Wednesday, July 06, 2016

Comparing Surface and Satellite Temperatures for El Ninos

In the comments of this article, Layzej asked why the peak lower tropospheric temperatures of the El Nino that just ended were about the same as the peak in GISS surface temperatures, whereas they were about 0.15 C higher for both the 1997-98 El Nino and the 2009-10 El Nino. He pointed to this chart from WFT, which uses version 5.6 of UAH's LT dataset.

Perhaps WFT is waiting for UAH v6 to come out of beta, and if so I can't blame them. Frankly I think UAH should not even have introduced a beta version of version 6, except internally, but should have done like NOAA with the Karl et al Science paper of 2015, when the data were released in a set version and published when the paper itself was published. UAH says their paper is still in peer review.

Anyway, here is a comparison between UAH LT v6beta6 and GISS. I've adjusted the GISS anomalies to have the same baseline as UAH, 1981-2010, and smoothed the data with a 12-month moving average. The results look significantly different (any thoughts why?):

6 comments:

Layzej said...

I think the biggest difference between 5.6 and 6.x is the trend. If you detrend both GISS and UAH6 you will still see as a general feature that UAH is quite a bit higher than GISS during El Nino and quite a bit lower during La Nina. Not so for this year though.

David Appell said...

Good idea. I'll do that when I get a chance.

DocRichard said...

Before 2000 UAH is reading warmer than GISS. 2000-2005 the two are in agreement. 2005 onwards UAH is reading cooler than GISS, apart from the peak in 2011. Possible explanation is either a deterioration in the sensitivity of the satellite sensors, or an increase in the sensitivity of the GISS data collectors. The former seems more likely, because there are more surface temperature sensors than satellites.

Layzej said...

Are they using different instruments in V6.5 vs V5.6? The discrepancy between satellite and land measurements is small in 5.6 but quite large in 6.5. If the instruments are the same between versions then the change must be related to how the data is being processed. I'm not quite sure what issue version 6 was meant to address. I guess we'll have to wait for the paper?

Unknown said...

The only explanation for the difference is the GISTEMP is more sensitive to Arctic global warming than UAH. This has been commented on before I think. I also wonder if the UAH measurements have actually peaked, or will they go up again?

David Appell said...

UAH LT has been down four months in a row now. It's looking like it has peaked, but....