[Go-essp-tech] ***CHANGE***: Connection info for GO-ESSP call on Tues 10/28/14
Karl Taylor
taylor13 at llnl.gov
Tue Oct 28 15:07:34 MDT 2014
Hi again,
A correction on my previous email:
I may have been confused when I said some of an endorsed-MIPs model runs
might be excluded from CMIP6. At the recent WGCM meeting we agreed to
define tiers of simulations within each endorsed-MIP, but I think
Martin's right that all of an endorsed-MIP's runs will be included in
CMIP6. So, if this is correct, the use case is slightly simpler than
stated: we just want the experiments contributed by an endorsed-MIP to
appear under that MIP's project name, but also under the CMIP6 project name.
[By the way the tiering of runs within each MIP is meant to help
modeling groups prioritize which experiments they want to perform. If
they want to contribute to an endorsed MIP, they will be expected to at
least perform all the experiments in the first tier (but they can choose
to participate in only some subset of the endorsed-MIPs). (There is
actually a slight caveat on this, which I won't bother to explain at
this point.)]
best regards,
Karl
On 10/28/14, 1:27 PM, martin.juckes at stfc.ac.uk wrote:
> Hello Karl,
>
> I was puzzled by your reference to MIPs doing some experiments "in" CMIP6 and other experiments "not in" CMIP6: I thought the idea was that once a MIP had been suitably blessed as part of the CMIP6 extended family all the experiments within that MIP would have the same status. Is this idea of having multiple classes of experiment spelt out somewhere?
>
> cheers,
> Martin
>
>
>
> ________________________________
> From: Cinquini, Luca (398G) [Luca.Cinquini at jpl.nasa.gov]
> Sent: 28 October 2014 19:55
> To: Karl Taylor
> Cc: wgcm_wip at wcrp-climate.org; go-essp-tech at ucar.edu; cmip-panel at wcrp-climate.org
> Subject: Re: [Go-essp-tech] ***CHANGE***: Connection info for GO-ESSP call on Tues 10/28/14
>
> Hi Karl,
> thanks for attending today, and for your continuos interest and suggestions for improvements. It might be worth discussing these use cases at our next WIP telco, but in the meantime here’s a few quick considerations:
>
>
> On Oct 28, 2014, at 12:14 PM, Karl Taylor <taylor13 at llnl.gov<mailto:taylor13 at llnl.gov>> wrote:
>
> Hi Luca and all,
>
> Thanks for the helpful demo this morning on the Go-ESSP tech telecon. The interface appears to be versatile and the search seems quite a bit better than what is now in place. I was particularly pleased to see your new option to make multiple selections under each search facet. I was one of the folks who were hoping to be able to do this, so I appreciate you making that a priority.
>
> Of course, the work is never done, and additional features will always be requested. Here are two search use cases that might be candidates for future work. (Before settling on how to proceed, however, I think it would be good to give the WIP a chance to refine the ideas and provide guidance, along with indicating whether this is a priority.)
>
> Case 1: I want to include in my search all the CMIP6 simulations, but no others.
> This would appear to be straightforward, but CMIP6 is being organized in a "distributed" way now, and it really comprises a subset of the experiments that have been performed as part of a number of coordinated, but quasi-independent contributing MIPs. The individual MIPs plan to perform some simulations which will be included in CMIP6, but others that are not. I think that each of the MIPs will publish all of their simulations under their own project name. For example, PMIP might publish an lgm, midHolocene, and past1000 simulations as "PMIP" project runs, but perhaps only the lgm run is included in the CMIP6 project. Another project, "xMIP" might publish results from 5 expts., but only 2 of these (say "exp1" and "exp2") are included in CMIP6. Suppose no experiments are originally published with the "CMIP6" project identifier (but instead are published under the name of the MIP directly responsible). Is there a way for us to construct a virtual project that includes only lgm and exp1 and exp2 (but not midHolocen, or past1000, or exp3, 4, or 5)? A few options to explore are:
> 1) publish the CMIP6 runs twice (i.e., both as belonging to its host project but also CMIP6).
> 2) publish the CMIP6 runs only once, but somehow identify them with both the host project and CMIP6
> 3) build on the concept of a "family" of experiments, which is part of the current ESGF search capability, and use it to collect different experiments into a single group.
> The point is to find a way to select and list all the CMIP6 simulations, but none of the other experiments.
>
> The above use case is certainly feasible, without a lot of additional development. Suggestion 1) above won’t really work because you either create duplicate copies of the data (if you change the ids), or you override the previous copy. But suggestions 2) and 3) can both be implemented by either submitting records with multiple metadata fields ("project=X and project=Y"), or by executing a little server-side processing (“if experiment=X then add experiment_family=Y” - we actually do this already for CMIP5).
>
>
> Case 2 (discussed briefly on the telecon this morning): I want to retrieve files containing output from only those models that have carried out *all* experiments and *all* variables of interest.
> For example, if I want to analyze correlations between temperature and precipitation differences between the historical and control run, then I want to exclude from my search models that didn't report *both* of these variables for *both* of these runs.
>
> This use case is more difficult because the search always targets datasets, and it would be wrong to tag a single dataset with multiple values of an experiment or a variable that it does not include. Without thinking it over too much, I think this use case could be enabled by some more sophisticated search workflow such as the following:
>
> a) Loop over all possible (experiment, variable) and find list of models that produced that combination
> b) Get the intersection of model results form step a). These are the models that produced both variables for both experiments
> c) Loop over all the (model, experiment, variable) combinations and download the files
>
> Not easy but it is possible, I guess…. Is the above right ?
>
> thanks again,
> Luca
>
>
> Something to begin thinking about, but perhaps waiting to act on until we get further input from others.
>
> Best regards,
> Karl
>
> On 10/27/14, 3:31 PM, Cinquini, Luca (398G) wrote:
> Hi all,
> please see this connection information from Cecelia regarding the GO-ESSP meeting tomorrow.
> thanks, Luca
>
> Dear all,
>
> The web share and call-in number for tomorrow's demonstration
> have changed. Please use:
>
> **Phone line:**
> US: 1-866-692-3582, 3443874#
> Additional numbers are below.
>
> **Web share:**
> https://www.mymeetings.com/emeet/join/index.jsp?customHeader=mymeetings&netId=741490701&netPass=9270604&netType=conference/en&acceptTerms=on&sigKey=mymeetings
> Please fill in your name and other info, check the Privacy Policy box, and click "Proceed."
> If you have trouble with this link, alternate instructions are at the very end of this mail.
>
> Best regards,
> Cecelia
>
> **INTERNATIONAL NUMBERS**
> ARGENTINA 0800-777-0515 (toll free)
>
> AUSTRALIA 1-800-659-380 (toll free)
> ADELAIDE +61-8-8121-4883
> BRISBANE +61-7-3102-0985
> CANBERRA +61-2-6100-1985
> MELBOURNE +61-3-9010-7754
> PERTH +61-8-9467-5264
> SYDNEY +61-2-8205-8123
>
> AUSTRIA +43-1-92-81-120, 0800-005-266 (toll free)
> BELGIUM +32-2-400-9824, 0800-3-8744 (toll free)
> CHILE 1230-020-2638 (toll free)
>
> CHINA +86-400-810-4783
> CHINA B 10800-120-1677 (toll free)
> CHINA A 10800-712-1677 (toll free)
>
> COLOMBIA 01800-9-156477 (toll free)
> CZECH REPUBLIC +420-2-25-98-56-73 800-700-235 (toll free)
> DENMARK +45-7014-0300, 8088-8304 (toll free)
> ESTONIA 800-011-1129 (toll free)
> FINLAND +358-9-5424-7178, 0-800-9-14607 (toll free)
>
> FRANCE 080-563-6110 (toll free)
> LYON +33-4-26-69-12-99
> MARSEILLE +33-4-86-06-00-99
> PARIS +33-1-76-74-24-94
>
> GERMANY +49-69-2222-4873, 0800-664-4254 (toll free)
> GREECE +30-80-1-100-0698 , 00800-12-7310 (toll free)
> HONG KONG +852-3001-3859, 800-933-921 (toll free)
> HUNGARY 06-800-13830 (toll free)
>
> INDIA B 000-800-001-6236 (toll free)
> INDIA A 000-800-852-1264 (toll free)
> INDIA C 1800-300-00439 (toll free)
>
> INDONESIA 001-803-011-3973 (toll free)
> IRELAND +353-1-247-5678, 1800-992-935 (toll free)
> ISRAEL 1-80-9216159 (toll free)
>
> ITALY 800-986-993 (toll free)
> MILAN +39-02-3600-6016
>
> JAPAN 0066-33-132424 (toll free)
> OSAKA +81-6-7739-4783
> TOKYO +81-3-5539-5187
>
> LATVIA 8000-3063 (toll free)
> LUXEMBOURG +352-27-000-1384
> MALAYSIA 1-800-81-3069 (toll free)
> MEXICO 001-866-295-6360 (toll free)
> NETHERLANDS +31-20-718-8596, 0800-023-1812 (toll free)
> NEW ZEALAND +64-9-970-4767, 0800-450-789 (toll free)
> NORWAY +47-21-590-072, 800-15308 (toll free)
> PANAMA 011-001-800-5072111, (toll free)
> PERU 0800-53721 (toll free)
> PHILIPPINES +63-2-858-3753
> POLAND 00-800-1212569 (toll free)
> PORTUGAL 8008-14054 (toll free)
> RUSSIA 8-10-8002-0114011 (toll free)
> SINGAPORE +65-6883-9226, 800-120-4675 (toll free)
> SLOVAK REPUBLIC +421-2-322-422-39
> SOUTH AFRICA 080-09-80424 (toll free)
> SOUTH KOREA +82-2-6744-1079, 00798-14800-7189 (toll free)
> SPAIN +34-91-414-25-40, 800-300-060 (toll free)
> SWEDEN +46-8-566-19-397, 0200-884-614 (toll free)
> SWITZERLAND +41-44-580-6402, 0800-120-039 (toll free)
> TAIWAN +886-2-2795-7375, 00801-137-796 (toll free)
> THAILAND 001-800-1206-66050 (toll free)
>
> UNITED KINGDOM 0808-238-6019 (toll free)
> BIRMINGHAM +44-121-210-9039
> GLASGOW +44-141-202-3239
> LEEDS +44-113-301-2139
> LONDON +44-20-3043-2499
> MANCHESTER +44-161-601-1439
>
> URUGUAY 000-413-598-3460 (toll free)
> USA +1-517-466-2232, 866-692-3582 (toll free)
> VENEZUELA 0800-1-00-3782 (toll free)
>
> **ALTERNATE WEB SHARE ACCESS**
> 1) Go to: mymeetings.com<http://mymeetings.com/>
> 2) On the right side of the bottom box is a blue header called "Participant Join".
> Underneath this label is a link: "Join an Event". Click on this.
> 3) Meeting Number is: 741490701
> 4) Meeting Passcode is: 9270604
> 5) Enter your personal information and check the box related to the Privacy Policy.
> 6) Click "Proceed".
>
>
> --
> ===================================================================
> Cecelia DeLuca
> NESII/CIRES/NOAA Earth System Research Laboratory
> 325 Broadway, Boulder 80305-337
> Email: cecelia.deluca at noaa.gov<mailto:cecelia.deluca at noaa.gov>
> Phone: 303-497-3604
>
>
>
>
> _______________________________________________
> GO-ESSP-TECH mailing list
> GO-ESSP-TECH at ucar.edu<mailto:GO-ESSP-TECH at ucar.edu>
> http://mailman.ucar.edu/mailman/listinfo/go-essp-tech
>
>
> _______________________________________________
> GO-ESSP-TECH mailing list
> GO-ESSP-TECH at ucar.edu<mailto:GO-ESSP-TECH at ucar.edu>
> http://mailman.ucar.edu/mailman/listinfo/go-essp-tech
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/go-essp-tech/attachments/20141028/270a2e1e/attachment.html
More information about the GO-ESSP-TECH
mailing list