Generate temporal high-resolution fields for the wind nudging

Strategy

• We compute the high-frequency variability, typically 6-hourly variability from a NorESM1 run. This high-frequency (HF_clim) does not contain the monthly variability.
• We compute the monthly climatology (LF_clim) for the variables we want to nudge from the same simulation.
  • We interpolate onto the 6-hourly resolution using a spline interpolation
• We compute the monthly fields (LF_ano) from the product we want to nudge (typically a reanalysis).
  • We interpolate onto the 6-hourly resolution using a spline interpolation
• Then we sum LF_ano, HF_clim and LF_clim

Main folder for the algorithm:

/cluster/projects/nn9039k/people/dalaiden/nudging/combine_HF_LF

Generate HF_clim and LF_clim

The first thing we need to do is to cat the daily files into a monthly files, using this tool:

/cluster/projects/nn9039k/people/dalaiden/NorCPM/tools/cat_daily_cam_files.sh

Then, the main script to generate HF_clim and LF_clim is compute_HF_plus_clim.sh. This script calls the python script make_highres_variability_and_clim.py and cat_HF_all-vars_final_files.py.

Several inputs are needed, which are given in compute_HF_plus_clim.sh:

# Parameters
period_full="1961-1990" # The period we want to export the high-frequency variability
period_clim="1961-1990" # The period for the climatology
indir_simu="/cluster/work/users/dalaiden/archive" # Simulation folder
case_id="past500y.hist.001" # Case ID
outdir="/cluster/work/users/dalaiden/nudging_work" # Folder for the outputs
ncpus=9
#----------------------

This script will generate HF_clim and LF_clim; one file per month including these two variables.

Configuration, resources needed

#SBATCH --account=nn9039k --job-name=HF
#SBATCH --partition=preproc
#SBATCH --time=04:00:00
#SBATCH --ntasks=9
#SBATCH --mem-per-cpu=23G
#SBATCH --error=HF_log_%j
#SBATCH --output=HF_out_%j

2h30 for 30 years.

Generate monthly anomalies from the simulation we want to nudge

Script: job_make_spline_monthy_reanalysis.sh which calls make_spline_monthy_reanalysis.py

Basically here, we remove the monthly climatology and interpolate the monthly anomalies at 6-hourly resolution

Configuration, resources needed

#SBATCH --account=nn9039k --job-name=spline_rea
#SBATCH --partition=preproc
#SBATCH --time=24:00:00
#SBATCH --ntasks=1
#SBATCH --mem-per-cpu=40G
#SBATCH --error=spline_rea_log_%j
#SBATCH --output=spline_rea_out_%j

For one member, 60 years: 2:40.

Generate the final files for the nudging at 6-hourly resolution

Script: generate_fnudging_rea.sh.

We sum everything.

Configuration, resources needed

#SBATCH --account=nn9039k --job-name=HF_nudging
#SBATCH --partition=preproc
#SBATCH --time=24:00:00
#SBATCH --ntasks=1 --cpus-per-task=1
#SBATCH --mem-per-cpu=3G
#SBATCH --error=HF_log_%j
#SBATCH --output=HF_out_%j

For one member, one year: 8 minutes.

Results when nudging monthly fields (twin experiments)