Skip to content

Latest commit

 

History

History
123 lines (103 loc) · 3.61 KB

notes.md

File metadata and controls

123 lines (103 loc) · 3.61 KB
Raw

September 9, 2022

  • Post-processing
    • S0

September 8, 2022

  • Had a mistake in S1
    • wasnt using the CO2 stream until 1850

September 6, 2022

  • Had a mistake in S2/S3
    • needed ./xmlchange DATM_CLMNCEP_YR_END=2021
    • restage 1901 restarts
      • cp /glade/scratch/djk2120/archive/TRENDY2022_f09_S2/rest/1901-01-01-00000/* ./

September 5, 2022

  • S2 and S3 crash
    • at different points
  • increasing S3 to 100 nodes
    • rm S3 case
    • qcmd -- ./S3.sh
  • nixing the 1860 switchout for both
  • S2 crashed coming out of 1900
    • was not removing .bin files correctly

September 4, 2022

  • create tx user.datm
  • S2 is broken into 5 parts
    • 1701-1850 (2x75) S2.sh
    • 1851-1860 (10) S2_part2.sh
    • 1861-1900 (40) ...
    • 1901-1960 (60) ...
    • 1961-2021 (61)
  • /glade/u/home/djk2120/tether/tether.sh joblist.txt ../../cheyenne.template &> S2_part1.log &
  • S3 is basically identical to S2
    • except for flanduse in the user_nl_clm
  • /glade/u/home/djk2120/tether/tether.sh joblist.txt ../../cheyenne.template &> S3_part1.log &

September 3, 2022

  • spinup looking good
    • <5% diseq (TOTECOSYSC 1g/m2)
  • running S0:
    • qcmd -- ./S0.sh > S0.log &
  • running S1:
    • qcmd -- ./S1.sh > S1.log & (creates,builds case)
    • bash mjl.sh (makes joblist.txt)
    • /glade/u/home/djk2120/tether/tether.sh joblist.txt ../../cheyenne.template &> S1_part1.log & (initiates 3job sequence)

September 2, 2022

  • Spinup is probably fine at 600yrs, will let it go overnight to 800

  • testing out S0 and S1 scripts

  • S1 is broken up into three parts

    • 100 years S1.sh
    • 49 years S1_part2.sh
    • 2x86 years S1_part3.sh

  • configuring tether for S1

    • /glade/u/home/djk2120/tether/tether.sh joblist.txt ../../cheyenne.template &> S1_part1.log &

September 1, 2022

  • runs die overnight due to exceeding scratch quota
    • went on an rm rampage
      • deleted an important PPE file
      • oops
  • resubmitted
    • progressing well

August 31, 2022

  • spinPostAD is progressing
    • <6% diseq after 200 years
    • drift of 0.015 PgC/yr
    • does seem to be spending a lot of time queued
  • preparing S0 scripts
    • scripts/S0.sh
    • recycle 1901-20 climate, 321 years, 1701-2021
    • made the executive decision to run from 1701, makes way more sense
    • probably will pickup from year 300 of postAD
  • preparing S1 scripts
    • uses a transient compset
      • double-check datm_in that taxmode='cycle'
    • preparing co2 timeseries
  • spun up with the wrong co2
    • accidentally used clm default 284.7
    • switched to TRENDY PI at postAD year300
      • ./xmlchange CLM_BLDNML_OPTS="-bgc bgc -crop -co2_ppmv 276.65"
      • easier to use ./xmlchange CCSM_CO2_PPMV="276.65"
    • resubmitted

August 29, 2022

  • spinAD has progressed 600 years
    • ~10% diseq at 1g/m2 (TOTECOSYSC)
    • drift = -0.05 PgC/yr
  • setting up postAD
    • postad.sh
    • qcmd -- ./postad.sh > postad.log &

August 27, 2022

  • Erik and Peter generated fsurdat, flanduse
  • performed fresh CLM checkout
    • checkout.sh
    • branch_tags/TRENDY-2019.n05_ctsm1.0.dev056
  • configured and submitted spinAD
    • spinad.sh
    • namelists/spinad
  • setting up git version control
  • next up:
    • monitor spinup progress
    • configure spinPostAD

August 2, 2022

  • Sean converted the input data to three stream
  • downloading ndep data from ftp site
  • converting ndep to CLM format via ndep.ipynb
  • edit user_datm.streams to point to the new files
    • also added direct/diffuse to solar
    • user_datm.streams/PI