Running an Experiment
The examples in the CCLM starter package are built in a way that everything will be stored under the directory tree starting with $SPDIR
and therefore will be on the same file system. This is often not convenient, it might be better to put scripts and I/O to different more appropriate file systems. This could be done by setting the main directory settings (see table) in the job_settings script accordingly.
Before you run your own experiment it is recommended to run the test cases sp001
and sp002
as described in the Install section.
Creating the first new experiments by yourself
Downscaling simulation with GCM data (gcm2cclm)
Make a copy of the template sp001
as starting point for your experiment (this is named mytemplate1
in the following, but you can choose a name by yourself).
cp -R ${SPDIR}/chain/gcm2cclm/sp001 yourpath/mytemplate1
Change into the yourpath/mytemplate1
directory and perform your modifications in the job_settings file and the other scripts. Then perform a clean to get rid of any unnecessary files that may have transferred during the copying of sp001
.
cd yourpath/mytemplate1
./subchain clean
Now you are ready to start your experiment
./subchain start
If the simulation is successful you can find the output data under the directories $ARCHIVE_OUTDIR
and as post-processed time series under $WORKDIR/mytemplate1/post.
Nested simulation (cclm2cclm)
Make a copy of the template sp002
as starting point for your experiment (this is named mytemplate2
in the following, but you can use a name by yourself).
Change into the yourpath/mytemplate2
directory and perform your modifications in the job_settings file and the other scripts. Then perform a clean to get rid of any unnecessary files that may have transferred during the copying of sp002
.
Now you are ready to start your experiment
If the simulation is successful you can find the output data under the directories $ARCHIVE_OUTDIR
and as post-processed time series under $WORKDIR/mytemplate2/post
.
Creating any further experiments
After you have created an experiment as in the examples above (mytemplate1
or mytemplate2
) you can use this experiment as a template (or as your new sp001
or sp002
so to speak) for your further experiments. This will save you some time since you have not to do a lot of basic changes again.