Social Channels


Additional Options

Date Range

Receive a Daily or Weekly summary of the most important articles direct to your inbox, just enter your email below:

Schareck ski region in Austria
View of the Schareck ski region in Austria. --- Image by © Gonzales Photo/Christoph Obersch/Gonzales Photo/Corbis
3 December 2016 13:13

Factcheck: Newspaper claim about global temperature is ‘deeply misleading’

Zeke Hausfather

Zeke Hausfather

Zeke Hausfather

Zeke Hausfather

03.12.2016 | 1:13pm
Global temperatureFactcheck: Newspaper claim about global temperature is ‘deeply misleading’

This is a guest post by Zeke Hausfather, a climate scientist and energy systems analyst at Berkeley Earth, an independent temperature analysis project.

It is all but certain now that 2016 will shatter historical records to be the warmest year ever by a wide margin. It was helped along the way by a large El Niño event, which tends to be associated with warmer temperatures globally. But, even without El Niño, 2016 would likely still be the warmest year ever. Now that the El Niño is fading, temperatures are dropping modestly down to around where they were before the El Niño started.

Recently, journalist David Rose published a deeply misleading article in the Mail on Sunday, a UK tabloid, claiming, “Global average temperatures over land have plummeted by more than 1C”.

He also argued that 2016 (and 2015) would not have been particularly warm years in the absence of El Niño, and that El Niño might be responsible for much of recent warming. These claims are incorrect. They are prefaced on cherry-picking an obscure temperature record, whose creator suggests it “should be used with caution” and which disagrees with other estimates by independent groups.

In reality, 2014, 2015 and 2016 have been the three warmest years on record not because of a large El Niño, but because of a long-term warming trend driven by human emissions of greenhouse gases.

The modest decline in temperatures in recent months from the peak of the El Niño event is completely in line with what has happened during past large El Niño events and was expected by scientists.

To better understand what’s going on with the Earth’s temperature, lets take a look at the various temperature records and what they tell us.

Global surface temperatures

We live on the surface of the Earth. It should come as no surprise that the easiest and most accurate way to measure the Earth’s surface temperature is with thermometers.

Various groups have put together long term temperature records using data from weather stations, ships, and floating ocean buoys. They largely agree, despite using differing approaches where data is lacking, and using somewhat different underlying datasets. Temperature records from five groups are shown in the figure below: NASA, NOAA, the UK’s Met Office Hadley Centre/UEA, the non-profit Berkeley Earth, and a record by independent researchers Cowtan and Way.

Global average surface temperature, 1879-2016

Global average surface temperature, 1979-2016

Here we can clearly see both the long-term warming trend over the past 30 years, as well as the effect of El Niño events (the small jumps up) and their sister La Niña events (the small jumps down).

The large El Niño of 2015/2016 clearly stands out, but it’s clear that the overall record is warming quickly, even without the El Niño event. In the past few months temperatures have fallen a bit from the El Niño peak, but are still quite high by historic standards.

Global tropospheric temperatures

In addition to the surface temperature record, a number of groups have attempted to put together a record of changes in atmospheric temperature using data from satellites. These satellites indirectly measure the temperature of the lower troposphere, a layer of the atmosphere about 2 miles (4 km) straight up in the sky.

These will necessarily differ a bit from surface records, since we wouldn’t expect the temperatures of the surface and miles up in the atmosphere to necessarily be identical.

Creating a reliable temperature record from satellites has proven somewhat difficult. There has not been a single satellite taking measurements since 1979; rather, new satellites are launched every few years as old satellites’ orbits decay and they fall out of the sky.

As the orbits change, the temperature readings from satellites will change a bit, and need to be adjusted. Similarly, when new satellites are launched their instruments need to be calibrated against the readings from old satellites. All of these introduce uncertainties that appear much larger than those in the surface record, as discussed in the video below:

Three groups provide satellite-based estimates of global lower tropospheric temperatures: UAH, RSS, and NOAA’s STAR. All use the same underlying satellite readings, but apply different adjustments and models to account for satellite transitions and orbital drift. The global average lower tropospheric temperatures from the three groups since 1979 are shown below:

Global average lower troposphere temperature

Global land-ocean lower tropospheric temperature anomalies by month, January 1979 through November 2016.

El Niño events tend to show up more prominently in tropospheric temperatures than in surface temperatures. We can see this in the figure above, where the spikes during the El Niño events (particularly the 1997/1998 event) are much larger than in the surface data. Temperatures have dropped a bit since the end of the latest El Niño event, as we saw in the surface record, but are still quite high by historic standards.

Recently one of the groups, RSS (which is the one used by David Rose in the Mail on Sunday article) identified an error in their record due to drifting measurement times that has been affecting observations in recent years. While they have updated some of their other records, their lower tropospheric record has yet to be corrected. They state on their website that the data suffers problems “with the adjustment for drifting measurement times” and “should be used with caution”.

Carl Mears, lead scientist at RSS, has also argued that the surface temperature records are more accurate than satellite-based data when trying to estimate surface temperatures.

Global tropospheric temperatures over land

Satellites have particular difficulty accurately measuring tropospheric temperatures over land. Over time, the satellites orbit drifts, resulting in measuring different locations at different times of day. Correcting for these differing observation times can be difficult, as different parts of the world will have very different day/night temperature cycles.

Over oceans these corrections are more straightforward, as the temperature difference is smaller and is fairly consistent. Over land, however, much larger uncertainties are present, and complex models are used to try and adjust for changing observation times. Which model is chosen and how it is applied can have a large impact on the resulting satellite tropospheric land temperature series.

This means that tropospheric land-only satellite-based records should be treated with caution. Unfortunately, it’s this data (which is rarely used in any scientific analysis) that David Rose chose to highlight in his article. Not global surface temperature data, which shows a normal modest decline from the recent El Niño, nor global lower tropospheric temperature data, which shows nothing out of the ordinary, nor land-only surface temperature data, which also shows no large drop. Rose chose land-only lower tropospheric temperatures to focus on.

Land only lower troposphere temperature

Land-only land-ocean lower tropospheric temperature anomalies by month, January 1979 through November 2016.

Land-only lower tropospheric temperature data from the three groups, UAH, RSS, NOAA STAR are shown in the figure above. Here there is less pronounced warming than in any of the other series we’ve examined, but there is also more disagreement between the records. In the last few months in particular, RSS shows a precipitous drop in temperatures not seen in the other two records. We can see this clearly if we zoom in on the past two years:

Land only lower troposphere temperature, 2015-2016

Land-only land-ocean lower tropospheric temperature anomalies by month, January 1979 through November 2016.

Here, the precipitous drop that David Rose makes such a big deal about only shows up in the RSS record. (And the very latest data, released after the Rose article was published, now shows an uptick in temperatures.) The other two records show no similar dramatic drop in recent years, just a normal decline after the El Nino event fades.

Was 2016 a record year without El Niño?

It should be clear by now that the modest decline in temperatures in the past few months is a normal reversion after the end of theEl Niño event. It has relatively little impact on the longer-term warming trend, which is the most scientifically relevant metric of climate change.

However, it is possible to disentangle the effects of El Niño (and La Niña) events from the underlying warming trend, and to predict how warm it would have been in the absence of El Niño.

Scientists use a number of approaches to determine the magnitude of El Niño’s impact and remove it. An example of what temperatures would look like with El Niño removed is shown in the figure from climate scientist Chris Colose below:

Land ocean temperature index (Gistemp)

Global average surface temperature from NASA with and without the effect of El Nino/La Nina removed. Figure courtesy of Chris Colose.

Multiple researchers have found that when the effect of El Niño is removed from the surface temperature record, 2015 and 2016 remain the second hottest and hottest years, respectively. We can also try a much simpler test that doesn’t involve any complex model or statistics. Let’s take the NASA record and simply remove the El Niño spike and see what happens:

NASA's surface temperature record, 1979-2016

Global average surface temperature from NASA and a variant with September 2015 to May 2016 held constant.

Here the blue line represents the actual NASA surface temperature record, while the red line is a variant, where the 2015/2016 El Nino is replaced by the temperature from the month prior to its onset.

If we remove the El Niño spike, 2016 (to date) is still the hottest year on record and 2015 is still the second hottest year on record. This simple test shows that the record warmth seen in surface records in recent is best thought of as a continuation of a long term warming trend amplified (but not caused by) a large El Niño event.

Sharelines from this story
  • Factcheck: Newspaper claim about global temperature is 'deeply misleading'
  • Kerry McNamara

    Thanks Dr. Zeke. Nice summary.

  • Rick20112

    Here is a climate scientist’s response to the article originally posted in the Daily Mail and summarized in Breitbart that caused so much consternation in the Washington Post and The Guardian. Here is a reasoned response, aided by several charts [at the article site] that describes Rose’s approach, the shortfalls that Hausfather perceives, what the shortfalls mean, and what the full set of data supports. A much more ‘scientific’ argument. And one that might still be interesting for and well understood by non-scientists.

    • Bongstar420

      Too bad no one admits to not understanding and what ever happens will be used to legitimize some unrelated political agenda.

      Confirmation of “global warming” apparently validates liberal agenda
      Confirmation against “global warming” apparently validates conservative agenda

      Never mind that both agenda’s are problematic and riddled with BS. I am a liberal if you are wondering.

      • john

        No you draw a long bow.
        The evidence points to changes and the only aspect of the outcomes points to humanity’s effect.
        This has zero to do with some political stance.
        Continued spread of FUD and total rubbish type articles are done to appeal to the ignorant and supports their preconceived idea of their self importance.

        It appalls me that we are honestly living in the age of total ignorance.

  • GothBoyUK

    Excellent article & a great rebuttal to the tripe posted by the Daily Mail.

  • Bongstar420

    Not the warmest year ever…the warmest year on record based on our records, yes.

    • SchroedingersDeplorableDog

      Yeah, it should not be written as ‘warmest year ever’. But,

      A. we don’t live in the world of 5 million years ago
      B. if you wanna do the science-is-baloney thingy, remember that carbon dating is ‘secular religion’ and the Earth in 6000 years old, and evolution is a fraud. So… there was no five million years ago, unless you believe all this science baloney.

      I’ll go with the science which provided the backbones supporting electric power, the phone system, computing and the internet. We’re using all of that at the moment, while the no-science crowd would have us stopped around the year 100 A.D. Ever see the amazing new tech built by the anti-AGW Discovery Institute? There is none: despite being around for several decades, the Discovery Institute has made zero scientific discoveries. None. Zip. Nada.

    • john

      Let us get this into perspective.
      Human habitation is the critical aspect.
      The title should not have to go into such an explanation or we well have every article going into every piece of evidence that has been built up for 100 years rather tiresome to do that.

  • Al Rodger

    It is quite remarkable how eager the deny-o-sphere has been to adopt this Dire David Rose Rubbish. We even get prize twit Delingpole being cited by a U.S. House of Representatives committee (or by its denialist chairman Lamar Smith on their behalf) by repeating Rose’s Rubbish.
    In the round, the thrust of this denial-speak is that the AGW-scam is over. Rose said ”Experts will be forced to eat their words.” Delingpole said “The last three years may eventually come to be seen as the final death rattle of the global warming scare.”
    It actually took Rose three weeks to spot his cherry-pick drop in temperature, its size possibly not entirely a product of the large size of the 2015/16 El Nino. But it wasn’t reflective of UAH TLT land temperatures and, as with similar drops in the past, and a week later with the November figures posted it has now entirely disappeared. RSS TLT(land) is still not returned to a toasty-hot level but the argument for picking TLT(land) was that this provides the signs of the future. RSS TLT(global) is the second warmest November on the RSS record while UAH TLT is the warmest.
    The low land temperatures of October were the flip-side of the unprecidented warm Arcitic this Autumn. Neither are good news for the future.

    My view on this deny-o-blather is that we should be sure to keep it for later. Which of Rose’s experts will be “forced to eat their words”? I will probably the denialists he cites, blog-mom Judy Curry and that esteemed climatologist and Gentleman Who Prefers Fantasy David Whitehouse. Let us make sure they never forget!!

  • rlhailssrpe

    This article, while interesting, is not a topic for a pro carbon or anti carbon publication such as Carbon Brief. It is a topic for a science conference, in which experts debate the data and analyses and in which the media and politicians are excluded. But it exhibits the characteristics of all data analyses, “If you torture the data long enough, it will confess.” – R. Coase in discussions on economics.

    It does NOT provide the certainty upon which to base a national energy policy. We face a hard fact of life: There is only one energy source, the exothermic reaction of carbon combustion, which can sustain human life by the billions. Nuclear fission can help, at the edges, for peaceful advanced nations. The problem is cost, which is not the purview of scientists. Cost and technology are the professional concentration of engineers. Total energy costs includes this subtotal plus taxes, laid on by politicians and profits, laid on by investors; I address neither but they are real.

    I, a retired engineered, worked on a score of nukes, two score carbon combustion power plants, and assessed advanced technologies, most in energy systems, for decades. I am not qualified to critique the article. However, IMHO, any major nation which cuts back on carbon combustion will destroy its economy. All of the green energies, for many reasons, cost too much for the masses to pay. The green energies will find niche markets but can not sustain the base load. Normal people can not afford it. America, if it tries, will lose the grid, the greatest democratic distribution of wealth in the history of man. In many nations, energy only exists in the guarded enclaves of the masters. The greatest break through, in energy, in recent times, is fracking which made vast hydrocarbon reserves accessible at low cost. It is realigning (hopefully peacefully) the wealth of nations. It is basically the wedding of robotics with miracles in material science. It is, in large measure, transferable to coal mining. Seven miles below the surface a robotic machine is better than John Henry. When, not if, this occurs America will have many centuries of the best recoverable coal reserves on earth. There are no realistic cost break-troughs in the green energies to be expected as far as experts can forecast.

    Thus we face an existential social problem. If burning carbon is a real and present threat to human existence then billions must be denied fire and they will die. This struggle has been normally resolved by war.

    I request that the petty ad hominid attacks on those of differing judgments cease. This conflict has the potential to destroy my nation, the United States of America.

  • Al Rodger

    Congratulations CarbonBrief. Your post here has been cited by the Gentlemen Who Prefer Fantasy for committing sins. As a result you have failed to spot‘ the return of the “hiatus” period’ which apparently will be approaching its third decade.
    This is of course a ding-dong that kicked off with Gentleman Whitehouse seeing the RSS & UAH TLT anomalies for land drop sharply in October after the recent El Nino. I see he was actually on to it prior to Rose. What Whitehouse has failed to spot yet (it took him 3 weeks to spot the October ones) is the rise of these anomalies for November with UAH TLT for land rising to become the warmest November on record.
    And do not worry. I’m sure the rise is entirely consistent with the rest of Whitehouse’s analysis in that it was a complete nonsense prior to the November data.

  • Pingback: Weather Channel Fires Back at Breitbart Over Climate Change Story – USSA News | The Tea Party's Front Page()

  • Gary Hall

    Satellite data is more accurate for measuring surface sea level than for the temperature of the lower


    Repeat that over and over again until you believe it. Oh right – they’ve already been doing that.

Related Articles


Expert analysis directly to your inbox.

Get a Daily or Weekly round-up of all the important articles and papers selected by Carbon Brief by email.


Expert analysis directly to your inbox.

Get a Daily or Weekly round-up of all the important articles and papers selected by Carbon Brief by email.