Arguments
Sense and climate sensitivity – more evidence we're in for a hot future
Posted on 12 May 2014 by , dana1981
Arguments that the climate is relatively insensitive to the increased greenhouse effect have become the last best chance for climate contrarians, but a new study from Texas A&M University hammers a big nail in the coffin of that argument.
With overwhelming evidence that humans are the main cause of global warming, and with arguments that modest warming is beneficial falling apart, those who oppose climate solutions are forced to put their eggs in the 'low climate sensitivity' basket. If the Earth's climate is less sensitive to the increased greenhouse effect than most of the available evidence indicates, then perhaps there's not quite as much urgency to tackle the threat of global warming, they argue.
The IPCC shifts on climate sensitivity
The latest IPCC report gave contrarians a glimmer of hope in this area. The fourth IPCC report in 2007 estimated that the planet will warm between 2 and 4.5°C warming in response to a doubling of the amount of CO2 in the atmosphere, with a best estimate of 3°C. At the time, all of the different approaches to estimate climate sensitivity (using historical data, climate models, and recent instrumental data) were in good agreement.
However, in recent years, a few studies using a combination of instrumental data and simple climate models have arrived at lower estimates. This caused the IPCC in its fifth report in 2014 to reduce the lower end of its estimated climate sensitivity range, back to 1.5 to 4.5°C in response to doubled CO2 (the same range as its reports estimated prior to 2007). Due to the newfound disagreement between the different methods, the IPCC also dropped its best estimate altogether.
The 'instrumental' climate sensitivity estimates were something of a puzzle. Why were they no longer in agreement with estimates from historical climate changes and climate models, which still consistently arrived at climate sensitivity estimates between 2 and 4.5°C? John Abraham reported on a study by Drew Shindell at NASA that may have hit on the answer.
Shindell suggests the source of the discrepancy
These 'instrumental' estimates were assuming that the Earth's climate is equally sensitive to all external temperature influences. However, while greenhouse gases are well mixed throughout the atmosphere, aerosols and ozone are concentrated more in the northern hemisphere. Here the temperature response to energy imbalances is more sensitive, as world-renowned climate scientist James Hansen noted in a 1997 paper,
“A forcing at high latitudes yields a larger response than a forcing at low latitudes. This is expected because of the sea ice feedback at high latitudes and the more stable lapse rate at high latitudes”
Shindell found that correcting for this faulty assumption brought the 'instrumental' climate sensitivity estimates much higher in the models he looked at.
Kummer & Dessler build on Shindell's results
In their study, Kummer & Dessler took Shindell's approach one step further. Dessler told me,
"I view my paper as a follow-on to Shindell's paper. What he showed in his paper was that climate models respond more strongly to forcing from aerosols and ozone. What we show in our paper is that if we take his result, and re-analyze the 20th-century observational record then we get a higher climate sensitivity than [studies] which assumed that all forcing was equally effective. Taking efficacy into account, our climate sensitivity is right in the middle of the values derived from other sources. So this allows us to bridge the gap between the various estimates of climate sensitivity and converge on a value around 3°C."
Dessler also explains the study's approach and results in the following 4-minute video.
In short, Shindell showed that according to models, the climate is significantly more sensitive to changes in aerosols and ozone than greenhouse gases, perhaps by as much as 50%. Kummer & Dessler showed that if the climate is 33% more sensitive to changes in aerosols and ozone, then the 'instrumental' estimates are right in line with those derived from historical climate changes and global climate models, with a best estimate of 3°C warming in response to a doubling of atmospheric CO2.
What does this mean for future global warming?
If we continue on our current business-as-usual path, we're on track for close to two doublings of atmospheric CO2. If climate sensitivity is 3°C for doubled CO2, that gives us a best estimate of about 5°C warming above pre-industrial temperatures by 2100, while even 2°C is considered dangerous.
Right after the Hansen quote, you have: "Shindell found that correcting for this faulty assumption in brought the 'instrumental' climate sensitivity estimates much higher in the models he looked at."
Drop the "in" or add something after it. Also, it would help me, at least, if you added a parenthetical clarification right after 'faulty assumption' to state clearly exactly what the nature of the faulty claim is.
Check the original of the Dessler quote. In the third sentence that starts "What we show our paper is that..." is there an "in" after show? If so, put it back in. If not, put it in [square brackets].
(In general, if you would like another pair of eyes to proof papers before posting, feel free to contact me. I'm not perfect--especially at proofing my own work--but I kind of do this kind of thing for a living.)
Great article, by the way.
Nic Lewis makes the following claim at Bishop Hill:
(Not sure how to link to the exact comment, but you can search for "Nic Lewis".)
I was able to check:
- Kummer and Dessler do cite Table 8.6 of AR5
- Table 8.6 is referenced to 1750 (source)
- Kummer and Dessler state:
It would be useful to get clarification on whether this is a genuine error, and if so how much it matters.
JoeK,
Reading to the end of the comments at Bishop Hill, another commenter states that Shindal's post at Real Climate (linked at Bishop Hill) addresses Nic Lewis complaints. I cannot review the matter, but it seems to me that if Lewis has a real point he would be more effective taking it to Real Climate where people are qualified to review it. Most of the comments at Bishop Hill were of the "Dessler used a model so it must be wrong" type. The OP at Bishop Hill disparages Dressler's paper based on a review of the abstract. It has no substantiative analysis or data.
JoeK @2: the answer may be in the comment stream of your link. 'John L' in that stream points to a similar argument made against Shindell's paper and links to Shindell's Realclimate response. To quote: "Lewis’ claim that the my TCR results are mistaken because they did not account for 1750-1850 aerosol forcing is incorrect because he fails to use consistent time periods for all forcing agents." You can use either reference, 1750 or 1880. As long as the reference is applied equally (to GHG as well as aerosol) the calculation yields the same result.
Really, I don't want to fill up this thread with irrelevancies, and the moderators should feel free to remove this comment if it's just a distraction, but I don't find michael sweet and ubrew12's responses very helpful.
michael sweet refers me to Drew Shindell's Realclimate post, which I was already familiar with. He states
The Shindell paper is not the same as the Kummer and Dessler paper. The Realclimate thread addresses the Shindell paper. This thread is about the Kummer and Dessler paper.
Leaving aside advice for Lewis, what should I do if I want to find an answer? Is sweet saying that Skeptical Science is not an appropriate venue for this question (and that Realclimate is)? It seemed to me that it would be somewhat more appropriate for Skeptical Science but if others agree not then I will take it elsewhere.
sweet continues
which may be true, but it doesn't help me. In fact, that was one reason I didn't persist in trying to get an answer there and came over here.
ubrew12 also refers me to Shindell's post, pointing out that one can use either a 1750 or an 1880 reference, which may be true. But Nic Lewis is claiming that Kummer and Dessler mixed two legimate options in an illegitimate way - using one reference for temperature and another for forcing. An initial check showed a prima facie contradiction in Kummer and Dessler. I thought someone here may be able to offer a quick clarification.
Nice article, though I miss some analysis of the carbon cycle feedback and how much of an effect this will have. You recently wrote about the vegetational response. But there is also the transient uptake by the ocean and of course the melting permafrost issues.
"...that gives us a best estimate of about 5°C warming above pre-industrial temperatures by 2100..." Does this take into account carbon cycle feedback or does it implicitely assume it is zero?
JoeK @2 and @5:
First, Kummer and Shindler are simply wrong in saying the forcing data is referenced to the late 19th century.
Second, here is the forcing data over time from AR5 (Fig 8.18):
And here is the temperature data over time, also from AR5 (Fig 2.14):
It is very clear that there is a large negative forcing over the period 1880-1900, primarily due to volcanoes, and in particular the Krakatoa erruption of 1883. It is also clear that the associated temperature excursion is nowhere near as large in relative terms. That is because of the thermal inertia of the oceans, which prevents temperatures from fully following large but short term forcing excursions.
One immediate consequence of that difference is that the energy imbalance at the TOA durring the period of 1880-1900 was reversed in sign relative to the average over the twentieth century. To estimate climate sensitivity starting from the base period of 1880-1900 you need a reasonably accurate estimate of that energy imbalance, which we do not have. To estimate it from the forcing and temperature data, you would need to know the climate sensitivity, which is what you are trying to estimate in the first place.
Rather than do that, you could take a simple fudge factor that approximates to a 0 W/m^2 forcing at the time by reducing the forcing excursion to match qualitatively the temperature excursion. It is very clear that such a fudge factor would reduce the forcing estimate below the underlying trend estimate, but that it would be far closer to zero than the simple average of the forcing as used by Nic Lewis.
I think there are better ways around this problem, but Kummer and Dessler's "solution", however accidental, is plainly superior to Lewis's solution. I doubt it is 100% accurate but the error may be in favour of a lower climate sensitivity, although it is more likely a small error in favour of a higher climate sensitivity.
It is important to remember that a 1.5 degree C rise of global average above the pre-industrial level is still the threshhold of concern regarding signficant difficult to predict and difficult to adapt to climate change consequences.
The change to a 2 degree target occurred in the 2009 Copenhagen meetings. That was when it was admitted that the 1.5 degree target was no longer achievable because of the lack of reduction of pursuit of benefit from burning fossil fuels, primarily by the most fortunate who refused to participate in leading toward a sustainable better future for humanity (less reliant on the ultimately unsustainable burning of fossil fuels).
The continued deliberate lack of concern by many of the most fortunate for anything other than their personal potential benefit, including their deliberate desire and efforts to discredit the best understanding of what is going on, is likely to make 2.5 degrees the target discussed in 2015, then 3.5 in 2020, then 4.5 if anyone even bothers to care after that.
I meant to end my previous comment with the importance of remaining focused on the scientific best understanding of the temperature increase that should be a concern (1.5 C), especially when referring to the potentially ever increasing "political popularity target".
As global leaders reset their "targets" out of "pragmatic expediency" balancing popular support for unacceptable actions against common well understood decency, there needs to be constant reminders of how much further they are from acceptable impacts of one group of people, the most fortunate at a point in time, on all others, especially on future generations (no group should be able to excuse getting benefit from the creation of negative consequences for others).
Even the IPCC reports regarding mitigation make the fatally flawed claim that discounting future costs is a legitimate methodolgy for evaluating acceptability of a lack of concern about future impacts that another group of people will face. That is deemed excusable because, of course, the economy will always grow, in spite of the fact that the only rational way for the economy to sustainably grow is for that growth to exclusively be in activities that are sustainable and not damaging to the future, meaning any perceived beneficial economic growth through increased burning of fossil fuels is a lie deliberately created by uncaring people.
The following comment was made by BC, and accidentally deleted:
Sorry for any inconvenience.
BC, I don't see how that implication follows from the OP. If you look at the first image in my post @7, you will sea the aerosol forcing has the largest error range. Indeed, from table 8.6 we find that while the mean estimate of aerosol radiation interactions has declined from -0.5 to -0.35 W/m^2 from AR4 toAR5, the uncertainty has risen from a 0.8 W/m^2 ranged (-0.1 to -0.9) to a 1.0 W/m^2 range (0.15 to -0.85) so that overall uncertainty is greater.
The key point on which the OP is based is not the uncertainty of the aerosol forcing, but the fact that the aersol forcing is regionally localized, specificly around areas of significant industrialization (Europe, North Eastern US, China). That, combined with the fact that the NH responds more rapidly to forcings simply by virtue of its large land mass (and consequently less thermal inertia provided by the large oceans of the SH) implies that there will be a larger response to a given aerosol forcing than to an equal GHG forcing, given that GHG forcings are globally homogenous.
Dean@6,
Climate models do take into account that part of carbon cycle feedback that is understood as much that it can be quantified. Specifically, the process of CO2 equilibration between atmosphere and oceans is understood quite well and the rate quantified. CO2 uptake by biosphere due to the fertilasation effect is less known but still quantified from the observations of CO2 mass balance between emissions and Keeling curve sine 1958.
The rate of last of the feedbacks you mentioned: melting permafrost, is poorly understood and not quantified. Therefore it is omitted from the models.
Tom Curtis @7,
Thanks for the explanation.
If I understand you correctly you are suggesting that:
1) Nic Lewis correctly found an error in Kummer and Dessler, namely that treat their forcing as referenced to 1880-1900.
2) Lewis may well have the arithmetic correct when he says that simply subtracting the 1880-1900 average from the forcing series would result in a significantly lower estimate of climate sensitivity.
3) However, because of the large volcanic forcing in the period 1880-1900 this is misleading and it would be better to use a more 'normal' period, where the forcing is closer to the pre-industrial 'average'.
4) Unfortunately, data for average surface temperatures become far less accurate going back before 1880, so the simple fix of plugging in the average surface temperature in 1750 is not available.
5) Tom Curtis suggests that if such a procedure could be carried out (i.e. we had good surface temeprature for 1750) then the result would be closer to Kummer and Dessler's 'incorrect' published result than to Lewis' 'corrected' result (by luck for Kummer and Dessler).
6) Eyeballing graphs, Tom Curtis suggests 'the error may be in favour of a lower climate sensitivity, although it is more likely a small error in favour of a higher climate sensitivity.'
7) Finally, Tom Curtis suggests 'there are better ways around this problem'.
If the above are correct I would draw the conclusions:
1) The method of Kummer and Dessler is promising. It is worth developing and implementing ways around the problem. Perhaps this involves taking an average over a longer period, such as 1850-1900. On the other hand there are problems with this, too. 1800-1850 looks like the forcing would be even more negative. The result should not depend on the arbitrary period chosen. I can think of other approaches but they may be more complicated, e.g. depend on GCMs, which has other drawbacks.
2) The results of Kummer and Dessler should not be quantitatively relied on. A correction should be issued addressing these issues.
I don't like the F scale presented in the video (evidently targetted at US audience), because the video does very poor job translating from/to C scale. It portrays ESC from models of efficacy 1 as 4F, wrongly equaling it to 2degC, where in fact it equals 2.2degC (according to the formula: dF/dC = 9/5). Then, the efficacy 1.33 yields 6F, which is again wrongly equaled to 3degC, when in fact it is 3.3degC. My nitpick is justified here, because in this range of ECS numbers, the inaccuaracies in the video make a difference.
Best of all (in the interest of KISS paradigm that I advocate everywhere) the US audience should learn that in climate science the golden rule is to measure temp in degC (or K) only and forget about F that confuse the picture. It's not hard to accept degC because the temps we are talking about (global averages) are somewhat abstract entities, different to those we are feeling and seing on TV forecasts. It seems that those who accept the science have no problem accepting different scale. So, accordingly, let's bring the numbers from the acual abstract (forgetting the last one minute of the video):
JoeK @13, your itemized points fairly represent my claims. With regard to your points in response, I largely agree with your first point. With regard to the second point, first, while the error regarding the forcing baseline is a problem it is almost certainly less of a problem than those that have afflicted other similar attempts using this method, including those by Nic Lewis. In particular, the use of integrated values to estimate climate sensitivity is a clear improvement in method over anything before. Further, no prior attempt using the general method has adequately determined the initial energy imbalance, and have generally used energy imbalances which on the face of it are the wrong sign. If we are to disregard Kummer and Dessler's result, therefore, we should disregard also all results from equivalent methods. Alternatively and more sensibly we should recognize it as one attempt to measure climate sensitivity among many and not assume that either it or its predecessors provide a perfectly accurate assessment.
With regard to errata, it is not yet clear that this is a significant error. It may be that the result is accurate or nearly so despite the mistatement about forcings, in which case an erratum or corrigendum need not be warranted. It may be appropriate to merely acknowledge the error in blogs or by some other means. Of course, if the error does significantly alter the estimate of climate sensitivity, then a corrigendum should be issued.
Tom @15,
With regard to your point about relying on Kummer and Dessler I think you somewhat miss the significance of their paper. If it were just one more estimate of climate sensitivity it would not be such a big deal. Importantly, they claim that their estimate is conditional on an estimate for efficacy (which may be a bigger source of uncertainty than points so far raised here).
To me the significance of Kummer and Dessler is to take Shindell's estimate of aerosol efficacy and ask what implications it would have for climate sensitvity? The important point is setting some scale to the relation between efficacy and sensitivity. If, for example, they had found that you need efficacies in the range of 2-3 before a big effect on sensitivity shows up then Shindell would have an interesting observation, but of limited relevance to the question of climate sensitivity. As it happens they find a problem in the other direction, that in their simple model Shindell's central estimate produces a rather high sensitivity to be realistic.
I think Kummer and Dessler's concluding paragraph highlights this.
(my emphasis)
Their final sentence is also appropriate. They do not end by concluding they have a better estimate of sensitivity, but rather that
Maybe a chart of modal estimated ECS vs. efficacy, and the slope of that dependence would be of interest? While you may be able to eyeball an estimate of the effect of the error on a point estimate of ECS, can you do the same for the ECS vs. efficacy curve?
As for a correction, it seems to me that this should not be such a big deal. We are in the 21st century now, with electronic publishing. If there's a simple error it should be noted for future readers who may come to the document (possibly in future years) without having to search the internet or blogs for errata to every paper. The fact that by co-incidence the numerical result may happen to be close to correct doesn't seem an important consideration. Developing methods correctly is what will allow this work to be built on in the long term.
Tom,
You state that "Kummer and Shindler are simply wrong in saying the forcing data is referenced to the late 19th century"
Shindlell says:
"The second substantive point Lewis raised relates to the time period over which the TCR is evaluated. The IPCC emphasizes forcing estimates relative to 1750 since most of the important anthropogenic impacts are thought to have been small at that time (biomass burning may be an exception, but appears to have a relatively small net forcing). Surface temperature observations become sparser going back further in time, however, and the most widely used datasets only go back to 1880 or 1850. Radiative forcing, especially that due to aerosols, is highly uncertain for the period 1750-1850 as there is little modeling and even less data to constrain those models. The AR5 gives a value for 1850 aerosol forcing (relative to 1750) (Annex II, Table AII.1.2) of -0.178 W/m² for direct+indirect (radiation+clouds). There is also a BC snow forcing of 0.014 W/m², for a total of -0.164 W/m². While these estimates are small, they are nonetheless very poorly constrained.
Hence there are two logical choices for an analysis of TCR. One could assume that there was minimal global mean surface temperature change between 1750 and 1850, as some datasets suggest, and compare the 1850-2000 temperature change with the full 1750-2000 forcing estimate, as in my paper and Otto et al. In this case, aerosol forcing over 1750-2000 is used.
Alternatively, one could assume we can estimate forcing during this early period realistically enough to remove if from the longer 1750-2000 estimates, and so compare forcing and response over 1850-2000. In this case, this must be done for all forcings, not just for the aerosols. The well-mixed greenhouse gas forcing in 1850 is 0.213 W/m². Including well-mixed solar and stratospheric water that becomes 0.215 W/m². LU and ozone almost exactly cancel one another. So to adjust from 1750-2000 to 1850-2000 forcings, one must remove 0.215 W/m² and also remove the -0.164 W/m² aerosol forcing, multiplying the latter by it’s impact relative to that of well-mixed greenhouse gases (~1.5) that gives about -0.25 W/m².
If this is done consistently, the denominator of the climate sensitivity calculation containing total forcing barely changes and hence the TCR results are essentially the same (a change of only 0.03°C). Lewis’ claim that the my TCR results are mistaken because they did not account for 1750-1850 aerosol forcing is incorrect because he fails to use consistent time periods for all forcing agents. The results are in fact quite robust to either analysis option provided they are done consistently."
It seems to me that you have oversimplified Shindell's position. Can you offer a citation to your claim that Shindell has not consided the forcings correctly? I do not see a reply to Shindells claims in your claims. Can you refer to Shindell's position above to support your claim that he is "simply wrong"? It appears to me that Dessler agrees with Shindell. Claiming that they are both wrong without strong references is a pretty strong claim.
Michael Sweet @17, I had read what Shindel wrote. It is, however, largely irrelevant because:
1) We are discussing Kummer and Dessler (2014), not Shindel et al (2010); and (more importantly,
2) Kummer and Dessler use 1880-1900 as there baseline period, not 1850.
Because of the later point, it is not possible to straightforwardly apply Shindel's comments to Kummer and Dessler. What is true of 1850 may well not be true of 1880-1900.
Further, the discussion of failure to use all forcings by Nic Lewis in criticizing to Shindell et al probably is not relevant to Lewis's superficially similar criticism of Kummer and Dessler. In fact Lewis does not show his working, so he may only have modified for the average anthropogenic forcing for 1880-1900. From the wording of his comment at Bishop Hill, however, it is far more likely that he has used the average of the net forcings (anthropogenic plus natural).
Finally, in a way I have argued that the first of Shindell's ways of resolving the issue for 1850 also applies for 1880-1900. It will almost certainly result in a small error of indeterminate sign, which is no criticism at all in science. I have, however, argued that his second method cannot be used in 1880-1900 because of the distorting effect of the volcanic erruptions, and in particular Krakatoa. That is an almost negligible consideration in 1850, where the closest preceding volcano was more than ten years prior and hence of negligible remaining effect.
An addendum for my prior post:
Kummer and Dessler state of the forcing data:
However, the IPCC AR5 WG1 table 8.6 gives values for various IPCC reports including:
It's caption reads:
That clearly proves that Kummer and Dessler were in error in claiming that the forcings tabulated on Table 8.6 were "...referenced to the late-19th century". It is about that point only that I have said that they are wrong.
Dessler has answered Lewis' critique at http://www.climatedialogue.org/climate-sensitivity-and-transient-climate-response/#comment-924
It seems like it was a technical mistake but that the effect was that the revised sensitivity became slightly higher.
[PS] Fixed link
Dessler says that the mistake is similar to Tom's description with a small increase in the calculated climate sensitivity (as Tom suggested). Apparently the paper has not yet been published so changes are being made in the gally proofs. Dessler states the conclusion that aerosol sensitivity needs to be addressed is unaffected.
Good job Tom!
Seconded. I only hope I'm never on the wrong side of a debate with Tom. He's downright scary...in a good way!
Stephen@22,
I assume you're joking. On a serious note, being on the "wrong side of a debate" (with Tom or anyone else) can be a rewarding experience if you're open minded and want to learn. If you're afraid of a rational debate, you never learn anything. The trick is to know what is logical & rational and follow rational argumentation while rebutting the irrational. If you're wrong, ratinal arguments eventually teach you to change your mind.
Michael Sweet, Stephen Baines, thankyou for the compliments. As, however, I now find myself disagreeing with Dessler, they may not be deserved.
Dessler states that "...radiative forcing is +0.3 W/m2 in the late 19th century", which is a fair estimate for the 1880-1900 value if you ignore the effects of volcanism. Including the effects of volcanism makes the value distinctly negative, however. More importantly, it reduced the global mean surface temperature. Consequently simply using the underlying forcing would overestimate climate sensitivity (unless you had an appropriate compensation for the 1880-1900 energy balance, which Kummer and Dessler does not).
Tom, critical appraisal of papers is always very welcome. That is how science progresses. Its the disinformation and spurious/medacious critiques that are unwelcome and which the site exists to debunk. I would say worth getting clarification from Dessler.
(I cant imagine there is anyone who actually desires AGW to be real, but some prefer not to live in a fantasy either).
Klapper@26,
I agree with your general point. Kummer 2014 can be used as a rebuttal to the claim "climate sensitivity from 20th century observations is low", because they showed with very simple math that observational estimates should not assume human aerosol efficacy of 1. We can be quite confident that said efficacy is > 1, therefore the feedback factor lambda becomes smaller, yielding ECS values higher than claimed before. But the uncertainties are so high (if you look at the last column of their Table 1) that talking about a meaningful value of ECS thus obtained is pointless. I do much prefer the methods that accually measure (via proxies in paleo methods) the dT in equilibrium, according to the definition of the term.
@Chriskoz #26:
I see my post #26 is missing in action. Your 2nd sentence is only valid if you believe that the net effect of anthropogenic aerosols is negative. Tom's post #11 indicates the range of AR5 estimates for aerosols forcing extends into the positive. As I pointed out in my original post #26, the fact that China has experienced both a faster than global SAT warming rate and extremely high local aerosol forcing is reason to question the "anthropogenic aerosols are a negative forcing" assumption.
[PS] Not moderating here, but you need to back assertions with evidence/references and skip the pointless remarks (eg pink unicorns) if dont want to be moderated for sloganeering. Try again with backing data.
Klapper @27, China is a big place. It's air pollution problems are not uniform, and neither are its temperature trends. In fact, the temperature trends are highest in the unindustrialized North of the country:
In contrast, the air pollution (aerosols) are to be found predominantly in the South East of the country, and around Beijing:
The correlation between lower temperature trends and higher aerosol polution in China is noticable (though not perfect). If you want to mount the argument that China's temperature trends prove aerosol pollution is a positive forcing, you are going to need a far more detailed analysis than merely mentioning national average trends.
Klapper@27,
In addition to Tom Curtis' response, please note that China is not isolated from the rest of the world. The heat (or negative change in heat in this case resulting from scatering of incoming shortwave radiation) is transported by the atmospheric circulation and can result in warming in other place. Therefore, before you start mounting nitpicky claims from the comparison of Tom's images, e.g. "Beijing area, where the pollution levels are the highest, is actually one of the warmest", you need to at least find the GCM models confirm the Cinese aerosol forcings indeed influence local weather, for your claim to be meaningful.
An additional point to add wrt pollution in China. A significant part of the 'pollution' around Beijing is dust. Beijing and its surrounds are very prone to dust storms coming in from the Gobi desert.
Further south, towards Nanjing & Shanghai you start to get interactions between pollution and humidity. Lots of 'sfog' - mists and fogs where pollution provides condensation nuclei for water vapour.
@Glen Tamblyn #30:
I looked at ozone as a proxy for aerial pollution and Bejing and the area west of the Bonsai Sea down to Shanghai seems the highest concentration. NO2 looks to be the same pattern.
The prevailing winds north of 30N would push higher level aerosols to the NW, so towards one of the fastest warming areas of China.
Anyway my calculation on the SAT warming rate of China used a window of 25 to 45N and 105 to 120E so it disincluded the west and NE most parts of China, but did include the major source areas for anthropogenic aerosols. Warming rate within that geographic window is about 0.3C/decade over the last 30 years.
Klapper @31:
1) The image I showed @28 shows the presence and intensity of particulate pollution - not the sources of emission. Adjusting the figure to suite your peculiar views of the prevailing wind direction in China is simply to fudge the data based on prejudice.
2) In any event, the prevailing winds in China, which are to the North East, not the North West, so that even your fudge in based on an error:
Upper tropospheric winds will be opposite in direction to surface winds, ie, blowing to the North East, thereby carrying the pollution over Korea, Japan, and eventually to Alaska.
3) Your latitude and meridionally restricted region leaves out the heavilly polluted southern coast of China, not to mention the heavilly polluted region around Chengdu (which has the lowest temperature increase of any area in China over recent times), but includes a large region of lightly, or unpolluted (and rapidly warming) area north of Beijing, stretching into the Gobi desert and Mongolia. In short, it seems well tailored to misrepresent the relationship between aerosols in and temperature trends in China.
4) You compare the trends found in that area to the global trends which include the low trends over ocean. Warming is much higher over land, so the comparison should be with global land temperature, where the trend is 0.296 +/- 0.08 C/decade (NOAA; 0.288 +/- 0.114 C/decade, BEST) since May 1984. That is, the global land surface temperature trend over the last thirty years is indistinguishable from the trend you find in your cherry picked latitude and meridionally restricted area. Had that area included the region around Chengdu and the south coast of China, and excluded the lightly polluted band north of Beijing, the trend would have been lower than the global land surfact temperature trend.
@Tom Curtis #32:
1. But only at ground level. The forcings occur at all levels of the troposphere. The sources, at least for some elements of anthropogenic aerosols are predonminately in the north of China, not the SE (like coal fired power plants).
2. Sorry my typo. You're correct the prevailing winds are to the NE, which is where the the higher warming rates are.
3. You're confusing me. Let's get back to basics. North of 30N the prevailing winds are to the NE (agreed). The maximum emission areas are the Bejing to Shanghai corridor and slightly SW of that. So the wind blow the upper level pollution to the NE of this area which is coincidently the area of max warming rate for China.
4. OK, land vs. land makes sense. But shouldn't China be cooling or at least warming a lot slower if aerosols are a negative forcing? I checked the BEST dataset for my prescribed geopgraphic window and the warming rate is 0.31 from the Berkeley dataset. Note this is slightly above the global rate, not below, or cooling despite the very high anthropogenic aerosol emissions from China. Chengdu is almost within by capture area and if the prevailing winds are to the NE, then its effects on the warming rate are within my window.
Klapper @33:
1) The map in question is generated from satellite data. Therefore, your claim that it is of surface pollutants only appears highly conjectural. In fact, it looks like a claim asserted to be fact only because your argument collapses if it is false.
Still, if you want to be certain, here is the global spring (MAM) anthropogenic aerosol optical depth averaged over the years 2003-2010:
You will notice:
a) That no red zones(0.4 +) are found north of Liaodong Bay (the northernmost extension of the Yellow Sea), so that the most intensive aerosol effect in China remains south of about 41 degrees north, is well south of your northern limit. Although you cannot see it here, Spring has the most northerly distribution of Chinese aerosols, with no green (0.2->0.3) showing north of Liaodong Bay in Autumn (SON) or Winter (DJF). Summer (JJA) shows more intense aerosol distribution around Beijing, with red intensities across the Yellow Sea, but the northern edge graduates more quickly so that there is less aerosols over north east China than in spring.
b) The area around Chengdu shows as dark red (0.5 +), as it does in all seasons. Chengdu is China's fourth largest city, a major industrial city and 104 degrees east, ie, outside your mapped area. Its pollution effects extend significantly to the west. As it happens, and unlike Beijing, its effect on aerosol optical depth is year round. Beijing's dissipates in autumn and winter bases on satellite observations.
c) The south China coast, which you exclude, shows red intensity. It shows red intensity also in autumn, yellow intensity in winter, but only green in summer. It is thus never has less aerosol concentration than north east China, and for much of the year has as intense an aerosol concentration as Beijing.
All in all, your claims about aerosol concentrations are not born out by the observations, and you have included areas in your field with consistently low anthropogenic contribution to aerosol optical depth, and excluded areas with consistently high anthropogenic contribution to aerosol optical depth.
2) The upper atmosphere winds are more or less the reverse of those in the lower atomsphere. Looking at the map you can see that means polution in China from Hong Kong north to Beijing will be carried primarilly east across Korea and Japan, with only a slight northerly set. Even in spring, with the most extensive northerly component of the aerosol drift, aerosols are largely confined south of a line from Beijing to the north tip of Sakhalin Island. Thus confined, it means only about a third of north east China (close to the border with North Korea) has a significant anthropogenic aerosol effect at any time of year, and for half of the year the effect is negligible.
3) Thoroughly dealt with above and below. Your assertions are, on the information available, simply wrong.
4) I sampled three 5 degree by 5 degree areas, using NCDC land only temperature data from the KNMI explorer. They were:
North West China (80-85 E, 45-50 N) 0.47 +/- 0.18 C/decade
Beijing (115-120 E, 35-40 N) 0.31 +/- 0.09 C/decade
North of Liaodong Bay (120-125 E, 40-45 N) 0.17 +/- 0.16 C/decade.
The strong warming is in the North West, which is not impacted by anthropogenic aerosols. In those areas impacted by anthropogenic aerosols, the warming is much less. Further, the your claim that the norht east is were the strongest warming is to be found is distinctly dubious.
You might reasonably claim the to north easter samples are also coastal, and so should have lower trends. They also may be more strongly effected by ENSO or the PDO. Those, however, are complexities you have simply ignored in your analysis. If you wish to apply them in defense of your claims, you must apply them when making the initial analysis as well, something you have completely failed to do.
Data on aerosols is from Bellouin et al, "Estimates of Aerosol Radiative Forcing from the MACC Re-analyis"
@Tom Curtis #34:
I think calculating warming rates by 5 degree lat/long boxes is too small. Chriskoz pointed out in #29 that temperature response to forcing is mixed well outside the forcing subareas in this case the Bejing/Shanghai aerosol zone. In any case my zone of 105 to 120, 25 to 45 degrees could be refined to 22.5 to 42.5 and 102.5 to 120 if I can find grid compilers with that resolution. Maybe I will try the RSS or UAH TLT datasets when I get home.
You are right that I ignored some complexities but I got the conversation started. You have addressed a lot of the detail, but my main point remains: there is no evidence for cooling from anthropogenic aerosols in China over the last 30 years.
Klapper - I'm really scratching my head over your last response. Tom Curtis has documented that the larger region warming in China is not correlated with aerosols, rather the contrary (actually less warming in high aerosol areas), and pointed out significant errors in your argument including spatial distribution and erroneous winds. The majority of your assertions and arguments in this regard have been demonstrated to be incorrect, that the data indicates aerosols are in fact a cooling factor.
And yet after all that you end by asserting that your original statement regarding aerosols was correct? I'm sorry to say so, but it appears to be simply an argument by assertion at this point.
Klapper @35, first, Chriskoz reffered to the effects of air bodies that are warmed, or cooled at one location being carried downwind thereby warming or cooling a location further downwind. As we are talking about the surface (2 meter) air temperature, it is surface winds we are talking about for this effect. Downwind for surface winds is West-SouthWest from the line from Beijing to Shanghai, not NorthWest (or indeed North-NorthWest) of it as required by your theory.
Second, you claimed @27:
That is a significantly stronger claim than the "no evidence" claim you now indicate was your main point. Further, from the evidence I have shown above, even the "no evidence" claim looks dubious. The strongest warming in China is north of the line in which there have been a substantial increase in anthropogenic aerosols. The weakest warming has been downwind at the surface from the area of major emissions from Beijing down to Shanghai, and around, and downwind of Chengdu, another major area of aerosol emissions. Qualitatively, that is what we would expect of aerosols were a significant negative forcing.
Further, there are good physical reasons to think sulfates, at least are a negative forcing - reasons confirmed by the impact of stratospheric aerosols from volcanoes. An argument to the contrary from Chinese data would need to be very convincing to overturn that data, not to mention data from the US and Europe that indicates the aerosol forcing is negative.
Having said that, there are complexities. The relative composition of emitted aerosols (BC vs NOX vs sulfates) probably varies regionally and by time and will have an impact on the overall forcing. Latitude, closeness to oceans, and ocean fluctuations are also complicating factors. Nor has it escaped my attention that the northern band of strong warming in China more or less coincides with the region where the easterly trade winds (ie, blowing to the west) are replaced by the Westerlies (blowing to the east), and hence represents the area where local ocean influence is minimized and where Chinese temperatures are most impacted by influences over all of Eurasia.
Given these complexities, I seriously doubt a blog scientist will have the resources to disentangle all the issues. They certainly will not if they use a single block grid to represent Chinese temperatures.
@Tom Curtis #37:
I disagree with your surface winds argument. Aerosols don't just hang close to the ground and their forcing is not just at ground level, it's all the way up the atmospheric column as volcanic episodes show. Your AOD map in post 34 shows dispersion to the NE, out over the Pacific, which is the direction of the prevailing winds on the maps I found and which you noted in your post 32.
Your PM2.5 map in post 28 shows the aerosol maximums in the Bejing/Shanghai corridor. I think you would agree to this? Also in post 28 the warming map for China in this same area shows "red", in the range of 0.4 to 0.6C, although I'm not sure over what time period this is.
Looking at the Bejing/Shanghai zone, it can be boxed by 110 to 122.5 and 27.5 to 42.5. If you take a land only dataset (I choose CRUTEM4.2 since the grid is 2.5 degrees not 5), the warming rate in this area is 0.30C/decade for the last 30 years. For the global land CRUTEM4 shows 0.28C/decade.
So the most aerosol polluted geographic area on the planet is warming slightly faster than the globe. Again I ask the question: how do you reconcile this basic fact with the interpretation that anthropogenic aerosols are a negative (cooling) forcing?
You are right that
Klapper @38:
First, you are confusing two issues. First, wind disperses aerosols from their source of emissions. That means the area over which the forcing applies does not exactly coincide with the area of emission. In China, the wind initially disperses aerosols to the WSW. but as the aerosols gain in altitude that reverses and they are dispersed to the ENE. These directions vary with season, as does the distance aerosols typically travel before exiting the atmosphere. In China, that means the dispersal is to the WSW for two seasons of the year, and largely static with a thin tail to the ENE for the other two seasons.
In addition to this effect, air that is warmed (or cooled) by a given regional forcing is then carried downwind. There it mixes with cooler (or warmer) air. Cooler (or warmer) air upwind of the regional forcing will also be carried down into hte area of the forcing. The consequence is that the temperature effect of the forcing is more dispersed than the forcing. That is, you have two stages of dispersion. In the first, the aerosol becomes more dispersed than its source; and in the second stage the temperature effect of the aerosol becomes more dispersed (and diffuse) than its source.
Chriskoz (@29) was referring to the later effect. That effect is purely a funtion of surface winds because it is the surface air temperature we are discussing. The more complicated case with the dispersion of aerosols has already been discussed above, and is a seperate issue. Your confusing of the issue looks like your trying to have two bites of the cherry. Having been comprehensively refuted on aerosol dispersion above, you appear to argue aerosol dispersion must be greater than observed, and more favourable to your case than is observed because of some effect of wind which is in addition to the observed dispersion of aerosols.
Second:
Seriously?
After all the discussion above, your just going to exclude Chengdu (104 E) from consideration, even though it is the only area to have red or greater aerosol concentrations year round in China; and restrict yourself to the Peking/Shanghai corridor despite the fact that for half the year it shows yellow or less aerosol concentration; and just totally ignore the temperature dispersion from prevailing winds even though you (following Chriskoz) brought it up?
Your new region of choice is a blatant cherry pick that simply ignores the prior discussion. Having been shown wrong on almost every point you simply attempt to start the argument again showing that you have learnt nothing, and that it is a waste of time continuing the discussion with you.
@Tom Curtis #39:
The most viewable map you've posted (PM2.5) shows the Bejing/Shanghai corridor to be the most polluted. I also looked at other air pollution proxies (ozone, NO2) and they showed the same thing.
Anyway here is a link which shows 2001 to 2010 PM2.5 on a grid which you can zoom in so it's somewhat better than the map you posted. The Bejing/Shanghai area is just as intense as the Chengdu area, but larger in areal extent.
http://sedac.ciesin.columbia.edu/data/set/sdei-global-annual-avg-pm2-5-2001-2010
If you are convinced that Chengdu is more representative than Bejing/Shanghai then do your own trend analysis of the Chengdu area.
@Tom Curtis #39:
I checked a box around Chengdu (25 to 35N, 100 to 110 E) and found that the warming rate in CRUTEM4.2 for the last 30 years to be 0.31C/decade, so Chengdu has a similar warming rate to the Bejing/Shanghai area.
Klapper@40,41
Despite my assertion @29 not to consider climate change in any chinese locality in isolation from the rest of the globe, you did precisely what I warned against (i.e. you "boxed" a convenient cherry area, where temperature trend supports your preconceived claim), therefore you've implicitly denied the validity of my assertion.
Maybe you should open your mind at this point, at the fact that climate change is not only about CO2 and aerosols forcings and cherries. Climate change is about variety of phenomenons, including but not limitted to athmospheric and ocean fluid dynamics, water vapour transport via winds and convection and associated rainfall patterns, etc.
Let's consider for example changes in rainfall over China in last 50y from the widely cited article: (Xu 2001). Xu found that due to the changes in heat equilibrium of the land surface (due to SO2 pollution), the summer monsoon belt has moved southward. How much? Check Xu's figure 4: Average latitudes of the central axis of monsoon rain belt in summer and mid-summer. I'm eyeballing that it moved by some 5 degrees since 1960, leading to an abnormal summer climate pattern of ‘‘north [your cherry Beijing] drought with south flooding’’ It does not take much imagination to conclude that in such situation of signifficant shift in precipitations, the temperature response will also vary. Specifically, we can expect the increase in temperature for the areas that became drier, because the rain cools things down in summer. And that's what Beijing may be experiencing: drier summers due to retreat of monsoon southwards.
This is only one example that I'm giving you, without even trying to quantify the issue because I don't have time and expertise for it. This is just to show that your method of "boxing" and isolating a cherry temperature trend is not how you should aproach the complex problem of climate change in China.
@Chriskoz #41:
So far I've looked at the temperature trends in 3 different boxes, the first from NCDC, the second and third from CRUTEM4. The last box covered the Chengdu area based on input from Tom Curtis. All were above the CRUTEM4 global average for land temperature. I haven't found any evidence from these data that anthropogenic aerosols have a cooling effect on SAT.
It's possible that the cooling effect of aerosols is non-linear, i.e. cooling at lower intensities but less so at higher intensities, or that the Chinese mix is richer in soot vs sulphate, or that 30 year trends are too long to capture the recent Chinese surge in aerosols. On the latter point, regional SAT is very noisy and trends shorter than 30 years probably wouldn't be meaningful statistically.
All that being said, data from China over the last 30 years don't appear to support the hypothesis that anthropogenic aerosols represent a negative SAT forcing. That brings us back to the question at hand, whether increasing the leverage of anthropogenic aerosols to recalculate climate sensitivity is appropriate or not. If anything, data from the last 30 years in China would indicate that decreasing the leverage of anthropogenic aerosols to recalculate climate sensitivity should be the next experiment.
Klapper@43,
You're still denying to acknowledge my assertion when supported by my citing of (Xu 2001) because you did not even bother to acknowledge/address my argument. And you're continuing to cherry pick the "boxes" confirming your preconceived idea. That falls IMO into the "excessive repetition" category of policy violations, therefore I would not be surprised if your comment was deleted (and my response herein can also be deleted). All I can conclude (after Tom@38) is that it is waste of time continuing the discussion with you
@Chriskoz #44:
I think what you are saying is that there are confounding factors to the cooling forcing of anthropogenic aerosols. You gave the example of changes in precipitation in turn causing local warming in Beijing, undoing the normal cooling effect of aerosols.
What about Chengdu then. My analysis of a 10 x 10 degree box around Chengdu shows the same warming rate over 30 years as the global rate for land. Is this area also effected by the same drying by sulphates as Being?
If there is a sulphates drying effect then perhaps this effect is also occurred historically in other areas meaning our current estimates for net aerosol forcing are too high.