CCPP Framework Meeting Minutes 2022 12 20

Agenda

• GitHub issues
  • notes
• Updates
• Other business
  • notes

---

Attendees: Mike Kavulich, Dustin Swales, Ligia Bernardet, Cheryl Craig, Grant Firl, Jordan Powers, Dom Heinzeller, Matthew Dawson, Lulin Xue, Courtney Peverley

GitHub issues

• 457: ccpp-framework not working with Python 3.11
  • Fix in PR 456; still not merged.
    • Will go in with fv3atm PR 611? Yes, no problems. Will be merged soon.

GitHub notes

Updates from last time

Previous notes

Other business

• Mike K. gave talk at AGU last week on CCPP Framework (https://agu.confex.com/agu/fm22/meetingapp.cgi/Paper/1184406)

  • Looks like videos are not posted yet, but those curious can see slides here: https://docs.google.com/presentation/d/1Yqaqo9QTJBlC2_1Rf6YTOJFmD1gA7roZ/edit#slide=id.p1

• Using abstract data types

  • ACOM group is bringing some aerosol using abstract data types (as compared to Fortran DDTs)
  • As we move forward, if abstract data types are used by aerosol community, it can become available in the CCPP Framework (it would be as usable as Fortran intrinsic types are today)
  • It would have methods as well
  • Github.com/NCAR/demo-radiation created by Steve G./Matt/ACOM which would be a good reference
  • Abstract data types would help interoperability for CCPP physics, esp. interstitials
  • But we need to be careful bcs abstract data types may end up being customized code for specific groups
  • Community would have to define the abstract data types that are needed for a community
  • Francis is implementing an abstract data type specific for CAM so that MAM and KARMA can be used interchangeably in CAM
  • DDTs accessible from the host? Or DDTs constructed by the Framework? The former hinders interoperability.
  • One possibility is to have a bunch of abstract types, e.g., for clouds. It goes through suite. Radiation then can query this type. This allows extending basic information so that a simple scheme can have simple information, while a complex scheme can have extended information.
  • Fortran 2003 is needed (Fortran 90 will not support this)

• Cheryl is not so optimistic about Steve G. joining the CCPP-Framework effort in the near future. May be just on a limited fashion.

• CIRA CCPP Framework job is posted: https://jobs.colostate.edu/postings/117554

• We need to keep an eye on this phys-dyn coupling discussion: https://github.com/NOAA-GFDL/GFDL_atmos_cubed_sphere/issues/230

  • Would returning tendencies help what Kevin is trying to do?
  • CESM/CAM requires returning tendencies

• CCPP Framework paper is moving through the review process: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-855/#discussion