User Tools

Site Tools


clipc:procedure

Procedure

(This is mainly a listing of the steps involved for myself. Scripts and templates: scripts.tar.gz)

  1. run ${VAR}/scripts/FULL_batch_bc_${VAR}_training.sh
  2. wait till jobs have finished
  3. run ${VAR}/scripts/FULL_batch_bc_${VAR}_application.sh
  4. wait till jobs have finished, check .out and .err files

  5. rename files according to CLIPC, BCIP and arcticdata standards: run ../scripts/BCIP_rename_${VAR}.sh #(in out directory)
  6. Edit NetCDF attributes according to CLIPC and BCIP standards: for file in *.nc;do ../scripts/BCIP_ncatted_${VAR}.sh $file;done #(in out directory)
  7. Do some simple check for reasonable values: for file in *.nc; do BCIP_basic_test_${VAR}.sh $file;done » ../test/basics.log &
  8. upload to arcticdata server: adjust ${VAR}/out/ftp_put.sh and run it

EVALUATION


  1. merge and make monthy sum/average for bias_corrected sims: run ${VAR}/test/merge_and_monmean_bc.sh resp. ${VAR}/test/merge_and_monsum_bc.sh
  2. make monthy sum/average for raw sims: run ${VAR}/test/monmean_raw.sh resp. ${VAR}/test/monsum.sh
  3. make monthy sum/average for obs files: cdo -r yseasmean -monsum -mulc,86400 ../in/${VAR}_EUR-11_SMHI-HIRLAM_reanalysis_v1d1-v1d2_SMHI-MESAN_v1_day_19890101-20101231.nc ${VAR}_EUR-11_SMHI-HIRLAM_reanalysis_v1d1-v1d2_SMHI-MESAN_v1_monsum_seas_19890101-20101231.nc
  4. make eval plots for all models: Rscript eval_plots_loop.R
  5. merge eval plots to one pdf: convert -geometry 2000×1200 bias*png bias_${VAR}_EUR-11.pdf; rm bias*png
  6. make taylor plot for with all models: Rscript eval_taylorplot.R
clipc/procedure.txt · Last modified: 2015-07-15 12:41:54 by andreasd