More on the Divergence Between UAH and RSS Global Temperatures

July 8th, 2011 by Roy W. Spencer, Ph. D.

After talking with John Christy, I decided I should further expound upon the points I made in my last post.

The issue is that the two main satellite-based records of global lower tropospheric temperature change have been diverging in the last 10 years, with the RSS version giving cooler anomalies than our (UAH) version in recent years, as shown in the following plot:
(the RSS anomalies have been re-computed to be relative to the 1981-2010 period we use in the UAH dataset)

Ten years ago, this meant that the AGW folks were claiming RSS was right and we (UAH) were wrong, since the RSS global warming trends were greater than ours.

But now the shoe is on the other foot, and the RSS linear trend since January 1998 has actually cooled slightly (-0.03 deg. C per decade) while ours has warmed slightly (almost +0.05 deg. C per decade).

John works hard at making our dataset as good as it can be, and has correctly reminded me that he and others have several peer reviewed and published papers recent years on the subject of the accuracy of the UAH dataset:

Christy, J.R. and Norris, W.B. 2006. Satellite and VIZ-radiosonde intercomparisons for diagnosis of nonclimatic influences. Journal of Atmospheric and Oceanic Technology 23: 1181-1194.

Christy, J.R., Norris, W.B., Spencer, R.W. and Hnilo, J.J. 2007. Tropospheric temperature change since 1979 from tropical radiosonde and satellite measurements. Journal of Geophysical Research 112: doi:10.1029/2005JD0068.

Christy, J.R. and Norris, W.B. 2009. Discontinuity issues with radiosondes and satellite temperatures in the Australia region 1979-2006. Journal of Atmospheric and Oceanic Technology 25: OI:10.1175/2008JTECHA1126.1.

Christy, J.; Herman, B.; Pielke, Sr., R.; Klotzbach, P.; McNider, R.; Hnilo, J.; Spencer, R.; Chase, T. et al. (2010). “What Do Observational Datasets Say About Modeled Tropospheric Temperature Trends Since 1979?”. Remote Sensing 2 (9): 2148. doi:10.3390/rs2092148

Douglass, D. and J. R. Christy, 2009: Limits on CO2 climate forcing from recent temperature data of Earth, Energy & Environment, Volume 20, Numbers 1-2, January 2009 , pp. 177-189(13) doi:10.1260/095830509787689277.

Randall, R.M. and Herman, B.M. 2008. Using limited time period trends as a means to determine attribution of discrepancies in microwave sounding unit derived tropospheric temperature time series. Journal of Geophysical Research: doi:10.1029/2007JD008864.

Bengtsson, L. and K.I.Hodges, On the Evaluation of Temperature Trends in the Tropical Troposphere, Clim. Dyn., doi 10.1007/s00382-009-0680-y, 2009.

These papers either directly or indirectly address the quality of the UAH datasets, including comparisons to the RSS datasets.

Based upon the evidence to date, it is pretty clear that (1) the UAH dataset is more accurate than RSS, and that (2) the RSS practice of using a climate model to correct for the effect of diurnal drift of the satellite orbits on the temperature measurements is what is responsible for the spurious behavior noted in the above graph.

Our concerns about the diurnal drift adjustment issue have been repeatedly passed on to RSS in recent years.

What Will the Next IPCC Report Say?

As an aside, it will be interesting to see how the next IPCC report will handle the various global temperature datasets. There have been a few recent papers that have gone through great pains to explain away the lack of long-term warming in the satellite and radiosonde data (the missing “hot spot”) by trying to infer its presence from upper tropospheric wind data (a dubious technique since geostrophic balance is a poor assumption in the tropics), or by using a few outlier radiosonde stations with poor quality control and spurious warming trends.

Call me a cynic, but I think we can expect the IPCC to simply ignore (or at most brush aside) any published evidence that does not fit the AGW template.

11 Responses to “More on the Divergence Between UAH and RSS Global Temperatures”

Toggle Trackbacks

  1. Jon-Anders Grannes says:

    I Think I have seen something similar in the Arctic temperatur series.

    If you adjust the sensivity of a measurement to increase the warming when you have a warming.
    If there is no more fudging, you will get increasd cooling if There is a cooling.

  2. Alastair says:

    What I have yet to see is a comparison between the models and your UAH results from 1979 until present. Do they match? Or is the claim by the RSS crowd that since they found a bug in your code just their wishful thinking?

    Nobody seems to speak out and say there is a problem with the models.

    Cheers, Alastair.

  3. No hotspot, no global warming says:

    “radiosonde data (the missing “hot spot”) ”

    The lack of hot spot not only disproves AGW, it also disproves any global warming at all.

  4. Jon-Anders Grannes says:

    The Science(“climate”)is mostly UNFCCC controlled and driven(Political).

    Models are made according to the Political UNFCCC convention.

    The real measurable World is not.

  5. Dennis Hlinka says:

    Dr. Spencer,

    In regards to your last statement about the IPCC not likely publishing evidence that does not fit the AGW template, when are you, and many of those on your side of the argument, finally going to acknowledge the fact that your own UAH data set disputes the anti-AGW template of a cooling trend going on since 1998? I am specifically referring to the fact that your own plot of the running 13-month average shows that the temperature anomalies for 2010 exceeded the earlier peak in 1998, therefore eliminating that false argument. I have never heard or seen any statements coming from your side that clearly points out that fact. When will you actually make that point very clear in the most factual terms to the public, or is that too much against your template to be that forthcoming with that level of acknowledgment because it may confuse your followers in what they have been told to believe for the past 10+ years?

  6. Andrew says:

    Hlinka, you are joking, right? The comments regarding a lack of warming trend since 1998 were made during periods when that is indeed what UAH showed. About a year or two ago it would have been possible to say there had been no warming since 1997! But with a recent El Nino we get a look at what the apparent “ongoing trend” is capable of doing…producing .05 degrees of tropospheric warming per decade, apparently. This is rather pathetic. As for Roy’s chart, I’ve looked at it, and the 1998 peak is indistinguishable from the 2010 peak, so what you are claiming is not quite true, in fact.

    You would also really like to obscure the fact that the trend is so small it can disappear for twelve years at a time just from variability (note that other such periods have occurred, but were associated with volcanic eruptions, and were shorter) and is much less than climate models “project”.

    Hotspot guy, your statement is factually erroneous. It is certainly true that the lack of amplification of surface warming in the tropical troposphere is an indication that the models are severely deficient, and/or the surface data exaggerate warming, but it is quite clear that there is a (small) global trend in spite of the lack of hotspot. I and other skeptics would appreciate it if strawmen arguments would only be made by men of straw, not actual people. Allowing comments such as these to appear to be representative of our thinking gives warmists an easy target for rebuttal to falsely make all of us appear to be wrong.

    • No hotspot, no global warming says:

      Andrew, your statement is factually erroneous. Amplification of surface warming in the upper tropical troposhere is a simple consequence of the slowing of the saturated adiabatic lapse rate as the tropical troposphere gets warmer. Complex models don’t change this simple law of physics. At least you’re right when you say (or) it’s an indication that the surface data exaggerates warming. No amplification -> no warming from any cause.

  7. Andrew says:

    Roy, Carl Mears’ comment regarding the TMT trends made me curious to see what those differences looked like. I came across some strange features. For one thing, there appears (from cursory examination) to be a significant seasonal cycle divergence in the AMSU era, similar to what was corrected for in the transition from v5.2 to v5.3 of the lower troposphere data. Was this feature not addressed in the TMT data? Other curiosities include the early nineties discontinuity appearing to be absent from these data, indeed the divergences are mostly before 1986 (and these look really odd) and after about 1998 (slightly after 1994) with the period from 1986 to 1994 being very close between the two. As with the TLT data, the recent, post about 2000 data show a cooling of RSS relative to UAH offsetting partially an earlier warming.

    Can you comment on these differences, or perhaps get John Christy’s opinion on these differences? I rarely use or make reference to the TMT data compared to the TLT data which I reference a lot, but I am still curious. Thanks!

  8. Andrew says:


    No, you are still wrong. If as you say there physically must be a greater warming in the tropical troposphere than the surface (I am not so sure that they “must” and don’t think that there is anything about the adiabatic lapse rate that represents a “simple physical law” but whatever) so if this is the case then if we know what the amplification rate “must” be, we can calculate appropriate surface change from the tropical troposphere data. There is a small, but positive trend in the tropical troposphere (in the lower troposphere, for instance, it’s about .07 degrees per decade) which “must” according to you and models, correspond to a surface trend of about .05 degrees per decade, which is NOT zero. Now using the global data and the global average expected amplification the .14 degree per decade global lower troposphere trend becomes a global surface trend of about .116 degrees per decade, which again is NOT zero. These trends do show that the surface data exaggerate warming (assuming that the amplification ratio is a “physical law”) but not that there is no warming whatsoever. Again, your statements are quite clearly factually erroneous. Even just the fact that there is a positive global trend in the lower troposphere shows that it is not necessary to have a large tropical tropospheric trend to have any global trend at all.

  9. No hotspot, no global warming says:


    “If as you say there physically must be a greater warming in the”


    “tropical troposphere than the surface (I am not so sure that they “must””

    So you think there shouldn’t necessarily be an upper tropical troposphere “hotspot”. In that case you don’t accept the view that the lack of a “hotspot” is proof that global warming cannot be caused by CO2 forcing or indeed proof of anything. Dr. Spencer appears to have a different view.

    “and don’t think that there is anything about the adiabatic lapse rate that represents a “simple physical law””

    The saturated adiabatic lapse rate just depends on physical properties of air and water vapor.

    “so if this is the case then if we know what the amplification rate “must” be, we can calculate appropriate surface change from the”

    difference between the upper

    “tropical troposphere data”

    and the surface data (which is the bottom of the troposphere).

    I don’t think you’re getting the point. The “hot spot” is the difference between the upper troposphere rise and the surface rise. No difference -> no rise (global warming) in the tropics.

  10. Martin says:

    @No hotspot, no global warming
    In the history of UAH, tropical region has warmed only slightly. So did southern hemisphere. Most of the warming occured in northern hemisphere.

    That´s what I call internal variability. And of course, tropics is mostly driven by ENSO