This section describes a typical directory structure the user will be working with when running the TIEGCM model. The working directory (workdir) is the “root” directory of the user’s project.
The model directory (modeldir) is typically a subdirectory under the working directory, and contains the model source code, supporting scripts, documentation, and scripts for running tests.
The data directory (datadir) may also be a subdirectory under the working directory, or it may be on a large temporary disk that is accessible from the working directory. The data directory contains start-up and input data files for running the model.
The user’s working directory will typically look something like this (the datadir can be on a large separate disk system):
workdir
|
-----------------------------------------------
| | | |
*.inp execdir/ modeldir/ datadir/
*.job |
*.out *.o
*.mod
*.nc
Make*
exec
Here, \*.inp are namelist input files, \*.job are job script s, and \*.out are are stdout files from model runs. The execdir is the build/execution directory (created by the first build), with object code (\*.o, \*.mod), model output history files (\*.nc), make files (Make\*), and an executable file (exec). Various other files may also be in the execdir. The modeldir and datadir directories are described below.
The job script in your working directory contains a shell variable specifying the modeldir, so it knows where to find the source code and supporting scripts for the build process. The namelist input file also refers to the datadir for start-up and other data input files (e.g., SOURCE, GPI_NCFILE, IMF_NCFILE, etc). These namelist parameters can use the environment variable TGCMDATA to specify the datadir (see section on namelist input files).
The model root directory is what you get when you download the model source code tar file. The model directory contains subdirectories with the model source code, supporting scripts, documentation, and test scripts:
modeldir
|
----------------------------------------------------------------------
| | | |
src/ scripts/ doc/ tests/
| | | |
Make.* userguide.pdf control/
*.F linux.job description.pdf climatology/
*.h ibm.job Release_Notes dec2006
default.inp diags.table nov2003/
tgcm_contents README whi2008/
tgcm_ncdump
etc
The src/ directory contains all source code, i.e., all *.F fortran files, and a single header file with grid definitions and dimensions (defs.h). The TIEGCM source code is f90 compliant, and most files are written in fixed-format Fortran.
The scripts/ directory contains various csh, perl and other scripts providing infrastructure support for the model. These include several Make files, sample job scripts, a default namelist input file, and several utility scripts.
The documentation directory doc/ contains the User’s Guide (this document), and Model Description (pdf and html files). These docs are also available in the documentation section of the main tgcm website tgcm website. The User’s Guide is written in Restructured Text, with the Sphinx markup tool. Source code (.rst and other files) for the User’s Guide are available in the “userguide” subdirectory. The Model Description is written in TeX, and its source code is available in the “description” subdirectory.
Also in the doc directory are Release Notes for the current version, a README file, and other relevant documents (e.g., table of available diagnostic fields from the model).
The tests/ directory contains subdirectories for various benchmark test runs, (seasonal, full-year climatology, storm cases, etc.) Each subdirectory contains namelist input files and job scripts that can be used to reproduce these simulations for validation and testing purposes.
The public TIEGCM data directory is what you get when you download the data tar file. This directory is typically referred to with the environment variable TGCMDATA. Subsequently, after the data download, you may obtain additional needed data files from the NCAR Community Data Portal. Here is a partial schematic of the datadir (where “tiegcmx.xx” is the desired model version):
datadir
|
-----------------------------------------------
| |
gpi*.nc tiegcmx.xx/
gswm*.nc |
imf *.nc TGCM.tiegcmx.xx.p*.nc
etc pcntr*smin*.nc
pcntr*smax*.nc
etc
Files listed on the left side refer to data input files that may be needed when running the model in different modes. These are netCDF files, specifically prepared for import into the TIEGCM model (they are not model start-up files). These files are version-independent (can be used by (almost) any version of the model). They are usually provided to the model as namelist input parameters:
The “tiegcmx.xx/” subdirectory refers to the version of the model that was downloaded. This subdirectory contains start-up SOURCE <SOURCE> files from benchmark runs executed by that version of the model (see Section on Benchmark Test Runs). These files can be used to remake the benchmark runs for testing and validation. Here is an example of start-up files provided for benchmark runs made by TIEGCM version 1.94:
TGCM.tiegcm1.94.p_dec2006_heelis_gpi_001.nc
TGCM.tiegcm1.94.p_dec2006_weimer_imf_001.nc
TGCM.tiegcm1.94.p_nov2003_heelis_gpi_001.nc
TGCM.tiegcm1.94.p_nov2003_weimer_imf_001.nc
TGCM.tiegcm1.94.p_whi2008_heelis_gpi_001.nc
TGCM.tiegcm1.94.p_whi2008_weimer_imf_001.nc
TGCM.tiegcm1.94.pclim_heelis_001.nc
pcntr_decsol_smax.nc
pcntr_decsol_smin.nc
pcntr_junsol_smax.nc
pcntr_junsol_smin.nc
pcntr_mareqx_smax.nc
pcntr_mareqx_smin.nc
pcntr_sepeqx_smax.nc
pcntr_sepeqx_smin.nc