Ingest ESG metadata into mindmaps
This wiki is for documenting ticket 247
04/06/2009: summary of dynamic core reorganisation (Bryan's e-mail after 04/06 ESG DynCore special telco with Bryan, Sophie, Marie-Pierre)
This is my record of the outstanding issues from our telco on importing Curator dynamic core material.
For everyone's benefit, in general we found that we covered a lot of the material in what we had,
but Curator did have some useful vocabs for properties that we've added in (and in doing so done some
"controlling" - as we and Curator both knew that their input method didn'd do much "controlling"). We now
believe we have all the "dynamical core" properties in (for the moment) the Atmosphere.
There are some outstanding issues though, some of which arose from what we got from Curator, and some from what we already had:
1) In the advection scheme we need to allow multiple advection components, not just for "tracers" + "momentum".
It might be that there are different advection schemes used for different "tracers", and we thought it might
be helpful to clarify that "tracers" includes "species which are not conserved, because other components control their budget, even if they are advected "by" the dynamical core.
Of course other species are "advected" within their own components - either vertically or horizontally - and this is a coupling issue of sorts.
The bottom line is that we need to make sure there is some clarity here, that we haven't quite got. - not quite sure who got the action on that :-)
2) The controlled vocabularies we currently have for top boundary condition and lateral boundary conditions (for RCMs) are not yet adequate.
- Bryan to consider the issue and potentially add values to mind-map.
3) There are some issues in the discretisation vocabularies that need settling, to do with overlap between terms e,g. Galerkin, finite-element, and spectral decomposition etc ...
- Bryan to consider the issue and propose modifications to mind-map.
4) There are a number of "reference names" in the Curator vocab for advection schemes. Many of those names are in fact names of discretisation method. We need to:
- comb the literature a bit to get a reasonable set of "proper name options" for well known advection schemes (Bryan, who used to do this sort of thing for a living),
- take the "reference names" that are in the Curator vocab (littered throughout), and
-- get a list of them (Marie-Pierre)
-- convert those names to "questionnaire format" references in csv (someone else, Gerry maybe?), and
-- load them into the questionnaire so that folks have some pre-canned references to choose from (Bryan from
a csv list)
5) There are a number of "grid type keywords" covering grid staggering etc. We've currently pushed those
to a "we think these are in gridspec bucket".
- at some point someone should check. Phil maybe?
6) There are a number of issues around what and how conservation is done in the advection scheme, and how we promote the conservation properties of the dynamical core.
- Bryan to work through what we have and resolve.