Bishop Hill El Reg on CLOUD
Aug 25, 2011 Climate: Models Climate: solar Andrew Orlowski has noticed a good quote in the CERN press release about CLOUD (emphasis added):
What has CLOUD discovered and why is it important for our understanding of climate? There are several important discoveries from CLOUD. Firstly, we have shown that the most likely nucleating vapours, sulphuric acid and ammonia, cannot account for nucleation that is observed in the lower atmosphere. The nucleation observed in the chamber occurs at only one‐tenth to one‐thousandth of the rate observed in the lower atmosphere. Based on the first results from CLOUD, it is clear that the treatment of aerosol formation in climate models will need to be substantially revised, since all models assume that nucleation is caused by these vapours and water alone. It is now urgent to identify the additional nucleating vapours, and whether their sources are mainly natural or from human activities.
I am slightly confused about this though - are we saying that the models include a factor for nucleation that is equal to the rate of nucleation currently observed, and which changes based on how we think sulphuric acid and ammonia levels in the atmosphere will change in future? Or are we saying that the level of nucleation in the models is 10--1000 times too small? I assume the former, but I had also believed that the models went back to first physical principles rather than using empirical measures.
Maybe somebody can put me right here?
Bishop Hill
David Whitehouse has more.
Reader Comments (61)
I too am confused. But the conclusion seems to be that existing models have been constructed with wrong assumptions.
ie, the climate change models are worse than we thought
BISHOP:
Models are miles away of using only physical laws, that would be practically impossible anyway. They require hundreds of parameterizations, which are empirical simplification of how a real process works. Some of these parameterizations are quite solid, some others are quite questionable, like the role of aerosols, specially (NH4)2SO4 . As the cloud experiment reveals current parameters are nowhere near reality, and that needs to be reconsidered. So your first interpretation is correct.
"I had also believed that the models went back to first physical principles"
You'd think so, wouldn't you? Perhaps a modeller can tell us...
Help yourself: GCM cloud parameterization
"Parameterization"?
Fiddle-factor adjustment, if you don't mind.
Patagon
Not quite understanding you.YOu say cloud parameterisations are nowhere near reality, but then say that my first interpretation is correct. But this was the one where I said that the parameterisation is correct.
This should give us all pause for thought. Richard Betts is surely right on a previous thread that we shouldn't try to apply such important findings at once to the 'big problem' of AGW and expect to get the answers right. What this really shows is another massive hole in the GCMs - or rather, that a hole many suspected was there has now been shown beyond doubt.
I'm sure the search will now be on to find nucleating vapours from 'human activities'. And so indeed it should. And equally from natural causes (and let's not kid ourselves that it's easy to separate the two in such a system). But science should be done better than the ignoring of Svensmark by Kirkby and the mainstream media. It's a red-letter day from him and Calder. Many congratulations.
"are we saying that the models include a factor for nucleation that is equal to the rate of nucleation currently observed, and which changes based on how we think sulphuric acid and ammonia levels in the atmosphere will change in future?"
both things are correct
The parameter makes: (clouds) ~ f(aerosols), f accounts for a function, which introduces a parameter that multiplies current observations of aerosols to make it fit current observations of clouds.
The problem is that we do not have a range wide enough to be sure if that parameter is working properly, especially when we extrapolate outside current observed conditions. It means that a parameter has been introduced to fit observations of clouds to observations of aerosols, but it is not sure whether the relationship is correct. That parameter may make the observations correct as long as the observations don't change too much
should say "That parameter may make the modelled results correct as long as the observations don't change too much"
I think today would be a good day to release Mann's e-mails in the hope that no one notices.
It would also be interesting to see how the Wikipedia entry for Svensmark is updated. I last looked about a year ago, and there were scathing, comments from William M Connelley
FWIW my reading of it is that the conclusion is that there is a large missing piece of the nucleation story. The models appear to have been using "observed nucleation rates" which had been assumed due to apparently quantified effects of SO2 and NH3, however these experiments do not corroborate the size of the observed effects. The paper posits that cosmic rays may be responsible for (some of?) the additional observed nucleation. I presume the success or failure of the search for the "additional nucleating vapours" will help support or undermine the case for cosmic ray effects according to what is found.
wrt the models - I recall informed comments from blog commentators that aerosol effects are used as a tuning factor to calibrate models to historical temps. If that is true, and this in turn is used to deduce "observed" nucleation effects, one implication is that the offset warming effect that an elevated nucleation rate would be mitigating is being overstated.
I may be way off here so it would be good to hear from others more knowledgeable - have just seen patagon's comments and reference. Thank you
Patagon
Got you. Thanks.
The David Whitehouse summary is excellent.
Many people in the nuclear industry tried for many years to develop models that could model simple closed sytem thermal-hydraulics problems from first principles. It couldn't be done. Every time they tried to model an experiment with a novel characteristsic they got the wrong results. Back to correlations and parameterisations (but well-validated against experiments).
I understood this to refer to the case when the chamber was isolated from GCRs (the initial part of the graph Nigel Calder reproduced) - i.e. UV alone cannot account for the amount of cloud we actually see, so the models are wrong by a large factor.
Add in the GCRs and you get the middle part of the graph; add in pretend GCRs from CERN and you get the large rise at the end of the graph.
What the CLOUD experiment has shown is that uncertainty over the role of clouds in the AGW hypothesis is much, much greater than previously thought and that once accounted for, e.g. using Svensmark's ideas, may well significantly diminish the projected impact of AGW in changing the climate.
We now have two competing theories - which one has had the greater impact on climate?
This is all a whole new ball game. Science is beginning to prevail over dogma.
my sense is that there are too many first principles that interact in strange, sometimes still unknown ways for there to be much chance of modelling an answer from first principles. My experience of modelling from first principles is that you dump it as fast as you can and revert to empirical results.
In producing business cases for new telecomms products, for example, the simple principles soon dictate that you require the population of the UK to be handling the call centre.
The main thing this paper tells us is that the science is not settled.
My son is a world class meteorologist in the private sector. He told me years ago that Svensmark's Theory should have been the most exciting theory in the modern history of meteorology because it provided a plausible hypothesis to explain many of the major climate changes in the entire history of the Earth ... except that it conflicted with AGW politics and was therefore to be labeled as heresy. I read Svensmark's book years ago; if you haven't read it ... your loss.
I think the key thing here is to ask the question: "what is the controlling factor in cloud formation".
The press release is trying to suggest that organic vapours at totally ridiculous concentrations prove that the controlling factor is organic vapours, which is a bit like checking the response of plants to environmental conditions like acidity and water and concluding that because when conc HCL was poured over them they all died, the controlling fact in the real world must be acidity: QED cabbages will grow in deserts, because cabages in conc HCL die, therefore acidity controls whether cabbages survive and water therefore is unimportant.
Likewise, they've decided that totally unrealistic levels of organic vapours are the controlling factor and therefore the level of cosmic rays must by inference have no effect.
But simple physics tells us that the rate of cloud formation must be proportional to the level of cosmic rays (assuming only a proportion of the air is ionised). Whereas, the high sensitivity to organic vapours does hint that even very small levels would be sufficient to provide more than enough to allow the development of cloud formation nuclei.
As well as his admirably clear summary, Patagon gave us a reference to 12,400(!) articles to read up on. Thanks, but ouch! Looking through the first dozen or so pages for decent looking freebie review papers, found these:
http://www.deas.harvard.edu/climate/seminars/pdfs/randall_etal_2003.pdf
http://journals.ametsoc.org/doi/pdf/10.1175/1520-0442%282004%29017%3C2493%3ARATCPP%3E2.0.CO%3B2
http://curry.eas.gatech.edu/currydoc/Randall_BAMS84.pdf
Anyone else got any other suggestions for quick priming on cloud modeling?
Surely cloud formation has to be a part of the feedback loop that tries to dampen or stabalise changes in climate in response to the input from the Sun.
When the Sun is more active then extra energy is pumped into the climate giving a rise in atmospheric CO2 and increased vegetation and other forms of life.
Dimethyl Sulphide emitted by various algae in the ocean is broken down in the atmosphere into sulfur dioxide, dimethyl sulfoxide (DMSO), dimethyl sulfone, methanesulfonic acid and sulfuric acid providing the basics for cloud formation from Cosmic radiation.
When the Sun becomes less active more Cosmic radiation enter the atmosphere increasing the amount of cloud further forcing the climate to cool thus reducing the amount of CO2 in the atmosphere and slowing the growth of life removing the basics for cloud formation.
Further dampening being provided by energy storage and transfer in ocean currents.
Life and climate interacting to provide the optimal conditions for life to flourish, didn't Darwin have something to say about life adapting itself and its environment to ensure survival?
Lord Beaverbrook
The DMS cloud feedback, and more generally the idea of "life and climate interacting to provide the optimal conditions for life to flourish", is more Lovelock than Darwin (it's the Gaia theory).
Coming back to the Kikby et al paper, and the relevance to model parametrizations, they cite Pierce and Adams 2009 who looked at this in a GCM and said:
"In our simulations, changes in CCN from changes in cosmic rays during a solar cycle are two orders of magnitude too small to account for the observed changes in cloud properties; consequently, we conclude that the hypothesized effect is too small to play a significant role in current climate change."
Kirkby et al seem to cite this without disagreeing with it (except that they say "may" rather than "is") so it's not clear whether they think they are overturning Pierce and Adams' conclusion or not - and the paper itself isn't really any help there as far as I can see. But anyway, I've written to Kirkby to ask what he thinks.
Richard Betts
If you can obtain his permission, please let us know the substance of his reply.
For what it's worth, my understanding is as per P&A (2009) - the effect is real, but too small to have climatically significant effects.
One takeaway, in particular from D Whitehouse;s excellent summary, is that GCM modellers appear not to fully understand cloud formation processes. Thus we cannot believe the output of their models, and indeed the reality of temperature trends is considerably lower than those predicted by such models.
It took this very basic experiment to show that sulfates, ammonia, and cosmic rays by themselves aren't enough to create the CCN needed for cloud formation. The next speculation is that some other chemicals in the atmosphere, perhaps volatile organic compounds, will give us what we need to form the right amounts of clouds, e.g., to simulate reality accurately enough. (Just a week or so ago, I seem to recall reading something about organics from trees in the Smoky Mountains -- so named because the aerosols emitted from trees in summer creates a haze -- being crucial to cloud formation in that part of the US.)
Whitehouse thus reminds us that we cannot currently believe either in the current GCM fomulation of how clouds form, nor can we at this time believe in Svenmark's ideas of the importance of cosmic rays, because it hasn't yet been demonstrated, in a system which includes all relevant chemicals, that the cosmic rays create enough new cloud at lower to mid latitudes to have to cooling effect he suggests.
I'm still leaning toward Svensmark's hypotheses being eventually demonstrated -- it is just that this elegant experiment at CERN reminds us that when we actually get around to doing the science, we can't predict what we will find.
BBD
OK will do
Richard Betts
Thank you.
Some will doubtless think this press story interprets things too far
http://tucsoncitizen.com/wryheat/2011/08/25/cern-experiment-confirms-cosmic-ray-effect-on-climate-another-blow-to-climate-models/
But there is an interesting statement in the closing sentence-
'The paper presented in Nature today was submitted in September 2010, so it does not reflect the latest results of the CERN experiment.' That more or less coincides with the date when Calder mothballed his blog to work on what seems to be an independent submission.
Therefore, re Richard Betts request, any hints of further results from Kirkby of ongoing CLOUD Experiment runs may be fascinating.
For those who are not afraid of Navier-Stokes, del, Laplacian, or convective derivative:
Turbulence, Clouds and Climate Models
Joao Teixeira, Caltech
http://online.kitp.ucsb.edu/online/turbulence11/teixeira/
It's found in much smaller populations than the whole of Gaia. Interacting colonies of species will even 'co-operate' to modulate the local environment to ensure survival.
anna v was on to the biosphere ultimately or penultimately determining climate.
===================
"'The paper presented in Nature today was submitted in September 2010, so it does not reflect the latest results of the CERN experiment.' That more or less coincides with the date when Calder mothballed his blog to work on what seems to be an independent submission."
Somehow I don't think the way the press are preoccupied with Libya, the releases of Mann's emails and the date of publication of this paper are any coincidence.
A good week to bury bad news?
I want to hear more from Calder about his suppressed graph conspiracy theory frankly. Something smells there.
Am I correct to understand this result in layman terms that:
1) the observed nucleation is 1000 times higher than that recreated in CERN;
2) other nucleating agents must be at work to account for observation in nature;
3) the presumed agents like aerosol and sulphates ate entirely off the mark and regulations to cut emission are not justified;
4) since most man made agents have already been blamed nature must have a big hand in fill this huge gap be it cosmic ray or otherwise;
5) the cosmic way theory has not been disproved but funny enough CERN did not see fit to explore its effect;
6) downplaying the effect or current assumed nucleating agents in the current climate model by a factor of 1000 would result in a very different climate prediction (or not?).
Really informative blog.I am quite amazed by the information in this blog,keep up the good work.
Kirkby has a good slide show depicting the background to and aims of CLOUD here:
http://www.slideshare.net/NEUROMON/cosmic-rays-and-climate-jasper-kirkby-cern-462009
'The researchers were also surprised to find that the DMS molecules completely refresh themselves after only three to five days. That means the plankton may react to UV rays quickly enough to impact their own weather. Toole and Siegel were surprised by the lightning-fast rate of turnover for DMS. '
http://www.nasa.gov/vision/earth/environment/0702_planktoncloud.html
The article and referenced post refer to the need to create a simulated atmosphere because the real atmosphere has too many pollutants. Because the results are so far away from observation, they say there must be something in the atmosphere that' causes the difference. Why not just use atmospheric air? If that produces better results, since its composition can be measured, won't there be clues to work on?
Beaverbrook:
Two useful references, thanks. I'd also just noted Pallab Ghosh's phrase
I love researchers being surprised. Ghosh would have been far better to have explained the surprise and left it at that. At least for now. As David Whitehouse says very well, with the surprise should also come humility. And that single quality would end CAGW, AGW and the rest in an instant.
The sun is very sultry and we must avoid its ultry-violet rays. H/t Coward in Plum Orchard.
We have hypotheses of UV working on ozone in the stratosphere impacting climate and we have hypotheses of UV working on plankton in the ocean impacting climate. It's layers of turtles from the sun to the bottom of the deep blue sea.
Ultry-violet radiation varies around 20%, I believe, compared to only 1/2% variation of TSI. There is enough variation in UV to vary climate as it does, if there are causal connections.
Is there any relationship between the stratospheric impact of UV and oceanic biologic impact of UV? Why wouldn't there be?
===========
Probably also worth reminding ourselves what the CERN press release says:
This result leaves open the possibility that cosmic rays could also influence climate. However, it is premature to conclude that cosmic rays have a significant influence on climate until the additional nucleating vapours have been identified, their ion enhancement measured, and the ultimate effects on clouds have been confirmed.
It sounds likely it may confirm P&A, modulo the mysterious vapours.
See also: http://indico.cern.ch/getFile.py/access?resId=0&materialId=slides&confId=52576 (2009, slide 30).
Here are some statements from the Randall et al paper (Breaking the Cloud Parameterization Deadlock) mentioned earlier (published 2003):
(near the start...)
Despite the best efforts of our community, and notwithstanding the achievement of significant advances, summarized below, the problem remains largely unsolved. At the current rate of progress, cloud parameterization deficiencies will continue to plague us for many more decades into the future.
(right at the end...)
The emergence of superparameterizations presents an opportunity for our community to undertake a "Manhattan Project" for cloud parameterization.
The development of all this stuff therefore looks set to go on for years, probably decades. Perhaps science may never straightforwardly determine the policy choices.
For these reasons and others, we are obliged to decarbonize IMO. Meanwhile, development and growth will freely continue. I think the idea we can't do both is simply another green daydream, and one that should be rejected.
For these reasons and others, we are obliged to decarbonize IMO
Aug 26, 2011 at 2:47 PM | Philip
Just in case again!!!!!!! Insane! Please show the alternative energy source you will provide. Put up or shut up!
I have to disagree with you there Philip, we would only be obliged to decarbonise if a reduction in carbom was a viable solution to any problem. To decarbonise because the science is still unclear is a waste of money time and effort, three things that some of us have decidedly less of than our younger friends.
Pete H, "Put up or shut up!"
It's not written in stone. But gas and nuclear for now, I would have thought, and as quickly as possible. Longer term, more R&D. There are several new types of nuclear in the offing. With more development, CCS technology and solar may be able to contribute. Possibly even fission if we are lucky, who knows what research may turn up?
What do you think?
Lord Beaverbrook, "To decarbonise because the science is still unclear is a waste of money time and effort"
We have already been decarbonizing for a hundred years or more. Remember, it means reducing the carbon intensity of energy, nothing more. If you like, we need to decarbonize because we need to provide more energy. How then can it really be a waste of money time and effort?
Just in case there's any doubt. ... The government's current energy policy is a shambles and a disgrace. It will wreck the countryside, and through brownouts and price hikes cause unnecessary suffering to the weak and old. By damaging the economy, it will leave us less able to help ourselves and others. I have written many times to my MP complaining about it and will continue to do so. I've added my name to all petitions against wind farms that I am aware off. I've engaged on hostile blogs, I've attended anti-climate bill meetings at the HoC. What more can anyone do? I completely stand by what I've said here over the past couple of days, there is no contradiction. We need to find ways to oppose the greens, but without shooting ourselves in the foot at the same time.
I'm afraid you've lost me there, Philip.
Coal has been the cheapest most efficient form of energy for the last hundred years and will be for the next. To choose not to use coal or to add extra cost to the use of coal is defeating the object of providing more energy to a growing population.
Philip
Absolutely agree with what you say above, culminating in this:
So, champion nuclear, force renewables proponents to show their numbers - all of them - and stick it to the energy fantasists (they'll be greens) at every opportunity.
Lord B
The science is clear enough: CO2 can and does cause energy to accumulate in the climate system through increased RF. Yes, there are uncertainties. No, they do not add up to a coherent argument for a low climate sensitivity, never mind a 'refutation of AGW'.
We need to decarbonise, and electricity supply is the logical place to start. Displacement of coal by nuclear for baseload generation is the obvious way to do this.
BBD:
Thank you! The energy situation is really starting to scare me now. I hope the government will see sense soon.
A question came up yesterday over whether it is possible for the private sector alone to carry the burden of energy technology R&D. Have you noticed any reasonable discussion of this question that you can point to?
Lord Beaverbrook:
The hundred years of decarbonisation is based on figures from A.Maddison and the U.S Department of Energy, which show that the emissions per unit of GDP have halved since 1910. I'm sorry, I can't find you a reference at the moment. You might though look here for a hint, http://eprints.lse.ac.uk/27939/1/HartwellPaper_English_version.pdf:
We share the common view that it would be prudent to accelerate the historical trend of reducing the carbon intensity of our economies, which has been a by-product of innovation since the late eighteenth century.
Philip
I'm not at all sure the private sector can bear the financial burden. Nor should it be expected to.
Some, eg Roger Pielke Jr, argue for a carbon tax (not Pigovian), and against cap and trade. The revenues from this tax to fund R&D into energy generation technologies including Gen IV IFR and renewables, particularly wind and solar.
The thinking behind this is set out in detail by Pielke and co-authors in the 'Hartwell Paper':
.Chapters 3 and 4 of Pielke's 2010 book The Climate Fix are also essential reading.
Sorry - I see that you referenced the Hartwell paper above...