Good reads on climate models

Two excellent and accessible papers on climate models have appeared today. Ross McKitrick, writing in the Financial Post, wonders about model evaluation:

So how do models do at predicting the spatial pattern of warming over land? Though the 2007 report of the Intergovernmental Panel on Climate Change (IPCC) devoted a whole chapter to model evaluation, it said almost nothing about this question. The IPCC talked mainly about static features, such as whether the model can make the tropics hot and poles cold, and so forth. But it was mostly silent on the spatial changes. A 2008 report of the U.S. Climate Change Science Program went a bit deeper, but only to report on tests of how daily and seasonal variations in models matched the real world (is winter a suitable amount colder than summer, etc.).

The reports weren’t ignoring anything: There just hasn’t been much work on the topic.

Meanwhile, Tamsin Edwards has this:

Models are always wrong, but what is more important is to know how wrong they are: to have a good estimate of the uncertainty about the prediction. Mark and Patrick explain that our uncertainties are so large because climate prediction is a chain of very many links. The results of global simulators are fed into regional simulators (for example, covering only Europe), and the results of these are fed into another set of simulators to predict the impacts of climate change on sea level, or crops, or humans. At each stage in the chain the range of possibilities branches out like a tree: there are many global and regional climate simulators, and several different simulators of impacts, and each simulator may be used to make multiple predictions if they have parameters (which can be thought of as “control dials”) for which the best settings are not known. And all of this is repeated for several different “possible futures” of greenhouse gas emissions, in the hope of distinguishing the effect of different actions.

NYT on clouds

Justin Gillis, the New York Times' eco-activist has an article on the role of clouds in climate and the dispute over their impact. It is, in essence, an extended pop at the work of Richard Lindzen.

Among the experts most offended by Dr. Lindzen’s stance are many of his colleagues in the M.I.T. atmospheric sciences department, some of whom were once as skeptical as he about climate change.

“Even if there were no political implications, it just seems deeply unprofessional and irresponsible to look at this and say, ‘We’re sure it’s not a problem,’ ” said Kerry A. Emanuel, another M.I.T. scientist. “It’s a special kind of risk, because it’s a risk to the collective civilization.”

Bob's book

One of the blogs I've been struggling to get the time to read for quite some time now is Bob Tisdale's. This is Climate Audit territory - lots of graphs, lots of statistics, lots of reading to do before you can understand the full story. Like Climate Audit it's a site that cries out for an introductory text to enable newbies to catch up with the story that has gone before.

Fortunately, Bob Tisdale has now produced a book, which I'm currently working my way through. There are still lots of graphs and lots of statistics, but it's written in a good accessible fashion and I'm getting a great deal from it.

Buy the PDF here or Kindle versions here (US) or here (UK).

Quote of the day

One climate modeller we interviewed explained that the climate is a ‘heterogeneous system with many ways of moving energy around from system to system’ which makes the theoretical system being modelled ‘tolerant to the inclusion of bugs'.

From the Pipitone and Easterbrook paper on validating climate model software, currently the subject of a guest post at Judith Curry's.

How are the statistics?

I'm not a statistician, but I have been hanging around with statisticians for several years now and I have picked up a certain amount along the way. I was therefore intrigued by the paper by Lohmann et al, which Luning et al discussed at WUWT a couple of days ago. The Lohmann paper is a comparison of climate model output with proxy sea surface temperature reconstructions for the Holocene and concludes that the correlation between the two is poor.

I had a leaf through the paper and was struck by the fact that they seem to have calculated a simple R2 for the correlation. Can any of my statistically qualified readers tell me if this is right? I had thought that both series would be highly autocorrelated and that any correlation measure would therefore be inflated. Shouldn't they correct for autocorrelation?

Continental hindcasts

I recently emailed Richard Betts, inquiring about evidence that climate models could correctly recreate the climate of the past ("hindcasts") at a sub-global level. Among other things, Richard pointed me to FAQ 9.2 from the IPCC's Fourth Assessment Report, a continental-scale comparison of model output with all forcings (red band), natural-only forcings (blue band) and observations (black line). This figure also appears in the Summary for Policymakers as SPM 4. There is a similiar analysis at subcontinental level in the same chapter of the report.

Click to read more ...

More on fraternisation

Interest in the ongoing "football match between the trenches" continues, with an article on Yale Climate Forum by Keith Kloor, who discusses the Met Office's outreach work as well as Scott Denning's visit to the Heartland Conference. This bit was somewhat irksome.

Reminded that much opposition to climate science (and dismissal of climate change) seems ideological in nature, perhaps limiting the amount of headway that can be made on the science if people are already predisposed against it, Denning agreed that the culture war dynamic presents a high hurdle to overcome.

“Almost everyone that dismisses climate change as a problem does it for ideological or political reasons, not for scientific reasons,” he said. “We scientists need to recognize that.”

I clearly have an ideological objection to the use of unvalidated climate models as policy tools. There must be an "-ism" for that. :-)

Another rebuttal

Richard Betts points us to this paper by a group of climatologists who seek to rebut Richard Lindzen's talk at the House of Commons the other day. The authors are, in the main, familiar names. John Mitchell and Brian Hoskins featured regularly in the Climategate emails and both were involved in the coverups too; Eric Wolff made a couple of brief visits to BH in the wake of the Cambridge Conference last year, but was put off by the over-hostile reaction from commenters; Tim Palmer has been mentioned on the pages of BH a couple of times. Keith Shine is less familiar to me although he too has been mentioned before as one of the members of the Royal Society's advisory panel on climate change (as indeed are most of the others).

With my current focus on climate models, here's an interesting excerpt:

At every stage models should be evaluated by exhaustive comparison with observations. The models encapsulate our understanding of the basic science of the climate system, including for example, Newton’s laws of motion, the laws of thermodynamics and the quantum theory of radiation. When deficiencies are found at one level then improvements are sought and the lessons learnt should cascade to models at other levels. This is, of course, the ideal: the actual development of the science is rather more irregular but very definitely in this direction. Even the models at the more complete and complex end contain many uncertainties and deficiencies, which are widely recognised within the modelling community, but they are the best guide we have as to how the climate system may change in the future. Their results are not to be accepted in an unquestioning manner; they should be analysed in detail, with the dominant processes behind any climate variability and change thoroughly investigated using observations and simpler models in the hierarchy.

I think the words "out of sample" need inserting in a couple of places in that paragraph. I think it would also have helped if Hoskins had reiterated his earlier clarification about the limitations of climate models - namely that they are "lousy".

Nordhaus and the sixteen

The war of words between economist William Nordhaus and the sixteen scientists continues in the New York Review of Books. Cohen, Happer and Lindzen (CHL) seem to have been delegated to speak for the sixteen.

It's all interesting stuff. I was struck by this bit from Nordhaus's response.

The final part of the response of CHL comes back to the economics of climate change and public policy. They make two major points: that the difference between acting now and doing nothing for fifty years is “insignificant economically or climatologically,” and that the policy questions are dominated by major uncertainties.

Is the difference between acting now and waiting fifty years indeed “insignificant economically”? Given the importance attached to this question, I recalculated this figure using the latest published model. When put in 2012 prices, the loss is calculated as $3.5 trillion, and the spreadsheet is available on the Web for those who would like to check the calculations themselves. If, indeed, the climate skeptics think this is an insignificant number, they should not object to spending much smaller sums for slowing climate change starting now.

If your climate models cannot hindcast regional climate and they have not been shown to be able to predict either global or regional climate, what is the point of discussing a cost-benefit analysis based on their output? You might just as well have used a ouija board.

Conveying truth 4

As my thoughts have turned increasingly to climate models in recent months, I thought I might take a look to see how Julia Slingo described these mathematical behemoths in the briefing she sent up to central government in the wake of Climategate - I have criticised this document on a number of occasions in the past (1, 2, 3).

Considering just how central climate models are to the case for DAGW* I was taken aback by how little Slingo had to say on the subject. And what she said was, to say the least, surprising. Here it is:

Models have been used that take into account all the factors that influence climate. These models have been able to simulate the historic changes in global and regional temperatures and have shown that most the warming over the past half century has been caused by the rise in greenhouse gas concentrations.

There are a couple of passim mentions elsewhere in the document, but in essence that is what the Met Office feels that people in government need to know about climate models.

It's hard to know where to begin. I wonder whether anyone will seek to defend Slingo's paper as a fair representation of the reliability and importance of climate models.

Readers should feel free to critique Slingo's words, but please avoid venting. I'm more interested in what policymakers should be told about climate models. I think we should allow ourselves slightly more space than Slingo - shall we say eight sentences?

*Dangerous anthropogenic global warming - "dangerous" being perhaps a less emotive term than "catastrophic".

Happer in the WSJ

Will Happer has an article in the Wall Street Journal reviewing global warming science. Bob Ward says it's unscientific. I must say it seems unobjectionable or even irrefutable to me.

What is happening to global temperatures in reality? The answer is: almost nothing for more than 10 years. Monthly values of the global temperature anomaly of the lower atmosphere, complied at the University of Alabama from NASA satellite data, can be found at the website http://www.drroyspencer.com/latest-global-temperatures/. The latest (February 2012) monthly global temperature anomaly for the lower atmosphere was minus 0.12 degrees Celsius, slightly less than the average since the satellite record of temperatures began in 1979.

The lack of any statistically significant warming for over a decade has made it more difficult for the United Nations Intergovernmental Panel on Climate Change (IPCC) and its supporters to demonize the atmospheric gas CO2 which is released when fossil fuels are burned. The burning of fossil fuels has been one reason for an increase of CO2 levels in the atmosphere to around 395 ppm (or parts per million), up from preindustrial levels of about 280 ppm.

Pinning down the debate

In his Radio 5 interview, James Delingpole correctly framed the argument over AGW as being over (a) how large the effect is (b) how much warming there will be and (c) how much of a problem it is.

Vicky Pope at the Met Office has taken a different approach in an article in the Guardian today.

You can see research by the Met Office that shows the evidence of man-made warming is even stronger than it was when the last IPCC report was published. A whole range of different datasets and independent analyses show the world is warming. There is a broad consensus that over the last half century warming has been rapid, and man-made greenhouse gas emissions are very likely to be the cause.

Click to read more ...

Hopeful fudging

I briefly met Chris Hope of the Judge Business School at the Cambridge Conference last year - Chris works in climate policy if I remember correctly. He has just started following me on Twitter and sent a tweet in response to reader Andrew's guest piece on mathematical models.

He was picking up on the statement that "invariably [models] require 'tuning' to real world measurements" and responds:

Aren't climate scientists criticised for this?

This seemed a reasonable point to me.

Mathematical models for newbies

Reader Andrew sent me his summary of the basics of mathematical models, which I think readers will find useful.

I have been devising mathematical models (simulations) of physical processes for over 20 years, and I just wanted to point out some of the basics that might help people understand these types of models:

1. The physics of the process (to be modelled) may be well understood, but although this helps it is somewhat irrelevent to the accuracy of all but the  most simple model (although you will almost certainly not get a good model if you don't understand the physics). Nearly all computer models are based on mathematical formulae, commonly binomial expansions, that are representative of the physical situation.

Click to read more ...

Climate models for politicians

Some weeks ago, I invited readers to improve upon parts of a summary of global warming science, written by Julia Slingo for the benefit of readers in central government. The ground covered was mainly about surface temperatures. At some point I may well write this up into something more formal.

I think it would be interesting to also say something about climate models and their uncertainties and I have been giving this some thought. My knowledge of climate models is somewhat sketchy, so some of my understanding may be incorrect, but here's the ground I think central government really ought to understand:

Click to read more ...