Decadal Trend in Temperature

We all know that many things affect global temperature. Some, like greenhouse gases, cause long-term changes and can create significant trends. Others fluctuate but don’t really go anywhere over the long term, so they cause temperature fluctuations but don’t create long-term trends.

At least two such fluctuating influences have been identified. One is solar variation. The energy output of the sun is variable, especially showing a cyclic variation with the roughly 11-year solar cycle. Another is the el Nino Southern Oscillation (ENSO), which when high warms the planet and when low cools it. It too fluctuates up and down, but hasn’t exhibited long-term trends which could account for the trend we observe in global temperature. You might wonder, what has been the short-term impact of these two influences recently? Let’s look at their temperature influence over a recent 10-year period, from January 2002 through December 2011.

Their temperature influence can be estimated by multiple regression of temperature as a function of long-term trend and exogenous factors. This enables us not only to correct the temperature data for such things as solar variation and ENSO, it also enables us to estimate their impact.

I recently re-computed the regression to update it through December 2011, using TSI (total solar irradiance) to represent solar activity and MEI (multivariate el Nino index) to represent ENSO. Here’s the influence of solar variation on global surface temperature (from NASA GISS) over the last 10 years:

Declining solar activity over this 10-year period has caused a drop in temperature, which certainly is not evidence that global warming has stopped or even slowed. But it has introduced a rather sizeable trend of -0.05 deg.C/decade over that time span — one which fake skeptics conveniently forget to mention when they discuss recent temperature trends. It’s one of the reasons that 10 years is far too little time to estimate long-term temperature trends reliably — a fact which fake skeptics also conveniently omit to discuss.

Despite the sizeable (but temporary!) trend caused by solar variation, then trend due to ENSO is even larger:

Note that the trend over these 10 years is dominated by starting with sustained el Nino conditions and ending with two major la Nina events of 2008 and 2011. Therefore the temporary influence of ENSO has caused a 10-year trend of -0.09 deg.C/decade.

If we include all the exogenous factors used here, then we must also add volcanic activity into the mix. It has actually caused a warming trend, but one so slight that it has only a tiny impact on the net temporary trend. Here’s the total effect:

Over the 10-year period from 2002 through the end of 2011, exogenous factors created a temporary trend of -0.13 deg.C/decade. That’s substantial — if not for the global warming due to greenhouse gases, we would have observed a sizeable global cooling trend.

Temperature in the troposphere (the lower layer of the atmosphere, where most of our weather takes place) is even more strongly influence by these factors than surface temperature. Here’s the estimated influence of solar variation on tropospheric temperature according to UAH:

Solar variation introduced a temporary trend of -0.06 deg.C/decade. The influence of ENSO is even stronger:

The temporary trend in UAH temperature due to ENSO is -0.16 deg.C/decade. If we include volcanic activity as well, we again find its influence over this 10-year period is tiny, and the total influence of all three factors is:

The temporary trend in UAH temperature due to these fluctuating factors is a whopping -0.21 deg.C/decade.

These known factors do indeed have a sizeable, but only temporary, influence on temperature. So sizeable, that we would indeed have witnessed substantial global cooling over these 10 years, except for the also-substantial global warming caused by … you guessed it … human activity.

23 responses to “Decadal Trend in Temperature

  1. Very interesting, thanks. Is it possible that the Earth is still gaining heat over this period despite the decline in surface and lower troposphere temperature? My understanding of La Niña conditions is that when surface temperature *falls* during a La Niña, that means the heat content of the oceans is *rising* (because the atmosphere is giving up its heat to the cooler water). NOAA’s 0 – 2000m Global Ocean Heat Content data shows continuing accumulation of heat over this period, not cooling. Also sea level has continued to rise over this period, and ‘Earth’s Energy Balance and Implications (Hansen et al 2011) says that “despite unusually low solar activity between 2005 and 2010, the planet continued to absorb more energy than it returned to space. “

    • W Scott Lincoln

      Of course the earth is still gaining heat energy. Measurements of surface temperatures and lower tropospheric temperatures are not showing the whole picture of where heat is stored in the climate system.

      • Measured over the whole decade, the Earth as a system gained more energy over the past 10 years than any decade in the past 40. ENSO and solar variations certainly provide the “noise” that might mask (or enhance) tropospheric temps over shorter periods, but the constant net energy added to the Earth as a system from net anthropgenic factors is about 0.6 wm2. The poor heat sink of the troposphere is not the place to look. When you look in the natural heat sink of the ocean, you find that just in the top 2000 meters of the ocean alone the Earth as an energy system has added some 10 x 10^22 joules of energy over the past 10 years. Those who say the Earth has cooled over the past decade certainly are speaking only of the fickle poor heat sink of the troposphere, and ignore the much larger and warming heat sink of the ocean.

  2. why only linear fits in your analysis? I’m not a chemist or physicist but I would expect the response to at least one forcing to be non linear.

    [Response: Linear fits to the estimated temperature influence tells us how much they have probably affected the linear trend in global temperature over the same time span, one of the most common summary statistics.]

  3. Rob Honeycutt

    Hang on to yer bloomers Betsy, it looks like the next decade may prove to be a real ride.

  4. Tamino,

    Have you seen this paper featured at Stoat’s place? I don’t have access to the full paper but one of the key points seems to be a significantly smaller influence of the solar cycle than found by yourself and Lean & Rind.

    • Gavin's Pussycat

      Hmm, does it matter? Tamino’s coefficients come from multiple regression in which both ENSO and volcanism are along. Their coefficients may both be wrong due to multicollinearity, but the wrongnesses will partly cancel out in the combined effect.

  5. First, a technical question.

    I’ve opened up the 6 image files above (GISS, UAH, GISS, …).

    In so doing, I see no difference in the 1st set of GISS/UAH TSI graphs (outside of the fact that the y-axes are different). Am I missing something rather obvious (to others) here? As the underlying GISS/UAH temperature time series are different.

    Let me hazard a guess. You’ve subtracted out the linear GISS/UAH trends from the TSI data (which themselves are scaled in some fashion from the external forcing (watts/m^2) to temperature (degrees C)). That way, the actual temperture time series would not come into play, per se?

    Similarly, for the 2nd set of GISS/UAH ENSO graphs, different y-axes, but one appears to be shifted (By one month?) relative to the other (UAH is to the right of GISS, one data point shift means one time series starts/stops slightly before the other one does).

    Anyways a bit confused at the moment on your presentation, probably need to go back and read your paper again (and/or previous posts on said paper).


    [Response: The estimated linear influence of solar (ENSO) is proportional to TSI (MEI). So the influence on different temperature series will differ only by a proportionality constant. The time offset between GISS and UAH for ENSO is due the fact that they give a different optimal time lag.]

  6. I don’t think you are handling the enso data correctly. Under your analysis, the temporary influence of ENSO has caused a 10-year trend of -0.09 deg.C/decade using your first GISS/enso figure. So to account for the ENSO effect one would have to add 0.09 deg C to the 2012 measured temperature. So assuming a flat zero anomoly over the 10 year period, the enso adjusted temperature in 2012 would be plus 0.09 deg C. But the overall enso data during the period averages out to be positive, and therefore the overall enso effect is warming even though the trend is downward, so the temperature in 2012 would actually have to be lowered to account for enso. The trend matters, but so do the actual values.

    [Response: A perfect demonstration of your willingness to contort reason in order to support your unwillingness to believe the truth.

    We leave it as an exercise to readers to demonstrate your fallacy.]

  7. Great post, Tamino! This should be mandatory reading at WUWT and other “skeptic” sites, but I doubt that many there will read this or especially take the time to understand what you have done.

  8. Ric Merritt

    The reasoning about solar variation seems obvious. While earthly conditions affect what happens to arriving solar energy, no one expects them to change what the sun sends our way.

    Similarly, volcano activity comes and goes on its own, becoming an independent input into conditions that affect the energy budget. Yeah, it’s conceivable that climate could somehow affect volcanism, but in the absence of evidence and, preferably, some good theory about mechanism, it’s farfetched.

    The ENSO bit seems a bit different though, not guaranteed exogenous, in that ENSO and global temperatures are part of the same system full of feedbacks. Isn’t there a concern that ENSO itself might be changing with the climate?

    I guess as long as a model of trend plus exogenous factors works well, that is evidence that the assumptions are working well, but that seems to me like a source of uncertainty.

    • David B. Benson

      The detailed timing of volcano eruptions is affected by mass distribution change as caused by ice sheet melting; NGISP contains good records of this. ENSO has long period variations: “Variability of El Niño/Southern Oscillation activity at millennial timescales during the Holocene epoch”

    • However farfetched, there is a long tradition of looking for linkages between geology and the weather, for example when a somewhat notable scientist spent a long paragraph on: “The connexion between earthquakes and the weather has been often disputed…such cases seem to bespeak some more intimate connexion between the atmospheric and subterranean regions.” — Charles Darwin, The Voyage of the Beagle, Chapter XVI. Darwin is perhaps commenting on Lyell (Principles of Geology, Volume 1, Chapter XXVI, p.533-4) who is in turn commenting on writings by Aristotle, Pliny, and Seneca.

  9. But, Roy Spencer has just shown that the US temperature rise is just caused by population growth, in
    “New U.S. Population-Adjusted Temperature Dataset (PDAT), 1973-2012”

    I’m not sure if there is such a thing as a crime against statistics, but if there is….

    • I think this qualifies as a war crime. He certainly tortured the data!

    • Lord, John, we know that CO2 emissions *don’t* grow with population growth, there can’t possibly be a correlation there that undermines Roy’s latest fantasy, right?

      (/snark, just in case)

  10. Given the level of commetary there, I’m having a hard time differentiating Poe from idiocy, i.e.:

    ” It would be quite a stunning revelation that all these indicators – including species migration and the shift of seasons and climate zones – which are well-explained by a system wide warming, instead are the result of local effects.”

  11. Great post Tamino, I have to say, lucid as always. It really show what’s coming down the pipe line regarding global warming.

  12. I think an enormous problem in public understanding would have been avoided if, back in the early 2000s, many more climate scientists had been more prepared to say (loudly) that we could expect periods of ‘cooling’ to disguise the underlying warming trend in the near future, but that we should not be distracted from the reality of the situation by such passing variation.

    Perhaps they did, but my recollection is of a stout denial of the possibility of ‘cooling’, allowing sceptics to allege that the theory of AGW posited continuous linear warming in line with rising CO2 – a condition which did not seem to have been met by the observations.

    I am scientifically ignorant to a shameful extent, but this piece says to me exactly what I assumed must have been happening some years ago, but was told (as I recall) was not happening at all by those most concerned to raise the alarm about anthropogenic greenhouse gas emissions.

    Perhaps I’ve misunderstood (and I won’t be in the slightest surprised if I have, and apologies if so), but I’m tempted to say ‘better late than never’.

    [Response: I haven’t seen any evidence of “stout denial of the possibility of ‘cooling'” by actual climate scientists. My impression is that instead, they have repeatedly and consistently stated that on short time scales noise can obscure the trend, but long enough time spans will consistently show an upward trend.

    And when they do, their words are likely to be misrepresented. One such case: on Sept. 1 2009, at a world climate conference in Geneva, Mojib Latif in a scientific presentation explicitly stated that due to short-term noise, “it may well happen that you enter a decade or maybe even two when the temperature cools relative to the present level.” He was speaking hypothetically in order to emphasize that natural variation could mask the underlying trend. In less than a week, Marc Morano reported it this way at his “Climate Depot” website: “Sept. 5: UN Fears (More) Global Cooling Commeth! IPCC Scientist Warns UN: We are about to enter one or even 2 decades during which temps cool.” On Sept. 19 Lorne Gunter said in the Calgary Herald: “Latif conceded … ‘that we are likely entering one or even two decades during which temperatures cool.'” Even though Morano changed his misquote from “are about to” to “may be about to,” he seems not to have let it sink in because on Sept. 29 he wrote “Mojib Latif of Kiel University in Germany told a UN conference earlier this month that he is now predicting global cooling for several decades.”

    It seems to me that the problem isn’t a failure of climate scientists to represent the situation accurately. It’s deliberate misrepresentation by those who wish to deny the reality, human cause, and/or danger of global warming.]

  13. John Mashey

    Charlie: without having to know a lot of the science, the book advertised here “Noise” is really a nice short introduction to these issues, especially noise in time series, which many people need to understand.

  14. Thank you both (Tamino presumably and John Mashey) for your responses. Much appreciated.

    I’ll have a look at the book ‘Noise’ – many thanks for the recommendation.

    I do see that climate scientists have been clear about the difference between noise and trend. I wonder if perhaps there’s been a problem with the journalistic representation of the issue.

    Thus, the sceptical argument has been that we have seen ‘cooling’ (or at least a lack of warming) since 1997/8, demonstrating the lack of connection between atmospheric CO2 and average global temperature.

    The mainstream journalistic response, presumably informed by properly qualified scientists, seems to have been to argue the toss about whether there is ‘cooling’ or not. I appreciate I’ve provided no reference for this – it seems to me a relatively uncontentious claim.

    I much prefer the approach taken by Tamino in this piece, which if I have understood, differentiates between a ‘cooling’ phase inherent in the natural variability of the planet and the longer term anthropogenic trend.

    I wonder if it’s now too late to persuade journalists to approach the issue in this way. If people have expected linear warming and can be convinced it’s not happened, saying ‘that’s not what we meant, or even said’ might simply not work. I think that’s all I was really trying to say!

    Thanks for your time though!