Earlier today I wanted to take a quick look at the global plots.
Invoke update here which gets new data, nothing happens. I’d forgotten, or it had not registered with me there is a version change and that means filename changes. No problem.
My only comment is wondering about the amount of month to month variation, an ongoing gripe with this kind of dataset (all global datasets suffer): annual variation should not be present. Such variation has the chickens in panic, watching every doom twitch is going to break eggs.
[Commented added 29th. Had a quick look at V5.4 post 2002 and it looks improved, less annual variation. This problem is very tough for satellite data given a dataset from 1978 to date is a mix of disparate datasets. Only looking at post mid-2002 with UAH is single satellite, but as a whole dataset compromising one period for another is not an option, the dataset authors have a difficult job]
Here are September through December 2010 equal area plots, plus Hovmoller, 1990 through end 2010.
Click on an image for full size.
The pattern, can be seen over many years in Hadcrut3, is heat bumps in the tropics which then flows in time toward the poles, heat shedding. This is always followed by a colder time in the tropics. Pretty obviously this is in progress now. Expect much like the sequence after 1998. The recent cold in the northern temperature zone doesn’t figure, too transient and local to make the bigger picture.
I suspect a good measure of global is just a narrow band in the tropics, not global as such which adds confusion and time delays as heat flows. Those people trying to link terrestrial and solar might find it a better metric. That said, is there a variation in the degree of temperature variation in the north and of so why? Also the ongoing tilt from south to north, why? On the latter I suspect it is about to start moving back, but my reasons are unstated here.