C3S_34e Meeting with Ouranos and CRIM: 30th September 2021

[agstephens]

Agenda - C3S_34e Meeting with Ouranos and CRIM: 30th September 2021

1. Minutes and actions from previous meeting

Previous minutes: #103

Previous actions

• TS: Share details of the segfault problem seen when using clisops in raven WPS.
• AS: Run the cf-checker on example files generated by Xarray to see if 1D coordinate variables with _FillValue set will fail cf-checks. Do we need to take action?

2. Project update: c3s_34e [Ag]

• CF-Checking NetCDF data in which coordinates and coordinate bounds have a _FillValue set (but no missing data):

  • this is/was default for some versions of xarray
  • cf-checker says:
    • nothing about coordinates with _FillValue attr
    • warning about coordinate bounds with _FillValue attr
    • WARN: (7.1): Boundary Variable lon_bnds should not have _FillValue attribute

• fixes for decadal (DCPP) runs for C3S

  • will go into production in October 2021

• clisops updates:

  • some CI errors resolved in PRs:
    • dask delayed errors in averaging
    • intake-related issues with dask/intake/fsspec versions
  • Martin looking at regridding of unstructured grids:
    • possible bug in ESMF/xESMF with masks

• STAC meeting with CMIP6 in Cloud meetings:

  • would be great to collaborate with Ouranos
  • starting to work on the management of controlled vocabularies for faceted search

• funding future for C3S:

  • still unknown!

3. Project updates: relevant Ouranos/CRIM projects [David/Tom]

Ouranos (DH):

CRIM (TL):

4. Discussion:

• Is the Raven install bug still an issue?

5. Agree next steps and next meeting

Next meeting date: Thursday 21st October (pushed back 1 week)

[agstephens]

Minutes - C3S_34e Meeting with Ouranos and CRIM: 30th September 2021

1. Minutes and actions from previous meeting

Previous minutes: #103

Previous actions

• TS: Share details of the segfault problem seen when using clisops in raven WPS.
  • See: roocs/clisops@50a1094
• AS: Run the cf-checker on example files generated by Xarray to see if 1D coordinate variables with _FillValue set will fail cf-checks. Do we need to take action?

2. Project update: c3s_34e [Ag]

• CF-Checking NetCDF data in which coordinates and coordinate bounds have a _FillValue set (but no missing data):

  • this is/was default for some versions of xarray
  • cf-checker says:
    • nothing about coordinates with _FillValue attr
    • warning about coordinate bounds with _FillValue attr
    • WARN: (7.1): Boundary Variable lon_bnds should not have _FillValue attribute

• fixes for decadal (DCPP) runs for C3S

  • will go into production in October 2021

• clisops updates:

  • some CI errors resolved in PRs:
    • dask delayed arrays errors in averaging
    • intake-related issues with dask/intake/fsspec versions
  • Martin looking at regridding of unstructured grids:
    • possible bug in ESMF/xESMF with masks
    • when calling the regridder: any variables that cannot be regridded are not changed:
      • could be fixed/treated in xESMF
  • Time and Level updates:
    • See: Create "subset-by-point" capability #105

• STAC meeting with CMIP6 in Cloud meetings:

  • would be great to collaborate with Ouranos
  • starting to work on the management of controlled vocabularies for faceted search

• funding future for C3S:

  • still unknown!

3. Project updates: relevant Ouranos/CRIM projects [David/Tom]

Ouranos (DH):

• Looked at STAC work at CEDA:
  • passed to CRIM
  • hope to add catalog creation logic to Richard's code
• Would be good for CEDA to look at adding Merdian-fix for:
  • Subset before averaging clisops#189
  • Plan to do some benchmarking for improvements to performance

4. Discussion:

5. Agree next steps and next meeting

Next meeting date: Thursday 21st October (pushed back 1 week)