the tar file into a directory, e.g., /seismo:
- untar the file tar -xvf mfast_package_v2.2.tar (may need to use gunzip first)
- read the manual in the documentation directory, and the JGR submission describing the results.
- add directory mfast/bin to path. do NOT use ~ for your home dir, always give the full path: path should be set in a file called .bashrc or path.cfg or something similar (e.g., in path.cfg add the line: PATH="${PATH}":/seismo/mfast_package_v2.2/mfast/bin
There are now sample scripts for setting the paths in the sample_data directory
- Install/compile SPLIT_mfm software (Teanby and Martha Savage: cd split_mfm/code; make all), add "bin" to path (e.g. /seismo/mfast_package_v2.2/split_mfm/bin ) and
- e.g., in path.cfg or .bashrc, set environment variable SPLIT_HOME to split_mfm directory (e.g. /seismo/mfast_package_v2.2/split_mfm)
- mfast/bin/mfast_config: set paths and parameters in this file
- mfast_precheck: run to see if programs are ready
- try to run all the scripts from BASH
- try the sample data as follows:
>cd sample_data
cp -r raw_data LHOR (copies all raw data into directory LHOR)
>./do_station_mfm LHOR (runs all scripts on data)
>>check that the results match those in the directory LHOR_example.
>>>don't expect exact matches if different operating systems are used.
- In particular, different systems may sort the output in different order, so you may need to use the unix "sort" command on both .summ files to make sure the order in the sample file and the new file are the same.
- You can do this easily using the script test_versions.bash in the sample_data directory now.
- Also there are now outputs from previous runs of do_station_mfm LHOR >& output_dostation.
>> You may want to try the archive.sh command that has been commented out of the do_station_mfm script to reduce directory clutter and file sizes. You can compare that output to the directory called LHOR_example_archived.
- read the documentation in the documentation directory to understand what the script do_station_mfm does and how to interpret the resulting plots.
- If you have high-frequency data with small earthquakes close to your stations you should try the scripts and data in very_local, which is a subdirectory of sample_data.
- There are hints in the documentation for trouble shooting.
-you are ready to try it on your own data. You will probably want to modify the script do_station_mfm for your own purposes.

-Note that the do_station_mfm code only makes some of the plots discussed in the JGR paper. The utilities directory has some codes to make some of the other plots in the JGR paper. The linux utility "awk" is also very helpful to change grading criteria. Examples are in the manual in section 5

Data Preparation:
- an S-Pick is needed (SAC header t0; or t5 or t6 if using a different script)
- SAC header variables RESP0, RESP1, RESP2,RESP3 KUSER0 are now reset to undefined in the scripts.
- some other SAC header fields are recommended: kstnm,evdp,slat,slon,...see manual for full discussion
 

- SAC header variable LCALDA must be set to TRUE
- SAC file names may need to be less than 50 characters and end in .e, .n, .z to run through all the scripts properly.
- To get angles of incidence set properly, modify the velocity file ak135_alp.tvel to include your own velocity model, set that velocity model in the mfast_config file (it is currently in the mfast subdirectory), and run the Taup command taup_create to make a new velocity model. You may need to run taup_create again if you use a different version of Taup or use it on a different type of computer.