CAM reanalysis bias correction for CLM-DART

braczka commented 1 year ago

Description:

Scripting has been added that corrects the meteorological bias in the CAM reanalysis forcing when performing a site-level assimilation with CLM-DART. This bias is rooted in the spatial mismatch between a CAM grid cell and the spatial coverage of a land site (typically 1x1 km2). The scripting includes nco commands that will cut out the CAM grid cell collocated with the land site of interest. Subsequent scripts perform the bias correction, while still maintaining necessary ensemble spread to generate model error covariance in CLM.

Fixes issue

This PR addresses issue #446 CAM reanalysis (site-level) bias correction tool

Types of changes

[ ] Bug fix (non-breaking change which fixes an issue)
[X] New feature (non-breaking change which adds functionality)
[ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
[ ] Documentation update

Documentation changes needed?

[X] My change requires a change to the documentation.
- [ ] I have updated the documentation accordingly.

I have provided some brief documentation as comments within the scripts themselves, and also a README file as part of the PR. Eventually I will add web-based documentation, but I still want to complete more testing before fully advertising the code.

Tests

Please describe any tests you ran to verify your changes.

These scripts have undergone significant diagnostic checks (some diagnostic figures are included in the PR), and also have been successfully run with several site level ensemble free runs. For example see CLM-DART case run: /glade/work/bmraczka/cases/cesm2.2.0/clm5_CAM4_BGC_analysis_adjusted_e80 or case run: /glade/work/bmraczka/cases/cesm2.2.0/clm5_NR1_BGC_analysis_adjusted_e80. The bias correction code still needs additional testing with an assimilation run, but the code is ready to receive software review for improvements/streamlining.

Checklist for merging

[x] Updated changelog entry
[ ] Documentation updated
[x] Update conf.py

Checklist for release

[ ] Merge into main
[ ] Create release from the main branch with appropriate tag
[ ] Delete feature-branch

Testing Datasets

[x] Dataset needed for testing available upon request
[ ] Dataset download instructions included
[ ] No dataset needed

The scripts include paths for an example run for flux tower site US-NR1.

For example:

% Input data files
 path_CAM_forcing = '/glade/work/bmraczka/CAM6_NR1/';
 path_towermet_forcing = '/glade/work/bmraczka/SIFMIP2/tower_met_forcing/';

braczka commented 1 year ago

I pulled off the draft status so it can be reviewed again. I think I addressed most comments in my updates, but please let me know if things still need to be addressed or better streamlined. No rush on merging this into main branch of DART. I am still actively working with site level land DA work which uses this ensemble met forcing as a driver. The code itself is fairly well commented, and I plan to forgo formal DART webpage documentation at this time, until I kick the wheels a bit more on this.

hkershaw-brown commented 1 year ago

@braczka This is is looking much better. Do you mind I go ahead and tidy up the indenting? Just to clear up the comments, see what is left from Kevin (any dealbreakers). Otherwise I think this is good to go.

braczka commented 1 year ago

@hkershaw-brown That would be great -- I have been swamped to this week, so feel free to correct indenting, check on Kevin's comments etc. Very much appreciate your feedback @kdraeder

NCAR / DART