Register LS3MIP data contribution from IIS-UTokyo

[durack1]

@gkrinner here is a start for getting the details for the LS3MIP data registered so we can get the data written to specification and published in input4MIPs. Can you please entrain Hyungjun, I am uncertain if he has a github handle?

First up we need to register IIS-UTokyo as an institution_id.
"IIS-UTokyo": "Institute of Industrial Science, The University of Tokyo, Hongo, Bunkyo-ku 113-8656, Japan"

Will there be a second institution contributing? (for existing institutions contributing to the project see input4MIPs_institution_id.json).

Next we need to define the source_id for the dataset, I suggest IIS-UTokyo-LS3MIP-1-0. This follows the format of other datasets that have registered and contributed, see input4MIPs_source_id.json

@hyungjun

[durack1]

The variables to be considered here include:
LWdown, PSurf, Qair, Qair.v1.09, Rainf, SWdown, Snowf, Tair, Tair.v1.09, Wind

Datasets are available from:
http://search.diasjp.net/en/dataset/GSWP3_EXP1_Forcing
http://www.dias.nii.ac.jp/gswp3/input.html

[durack1]

@hyungjun I was unable to figure out how you had partitioned the code in hyungjun/cmor4alma, so I started again with the MRI demo and now have a placeholder ready to go for the variables huss, prra, prsn, psl, rlds, rsds, and tas. If you wanted to continue with your branch, the one line to be added after this line would be:
time.setBounds(time.genGenericBounds()) ; # Add bounds to time axis

To achieve the successful demo I needed to update the tables, and registered the institution_id = IIS-UTokyo and source_id = IIS-UTokyo-GSWP3-LS3MIP-1-1. These will both need tweaks if you're unhappy with these identifiers - I have attempted to most closely follow the format of existing registered content.

The working demo is now writing out all datasets for the 1901 year (aside from the sfcWind which needs further info), a demo ncdump -ct output is attached for tas.
190120_tas_input4MIPs_atmosphericState_LS3MIP_IIS-UTokyo-GSWP3-LS3MIP-1-1_gn_190101010000-190112312100.txt

We need to double check what these variables are. The A3hrPt tables reflect data "3hrPt":"sampled 3 hourly, at specified time point within the time period" (see the PCMDI/input4MIPs-cmor-tables/input4MIPs_frequency.json file for all the entries available), so temporal snapshots, whereas the A3hr table reflect time-averaged data "3hr":"sampled every 3 hours". Are all these correct, i.e. huss, psl and tas are temporal snapshots, rather than 3-hour time averages?

All CMIP6 defined variables are available in the data request - see http://clipc-services.ceda.ac.uk/dreq/index/CMORvar.html - the sfcWind variable is available for 3hr (time average), daily and monthly (with variables also available for differing vertical coords, e.g. 6hrPlev). To get this correct, we may need to define the specific variables into the input4MIPs Tables specifically. There are also different variants that provide zonal and meridional values (uas and vas).

[durack1]

@hyungjun, just wondering where we're at with this data rewrite?

[durack1]

@hyungjun, @gkrinner just wondering where we are with the LS3MIP data contribution for input4MIPs?