Parts of the code are setup to run on T.Carman's local workstation, and parts are to be run on tcarman@modex.bnl.gov.
Overall workflow is like so:
- Update and compile desired versions of PEcAn and dvmdostem.
- Update dvmdostem-input-catalog (use dvmdostem update-mirror.sh script).
- Run model (via pecan) over the various experiment cases.
- Adjust pecan xml file(s) for run(s)
- submit with nohup_wrapper.sh
- preprocessing (pecan meta analysis) happens on head node
- model runs are submitted by qsub and happen on other nodes
- model2netcdf I think runs on head node
- run posthoc_easa.R
- use the ph_nohup_wrapper.sh
- this does not use qsub - all work is done on head!!
- Convert Rdata to csv files using rdata2csv.sh
- rsync to local
- change into ~/Documents/SEL/NGEE_Dec_2018_followup
- rename (move) the ngee_dhs_runs directory to archive
- then get the new stuff
$ rsync -avz --exclude "*/out/*" --exclude "*/run/*" --exclude="*/yearly_runs/*" modex.bnl.gov:/data/tcarman/ngee_dhs_runs .
- run plotting with ppp.py
Cite this code:
