Bishop Hill New study finds low climate sensitivity
Jan 25, 2013 Climate: sensitivity Via Leo Hickman, more evidence that aerosols have a small impact and that climate sensivity is low.
When the researchers at CICERO and the Norwegian Computing Center applied their model and statistics to analyse temperature readings from the air and ocean for the period ending in 2000, they found that climate sensitivity to a doubling of atmospheric CO2 concentration will most likely be 3.7°C, which is somewhat higher than the IPCC prognosis.
But the researchers were surprised when they entered temperatures and other data from the decade 2000-2010 into the model; climate sensitivity was greatly reduced to a "mere" 1.9°C.
Professor Berntsen explains the changed predictions:
"The Earth's mean temperature rose sharply during the 1990s. This may have caused us to overestimate climate sensitivity.
"We are most likely witnessing natural fluctuations in the climate system – changes that can occur over several decades – and which are coming on top of a long-term warming. The natural changes resulted in a rapid global temperature rise in the 1990s, whereas the natural variations between 2000 and 2010 may have resulted in the levelling off we are observing now."
I wonder how much more evidence we need of low climate sensitivity before policymakers are forced to take notice?
Reader Comments (82)
Geronimo I had not seen your post of 8.21 am when I crafted my own meagre rejoinder which is rendered otiose.
Jan 26, 2013 at 9:08 AM | N.Tropywins
Not at all, (I'm pretending I know what "otiose" means.) In fact the major contributor to my post was Dr. Trenberth.
(a) Has this work been published ?
(b) I don't fully understand how they can say anything about equilibrium climate sensitivity on the basis of 20 years of data.
(c) Even if they can, there seems to be so much variability in the estimate right now (3.7 to 1.9 on the basis of an extra 10 years data), that this can hardly been seen as any kind of a definitive figure.
Bruce Cunningham
Isn't the main point that the contribution of the small heater in increasing, while that of the large heater is stable. Although if there is an amplifying feedback loop, the contribution of the large heater will actually increase.
HaroldW re the Aldrin et al (2012) study
"Their figure 6a shows the posterior distribution with this uniform prior: mean of 2.0 K, 90% interval 1.2 to 3.5 K. Figure 6f shows the result with a 1/S prior (which Nic says is the Jeffreys' prior): mean of 1.5K, 90% interval 1.1 to 2.5K. [by eye] The mode doesn't change much -- from ~1.6K to ~1.4K (again, by eye)."
I said that that a uniform-in-1/S prior would usually be the Jeffreys' prior where S (climate sensitivity) was the only parameter being estimated. Confusingly, because you have to compute the derivative on a change in variable from 1/S to S, a uniform-in-1/S prior has the form 1/S^2 when expressed as a prior for S.
I should, however, have made clear that was only true when S was being estimated directly from measurements, as in Forster & Gregory 06 and Gregory 02. Where, as in Aldrin et al (2012), S is being estimated by comparing observed temperatures or other variables with model-simulated versions thereof at varying settings of the model parameter(s), then the Jeffreys' prior may be completely different. Where other parameters are being varied and estimated as well as S, the Jeffreys' prior will be a joint prior for all parameters, and its shape along the S axis may vary with the values of the other parameters.
I certainly think that the uniform-in-S prior used by Aldrin was not close to the Jeffreys' prior, and that it results in the risk of high S being overstated. But a uniform-in-1/S prior may go too far the other way.
His paper does show, incidentally, that the mode is a much more stable central estimate than the mean.
BTW, I'm not sure that the press release was about Aldrin et al (2012), the study cited in the SR5 SOD. I think he had a second study on the go; this may be what it refers to.
Alan Kennedy
"If your account is correct, and given water vapour is a powerful GHG, the setup you propose would obviously be very unstable indeed. It doesn't seem that way in reality. Isn't that an issue?"
In the short term it is horribly unstable. In the UK we can get temperature swings of 20C in a week, under normal conditions! My own local station has recorded 30C and -16C in recent years.
In 2012 it took less than a month to go from unusual drought to unusual flooding. The sceptics are convinced such variation is within normal limits. Among those studying climate is an increasing concern that the extra energy accumulating in the system is making it even more unstable, with increasing probabilities of extreme weather of all types.
Entropic Man
What is the supposed mechanism behind this "increasing probabilities of extreme weather of all types.". Is it one of Hansen's inventions ?
It puts us all at the mercy of journalists. It's like putting kids in charge of a chocolate factory. They will exaggerate literally everything for money. I saw a headline in the Groaniad the other day.
"Antibiotic-resistant diseases pose 'apocalyptic' threat, top expert says
There's always a geek who wants his 15 minutes of fame.
"Among those studying climate is an increasing concern that the extra energy accumulating in the system is making it even more unstable, with increasing probabilities of extreme weather of all types."
There is no evidence for an increasing incidence of extreme weather of all types. There is no evidence for Trenberth's "missing heat". It's a travesty!
Nic Lewis (4:29 PM) -
However, you write that uniform-in-1/S prior gives a prior for S inversely proportional to the square of S...does this mean that you believe they mis-calculated that prior? [Notwithstanding your comment that uniform-in-1/S may tilt the scales too far in the direction of low S.]Thanks for the clarification. Not being a statistician by trade, I wasn't aware of those subtleties in constructing a prior, so I appreciate the extra information.
.
I note that Aldrin says, in section 4.6,
How does "extra energy accumulating in the system" cause drought exactly?
Sensitivity may be low? Someone needs to tell Nicholas Stern:
http://www.guardian.co.uk/environment/2013/jan/27/nicholas-stern-climate-change-davos
HaroldW
"However, you write that uniform-in-1/S prior gives a prior for S inversely proportional to the square of S...does this mean that you believe they mis-calculated that prior?"
Good question. I'm unsure whether they mis-calculated the prior or simply mis-stated it, saying 1/S when they should have said 1/S^2.
Measuring off the graphs, the ratio of the PDF at S=2 to that at S=2 is 3.1x for the main graph, but 1.0x for the 1/S graph. That 3.1x/1.0x = 3.1x ratio-of-ratios is halfway between a 1/S and 1/S^2 relationship (for which the S=2:S=1 ratios would respectively be in the ratio 2:1 or 4:1 between the two graphs). Maybe soemthing else that I have overlooked is going on. I'll ask Magne Aldrin.
Sorry, a typo in my last comment.
"Measuring off the graphs, the ratio of the PDF at S=2 to that at S=2 is 3.1x for the main graph, but 1.0x for the 1/S graph."
should read:
"Measuring off the graphs, the ratio of the PDF at S=2 to that at S=1 is 3.1x for the main graph, but 1.0x for the 1/S graph."
"In 2012 it took less than a month to go from unusual drought to unusual flooding. The sceptics are convinced such variation is within normal limits. Among those studying climate is an increasing concern that the extra energy accumulating in the system is making it even more unstable, with increasing probabilities of extreme weather of all types."
I think it depends on how you classify unusual. For example, in 1979 there was a snowfall of 4 inches of snow in South London that lasted for an entire month. I know because i walked to work through it. The present "arctic conditions" foretold by Piers Corbyn are melting away after one week.
And, of course, Hansen's latest paper should not be understood as meaning that he is suggesting that added heat will increase the probabilities of extreme wether of all types - if you are following the climate etc post on the latest BEST findings.
diogenes
Thank you. I had'nt seen the new BEST data.
http://berkeleyearth.org/results-summary/
Remind me to show it to all those people telling me no warming has taken place since 1998. :-)
Presumably 1998 does not stand out because, being a record of land temperatures only, the BEST data shows less effect from the intense El Nino of that year.
Regarding the extreme weather debate. we'll have to wait for a longer baseline to be sure. Non-linear systems often show a period of increased instability before tipping over into a new stable state. It will be interesting to see if something similar is happening to the climate.
40 years of reading climate change research has left me with a "fighting the hydra" feeling. You answer one question and three more pop up in its place. What fun!
For those asking me about increasing energy and extreme weather, there's a summary designed for politicians at
http://www.ipcc-wg2.gov/SREX/images/uploads/SREX-SPMbrochure_FINAL.pdf
"It is impossible to predict whether the climate will be warmer cooler or just the same in 2100. "
No, it is possible to make the prediction. And one doesnt even need a computer. I predict it will be warmer. There, I did it.
Paul Butler,
Water vapor and CO2 are about equal in their effectiveness as GHGs, with water vapor being slightly better (gases such as methane are much better GHGs). In the warm tropical and sub-tropical areas of the globe, water vapor can make up as much as 5% of air. CO2 is a constant .04%. That is 125 times more water vapor than CO2. In the polar regions, the amount of water vapor is much lower, less than 1%, but it is still many times that of CO2 (after all, it still snows massive amounts there at times). This comparison is using the present value of CO2. In past years, the inequality would be even more profound. Moreover, the places on the globe where water vapor is at its highest, happens to be the same locations where solar radiation is also the highest! Yes, the contribution of CO2 is increasing, but it has a long way to go to have as much effect as water vapor. When we reach the point where we have doubled the amount of CO2 in the air, we will have increased the total amount of GHGs in the air by about 2%.
Keep in mind that contrary to what some are saying, water vapor is never completely absent from the air and is always at a much higher concentration than CO2.
Entropic Man
When I mentioned instability, I meant soemthing more extreme than your examples. What Motl was claiming was that if a system involves strong positive feedback it becomes REALLY unstable. It just doesn't seem that way. I live in France. True, a few years back it was 42 degrees. But it has never been 52 degrees. Why not?
I read the useful paper you linked to. But looking for increased variability by counting the incidence of rare events seems a misguided (and circular) way of going about the task, One of the things it is relatively easy to test for is a change in variance. Take the past 15 years or so, when the mean hasn't changed (much): has the variance increased or not ?
This quote :" It is likely that anthropogenic influences have led to warming of extreme daily minimum and maximum temperatures at the global scale" is either a typo or it is a statement about a change in the mean, not the variance.
If it was really getting warmer (or in the presence of flat temps, energy accumulating somewhere else) wouldn't we see undisputed unprecedented extreme weather. Not just things that don't happen very often, not just hottest/wettest/coldest since the year XXXX? Real new records. Actual unprecedented events? We really are not, and my intuitive feeling is that there would be far more of them (or, indeed some of them) if anything unprecedented in the way of energy balance was going on. The very absence of those events is a real indicator of 'not much going on'.
EM, you really have to tell us whether the extra heat is going down a hole in the ocean and coming back in 800 years or not. And how we'd tell if it were? Or was it 200 years? Or 2,000?
The variance has increased. The graphs under discussion are for temperature, mainly because it it easiest to quantify. They are starting to show an asymetric increase in the total temperature range.
The minimum temperature boundary hardly changes, but the mean and mode increase.
At the upper temperature boundary, what were previously extremely rare temperatures become more frequent. As the upper temperature boundary extends, previously unheard of temperatures become possible and new high temperature records are set.
One way of testing for this would be to take a large sample of weather stations and plot the number of new records set against time.
If climate were stable, you would expect a small and fairly constant rate of new records, high and low.
If the low temperature end was extending more cold records would be seen
If the high temperature end were extending, we would see more hot records would be seen.
What is starting to emerge is an increase in high end records, while the number of cold end records stays constant. This is consistent with a warming trend extending the high end of the frequency curve while the cold end stays in place.
This is aan extreme climate index from NOAA. Judge for yourself.
http://www.ncdc.noaa.gov/extremes/cei/graph/1/01-12
Is that your final answer? That we are seeing high end increases? Now, where is the unprecedented stuff? I don't mean that the temp on July 4th in Podunk was higher by the width of a thermometer division line than it was in 1934. I mean the temp somewhere was higher than it had ever been, corrected for UHI and land use and reflected in nearby temps. Or rainfall or storms or whatever, ditto. Are we seeing that in the CET series, the longest comprehensive series there is? Because what I can see fits my conjecture that nothing much is happening.
rhoda
"EM, you really have to tell us whether the extra heat is going down a hole in the ocean and coming back in 800 years or not. And how we'd tell if it were? Or was it 200 years? Or 2,000?"
The thermohaline circulation is decreasing.
ftp://kakapo.ucsd.edu/pub/sio_220/xx%20-%20Formation%20of%20Greenland%20Sea%20Water/Wadhams_et_al.jpo_02.pdf
That decreases the amount of heat leaving the Arctic into the deep ocean. The consequent heat accumulation would correalate with the rapid Arctic warming and the recent record ice loss.
It's not enough to account for the mismatch between the extra incoming energy and the static surface temperatures. Something elsewhere is acting as a heat sink.
The deep ocean is still the best candidate. While temperatures remain static at the surface the sea level continues to rise.
http://sealevel.colorado.edu/
That 3.2mm/year is, at 360 cubic kilometres per mm, 1150 cubic kilometres extra volume annually. The increase is mostly due to to temperature induced expansion. None of the alternatives can be demonstrated to produce as much effect.
rhoda
"Now, where is the unprecedented stuff?"
The 48 contiguous United States recorded its highest annual average temperature on record in 2012.
http://www.ncdc.noaa.gov/news/ncdc-announces-warmest-year-record-contiguous-us
Australia has been quite warm.
http://www.bbc.co.uk/news/world-asia-21072347
The US, again, has been rather dry of late.
http://www.climatecentral.org/news/drought-expands-again-as-forecasts-show-little-relief-ahead-15024
I could go on, but I'd prefer to avoid a gish gallop.
Yes, while the US had high temps, we were freezing our asses off. The US is not much of the world, and the techniques involved in making that a record average temp are suspect. But let's not argue about that now. We need more frequent true extremes than that. We need to have past records stay the same, not decrease as time goes on. We need it to be warm where I am occasionally. Not much going on is still a valid intuitive summation at least for me. I guess we can wait and see. But on your side there is always going to be something to point at, where what we all see as the range of natural variation produces a result at some boundary or other. (Noticed your straw man of stable temps in there. Nobody is claiming climate is stable. Plenty of us are claiming it has been both warmer and colder in the past)
Entropic Man
I'm not sure the figures you refer to address the point head on.
This website plots changes in global temperature over a long period:
http://www.sjsu.edu/faculty/watkins/globaltemp1.htm
One conclusion is the following:
"It is notable that although the standard deviation of the temperature changes increased from 0.116°C/century to 0.129°C/century between the first seventy years and the second seventy years this is only an 11 percent change and not significantly different from no change."
It really doesn't look all that dramatic. Of course, the figures are only up to 2003, but a simple "eyeball test" shows that the year on-year-variation seems negligible.
"The 48 contiguous United States recorded its highest annual average temperature on record in 2012"
1934 ?
The rest of your thesis is 1984: "He who controls the past controls the future. He who controls the present controls the past."
Rhoda Klapp
The problems with the intuitive approach is its unreliablity. That's why you need data and statistics to analyse it. I live in Ireland, where the weather changes so often that we dont have climate :-)
Straw man?, I just neded a term for a a graph in which there was no short term change going on.
Alan Kennedy
I've done my share of statistical analysis down the years, but I made very little sense of your link. Using frequency distributions of changes between one year and the next seemed a very roundabout way of deriving standard deviations. Perhaps you should repost it on one of the statistics threads here and see if someone there can explain it.
I don't seek to convince anyone with intuitive assertions. As you say, that is not a valid argument. But then I don't seek to convince anyone of anything. Indeed that would violate my first precept, 'you can't tell anybody anything'. (I have others). I say not much is going on, and I set a high threshold on changing my mind. Unreasonably high? Maybe. But I kinda think an actual change of climate of any significance would be obvious. You and I would not need to argue over invisible trends or unprecedented things which have happened before. There is no unprecedented extreme event. Warmists representing things as extreme when they are merely unusual has rather an inverse effect on me. I wonder why they are lying. I wonder what they are at. In my world you don't lie to make any case, no matter how much you believe. If you are chief scientist at the MO, you set an example, you don't tell porkies to the pblic to get funding. Other ethical codes are available.
rhoda
We share an aspiration for a world where everyone can and does speak straightforwardly. If this discussion was about the production of heavy elements in stars there would be no difficulty, apart from the usual academic disagreements.
Unfortunately the science of climate change has become entangled with a variety of political, idealogical and economic interests, all producing propoganda, ad hominems and all the other tactics learned from years of political debate. Even the scientists have had to become politicised in order to function, retain funding and communicate their findings.
Ideas on how to dig ourselves out of this hole would be welcome. I would love to be able to discuss climate change purely as an academic scientific study, but here only a few people will engage with me on that basis.
Though it is more fun than just talking to people who agree with you. :-)
Entropic Man
It's not as subtle as all that. The link was to plots of changes in global temperature year on year. It is pretty obvious that the *rate of change* is not, itself, changing (or, if you want to be super cautious, it is not changing very much.) This is clear from just looking at the plots, but you can also compute the variance in the rate of change and do a t test betwen two chosen periods: that is what I reported. There is no difference in the slope. My point is this sits uneasily with claims about increasing rates of change (and more recent claims about increasing incidence of extreme events).
It would still be interesting to know whether the variance in global temperature has changed over the recent period during which the mean has not changed very much. Has it?
Alan Kennedy
"It would still be interesting to know whether the variance in global temperature has changed over the recent period during which the mean has not changed very much. Has it?"
I do not know a direct source of variance figures for climate data.
The Met Office publish 95% confidence limits for their recent annual global average temperatures, at the bottom of this page.
http://www.metoffice.gov.uk/research/monitoring/climate/surface-temperature
The last graph on the page, the overall global annual averages, includes 95% confidence limits for individual years and for the 5 year smoothed averages. Enjoy.
Entropic Man
Thanks - very useful. A few minutes with excel and you would be hard pressed to say the shape of the distribution has changed very much at all since 1999 - certainly nothing to get worked up about. Nobody could claim there has been a massive increase in variability on the basis of these data.
(It seems we are now alone in carrying on this discussion. It has been informative and interesting.)