Bishop Hill Extreme weather
Jul 18, 2012 Climate: other An interesting article about the wet British summer comes to us from Michael Hanlon in the Mail.
Perhaps the most dramatic and visible impact of climate change to date has been the reduction in Arctic sea ice cover and, particularly, thickness, seen in the last 20 years or so. The ice grows back in winter of course but, the evidence suggests, each year (on average) perhaps a little thinner than before. This makes the North Atlantic a little warmer than otherwise, reducing the temperature gradient between Polar and Tropical air and hence taking some of the wind (literally) out of the jetstream’s sails.
So, the overall effect of global warming will be to make our summers cooler and damper. The trouble is, this contradicts what most of the computer models have been saying to date, namely that in Britain we can expect hotter, drier summers and milder, damper winters. I spoke to Kate Willett, a climate scientist at the Met Office who agrees that the picture is confusing. “Yes, this contradicts the model of wetter winters and drier summers,” she says. It is also true, she adds, that since 2007 Arctic sea ice levels have been exceptionally low but it is not true that the last five summers have been exceptionally bad – those of 2010 and 2011 were average or a little above average in terms of sunshine and temperature.
This is almost beyond parody, so I'm not even going to try.
Interestingly, Hanlon says that there is a consensus that weather will become more extreme in an warming world. Is this right?
Reader Comments (109)
Hi Marion
No, we stand by the UHI stuff. The brochure was pulled because the uncertainties in paleoclimate reconstructions were not well represented, and because of highlighting of a statement from AR4 that was controversial and only based on one paper, and because of too much focus on the old UKCIP02 "drier summers" idea, which as I said at the start of this thread is rather out of date due to wide uncertainties in precip projections.
AR5 reviewers are indeed asked to keep the draft content confidential until it is released at the end of the process. NB I don't make the rules, I'm just telling you what the IPCC say!
I just thought Philip has quite a talent for spotting key issues in the peer-reviewed literature which remind us that models should be used with caution. Whether he chooses to put this talent to use in helping make sure that IPCC authors consistently remember to use such caution is up to him. Philip, if you want to review the WG2 FOD, email tsu@ipcc-wg2.gov
Sorry I can't stick around to chat today, but thanks for the discussion.
Cheers
Richard
Re: Jul 20, 2012 at 1:16 PM | Richard Betts
Thanks for the input, Richard, and the confirmation that Steve McIntyre's observations were correct (as usual!) ie
Steve McIntyre
Posted Jan 13, 2012 at 12:45 AM
"I’m accepted as a AR5 reviewer but have refrained from downloading First Draft documents since I am not prepared to agree to the confidentiality terms. IPCC’s main enforcement mechanism seems to be their threat to expel someone as a reviewer. Since they ignored my review comment. I dont see that this thread has much downside for me.
Otherwise I’m not sure what they can do against someone who doesn’t rely on government grants. (Anyone relying on government grants who defied IPCC would pay the price when he sought new funding – that’s for sure.)
Would they sue me for commenting publicly on their documents? If so, for what? Offhand, it seems an unattractive course of action for them, but you never know."
http://climateaudit.org/2012/01/12/stockers-earmarks/
And interesting that the statement
"The urban heat island effect already warms central London by more than 10C on some nights. Increased urbanisation and release of waste heat would increase this still further - on top of the effects of global warming"
still stands.
It would seem to be quite contradictory to Phil Jones 1990 study that the IPCC relied so heavily on, the study that untruthfully claimed the Chinese temperature stations it used "were selected on the basis of station history: we chose those with few, if any, changes in instrumentation, location or observation times” when it seems the opposite may have been the case
See the evidence provided by Douglas Keenan
http://www.informath.org/apprise/a5610/b101201.htm
And yet another example of where an IPCC author cites his own study in an IPCC Chapter ie “Surface and Atmospheric Cimate Change”
Leading the IPCC to claim -
"In summary, although some individual sites may be affected, including some small rural locations, the UHI effect is not pervasive, as all global-scale studies indicate it is a very small component of large-scale averages. Accordingly, this assessment adds the same level of urban warming uncertainty as in the TAR: 0.006°C per decade since 1900 for land, and 0.002°C per decade since 1900 for blended land with ocean, as ocean UHI is zero."
http://www.ipcc.ch/publications_and_data/ar4/wg1/en/ch3s3-2-2-2.html
Yet more utterly worthless 'science' from the IPCC.
Steve McIntyre comments -
"Having more or less satisfied myself that the analysis of the Chinese network in Jones et al was worthless, my last post in the series on the China network was on April 19. I later took a look at other UHI analyses that supposedly proved that UHI was unimportant , with several posts on Peterson 2003 here and Parker 2006 here here here, also concluding that this analysis didn’t prove anything. Whatever the true contribution of urbanization to 20th century trends, it seemed to me that these various studies all suffered from conceptual defects. The failure of the Jones et al 1990 China network to define a “rural” network was typical, rather than exceptional."
http://climateaudit.org/2010/11/04/phil-jones-and-the-china-network-part-2/
Jul 19, 2012 at 3:19 PM | Richard Betts
Hi Richard,
Thank you again for your generous response, and especially for acknowledging that the low-frequency variability of unforced runs of HadCM3 is lower than observations.
Thank you also for pointing out the Stott et al diagram, which I agree does give a handle on the simulated variability in the natural forcings case as well. I'd add that this diagram uses the Lean estimate for solar forcing; if the more recent Svalgaard / Preminger reconstructions are more accurate, then the simulated variability in the natural-only case is likely to be even smaller than that shown in the diagram, especially for the first half of the 20th C.
Eye-balling the diagram, the natural-only 20th C variability for HadCM3 looks quite similar to that shown in the Collins et al for the unforced case. It is therefore reasonable to think that in common with the other models tested, HadCM3 also shows too little variability when forced by realistic natural forcings. If so, then the spectrum in the natural-only case would also likely be flat after 20/30 years and the real-time fluctuations will decrease in size with increasing time scale, rather than increasing (i.e. the short term fluctuations will dominate the long term).
So while I accept that HadCM3 shows far less than the recent observed warming when anthropogenic forcings are excluded, I also think it shows far less than the observed variability for the period before anthropogenic forcings became an issue.
I can try to quantify this a little. If the simulated natural-forcings variability over the 1,000 year time scale is similar to the unforced case from Collins et al, then by eye the average fluctuation size for HadCM3 flattens out after the 20/30 year mark to around 0.2 K. In comparison, the observed real-time fluctuations continue to increase after the 20/30 year mark. If the scaling rules are taken seriously, then they can be used to estimate the average size of these fluctuation as a function of their time-scale. The following estimates have been read off from the climate regime reference slope in figure 3:-
Dt(yr) => DT(K)
20 => 0.30
50 => 0.44
100 => 0.57
200 => 0.75
500 => 1.08
1000 => 1.43
10000 => 3.58
25000 => 5.17 (roughly, the glacial/interglacial variability)
Once you get beyond the 50 year mark, the observed averaged natural variability becomes significantly larger than the variability simulated by HadCM3 and other models using realistic natural-only forcings. The observed average fluctuations for 50 and 100 years are comparable with the fluctuations in the instrumental record (although slightly smaller I agree). Over the 1,000 year time scale, the average fluctuations look compatible with the recent Christiansen and Ljungqvist multi-proxy reconstruction (figure 5).
To be honest, I don’t understand how it would be possible to account for the Stott et al discrepancy between observations and models simply by interpreting the simulated time series. I think the evidence tends more to confirm that the variability seen during the 20th C is not particularly exceptional compared to that seen in earlier times and that the reason for the discrepancy in the Stott at al diagram is simply the weak long-term variability simulated by HadCM3 and the other GCMs.
Philip, you don't seem to understand that Richard's model is sufficiently omniscient to be used as a control for comparison with actual observations in the attribution problem.
Your approach looks interesting, although I am not convinced the climate is not too chaotic to use this as a basis for anything beyond discussion of model effectiveness.
Hi Rhoda,
"...use this as a basis for anything beyond discussion of model effectiveness"
You're right! All I've done is pointed out some statistics that are seen in the observed temperature series, but probably not in the model simulations. This discrepancy has been noticed in a number of scientific papers that investigate natural variability in the climate system. The numbers I quoted for DT vs Dt are averages calculated over the observed temperature series. If you'd like to find out a little more about this angle, here is one good place to start.
Re: Jul 23, 2012 at 12:42 PM | Philip Richens
Thanks for the link, Philip, and neat response to RB
"I don’t understand how it would be possible to account for the Stott et al discrepancy between observations and models simply by interpreting the simulated time series. I think the evidence tends more to confirm that the variability seen during the 20th C is not particularly exceptional compared to that seen in earlier times and that the reason for the discrepancy in the Stott at al diagram is simply the weak long-term variability simulated by HadCM3 and the other GCMs."
Rhoda,
Perhaps you should include also the best evidence in defence of CO2, along the lines of models vs reality, just a thought.....
Models. Well, one can learn from models, and what you usually learn is that you were wrong. Then you tune the model and start again. A model which makes correct predictions is necessary, but not sufficient . What one can say is that only the models that predict this or that measuremnt are correct. The rest may be sent back to try again. What happens in some branches of climate science is that they use an ensemble of models and find the one that got nearest and make much of that result. But you can only do that once, as you have now selected your model. The others are plainly not correct.
But with the best will in the world they can't be used as proof of anything. If they got the last hundred years right they might not get the next ten. So I'm not doing models. If scientists want to compare reality with model output good luck to them, but it isn't ever going to be evidence.
I am not in favour of going over measurements to make them line up with model output either, but I acknowledge that sometimes that may be valid. Maybe it was in the case of the sondes, although changing methods in the middle then looking at the whole sequence for trends is fraught with danger and the temptation to find what you wanted to find.
"But with the best will in the world they can't be used as proof of anything. If they got the last hundred years right they might not get the next ten. So I'm not doing models. If scientists want to compare reality with model output good luck to them, but it isn't ever going to be evidence."
Totally agree, and if models, vulnerable to subjective assumptions as they are, are cited as 'best evidence' then isn't that proof that they have no 'proof' at all.
Especially when such models are politicised and it seems likely that political decisions have driven some of the parameters involved. See too the latter part of my response to EU at
http://www.bishop-hill.net/blog/2012/7/21/a-trickle-of-further-information.html
Jul 23, 2012 at 11:02 AM | Marion
Jul 23, 2012 at 1:10 PM | Marion
Hi Marion,
I hope you enjoy the link. I think this research is very interesting and relevant and yet it tends to be somewhat neglected. Just to clarify my response to RB ... it reflects a genuine perplexity, and Richard may easily have a good resolution. I hope he will return to us with a further clarification of his position.