The CCSM Graphical User Interface (GUI) will generate example scripts to run the CCSM for various time periods in configurations different from the default ``all active models'' configuration in the CCSM2 distribution. To do this, the GUI modifies the scripts/test.a1/test.a1.run example script so that it will run the selected model configurations.
The GUI is a perl wrapper for some of the tools created by the CCSM developers at NCAR for automated generation of different model configurations. While the GUI is designed to generate scripts that run at NCAR, the resulting scripts are a good starting point for off-site users as well.
Currently, the only officially supported configuration and resolution combination is all active components (Configuration B) at the high and low resolution. While the GUI will generate test scripts for other combinations, the results have not been scientifically validated.
The CCSM GUI can be found in the .../ccsm2.0.1/scripts/gui_run directory. The GUI is designed so that it does not have to be run on the same machine that the CCSM model will be run. PerlTK is required to run the GUI. Start the GUI by typing ./ccsm_gui and a window will be created with a few pull-down menus and some input and check boxes. The GUI will search for any previous GUI settings archived in the file /ccsm_gui.
First select the platform that the CCSM will be run on: aix for IBM SP AIX and sgi for SGI Origin 2000-type machines. If the target machine is neither of these, chose 'aix' if the target machine is a distributed memory architecture and choose 'sgi' if it is a shared memory architecture. Then select the desired resolution.
Information on the directory structure of the target machine where the CCSM will be run needs to be supplied. Given the proper information, the CCSM gui will generate a new directory containing a CCSM model run script corresponding to the options that were selected.
| 1. | Select CCSM climate period specifies the climate period to configure the CCSM for. Standard is present day climate (recommended). Paleoclimate reproduces the T31_gx3v4 b20.604gm paleoclimate configuration control run (which used 1990 orbital parameters and a fix to the Gent-McWilliams eddy transport parameterization and isopycnal diffusion file (hmix_gm fix in scripts/test.a2/src.ocn/hmix_gm.F). |
|---|
| 2. | Select target architecture specifies the computer platform that the model will be run on. Selecting IBM results in a version compatible with a 4-processor per node IBM-SP. Selecting SGI creates a version for SGI Origin 2000 or 3000 models. |
|---|
| 3. | Select CCSM resolution The CCSM2.0 release officially supported only one resolution (referred to as "T42_gx1v3"), and one configuration (fully active (B)) with respect to scientific results. This combination was used in the CCSM2.0 Control Run and is the default configuration in the test.a1 case. The CCSM2.0.1 release supports a second, coarser resolution, "T31_gx3v4" version of the fully active coupled model. "T31" and "T42" are short-hand references to the resolution of the atmosphere and land models; "gx1v3" and "gx3v4" are short-hand references to the resolution of the ocean and ice models. In addition to these two scientifically validated configurations, a number of other configurations are available for scientific testing and research. More information on the CCSM resolution can be found at http://www.cesm.ucar.edu/models/ccsm2.0.1/ccsm2.0.1.faq.html |
|---|
| 4. | Select a Configuration selects the desired combination of active and data model components. The B configuration (all active components) is used for most CCSM runs. The G (active ice, active ocn, data latm and data lnd components) and M (active ice, mixed-layer ocn, data latm and data lnd components) configurations are used to test the sea-ice model and require a T62_gx1v3 resolution. Any of the other configurations can be built at either T42_gx1v3 or T31_gx3v4. Results for configurations other then a B configuration have not yet been scientifically verified. |
|---|
| 5. | Select stop time for the run allows the user to specify the length of the model run. The choices are "stop the run after 5 days", "stop the run after 10 days", "stop the run at a new month", "stop the run at the half year", "stop the run at the new year" |
|---|
| 6. | Enter ccsm root directory (CSMROOT) (string) specifies the top level directory where the ccsm code and
scripts reside. The scripts will expect that the complete CCSM2.0.1 code
distribution can be found under this directory.
For example, if CSMROOT is set to /home/user/ccsm2.0.1 then the scripts expect that the following directories exist: /home/user/ccsm2.0.1/models and /home/user/ccsm2.0.1/scripts CSMROOT must be entered by the user. |
|---|
| 7. | Enter case name for run (CASE) (string) specifies the CASE name that identifies this CCSM run. Two scripts will be generated in the directory CSMROOT/scripts/CASE: CSMROOT/scripts/CASE/CASE.run and CSMROOT/scripts/CASE/CASE.har. CASE.run corresponds to the one month startup run script. CASE.har corresponds to the harvesing script Archiving and harvesting are by default turned off in CASE.run (the lines are commented out). To active these features, the user must edit the script appropriately If the file CSMROOT/scripts/CASE/CASE.run already exists, the gui will not generate the run script and will issue a warning. |
|---|
| 8. | Enter ccsm data directory (CSMDATA) (string) is the root directory for the CCSM input and boundary datasets. This directory differs from the others in that it is not created by the CCSM run script. It is assumed that CSMDATA already exists and contains the data files contained in the CCSM2.0.1 input data distribution tar files. For example, if CSMDATA is set to: /home/user/input_data Then the scripts will expect the following input data directory structures to exist underneath /home/user/input_data: atm, cpl, ice, lnd, ocn. CSMDATA is where the input data tar files are untarred to. |
|---|
| 9. | Enter ccsm executable directory (EXEROOT) (string) is the directory where the model will run. Subdirectories for each of the model components will be created under EXEROOT. Typically, this is a local, fast, and large temporary disk space. For example, if EXEROOT is set to /ptmp/user/CASE Then, the following directories will be generated by the script at run time below EXEROOT:all, atm, cpl, esmf, ice, lib, lnd, ocn. The model will be run in these directories. |
|---|