2015: record hot or not?

With the 2015 lower troposphere temperatures in, NASA and NOAA have declared without any caveats or doubts, that 2015 was the warmest year since records began. This short post updates my charts that compare satellite with surface thermometers temperatures. The differences between the two methodologies are in fact tiny and subtle. According to satellites, 2015 was the third warmest year, lagging 1999 by a fair margin. The main material difference, therefore, is that the satellite record would deprive the scaremongers at NASA and NOAA of their eye popping headlines.

Figure 1 Surface thermometer (GISS LOTI and HadCrut4) and satellite (UAH and RSS) records compared. It is plain to see that surface thermometers set a new record in 2015 while the satellites did not.

In this post I am looking at two of the surface thermometer records (GISS LOTI and HadCrut4) and the only two versions of the satellite record (UAH and RSS). The surface thermometer models are based to a large extent on the same surface thermometer data base where air temperature is measured over land and sea surface temperature (SST) is measured over the oceans. The results are area weighted with the SSTs contributing about 70% of the total. The satellite models are based on the exact same satellite recordings. We will see that there is no material difference between GISS LOTI and HadCrut4 and no material difference between UAH and RSS.

The satellite record begins in 1980 and it is only the post-1980 parts of the records that are considered here. The records use different base periods from which temperature anomalies are calculated and this hinders direct comparisons. This is overcome by rebasing each series to 1980-1984 = 0˚C. The online data sources and how to find them are detailed in the appendix.

Figure 2 The surface thermometer records from NASA GISS and UK Met Hadley are in excellent agreement. The notorious pause, 1997 to 2014 has now been busted and these records now present a picture of continuous warming across this 36 year period.

Figure 3 The satellite record as interpreted by the University of Alabama (UAH) and Remote Sensing Systems (RSS) are also in excellent agreement. As we shall see, the difference between this and surface thermometers is small and subtle. The satellite record is in two parts. 1980 to 1997 is largely flat. 1998 to 2015 is largely flat. A step change took place 1997 to 1998. The material difference with thermometers is that 2015, an el Nino year, has not come close to setting a new record.

Figure 4 Averaging the two thermometer records and the two satellite records provides this simplified picture. The regression shows that thermometers are warming at 0.45˚C per century more than satellites. This really is a trivial amount.

Figure 5 Subtracting satellites from thermometers provides this picture of diverging records. The divergence should make all scientists curious since it raises some vital questions. Is the near surface (thermometers) actually warming more rapidly than the bulk lower troposphere (satellites)? Or is there a systematic bias / error in one or both of the data recording mechanisms?

Standing back to take a broader view, one thing strikes me as very odd and that is the different response of the satellites and thermometers to the three large el Nino events in 1998, 2011 and 2015. In particular, the satellites registered much higher temperatures during the 1998 event than thermometers, reversing the normal bias. In 2011, satellites recorded slightly higher temperatures. And in the 2015 el Nino thermometers registered significantly higher temperatures than satellites.

For those thinking that my introductory comments about eye popping headlines was over the top, here it is.

NASA, NOAA Analyses Reveal Record-Shattering Global Warm Temperatures in 2015

It’s not exactly measured scientific jargon from the pride of US science and engineering institutions, is it?  The press release does at least mention satellites:

NASA monitors Earth’s vital signs from land, air and space with a fleet of satellites, as well as airborne and ground-based observation campaigns.

You’d have thought NASA and NOAA, who have one foot on the ground and the other in space, would at least have mentioned the discrepancy between land based and space based measurements and offered some explanation.

I’ll leave you with this video of Dr Gavin Schmidt (Director of Goddard Institute of Space Studies) and Dr Roy Spencer who runs the UAH satellite data set together with Dr John Christy.


Acronyms, Abbreviations and Data Sources

HadCRUT4 stands for Hadley Research Centre at the UK Met office and Climate Research Unit of the University of East Anglia and Temperature and version 4. Data downloaded from this link:


NASA is the US National Aeronautics and Space Administration. GISS is the Goddard Institute for Space Studies. LOTI is the Land Ocean Temperature Index. Data downloaded from this link:


UAH is the University of Alabama, Huntsville where Roy Spencer and John Christie are the curators of the satellite microwave data. Data for V6 were downloaded from this link:

Selecting the file tltglhmam_6.0beta3

RSS is Remote Sensing Systems, the other curator of the satellite microwave data. Data were downloaded from this link:


Clicking on FTP Air Temp Time Series and then selecting the first file called


This entry was posted in Climate change, Political commentary and tagged , , , , , , , , . Bookmark the permalink.

75 Responses to 2015: record hot or not?

  1. A few additional points or clarifications.

    Satellites do not measure temperature directly, but have to infer temperature from microwave brightness. This requires some serious calculations. Carl Mears, who is responsible for the RSS data, says that his analysis of his own satellite data has five times the margin of error of ground measurements (in this news item).

    As Tamino showed, balloon data (which uses thermometers) deviates from RSS around 2000. I’ve seen a comment that suggests the switch from MSUs to AMSUs around that time may play a part.

    So should we go by surface measurements, as much as possible? After all, this is where people and other species live, not many thousands of feet up in the troposphere (which is what the satellite data provide estimates for). There are many surface temperature datasets (more than the 2 mentioned here) and they all give 2015 as the record year, by a wide margin. Carl Mears, again, from that linked article: “The satellite measurements do not measure the surface warming. They are measurements of the average temperature of thick layers of the atmosphere (about 50,000 feet off the ground). … For impacts on human society and the environment, the surface data are more important.”

    On El Nino, the current one has been compared to the previous big one in 1997/1998. In that previous one, the second year was significantly warmer than the first. The satellite data seemed to lag the ground effects by several months. If the current El Nino follows the same pattern, we should expect this year to be even warmer (perhaps by a big margin) and the satellite data to show a significant spike this year.

    • michael hart says:

      “So should we go by surface measurements, as much as possible? After all, this is where people and other species live, not many thousands of feet up in the troposphere (which is what the satellite data provide estimates for).”

      The troposphere is where the effects of CO2 emissions should, according to the theory, have their largest measurable effects. Satellites and radiosondes measure temperature there. Surface thermometers do not.

      • From the radiosonde data, it looks like the troposphere is warming faster than the surface. The radiosonde measurements and the satellite temperature calculations don’t seem to match, though.

        However, the surface is where we live, so we should take more notice of surface measurements, IMO.

        • David n says:

          Should we also discard ocean data? Since few to no humans live on the ocean? And last I checked, the problem with climate change isn’t “humans have to deal with a little bit warmer temperatures,” it was “This will destroy our natural environment.” And the Troposphere is part of “The Environment”.

          • Mike Roberts says:

            Humans should be more concerned with land surface temperatures but we’re part of a complex ecosystem thar exists in a thin layer around the surface of this planet.

    • Dave Rutledge says:

      Hi Mike,

      “Satellites do not measure temperature directly, but have to infer temperature from microwave brightness.”

      It is clear from following this discussion over the last twenty years that the only reason there is concern about the satellite temperature calculations is that they are not under control of the climate change advocates. It is a political objection rather than a scientific one.

      In my experimental work over the past 40 years, I have had many occasions to measure temperature by the electromagnetic radiation emissions (brightness) and by contact measurements (thermometers). They both can be appropriate ways to measure temperature and they both need careful calibration to make accurate measurements.

      Ordinarily my preference is for measuring the temperature by the emitted radiation, precisely because it avoids the many errors associated with contact measurements that include poor sampling, convection effects. contact thermal resistance, and the heat capacity of the thermometer. Many scientific and engineering fields use the emitted radiation to measure the temperature including radio astronomy, microwave background measurements, oven measurements and electronic circuit inspection

    • Javier says:


      We have no way of measuring temperatures directly, as temperature is simply a kinetic state of atoms. So a liquid thermometer (liquid expansion) is as indirect as a measure of conductivity or radiation.

      Surface thermometers have some advantages, and are the best way to determine local temperatures. Satellites are the best way of measuring planetary temperatures due to their nearly global coverage and immunity to certain biases like Urban Heat Island.

    • ducdorleans says:

      Mike, hi …

      ” If the current El Nino follows the same pattern, we should expect this year to be even warmer (perhaps by a big margin) and the satellite data to show a significant spike this year.”

      you make a predicton here, one that we can check in a year’s time … I hope we will be around to try and do that !

      • Not so much a prediction as an observation. If past el Nino behaviour is a guide then what I wrote is exactly what we should get. Not trying to step back from that but if we don’t get the same as 97/98, or something very similar, then that would be useful information in itself and would, hopefully, spur some research into why the two big el Nino’s caused different results.

        But, yes, we can certainly check what happened, in a year’s time. I’ve got no problem with that.

        • Euan Mearns says:

          I’m with Mike on this one and see a distinct possibility that the satellite temperatures have not yet peaked. To be clear Mike, are you saying that you expect the behaviour either side of a big el Nino to be the same?

          • Not sure what you mean by “either side”. As an el Nino unfolds, I’d expect it to have similar impacts as previous el Ninos, though I think I’ve read somewhere that the size of the el Nino could cause it to behave differently. As far as I’m aware, the current one is a similar magnitude to the 1997/1998 one, so I would expect similar effects. Having said that, I’m no expert, so just going by the last big one; maybe there are other factors.

  2. Excellent analysis of a topic that the press love to ignore, apart from “records”.

    The long term measurements show a gradual increase in temperature since about 1850; about .6 degK / century I think. Would it help to use this long-term trend and look for deviations from that as even the “hiatus” debate ignores the long term trend?

    • Euan Mearns says:

      This post endeavoured to do just that:


      Figure 1 De-trended HadCRUT4 compared with the AMO index that is based on de-trended NA SSTs. AMO index from NOAA. (AMO unsmooth long)

      Figure 2 HadCRUT4 temperature data oscillate around the liner regression (solid red line). Subtracting the linear regression from HadCRUT4 de-trends the data as shown in Figure 3.

      Running the regression from 1980 (satellite era) gives a false picture of long-term warming since it just catches the latest up-leg.

      • Nial says:

        Figure 3?

        • Euan Mearns says:

          Figure 3 Subtracting the linear regression from the actual HadCRUT4 data provides this distribution of residuals and the quasi 66 year cycle emerges. The dates for the peaks and troughs are labelled. Starting in 1878+33=1911+33=1944+32=1976+33=2009.

          The next cold trough should be in 2042 (I’ll be 85 years old). With La Nina due, the AMO set to change and the Sun falling sound asleep, the stage is set for the battle between hot and cold to be played out.

      • Yvan Dutil says:

        The way AMO is defined, it is not orthogonal with global temperature. Hence, if you compare AMO with global temperature, you will get spurious correlation.

        • Euan Mearns says:

          Yvan, can you explain a bit more please.

          • Sam taylor says:

            What he means is that the amo index is basically defined as the residual of a the detrended Atlantic temperature time series. The two time series are not independent of one another, so it’s not surprising there’s a correlation. In a statistical sense the term “orthogonality” refers to whether or not the dot product of two variables is equal to zero. If it is, then they are independent.

            The other issue with your analysis is that if the warming signal is nonlinear (likely) then by only removing a linear tend some of that signal will leak into your amo series.

          • Euan Mearns says:

            You should write to NOAA and tell them they got it wrong.

        • Euan Mearns says:

          Here’s what NOAA have to say:


          What is the AMO?
          The AMO is an ongoing series of long-duration changes in the sea surface temperature of the North Atlantic Ocean, with cool and warm phases that may last for 20-40 years at a time and a difference of about 1°F between extremes. These changes are natural and have been occurring for at least the last 1,000 years.
          How much of the Atlantic are we talking about?
          Most of the Atlantic between the equator and Greenland changes in unison. Some area of the North Pacific also seem to be affected.
          What phase are we in right now?
          Since the mid-1990s we have been in a warm phase.
          What are the impacts of the AMO?
          The AMO has affected air temperatures and rainfall over much of the Northern Hemisphere, in particular, North America and Europe. It is associated with changes in the frequency of North American droughts and is reflected in the frequency of severe Atlantic hurricanes. It alternately obscures and exaggerates the global increase in temperatures due to human-induced global warming.

  3. JerryC says:


  4. wehappyfew says:

    RSS and UAH are not the “only two” versions of the satellite record. See UW (Po-Chedley) and NOAA STAR. Their versions differ significantly from the other two.

    Also, you said:

    “Standing back to take a broader view, one thing strikes me as very odd and that is the different response of the satellites and thermometers to the three large el Nino events in 1998, 2011 and 2015. In particular, the satellites registered much higher temperatures during the 1998 event than thermometers, reversing the normal bias. In 2011, satellites recorded slightly higher temperatures. And in the 2015 el Nino thermometers registered significantly higher temperatures than satellites.”

    This is not hard. El Ninos peak in austral spring and summer – straddling two calendar years. The first calendar year, surface temps are often higher than troposphere temps. the 2nd year, troposphere temps rise more than surface temps, often (but not always) peaking much higher than the surface maximum.

    The 3 largest El Nino events are:
    … 1997/1998 …the current one (2015/2016)… and 1982/1983

    (see: http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml )

    2009/2010 was a medium El Nino event and also follows this pattern.

    The peak troposphere temps of the current El Nino will occur in the next few months. Unless it fizzles out extremely rapidly, or a huge volcano intervenes, 2016 will follow the same pattern – record high troposphere temps and surface temps even higher than 2015 (with the troposphere much higher than the surface).

    See: http://www1.ncdc.noaa.gov/pub/data/ratpac/ratpac-a/RATPAC-A-annual-levels.txt

    Which brings up the radiosonde temperature data. When the troposphere is measured with actual thermometers, the data at the link above show 2015 already reaching record highs at the 500 and 400 mB levels, and at the surface. 2016 will be even higher, especially in the mid-troposphere, due to the unfolding El Nino.

    • A C Osborn says:

      I am really glad that you mentioned NOAA STAR as Paul Homewood has plotted the data from that source and you are correct it does vary considerably from the other 2 Satellite sets and the Surface based Thermometer sets as well.
      In fact it appears to show that not only is 2015 not the hottest year ever, it actually shows that one level is Cooling from 1980, one level shows virtually no movement at all since 1980 and one level shows No increase from 1980 to 1997/8 then a step change for the el nino and then no increase to date.
      So we have on level showing COOLING, one level a PAUSE for the whole period and one level a STEP Change for the El nino.

      Kind of suggests that the Sattelites confirm that the whole Surface based Thermometer recent upward Trends are due to MAN MADE ADJUSTMENTS to the original Raw Land and Sea Surface Data.
      But then we already new that.

      The only dataset, with the best Weather Stations, that has not been adjusted also shows no upward trend for the contiguous USA.

      • wehappyfew says:

        I’m glad you are looking at the data. Now you need to work on understanding what it means. Do some research on “stratospheric cooling” due to greenhouse gasses. Also, “stratospheric contamination of the TMT” (Mid-Troposphere) temperature estimate.

        For example:


        Your enthusiasm for “raw data” is so adorable. Did you ever wonder what the “raw data” for the satellite temperature calculations look like?

        The satellites measure microwave brightness, not temperature. Converting it into a temperature estimate requires dozens of adjustments, calibrations, and corrections. It even requires the input of a GCM (Global Climate Model) to be able to accurately adjust the diurnal drift of the satellite measurement time as the orbit decays.


        See figure 3 to get an idea of what the “raw data” look like for the satellite microwave data before and after adjustments, correction, etc..

        Now look again at the global radiosonde data:


        Those are real temperature measurements made with actual thermometers, not estimates from microwave sounders.

        Actual thermometer temperature measurements of the mid-troposphere show it is warming faster than the surface, exactly as predicted by the GCMs.

        • Euan Mearns says:

          And thermometers are usually measuring the volume expansion / contraction of mercury, sometimes in the Sun, sometimes in the shade, sometimes max / min, sometimes 11:00 am. Sometimes in a bucket, sometimes in cooling water getting pumped around a steel hulled ship.

          Here’s the plot that AC refers to posted by Paul Homewood:

          What we see in the T Lower Stratosphere (TLS) is a completely flat line since about 1993. And the chart of TLT (temperature lower troposphere) shows 1.2˚K per century, identical to what I report here from UHA and RSS. Now you have told us that NOAA STAR differs significantly from “the other two”. Evidence please, otherwise your credibility is going to go down the tubes, which is a pity since you are providing links to some interesting sources.

          The NOAA microwave sounding unit page simply links to UHA and RSS:


          You need to note that in this post I do not criticise the thermometer record at all, which makes it more interesting to note your desire to defend it. You cite multiple NOAA sources of satellite measurements – do you not think they should be obliged to at least refer to these in preference to sensational headlines?

          • wehappyfew says:

            I totally agree that accurate, long-term trend estimation using thermometers is very hard. All those effects you mention require adjustments, corrections, homogenization, etc.

            It’s obvious from the multiple version changes by UAH and from the very different trends estimated by the different satellite microwave data groups that satellite trends are much, much harder; and have far higher uncertainty.

            The RSS data series author Dr Mears agrees that satellite uncertainty is much higher, and has published his best estimates of all the many sources of errors in the satellite temperature estimation process – he found the uncertainty/confidence interval to be much wider for the satellite data compared to the surface thermometer readings.

            UAH does not publish any estimate of their uncertainty/confidence interval. I wonder why?

            Then you provide us with a beautiful example of confirmation bias…

            You looked at Paul Homewood’s data plot from NOAA STAR which is clearly labelled TMT…

            Then you compared it to your plot of TLT data from RSS and UAH v6.0.

            That’s with the letter “L” …. not “M”

            You saw what you wanted to see, which is that the NOAA data corroborate the RSS and UAH TLT data with similar trends of 0.12K/decade. They do not, because they are not measuring the same layers of the atmosphere. Investigate the difference between TMT and TLT, look at the RSS and UAH trends for TMT and TLT, compare to the NOAA trend for TMT (NOAA does not compute a TLT synthetic series).

            Then look at Po-Chedley’s computed trends for TMT compared to UAH or RSS. It’s in the link I provided in my 2nd comment.

            Paul Homewood correctly compared RSS to NOAA. He showed the RSS TMT trend of 0.08K/decade and the NOAA STAR TMT trend of 0.12K/decade. He did not deign to notice the discrepancy, nor did he comment on the radiosonde data he posted in his last graph that showed record high temps for 2015 at the 500mb level of the troposphere. That did not fit the narrative of his article that tropospheric temps are still “paused”.

            As far as sensational headlines… the surface temps spiked this year so that’s what they talked about – the troposphere temps will spike next year. I’m sure NASA and NOAA are aware that the satellite/troposphere temps don’t normally spike until the 2nd year of the El Nino… maybe they are saving that for another sensational headline next year.

        • wehappyfew:

          Take out the gratuitous insults and that was quite a valuable contribution – it actually presents some data. I have just a couple of comments.

          Did you ever wonder what the “raw data” for the satellite temperature calculations look like?

          I’ve been analyzing raw data for over 25 years, both professionally and for the hell of it, and the UAH raw data shown in Figure 3 of the article you link to is some of the cleanest I’ve ever seen. It’s certainly a lot cleaner than the spaghetti tree ring and other proxies that paleoclimatologists have no problem converting into hockey sticks:

          Actual thermometer temperature (RATPAC radiosonde) measurements of the mid-troposphere show it is warming faster than the surface, exactly as predicted by the GCMs.

          You’ve got it backwards:

      • A C Osborn says:

        It is quite interesting that you think I need to learn what the data means and how it is derived in the first place.
        You have an extremely condescending attitude to other posters, without having any clue as to their backgrounds or knowledge.

        I looked in to Sattelite Temperature, Cloud Cover and Ice Cover Instrumentation years ago.
        I have also spent quite a few years looking at Surface Thermometer data and adjustments as well, including the so called “BEST” dataset
        I have been watching the ever changing Temperature History being re-written until it bears no resemblance to the original data.
        The original data has masses of Anecdotal Evidence to back it up, whereas the “Adjustments” are arbitrary Algorithms that are used ad hoc and based on extremely poor evidence of “site” characteristics, classification and quality.
        Add to that the replacement of “good raw data” with “Estimated” data and the addition of Estimated data for periods when none existed means that I have absolutely no faith in the quality of the final output.

  5. Pingback: 2015: record hot or not? – Olduvai.ca

  6. donb says:

    Here is the main issue. Troposphere temperature measurements are the most accurate because of relatively large changes in temperature and because of their widespread and rapid sampling via satellite. But the atmosphere contains the least proportion of Earth’s solar heat budget. The oceans contain by far the greatest proportion of Earth’s heat budget, but their temperature variations are quite small and difficult to measure, and the ocean temperature is not well monitored. Land temperature falls intermediate in these characteristics between ocean and atmosphere.

  7. Javier says:


    I disagree that the difference between satellites and thermometers are tiny and subtle. The degree of agreement between both since 1979 to 2005 is surprisingly good, specially if we make room for the extra warming and cooling of the atmosphere during strong El Niño and La Niña years.

    Since 2005 thermometers and satellites start to diverge and this difference is becoming ridiculous. We know that this differences are due to changes introduced to surface temperatures, because HadCRUT3 continued agreeing quite well with satellites until it was replaced by HadCRUT4 to get in line with the rest.

    This divergence cannot continue much longer. We are already witnessing attacks to the credibility of satellites like this recent 10 min. video:
    How reliable are satellite temperatures?
    perpetrated by Michael Mann, Ben Santer, Kevin Trenberth and Andrew Dessler.

    If in 2017-18 we get a strong La Niña as sometimes happens after a strong El Niño, satellites could show not only the continuation of the pause, but even some cooling for the entire 21st century.

    So what we have is 26 years of very good agreement between satellites and thermometers, followed by a continuation of the pause in satellites and a return of global warming to thermometers after their databases were modified.

    • wehappyfew says:


      “We know that this differences are due to changes introduced to surface temperatures, because HadCRUT3 continued agreeing quite well with satellites until it was replaced by HadCRUT4 to get in line with the rest.”

      Of course, we do not know this at all. The satellite data sets have been amended, corrected, and revised numerous times. UAH v5.6 agreed quite well with the surface temp series. Now v6.0beta does not. Previous versions had wide swings in their trends, both up and down, over the entire history of satellite data collection. This is not new, this is not a recent “attack” against the credibility of satellite data sets… it is just the normal process of science working to minimize errors and discrepancies as they are found – and lots of errors and drifts have been found.

      The satellites data sets are routinely calibrated against radiosonde data. Since the trends from the radiosondes have been recently diverging from the satellites, no doubt more adjustments to the satellites are in the works in the coming years to bring them back to the ground truth as measured by actual thermometers in radiosondes. Then you will squawk about the outrageous manipulation of the “raw data” when the satellite trends swing the other way in the next version.

      Also, the replacement for AMSU… ATMS, I think?… has been in orbit for what? … 4 years?… is already producing data… but the process of calibrating it against AMSU and integrating it into the UAH and RSS data sets is “in progress”. Satellite temperature measurement is very hard. After the hot target calibration fiasco for AMSU vs MSU, they may cogitate on that process for quite a while.

      If you want to speculate on the future trajectory of tropospheric temps, how about a bet that the average of satellite and radiosonde lower tropospheric temps will reach a new high in 2016, and the surface temps will be slightly higher than 2015?

      Let’s say tropospheric temps 0.20K higher than 1998 for the over/under.

      2nd bet: any upcoming La Nina will bottom out 0.30K higher than the 1999-2001 La Nina – same average of radiosonde and satellite data.

    • Euan Mearns says:

      I of course agree with what you say here. If the divergence continues something will have to give.

  8. Jim Brough says:

    While you are talking about the mathematical interpretation of the recent temperature records and its meaning. There are many, many angels dancing on the head of a pin called catastrophic anthropogenic global warming caused by CO2 emissions. That morphed into climate change.
    Climate change has always been part of the recent history of the earth, regardless of the CO2 concentration in the atmosphere.
    It is obvious from the depiction in 6,000 yr old caves of giraffes and other animals which no longer exist in the desert of Northern Africa.
    Climate change long before the discovery of coal as an energy source.

    • Mike Roberts says:

      That climate always changes is not really the point. It’s happening very quickly now and primarily caused by human behaviour. The rate of change is one of the main problems

  9. wehappyfew says:

    … thank you for the kind words,

    You do not give a data source for your graph, nor do you calculate a regression slope.

    Using RATPAC-A, I get these trends for the full global dataset:

    Surface = 1.4K/century
    850mb = 1.6K
    700mb = 1.5K
    500mb = 1.6K
    400mb = 1.7K
    300mb = 1.3K

    Over shorter times spans, one can find greater or lesser trends, including some with surface greater than troposphere.

    The important congruence with GCMs is, of course, for the tropical tropospheric hotspot, not the full global average. That was a mistake on my part. The ratio of the troposphere/surface trends for the tropics only is much stronger.

    Surface 1.1K
    850mb 1.6K
    700mb 1.6K
    500mb 1.7K
    400mb 1.9K
    300mb 2.0K

    This hotspot amplification is confirmed by Po-Chedley’s work using the Satellite data. (see Table 4: http://www.atmos.uw.edu/~qfu/Publications/jtech.pochedley.2015.pdf )


    Overall, the big picture remains the same. Many lines of evidence corroborate continued global warming:

    * Stratospheric cooling due to GHGs as predicted by radiative transfer theory (slows IR heat loss to space).

    * Ocean heat content at record high, rate of accumulation is accelerating, no pause evident.

    * Ocean surface temps at record high, after a pause from 2002 to 2013.

    * Global Mean Sea Level is rising, and accelerating.

    * Surface thermometer series are at record high.

    * Radiosonde series is at record high for the surface.

    * Troposphere radiosonde data at or near record highs.

    * 2 out of 4 satellite datasets show a trend consistent with radiosonde and surface datasets, but not yet at record highs.

    * 2 out of 4 satellite datasets show a weak trend, not consistent with the others.

    * The clock is ticking for the release of accumulated ocean heat due to El Nino. This will push all the tropospheric data series into record territory by a wide margin in the next few months.


    All the anguish about the record highs at the surface versus no records in some of the troposphere data is due to the invalid comparison of 1998 data dominated by El nino to 2015 data which doesn’t show the El Nino influence yet. The better comparison is between 1997 and 2015.

    1997 is to 2015… as 1998 is to 2016.

    Extrapolating, assuming the current El Nino is exactly as strong as the 1997/1998:

    Year….. GISS… RadioSurf.. RadioTropo
    1997….. 0.47…… 0.26 ………. 0.15
    1998….. 0.63…… 0.44 ………. 0.74

    2015….. 0.87…… 1.02 ………. 0.80
    2016est..1.03….. 1.20 ………. 1.39

    I think this simplistic extrapolation is a bit too high, for two reasons:
    1. It looks like this El Nino is slightly weaker than 1997/1998
    2. El nino 1997 started later and from a lower ENSO starting point than 2015, so 2015 got a bit of a head start relative to 1997.

    So reduce the 2016 estimates by a tenth of a degree. That’s my guess.

    2016est..0.93….. 1.10 ………. 1.29

    • You do not give a data source for your graph, nor do you calculate a regression slope.

      The data source was of course the one you kindly supplied me with, although I probably should have specified that the plot was for global data. And we don’t need a regression line to see where the warming is.

      This hotspot amplification is confirmed by Po-Chedley’s work using the Satellite data.

      This conclusion is disputed, but I don’t have time to look up links.

      All the anguish about the record highs at the surface versus no records in some of the troposphere data is due to the invalid comparison of 1998 data dominated by El nino to 2015 data which doesn’t show the El Nino influence yet. The better comparison is between 1997 and 2015.

      There’s a lot more to it than that. Maybe a post coming up.

      The rest of your comments were what I refer to as “reciting the AGW liturgy”. Similar recitations have appeared in comments several times before on this blog without so far as I know converting anyone.

      • wehappyfew says:

        Thank you for responding, and for your implicit acknowledgement that my data and trend regressions are correct (though I still don’t understand why your data points do not match up with RATPAC-A Global data values). Eyeballing a graph is better for you, maybe, but I prefer a more rigorous treatment of data.

        I look forward to your next post on these data – I am ready to be “converted” by cogent analysis of all the available data.

        Specifically, I am curious how you will reconcile the strident claims by commenters here that RSS and UAH are more accurate and all other data are suspect with the logical progression of:

        Energy imbalance at TOA -> heat accumulation (mostly in the ocean) -> higher SSTs (but with large medium term variability) -> heat transfer from ocean surface to atmosphere.

        Implicitly, those who acknowledge that SSTs are at record high should logically expect the surface air temps and eventually the troposphere to follow along, and also reach record high temps. The fact that most data series confirm this correlation should be a compelling reason to trust the consilience of the majority of the data, and to distrust the few data series that disagree – RSS and some versions of UAH.

        Otherwise, they are arguing that the atmosphere temps can move independently of SST, despite their very high correlation in the past, and their obvious connection through evaporation/condensation/convection across an extremely large surface area.

      • Javier says:


        If you make a post about the lack of records in satellite temperatures, the difference in trends since 2003 would be a good topic to discuss.


  10. Euan Mearns says:

    This is a work in progress. Can someone remind me why GHGs cool the stratosphere.

  11. Pingback: 2015: Record hot or not? | Climate Collections

  12. wehappyfew says:

    Saw this in the comments on Nick Stoke’s site



    from ehak …

    “Olof: Agree. When everything else disagree, there is very slim reasons not to conclude AMSU on NOAA-15 has a wrong diurnal drift correction. There might of course be other biases. I believe (that is: I think I remember) that UAH5.6 did not have a diurnal drift correction because they used the non-drifting Aqua. When the Aqua AMSU could no longer be used, they reintroduced a diurnal drift correction.

    A smoking gun for me in this is the divergence between land-stations and TLT land:

    Diurnal drift corrections have most effect over land. And no: I do not think the land stations can poosibly have such a big warming bias after 2000.”

    Olaf’s reply:

    “Ehak, UAH v6 TLT and RSS TLT are misfits over oceans as well, not with each other, but with all other indices, defined by the lack of trend during the AMSU-era.


    This reinforces/supports my point that multiple lines of evidence and data point to record new highs. There is consilience between many different indices, measured many different ways. Water vapor measured by RSS using satellites, and SSTs measured by buoys and ships are both at new records highs. These two must move in lockstep, there is no way they can diverge – they are intimately linked by evaporation at the ocean surface. The ocean troposphere temps are linked to those two by convection/lapse rate, and the relatively constant relative humidity.

    If UAH v6.0beta and RSS TLT-ocean temps are not tracking water vapor and SSTs, then there is a problem.

    • Euan Mearns says:

      I’m not sure if you have read my post and watched the great vid? It is as much a complaint about warmists salivating about record high temperatures and the disaster this is gong to bring to Mankind than anything else. And as a commenter here you are lead salivator. While I have respect for Nick Stoke’s (though he allows unmoderated dross comments on his blog) do you really think we give a toss about what “ehak” and “Olaf” have to say?

      Simply saying things here without supporting links to data and charts carries now weight.

      I strongly suggest you start posting under your own name (as opposed to the clown pseudonym) and provide a brief CV of who you are and the technical expertise that you bring to this complex debate.

      I’ve been approving virtually all comments from moderated commenters (and deleting a few from unmoderated) because its fair to say that you and others do bring important sources and balance to the debate. And, for example, I’ve spent the last 2 days looking at RATPAC A data that is far more complex than warmist salivators would have us believe.

  13. Mike Roberts says:

    This article echoes Carl Mears’s comments about the uncertainties in the satellite data. The only reason I can think of for why contrarians prefer the satellite data (or two such sources, anyway) is that they appear to show what they want to see.

    • javier says:

      I don’t see why your argument cannot be used backwards. The only reason I can think of for why warmists prefer the surface data (or several such sources, anyway) is that they appear to show what they want to see.

      I think you have to sit down and analyze where the differences are coming from to reach an understanding. It is surprising to see that UAH v6 and GISSTemp LOTI agree quite well in the 55N-55S band, and both agree quite well also with CERES Ebaf ToA Outgoing Longwave Radiation (OLR).

      Figure 4. In this post: https://okulaer.wordpress.com/2016/01/27/why-gistemp-loti-global-mean-is-wrong-and-uahv6-tlt-gl-is-right/

      If we move to the 55N-90N band, then we see that most of the differences between GISSTemp LOTI and UAH v6 warming come from warming in the polar regions, where surface temperatures rely on a series of assumptions, for example GISSTemp LOTI does not use SST data and extends land data over the entire Arctic region from a very few measurements.

      But this time the CERES Ebaf ToA OLR agrees with the warming measured by UAH v.6 and not with GISSTemp LOTI polar assumptions. See figure 6 at the same link.

      So the facts are:
      – Since the pause started surface records have been warmed exclusively through polar warming that is debatable since it is not based on measurements.
      – Satellite warming (or more precisely lack of) agrees well with partially independent techniques like ECMWF Era interim reanalysis, or independent techniques like satellite measurements of OLR.

      I just don’t see why anybody would choose to trust better a warming that comes mainly from areas where we don’t measure temperatures and that doesn’t agree with the Earth’s radiative budget, unless is that it appears to show what they want to see.

      Another blow is that during the pause all the ocean warming according to the Argo system is coming from the 20-60°S oceans. 20°S-60°N oceans show no warming. See figure 5 in:

      Fifteen years of ocean observations with the global Argo array

      Very surprising. According to the Argo evidence warming on the oceans is taking place only in the Southern Hemisphere. According to surface records warming is taking place only at polar regions. This is looking less and less as a warming planet.

      • I don’t think the argument can be turned around. I’d have to disregard multiple surface temperature datasets from multiple organisations and multiple nations, plus several satellite datasets (as mentioned by another commenter) and the radiosonde data, I’d have to disregard the ocean warming, sea level rise, melting of ice sheets and glaciers, plus a few other pointers, to believe that the two satellite data sets showing little warming (some believe no warming) over the last 18 years or so are the best record of surface warming. The only reason I can think for doing that is because those datasets tell a story that I want to believe. Now, I don’t think the reverse argument can be made, as I’d only need to disregard the UAH dataset (since Carl Mears of RSS states that the surface record is the best).

        Interesting post you linked to though I couldn’t find a source for the GISTEMP LOTI data between 55N and 55s (the GISTEMP LOTI data only seems to be available for different latitude bands), so I could do any checking. However, as I mentioned, there are multiple lines of evidence that the planet has continued to warm over the last 18 years, so that’s what I must continue to conclude.

        • javier says:

          Mike Roberts:

          “I’d have to disregard multiple surface temperature datasets from multiple organisations and multiple nations, plus several satellite datasets (as mentioned by another commenter) and the radiosonde data, I’d have to disregard the ocean warming, sea level rise, melting of ice sheets and glaciers, plus a few other pointers, to believe that the two satellite data sets showing little warming (some believe no warming) over the last 18 years or so are the best record of surface warming.”

          That is actually not correct. The elimination of the pause from the surface temperature datasets is not based on data but on data analysis and therefore due not to what thermometers are showing but to what some scientists judge it to be the best way to analyze that data. The most clear example is the paper by Karl et al. 2015 “Possible artifacts of data biases in the recent global surface warming hiatus”.

          These new interpretations of the data are quite controversial between climate scientists. See for example:
          Has NOAA ‘busted’ the pause in global warming?

          So the choice is not as you presented it. The choice is between data that is being collected in the same way and treated in the same way and that has given the correct warming for decades (satellites, RSS) or data that is now being treated in a different way and interpreted in a different way that is controversial.

          And so you get it right, satellite data is not alone in displaying a prominent pause. ECMWF Era interim reanalysis data is also showing it, and this is the best weather prediction tool in the world, with a much better surface temperature coverage than NOOA, GISS or BEST surface temperature GHCN based datasets, and a critical mission since weather prediction is used by planes and ships and lives depend on it. And Outgoing Longwave Range satellite measurements also agree with satellites in that there is a pause.

          • As I understand it, the ECMWF interim data are a combination of observations and models. Has anyone enquired of ECMWF whether it’s a good source for looking at warming trends? Yours is the first mention I’ve seen of it, among contrarians. I can’t think why.

            All data need adjustments and that link has a graph of raw NOAA data overlaid with adjusted data. The warming trend is reduced after adjustment and no pause is visible in the raw data. At what point would you like data to stop being adjusted?

            As I say, I’d have to disregard all other lines of evidence for the lack of a pause, and also disregard what the RSS scientist says about his data, to accept RSS data as the best there is for determining surface temperature.

          • javier says:

            Mike, I do not see why we should ignore every possible source of information. I am not saying that ECMWF ERA data is better or more adequate, because I don’t know it, but it is our state of the art weather prediction tool and its source of data is both independent, inclusive and differently treated than both surface records and satellite records. It is also a multinational effort from like 22 countries.

            It shows a rate of warming slightly higher than satellites:

            The surface record is reconstructed from too many irregular sources, so yes it needs adjustments, but when most of the warming is coming from the adjustments, then we have a problem.

            Climate4you tracks the difference in anomaly for Jan 1910 and Jan 2000 since April 2008 for several databases. In the case of GISS that difference was 0.45°C in April 2008, but by January 2016 that difference has grown to 0.7°C. In only 8 years of the 21st century when barely no warming took place, the warming in the 20th century had increased by 0.25°C. This gives a whole new meaning to anthropogenic warming.

            Regarding the pause, the problem is that scientific consensus is that the pause exists and is real. For example you can check this Nature Climate Change Focus: Recent slowdown in global warming, where you can find over 30 editorials and articles just from the journals Nature Climate Change and Nature Geoscience on the pause, most of them from 2013 and 2014. So you can calculate the huge bibliography that exists on the pause.

            Are you going to disregard all the scientific evidence published over the years on the pause? Most scientists clearly are not.

          • So, are you saying that there has been a slowdown in surface warming over the last 18 years, or that there has been a pause over the last 18 years? The ERA-Interim data set isn’t purely observations but has “data” filled in by a model which is constrained by observations. But even that shows a warming trend. Heck, even the UAH and RSS satellite data sets show a warming trend (though much lower). The UAH trend lowered after the last set of adjustments (still in beta), so should we disregard that, because it’s now different from what it showed previously? As I said, adjustments are necessary and some of recent adjustments to the SST data have actually decreased the long term warming trend, whilst almost eliminating the recent slowdown. So it goes. Out best surface temperature data sets now show that there was never even a surface warming slowdown (or just barely), though surface temperature isn’t the only metric, of course, and the ocean heat content seems to have risen even more strongly during the surface slowdown.

            I can’t do the adjustments and analysis myself so I rely on others. At the moment, it looks like there was never any slowdown in surface warming and certainly no slowdown in global heat content increase. You may think otherwise. I guess it doesn’t really matter, because I don’t see that any significant actions will be taken to forcibly slow the increase or to limit warming to 2C over pre-industrial. I suspect the lack of action will please the contrarians more than those who accept the consensus but at least it will give us a chance to see what actually happens!

  14. Pingback: RATPAC – an initial look at the Global Balloon Radiosonde Temperature Series | Energy Matters

  15. Pingback: Surface versus satellite; the temperature data set controversy | Energy Matters

Comments are closed.