Arguments
Global warming...still happening
Posted on 30 August 2013 by Ari Jokimäki
The apparent lack of warming in Earth's surface temperature measurements since 1998 is not yet significant from climatic perspective. Surface temperature also seems to be changing according to IPCC projections. Climate model simulations show similar warming breaks, and have done so even before current break started, even if they include the effect of carbon dioxide. Models also can re-create the current break and the cause for the break seems to be known: warming has gone to the oceans instead of warming the surface. The ocean warming has been observed. Also the continuing warming effect of greenhouse gases has been observed. Global warming as a whole seems to continue despite the apparent break in surface measurements.
Several decades of long-term warming is evident in recent surface temperature measurements. However, since 1998 surface temperature records don't show clear warming. This is not very clear because the time period is not very long and the possible trends might not be statistically significant. Skeptical Science trend calculator shows warming trends since 1998 but they are not statistically significant. Santer and others (2011) estimated that it takes 17 years of satellite measurements, before effect of mankind to lower atmosphere can be detected. In some cases, 15 years has been mentioned for the limit of statistical significance, so the situation seems to be quite borderline. For example, from 1983 and 1998, which was time of rapid warming, SkS trend calculator still shows a trend that is not statistically significant. Trend is significant from 1982 to 1998. (Even if statistically enough time would pass without warming, it still wouldn't mean that increases in greenhouse gases wouldn't have a warming effect. This we will see below in more detail.)
Climate is usually considered as average weather over longer period of time. Standard length for the climatic time period is 30 years. Let us see what this means for the surface temperature. Following figure shows Earth's surface temperature as a running 30 year mean (this means that the value of each point in the graph is the average of the surrounding 30 years of temperature values, for example, the running 30 year mean value for year 1990 is the average of temperatures of 30 years between 1975 and 2004):
As we can see from the graph, there are no signs of global warming stopping or even slowing down in this kind of inspection. From climatic perspective global warming still continues. Those with sharp eyes notice that the time shown in X axis stops in the graph before 2000. This is because in 30 year running mean the year 1998 is currently last one that can be shown. However, the temperature evolution after 1998 is included and is affecting the graph starting from 1983. The fact that graph stops at 1998 means that we have to wait 15 years before we know how climate has evolved since 1998.
Although global surface temperature evolution since 1998 is not yet climatically important, there has been quite a lot of research on the issue. One question appearing in public has been that does the temperature follow earlier projections. Rahmstorf and others (2012) have analyzed how IPCC projections match the surface temperature measurements. Here is a graph of their results:
The evolution of Earth's surface temperature and IPCC projections. Surface temperature without corrections is shown in pink (one year running mean averaged from all surface temperature analyses) and corrected surface temperature is shown in red (corrections are explained in text). Blue area and blue lines are IPCC third assessment report projections. Green area and green lines are IPCC fourth assessment report (AR4) projections.
As we can see from the graph, surface temperature changes shown in pink sometimes go outside the projections. This is because some factors are not included in IPCC projections. Such factors are solar activity changes and eruptions of volcanoes. Additionally, the variation of El Niño/La Niña is random and therefore it doesn't change in simulations at the same time it does in real life. IPCC projections are combined results from many simulations, so the El Niño/La Niña variations of different simulations tend to cancel out when simulation results are combined. This means that the projections don't actually include El Niño/La Niña variation either. It should be noted that even if surface temperature shown in pink doesn't stay within limits of projections, it does stay within limits of all individual simulations (not shown in the Figure above).
The effects of the Sun, volcanoes, and El Niño/La Niña variation have been corrected for in the temperature evolution shown in red. This graph stays quite well within limits of projections especially during last few years. It sometimes goes outside the projections in the beginning of the time period. There still might be some factors which would need to be corrected for. On the average it does seem to follow the projections even in the beginning of the time period (at least it doesn't deviate permanently to one direction).
Already happened temperature evolution can be recreated with models also so that internal variability of climate system and changes in solar activity and in volcanoes are included (afterwards we have knowledge for example when a volcanic eruption has happened). Lean & Rind (2009) have done this and the following Figure shows their results:
a) The observed surface temperature (black) and simulation of surface temperature from a simple model (orange). b) The factors affecting surface temperature. Graphs are from Lean & Rind (2009).
The result of the simple model shown in the graph is so close to observed surface temperature evolution that we have good reason to suspect this study might be able to answer us why surface temperature has not apparently increased since 1998. Lower part of the Figure shows the factors affecting surface temperature, and from there we can see that ENSO (that is, the variation of El Niño and La Niña) seems to have varied in quite a similar manner than surface temperature in the period in question. Also the cause of the longer-term temperature rise seems to be clear: the effect of mankind is the only one of the factors, which shows long-term increase.
Also the future projections of climate models show periods where surface temperature doesn't increase, even if models have the effect of greenhouse gases included. Here are some examples of such simulations:
Projections of climate models show periods of slowed/paused warming similar to that in the observations since 1998. On the left: simulations with three different emission scenarios from IPCC AR4. Upper right: simulation example from Easterling & Wehner (2009). Lower right: simulation example from Meehl and others (2011).
All shown simulation examples show periods where long-term warming trend is paused even for decades, and warming continues after that. The Figure also shows simulation examples from model runs for IPCC AR4 projections. These show similar pauses. AR4 discusses the expected temperature evolution rather carelessly. From the texts of the report one might get an impression that surface temperature should rise certain amount in each decade. They of course mean that on average certain rise in temperature is expected per decade, even if it doesn't occur during each decade. Some people have used the carelessly worded texts in IPCC AR4 to distort the issue, even when the same report shows the simulations presented above, where the truth in the matter can be seen.
Simulation examples shown above are all quite recent. Model simulations have shown similar features also earlier. Here we see an example from IPCC second assessment report (SAR), which was published in 1995, before current warming pause apparently started:
Model simulations from IPCC second assessment report. Year 0 means year 1990.
One interesting detail in the SAR model simulations is that one of them shows strong spike around 1995. This is comparable to the 1998 peak in observed temperatures. Even if the post-1995 evolution is difficult to see in the graph, we can be sure that the simulation in question shows quite long pause in the warming after 1995. In principle, we could say that the simulation in question predicted the warming pause, albeit being off by few years. It's not genuine prediction of course, but just a coincidence. Nevertheless, it's a quite curious detail.
So, the simulations of climate models show clearly that while the increase of greenhouse gases in the atmosphere increases Earth's surface temperature in the long run, other factors cause pauses to the warming every now and then. Similarly, those other factors also speed up the warming in other times. This has been explicitly stated in Easterling & Wehner (2009): "We show that the climate over the 21st century can and likely will produce periods of a decade or two where the globally averaged surface air temperature shows no trend or even slight cooling in the presence of longer-term warming".
At the moment it seems that the factor causing the pause has been ENSO, which in practice means that the warming effect of greenhouse gases has gone deeper to the oceans instead of warming the surface. This has been the subject of some recent studies.
Already shown above were the results of Lean & Rind (2009) and Meehl and others (2011). Similar results have also been reported by Kaufmann and others (2011), Hunt (2011), Guemas and others (2013), and Watanabe and others (2013). According to all these studies, the primary cause for the apparent pause in the surface warming is that the warming has gone to the oceans. Solar activity has also been said to have played some role on the issue. Also, Solomon and others (2010) have suggested that changes in water vapor content in the atmosphere might have speeded up the warming during the 1990's and slowed the warming during 2000's.
The warming going into oceans has also been observed. Following Figure shows the ocean heat content in the top layer (0-700 m) of the oceans (from Lyman and others, 2010):
The graph shows that after 1998 the heat content in the oceans has increased substantially.
But when can we expect surface warming to continue? Surface warming continues when the sum of all factors affecting Earth's surface temperature has a warming effect. It can take decades and decades, as long as there are other factors that have large enough cooling effect to mask the warming effect from greenhouse gases. However, our current knowledge suggests that there are no such factors that could have large enough cooling effect in order to make this pause much longer.
There are reasons to think that warming might continue soon. Earth's surface temperature has been very high recently, close to record temperatures, while solar activity has been very low and La Niña has been the prevailing state of ENSO. Without the effect of greenhouse gases these factors would have cooled Earth's surface substantially. We haven't seen such cooling. When La Niña changes to El Niño, it is expected that warming will continue.
So it seems that the warming effect of greenhouse gases seems to be still there. Fortunately, we don't need to guess this, as we also have observations of the effect of greenhouse gases, as we see next.
A group of researchers have studied spectral measurements of outgoing long-wave radiation taken from satellites (Chapman and others, 2013). They found out that the warming effect of carbon dioxide has continued to increase during the 2000's. They calculated from the spectral measurements that between 2002 and 2012 the amount of outgoing long-wave radiation decreased in the characteristic absorption frequencies of greenhouse gases. This was the case at least for carbon dioxide, ozone, and methane. Largest warming effect was from carbon dioxide. Observed decreases in the outgoing long-wave radiation matched the expectation from the increased greenhouse gas concentrations during the study period. Here are their results in a graph:
It should be noted that the study of Chapman and others was presented in a conference in April 2013, and apparently official research paper has not been published yet. Information presented here is from the conference paper.
References:
IPCC AR4 simulations: Figure 10.5 with caption.
D. Chapman, P. Nguyen, M. Halem, A decade of measured greenhouse forcings from AIRS, Proc. SPIE 8743, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIX, 874313 (May 18, 2013); doi:10.1117/12.2017019. [abstract]
John A. Church, Neil J. White, Leonard F. Konikow, Catia M. Domingues, J. Graham Cogley, Eric Rignot, Jonathan M. Gregory, Michiel R. van den Broeke, Andrew J. Monaghan, Isabella Velicogna, Revisiting the Earth's sea-level and energy budgets from 1961 to 2008, Geophysical Research Letters, Volume 38, Issue 18, 28 September 2011, DOI: 10.1029/2011GL048794. [abstract, full text]
David R. Easterling, Michael F. Wehner, 2009, Is the climate warming or cooling? Geophysical Research Letters, Volume 36, Issue 8, April 2009, DOI: 10.1029/2009GL037810. [abstract, full text]
Virginie Guemas, Francisco J. Doblas-Reyes, Isabel Andreu-Burillo & Muhammad Asif, Retrospective prediction of the global warming slowdown in the past decade, Nature Climate Change, 3, 649–653 (2013) doi:10.1038/nclimate1863. [abstract]
B. G. Hunt, The role of natural climatic variation in perturbing the observed global mean temperature trend, Climate Dynamics, February 2011, Volume 36, Issue 3-4, pp 509-521, DOI: 10.1007/s00382-010-0799-x. [abstract]
Robert K. Kaufmann, Heikki Kauppi, Michael L. Mann, and James H. Stock, Reconciling anthropogenic climate change with observed temperature 1998–2008, PNAS July 19, 2011 vol. 108 no. 29 11790-11793, doi: 10.1073/pnas.1102467108. [abstract, full text]
John M. Lyman, Simon A. Good, Viktor V. Gouretski, Masayoshi Ishii, Gregory C. Johnson, Matthew D. Palmer, Doug M. Smith, & Josh K. Willis, Robust warming of the global upper ocean, Nature 465, 334–337 (20 May 2010) doi:10.1038/nature09043. [abstract, full text]
Gerald A. Meehl, Julie M. Arblaster, John T. Fasullo, Aixue Hu & Kevin E. Trenberth, 2011, Model-based evidence of deep-ocean heat uptake during surface-temperature hiatus periods, Nature Climate Change, 1, 360–364 (2011) doi:10.1038/nclimate1229. [abstract, full text]
Stefan Rahmstorf et al 2012, Comparing climate projections to observations up to 2011, Environ. Res. Lett. 7 044035 doi:10.1088/1748-9326/7/4/044035. [abstract, full text]
B. D. Santer, C. Mears, C. Doutriaux, P. Caldwell, P. J. Gleckler, T. M. L. Wigley, S. Solomon, N. P. Gillett, D. Ivanova, T. R. Karl, J. R. Lanzante, G. A. Meehl, P. A. Stott, K. E. Taylor, P. W. Thorne, M. F. Wehner, F. J. Wentz, 2011, Separating signal and noise in atmospheric temperature changes: The importance of timescale, Journal of Geophysical Research: Atmospheres (1984–2012), Volume 116, Issue D22, November 2011, DOI: 10.1029/2011JD016263. [abstract, full text]
Susan Solomon, Karen H. Rosenlof, Robert W. Portmann, John S. Daniel, Sean M. Davis, Todd J. Sanford, Gian-Kasper Plattner, Contributions of Stratospheric Water Vapor to Decadal Changes in the Rate of Global Warming, Science 5 March 2010: Vol. 327 no. 5970 pp. 1219-1223, DOI: 10.1126/science.1182488. [abstract, full text]
Masahiro Watanabe, Youichi Kamae, Masakazu Yoshimori, Akira Oka, Makiko Sato, Masayoshi Ishii, Takashi Mochizuki, Masahide Kimoto, Strengthening of ocean heat uptake efficiency associated with the recent climate hiatus, Geophysical Research Letters, Volume 40, Issue 12, pages 3175–3179, 28 June 2013, DOI: 10.1002/grl.50541. [abstract]
Ari,
This paper is causing some current stir ... it may be one you would like to comment on in the light of this excellent review of papers & perhaps add to your list.
Recent global-warming hiatus tied to equatorial Pacific surface cooling
Thanks.
I think the meaning of statistical significance has been misapplied to the current hiatus/pause.
As I understand Santer (please correct me if I'm wrong) defines statistical significance as a 95% probability of a positive trend over 17 years. As applied to a "pause" in surface temperatures during a period of rising Co2, let's define that as
"During a period when CO2 rises at 1.5ppm per year, we expect, for any given year, a 5% probability that the trend across the preceeding 17 years will display a zero or negative value"
Let's define modern warming as being over the past 50 years (during which time Co2 has risen at ca 1.5ppm/year)
We then *expect* one in twenty of the past 33 years (those where we can measure the previous 17) to show a negative or zero trend for the preceeding 17.
Our prediction is then that roughly one or two years of the past 33 will have a negative trend.
Feedback on if this is a correct interpretation of the science would be much appreciated.
Thanks, Shoyemore. I think that there might be a separate article on that paper in SkS in near future.
Let's then look at the trends recently.
I used GISS, and by my reckoning, there are *no* years in the past 33 where the trend is negative.
This applies whether the trend is defined as end – to – end, or least squares best fit.
Going for the least squares best fit approach, which seems more reliable, the last time there was a 17 year negative trend was 1976, and there are only 3 in the past 50 years (’67, ’68, ’76)
On that basis, there is nothing unusual whatever about the current “pause”, indeed it is expected and could even be argued as overdue.
Ari,
Last figure is unreadable, especially the axis legends.
At least I can imagine on horizontal axis the values of spectral numbers where known CO2 and H2O absorption bands are. But I have no clue what is on vertical axis.
And how do you read from said figure that total outgoing IR in the spectrum shown has decreased over last decade as implied by the caption?
The indepth presentation is interesting and infromative. However, I prefer a simpler statement.
The warming since 1998 can best be evaluated after the next signifcant El Nino of the magnitude that occurred in 1998 occurs. There are other variable factors influencing the surface temperature, but the global average surface temperature history shows very strong "bumps" in the global averages when significant El Ninos are acting.
So the implication that a specific minimum number of years of data is required to establish statistical significance is not correct. The major variable influencing factors need to recur in the data being evaluated, whatever time that takes.
VeryTallGuy #2: I think your interpretation of Santer et al. makes sense, but I'm not a statistics expert.
Chriskoz #5: Sorry, the vertical axis is indeed insufficiently labeled (but it's like that already in the Chapman et al. paper). The vertical axis shows Brightness Temperature.
I don't think I said anything about total outgoing IR, but only the "characteristic absorption frequencies of greenhouse gases". The study of Chapman et al. shows that the IR absorption of said greenhouse gases has increased over the ten year study period.
Regarding Santer et al and time required for trend significance, Tamino posted something very interesting a few years ago:
Cherry Snow - It turns out that if you cherry-pick an extrema starting point for your trend identification (as opposed to the more frequent values near the trend), the t-value needed to establish trend significance almost doubles.
In other words, if you select a trend point such as the 1998 El Nino as your start, as done over and over by 'skeptics', it distorts the statistics such that the trend over that period is less significant, so that what would be reasonable (say, 17 years) is insufficient due to that cherry-pick.
Cherry-picking extrema end points weakens trend significance, making the oft-repeated "16 year" claims even less meaningful - they have the least trend significance of any possible data selection of that length.
Tamino gave an even better summary later in the thread:
For me, it is pretty obvious that the graph of Chapman et al (2013) will become a classic to illustrate the greenhouse gaz effect.
Solomon isn't seeing any recent increase in stratospheric water vapor, and Chapman's contribution spectrogram seems to confirm this. Over a longer period then we need to look to Dessler (2010) for long term water vapor feedbacks. You can just bet that some of the denierrati will jump on this to argue that there are no positive feedbacks, that we're limited to forcing from increased CO2 only, ECS is about 1C and thus all is right with the world.
Very nice article, but there's one thing I don't understand. The pink line in the second figure suggests that the uncorrected global temperature in 1998 was higher than for any succeeding year. In contrast, other global temperature measurements, such as the four measurements graphed on the "Berkeley Earth" summary page suggest temperatures higher than 1998 in several of the succeeding years. My guess is that a different global temperature data set was used to generate the pink line, different from the data sets used to create the lines in the Berkeley Earth summary graph (NASA GISS, NOAA/NCDC, Hadley/CRU, and Berkeley Earth). Am I right? In any case, it would always be a good idea to provide the source of any data set used to make graphs.
[DB] The paragraph immediately following the legend for that graphic fully explains the corrections done:
Once these exogenous and transitory factors are removed, the underlying trend emerges, shown in red.
The BEST version, using a similar accounting:
Joel, the caption of the figure in question mentions the data-source for the pink line: "one year running mean averaged from all surface temperature analyses". Further details are available in Rahmstorf et al. paper to which there's a link in the references section.
Thanks to both Ari and the moderator (DB) for the explanation. I hadn't realized that the data shown in the Berkeley Earth summary graph had been corrected in the same way that the red line in Ari's graph had been corrected. The explanatory legend at the Berkeley Earth site didn't make that clear. But it makes sense, because the red line in Ari's graph resembles all four lines in the Berkeley Earth summary graph. I conclude that the uncorrected global temperature in 1998 has not yet been exceeded by any subsequent uncorrected annual average global temperature.
@VeryTallGuy for #2, @Ari Jokimäki for #7,
I am not familiar in depth with the Santer et al (2011) calculations, but a quick look at their Appendix and their Equation (A1) and runup to that suggests these a pretty ad hoc. They themselves say, in their paragraph numbered "68":
I can't comment on what they are trying to do here, but it's fair to say this is no conventional kind of statistical significance being calculated. Worse, statistical "significance" cannot be properly used in the way VeryTallGuy seeks to do so, and as many do. "Significance" has a specific technical meaning, and I would strongly recommend reviewing that, e.g., at http://en.wikipedia.org/wiki/Statistical_hypothesis_testing#Interpretation See also http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1756718/
The problems classical significance testing or hypothesis testing has in complicated situations is one of many reasons why modern statistical inference is couched in the Bayesian view. See the text, J. Kruschke, Doing Bayesian Data Analysis for a great introduction. Alternatively, check out http://hypergeometric.wordpress.com/2013/08/22/soon-you-will-be-one-of-us/
Joel, depending on the datasets used, 2005 and/or 2010 each equalled 1998 as the warmest year for the surface temperature record. However, when viewed as decadal averages, the most recent decade has been the warmest, by far:
Source
Thanks, Daniel (#16), for the very nice graph from Climate Central. One from NASA that conveys a similar lesson can be found here. In the NASA graph, small differences can be seen among the four datasets/interpretations. The dataset/interpretation from the Japanese Meteorological Society shows 1998 as the warmest year on record, but the datasets/interpretations from the other three sources show 2005 and 2010 as being slightly warmer. Consistent with the graph from Climate Central, the graph from NASA clearly shows (by simple eyeballing) that, regardless of dataset/interpretation, the decade 2001-2011 was warmer than any preceding decade.
I still don't understand why this argument hasn't already been put to bed. I've seen several news articles and papers in recent months talking about the pause in warming being linked ot the enso cycle, as if this was news? Surely this was patently apparent ever since this chart was first circulated:
http://skepticalscience.com/graphics.php?g=67
All trends continue upward, it's just we're still in the la nina and neutral territory of the spectrum.
It seems obvious to me that the next El Nino is gonna prove decisively what this chart already shows.
I'm imaging the climastrology reponse to Dan Bailey's repost of the excellent Climate Central graphic-
begin simulation exercise
'Ya see...we had 4 steps up and a plateau from 1900-1940, so we've had 4 steps up now, so we're due for another plateau!!! If only climate scientist could count with the fingers on one hand'
(end simulation exercise).
Global warming is produced by an energy imbalance at TOA, an imbalance which is evidenced most clearly in accumulating ocean heat content. I would suggest that you use the latest NOAA values for OHC (http://www.nodc.noaa.gov/OC5/3M_HEAT_CONTENT/basin_data.html) which show, through the first quarter of 2013, an unabated rise in heat accumulation on earth (as oceans absorb ~93% of that imbalance). The atmospheric temperatures, due to the far, far lower heat capacity of the atmosphere relative to the oceans, and driven by the current set of ocean-atmosphere heat transfer mechanisms (ENSO, PDO, AMO, thermohaline circulation, etc), provide the least reliable assessment of an energy imbalance at TOA is derived from atmospheric temperatures. A ton of statistics on a poor indicator is not particularly helpful.
Daniel @ 16.
I would caution against the presentation of decadal averages without also showing the trend line of rolling averages of a decade of data. A new decadal average can be claculated for each new month of data. However, some may claim the need to "wait for the next decade of data" before having to admit anything. Also, a claim could be made that the decadal rolling average has stopped climbing, because it has. However, averages of larger sets of data that would include the recurrance of signifcant variable factors such as El Nino continue to climb.
The numbers and statistics game can be played in many ways. Even the 30 year average shows fluctuations resulting from the many variable and random factors that affect Global Average Surface Temperature.
What is important, and challenging, is the presentation of information in a way that will make it harder for people to "believe what they prefer to believe".
Some still claim the Earth to be flat. Empirical, observational evidence and physics demonstrate otherwise.
Circular reasoning bordering on sloganeering. The ongoing energy imbalance at the Top Of our Atmosphere (TOA) amply demonstrates the utter, physics-less falsehood of this statement. A claim could be made that sheep's bladders could be employed to prevent earthquakes. So what?
Daniel,
There are many people who are looking for any possible excuse to believe that activity they benefit from is not a problem. There is a very sucessful deceptive marketing effort that preys on that desire. It uses any means to delay global action that would actually significantly reduce the burning fossil fuels.
So my comment remains that using a comment like "the past decade was the warmest" will be successfully abused by the decievers and delayers saying "we need to wait for the next full decade of data before we do anything". If you are not careful about what you try to argue with you can make it easier for someone to "beleive what they prefer to believe", even if you know they have misinterpretted the point you were tryng to make.
The warming of the earth due to human activity is indeed beyond scientific dispute. It has been for more than a decade. This issue is now political. How something is said matters more than the actual facts in politics. The best way to deal the decievers is to present infromation that is very difficult to "take advantage of". You have to remember that much of the population has "a vested interest in not properly understanding this issue". Much of the population cares more about getting more benefit for themselves than they do about future consequences others will face.