Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.


Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Bluesky Facebook LinkedIn Mastodon MeWe

Twitter YouTube RSS Posts RSS Comments Email Subscribe

Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...

New? Register here
Forgot your password?

Latest Posts


Working out climate sensitivity

Posted on 23 October 2009 by John Cook

In the last post on past climate change, we looked at just two determinations of climate sensitivity. I started with Hansen 1993 as it used a relatively simple methodology in using temperature change and radiative forcing to calculate climate sensitivity. Hegerl 2006 was chosen because it covered the most famous of past climate change periods - the Medieval Warming Period and Little Ice Age. But to be fair (and consistent with my 'whole picture, not a single piece of the puzzle' philosophy), you get a clearer idea of our understanding of climate sensitivity by perusing the whole range of peer reviewed scientific literature on the subject.

The equilibrium sensitivity of the Earth’s temperature to radiation changes (Knutti & Hegerl 2008) presents a comprehensive, concise overview of our scientific understanding of climate sensitivity (H/T to Things Break for the heads up). In particular, it presents a visual summary that very neatly encapsulates how various methods examining different time periods have yielded consistent results (and I'm not just saying that because it uses lots of bright, shiny colours).

Figure 1: Distributions and ranges for climate sensitivity from different lines of evidence. The circle indicates the most likely value. The thin coloured bars indicate very likely value (more than 90% probability). The thicker coloured bars indicate likely values (more than 66% probability). Dashed lines indicate no robust constraint on an upper bound. The IPCC likely range (2 to 4.5°C) and most likely value (3°C) are indicated by the vertical grey bar and black line, respectively.

There have been many estimates of climate sensitivity based on the instrumental record (eg - the past 150 years). Several studies used the observed surface and ocean warming over the twentieth century and an estimate of the radiative forcing. A variety of methods have been employed - simple or intermediate-complexity models, statistical model or energy balance calculations. Satellite data for the radiation budget has also been analysed to infer climate sensitivity.

Some recent analyses have used the well-observed forcing and response to major volcanic eruptions during the twentieth century. A few studies examined palaeoclimate reconstructions from the past millennium to gain insight into climate sensitivity. Direct estimates of the equilibrium sensitivity from forcing between the Maunder Minimum period of low solar forcing and the present are also broadly consistent with other estimates.

What can we conclude from this? We have a number of independent studies covering a range of periods, studying different aspects of climate and employing various methods of analysis. They all yield a broadly consistent range of climate sensitivity with a most likely value of 3°C for a doubling of CO2. When I say doubled CO2, what is meant specifically is a radiative forcing of 3.7 Wm-2. A variety of forcings will have actually influenced climate over the periods examined, not just CO2.

So the combined evidence indicates that the net feedback to radiative forcing is significantly positive. Knutti concludes there is no credible line of evidence that yields very high or very low climate sensitivity as a best estimate. However, there is still some uncertainty about the exact value of climate sensitivity with a likely range (more than 66% probability) of 2 to 4.5°C.

There is more certainty regarding the lower bound of climate sensitivity. The probability of lower climate sensitivity drops off sharply while it's harder to rule out larger climate sensitivities. This is the inevitable consequence of a climate system with positive feedback. Constraints from observed recent climate change indicate that climate sensitivity is very likely (more than 90% probability) to be larger than 1.5 °C.

Where the rubber hits the road in all this analysis is how future temperature rise will affect humanity. Figure 2 shows temperature rise for a given CO2 level. The dark grey area indicates the climate sensitivity likely range of 2 to 4.5°C.

Figure 2: Relation between atmospheric CO2 concentration and key impacts associated with equilibrium global temperature increase. The most likely warming is indicated for climate sensitivity 3°C (black solid). The likely range (dark grey) is for the climate sensitivity range 2 to 4.5°C. Selected key impacts (some delayed) for several sectors and different temperatures are indicated in the top part of the figure.

If we manage to stabilise CO2 levels at 450ppm (current value 384ppm), we have a probability of less than 50% of meeting the 2°C target. The key impacts associated with 2°C warming can be seen at the top of Figure 2. The tight constraint on the lower limit of climate sensitivity indicates we're looking down the barrel of significant warming in future decades.

0 0

Printable Version  |  Link to this page


Comments 1 to 23:

  1. The caption for Fig 1 is confusing: how could the thicker bars represent higher integrated probability than the thin bars (which cover a broader range)? Surely it's the narrower range that is "likely" and the broader range that is "very likely". I also find interesting the circles for the most likely value. Instrumental record, last millenium proxy, and combined lines of evidence are less than 3 Celsius; only general circulation models indicate more. So it's hard to see why the IPCC point estimate is 3 C rather than, say, 2.8 C (like for "combined lines of evidence"). Finally, it's interesting to see that "expert elicitation" gives exactly the same likely range as IPCC (I don't really see how they would differ) while surely with several estimates producing similar ranges the confidence in that range should increase (and thus the narrow bar for "combined lines"). And finally, er, really finally, I think we must be over 384 ppm in CO2 equivalents now. Just CO2 was 385 ppm so I imagine adding anthropogenic methane and such would push us to ... maybe 395?
    0 0
    Response: The caption for Fig 1 was confusing because it was wrong - I've updated the text, swapping thicker and thinner. Thanks for pointing that out.
  2. What about the latest study by Lindzen/Choi (see that pegs sensitivity at .5 degrees C? Is there any validity to the theory that increased ocean temperatures increases the amount of radiation that is lost to space?
    0 0
    Response: Increased ocean temperatures do increase the amount of radiation lost to space - if the Earth is in positive energy imbalance, the planet will accumulate heat, oceans will warm, the earth will radiate more energy to space until it approaches radiative equilibrium again. This is discussed in more detail in the Climate Time Lag post.
  3. djb95054, Increased ocean temperature absolutely increases the amount of radiation emitted by the oceans. According to the Stephan-Boltzmann Law the energy emitted across all wavelengths by a black body increases as the 4th power of the temperature. A portion of this is lost directly to space out what is called the infrared window where greenhouse gases do not intercept the radiation, while a much larger portion is absorbed by the atmosphere. Thus, warmer ocean's maintain a warmer lower atmosphere which expands due to the added warmth. The layer of emissivity, where Earth's effective temperature of 255K is met, reaches to greater height where the radiation balance between incoming and outgoing radiation is attained, no extra energy is lost from there. So, most of the additional energy emitted by a warmer surface (oceans) goes to warming the atmosphere to higher temperature, while an increased but much smaller amount is lost out the atmospheric infrared window.
    0 0
  4. Some issues. Hansen 1993 - This is a National Geographic publication? Not peer-reviewed? I can't access it as a journal thru my university. I've nothing against non-peer review but it is often a stick used to beat 'deniers' quote from this article "you get a clearer idea of our understanding of climate sensitivity by perusing the whole range of peer reviewed scientific literature on the subject." hegerl 2006 in their abstract "A number of observational studies3–10, however, find a substantial probability of significantly higher sensitivities, yielding upper limits on climate sensitivity of 7.7K to above 9 K (refs 3–8)." Looking at some of these references (often written in less politically ardent times) you get the following quotes more quotes from hegerl "The dominant uncertainty in the calculation of climate sensitivity is clearly that pertaining to the estimates of radiative forcing" radiative forcing is the mechanism most strongly accosiated with climate change from CO2? "Improved understanding of physical processes of climate change and refinement of climate models is essential to reducing uncertainty in climate prediction." Speaks for itself. and from some of hegerls references "Recently Tett et al. [1999] found the increase in global mean near-surface temperature during the first half of the twentieth century may be due to variations in the sun's irradiance. This supports the earlier findings of Kelly and Wigley [1992] and Schlesinger and Ramankutty [1992]; further support is provided by Marcus et al. [1999], Drijfhout et al. [1999] and Beer et al. [2000]." I wonder if you include those 6 sun irradiance papers in your 'whole range of peer reviewed scientific literature' "Radiative forcing is the greatest source of uncertainty in the calculation; the result also depends somewhat on the rate of ocean heat uptake in the late nineteenth century, for which an assumption is needed as there is no observational estimate." "climate estimated to lie between 1.5 and 4.5K (Cubasch et al. 2001), largely on the basis of experiments with general circulation models (GCMs)." This wikipedia entry discusses some "potentially serious inconsistencies" with these models many of the papers are also full of words like assumption, maybe and uncertainty. Yet you seem to go to definite statements based on these papers. Remember all these quotes come from the climate change science papers not deniers.
    0 0
    Response: I gleaned two "take-homes" from Hegerl's paper:
    1. Despite the many determinations of climate sensitivity, there hasn't been much progress in constraining the likely values - it's gone from initial estimates of 2 to 4°C to current estimates of 2 to 4.5°C. You call that progress?! A large part of this is uncertainties associated with radiative forcings, as you point out.
    2. At least there has been progress in constraining the lower bound being very likely greater than 1.5°C. Anything over 1.2°C. indicates the climate has net positive feedback.
    It's interesting you should mention the uncertainty with radiative forcing in relation to CO2 forcing - I will be posting on that very subject in the next post.

    Re the sun irradiance papers, I'm fairly confident they would be included in the 'Instrumental Period' results but feel free to check this out yourself (and report back please). We examine elsewhere the role of the warming sun in the early 20th century.
  5. A few general comments on Lindzen and Choi paper. The paper, in the first part, focus on the comparison between ERBE data and model predictions. The simulations used were taken from the AMIP project which are atmospheric only GCM with prescribed sea surface temperature. The goal of the project is model intercomparison, not state of art simulations of real climate for which a fully coupled AOGCM is required. Hence the use of AMIP models is at least questionable. Another weakness of the paper is the use of ERBE data for the tropical latitude band to infer a global climate sensitivity. This is aknowledged by the authors themselves in the conclusions. Finally, there are already many different values of the climate sensitivy in the scientific litterature, as clearly pointed out in this post. Only rarely a single piece of work makes huge differences. So, scientifically this work should be first confirmed and then compared to other data and pluged in the big picture; and I'm sure this is exactly what the science community will do. People should then refrain from simply stating that the climate sensitivy is half the accepted value.
    0 0
  6. re #4 HumanityRules, you've done an interesting bit of quote mining, but it doesn't really reflect current understanding. In relation to your highlighting some papers on solar contributions to early 20th century warming, it's useful to know that in the 10 years since the work cited in your quote (from the Introduction to Andronova and Schlesinger (2000) Geophys. Res. Lett. 27(14), 2137-2140), that there has been a significant reassessment of solar irradiance changes as indicated from a number of methods. The contribution of solar irradiance to early 20th century warming is small (of the order of 0.1 oC or less). This is taken into account in more recent analyses of climate sensitivity and is likly part of the reason that the low end of the climate sensitivity is rather better constrained now than in the past, as indicated in John Cook's summary: see for example: Wang YM, Lean JL, Sheeley NR (2005) Modeling the sun's magnetic field and irradiance since 1713 Astrophys. J. 625, 522-538. Krivova NA, Balmaceda L, Solanki SK (2007) Reconstruction of solar total irradiance since 1700 from the surface magnetic flux Astron. Astrophys. 467, 335-346 Lean JL, Rind DH (2008) How natural and anthropogenic influences alter global and regional surface temperatures: 1889 to 2006 Geophys. Res. Lett. 35, L18701 Benestad RE, Schmidt GA (2009) Solar trends and global warming J. Geophys. Res. 114, D14101 Steinhilber F, Beer J, Frohlich C (2009) Total solar irradiance during the Holocene Geophys. Res. lett. 36, L19704
    0 0
  7. I would ask a simple question: why is there an assumption that all forcings should result in the same climate sensitivity? At the simplest level, I could suggest that if a large part of a change in TSI is absorbed as additional (UV absorbed) energy in the stratosphere, and an internal forcing results in ocean heat release and hence cloud formation, and a GHG forcing results in an increased pathlength for photons riding a CO2 principal frequency, then why shold we assume that the power to temperature ratio is the same for all three?
    0 0
  8. Regarding Lindzen/Choi: "Increased ocean temperatures do increase the amount of radiation lost to space - if the Earth is in positive energy imbalance, the planet will accumulate heat, oceans will warm, the earth will radiate more energy to space until it approaches radiative equilibrium again. This is discussed in more detail in the Climate Time Lag post. ]" Does that mean that Lindzen/Choi is correct when they calculate sensitivity to be 0.5 degrees C?
    0 0
  9. PaulK, it's not an a priori assumption; it is consistent with at least three thing (the first three i can think of :) ) 1) it's consistent with what is known on paleoclimate 2) it's what you expect in a strongly coupled system 3) this behaviour is reproduced in climate models Having said this, it's clear that it's just a usefull aproximation with a limited range of validity to compare different forcings.
    0 0
  10. Re #8: No, it just means that hotter things give off more heat, which is the answer to your second question in #2.
    0 0
  11. From what I can tell, the sensitivity relationship being sought here relates to a simplified control model, basically one giant block that lumps all forcings and feedbacks into a single transfer function (for the entire planet!). Would'nt it be more accurate to dissect this problem a little by breaking out the forcings and feedback into separate paths, and isolating each factor, each with its own particular coefficient? Or would this reveal too much about what is really going on? I suspect that this approach was not taken due to the real complexity of the problem. Furthermore, the choice to describe things probabilistically only diffuses the truth of the matter that much more. Ironically, the math uses incertainty to establish the certainty of an idea, or allow some wiggle room if things dont pan out, etc. (In actually, the lumping does not lump all the forcing factors, it basically makes CO2 the ONLY factor, which is fairly unrealistic.) Would we all then be freezing to death today if there was no CO2 in the atmosphere?
    0 0
  12. You're going round in circles RSVP. Of course the forcing from each factor has been isolated "each with its own coefficient". There's dozens, if not hundred's, of papers that parameterize individual forcings. You can read examples of this here: or here: J. L. Lean and D. H. Rind (2008) How natural and anthropogenic influences alter global and regional surface temperatures: 1889 to 2006 Geophys. Res. Lett. 35, L18701 and some of the recent threads on this site address this specific question, as you surely must have noticed; see for example, Figure 2 here: Notice that "lumping all the forcing factors" obviously doesn't "make CO2 the ONLY factor". In fact you've made a rather meaningless statement. Lumping all the forcing factors gives each of the factors its own contribution according to the determined parameterization. I think you might be referring to the definition of a generalized "climate sensitivity" which is defined as the warming response at equilibrium to a given radiative imbalance (or forcing). This climate sensitivity is essentially independent of the specific forcing associated with the radiative imbalance (solar, volcanic aerosol, anthropogenic aerosol, greenhouse gas etc.). So you can’t pretend that this somehow makes CO2 “the ONLY forcing”, when it obviously doesn’t.
    0 0
  13. chris It is clear that you are simply trying to confuse other readers about what I said. Are you a lawyer?
    0 0
  14. not really RSVP. "Other readers" can read your post and determine themselves what you said! You said: " Would'nt it be more accurate to dissect this problem a little by breaking out the forcings and feedback into separate paths, and isolating each factor, each with its own particular coefficient?" and the answer is that this has been done to a rather extraordinary extent (see citations/links in my post above) and you said: "...the lumping does not lump all the forcing factors, it basically makes CO2 the ONLY factor" however, the use of a generalized "climate sensitivity" certainly doesn't "make CO2 the only forcing".. it's pretty simple RSVP. If you think I've misrepresented your post why not explain why you think so. (I'm not a lawyer btw...I'm a scientist, and do try to communicate meaning carefully. If I've messed up in this case you can simply explain why).
    0 0
  15. Chris The paper seeks a single number to describe "climate sensitivity" which according to the author appears to be around 3 degrees C, (although a lot of variance is admitted). Even though the narrative BEGINS with "climate sensitivity", the only forcing being considered is the effects of CO2 operating as an atmospheric greenhouse gas. I am not inventing this. The UNITS of this elusive number represents: degrees C of the "Earth's" temperature rising for a doubling of CO2 concentration. I do not find the information you are referring to in this article, regardless of whether 100 or 1000 other papers cover this subject, indicating a difference between what one imagines is writen and what actually is writen. (actually now a days it doesnt matter, since all you have to do is press REFRESH and you might find the content altered). If the narrative was implying that this forcing was considered in isolation,
    0 0
    Response: The strict definition of climate sensitivity is the equilibrium change to global temperature if the climate experienced a radiative forcing of 3.7 Wm-2. This doesn't mean the forcing comes solely from CO2. In fact, it certainly doesn't as CO2 is not the only driver of climate. But the term 'doubling of CO2' is used because it is more meaningful than the figure 3.7 Wm-2.
  16. I don't think that's quite right RSVP. If you read the early sections of the review under discussion (Knutti and Hegerl, 2008), the authors make it clear up front that the climate sensitivity is broadly independent of its source. They state:
    Anthropogenic emissions of greenhouse gases, aerosol precursors and other substances, as well as natural changes in solar irradiance and volcanic eruptions, affect the amount of radiation that is reflected, transmitted and absorbed by the atmosphere. This externally imposed (naturally or human-induced) energy imbalance on the system, such as the increased long-wave absorption caused by the emission of anthropogenic CO2, is termed radiative forcing (F). In a simple global energy balance model, the difference between these (positive) radiative perturbations F and the increased outgoing long-wave radiation that is assumed to be proportional to the surface warming T leads to an increased heat flux Q in the system, such that...
    As I said above, there are hundreds of papers that describe the ongoing efforts to refine the specific contributions (parameterizations) of each of the forcings. Since there are examples of these in recent threads on this site, anyone who posts here should be aware of the fact that all of the sources of radiative forcings are considered in attribution of contributions to historical and contemporary warming. Scientific papers cannot be expected to regurgitate the entirety of a subject as background, although the work should be placed in context...a certain amount knowledge and undersanding is always assumed. One has to try to be aware of the broader subject when assessing every piece of science...
    0 0
  17. Skeptics have suggested that there may be feedbacks that only react to certain forcings. They base this on Shaviv 2003, claiming that the ocean heat flux correlated to the 11-year solar cycle needs 5-7 times more radiative forcing than the radiativee forcing associated with just the TSI variation. If someone could give me some context on this, I'd be very grateful.
    0 0
  18. PeterPan, maybe the comments starting with Chris's on a different thread on Skeptical Science will have enough pointers? Dunno, I haven't read them all.
    0 0
  19. Thanks Tom! I read that post a couple of months ago and I didn't connected it to this, but I think that the clue is in the main article: Given that there's no trend in TSI since the 50s, the radiative imbalance at the top of the atmosphere should have been decreasing since then until reaching a new equilibrium. However, what we have observed is just the contrary (more energy in than out). I guess that should be true regardless of any amplfication... Thnks again!
    0 0
  20. Could someone please explain to me why the Lindzen and Choi paper makes a vague reference to the climate sensitivity of doubling CO2 being 0.5 C? This seems to be inconsistent with the warming (over 0.7 C) that has already been observed with "only" ~390 ppmv of CO2. Does the CO2 sensitivity implicitly include the net impact of all feedbacks arising once the warming is initiated? That is, it is not only a metric of the response arising solely to CO2, but rather of the integrated response of the entire climate system (oceans, atmosphere, cryposhere) to doubling CO2. Thanks in advance. If the latter, then Lindzen and Choi's estimate seems way off the mark. Thanks in advance.
    0 0
  21. Albatross, there might be problems with how Lindzen managed the ERBE data. The ERBE team recommends to analyze data at 36 (or multiples of 36) days intervals due to drifting of the satellite. Lindzen and Choi apperently didn't follow the advice. Dr. Roy Spencer, not an AGW supporter for sure, repeated the analysis and found significantly smaller feedback parameter. So probably it is Lindzen and Choi's estimate "way off the mark".
    0 0
  22. Thanks Riccardo!
    0 0
  23. I'm trying to get this all straight. So, we look back at T and CO2 correlations over time. Using the correlation relationships, recognizing that the correlations are confounded by obvious additional factors, we build probability distributions to estimate the likely correlation between just CO2 and T. I'm OK with that. Seems pretty straight forward. But stay with me here. Now, in a separate conversation we look into some principles from basic physics describing how CO2 molecules interact with IR. For instance, we know that CO2 absorbs well at characteristic frequencies, like most similar molecules. I'm OK with that too, straight forward stuff. But stay with me here. So there is a correlation between T and CO2, we can describe it with some set of probability distributions, AND, we know that CO2 interacts with IR radiation, or heat, which is also correlated with temperature. So, here is where I get lost. All of a sudden, because we know that Temperature is a covariate of IR radiation, and that Temperature is also a covariate of atmospheric CO2 concentration, we somehow get causality or "Forcing". Transitive causality in atomospheric science. Maybe for some it's "common sense", but with the myriad complex variables comprising atomospheric events, "common sense" doesn't cut it. In addition, the concept that there is a single "forcing" number contained within the probability distributions is offensive to my sensibilities as a scientist. IF..., and this is a big IF in my interpretation, this so-called CO2 forcing does occur, isn't it highly likely that it would vary in a more complex manner than a simple logarithmic function. I am new to this topic in the last year or so, and only recently have I begun to read the foundational work in detail. I am finding it really hard to swallow.
    0 0

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.

The Consensus Project Website


(free to republish)

© Copyright 2024 John Cook
Home | Translations | About Us | Privacy | Contact Us