Which basins are in the 'basin' dimension?

[jypeter]

@martinjuckes @taylor13 @dnadeau4 can somebody tell me where I can find the official definition of the basin dimension? More precisely which basins we are talking about? I have tried to find this information after talking to @oliviermarti about https://github.com/cmip6dr/CMIP6_DataRequest_VariableDefinitions/issues/190

I have failed to find that information on the CMIP6 Data Request web site, but it's quite possible I did not follow the right links.

The Spreadsheet view of the variable definitions does not have the convenient dims panel that was present in the CMIP5 standard output document, so you need to find the basin information somewhere else

As a last resort, I have looked at the CMIP6_coordinate.json for CMOR and found the following:

            "requested": [
                "atlantic_arctic_ocean", 
                "indian_pacific_ocean", 
                "global_ocean"
             ], 

I have also looked for basin in OMIP contribution to CMIP6: experimental and diagnostic protocol for the physical component of the Ocean Model Intercomparison Project but the paper does not mention the basin dimension. It only mentions the basin region variable (that defines a mask for 10 regions)

[jypeter]

I have eventually followed the right sequence of links and found a DR web page defining the same basin dimension as the json file mentioned above: http://clipc-services.ceda.ac.uk/dreq/u/dim:basin.html

Before finding that information, I also asked James Orr, our local OMIP paper author, and he was not able to give me the information. It's strange that the definition of the basin dimension is missing in the OMIP paper. And the fact that the basin dimension has 3 basins that are different from the 10 basin regions in the paper may be confusing for some people

[martinjuckes]

True, but this is unchanged from CMIP5, so I think it is as intended. It would perhaps have been clearer to call the basin dimension something like basinTransects -- because it is actually referring to specific transects of special interest which happen to be commonly referred to by the name of the basin they sit in. I'll put a note in the CMIP7 issues (https://github.com/cmip6dr/cmip7_forward_look/issues/11).

[taylor13]

These are not north-south transects within a basin (as recorded, e.g., by research vessels). Rather these are "zonal means" limited to the longitudes included within the basin. The comment in CMIP5 standard_output provided instructions on how to compute zonal means for ocean models with grid cells no aligned with latitude circles; thus, the reference to zigzag paths. Those are paths roughly following a latitude circle used to calculate the zonal mean.

[martinjuckes]

Thanks for the clarification .. I'll adjust the forward look comment: https://github.com/cmip6dr/cmip7_forward_look/issues/11