For some functionality the docker container might need the display, and so it is reccommended that the container is run with the following options:
docker run -it --network=host -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix --privileged --rm fcoherreazcue/ripdocker
In csf docker is not available, but Singularity is, and so the image can be imported and converted. From the csf run:
singularity pull docker://fcoherreazcue/ripdocker:latest
The above command should create a ripdocker_latest.sif file. To run the image in a similar way to Docker, we can use the command:
singularity shell --contain --cleanenv --pwd /Sample ripdocker_latest.sif
To be able to activate the ncl_stable conda environment inside singularity use:
source /miniconda3/etc/profile.d/conda.sh && conda activate ncl_stable
The following information on how to use RIP is a very succint summary of the docs and online tutorial from ucar, that can be followed directly from the /SAMPLE directory in the docker container. The sample data to test is already at /SAMPLE/WRFData, but can also be directly downloaded from ucar's website .
It is important to make sure that your ncl_stable environment is active before using rip. To activate it, run
conda activate ncl_stable
Also, make sure that you are in the directory with your .in files. For the sample run, this is
cd /Sample
The wrfout files will be pre-processed with ripdp. To run ripdp you first need a "namelist" file, that configures what ripdp will do. Since ripdp will generate a lot of files, it is better to first create a directory for them. For the case of the sample run, the directory /SAMPLE/RIPDP has already been created.
A sample namelist file is already saved in the RIPDP directory, under the name rdp_sample. We can then run ripdp as:
ripdp_wrfarw -n RIPDP/rdp_sample RIPDP/rdp_sample all WRFData/wrfout_d01_*
This should have created a lot of files inside RIPDP, called rdp_sample_XXX_YYY, where XXX is the time and YYY the name of the WRF variable in each file. This is the data that RIP needs to run, so we are now ready to use it.
Now that we have pre-processed data, we need to configure rip. We do this using a "User Input File". A sample file called rip_sample.in is already at /SAMPLE.
Using this file, we can run rip as:
rip -f RIPDP/rdp_sample rip_sample.in
This should create a log file called rip_sample.out, and a rip_sample.cgm file, with the generated plots.
The plots can be saved to a pdf format instead of cgm if in the user input file we set ncarg_type='pdf' (line 9)
Sample input files to compute back-trajectories can be found in /SAMPLE/BackTraj. Note that there are 3 almost identical files called traj_.in. These files have the flag itrajcalc=1, which tells RIP that they will calculate trajectories. Running
cd /Sample/BackTraj
rip -f /Sample/RIPDP/rdp_sample traj1.in
rip -f /Sample/RIPDP/rdp_sample traj2.in
rip -f /Sample/RIPDP/rdp_sample traj3.in
will create the trajectories from the 3 different starting points (note that zktraj is different in each file). Files with endings '.diag', '.out' and '.traj' should have been generated.
We can now plot these trajectories with traj_plot.in by running
rip -f /Sample/RIPDP/rdp_sample traj_plot.in
To be able to see the plots in cgm files, we need to use idt, by running
idt /Sample/rip_sample.cgm
or
idt /Sample/BackTraj/traj_plot.cgm