Register 4 new PMIP experiments

[durack1]

@charliepascoe 4 new PMIP experiments are currently in discussion at WCRP-CMIP/CMIP6_CVs#898.

Discussions are ongoing, however it is likely the registration will be completed later this week, early next week

Ping @taylor13 @davidhassell @martinjuckes @jypeter Ref cmip6dr/Request#123

[durack1]

@charliepascoe please note there are now 316 registered experiments for CMIP6, see CMIP6_experiment_id.html

[jypeter]

@charliepascoe let me know if I should provide extra information here, once @durack1 has fixed two copy/paste problems in the CV

[charliepascoe]

@jypeter I've added past1000-volc-cluster as an alias to the volc-cluster-mill experiment.

[charliepascoe]

Experiment info can be found at https://github.com/WCRP-CMIP/CMIP6_CVs/issues/898#issuecomment-613969369

[charliepascoe]

@jypeter Is the esm-past1000 experiment still called past1000-esm in the past1000 GMD paper? We can make the connection with either an "alias" (where an experiment is known by a different name elsewhere) or a "previous_name" (where an experiment was once called something else) depending on whether past1000-esm is still in use or not.

[charliepascoe]

@jypeter past1000 has the following forcing descriptions in the ES-DOC database:

past1000 WMGHG past1000 astronomical parameters pre-industrial ice sheets pre-industrial land-sea mask past1000 land use past1000 solar variability past1000 volcanic aerosols

For the -solaronly and -volconly experiments, should the rest of the forcings be set to a pre-industrial state?

[charliepascoe]

@jypeter Does the second half of the past2k experiment duplicate the past1000? Does it need to?

[charliepascoe]

Waiting for input from PI before uploading for publishing

[jypeter]

@charliepascoe

• Thanks for creating a past1000-volc-cluster alias (the original name in the GMD paper) pointing to past1000-volc-cluster (the existing CMIP6 experiment_id) ! I did not know about this aliasing feature
• I think we can use this feature a second time and also create a past1000esm alias (NO DASH, the original name in the GMD paper) pointing to esm-past1000 (the name recommended by Paul, in order to be consistent with other existing esm- experiment_ids). I don't think groups have already run this experiment, so we probably don't need to use _previousname
• I will ask Johann Jungclaus' advice (the PI) to check the list of documented forcings. Is there a web page that he can open to see how the existing forcings appear? I have tried to open https://view.es-doc.org/, but it seems to stay stuck on Downloading documentation ... please wait
• The second half of the past2k experiment is a past1000 experiment. What do you mean by Does it need to?

[charliepascoe]

@jypeter

• great, I'll use those aliases.
• try https://documentation.es-doc.org/cmip6/experiments/past1000
• I noticed that the past2k experiment runs from 0 to 1850, and that from 850 to 1850 past2k is very similar to (if not the same) as the past1000. I suspect my question is naive but why does past2k need to duplicate past1000?

[taylor13]

It seems to me if the second half of the past2k experiment is forced with the same forcing as the past1000 experiment, then in the data archive, the 2nd half should be stored with experiment_id="past1000" and its parent_experiment_id="past2k". Of course those running only the past1000 experiment would haveexperiment_id="past1000" and its parent_experiment_id="no parent" (according to the current CV).

So the metadata would indicate that different initialization procedures were used for different past1000 experiments, but I think most users would just like to get all the simulations of the past1000, and not have to get some of them by downloading a part of past2k experiment.

Of course other folks interested only in the runs covering the full 2k period would then have to download the past2k data plus the past1000 data for those models who had run the full 2000 years.

Thoughts?

[charliepascoe]

Hi @jypeter, I'm following up on the question of baseline forcings for the -solaronly and volconly experiments. Did you get a response from Johann Jungclaus yet?

[charliepascoe]

Hi @jypeter I'm following up on the -solaronly and volconly forcing specifications. Should the non-solar and non-volcanic forcings be set to pre-industrial conditions? If I don't hear back from you I will assume that these other forcings should be pre-industrial.

[charliepascoe]

I specified pre-industrial conditions for the non-volcanic and non-solar forcings.