RealClimate logo

The most popular deceptive climate graph

The “World Climate Widget” from Tony Watts’ blog is probably the most popular deceptive image among climate “skeptics”.  We’ll take it under the microscope and show what it would look like when done properly.

So called “climate skeptics” deploy an arsenal of misleading graphics, with which the human influence on the climate can be down played (here are two other  examples deconstructed at Realclimate).  The image below is especially widespread.  It is displayed on many “climate skeptic” websites and is regularly updated.


The “World Climate Widget” of US “climate skeptic” Anthony Watts with our explanations added.  The original can be found on Watts’ blog

What would a more honest display of temperature, CO2 and sunspots look like?

  1.  It is better to plot the surface air temperature.  That is what is relevant for us humans: we do not live up in the troposphere, nor do natural ecosystems, nor do we grow our food up there. By the way, the satellite-based tropospheric temperatures shown by Watts show almost the same climatic warming trend as those measured by weather stations near ground level (in both cases 0.16 C per decade over the last 30 years).  However, variability in the tropospheric data is considerably larger, especially because of higher sensitivity to El Niño (as happened in 1998) and the solar cycle (we showed that in Foster and Rahmstorf ERL 2011 – when corrected for those factors the surface and troposphere data agree closely).  Because of increased noise, the trend is less obvious to the eye, especially if one shows monthly values which adds yet more noise.  Let us thus use the GISTEMP global annual temperature record from NASA’s Goddard Institute for Space Science (all surface data sets agree to better than 0.1 °C, see comparison graph).
  1.  One needs to scale the CO2 data correctly for an honest comparison with temperature, so that it can actually be used to evaluate climate scientists’ predictions of the CO2 effect.  You can calculate this with a complicated climate model, but one can also use a back-of-envelope estimate.  A CO2 increase from 280 to 400 ppm (equivalent to 2 Watts/meter2 radiative forcing) produces about 1 °C of global warming (at the time when 400 ppm is reached – some further warming will follow with delay). Thus, an increase of 100 ppm CO2 on the right hand side of the graph corresponds to a temperature increase of 0.8°C on the left hand side. That matches the IPCC’s estimate of the “transient climate response (TCR)” of ~2°C at the time of CO2 doubling (see Technical Summary of the IPCC WG1 report, p. 84). The TCR is smaller than the equilibrium climate sensitivity (about 3°C for doubled CO2) because it takes time to warm the oceans. The full equilibrium warming is thus only reached after a time delay. We are going to use the annual values from the famous CO2 measurements which began in 1958 on Mauna Loa in Hawaii.
  1.  And last but not least one should show honest sunspot data (annual time series), not just a snapshot of the number of spots on the sun today (which is completely uninformative for climate purposes – it’s apparently been added to the widget simply to insinuate an important role of the sun). Here also there is a question of the proper scaling (which is actually not that important because solar activity is cyclical and shows no significant trend over the period of the graph).  We will chose the scaling from the correlation analysis of Lean and Rind (2008) from which one can find a measurable effect on global temperature with an amplitude of 0.05°C.

When done this way the graph looks like this:


One of the readers of our German sister blog KlimaLounge, Bernd Herd, has programmed a widget for this graph so it can be added to any website at a size you like, automatically updated annually.

The trends in the CO2 and temperature anomaly curves agree very well with each other.  This is surprising at first because CO2 is of course not the only factor that influences global temperature. There are two reasons for this agreement:

(1)  Of the other anthropogenic factors, some have a warming effect (other greenhouse gases such as methane) while others have a cooling effect (air pollution). These roughly balance in global average. The IPCC AR4 report found a radiative forcing of 1.7 W/m2 from the CO2 increase alone, while the total from all anthropogenic factors amounted to 1.6 W/m2.

(2)  Natural factors (volcanoes, solar cycle) influencing the trend are very small in comparison to anthropogenic CO2 (as e.g. standard correlation analyses show, see for example Lean and Rind 2008Foster and Rahmstorf 2011). The IPCC AR5 found their contribution to global temperature change since 1951 to be in the range of −0.1°C to 0.1°C.

It requires quite some skill to produce a misleading graph like Watts’ global climate widget, which hides the actual connections between global temperature, CO2 and the sunspot cycle. Watts’ widget is quite a useful indicator though: whenever you see it on a website, you know they are trying to fool rather than inform you there.


A quick ‘n dirty guide to falsifying AGW

Dot Earth: Warming Trend and Variations on a Greenhouse-Heated Planet

217 Responses to “The most popular deceptive climate graph”

  1. 201
    Phil Scadden says:

    So bobbyv, what’s your theory for the increasing Antarctica seaice since you dont like the science version and how well does it stack up against data? (eg temperature trends in place new sea ice is forming). What do you think the climatic significance of the increase is?

  2. 202
    Douglas McClean says:

    Dave Walker at 174:
    “Given that we seem to have focused matters around the ice caps, lets stay with it. I think, that the general public thinks. that both the ice caps are melting – when they are not. One is, one isn’t (source: the link referenced in #170 above). – See more at:

    This is spectacularly wrong. The “ice caps” are both melting. For some reason you are restricting your attention in the southern “ice cap” to sea ice only, ignoring the much larger and more significant enormous continent-wide glacier, which is melting. If the public “thinks that both the ice caps are melting”, they are correct. If they were to think that “sea ice near both poles is melting” they would be wrong. Not the same thing by a long shot.

    [Response: Part of the problem here is the term “ice cap” which is pretty much wrong for Earth. Scientists talk about the “ice caps” of mars, but an “ice cap” on earth means a small ice mass not confined to a valley, but on land. An accurate statement would be that 1) most ice caps are melting, two of the three continental-scale ice sheets (Greenland and West Antarctica, but not really East Antarctica) are melting, virtually all glaciers are melting; and 2) the Arctic sea ice is melting, but the Antarctic ice sea ice, on average, increasing (though with large areas declining). –eric]

  3. 203
    Dave Walker says:

    To Ray @ 195 and Hank at 196

    … I have been advised to move on from this subject by the moderators so I am not going to respond other than to say Ray (re your para 3) – the graph does show Antarctic sea ice growing – and Hank, your para 3 is irrelevant to the graph – its the 25 year trend not the melt/freeze within a year.

    Based upon this thread it would appear that the graph at:

    … is the most deceptive graph of the year!!!

    [Response: I would agree it is deceptive, because it implies that averaging sea ice extent globally is meaningful. It isn’t. It’s like averaging Denmark and Iraq, and using that to claim the world is at peace!–eric]

  4. 204

    #203–Appreciate the ‘moving on, ‘ Dave–I think we’ve covered this ground (or ice) pretty thoroughly for now. But I’d like to clarify one last thing about what Hank said in his #200. When his source talks about ‘winter trend’ that *is* a long-term trend, not the ‘melt/freeze within a year’ that you thought he intended.

    It’s common to break down trends seasonally, as well as to consider them on an annualized basis. So, for the case of Arctic sea ice as an instance, the winter ice trends show a slower decrease than do autumn trends. (The largest decrease is for the annual minimum, which generally occurs around the middle of September.) For that matter, when NSIDC reports the monthly sea ice data, they always give the trend in terms of that particular month. So, for instance, if you read the linked page (link below), you’ll see that the trend for November is -4.7%/decade.

    I’m also thinking–and I may just be wrong about this–that the ‘winter trends’ idea is local to each hemisphere, so that the ~ -30,000km2 for the NH might be December-February, while the smaller SH trend is probably June-August.

  5. 205
    Dave Walker says:

    Ref: all the comments regarding the graph showing Global, Antarctic and Arctic Sea Ice levels over the last 25 years.

    Final comments from me on this subject.

    Eric The Moderator (is he one of the Marvel Comic Super Heroes?!?!?) summed up nicely at #203 above. The referenced graph shows receding Arctic sea ice area levels and increasing Antarctic sea ice area levels – but adding them up and jumping to a conclusion that therefore all is well- is totally wrong.

    The discussion that took place to get to that conclusion is proof of the point that I raised initially.

    All Hail Eric The Moderator!

    [Response: I would tell you *which* Marvel Comics Superhero I am, but I can’t as that would reveal his true identity! –eric]

  6. 206
    Jim Eager says:

    The graph itself is not deceptive, Dave, the ice extent is what it is. What is deceptive is that some people are using the graph without context to assert that sea ice extent has returned to 1979 levels, that Antarctic sea ice growth offsets Arctic sea ice decline, and that the polar “ice caps” are not melting. All three of those assertions have been shown to be demonstrably misleading or flat out false.

  7. 207
    Hank Roberts says:

    For Dave Walker:
    Have you noticed that you’re never linking to the original chart?
    There’s a reason to do that. You know where to find it, right?
    Using all the available information is honest, but inconvenient to those who copy images onto their own pages, out of context to spin them.

    Find the source, Dave.

  8. 208
    Jim Eager says:

    At the risk of carrying the subject on further, there is a factor that none of us has mentioned, or if they have I missed it. Extent and area are only a two dimensional measure of sea ice when in reality sea ice has a third dimension, depth, which multiplied by the other two yields volume, and from it ice mass can be estimated.

    The cited graph, whether the original or the copy, do not show ice volume, and therefore do not show how much sea ice is actually in either polar region. We know that large areas of Arctic sea ice are multi-year ice, and thus thicker than annual ice, for example. But we also know that ice volume in the Arctic has declined even more precipitously than ice area or extent. Most Antarctic sea ice is annual, on the other hand, meaning most of it melts back to the edge of the shore or fixed ice shelves each summer. Do we know if and how Antarctic sea ice volume is changing verses in the Arctic?

    [Response: We know very little about Antarctic sea ice volume. I would bet that it is thinning (which goes along with expansion) but there are too little data to demonstrate this. –eric]

  9. 209
    Hank Roberts says:

    For Kevin (and Dave) — to tie up the loose end:
    Kevin pointed out that Dave misunderstood my post at 200 above.

    Dave, look at the link and actually read what it says at Fact Checker: Does record Antarctic sea ice refute global warming?

    The caption identifies the data points — there’s one dot for each year, for each pole. That’s the longterm trend. Could that be clearer?

    The caption there says explicitly: it charts

    “Sea Ice at Minimum Months”

    the minumum at each pole for that pole’s season, as Kevin surmised.

    That’s information they quote as coming from from

    Ted Scambos …. lead scientist at the National Snow and Ice Data Center at the University of Colorado, where he works on Antarctic ice dynamics and polar climate change effects.

    Dave’s experience makes clear: it’s easy to fool people with copies.
    Dave, I recommend More for the annals of climate misinformation — rather dated and old but cautionary on this.

    Don’t be fooled again.
    The lesson is always — you may trust, but you must verify.
    Check the facts. Read closely. Find the source.

  10. 210
    MARodger says:

    Jim Eager @206.
    You are correct that they do tend to “assert that sea ice extent has returned to 1979 levels” although our recent visitor usually added the word “roughly” which is usually required because these folk are such poor with the analytical skills that they rely on the poor old eye-ball to provide them with evidence. I would suggest that even “roughly” it still has yet to return to “1979 levels.”

  11. 211
    MARodger says:

    Assuming nothing really dramiatic occurs in the next couple of weeks, Annual Average Global Sea Ice Areas 1979-2014 (Usually 2 clicks to “download your attachment,”)

  12. 212
    Jim Eager says:

    Thanks Eric (in line at 208), I thought that might be the case as I could not find anything solid on Antarctic sea ice thickness or volume. If, as you suspect, Antarctic ice thickness is declining as it spreads out then the increase in ice mass would not be as large as the observed increase in extent would suggest.

  13. 213

    #208–Volume, yes! I wrote about it here:

    Most of the regulars will probably have seen this, but others may want to give it a go.

    And, for anyone interested in sea ice matters, there is a great place to feed what can become an addiction:

    And finally, here’s the most recent update of the PIOMAS volume update:

    You can clearly see that the volume has been rebounding from the shocking low of the 2012 minimum. It’s good news, but it’s not nearly as significant as some of the usual suspects would like to believe; we had a couple of consecutive Arctic summers that were relatively cool, so there’s not much mystery about why we’re seeing what we are.

  14. 214
    Dave Walker says:

    To the moderators

    recently, whenever I land on RealClimate page, I get music of various sorts. Yesterday it was some sort of hi tech sounding dance track with the lyric “give them what they want”, today it is a bouncy big band sound from the 60s era. I dont get it when I land on any other site.

    Any ideas what is going on?

  15. 215

    #214–Yes, I’ve had that experience several times, too. Seems pretty random.

  16. 216
    brandon sheffield says:

    How does getting an observed average of temperatures aka a single number for the globe for each month create noise? My understanding is very little on this. In my opinion it should be easy to plot a single Temp on a monthly graph Excel style. If “noise” is the squiggly lines due to the ups and downs between the points indicating the temperature, please explain why that is noise. If it is something different please explain. Thank you

  17. 217
    Hank Roberts says:

    for Brandon Sheffield, LMGTFY:

    Measure anything around you every day.

    You won’t get exactly, precisely, the same number each time.
    But, over a month or two, you’ll find out if the variation you see is noise, or if there’s something that looks like it might be a trend.