Ticket #683 (assigned Task)

Opened 4 years ago

Last modified 4 years ago

Document TAMIP experiments in questionnaire

Reported by: charlotte Owned by: charlotte
Priority: critical Milestone: V1.2 Questionnaire release
Component: WP6 - CMIP5 Questionnaire Version: 1.0
Keywords: Cc: gerry, mark, rupert
Requirement:

Description (last modified by charlotte) (diff)

TAMIP (Transpose AMIP) experiments are likely to be included with CMIP5 so we need to document them in the questionnaire.  http://hadobs.metoffice.com/tamip/

We're using ticket #684 to establish how we document the TAMIP experiments (and 1.1 DecadalHindcasts?).

Attachments

tamip.xml Download (5.7 KB) - added by charlotte 4 years ago.
TAMIP experiment requirements - from Keith Williams

Change History

comment:1 Changed 4 years ago by charlotte

  • Status changed from new to assigned
  • Requirement modified (diff)

Mark Elkington Wrote: 
 The TAMIP folks are requesting that PCMDI add 64 new experiments to CMIP5.  These experiments are identical 5 day hindcasts with a different starting point (i.e. start and end date/times are the only difference).  The time resolution for the starting point on each 'experiment' is hours not days.

This means that each TAMIP 'experiment' has to have its own set of experiment requirements (ok - just a simple cut and paste when setting up the requirements xml). But it means that the cmip5Q's list of experiments will have a lot more TAMIP experiments than CMIP5 experiments, and the organisations involved in TAMIP will have to do an identical conformance process for each of 64 experiments. Not sure what the impact on Curator would be - I assume they could somehow partition the TAMIP and CMIP5 information to make the interface look OK

My thought was that we could have a single experiment and single simulation with 64 ensemble members, but its not really an ensemble in the normal sense unless we can somehow define the period covered as an 'initialisation method'.

Second thought was we could have a single experiment with multiple simulations. We can specify the start and end of each simulation in the cmip5Q. The simulation duration is not reflected directly in the DRS, except that in the case of these very short runs we can probably make the assumption that an atomic dataset will be in a single file and the DRS <time-period> would record the different simulation period. Its not clear to me that the DRS supports hours in the time period however.

Any other solutions I haven't considered

The request to change the MIP tables/CMOR to support the 64 TAMIP experiments went to Karl yesterday. If we want to consider options other than having a list of 64 TAMIP experiments in the questionnaire we probably need to get something to Karl in the next day or so. It just seems wrong to me from a modelling perspective that these are considered as separate experiments when it is a single experiment with some related simulations.

I don't think the TAMIP guys care one way or another - they are just following recommendations from PCMDI.

comment:2 Changed 4 years ago by charlotte

  • Description modified (diff)

comment:3 Changed 4 years ago by charlotte

The TAMIP experiment design can be found  here. This web page lists all the TAMIP experiment requirements.

Changed 4 years ago by charlotte

TAMIP experiment requirements - from Keith Williams

comment:4 Changed 4 years ago by charlotte

  • Cc gerry, mark added; gerry mark removed

Our methodology for creating the experiment requirement documents has been to deal with only one thing in each numerical requirement. If we continue this methodology for TAMIP it would mean separating out the numerical requirements into separate entities. So the example below would be separated into two numerical requirements – the cludgy solution.

<numericalRequirement xsi:type="InitialCondition"> 
            <id xsi:type="Identifier">ic.010</id>
            <name>ic.non_stat_variables</name>
            <description>Model non-state variable prognostics to be initialised from zero, or initialised using the nudging method of Boyle et al. (2005). (Please specify approach used).</description>
</numericalRequirement>

becomes

<numericalRequirement xsi:type="InitialCondition"> 
            <id xsi:type="Identifier">ic.010a</id>
            <name>ic.non_stat_variables_zeroInit</name>
            <description>Model non-state variable prognostics to be initialised from zero.</description>
</numericalRequirement>
<numericalRequirement xsi:type="InitialCondition"> 
            <id xsi:type="Identifier">ic.010b</id>
            <name>ic.non_stat_variables_nudged</name>
            <description>Model non-state variable prognostics to be initialised using the nudging method of Boyle et al. (2005). </description>
</numericalRequirement>

I think it would be better if we could utilise the “How this simulation conformed” drop down box to include the two possible initialisation strategies so that this example can be kept as a single requirement.

Gerry – What are your thoughts on this, do you think it will be possible?

If it is possible we could use the additions to the drop down menus to make the conformance pages shorter (more elegant and less daunting to users) and also clear up things like the emissions vs concentrations ambiguity too.

Cludgy will work but elegant would be better.

comment:5 Changed 4 years ago by charlotte

  • Cc rupert added

I added Rupert to the cc list because if we decide to change the way we do conformances it will mean Rupert will have to adjust the way the questionnaire outputs conformance information.

comment:6 Changed 4 years ago by charlotte

Proposed naming convention for TAMIP: TAMIP This wiki also describes how staggerd start ensembles can be applied to the decadal hindcast/forecast experiments

Note: See TracTickets for help on using tickets.