687-WORKSHOP-viirs-hdf5

[AJThorpe]

Reviewer Checklist

PR author: Please ensure you meet all of the below requirements, and check boxes appropriately.

Reviewers: Please confirm all required testing/documentation has been completed prior to approving.

Remove lines that are not applicable, explain if you select "NO REQUIRED"

• Required existing tests pass (ie full_test.sh, others as appropriate)
• NO REQUIRED existing tests (explain why not required)
• Required unit tests added and pass for new/modified functionality
• NO REQUIRED unit tests (explain why not required)
• Required integration tests added and pass for new/modified functionality
• NO REQUIRED integration tests (explain why not required)
• Required documentation added for new/modified functionality
• NO REQUIRED documentation (explain why not required)
• Required release notes added for new/modified functionality
• NO REQUIRED release notes (explain why not required)
• Required updates to other repos complete
• NO REQUIRED updates to other repos (explain why not required)

https://github.com/NRLMMD-GEOIPS/.github/blob/main/.github/review-template.md

Related Issues

• Creates VIIRS SDR HDF5 reader for geoips workshop

Testing Instructions

Summary

Output

[AJThorpe]

@mindyls @jsolbrig @cpcamacho

Why are there corrections done to the Visible products for VIIRS netcdf files? When looking at two files from the same location, and looking at the visible bands I get separate values for the SDR and the netcdf files: (hdf_xrt is SDR files with viirs_sdr_hdf5 and nc_xr is Netcdf files opened with viirs_netcdf)

In [21]: hdf_xrt['MOD-Vis'].M05Ref.max(), hdf_xrt['MOD-Vis'].M05Ref.min(), hdf_xrt['MOD-Vis'].M05Ref.mean()
Out[21]: 
(<xarray.DataArray 'M05Ref' ()>
 array(121.8136, dtype=float32),
 <xarray.DataArray 'M05Ref' ()>
 array(2.361164, dtype=float32),
 <xarray.DataArray 'M05Ref' ()>
 array(28.562603, dtype=float32))

In [22]: nc_xr['MOD-Vis'].M05Ref.max(), nc_xr['MOD-Vis'].M05Ref.min(), nc_xr['MOD-Vis'].M05Ref.mean()
Out[22]: 
(<xarray.DataArray 'M05Ref' ()>
 array(1.2865295, dtype=float32),
 <xarray.DataArray 'M05Ref' ()>
 array(0.02233079, dtype=float32),
 <xarray.DataArray 'M05Ref' ()>
 array(0.25161928, dtype=float32))

For the visible product there are several data manipulations done; a scale factor of 100, gamma correction of 1.5, and a Sun Zenith correction. Do we have to apply these to the Visible bands of the SDR files? Or is there a paper discussing when this is needed?

I think we have two clear options to take:

• Figure out what differences are between the radiances & reflectance for the visible and infrared bands between netcdf and SDR files (are netcdf L2 or L1b?) and correct accordingly.

OR

• Create separate products for VIIRS SDR files (not preferred for transition reasons, web products would be confusing, etc....)

[AJThorpe]

Per Mindy:

• Leave it at 0-100 range (we'll do 0-100 for reflectance and 0-1 for albedo, and this is ref)
• Apply solar zenith angle correction in the hdf5 reader to the visible reflectance
• Remove the gamma correction

[AJThorpe]

VIIRS radiance/reflectance correction will be done later! See issue #733