Temperature Data Update

Since last year was the hottest on record, many people have trained a keen eye on this year’s temperature data so see how it will compare with last year’s record-breaker.

So far, it’s a hot one indeed. How hot? NASA has just released the global temperature data for April, and although we only have four months of data for the year so far, some of the folks at home are wondering how this-year-so-far compares to previous years’ temperatures. Here you are:


56 responses to “Temperature Data Update

  1. Reblogged this on Gra Machree.

  2. Richard Mallett

    Is that January to April compared to some standard January to April, or May to April compared to some standard May to April ?

    [Response: It’s Jan-Apr compared to previous years (i.e., Jan-Dec)]

    • Please see my comment below. It seems to me that, due to changes in the seasonal cycle that have occurred, there are problems with comparing anomalies from different sets of months. But your expertise in this is far beyond mine (and, indeed, I think you’ve blogged about changes in the seasonal cycle previously).

      • Richard Mallett

        Yes, I entirely agree. There are many reasons why we should wait until the complete figures for 2015 are in before we compare it with 2014. We don’t have to write newspaper headlines, so we can afford to be patient :-)

  3. How does this compare if you only take the 4 first month of the years other than 2015? Does it change anything at all?

    • hotwhopper.com has produced a graphic which is progressive for the last 11 years. What you see is about what you’d expect given random natural variation: Regression is certainly seen given extreme values (hi or lo) early in the year.

      How much regression is a different issue related to autocorrelation.

    • Sou has a more complete plot here of the running YTD averages.

    • Greg Laden has plotted that. 2015 still top but 2010 close.

  4. Will this come down to a fight between the anticipated increasing El Nino and the Chilean volcano?

    • Calbuco released about half a million tonnes of SO2; Pinatubo released about 20 million tonnes. Calbuco’s impact on the climate is likely to be negligible, and dwarfed by any El Niño event.

    • Carbon Brief interviewed some climatologists and while the altitude reached by the eruption is nearly to the stratosphere (or higher) where sulfur dioxide would tend to remain in the atmosphere for a few years rather than a couple of weeks, we don’t expect much sulfur dioxide from this volcano and so far the data is bearing this out. Furthermore, this volcano is to far south to have much of an effect. If it were closer to the equator reflective aerosols would have a chance to enter both hemispheres of the stratosphere. Additionally, I suspect that as there is more ocean in the southern hemisphere and as ocean has greater thermal inertia, the reduced sunlight at the surface should have less of an effect on global temperature. Finally, they mention that volcanos can have the effect of strengthening an El Niño.

      Please see :


      • it’s not generally true that volcanoes need to be tropical to have a big impact, it’s just a different set of impacts. The spatial structure matters, and I’m spending my whole phd thesis looking at that. But, for Calbuco, the SO2 emissions were indeed small.

  5. I’m not sure if this has any significance or not (probably not), but for what it’s worth:

    If you compare Jan-Apr 2015 with historic Jan-Apr rather than historic full years, 2014 is still the warmest–but just barely. The second warmest is 2010 rather than 2014, and 2015 is only 0.01C warmer than 2010. (To be sure, I’m hoping that my quickie calculations are right!)

    I mention this only because I have no idea whether there is any sort of seasonal consideration that might be involved in comparing partial years to full years.

    • Windchasers


      Remember that El Nino events, when they happen, commonly start in about July to August of one year, and taper off in May to July of the following year. Most of the previous “warmest years ever” were those ‘following’ years, years on the back half of an El Nino. E.g., 1998, 2010, 2005, 2007, etc.

      2014 was unusual for a warmest year in that it didn’t start with an El Nino, and heck, didn’t even really end with one. In 2015, we’ve got an El Nino kicking up early, which might be its own rather unusual thing, or might be related to 2014’s half-aborted El Nino attempt.

      In any case, given that El Ninos typically make the Earth’s surface warmer than usual, this explains why the previous warmest-ever years started extra warm and then cooled off some as the El Nino subsided. But in 2015, it looks like the El Nino is just getting started, and could stay all the way through the summer and into the fall.

      So: if this rather unusual El Nino sticks around, then I expect the warm temperatures to stay through the year, too.

  6. Sorry, that should read “2015 is still the warmest”–not 2014.

  7. Is there a tendency for the first few months to be relatively less strong in the expression of anomalous readings considering it’s the NH’s cold season?

  8. Jeffrey Davis

    That looks to be almost 1C of warming since 1960.

    • It’s about 0.8C (55 years, OLS trend = ~0.015C/year).

    • For your consideration; the 1 degree C above baseline has already – and in some areas far exceedingly- been a reality, for quite some time now, for most of the populated world i.e. the Northern Hemisphere where the land to sea ratio is higher than it is for the southern one. Another reason why the cogitated 2 degree threshold to the end times of civilization as we know it today, is so bewildering.

      • michael sweet

        That is the point. We have seen less than half the projected warming for 2C and we already have major issues. For example: the drought in Syria and California. The unstoppable degradation of Antarctic ice. Sea level rise is estimated at 10+ meters per degree C increase. Once we hit the 2C average increase globally it will be worse. Your suggestion that since it is currently 2C warmer some places and the world has not ended is non sequitur. The globe has not even reached equilibrium with current temperature.

      • namnack,

        “Another reason why the cogitated 2 degree threshold to the end times of civilization as we know it today, is so bewildering.”

        And who is allegedly saying that this was the threshold for the “end times of civilization”, according to you? This isn’t what you find, e.g., in the IPCC Report or other UN documents. The phrase isn’t “end times of civilization” that supposedly would come over us, if the threshold was exceeded. Instead the statement is that this was the threshold of “dangerous anthropogenic interference with the climate system”. The statement refers to a risk, which is a combination of the probability of occurrence of an event and severity of impact of the event, if the event occurs. The risk that climate change will have severe, undesirable effects on ecosystems, society etc is considered to be unacceptably high above a globally averaged temperature increase of 2 deg. C relative to the surface temperature of preindustrial times (i.e, something like the average temperatures of the 19th century).

        Also, the 2 deg. threshold refers to the globally averaged surface temperature increase. If you have such an increase, then the increase, e.g., in continental regions of the Northern Hemisphere won’t be only 2 deg. anymore. Instead, it will reach 3 or 4 deg for the average temperature. (The Arctic has reached this already in parts today. There, the increase will be 6 to 8 deg. C, or even higher.) And then you have the impact of these changes. Here, one question is, will those impacts be linear, or will they be nonlinear? Including the impact of extreme events. It’s not just the average increase that matters, it also matters what the increase in extreme events will be and the impact of those. Thus, your bewilderment about the 2 deg. threshold, because you don’t see the impact in some continental regions, where it already has become 2 deg. warmer, is based on logically flawed thinking on your side.

        Consider this. The CO2 levels of 400 ppm in the atmosphere today is at a level that Earth has seen in the Pliocene about 3 million years ago the last time. Back then, surface temperature was 2-3 deg. higher than today, and the sea level was about 25 meters higher. Or, take the Eemian (the interglacial about 120000 years ago). We are just reaching the temperatures of the Eemian, or we will be there in only a few decades. Sea level during the Eemian was at least 4 meters higher than today. Now tell me, that this kind of change wouldn’t have any severe impact on human society, which should be avoided, if still achievable.

      • Richard Mallett

        We have only been taking global temperatures since 1850 (HadCRUT4) when the anomaly was -0.376 C. In 2014 it was +0.563 C so a gain of +0.939 C. That would give us 285 years from 1850 = AD 2135 to reach a warming of 2.0 C.

        If we want to use linear trends, since 1850 the linear trend has been +0.49 C per century, so 2.0 C warming would take 408 years from 1850 = AD 2258.

        Regarding sea level, http://sealevel.colorado.edu/ gives the rate as 3.3 +/- 0.4 mm. per year since 1992, so a 4 metre rise would take 4000 / 3.3 = 1212 years.

        [Response: Your propensity for taking a long-term linear trend of something that we can prove has accelerated over time, then extrapolating it to give the impression that the impacts are a long way off, is truly foolish. You might ponder the reason for your need to minimize future changes in such an obviously silly way.]

    • Richard Mallett

      NASA GISS 1960 anomaly -0.04 C.
      NASA GISS 2014 anomaly +0.68 C.
      Difference +0.72 C.
      Of course, 1960 was soon after the Earth started its 20 year pause after the 1944-1956 cooling – remember the warnings around that time of the next little ice age ?

      • Richard, that’s not a valid way to do it. It’s depends too much on how warm or cold the start year happened to have been, which is weather.

      • Richard Mallett

        I know, but the comment to which I was replying spoke about 1960, which (as I pointed out) was particularly cool.

      • 1960 may have been cool, but you don’t talk about the rise/fall in a linear statistical trend by addressing the start point datum and the end point datum. You talk about the rise/fall by noting the predicted start point value and the predicted end point value.

        This is Stats 101 or maybe high school.

      • Richard Mallett

        What do you mean by ‘the predicted start point value and the predicted end point value’ when we are talking about measured temperatures in the past ? The NASA GISS trend from 1960 to 2014 was 1.45 degrees per century if you are talking about the linear statistical trend.

      • “What I mean” is totally basic. WRT describing the magnitude of the warming between 1960 and 2014, chrisd3 is correct and you are wrong. If you do not understand why, buy Tamino’s book from the right hand frame.

      • Richard,

        They are talking about the way a time trend is determined in statistics–you figure a “linear regression” of your variable against elapsed time. That way you use all the points, not just the end points, and avoid having your trend distorted by outliers.

      • Richard Mallett

        That is precisely what I did when I said that the trend in global temperatures from 1960 to 2014 according to NASA GISS is 1.45 C per century. NCDC and HadCRUT4 trend from 1960 is 1.35 C per century. (I repeat that 1960 is not my choice of start point, and not one that I would use)

      • @Richard: That may be “what you said,” but it is certainly not how you calculated the 1960-2014 temp change. You just subtracted the 1960 temp from the 2014 temp. Your calculation did not involve the 1960-2014 trend at all.

        I should’ve known from your reference to “the warnings around that time of the next little ice age” that trying to help you on this point was not going to be easy.

      • Richard Mallett

        You must have missed the follow up where I said that the trend since 1960 was +1.45 (NASA GISS) or +1.35 (HadCRUT4 and NCDC) degrees per century.

        The original statement from Jeffrey Davis was that there had been about one degree of warming since 1960. I initially said in my reply to him that the difference in temperature was +0.72 C, which is way less than one degree. If you are trying to explain temperature differences between two dates to somebody, you are not immediately going to start talking about linear trends.

        I can quite understand how you could have missed my follow up, when there are so many conversations flying about.

      • I couldn’t imagine it would be this difficult.

        Yes, you mentioned the trends. But you didn’t use them in your calculation of the 1960-2014 temp change–a question to which you provided an incorrect answer, as has been explained several times.

        And, yes, if you are trying to explain the temp difference between two years and you’re talking about climate rather than “Wow, it’s 10 degrees warmer than it was yesterday,” then yes, you absolutely “start talking about linear trends,” because that’s the right way to do it. You never, ever just subtract the start temp from the end temp. That is completely useless information. It’s the sort of crap that people like David Rose use to claim that there’s been no warming at all.

      • Richard Mallett

        Well, now it turns out the person who types in italics within square brackets doesn’t like linear trends either; so maybe he will tell me what he does like to use to prove his case that global warming has accelerated.

        [Response: You’ve gone beyond the limits of actual skepticism, to genuine trolldom. Goodbye.]

  9. If this year turns out to be a really hot one, I can predict that the hottest denier argument for the next 20 years will be “global warming stopped in 2015”.

    • I doubt that. It will remain to be the ‘invisible hand’ argument; a negative that can be argued ad nauseam and will take more than a smile to debunk. Let ‘m deny it. Who cares, really. It’s not like public opinion makes a difference concerning this or any other large scale policy. We wouldn’t be in this mess if it ever did.

  10. Well, I think we can safely say that if that magnitude of anomaly remains until the end of the year, there will be no argument about what % chance 2015 had of being the hottest year versus other years – it will be 100%.

    • The deniers will claim that it’s an artefact of the satellites, most probably. They’re nothing but practiced liars.

    • Doubtful it will be 100%. Consider the difference between 2014 and all other values was about 37% figure or even worse with the 40+% calculated by NASA/NOAA. We would need differences of a much greater magnitude to get the probabilities down below 5% let alone down below 0.5%

      • We would need differences of a much greater magnitude to get the probabilities down below 5% let alone down below 0.5%

        And right now, the difference IS of a much greater magnitude. 2014 exceeded the previous record by about 0.01C. If the rest of 2015 matches its year-to-date average, it will exceed 2014’s record by about 0.1C. In other words, the difference is about an order of magnitude larger.

      • Richard Mallett

        I would suggest that we wait until we compare a complete year (when all the results have come in, which takes a couple of months) with a previous complete year before we jump to conclusions.

      • The relevant figure isn’t the temp difference but rather the standard error of each annual point. I don’t know that figure–possibly someone does–but to get to an effective 100% (i.e., 99.7%) you need to get to three standard deviations about that figure. We could easily see a year not outside that probability area.

      • jgnfld writes: “We could easily see a year not outside that probability area.”

        No. You are missing the point. Cgs’s statement above was:

        IF the rest of 2015 maintains the same anomaly as the first four months, THEN the uncertainty about whether it is a new record will be (negligibly small).

        According to Hansen’s 2010 paper, the 2-sigma uncertainty around the annual data points is +- 0.05 C for recent years. Thus, the 2014 record was less than 1-sigma above the previous record. By contrast, the 2015 year-to-date average — which cgs was hypothetically extrapolating to the entire year, as a thought experiment — is almost 0.1 above 2014.

        So if, as cgs hypothesized, the annual average for 2015 is similar to the current year-to-date value, then it would be about 4 standard deviations above the previous record. That’s close enough to a probability of 100% for most purposes.

      • Could happen. But looking at the record since 1900 there have been a total of 18 new all time highs set. In that set, precisely 2 were new records .08 or more degrees C higher than all previous highs: 1937 (.10) and 1998 (.15). Three more were .07 C higher: 1926, 1980, and 1988 but .7 would not be 3 full SDs out.

        The distribution of new highs is highly skewed therefore likely not a product of pure random variation which of course changes the probabilities. The median new high that is set is .04C higher than all previous highs.

        Could happen, then. But on purely historical, atheoretical grounds, it is an improbable occurrence. That said, the 1998 high was related to a set of physical factors that may be rehappening and so history may not be the appropriate guide. 1937 was ENSO neutral, BTW.

  11. Harold Brooks

    GiSS ties a record for most months in a single calendar year with an anomaly of at least 0.75 with 4 (set in 2010, 2014). 18 total on record.

  12. Nitpick: 2015 should not have a dot, because the dots mark the turn of the year — the line isn’t pinned down to a dot yet.

    In other news, the deadly polynomial trendline has emerged again from under the Arctic ice, accelerating all the way; http://3.bp.blogspot.com/-Wwjm5gFrTIY/U2ZAL1cWTBI/AAAAAAAANMI/aUOlYGCJpg8/s1600/methane-2.png

    And — congratulations, the donations link for “Peaseblossom’s Closet” for “Mistletoe” is live and working, first time in years it’s been working (for me)

  13. I urge all to take a look chez le Rabett and read about Bjorn Lomborg’s “scholarship”.

  14. We’ve had alternative versions of graphs in the thread above (year-on-year plots from hot whopper & Jan-April comparisons from Greg Laden) but no sign of my favourite – rolling 12-month averages. I’m presently unable to upload my version of this (which is using NCDC rather than NASA GISS) but from it, I come to the following conclusions.
    I’ve always argued that this ‘pause’ began in 2007 because prior to that data, a linear regression 1980-to date results in a larger upward trend in global temperatures with time, an increasing slope (as plotted in the red trace here- usually 2 clicks to ‘download your attachment). Such a finding is incompatable with a global ‘pause’ starting before 2007.
    An updated version of this graphic using NCDC data shows temperatatures since December have risen to the point of again accelerating the rate of temperature increase (that is it is strongly above above the black trace) for the first time since 2007. On the strenght of that finding, I think we can declare the ‘pause’ paused.
    If the full annual 2015 NCDC anomaly were to average above +0.787ºC, the rate of temperature increase 1980-to date would have exceeded the maximum reached back in 2007 which would suggest this particular ‘pause’ is actually ended.

    • Richard Mallett

      I think talk of a pause has to be taken in the context of the overall trend. We have had (roughly) :-

      cooling from 1880-1911
      warming from 1911-1944
      cooling from 1944-1956
      pause (slight warming) from 1956-1976
      warming from 1976-1998
      pause (slight warming) from 1998-2014

      – but the overall trend in degrees per century since 1880 is :-

      0.65 C (NOAA NCDC)
      0.66 C (NASA GISS)
      0.64 C (HadCRUT4)

      It seems to me that taking the slope of a graph that starts in 1980 is pretty meaningless in terms of climate.

      [Response: It’s hardly meaningless, when one computes the uncertainty it’s useful information.

      The claim “pause (slight warming) from 1998-2014” has no statistical validity. I’ve demonstrated this repeatedly. “It sure looks like” doesn’t cut it, even “It looks so strong it’s obvious” is just a recipe to fool yourself. If you have genuine statistical evidence that the trend from 1998 to 2014 is any different than that from 1970 to 1998, do tell.]

  15. It seems a little problematic to me to compare the first four months of 2015 to the annual average from other years, because my impression is that the seasonal cycle has changed over time, so GISTEMP’s “anomaly” calculation is not functioning optimally.

    I did a quick back-of-the-envelope attempt to predict the annual value based on the first four months’ average, using the past few decades’ data for a model, and IIRC the expected value for the annual average would be somewhat lower than the four-month average (but with a pretty wide range of uncertainty).

    Of course, if ENSO keeps building, that might counteract the default pattern of the annual mean being a bit lower than the JFMA mean.

    But still, on general principle, I think one ought to take into account changes in the seasonal cycle before comparing anomaly data from different sets of months.

  16. Richard Mallett

    I see that somebody has been posting replies in italics and within square brackets that do not appear in my email client.

    One claims that we can prove that warming has accelerated over time. Since we have had warming and cooling periods as shown below, I would be interested to see how one can prove from such varying data that the warming has accelerated over time.

    http://www.realclimate.org/index.php/archives/2014/12/recent-global-warming-trends-significant-or-paused-or-what/ identifies the following warming and cooling periods in GISS since 1880 :-
    1880-1912 cooling 32 years
    1912-1940 warming 28 years
    1940-1970 cooling 30 years
    1970-2014 warming 34 years

    The second asks me for statistical evidence that the trend from 1998 to 2014 is different from the trend from 1970 to 1998.

    So, since these two comments are related, let’s examine the linear trend from 1970 to 1998 and from 1998 to 2014 on the three main data sets :-

    NASA GISS 1970-1998 +1.67 C 1998-2014 +0.83 C per century.
    HadCRUT4 1970-1998 +1.73 C 1998-2014 +0.62 C per century.
    NOAA NCDC 1970-1998 +1.65 C 1998-2014 +0.56 C per century.

  17. I calculated the trends with 2 sigma error for the annual Berkeley Land + Ocean anomaly using air temperature above sea ice.
    Because the year to year chance are correlated with the ENSO, I did also an ENSO correction to reduce the error using the MEI and extended MEI data sets.

    Here are the results for the periods mentioned above.

    period trend +-2-sigma-error [95% CL]
    BEST Land+Ocean, without ENSO (MEI) correction:
    1880-1912 -0.24 +- 0.40 [-0.64,+0.16] C per century.
    1912-1940 +1.37 +- 0.42 [+0.95,+1.79] C per century.
    1940-1970 -0.22 +- 0.36 [-0.58,+0.14] C per century.
    1970-2014 +1.68 +- 0.21 [+1.47,+1.89] C per century.
    1970-1998 +1.65 +- 0.46 [+1.19,+2.11] C per century.
    1998-2014 +0.99 +- 0.79 [+0.20,+1.78] C per century.
    1880-2014 +0.72 +- 0.06 [+0.66,+0.78] C per century.

    BEST Land+Ocean, with ENSO (MEI) correction:
    1880-1912 -0.25 +- 0.34 [-0.59,+0.09] C per century.
    1912-1940 +1.33 +- 0.33 [+1.00,+1.66] C per century.
    1940-1970 -0.14 +- 0.30 [-0.44,+0.16] C per century.
    1970-2014 +1.67 +- 0.21 [+1.46,+1.88] C per century.
    1970-1998 +1.17 +- 0.41 [+0.76,+1.58] C per century.
    1998-2014 +1.25 +- 0.46 [+0.79,+1.71] C per century.
    1880-2014 +0.69 +- 0.06 [+0.63,+0.75] C per century.