Model Run Checklist
This blog post serves as a checklist for all the little loose ends that I need to tie up before launching the new model runs.
- Set up header file and compile ROMS on klone (x11ab)
- Define AVERAGES and undefine PERFECT_RESTART
- Use new fennel.h with 50% particle burial and corrected carbon burial
- Get latest version of ROMS
- Delete older varinfo files on local pc/perigee/apogee
- Git pull latest version of ROMS on klone
- Recompile ROMS after git pull
- Increase nudging to climatology
- Create nudgcoef.nc file for all biogeochemistry variables
- Update nudgcoef.nc file name in BLANK.in (to nudgcoef_tracer100.nc)
- Copy nudgecoef_tracer100.nc file into LO_data on apogee
- Set up dot in folder
- Copy Parker’s cas7_t1_x11ab folder to LO_user and rename as cas7_newtraps_x11ab
- Use newest BLANK.in version with shifted tides
- Use MPDATA advection scheme for bio tracers in bio_Fennel.in
- Adjust forcing list (atm00, ocnG00, tide00, trapsN00)
- Ensure make_dot_in.py is compatible with 320 cores for tiling
- Make sure make_dot_in is set up for daily averages with continuation
- Make another copy called cas7_newtrapsnoN_x11ab
- Specify trapsN01 forcing
- Specify trapsN01 forcing
- Generate new river/WWTP forcing using updated TRAPS for October 2012 - December 2017 on apogee
- Final code edits to eliminate any remaining bugs
- Adjust forcing code to discharge zero NO3 and NH4 from WWTPs
- Generate one set of forcing with DIN in WWTP effluent (trapsN00)
- Generate another set of forcing with zero DIN in WWTP effluent (trapsN01)
- Check that forcing is correct
- Make an adjusted driver_roms4 script in LO_user/driver
- Set up file to look for atmospheric, tidal, and ocean forcing in Parker’s apogee account (dat1/parker/LO_output/forcing/cas7)
- But look for river/WWTP forcing in Aurora’s apogee account
-
SCP files in LO_data/grids/cas7 into klone?
- Launch model run
- Email Parker before-hand to get final confirmation of resources
- Check with Jilian about resources
- Use start type “newcontinuation”
- Use 10 ckpt-g2 slices for each test condition
- Save output on dat1 in apogee
-
Set up crontab so I receive email updates of model progress every few hours
- Generate script to calculate bit-reproducibility after one day of run time