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


A detailed look at climate sensitivity

Posted on 8 September 2010 by dana1981

Some global warming 'skeptics' argue that the Earth's climate sensitivity is so low that a doubling of atmospheric CO2 will result in a surface temperature change on the order of 1°C or less, and that therefore global warming is nothing to worry about. However, values this low are inconsistent with numerous studies using a wide variety of methods, including (i) paleoclimate data, (ii) recent empirical data, and (iii) generally accepted climate models.

Climate sensitivity describes how sensitive the global climate is to a change in the amount of energy reaching the Earth's surface and lower atmosphere (a.k.a. a radiative forcing).  For example, we know that if the amount of carbon dioxide (CO2) in the Earth's atmosphere doubles from the pre-industrial level of 280 parts per million  by volume (ppmv) to 560 ppmv, this will cause an energy imbalance by trapping more outgoing thermal radiation in the atmosphere, enough to directly warm the surface approximately 1.2°C.  However, this doesn't account for feedbacks, for example ice melting and making the planet less reflective, and the warmer atmosphere holding more water vapor (another greenhouse gas). 

Climate sensitivity is the amount the planet will warm when accounting for the various feedbacks affecting the global climate.  The relevant formula is:

dT = ?*dF

Where 'dT' is the change in the Earth's average surface temperature, '?' is the climate sensitivity, usually with units in Kelvin or degrees Celsius per Watts per square meter (°C/[W m-2]), and 'dF' is the radiative forcing, which is discussed in further detail in the Advanced rebuttal to the 'CO2 effect is weak' argument.

Climate sensitivity is not specific to CO2

A common misconception is that the climate sensitivity and temperature change in response to increasing CO2 differs from the sensitivity to other radiative forcings, such as a change in solar irradiance.  This, however, is not the case.  The surface temperature change is proportional to the sensitivity and radiative forcing (in W m-2), regardless of the source of the energy imbalance. 

In other words, if you argue that the Earth has a low climate sensitivity to CO2, you are also arguing for a low climate sensitivity to other influences such as solar irradiance, orbital changes, and volcanic emissions.  Thus when arguing for low climate sensitivity, it becomes difficult to explain past climate changes.  For example, between glacial and interglacial periods, the planet's average temperature changes on the order of 6°C (more like 8-10°C in the Antarctic).  If the climate sensitivity is low, for example due to increasing low-lying cloud cover reflecting more sunlight as a response to global warming, then how can these large past climate changes be explained?

ice core temps

Figure 1: Antarctic temperature changes over the past 450,000 years as measured from ice cores

What is the possible range of climate sensitivity?

The IPCC Fourth Assessment Report summarized climate sensitivity as "likely to be in the range 2 to 4.5°C with a best estimate of about 3°C, and is very unlikely to be less than 1.5°C. Values substantially higher than 4.5°C cannot be excluded, but agreement of models with observations is not as good for those values."

Individual studies have put climate sensitivity from a doubling of CO2 at anywhere between 0.5°C and 10°C; however, as a consequence  of increasingly better data, it appears that the extreme higher and lower values are very unlikely.  In fact, as climate science has developed and advanced over time , estimates have converged around 3°C.  A summary of recent climate sensitivity studies can be found here

A study led by Stefan Rahmstorf concluded "many vastly improved models have been developed by a number of climate research centers around the world. Current state-of-the-art climate models span a range of 2.6–4.1°C, most clustering around 3°C" (Rahmstorf 2008).  Several studies have put the lower bound of climate sensitivity at about 1.5°C,on the other hand, several others have found that a sensitivity higher than 4.5°C can't be ruled out.

A 2008 study led by James Hansen found that climate sensitivity to "fast feedback processes" is 3°C, but when accounting for longer-term feedbacks (such as ice sheet
disintegration, vegetation migration, and greenhouse gas release from soils, tundra or ocean), if atmospheric CO2 remains at the doubled level, the sensitivity increases to 6°C based on paleoclimatic (historical climate) data.

What are the limits on the climate sensitivity value?


The main limit on the sensitivity value is that it has to be consistent with paleoclimatic data.  A sensitivity which is too low will be inconsistent with past climate changes - basically if there is some large negative feedback which makes the sensitivity too low, it would have prevented the planet from transitioning from ice ages to interglacial periods, for example.  Similarly a high climate sensitivity would have caused more and larger past climate changes.

One recent study examining the Palaeocene–Eocene Thermal Maximum (about 55 million years ago), during which the planet warmed 5-9°C, found that "At accepted values for the climate sensitivity to a doubling of the atmospheric CO2 concentration, this rise in CO2 can explain only between 1 and 3.5°C of the warming inferred from proxy records" (Zeebe 2009).  This suggests that climate sensitivity may be higher than we currently believe, but it likely isn't lower.

Recent responses to large volcanic eruptions 

Climate scientists have also attempted to estimate climate sensitivity based on the response to recent large volcanic eruptions, such as Mount Pinatubo in 1991.  Wigley et al. (2005) found:

"Comparisons of observed and modeled coolings after the eruptions of Agung, El Chichón, and Pinatubo give implied climate sensitivities that are consistent with the Intergovernmental Panel on Climate Change (IPCC) range of 1.5–4.5°C. The cooling associated with Pinatubo appears to require a sensitivity above the IPCC lower bound of 1.5°C, and none of the observed eruption responses rules out a sensitivity above 4.5°C."

Similarly, Forster et al. (2006) concluded as follows.

"A climate feedback parameter of 2.3 +/- 1.4 W m-2 K-1 is found. This corresponds to a 1.0–4.1 K range for the equilibrium warming due to a doubling of carbon dioxide"

Other Empirical Observations

Gregory et al. (2002) used observed interior-ocean temperature changes, surface temperature changes measured since 1860, and estimates of anthropogenic and natural radiative forcing of the climate system to estimate its climate sensitivity.  They found:

"we obtain a 90% confidence interval, whose lower bound (the 5th percentile) is 1.6 K. The median is 6.1 K, above the canonical range of 1.5–4.5 K; the mode is 2.1 K."

Examining Past Temperature Projections

In 1988, NASA climate scientist Dr James Hansen produced a groundbreaking study in which he produced a global climate model that calculated future warming based on three different CO2 emissions scenarios labeled A, B, and C (Hansen 1988).   Now, after more than 20 years, we are able to review Hansen’s projections.

Hansen's model assumed a rather high climate sensitivity of 4.2°C for a doubling of CO2.  His Scenario B has been the closest to reality, with the actual total radiative forcing being about 10% higher than in this emissions scenario.  The warming trend predicted in this scenario from 1988 to 2010 was about 0.26°C per decade whereas the measured temperature increase over that period was approximately 0.18°C per decade, or about 40% lower than Scenario B.

Therefore, what Hansen's models and the real-world observations tell us is that climate sensitivity is about 40% below 4.2°C, or once again, right around 3°C for a doubling of atmospheric CO2.

Probabilistic Estimate Analysis

Annan and Hargreaves (2009) investigated various probabilistic estimates of climate sensitivity, many of which suggested a "worryingly high probability" (greater than 5%) that the sensitivity is in excess of than 6°C for a doubling of CO2.  Using a Bayesian statistical approach, this study concluded that

"the long fat tail that is characteristic of all recent estimates of climate sensitivity simply disappears, with an upper 95% probability limit...easily shown to lie close to 4°C, and certainly well below 6°C."
Annan and Hargreaves concluded that the climate sensitivity to a doubling of atmospheric CO2 is probably close to 3°C, it may be higher, but it's probably not much lower.

Figure 2: Probability distribution of climate sensitivity to a doubling of atmospheric CO2

Summary of these results

Knutti and Hegerl (2008) presents a comprehensive, concise overview of our scientific understanding of climate sensitivity.  In their paper, they present a figure which neatly encapsulates how various methods of estimating climate sensitivity examining different time periods have yielded consistent results, as the studies described above show.  As you can see, the various methodologies are generally consistent with the range of 2-4.5°C, with few methods leaving the possibility of lower values, but several unable to rule out higher values.

sensitivity summary

Figure 3: Distributions and ranges for climate sensitivity from different lines of evidence. The circle indicates the most likely value. The thin colored bars indicate very likely value (more than 90% probability). The thicker colored 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.

What does all this mean?

According to a recent MIT study, we're currently on pace to reach this doubled atmospheric CO2 level by the mid-to-late 21st century.

Figure 4: Projected decadal mean concentrations of CO2.  Red solid lines are median, 5%, and 95% for the MIT study, the dashed blue line is the same from the 2003 MIT projection.
So unless we change course, we're looking at a rapid warming over the 21st century.  Most climate scientists agree that a 2°C warming is the 'danger limit'.   Figure 5 shows temperature rise for a given CO2 level. The dark grey area indicates the climate sensitivity likely range of 2 to 4.5°C.
key global warming impacts 
Figure 5: 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 stabilize CO2 levels at 450 ppmv (the atmospheric CO2 concentration as of 2010 is about 390 ppmv), according to the best estimate, 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 5. The tight constraint on the lower limit of climate sensitivity indicates we're looking down the barrel of significant warming in future decades.

As the scientists at RealClimate put it,
"Global warming of 2°C would leave the Earth warmer than it has been in millions of years, a disruption of climate conditions that have been stable for longer than the history of human agriculture. Given the drought that already afflicts Australia, the crumbling of the sea ice in the Arctic, and the increasing storm damage after only 0.8°C of warming so far, calling 2°C a danger limit seems to us pretty cavalier."

This post is the Advanced version (written by dana1981) of the skeptic argument "Climate sensitivity is low". Note: a Basic version is on its way and should be published shortly.

0 0

Printable Version  |  Link to this page


Prev  1  2  3  Next

Comments 51 to 100 out of 116:

  1. Berényi Péter, I am and was sure you know the problems quite well and that you're well aware that it does not have "immediate consequences either on numerical integration of these beasts ..." so widely used in many fields. For the same reason, I think that writing of that mathematical problem was just to make noise and confuse people.
    0 0
  2. I agree entirely that humans are causing global warming, and that solving the problem should be our highest priority. However, I am very skeptical of the accuracy of climate models and their global temperature projections for the 21st century. I think a global temperature increase of 3 degrees celsius is a bit wide of the mark. Since 1880, global temperatures have only risin .8 degrees celsius.In order for global temperatures to rise by 3 degrees celsius by 2100, global warming would have to undergo a very rapid acceleraton. Apparently this isn't showing any signs of happening. Global temperature increases over the past decade have been on the low end of climate model projections. I'm not saying that warming has stopped, but I am saying its falling short of climate model projections. I find it very hard to believe that we will have a global temperature increase of 3 degrees celsius by 2100. Could you please explain this to me?
    0 0
  3. Karamanski, the rate of CO2 emissions has been increasing since 1880, so the rate of CO2 concentration increase has been increasing, so the rate of temperature forcing from CO2 has been increasing, so the rate of temperature increase since 1880 has been increasing. Therefore the temperature increase rate can't be extrapolated as a linear increase from 1880 to the present. CO2 emissions will continue to increase. Perhaps the most convincing picture of the consequence is a graph of the temperature increase if CO2 emissions instead remained at their current level ("constant emissions"). That does not mean business as usual, because business as usual means the same amount of CO2 per unit of energy produced. Business as usual means an increasing amount of CO2 emitted, not merely an increasing amount of CO2 concentration resulting from constant emission. That's because increasing industrialization and increasing population will result in an increasing number of sources of and amounts of emissions. Business a usual will result in an increasing rate of CO2 emission. The lower bound of the business as usual scenario is the "constant emissions" scenario, in which the increasing number of sources is offset by reduced rates of emissions from each of the sources. That is the curve labeled "4. Constant emissions" in this figure reproduced at RealClimate. Even in that possibly unreasonably optimistic case, the temperature climbs dramatically because of the continued increase of amount of CO2 accumulating in the atmosphere. The business as usual case will be worse than that--much worse.
    0 0
  4. Whoops, I should have linked to a graph right here on Skeptical Science--a graph that includes as well some of the IPCC scenario projections for scenarios worse than constant emissions.
    0 0
  5. Karamanski #52: "Since 1880, global temperatures have only risin .8 degrees celsius." Only? You do realize that's the most extreme temperature increase, in both amount and rate, since the end of the last ice age right? "In order for global temperatures to rise by 3 degrees celsius by 2100, global warming would have to undergo a very rapid acceleraton. Apparently this isn't showing any signs of happening." It isn't? The actual temperature data seems to show that it is. The graph below shows both linear and quadratic 'best fit' trend lines against the GISS data. Note that the best quadratic fit is an increasing temperature slope. Not particularly surprising given that 2000-2009 was the hottest decade on record and 2010 is thus far the hottest single year on record. "Global temperature increases over the past decade have been on the low end of climate model projections." That's like saying, 'there were ten days at the end of Spring that were on the low end of average temperatures... therefor it seems unlikely that temperatures will increase this Summer'. You're looking at a short term fluctuation, no greater than half a dozen such since 1880, and ignoring the long term trend.
    0 0
  6. When I said "only .8 degrees celsius" I was comparing the global warming so far to what is projected for the future. A rise of .8 degrees celsius a very steep increase relative to the known paleoclimate record. I also was not saying that warming has stopped. This decade was indeed the hottest on record and 2010 is on track to tie or exceed 2005 as the warmest calender year on record. Even though warming may continuing the rate of global temperature increase is steady and is not accelerating. From what I think, assuming that climate sensitivity isn't super high, I find it diffult to think that we will have a temperature increase of 3 degrees celsius by 2100. For example, thermal inertia might be much larger and slower than expected. I'm still very skeptical of the accuracy of climate models. For instance, they did not predict the halt in the increase of methane concentrations over the past decade. Arctic sea ice predictions by climate models are 40 years behind real observations. Please explain this.
    0 0
  7. @Karamanski: "Even though warming may continuing the rate of global temperature increase is steady and is not accelerating." CBDunkerson's graph seems to suggest otherwise. Instead of restating your original argument, you should respond to that counter-argument. "For example, thermal inertia might be much larger and slower than expected." It seems we would have observed this already. Do you have any scientific article that suggest this may be happening?
    0 0
  8. Karamanski #56: "Even though warming may continuing the rate of global temperature increase is steady and is not accelerating." Again, the data suggests that the rate IS increasing. In addition to the up-slope of the quadratic fit note that the linear trend line has been consistently below the actual results for 15 years (since 1995). With a steady trend you see the random 'noise' causing the actual results to frequently fall both above and below the linear trend. When actual results start to fall only on one side of the linear trend line it is a clear indication that the actual trend is NOT linear (or 'steady' as you suggest), but rather curving in the direction of the variance. "For instance, they did not predict the halt in the increase of methane concentrations over the past decade." The slight slowing in the rate of increase of atmospheric methane levels a few years ago was generally attributed to changes in human emissions... humans are somewhat difficult to predict with climate models - which is why they are generally run with a range of different GHG emission assumptions. That said, methane levels have resumed rising more quickly and there is some indication that this is becoming a positive climate feedback rather than primarily an issue of human emissions. See here. "Arctic sea ice predictions by climate models are 40 years behind real observations." The IPCC tends to be fairly conservative. Some other examples of this can be found here and here. In the case of Arctic sea ice extent the primary factors not taken into account seem to have been increased ice export out of the Arctic region as the ice breaks up and greater than expected 'melt from the bottom up' as ocean temperatures increase. "I find it diffult to think that we will have a temperature increase of 3 degrees celsius by 2100." It should be noted that the 3 C estimate is not just a matter of models... the positive feedbacks assumed for that figure (primarily increased atmospheric water vapor and decreased Arctic ice) are evident in direct observations of current conditions. The figure is also consistent with paleoclimate reconstructions. So... past (paleoclimate), present (observations), and future (models) analysis are consistent with the 3 C figure.
    0 0
  9. To RW1 on the Lindzen thread, muoncounter recommended that you visit this thread for evidence of high sensitivity. I second that, but IMO the arguments here come up short in several respects. One is that the average measurement of higher water vapor do not take into account the distribution of WV. If it is higher and evenly distributed then it is a positive feedback to CO2 warming. But if WV is unevenly distributed in a world warmed slightly by CO2, then an average increase in WV will result in less or no amplification. Second, the derivation of sensitivity from paleo studies routinely ignores unmeasured confounding factors. I gave one possibility here: cosmic-rays-and-global-warming.htm but there are others. Typically the response is to treat solar geomagnetic variations as a proxy for TSI and then dismiss it because of poor correlation and low amounts of TSI change. Also the last 30 years of detailed measurements don't show much in the way of GCR related climate effects. However the penultimate interglacial coincides with an abrupt decline in GCR so a relatively small TSI increase could be amplified without the need for CO2 feedback.
    0 0
  10. Hopefully this will answer chris on the Lindzen thread. Disregarding my critique of paleo studies of sensitivities above, I still do not believe that we can take a sensitivity calculated in paleo records and use it in a linear fashion. For one thing the paleo sensitivities reflect long term correlation which may be somewhat linear. For example as oceans warm over hundreds or thousands of years, CO2 is released in a more or less linear fashion. But the short term is nonlinear. Short term sensitivity is based on water vapor feedback. But water vapor feedback is highly nonlinear as evidenced by daily tropical weather cycles and seasonal changes in weather (larger NAO fluctuations in winter than in summer is just one of many examples). The sensitivity that was based on long term factors shown in the ice cores has nothing to do with a sensitivity based on the short term factors. Furthermore, neither sensitivity is applicable to our current interglacial regime. The longer term sensitivity only applies to glacial to interglacial transitions. So an attempt to use that sensitivity for a current increase (50%, doubling, or other) in CO2 requires waiting for the long term responses (centuries at least) and won't show up in a few decades of data.
    0 0
  11. Eric, Paleo studies were not the only source for calculating climate sensitivity discussed in this post. Also discussed were measurements based on recent volcanic eruptions and on the modern warming trend itself. Both of these reflect short term sensitivity (and are obviously applicable to the current climate regime). The results from these studies agree with those from the long term paleo studies. Furthermore, the bottom-up physical simulations agree with the top-down statistical studies. Thus, the evidence does not support the suggestion that modern short term climate sensitivities vary substantially from those derived from paleo studies.
    0 0
  12. e, thanks for responding. I will obviously have to look at each study. The biggest difference is that short term amplification from water vapor feedback cannot be compared with long term amplification from a CO2-temperature feedback cycle. Even if we argue that the paleo studies were correct (no other unmeasured factor caused the majority of the temperature increase which then caused the CO2 increase by many centuries of ocean warming), that CO2 to T to CO2 amplification factor does not apply to water vapor for many reasons (e.g. it is not seasonal or geographic unlike water vapor which is). Another key factor is that the ice age to interglacial transition used to estimate sensitivity encompassed a nonlinear change in the warming effect of water vapor (in the present climate there is a much larger ratio of cooling effects from water vapor to warming effects than in the ice age climate). The only valid use of paleo sensitivity is the argument that CO2 increases have short-circuited the long slow feedback process and will emerge in 3C warming after ocean inertia. My very simple answer to that dilemma is that CO2 is sequestered with ocean turnover just the way heat is sequestered. So if the ocean has quickly absorbed the heat, it will also quickly absorb the CO2 as soon as we stop producing it. You may see that as a good argument for the many ideas on the solutions threads here and I agree and I agree with those threads. But I also realize that the paleo record shows a distinct nonlinearity in the region we are in that indicates the opposite of "tipping points" but rather a stabilization in temperature.
    0 0
  13. #62: "So if the ocean has quickly absorbed the heat, it will also quickly absorb the CO2 as soon as we stop producing it." The oceans do nothing quickly; see the graphs here. Oceans are absorbing CO2 now, as they are acidifying. You can't just declare oceans 'quickly absorb the CO2 as soon as we stop', as the oceans might might just give it right back to us (if we stop; that's a good one). See the ocean acidification page; you will find some parts of the ocean are sourcing CO2, while others are sinking it. "the paleo record shows a distinct nonlinearity in the region we are in that indicates the opposite of "tipping points" but rather a stabilization in temperature." In this and in #60, you've made a lot of grand generalizations. In order for general statements like as oceans warm over hundreds or thousands of years, CO2 is released in a more or less linear fashion and longer term sensitivity only applies to glacial to interglacial transitions to be accepted into the debate, you must provide some evidence.
    0 0
  14. Eric the Red here, you can simply combine the formulas by multiplying the constant in the Radiative Forcing formula by λ. It is unwise to do so, however, because the formula for other radiative forcings is quite different, while to a first approximation, the formula for climate sensitivity is the same across all forcings. Thus the formula for solar forcings 0.25*0.7*dI where I is the incident solar radiation on a meter squared area perpendicular to the solar radiation at the top of the atmosphere. Other formulas are given by the IPCC.
    0 0
  15. Tom, Sorry for the delay. Yours was a better analysis than mine, but it looked like John was confusing the equation by substituting oC for W/m2. Hence, he was arriving at a higher figure for climate sensitivity.
    0 0
  16. What's missing from the paleo paragraph in the OP is that paleo estimates of uncertainty contain one or more of the following: "No, poorly understood, large uncertainties, very few studies or poor agreement, (un)known limitations, low confidence" for most estimation criteria, particularly similar climate base state, similar forcings and feedbacks, etc. As I explained above about a year ago, the transition from glacial to interglacial conditions have starting conditions that don't apply to today's climate. Even if a 3C estimate of paleo-sensitivity is accurate (considering it uses the same models as other estimates so it is not an independent estimate), it doesn't apply to modern climate.
    0 0
  17. Eric @66, "Even if a 3C estimate of paleo-sensitivity is accurate (considering it uses the same models as other estimates so it is not an independent estimate), it doesn't apply to modern climate. " I think you meant to say " doesn't necessarily apply to modern climate". You do not unequivocally know that a priori.
    0 0
  18. Eric @66, "As I explained above about a year ago, the transition from glacial to interglacial conditions have starting conditions that don't apply to today's climate." Thanks, but we do not need you to explain this matter to us Eric. Besides, you are floating a red herring, b/c what you say is not necessarily true as shown in Fig3a of Knutti and Hegerl (2008). You need to look at Fig 3a again very closely. There is very close agreement between the 66% probability ranges (and medians) for EQS derived from the "General circulation models" and those derives from "Proxy data from millions of years ago". In fact, multiple independent lines of evidence (including those from modern/recent times)are largely consistent with the 66% confidence interval and median reported in AR4.
    0 0
  19. Albatross, applicability is what part b of figure 3 is all about. For both the LGM and millions of years old data, the first two boxes are red because the base state and feedbacks and forcings are not similar. The text explains further that "simple calculations" relating the cooling to changes in radiative forcing yield sensitivity ignoring all the nonradiative changes (e.g. convection, weather in general) that determine equilibrium temperature. The other choice is using a GCM with LGM or older conditions but that does not produce an independent sensitivity estimate.
    0 0
  20. Eric @69, You are ignoring my key points. I'm not even sure what your post @69 is meant to be in response to. I was speaking to Fig. 3a. And again, I do not need for you to explain to me how to read a graph, you can quit being condescending. "the first two boxes are red because the base state and feedbacks and forcings are not similar." Yet the the fact remains that "There is very close agreement between the 66% probability ranges (and medians) for EQS derived from the "General circulation models" and those derives from "Proxy data from millions of years ago". Do you deny what those data shown in Fig. 3a are showing? Do you deny that?
    0 0
  21. Regarding the value of climate sensitivity and the paleo data. This is what respected scientist Dr. Matt Huber (Purdue University) whose research is "focused on past, present and future climate, the mechanisms that govern climate, the different forms that climates can take on Earth, and the relationship between climate change and life [Source]: "Climate scientists don’t often talk about such grim long-term forecasts, Huber says, in part because skeptics, exaggerating scientific uncertainties, are always accusing them of alarmism. “We’ve basically been trying to edit ourselves”, Huber says. “Whenever we we see something really bad, we tend to hold off. The middle ground is actually worse than people think. “If we continue down this road, there are really is no uncertainty. We’re headed for the Eocence. And we know what that’s like." Dr. Matt Huber, October 2011. Now some might be so arrogant to believe that they know more about this issue than does Dr. Huber; I for one do not. Also, to most reasonable people his observations and expertise should be quite worrisome.
    0 0
  22. Eric (skeptic) @66, I presume this is a continuation of the discussion from here, and is a reference to the information in figure 3b from Knutti and Hegerl, 2008: That being the case, as presented your comment is nonsense. It is not possible for a method to be (or for the uncertainties to contain) "No, poorly understood, large uncertainties, very few studies or poor agreement, (un)known limitations, low confidence". For your comment to make sense, you need to split the analysis into the respective categories, and then apply the appropriate descriptor. If you do not do that, then you would need to apply all three cluster descriptions, for (for all paleo methods) they have both red, yellow, and green classifications. More importantly, that analysis does not recognize the concordance across a range of paleo studies. Studies of both the Last Glacial Maximum (20 thousand years ago) and of Pliocene (5.3 to 2.6 million years ago, and when CO2 levels where last at their current levels) both study periods when the Earth did not have a "Similar climate base state" to its current condition, the LGM being about 6 degrees C colder, and the Pliocene was about 2-3 degrees warmer. Never-the-less, studies of both periods have come up with climate sensitivities in the range of 2 to 4 degrees C per doubling of CO2. The supposition that climate sensitivity for the current climate base state will some how be much lower than that range, when both warmer and colder conditions have a higher climate sensitivity is magical thinking. In short, paleo-climate studies show the climate sensitivity to be very robust with respect to climate base state. That is not evident from looking at just studies of the LGM, or just studies of the Pliocene. Hence looking at the seperate categorization provided by Knutti and Hegerl with regard to climate base state is misleading if you do not recognize the robustness of the results. While I have focused on the LGM and Pliocene as the best understood paleo-eras, the robustness of climate sensitivity has extended across a range of conditions from snowball Earth through to Saurian Sauna. There is undoubtedly some variation of climate sensitivity across that range, but that range is very likely to be smaller than the uncertainty in determination of climate sensitivity, ie, the range of climate sensitivities is likely to fall in the 2 to 4 degree C per doubling of CO2 range. Finally, even if we suppose ourselves to be in a goldilocks zone for climate sensitivity, that is bad news. Suppose the modern climate sensitivity is in fact 2 degrees C per doubling, but that this is due to the fact that we are in a goldilocks zone. At that climate sensitivity, business as usual will still lift global temperatures by about 2-3 degrees C by the end of this century. But that is enough to lift us into a Pliocene base state with its higher (on this supposition) climate sensitivity but with our having much higher than Pliocene CO2 concentrations. In other words, we would face accelerating global warming even if we had managed to stop further emissions. In this scenario, the equilibrium temperature would still be that determined using a Pliocene (or even Cretacious) base state because our current activities would lift us to that level.
    0 0
  23. Albatross and Tom, thanks for your responses. I have to go to meeting right now, but I will try to give a thorough answer later.
    0 0
  24. Tom @72, "The supposition that climate sensitivity for the current climate base state will some how be much lower than that range, when both warmer and colder conditions have a higher climate sensitivity is magical thinking." That was pretty much the gist of my post @68. That is not my opinion of course, but what is borne out by multiple, independent studies by people who specialize in this field. I have a sneaking suspicion though, that Eric thinks he knows better than KH08 and Dr. Huber and the collective body of (robust) evidence.
    0 0
  25. Albatross, and Tom, yes, this discussion is about K&H08, figure 3b. I'd like to ask 1) whether the red squares in 3b in your comment are useful added information for threads such as this one (to wrap up the discussion from Serial Deleter thread), 2) whether the lines of evidence in K&H2008 fig 3b are independent and thus have meaningful agreement (to address Albatross's post 68), and 3) about paleoclimate sensitivity estimates (Tom's robustness claim in 72). I'd also like to discuss Dr. Huber's "heading for the Eocene" claim after reading a journal article or two (I found some links with google scholar and will choose some of those if necessary). On 1, the link to an unrestricted PDF of K&H08 is a good way to provide a lot of context including the added information on the sensitivity estimates in figure 3a that are shown in 3b. I would point out that K&H refer mostly to figure 3 rather than "3a" to point out various aspects of PDFs. But they separately call out fig 3b: "Figure 3b is a partly subjective evaluation of the different lines of evidence for several criteria that need to be considered when combining lines of evidence in an assessment." Thus the threads at SkepSci that "combine lines of evidence in an assessment" should utilize the information from fig 3b in some way, probably in the text. I think the OP in this thread, for example, is combining lines of evidence to some extent. Specifically I think the caption of figure 3 above or elsewhere in the OP should describe some of the red boxes. I think that would match the authors intent for assessments. On 2, the primary issue is the uncertainty of the climate regimes in paleoclimate. The radiative equilibrium is at TOA and that depends on many climate variables including convection in various zones, jet stream position and other factors that affect the effective radiating temperature. Those in turn are affected by ocean circulation patterns and other short term factors. These factors are not radiative forcings and can not be used to calculate sensitivity. There are many medium term factors (e.g. tilt and eccentricity) that are also not simple radiative forcing changes. While some of these can be simplified into a number like an albedo change, mostly these require a GCM. With any climate regimes and nonradiative forcing there is a CO2 amplification. The amplifier uses rising temperature to create more GHG and then more warming. We currently are short circuiting the amplifier by raising CO2 directly. This is not hard to model and can be used to predict the immediate temperature response (e.g. line-by-line applied to the average cloudy and clear day and nights over various zones. The problem comes when the regime changes, atmospheric (clouds and other weather) and oceanic weather (circulation) changes in a warmer world means that a GCM is required. Once a GCM is required to determine the temperature response to a radiative forcing, the evidence for forcing is the same, the GCM, not the data since the GCM is tuned to match the data (applies equally to paleo or present data). So to answer your question Albatross, yes, there is agreement, but it is because the same models are used. On 3, the sensitivity calculation for paleoclimatic data range from "Simple calculations relating the peak cooling to changes in radiative forcing..." to GCMs. Yet, transitions can be forced a number of nonradiative factors that can create metastable states, e.g. In another example the temperature changes come from a number of factors that are difficult to evaluate and measure in paleo data. At this point I do not have enough information to know how to define robustness in such estimates. Can you define that for me and I will try to read more on how such estimates are made. Unfortunately I'm not going to be able to do much today, until later this evening.
    0 0
  26. Eric (skeptic) @75, using your numbering: 1) On the Serial Deleter thread, I gave two criteria under which SkS practice should be considered sub-par: a) That "...not including figure 3b of Knutti and Hegerl has resulted, or facilitated, in misrepresentation of some part of Knutti and Hegerl by SkS"; or b) That "...not including figure 3b of Knutti and Hegerl has resulted in a failure to canvass issues that should have been canvassed in a post on SkS." In case (a), you would have found an example in which SkS practice could be (and should be) criticized in much the same manner as Michaels is criticized in the OP of the "serial deleter" thread. In case (b), you would have found an example in which SkS practice did not rise to the level of misrepresentation, and hence was not so criticizable; but in which it still represented a serious lapse from the standards which we would expect. It is noticeable that you do not criticize the OP in this article on either of these grounds. Instead you criticize it on the much weaker grounds that it contained relevant information. Of course, it does contain relevant information. Fortunately, however, we do not and should not expect articles of any nature to contain all relevant information. Where they to do so, nothing short of a expanded exposition of any scientific paper referred to would be acceptable; for, of course, all the information in those papers is relevant, and so is the background knowledge that could be expected of the scientists to whom the papers are addressed but which cannot be expected of the general public (hence the expanded exposition). Such a requirement would make any blog post extraordinarily long and turgid even by my long winded standards. You may suggest that the information in this case is so relevant that it should have been included, or that the information could have been included simply by including the figure (or both). The second, however, is false simply because the figure assumes background information, part of which I expounded after the bolded "more importantly" in my post 72. Because the figure is easy to misinterpret without background knowledge (and doubly so given the presence of people who would IMO willfully misinterpret it), its inclusion would have required adding at least two paragraphs to the two paragraph section on paleoclimate above, plus additional paragraphs in other sections in that the figure is not restricted to paleoclimate. Given an intent to communicate succinctly and clearly, the additional text required to include the relevant information from the graph was too high a load for the additional information required. And of course, much of the information about certainty is already included in the error bars from figure 3a, so it is not true that the additional relevant information should have been expanded on in the original article. Of course, the additional information is included in the article in the minimal form of a link to the original article by Knutti and Hegerl. Given this, and given the factors mentioned above, while the information in figure 3a is relevant, it was certainly not sufficiently relevant that it needed to be canvassed explicitly in the OP. That fact should lay to rest any suggestion that SkS is applying a double standard with respect to the use of graphs in its criticism of Michaels. 2 & 3) The issue of uncertainty in climate regimes is a major factor when considering a sensitivity analysis in relation to a particular period. It ceases to be a significant factor when a large number of paleo sensitivity analyses are made across a large range of different climate regimes. That is because a robust result found across all paleoclimate regimes almost certainly applies in any particular regime, including our current regime. I know that is simply repeating my point from 72 above, but I do not see anything in what you write that tends to rebut it. Yes there are other non-forcing related factors which adjust the base global average temperature for a given radiative forcing, either by altering the strength of feedbacks, or by altering the heat distribution of the Earth (with greater variation allowing a lower Global Mean Surface Temperature). Changes in land cover and land/sea proportions could also make minor changes to emissivity which would also adjust GMST relative to a particular forcing. But it is extraordinarilly odd that these factors should some how combine to produce the same apparent climate sensitivity when comparing either the pliocene (3 C warmer, low ice, higher sea levels) or the LGM (6 C cooler, much ice, lower sea levels) to the present day. It is even odder that with a climate sensitivity for doubling CO2 of between 3-4 degrees C during non-glacial periods and double that during glacial periods (due to the increased feedback from ice albedo effects) explains 85% or more of the variation in CO2 concentration, as found by Park and Royer (2011). Please note that this is the long term climate sensitivity, ie, the climate sensitivity allowing for meltback or growth of "permanent" ice sheets. In the absence of such ice sheets, that should approximate to the Charney Climate sensitivity discussed in the article above. Hence Park and Royer tends to confirm that the Charney climate sensitivity is in the range of 3 to 4 degrees C, and has been across all geographical conditions encountered in the last 540 million years. Once weatherable land surface is factored in, the primary determinant of CO2 levels is temperature. As Park and Royer show, factoring in weatherable land area allows an even greater explanation of the variance in CO2 levels with a presumed non-glacial climate sensitivity of 3-4 degrees C (figure 10). Hence a climate sensitivity given feed backs excluding long term changes in "permanent" ice sheets of 3-4 degrees C is a very robust feature of the phanerozoic.
    0 0
  27. Thanks Tom, for the answers and the link to the paper. I had an unexpected trip and am 4 pages of comments behind. I don't think that potential willful misrepresentation of fig 3b is a valid reason to exclude it, so the bottom line is it was deleted for space reasons. The paper with Royer has this claim: "We confirm the conclusion of Royer and others (2007): the necessity for greenhouse-weathering feedbacks in Earth’s long-term carbon cycle makes low values for Earth’s long-term climate sensitivity (delta)T2x highly unlikely." I understand that the driving factor for weathering is geography, not temperature, For example it is the explanation for The End of the Hothouse The drop in CO2 at the end of the hothouse explains the drop in temperature, but is difficult to resolve to a sensitivity number due to the large amplification from the newly formed Antarctic ice sheet. I don't believe their "empirical PDF" is a PDF, it appears to be a result of multiple runs an oversimplified model that leaves out factors that cannot be determined from the paelo record or are not included. The way I read the paper is that GEOCARB/SULF models are temperature to weathering models, but do not include geographic changes, is that correct?
    0 0
  28. Eric (skeptic) @77, your assumptions about the GEOCARBSULF model are incorrect. It includes geographic changes including rate of erosion due to expose land area and orogeny, rate of vulcansim, effects of erosion rates due to glaciation, and the effects of vascular and non-vascular plants. More details can be found in the description of the GEOCARBSULF model by Robert Burner. For some details you will need to consult the description of the GEOCARBIII model, which preceded GEOCARBSULF. The later reference shows the sedimentation rates and the "weathering uplift parameter" for various epochs of the phanerozoic, and discusses how they are determined. The important point at this level of discussion is that they are determined empirically. Given these factors, and given the fact that temperature plus CO2 concentration control the rate of chemical weathering, and given a particular ratio between CO2 concentration and temperature, it is possible to retrodict the CO2 concentration in a given period using the model. By varying the ratio of CO2 concentration to temperature, you can determine which ratio gives the best fit to the geological record of CO2 concentration. Of course, as the CO2/temperature ratio is just the climate sensitivity, you at the same time determine the best fit climate sensitivity over the entire phanerozoic. This was first done by Berner, Royer and Park (2007), which explains the methodology. It is true that GEOCARBSULF does not model specific geographical distribution of continents. This means there are significant factors which effect temperature, but which are not modeled. The question is, however, how significant? If their impact relative to CO2 concentration and temperature is large, then it will be impossible to get a good match between predicted CO2 levels and CO2 levels as observed in the geological record using this technique. Contrary to that, however, the fit is quite good: (From Park and Royer 2011, fig 9d. Alternative fits under different assumptions in figures 9 a-c and figure 10 should also be examined) There are, of course, to periods of significant mismatch. That may be because of problems in the record of erosion (see figure 10 and related discussion). More probably, IMO, it is because particular geographical configurations changed the climate base state. Or it could even be because the climate sensitivity was significantly different in those periods (which is a distinct possibility from the geographical change of the climate base state). Finally, the PDFs are indeed PDFs. Given a set threshold for explained variance in the CO2 concentration, the PDF maps the probability that a particular climate sensitivity (or climate sensitivity pairing, where glacial is distinct from non-glacial) will explain that degree of variance. However, like all statistical measures, a simplistic interpretation can be risky (and I am not the one too explain the risks of over interpretation in this case). However, it is not over interpretation to say that given the evidence in this study, "...the empirical PDFs for glacial climate sensitivity predict T2x(g)>2.0 °C with 99 percent probability, T2x(g)>3.4 °C with 95 percent probability, and T2x(g)>4.4 °C with 90 percent probability", and that "[t]he most probable values are T2x(g) 6° to 8 °C."
    0 0
  29. Re: sensitivity, assuming our peak CO2 is 560 ppm. I cannot develop a probability distribution based on evidence because there is no evidence to support one, i.e. there is no evidence for any particular distribution such as a normal distribution and uncertainties about feedback are discrete, not based on some continuous function. Running any particular model or set of models multiple times produces a distribution WRT those models not WRT reality. I would probably just choose one method and one number although I would prefer using two: modern temperatures and LGM with paleo data as it becomes available. (I rule out using glacial to interglacial for reasons I discussed here Using the modern temperature rise I would take the 4/10 of a doubling in CO2 that we have had so far, the 0.7 rise of temperature (assume 1/2 the pre-1940 was natural) then 0.7/0.4 is about 2 C. Uncertainties include GHG feedbacks from warming (tends to be longer run), weather feedback (could be + or -) and exogenous factors that could go either way. For example an active sun would likely produce a greater multiplier of CO2 warming than a quiet sun. We can't really predict the sun past a decade or two.
    0 0
  30. 79, Eric, From the other thread, you are assuming at least a doubling of CO2 as a certainty, correct? So the question is narrowed to one of sensitivity per doubling, and the danger presented by that sensitivity. I understand your reluctance to offer a probability, due to the complexity of the issue, but you should be able to recognize a few things. Please tell me which of these statements you agree to: 1) Sensitivity is likely, in the best case, to be no less than 2˚C per doubling. 2) A chance of a sensitivity of 2.5˚C per doubling must be considered to be at least 30%. 3) A chance of a sensitivity of 3˚C per doubling or higher must be considered to be at least 10%. Do you accept any or all of these statements as very likely to be true?
    0 0
  31. Sphaerica, yes I assumed doubling to 560ppm. I think all three of your statements are very likely to be true based on various possible positive feedbacks. I also think there are chances of technological progress in 50-100 years to offset those possibilities, see my post on this thread It could also be that we get larger positive feedback and fail to attain sufficient technological progress, but I think that is a very small probability because progress in science and technology is not contingent on political will or economic incentives (although they both help).
    0 0
  32. 81, Eric (skeptic), First, I don't want to derail from our train of thought, but as far as solutions go... once the CO2 genie is out of the bottle... you need to realize that it was very easy to extract 337 gigatons and counting of carbon in liquid form from the ground, burn it, and release it into the atmosphere as a widely scattered gas. See this past comment of mine on the numbers to see exactly how gigantic that problem is now (let alone after we reach 560ppm). Also note that your hopes for technological progress are in fact very dependent on maintaining the robust nature of our civilization. If climate and resource pressures grow too great, if some countries see their infrastructure collapsing while others invest their energies in other directions (defense in an increasingly unstable world, the need to maintain dwindling food supplies, the need to find new energy sources), then the resources available to dedicate to the difficulty of correcting the problem will be less. We might have been able to do so, if everything stayed the same, but will we be able to when civilization is under severe pressure exactly caused by our lack of solutions today? But that's a digression... more to the point: You accept certain guesstimated probabilities for higher levels of sensitivity. We have as yet not quantified the chances of future technological miracles which allow us to ignore the simple, available solutions we have at present. We will get to this eventually. But, given possible climate sensitivities of 2˚C, 2.5˚C or 3˚C or more, and recognizing that at least some of the extreme fire, drought, flood and temperature events that we see today are almost certainly connected to the meager global temperature change that we have achieved to date (which, because of lag time, is far less than the change to which we are already committed, even if we were to stop all emissions completely today), and that those events point to the expense and hardship their continued and increasing existence would pose... Can you put a number (in lives, dollars, whatever) to what you think the impact of a 2˚C, 2.5˚C or 3˚C or more climate change will be on the citizens of your own country, and on various people around the world in general? Can you in any way (just for the purposes of ball-park decision making) quantify the danger that a higher climate sensitivity implies?
    0 0
  33. Sphaerica, nobody should die from floods. If the frequency increases due to global warming, that does not change the needed preparations. But the magnitude will also increase. In Thailand that means about 2.5% of GDP for flood control and other costs, see for the costs, about 15 billion US dollars. For the US, our costs will be a bit smaller percentage, but for developing countries a much larger percentage of GDP. Bear in mind that there is some cost regardless of global warning. Temperature can be mitigated and there are savings from lower heating costs (US on average spends 1/2 as much on cooling). Droughts will be harder to mitigate than floods, maybe only with a long term change from agriculture to some form of industry. Like with floods the third world countries will have the largest impact relative to their economies. My ballpark cost for both flood and drought is 3-5% of GDP versus 1% (mostly Army Corps funds) without global warming.
    0 0
  34. 85, Eric, Okay, so we have anywhere from 3-5% of USA GDP, which in turn is about 25% of the world GDP. The "civilized" world (USA, EU, Japan, China) all account for about 63% of the world GDP. Can we assume that all of those will be affected in roughly the same proportion, so that by ignoring the developing world, climate change will cost, per year, about 3-5% of 63% of the world's current GDP of about 63 trillion US $? That would mean an annual cost, not counting the effects of suffering and lost lives as being priceless, equal to about $1.2 trillion to $2 trillion dollars per year, every year, for fifty or more years, and potentially a whole lot more if it takes that long to clean up the mess, which is assuming that the mess can be cleaned up (that the American Southwest doesn't become a permanent dust bowl, that sequestration technology can draw down atmospheric CO2 levels on what amounts to a Herculean scale, etc.). Do you agree with this appraisal?
    0 0
  35. 85, Eric, Also... does your number include sea level rise? If not, no matter... no need to argue about how great that will be and how fast. We'll just stick with 3%-5%... that's more than enough for our purposes.
    0 0
  36. Sphaerica, I don't know whether the Thai estimate was one-time or ongoing annual, but assuming the magnitude keeps increasing, your estimate sounds reasonable. In that case I would point out that the extra precipitation is a negative feedback, so it has the benefit of limiting warming, see my explanation here. The American SW is already partly a permanent desert. Perhaps Texas will end up in the same condition. The expanding Hadley cell theory is sound, but mainly applies to summer. Texas got a lot of unpredicted rain this past winter when the drought was predicted to continue.
    0 0
  37. No, I don't include sea level rise. If you want to pick a thread for that, I will explain why I don't think I need to include it.
    0 0
  38. A correction to the above, the Army Corps budget for flood control is about 0.1% of US GDP not 1%.
    0 0
  39. But I note Eric, that you seem very keen in past posts that the costs of adaption/geoengineering be paid by those affected by the issues, rather than those who are causing the problem. I would suggest though that this discussion belongs elsewhere. This doesnt seem like a discussion of why sensitivity could be lower.
    0 0
  40. 88, Eric, So... 30% chance of 2.5˚C or more 10% chance of 3˚C or more 40% chance of a cost of $1 trillion to $2 trillion per year for decades to centuries (or more, with higher sensitivity). And your position is that technology is certain to improve and save us from this, so there is no need to take simpler, cheaper, and more conservative action now?
    0 0
  41. Eric, What a wonderful suggestion. All the third world can change their economies into manufacturing from agriculture. Then they can eat the cars they build!! Think through your suggestions. What will people eat after their agriculture fails due to drought? You have been making a lot of these types of suggestions lately.
    0 0
  42. scaddenp, I have owed you a discussion of sensitivity for quite a while now. Here's a link to Scafetta and West 2006 containing estimates of climate sensitivity to solar energy increases. Scafetta in particular has been critiqued here on several threads for proposing terrestrial climate cycles based on solar system cycles which don't appear to be supported. Also critiqued for ascribing too much of the recent warming to a more active sun (both increased TSI and other less well-defined effects like GCR changes). If Scafetta is wrong about the contribution of solar changes to recent warming (i.e. he overestimated the contribution) then his low estimates of climate sensitivity to solar changes in the paper linked above should be even lower. I am assuming that climate sensitivity to solar secular changes (i.e. 100 years or so, not the solar cycle) is the same as climate sensitivity to CO2 forcing changes. If I am wrong, please correct that. If not, then a climate sensitivity of, on the low end, 0.2K per W per m2 leads to a 0.74 C per CO2 doubling. On the high end in that paper, 0.57K/W/m2 leads to 2.1C for a doubling of CO2. Wikipedia says 0.7K/W/m2 for solar and rounds up 3.7 to 4.0 to arrive at 2.8C which they round up to 3C.
    0 0
  43. Eric (skeptic) - Scafetta's work really isn't about climate sensitivity, but about causality; he attributes mid-term climate change to solar/planetary cycles with ill-defined relationships, dismissing the influence of GHG changes and other forcings. The proper threads for discussing Scafetta and West would be Scafetta's Widget Problems or Loehle and Scafetta find a 60 year cycle causing global warming, rather than dragging this thread off-topic.
    0 0
  44. Here's a paper Sea-salt aerosol response ... from the Cornell website. The paper uses a model to compare climate (precip, wind and salt aerosol loading between those four climate regimes. The precipitation shows a steady increase from glacial to preindustrial to present to future with doubled CO2. That represents a negative feedback as the increased latent heat transfer offsets some of the increased GHE. The wind shows a fairly drastic difference from glacial to the other three regimes indicating that measurements such as sensitivity of climate to a forcing change from the glacial to interglacial can't be applied to interglacial to future. Tom in #72 replies that even if we are in a "Goldilocks" climate state, BAU will lift temperatures by 2-3C. My answer is there are not sufficient changes in the climate regime (compared to LGM) to do that. There are large decreases in wind (and consequent dust and other aerosols) that will not be duplicated to any significant extent in the forthcoming change from current to doubled CO2. The precip increases will work against temperature increases. In fig 3b (post 72) the left two squares are red because the starting conditions (LGM) are too different from today's starting conditions and the sensitivity estimate cannot be applied. I am trying to show further that the sensitivity estimate is also an overestimate due to the greater magnitude of climate changes from LGM to present than present to doubled CO2. People may argue that the increased difference is accounted for simply by the increased GAT change, but that change was only 4-5C.
    0 0
  45. Eric, Knutti and Hegerl summarize the climate sensitivity by a variety of methods. The overwhelming consensus is 3C with a long tail upwards. Anything less than 2C is unrealistic. Please produce a link supporting your claim that 1C is a number that is capable of being considered. Here is the key graphic. Image and video hosting by TinyPic Please describe using the graphic how you conclude 1C is reasonable. The lowest high confidence intervals stop at 1.5C and most stop at 2C. There are long high tails. This graph has been posted many times to SKS since you started posting. Can this graph be added to the climate graphics page?
    0 0
  46. Michael, thanks for the challenge. I don't think "1C is reasonable" but it is plausible. As I explained here and previously, the apparent probability density functions are not actually PDFs but model run density functions. The choice of ranges of parameters to mimic natural variation determines the shape of the distribution. So I don't think it is reasonable to point to the center of "model density distribution functions" and claim that 3C is reasonable. The fundamental problem is sub-grid scale physical parameterizations which determine the density function, not reality or even simulated reality. More apropos to my previous post is that the paleo bars do not have any density function because there is really no way to know the distribution due to a lot of measurement error, localized measurements (ice cores) rather than global, changes in geography with large unknown effects, etc. One way to deal with some of the unknowns is by using models (e.g. feeding changed geography into the models). The full chart from Knutti and Hegerl is reproduced in post 72 above. It shows that the paleo estimates do not have a similar base state (far left red square in 3b) and do not have similar feedbacks and timescales (next red square). Those two red squares make the model estimates inapplicable to today's climate. To make them applicable, one must remove the uncertainties due to difficult-to-measure feedbacks. The result, at least in K&H08, is that the estimates of sensitivity from paleo data do not have a distribution, they could be uniform or more likely skewed left since other feedbacks also amplified the warming from glacial to current interglacial along with CO2. Today those other feedbacks are missing (e.g. dust, large weather changes, etc) As I said in the other thread there may be new positive feedbacks that were not in play in the glacial transition so that adds uncertainty on the high side.
    0 0
  47. Eric, Your position is utter nonsense. First... you don't see error bars? Second... all you see is models? Third... so if the range is 1C to 5C, and anything from 2C up is dangerous (the higher the more dangerous), and it is possible, based on current events, that 1.5C is dangerous... ...then tell me again how you wind up at "let's be patient, because climate sensitivity is probably 1C"? Tell me again, too, how you can look at the K&H chart, with a marginal probability of a climate sensitivity below 2C, and see 1C as a shining beacon? Oh, that's right, they are "not actually PDFs" and you can't seem to see the bars and lines and words like "likely" and "very likely." And they're all models... including the "Instrumental Period" and the "Current Mean Climate State" and the "Last Millenium" and the "Volcanic Eruptions" and the "Last Glacial Maximum" and "Millions of Years Ago" and...
    0 0
  48. Eric (skeptic) The "model run density functions" are probability density functions, and have a perfectly reasonable interpretation within a Bayesian framework. "The choice of ranges of parameters to mimic natural variation determines the shape of the distribution." Well of course they do, the "model run density function" is an indication of the relative plausibilities of different outcomes, given what we know, so naturally this would involve using realistic parameters rather than unrealistic ones!
    0 0
  49. Sphaerica, here are no error bars in the depiction above. All the depictions with a model run density functions are from models, the others are volcanoes and paleo-derived. Both have forcings other than CO2 and may not be applicable. The category of "very likely" (90%) encompasses the range from 1C to well over 5C in most cases, so it doesn't help much. The weaker "likely" category is mostly 2C and above. As for dangerous, I noted here that stronger storms are cooling (although I made a mistake, subsidence is cooling on average, not warming, so if there are fewer but stronger hurricanes on average, that means negative feedback). I don't think current sea level rise is dangerous other than model projections. Dikran, how can we say that a critical parameter (say convection) is reasonable or not? A sensitivity of 2C might be realistic, given a number of simplifying assumptions, see here But the required calculations from physics or empirical data to model parameter values are difficult to determine in any case.
    0 0
  50. Eric (skeptic) You are merely changing the subject. In a Bayesian analysis you start of with a prior state of knowledge (the ranges of the parameters) which you then combine with observations (in this case the model runs) which gives you the posterior state of knowledge. Thus the range of parameters is merely the prior state of knowledge, so it is hardly surprising that it is included in the analysis. The reason for specifying a plausible range is exactly because we don't know the correct value with high certainty. P.S. You might also want to investigate the purpose of "perturbed physics experiments".
    0 0

Prev  1  2  3  Next

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