[Go-essp-tech] grids in metafor, and the esg catalog

[Sylvia Murphy]

Bryan,

I have a question below...

On Nov 5, 2009, at 8:16 AM, Bryan Lawrence wrote:

> Hi Folks
>
> Balaji and I had a conversation yesterday, prompted by my barrage of  
> emails
> about grids which were themselves prompted by three things:
>
> 1) I had looked at what was in the spaceConfiguation of key  
> properties in the
> questionnaire, and
> 2) Phil had prompted me to compare a gridspec file and a CIM grid  
> xml file, and
> 3) In the go-essp-tech call on information flow, we made some  
> statements
> about code needed to handle grid descriptions ...
>
> This email is by way of summary of that conversation and of a  
> subsequent
> conversation with Phil (please both of you correct anything below  
> which
> is a bryan'ism).
>
> Before takeoffi:
>
> We had noted that in the previous version of the mindmaps (i.e.
> the one in the alpha8 running now), there were some horizontal grid
> definitions in the key properties of atmosphere and oceans, but
> they differed in good and bad ways...
>
> We expect gridspec files to be created at different times than the
> questionnaire (many by Balaji himself); so we can't expect them to  
> exist
> before q'naire input ... so we can't assume that we would create a CIM
> grid xml document  as output from the questionnaire by reading  
> gridspec
> to get (some) attributes.
>
> In flight:
>
> It became clear that we don't have any real use case (right now) for  
> mixing
> gridspec (which describes the grid the data is output on) and cim xml
> which is primarily about describing (amongst other things, the  
> resolution)
> of the model components for "understanding" and "discovery".
> E.g. "Which model ran with the highest atmopsheric resolution?
> (Never mind whether it wrote the data out at that resolution?) or
> "Find me all the models using a cube-sphere ..."
>
> The grid questions in atmopshere and ocean key properties were
> different, and very different from the other realms.
>
> We thought that the concept of resolution should be expanded
> to include three subtypes:
> - a short name using "accepted notatations" eg N96L52
> - a numeric resolution in degrees for the mid-latitudes (to allow  
> resolution ordering)
> - a textural comment on how the resolution varies (e.g. "higher  
> resolution
> in the tropics, with polar filtering").
>
> (But note the mixing of horizontal and vertial resolution in the  
> standard
> short name, so we should be careful about layout of horizontal
> and vertical resolution ... Phil and I had a chat about layout ...)
>
> Landing: Conclusions
>
> We should ask the same sort of grid questions in each realm, but
> - It should be acceptable to answer, say, in the case of aerosol,
> "same as atmosphere"
> - there might be realm level specifications.
>
> Currently, there is a risk of confusion between grids and components.
>
> Complete CIM XML *as currently defined* could only be created
> later, much later, and constructing them, then passing them to
> ESG would be nightmare. Better to modifiy the logical model
> for the CIM .... then we can export these to ESG, and then they can
> (if they want) handle grid definitions (from "newCIMgrid")  
> independently
> of gridspec harvested (if they want) from the data archives.
>
> Post flight: Actions Necessary:
>
> 1) Understand why the mindmap grid definitions have changed
> again.
>
> 2) Build a grid page for each realm level entry, and subsume the
> content from the mindmaps. (Plus some extra content from
> Balaji).
>
> (Why build  this? Because then we can more easily say "same as
> the atmosphere "....)
>
> Essentially come up with a new content model for a "simple" grid
> CIM.
>
> (Step 2 will probably take about a day of coding).
>
> 3) Remove the grid config stuff from the mindmaps.
>
> --- we can do stuff up to here before launch ----
>
> 4) Modify the ConCIM and ApCIM to conform to the new,
> simpler Grid definition ...
>
> 5) Rupert can then export it.
>
> --- We ought to be able to do that within a few weeks
> of step 2 ---
>
> 6) Sylvia can then load it.
>
> --- and Sylvia could tell us what the timescale implications of this  
> are ---


Bryan,  I am somewhat confused by the deliverable that you are  
composing here...

Are we talking about a separate grid XML sent to us sometime after the  
questionnaire?

At one point we were anticipating that GFDL would be writing a program  
to harvest the metadata contained within the gridspec files.  There  
was an open question about what form that output would be in.  Are we  
now talking about Metafor providing this info or is this a different  
set of grid info?

Thanks,

Sylvia




>
> Final question:
>
> Should we begin down this path? Or is there something we didn't know?
>
> Bryan
>
> -- 
> Bryan Lawrence
> Director of Environmental Archival and Associated Research
> (NCAS/British Atmospheric Data Centre and NCEO/NERC NEODC)
> STFC, Rutherford Appleton Laboratory
> Phone +44 1235 445012; Fax ... 5848;
> Web: home.badc.rl.ac.uk/lawrence
> _______________________________________________
> GO-ESSP-TECH mailing list
> GO-ESSP-TECH at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/go-essp-tech

-----------------------------------------------------
Sylvia Murphy

NESII Project Manager
NOAA/OAR/ESRL
sylvia.murphy at noaa.gov
303-497-7753