2010-11-28 18:57:28Is there any peer reviewed papers comparing surface measurements of temperature to satellite data?
John Cook

I mention in passing that surface data shows the same warming trend as the satellite data. Eg - GISS, NOAA and HadCRUT show around .16 per decade, RSS show .16 while UAH shows around .14. But would be good to reference a peer reviewed paper that compare the two types of measurements. I don't recall any such paper of the top of my head - I don't suppose anyone is aware of any such paper?
2010-11-28 20:57:52

I don't think I've ever seen this comparison. I also have never seen any attempt to extract surface temperature from lower tropospheric temperature. The most comprehensive temperature comparison I know is the U.S. Climate Change Science Program Synthesis and Assessment Product 1.1, in particular chapter 3.
2010-11-29 08:52:04
Glenn Tamblyn


I don't know of any specific papers comparing this although you would expect there to be some somewhere. I know the satellite data is not regarded as reliable for extracting a surface record - too much signal originating from the ground itself rather than the surface air.

Here is an interesting paper from the mid 2000's looking at correcting for biasing of the satellite data for the troposphere from signal originating in the cooling lower stratosphere.


It looks like the sat temps for the stratosphere are under reporting temps due to biasing from the stratosphere.

2010-11-29 10:30:10My paragraph on surface vs satellite temperature
John Cook


This is my paragraph in the Guide re surface temp and satellites (page 6):

  • Another way to check thermometer measurements is to compare them to satellite data. Satellite measurements show the same rate of global warming.35 This is confirmation that thermometers are giving us an accurate picture.

The reference I use is Riccardo's link:

  • Karl, T. R., Hassol, S. J., Miller, C. D. and Murray, W. L. (2006). Temperature Trends in the Lower Atmosphere: Steps for Understanding and Reconciling Differences. A Report by the Climate Change Science Program and the Subcommittee on Global Change Research, Washington, DC.

Is this a robust summary of the situation? In the referenced material, Chapter 3 Page 58, they say:

  • However, closer examination shows that two of the products (UMd and RSS satellite data) indicate considerable temperature increase over the period of record, whereas the other two (UAH satellite and HadAT2 radiosonde) indicate slight warming only.

So while RSS is consistent with surface temp, UAH shows a smaller trend. So I could be accused of the same cherry picking accusation I throw around throughout the guide, by picking RSS and ignoring UAH. There's the additional complication that the satellites measure "Lower troposphere" which is a different metric to surface temperature.

So my original statement is broadly correct - thermometers shows .16C warming while satellites show between .14 (UAH) and .16 (RSS). Would a more appropriate wording be "Satellite measurements show a consistent rate of global warming"?

2010-11-29 18:35:35
Ari Jokimäki


Here are some papers on the subject:


Here is a discussion and links to Karl's recent lecture where he compares surface and troposphere temeprature trends (see the second figure, bottom left panel):


2010-11-29 20:44:02
Glenn Tamblyn



The question of Tropospheric Temps from the satellites is murky. Really a bit of a spectrum of results.

The lowest signal is UAH. But here you have to be careful about what is meant by UAH. The most common sceptic cited series is UAH T2LT. This is an analysis of data from the MSU2 sensor on the earlier NOAA satellites - AMSU4 in the later ones. They are using unusual signal processing, looking at data from very wide in the scan angle to try and extract a signal for the lower troposphere without ht ebiasing from stratospheric cooling. This technique has been criticised.

RSS use a different approach, using a different weighting function to try and extract a lower troposphere signal with not too much stratospheric contamination.

Other issues. Both groups, although I think RSS was first have identified issues with the NOAA 16 satellite and switched to newer staellites as soon as they were available. Others have identified a divergence between UAH & RSS since the NOAA 9 Satellite. I don't know the reason for this.

Them Fu et al used Radiosonde data to try and establish a profile through the air column allowing some estimation of the stratospheric contribution to the tropospheric signal. They derived an equation that allowed them to calculate the tropospheric signal by deducting a proportion of the stratospheric signal from the MSU4/AMSU9 sensor to try and remove the stratospheric bias. The method seems reasonable but it is very coarse - poor radiosonde coverage and not usable to remove the stratospheric bias from the data on a daily basis. But useful for a broad look at the impact on overall trends. Their paper includes an interesting bar graph of trend data for UAH & RSS with their algorithm applied. Jo Nova's missing Hotspot shows up pretty clearly

They published a later paper in 2005 attempting to do the same thing but using data not from radiosondes but the much less well used MSU3 sensor. This has a weighting function that is a mid point between MSU2 and the stratospheric MSU4. I haven't seen any other reference to this.

NOAA maintain temp series from UAH & RSS and also UAH & RSS with the Fu 2004 adjustment applied. This only seems to go up to 2008.

Finally Vinikov & Grody have published a couple of papers suggesting problems with other aspects of the data analysis. There analysis gives trends over 0.2/decade. Fu applied to V&G would be even higher since they address differing non-overlapping issues.

So there seems to be a range from UAH at the bottom (Spencer & Christy being the authors of that couldn't have anything to do with that, SHUT YOUR MOUTH BOY!, Go wash your mouth out with Soap for suggesting such a thing), through RSS, UAH + Fu, RSS + Fu, G&V, G&V + Fu.

But any clear resolution of this, let alone Peer-reviewed work that does! Keep waiting.

So, given the reasonable the arguments of Fu that the stratospheric bias does need correction and isn't, picking the lowest series from all those available doesn't seem reasonable. RSS isn't a bad compromise between peer-reviewed certainty vs reasonable grounds for thinking it is higher.