Identified issue with NorESM2-LM and NorESM2-MM 'siareas' CMIP variable

[YanchunHe]

Quote from C.Holmes at bas.ac.uk

I have been using the variable 'siconc' to produce sea ice area, and the ready-produced 'siareas' (sea ice area south) where available as a cross-reference.

For both NorESM2-LM and NorESM2-MM, I noticed a discrepancy between these two values and concluded the published 'siareas' dataset is erroneous. I concluded this for the following reasons
The February minimum of the NorESM siareas in the published dataset is around 15Mkm2 (obs ~2, CMIP range ~0-6). However, the siconc distribution shows that the siarea values should not be a high outlier (see e.g. supplementary Figure S3 panels an and ao in https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2019GL086729)
The seasonal cycle of the NorESM siareas in the published dataset has a double peak.
Investigating further, the erroneous values are what I obtain if siconc is aggregated over the full globe and not only the southern hemisphere.

This error exists in historical and each SSP under ScenarioMIP, I've not checked elsewhere.
The "sivols" I have calculated appear to be the same as those provided.

I hope this is helpful, easily rectifiable, and that I have informed the right person.

C.H

[YanchunHe]

Quote J. Debernard.

Thank you for reporting this issue, and the very helpful information.

I can confirm that the published siareas-values are aggregated over the whole globe, as you suggested.
When taking the difference between the published siareas values and the published siarean values, (global - Nh), the resulting Sh-values are identical (within run-off) to Sh-values I calculate independently.

I can also confirm that this is only an issue with the siareas field, the siexts and sivols fields (which are calculated in a similar manner) look correct.
I also think I have spotted the reason for this error in a namelist file for our cmorization procedure, so unfortunately, I think this error is present in all our CMIP6 data.

Many thanks for your effort concerning this matter.

[YanchunHe]

Alo, @monsieuralok, please retract all the siareas variable, for both NorESM2-LM and NorESM-MM. Thanks!

[monsieuralok]

@YanchunHe I have removed all siareas from NorESM2-LM and NorESM2-MM. Are we planning to republish them?

[YanchunHe]

@YanchunHe I have removed all siareas from NorESM2-LM and NorESM2-MM. Are we planning to republish them?

Yes, @monsieuralok

corrected siareas from some of the experiments will be published together with other fields!