Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.


Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Bluesky Facebook LinkedIn Mastodon MeWe

Twitter YouTube RSS Posts RSS Comments Email Subscribe

Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...

New? Register here
Forgot your password?

Latest Posts


Satellite measurements of warming in the troposphere

What the science says...

Select a level... Basic Intermediate Advanced

The most recent satellite data show that the earth as a whole is warming.

Climate Myth...

Satellites show no warming in the troposphere

"Satellite measurements indicate an absence of significant global warming since 1979, the very period that human carbon dioxide emissions have been increasing rapidly. The satellite data signal not only the absence of substantial human-induced warming but also provide an empirical test of the greenhouse hypothesis - a test that the hypothesis fails." (Bob Carter)

John Christy and Roy Spencer of the University of Alabama published a series of papers starting about 1990 that implied the troposphere was warming at a much slower rate than the surface temperature record and climate models indicated Spencer and Christy (1992). One early version of their data even showed a cooling trend (Christy et al. 1995).

Several groups of scientists began looking closely at this discrepancy. With so many other pieces of evidence indicating warming, it seemed unlikely that the troposphere would not be warming. Errors were discovered in the methods the UAH group used to adjust the data.

To understand what was wrong: The satellites must pass over the same spot on Earth at the same time each day to get a temperature average. In reality the time the satellite passes drifts slightly as the orbit slowly decays. To compensate for this and other orbital changes a series of adjustments must be applied to the data.

Temperature trends of the troposphere now match well with the surface based trend.

The MSU satellite data is collected from a number of satellites orbiting & providing daily coverage of some 80% of the Earth's surface. Each day the orbits shift and 100% coverage is achieved every 3-4 days. The microwave sensors on the satellites do not directly measure temperature, but rather radiation given off by oxygen in the Earth's atmosphere. The intensity of this radiation is directly proportional to the temperature of the air and is therefore used to estimate global temperatures.

There are also differences between the sensors that were onboard each satellite and merging this data to one continuous record is not easily done. It was nearly 13 years after the orginal papers that the adjustments that Christy and Spencer originally applied were found to be incorrect. Mears et al. (2003) and Mears et al. (2005).

When the correct adjustments to the data were applied the data matched much more closely the trends expected by climate models. It was also more consistent with the historical record of troposphere temperatures obtained from weather balloons. As better methods to adjust for biases in instruments and orbital changes have been developed, the differences between the surface temperature record and the troposphere have steadily decreased.

At least two other groups keep track of the tropospheric temperature using satellites and they all now show warming in the troposphere that is consistent with the surface temperature record. Furthermore data also shows now that the stratosphere is cooling as predicted by the physics.

All three groups measuring temperatures of the troposphere show a warming trend. The U.S. Climate Change Science Program produced a study (pdf) in April 2006 on this topic. Lead authors included John Christy of UAH and Ben Santer of Lawrence Livermore National Labs. The first page has this quote:

Previously reported discrepancies between the amount of warming near the surface and higher in the atmosphere have been used to challenge the reliability of climate models and the reality of human-induced global warming... This significant discrepancy no longer exists because errors in the satellite and radiosonde data have been identified and corrected. New data sets have also been developed that do not show such discrepancies."

There are still some discrepancies between satellite measured temperatures in the tropics and those measured by radiosondes. Most researchers believe this difference is likely due to instrument errors.

The original discrepancy is an excellent example of how science works and of critical thinking. With many different indicators showing warming, it did not make sense that the troposphere would be cooling. This discrepancy was taken very seriously by the scientific community, and the consistency and accuracy of all relevant data were examined intensely.

Science advances by trial and error. The result is an increased knowledge of how to measure the temperature of the troposphere from space.

Last updated on 24 July 2015 by David Kirtley. View Archives

Printable Version  |  Offline PDF Version  |  Link to this page

Argument Feedback

Please use this form to let us know about suggested updates to this rebuttal.

Further reading

Where Bob Carter got his data

When I emailed Bob Carter querying about his data in the article above, this was his reply (28 Jun 2007):

"By mistake the graph that was reproduced in the Telegraph article was for the middle troposphere. Though it does not materially affect the argument or conclusions, I am embarrassed by it because it can be made to look as if I was pulling a swiftie - which I wasn't (intending to)."

I'll take Carter at his word that it was an honest mistake, although I've noticed he continues to state satellites show little to no tropospheric warming.


Prev  1  2  3  Next

Comments 26 to 50 out of 61:

  1. knaugle - The reason for starting in 1982 was the 60-month averaging Spencer used, which leads directly to some of the distortions in the graphing. The satellite temps are more sensitive (higher swings) to ENSO and yearly variation than the surface temperatures, and the 5-year average for 1982 (1979-1984) is one of the highest peaks above trend for the UAH record. 

    'Baselining' to a single point of the 60-month average, as Spencer did, gives an artificial negative offset to the UAH record, as below:

    60-month Temperatures and Offsets


    Blue: HadCRUT4 60 month average, Red: UAH 60 month average baselined to the common average 1980-1999, and Green: UAH using Spencers 60-month average and a single point (mis)alignment. 

    As you can clearly see, the 60-month UAH alignment shifts the entire record down, artificially offsetting the satellite temps and making surface and model temperatures seem much higher. When in reality UAH swings both above and below the surface temps, as per the graphs baselined for comparison over a period longer than their internal variability

    I don't know whether Drs Spencer and Christy did this alignment out of an error of confirmation bias or as a deliberate 'trick', but either way the graphs they presented with that 60-month baselining are deceptive. 

  2. knauble - Here's the same graph along with trends, and you can clearly see that the 60-month single point offset Spencer/Christy used results in the UAH trend sharing no common period with HadCRUT4 whatsoever. 

    60 Month averages with trends


    If trends don't cross in the baseline period, you cannot reasonably judge any divergence - this is a clear error on their part. And quite deceptive when their claim is that similarly baselined models are running too high - as they've artificially introduced an offset. 

  3. Hi

    I hope I can get some help understanding the conflict in RSS TLT data.

    The satellite data from RSS seems to fall well below all 33 IPCC data models for Temperature Lower Troposphere (TLT). As can be observed below. The thick Black line is the averaged RSS satellite data and the light blue lines are the 33 IPCC (TLT) models.

    RSS Data Graph

    I understand SKS has the Troposphere graph but I would expect the RSS Green line to be well under the Direct Measurement models. As is shown in the RSS website graph above.

    SKS representation of RSS data

    The actual response from RSS doesn't seem to nail the exact problem of the lack of correlation.

    RSS report on data discrepency

    Any assistance would be greatly appreciated.

  4. MEJ...  For one, the 33 models in the upper panel are not (to my knowledge) models of TLT. The heavy black line is, the blue-green lines are modeled surface temps.

    Next, Dr. Mears does a fine job of answering your questions in the article you linked to.

  5. Rob - the plot is indeed of the lower troposphere, and it was created by Carl Mears. See here. So it is genuine.

    Some of the issues are that:

    1. The CMIP5 collection of climate model simulations use observations up to 2005 and projected forcings thereafter.

    2. Recent observations indicate that the greenhouse gas forcing in the 21st century was smaller mainly due to increased light-blocking volcanic sulfates and a weaker-than-expected solar cycle.

    3. Climate models cannot accurately predict the timing of natural variation, especially the multidecadal variation associated with the warm and cool phases of the Interdecadal Pacific Oscillation (IPO). The IPO has been negative (cool surface temps) since 1999/2000 even though the ocean has been soaking up heat. The strong warming of the subsurface ocean is a testament to this.

    So the issues are the same as surface temperature observation versus naive projections of the near-future forcings. When these are accounted for, the climate models are seen to be doing a very good job of simulating recent surface temperatures  - within their obvious limitations.


    Clearly RSS has some additional problems, as UAH and the surface temperature data sets are broadly in agreement, even though they are not measuring the same thing (radiative brightness of oxygen in the lower atmosphere vs surface temperature measured by thermometers).

  6. The lines drawn on the graph exclude initial data that would appear to produce a flatter interpretation. The second point is that if you are arguing for a linear increase, it suggests a problem trying to correlate against CO2 increase which is non linear.

  7. ginunn:

    Regarding your second comment, in point of fact temperature increase is linear with logarithmically increasing CO2: climate sensitivity, you may recall, measures global mean surface temperature increase per doubling of atmospheric concentration of CO2.

  8. ginunn - Also note that CO2 is not the only driver of climate. Nobody with even a basic understanding of the science would argue for a simple monotonic linear increase in temperatures in the presence of multiple forcings and internal climate variation. 

  9. Thank you all so much for the feedback. A lot of great information. I particularly like the moving graphic on surface temperature.This is proving to be a VERY complex issue to get my head around. It has struck me that I may be comparing apples with oranges.

    The SkS graphic is specifically comparing Surface and Satellite temperatures. Three sets of data are being compared. 'Direct Surface Measurements' (Blue Line) and two sets of Satellite(?) data from UAH (Red Line) and RSS (Green Line). In respect to 'Surface Temperature' and I presume once certain adjustments are made they show a very close correlation.

    Surface and satellite temperatures

    My limited understanding is that if I take a temperature measurement at a ship on the Pacific Ocean (Surface Temperature) then fly up to the Lower Troposphere some 8 to 15 Klms above my position the Temperature measurement will be cooler.

    So can I compare Carl Mear's graph of Temperature Lower Troposphere with a graph based on Surface Temperature data?

  10. 8000 to 15000 meters does not really describe lower troposphere, in my understanding. It sounds more like the tropopause height.

  11. Geert Jan van Oldenborgh has an excellent guest post on Climate Lab Book:  "Is there a pause in the temperature of the lower troposphere?"

  12. UAH has apparently completed a reivew of their methods and data, and re-written their software. Their new calculations seem to indicate a lower warming trend than they had been reporting. I don't know enough math or physics to make sense of their recent claims. Can someone comment?

  13. Spencer's post does not link to any peer reviewed paper.  Will they submit it for review?  The blog post is written in the style of a scientific paper.  I wonder if they have allowed anyone to look over their new mothods to find errors or if they will wait for others to find them later.  In the past other scientists have found major errors in UAH data processing.  Most of those errors have caused the trend to be falsely low.

    The surface record does not have a history of making major adjustments to the data like the atmosphere data do.  It is more reliable.

  14. Thank you, Michael. I know that the people at UAH have had a lot of trouble, and have a history of unreliable calculations, as this article shows. I suspect the contrarian crowd are about to seize on the latest UAH numbers, and I’m hoping to have some responses I can give.

  15. dcpetterson @37, truly informed comment on UAH v6 will have to await release of the code and examination by experts (of which I am not one).  Several of Spencer's graphs stood out to me, however, as allowing some form of preliminary assessment of the difference between v5.6 and v6.

    Of these, the most important is that showing weights by altitude:

    Because satellite data does not come from just one layer, satellite temperature products show the weighted average of temperature data from different altitudes.  Because of their new method, UAH v6 gives a much lower weight to near surface temperatures and a higher weight to (cooling) lower stratosphere temperatures.  This is partially balanced by an increased weight to warmer upper tropospheric temperatures (can any body say "upper tropospheric hotspot").  Spencer says this cools the record by " less than 0.01 C/decade".  As the difference between v 5.6 and v 6 trends is 0.026 C/decade, that represents about a third of the difference just by changing the weighting profile.

    Also of interest is the chart of trends by zone:

    Ignoring the inappropriate use of a line chart rather than a bar chart, what stands out is the huge reduction in Arctic temperature trends.  The other note worthy cool changes in the record are, in order of magnitude, the NH north of the Tropic of Cancer (NExtraT), the USA 49, the USA 48, and the NH.  Clearly almost the entire change comes from high latitude NH temperatures.  That is clearly inconsistent with surface temperature records by GISS, or DMI:

    It also does not pass the smell test in relationship to changes in Arctic ice (sea ice, glaciers and ice sheets).

    Finally comes the data excluded from the analysis:


    "Fig. 1. Local ascending node times for all satellites in our archive carrying MSU or AMSU temperature monitoring instruments. We do not use NOAA-17, Metop (failed AMSU7), NOAA-16 (excessive calibration drifts), NOAA-14 after July, 2001 (excessive calibration drift), or NOAA-9 after Feb. 1987 (failed MSU2)."

    If you look at the excluded satellites, that means that from late 1998 to mid 2003, NOAA 14 was the only satellite from which Spencer and Christy would accept data.  Curiously, a general warming trend of UAH v6 relative to UAH v5.6 reverses in 1998, and flattens out in 2005 (ie, the onset of NOAA 18).  That is a noteworthy coincidence at least, and suggest their diurnal temperature correction has problems.  Whether it does or not, of course, can only be answered by detailed analysis which I am unable to provide.

  16. Tom Curtis"...truly informed comment on UAH v6 will have to await release of the code and examination by experts..."

    Of course, Christy and Spencer have never released their code or algorithms (unlike the other temperature records). Making such examination difficult in the extreme. External corrections have come from various groups reconstructing the UAH record without assistance from UAH, discuvering where they may have gone wrong, and then after some (often considerable) period of time the UAH group has updated their results. 

    It's been 26 years so far, and the UAH group has yet to release their methods. I don't expect that to change any time soon. 

  17. Tom Curtis and KR ---

    Thank you for the informative replies. Very revealing!

    I was not aware that Christy and Spencer have never revealed their code or algorithms. That rather makes a mockery of scientific rigor, doesn't it? Or am I mistaken?

  18. KR @41, it has emerged lately that Spencer and Christy have revealed their code here (under Mean Layer Temperatures - UAH).  William Connelley recently bloged about it.  Apparently Eli mentioned it in 2013, but Spencer and Christy have certainly not advertised it.

  19. Tom Curtis - Thank you, I had not been aware of that code release. 

    Although from Connelley's report it seems to be rather horrifying (gah!) from a software perspective, with code changes and recompilation needed every month due to fixed arrays that have to be resized for each additional data point, huge chunks of code commented out or unreachable due to gotos and program logic, etc. 

    I suspect the most useful part of that information will be the algorithm description, which should permit cross-checking by other investigators, for example Po-Chedley and Fu (relevant 2012 article  on UAH reconstructions here). I hope they'll put out an update that includes the V6 changes - the linked documentation refers to multi-view angle data, which has been abandoned in favor of multi-channel data for the computation of lower troposphere temperatures.

    I'm not entirely certain how comparable the 5.6 and 6.0 UAH versions are as a result of those method changes. 

    I'm also curious to see if the new version of UAH temperatures has a significant effect on the Cowtan and Way temperature reconstruction - the changes in the UAH V6 Arctic temperature trend (by a factor of almost 2) is quite significant. 

  20. Tamino has just now shown that RATPAC balloon thermometer data of the lower troposphere does not support a "pause" in warming over the past 18 years as fake skeptics claim RSS data do.

  21. Tamino showed difference between RATPAC balloon measurements and satellite measurements, which sharply increases after the year 2000.

  22. David Arthur @ 17 "Water molecules absorb sensible energy at the surface when they evaporate, and re-release that energy at whatever tropospheric height when they re-condense. That is, intensification of the water cycle results in energy "bypassing" tropospheric greenhouse gases as it departs earth." is incorrect per the following table. Note that the *current* increasing greenhouse gas effect is mostly above the point at which 90% of water vapour resides (<~4km altitude) so there's lots of that +CO2 at 4km-~16km altitude to send some radiation back down. (the tabs didn't work out, figure it out yourself):

    alti- air air
    tude tempe- density --— atmospheric ppmv --
    km rature g/m3 CO2 water vapour
    0 16 1,290 400 14,000
    0.5 12 1,235 399 11,000 **84% already caught and shimmering around**
    1 9 1,180 398 8,500
    1.5 5 1,130 398 6,400
    2 2 1,075 397 4,900
    3 -6 965 395 2,900
    4 -13 860 394 1,700
    5 -20 750 392 1,000
    6 -27 680 390 600
    ** zone above approx. this height has more CO2 than H2O **
    ** zone above approx. this height is not "saturated" with GHGs **
    ** zone above approx. this height produces most +CO2 & +CH4 warming **
    7 -34 610 389 350
    8 -42 540 387 200
    9 -49 470 386 120
    10 -56 420 384 70
    11 -56 370 384 40
    12 -56 320 381 25
    15 -56 200 376 10 ** 37.6x as much CO2 as H2O **
    ** not much +CO2 & +CH4 warming above approx. this height, air too thin **
    20 -56 90 368 8
    25 -52 40 360 8
    30 -47 20 352 8
    40 -25 5 344 8
    50 -3 1 336 7
    60 -18 0.39 328 7
    70 -50 0.125 320 5.5 to 6.5
    80 -83 0.027 312 2.5 avge (2 to 4.5)


  23. Can anyone estimate when the RSS data is likely to show peak reaction to the current El Nino?

    So far as I can tell, during the '98 event SST's peaked in around Nov/Dec 97 and RSS peaked around March/April. To judge from the current (9th Oct) very strong WWB west of the dateline this event is not going to peak until around Dec/Jan. So am I right in assuming that the RSS data is likely to break the '98 peak temperature record sometime around April 2016?

    I'm curious as to how the denial industry will spin a new record high temperature on the RSS graph.

  24. They'll just start a new "pause" in 2015, no worries. It's not like their audience is critical anyway. You can say anything to someone who wants to believe a certain way, as long as it goes that way, no matter how unreal.

  25. Relevant to the lower troposphere and satellites:

Prev  1  2  3  Next

Post a Comment

Political, off-topic or ad hominem comments will be deleted. Comments Policy...

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.

Link to this page

The Consensus Project Website


(free to republish)

© Copyright 2024 John Cook
Home | Translations | About Us | Privacy | Contact Us