I continue to get lots of e-mails asking how global average tropospheric temperatures for January, 2010 could be at a record high (for January, anyway, in the 32 year satellite record) when it seems like it was such a cold January where people actually live.

I followed up with a short sea surface temperature analysis from AMSR-E data which ended up being consistent with the AMSU tropospheric temperatures.

I’m sure part of the reason is warm El Nino conditions in the Pacific. Less certain is my guess that when the Northern Hemisphere continents are unusually cold in winter, then ocean surface temperatures, at least in the Northern Hemisphere, should be unusually warm. But this is just speculation on my part, based on the idea that cold continental air masses can intensify when they get land-locked, with less flow of maritime air masses over the continents, and less flow of cold air masses over the ocean. Maybe the Arctic Oscillation is an index of this, as a few of you have suggested, but I really don’t know.

Also, remember that there are always quasi-monthly oscillations in the amount of heat flux from the ocean to the atmosphere, primarily in the tropics, which is why a monthly up-tick in tropospheric temperatures is usually followed by a down-tick the next month, and vice-versa.

So, it could be that all factors simply conspired to give an unusually warm spike in January…only time will tell.

But this event has also spurred me to do something I’ve been putting off for years, which is develop limb corrections for the Aqua AMSU instrument. This will allow us to make global grids from the data (current grids are still based upon NOAA-15, which we know has a spurious warming over land areas from orbital decay and a changing local observation time). Since the Aqua AMSU is the first instrument on a satellite whose orbit is actively maintained, there will be no problem with those data since Aqua came online in mid-2002.

[Don’t get confused here…we use NOAA-15 AMSU ONLY to get spatial patterns, which are then forced to match the Aqua AMSU measurements when averaged in latitude bands. So, using NOAA-15 data does not corrupt the global or latitude-band averages…but they do affect how the warm and cool patterns are partitioned between land and ocean.]

I might also extend the analysis to specifically retrieve near-surface temperatures over land. I did this several years ago with SSM/I data over land, but never tried to get it published. It could be that such a comparison between AMSU surface and near-surface channels will uncover some interesting things about the urban heat island effect, since I use hourly surface temperature observations as training data in that effort.