time_series_analysis

Here are the instructions for running global line plots on Chrysalis:

Script:  /home/ac.ricciuto/models/CBGC_analysis/analysis_CBGCv2.py

Edit the script (lines 32-48) to change the case names, usernames (who ran the cases), short case names and run directory locations as needed.  It is currently set up to plot 2 cases but should work to plot up to 5.

Then start an interactive job (you’ll need at least 30 minutes to run the script), this allows for one node for 100 minutes:

srun -N 1 -t 100 --pty bash 

Load e3sm_unified:

source /lcrc/soft/climate/e3sm-unified/load_latest_e3sm_unified_chrysalis.sh

Then run the script:

srun -n 40 --oversubscribe python analysis_CBGCv2.py --year1 1850 --year2 2013 --varfile default_vars.txt --machine chrysalis

You will see some mpi errors but it runs ok despite this.  The file default_vars.txt (in the same directory; you will need to copy this too) has 60 commonly used variables.  It will take about 30 minutes to run.  You may specify your own variables by replacing “--varfile default_vars.txt” with the argument “--vars var1,var2,varn” where var1,var2… are the variables of interest.  The output will go to

https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/<USER>/CBGC_lineplots/<CASE1_CASE2>

For example, see here:  Index of /public/e3sm/diagnostic_output/ac.ricciuto/CBGC_lineplots/20TRWCCNP_20TRWCCNPPHS/