Skip to content

CCPP Framework Meeting Minutes 2020 03 12

Jump to bottom
Dom Heinzeller edited this page Mar 12, 2020 · 14 revisions

Agenda

  • Created new Standard Name repository, CCPPStandardNames, which includes code to parse the database and generate output. The repo also includes a generated Markdown file (Markdown is only currently-supported output format).
  • Speaking of standard_names, GFS physics uses geopotential for a quantity that is really gepotential_per_unit_mass. Thoughts on changing it? Has anyone ever seen an actual geopotential quantity pass between schemes?
  • What are feelings about having a flag in the <scheme> tag to optionally call a scheme? A possible syntax could be:
  <scheme run_if=run_var_standard_name>scheme_name</scheme>

       Another proposed syntax would be to add a new tag which could surround one or more schemes:

  <execute test="run_var_standard_name" default=".false." >
    <scheme>scheme_name</scheme>
  </execute>

       This adds complexity and uncertainty (is this scheme run or not?) to the suite definition file but may prevent a combinatoric explosion of SDFs. Thoughts?

  • We have another issue where a scheme needs to be inserted in several places in a suite but with different input values. An example is a diagnostic scheme which needs a different name to distinguish where these diagnostics are being produced. Does anyone else have anything like that?

  • Update needed to Wiki? Wiki is mostly good. I would like to know about the page on Development and the nemsfv3gfs tests. Is this information up to date? If not, can we do a quick fix this week?

  • Update needed to README.md file. Outdated. Since we already tagged the code for the upcoming v4 release, I suggest a post-release update. I don't think I (Ligia) am qualified for this. Can we identify a person/persons and put this on a to-do list, not super urgent, just get it done?

Meeting notes

  • CCPPStandardNames because of ESCOMP standards; would be better to use ccpp-standard-names

    • this repo may be temporary, discussion around combining it with ESMF field dictionary
    • need admin rights for DTC CCPP managers (Laurie, Grant, Dom)
    • we all need to work on this, fix bugs and loopholes, add new variables
    • EMC developers have asked for such a "database" of standardd names and rules to derive new ones
    • need to involve scientists from various organizations in the development of this database
    • geopotential: don't change anything, assume convention geopotential = geopotential_per_unit_mass in meteorology

  • Transition to cap_gen.py at DTC/EMC:

    • this will likely be delayed and trickle into the next period of performance
    • several steps need to be taken for the UFS models in preparation for the transition: remove dynamic build, clean up how derived data types are defined, ... - Dom will be working on these in the next weeks to months

  • Suite definition file proliferation:

    • the current philosophy is "what is in the SDF gets called, nothing else" --> clear
    • try to reduce the number of suite definition files in the authoritative repositories rather than implementing something complicated right now
    • how to handle calling schemes twice with different input data in the suite definition file? best to use interstitial schemes for now

  • Subgrid column / partition:

    • use for subgrid column physics calls
    • use for fractional landmask

CCPP Framework wiki

Clone this wiki locally