EQ
is an earthquake forward-modelling framework based on the GTdef
(Georgia Tech
Deformation) code. It can be configured to model earthquakes along a given slab geometry
(in our case Slab1.0), or it may otherwise be used to create a fault plane based on given
input of (longitude,latitude,depth,strike,dip), or lastly along a horizontal plane
centered around a point (longitude,latitude,depth). This file is designed to help the
user set up the code for their own use, with some tips on how to customise model input
parameters.
Tests indicate that despite the limits of forward modelling data, we are able to relocate events with sparse GPS data closer to regional seismic estimates better than their initial teleseismic inputs. This therefore allows those interested to quickly obtain improved constrained estimates of event location without requiring detailed seismic data.
Any questions related to the main EQ
code may be directed to me. Questions relating to
the GTdef
code dependency should be directed to Dr. Lujia Feng, currently at NTU.
Authors:
- Nathanael Zhixin Wong (Main
EQ
code): nathanaelwong@fas.harvard.edu - Dr. Lujia Feng (
GTdef
dependency): lfeng@ntu.edu.sg
The first step is always to clone/fork the repository
git clone https://github.com/natgeo-wong/Sumatra-ModEQ.git
Running EQ
requires MATLAB 2016b and above, at least. It also requires the following
toolboxes:
- Parallel toolbox
- Optimization toolbox
The basis of our forward models is that we must already have a set of predefined inputs.
We generally take this from gCMT teleseismic estimates. Based on these estimates, we
create a .eq
input file by running the EQ_create.m
script. This script creates a
folder in the current working directory that contains the relevant .eq
script. Both the
working directory and the .eq
input script are named based on the date, magnitude,
regression used and the rigidity.
It is advised that the user create a separate directory that contains all these input
files and directories, because EQ
will run the modelling and analysis inside these
directories.
After this, it is necessary to copy the contents of EQ_sub
into the working directory.
EQ_sub
contains both the master script EQ.m
and submission scripts (e.g. *.pbs
) that
are necessary to run these scripts on a server. The submission scripts will vary
depending on the different servers and user style of inputs. It is also necessary to
modify the startup.m
file, to add the paths of the main EQ
codes, slab files and GPS
velocity input files.
For example, in the server, my codes are found in /home/nwong002/Codes/EQ/
. Therefore,
in my startup folder, I add the following paths:
/home/nwong002/Codes/EQ/ # Folder containing main EQ codes
/home/nwong002/Codes/EQ/Slab/ # Folder containing slab/fault input information
/home/nwong002/Codes/EQ/GPS_vel/ # Folder containing GPS velocity information
/home/nwong002/Codes/EQ/GTdef/ # GTdef code directory, requires a separate download
In the directory, you should have at least the following files:
*.eq
startup.m
*submission script
Once the script has been run, you should see at least the following additional files and folders:
./GTdef_L_Input/ # Directory containing input information for L loop
./GTdef_L_Output/ # Directory containing output information for L loop
./GTdef_L_Misfit/ # Directory containing misfit information for L loop
./GTdef_SV_Input/ # Directory containing input information for SV loop
./GTdef_SV_Output/ # Directory containing output information for SV loop
./GTdef_SV_Misfit/ # Directory containing misfit information for SV loop
./GTdef_SW_Input/ # Directory containing input information for SW loop
./GTdef_SW_Output/ # Directory containing output information for SW loop
./GTdef_SW_Misfit/ # Directory containing misfit information for SW loop
*bfit*.txt # Textfile containing bestfit information
Based on this information, you can then run the analysis.m
script. Once the script has
been run, you should see the following additional folders:
./Analysis_ve1/ # Contains information required to plot patchfiles in GMT4 (from SV loop)
./Analysis_ve2/ # Contains information required to plot patchfiles in GMT4 (from SW loop)
Generally, I choose to use the SW (GPS-weighted variance explained) information in my plotting and analysis, as it tends to put more weightage on the horizontal displacement observed, and measurements of horizontal displacements are more accurate than those of vertical displacements for GPS.
The GTdef code was created by my mentor Dr. Lujia Feng and her advisor Prof. Andrew Newman at Georgia Technological University. Though the EQ
code requires the GTdef
code as a dependency, the GTdef
code is not opensource and therefore I do not have the rights to distribute it in public. However, those who are interested in the GTdef code may email Dr. Feng at lfeng@ntu.edu.sg to attain a copy of the code for your use. If GTdef
is used in your work, please acknowledge Dr. Feng and her advisor Prof. Andrew Newman in your work. Any questions relating to the GTdef
code may be directed towards her.