Many thanks for the very interesting links.

>I’m not sure what’s going on here. However, this blog post suggests it could be an artifact of the NOAA/NESDID SST anomaly data product (see also here).

What I understand from both posts is that the above NOAA images are created from NOAA/NESDIS data, which doesn't
include the buoy measurements, but only satellite data.
What follows from the blog posts is that the hot spot -if at all- is mostly only that strongly visible in the NESDIS satellite data
and not in the olv2 anomaly data, which includes buoy measurements.

The blog author found that the olv2 anomaly data displays a very cold peak in Aug 2008 for the Barents Sea, which seems
not visible in the NESDIS data, however I find the peak (around Aug 25) seems also somewhat visible in the NOAA data
if one compares those two images (and the following to Aug 25):




Moreover there seems to be a scale problem with the olv2 anomaly data though if you compare e.g. 04 Aug 1999 with Aug 2003:
HERE WAS AN EMBEDDING WHICH WAS NOT SWALLOWED BY THE HTML

The finegradedness changes there also from 16 quantization steps down to 9, which is only half of that of the NESDIS.

The NESDIS displays anomaly differences of around 5 degrees between the hot spot and the cold spots next to it. A global scale could make those differences smaller and less visible in a coarse quantisation.

Anyways but still even the the olv2 data reveals a kind of hot spot, if one looks at the january data:
HERE WAS AN EMBEDDING WHICH WAS NOT SWALLOWED BY THE HTML

So concluding it seems there are some artifacts, especially in the olv2 data, but these may be mostly due to the mapping (like the finegradedness, moreover in a comment in here it was remarked that some (scaling?) labels in the dataset where inconsistent).
Overall the hot spot seems still to be there in the images (especially in winter). Apart from the mapping differences I could however also imagine that
there are real differences also in the datasets. That however would point to some old suspicion of me, that there is something strange with the assesment of the global warming potential of methane.

>Also, I don’t really see how clathrate dissociation could significantly warm the regional surface waters.

That is if there are "considerably cold " methane leakages (like the olv2 data set and you seems to suggest) then due a high global warming potential (and as the
NESDIS data suggests this looks ("this is just very speculative statement though) much higher then as the official values would suggest) the daytime satellite data could reveal
much higher temperature values due to methane. But I don't know how the clouds and the methane rising up into the stratosphere would influence the satellite data.

>I can’t imagine there’s any hemispheric difference in top-of-atmosphere solar irradiation, but there can be a difference in surface insolation due to cloud albedo.

Yes I think there is quite certainly a difference in surface insolation due to cloud albedo. I am not sure but due to the ellipticity of the earth orbit around the sun I could though also
imagine that there could be differences in solar irradiation for the two hemisspheres. I was even almost at the point to try to do the calculations myself, but then I was missing data about the actual orbit (like exact ellipticity etc.) and apart from this this is really tedious work, moreover there should be tons of similar calculations somewhere, which I just sofar haven't found.