A recent paper in PNAS by Hansen et al. (there’s also a recently released discussion paper on the topic) has caused quite a stir. The essential result is that extreme heat (beyond the 2-sigma and even 3-sigma level) has become so much more commonplace, that the only plausible explanation is global warming.
In a sense, this paper doesn’t tell us much we didn’t already know. What it does accomplish is to show in practical terms the observable result of man-made global warming, which is not just to make the average temperature hotter, but to make extreme heat so much more common. What was once 3-sigma heat — which at any given time we would expect to cover less than 1% of the globe — is now at least 10 times more prevalent. There have been 3-sigma events before, it’s true, and because of that we know that such extremes have consequences. When they’re as rare as they should be, life can recover from those consequences. When they’re 10 times more common …
I have two criticisms of this paper. First, the exposition is not always as clear as it could be — but that’s a matter of style more than substance. Second, it gives the impression that variability of temperature has increased recently. I’m not convinced that’s the case when considering local temperature because part of the increased variability in “standardized” (i.e., scaled by the local standard deviation) temperature anomaly is due to spatial rather than temporal changes — different amounts of overall warming in different locations (i.e., different trends) — as I stated here. I admit I haven’t analyzed hemispheric data, nor did I (in the previous post) consider seasonal (principally summertime) temperature specifically.
But in another sense, temperature variability has increased precisely because of spatial as well as temporal variability. The point of Hansen et al. 2012 is that what used to be rare extreme heat is now much more common. Much. This is made even more true by the fact that some regions have warmed (trend-wise) more than the global or hemispheric average, so they’re even more susceptible to extreme events (“extreme” by the standard prior to 1980).
Even hot times in earlier years don’t stack up to what we’re seeing today. In their more recent discussion paper, they show standardized anomalies (i.e., anomalies divided by the local standard deviation) for summertime in the northern hemisphere, using a longer baseline period than in the original paper (in response to some critics). Here’s the color legend (units are standard deviations):
Here’s the map for the very hot summer (in the U.S.) 1936:
Note the strong heating over much of the American midwest, with a small region even showing 3-sigma (or more) heat (dark brown color). It was hot back then in the USA, but only 1% of the northern hemisphere was in the 3-sigma or more extreme range. Now look at what happened in the summer of 2010:
Not only is there a region of 3-sigma heat along the east coast of the USA and another along the north coast of South America, there’s a giant area from Russia down through the middle east. Fully 18% of the hemisphere is in the 3-sigma range. That’s not “natural variation.” It’s global warming.
This is, I believe, an important way to characterize the simple temperature effect of global warming because it puts it in the context of what we’ve seen before, of the conditions on which we have based building our modern civilization. The baseline period for their recent analysis is 1931 to 1980. That’s when we layed out the infrastructure which drives modern high-tech civilization. Those are the conditions from which we derived our expectations. But because of global warming, conditions today exceed expectation much more often than they used to. Much more often than we’re prepared to deal with. Much.
I did a similar analysis for the lower 48 states of the US only, using summertime-mean data for climate divisions from the National Climate Data Center. I compared the distribution of standardized (i.e., scaled by the local standard deviation) temperature anomalies prior to 1980, to those since 2000 (using 1930-1980 as a baseline period):
The most important aspect of this comparison isn’t the higher mean value for the most recent temperatures. It’s the fact that extreme high values — above 2-sigma and especially above 3-sigma — are so much more frequent. Much. And that’s the real problem. When a 3-sigma event happens, it’s a problem but we can deal with it and recover from it. When 10 (or more) times as many 3-sigma events happen … we have a problem.
That means we’re already in trouble. The really bad news is that we’re already in trouble from just the warming we’ve already experienced, but it’s going to get worse because it’s going to get hotter. You think the 2011 Texas-Oklahoma heat wave was bad? You think this year’s corn-belt heat wave was bad? You think the 2010 Russian heat wave was very very bad? You ain’t seen nothin’ yet.
That’s why Hansen et al. is so important. From a purely scientific perspective it doesn’t really add to our knowledge. But from a human perspective, it lays it on the line. We’ve had bad heat events in the past but now they’re so much more common they’re vastly more difficult to deal with, so stop kidding yourselves, it’s already a bad problem and it’s just gonna get worse.
All this reveals the utter foolishness of Cliff Mass’s distorted view that global warming has little to do with the extreme heat witnessed in recent years in many places. His argument is that global warming has raised temperature in the U.S. by about 1 degree F, but last year’s Texas-Oklahoma heat wave was 7 to 8 deg.F over large portions of TX and OK, so global warming is only responsible for a small portion of that heat wave.
Even if his result were correct (which it is not), he utterly misses the point. Rather than sum up the situation the wrong way as he does, Hansen et al. did it right, showing that global warming doesn’t just make heat waves hotter. What’s much much much more important is that it makes heat waves more frequent. Cliff Mass gives the impression that there’s nothing to worry about because our “3-sigma” events — the real killers — will only be one degree hotter, quite ignoring the fact that we’ll get 10 times as many of them.
Shouldn’t we compare timespans of similar lengths? Does that have any influence on the frequency of extreme events?
[Response: The graph shows frequency of occurrence per unit time.]
“Second, it gives the impression that variability of temperature has increased recently. ”
Ah. I had missed your variability post from July 21 (that should teach me not to go on vacation) – I found it very interesting. I encourage you to publicize that result further: I consider myself a climate change professional, yet I had taken the Hansen et al. type graph at face value in terms of showing an increase in variability in addition to the mean, and I think your July 21 post is pretty convincing that this sort of skew would be expected from averaging across a region with differing trends (not that there might not be an increase in variability anyway, only that you have to do a more sophisticated analysis to detect it).
I’m wondering if your USA48 graph could be looked at in another way: the percentage of the country that is likely to be X standard deviations above the local mean? Eg, it isn’t that in any given location in the US, you’re much more likely to see a 3-sigma event…
Also, while I no doubt disagree with Cliff Mass’s conclusions, I do wonder: if we originally have 5 days of 3 degrees above normal, and 1 day of 4 degrees above normal in our base case, which is the better way to portray a 1 degree warming: a 5-fold increase in 4 degree events, or a 1 degree increase in what used to be 3 degree events? The former sounds much more disastrous than the latter, even though they describe the same situation… implying that there’s a subtle issue going on here. Perhaps we need to think about how non-linear the damages are: if there is a threshold, or high non-linearity, then the former (five fold increase) is a better way to think about it, whereas if damages are fairly linear with temperature then the latter (just add 1 degree to any historical temperature) gives an impression closer to reality…
Corn pretty much can’t grow (at least the current strains) if the average daily temp is above 86F. That is a distinct non linearity.
Corn can survive short exposure to temps around 112. Temps in Kansas this summer were above that. So, basically we’re courting with killing off a huge chunk of our corn planting.
MMM, considering that we’re moving away from an optimum, I think it is a fair first guess, assuming the damage function to be “well behaved” (i.e., smooth and a few times differentiable), that it will be parabolic about the optimum, i.e., a square function of temperature offset.
That is, if we don’t meet any complications on the way, AKA “feeling lucky”…
MMM, there is a definite non-linearity of damages brought about by the fact that homeotherms have an effective upper temperature above which they cannot survive because they cannot maintain their core body temperature because they cannot cool fast enough. As the environmental temperature approaches that temperature, the percentage reduction in cooling ability increases more with each degree C increase in temperature. Treating the upper temperature limit as 40 degrees C, an increase in temperature of 7.5 C will effectively halve the ability of homeotherms to cool themselves, but it will only take an increase of 3.75 C to halve it again, and 1.875 C to halve it a third time. Clearly I have simplified because humidity is a major factor for cooling for most homeotherms, so wet bulb rather than dry bulb temperature should apply, but the principle stands. This applies not only to humans, who in Western societies can at least find air conditioned locations, but to our live stock as well, who effectively cannot, and to homeotherms in the wild as well, who certainly cannot.
There is an additional non-linearities related to evaporation, and (no doubt) other factors.
What is more, Cliff Mass’s application of his reasoning is simplistic. He provides no evidence that the mean temperature increase during weather events that introduce heat waves is not greater than the mean annual increase. The mean night time, day time, summer and winter increases all differ from the mean annual increase. Given that, he is not entitled to an assumption of uniform temperature increase in assigning effects to either weather or global warming.
I would think that the biggest thing Mass is missing is the latitude. The one degree F is the global average, and I believe it includes the ocean.But land is warming more quickly than water and the higher latitudes more quickly than the lower.Global warming is going to hit you where you live.
Timothy Chase, Mass shows the June – August temperature data which shows a very low trend. Significantly less than the 1 F that he uses. Technically using his method he should have ascribed even less of the warming to global warming than he in fact does. Of course, his method is rubbish, so that is of no consequence.
the exposition is not always as clear as it could be
Wow, that’s unusual for an academic paper. :)
As planers and policy makers plan critical infrastructure, they have to plan for much higher temperatures, and that results in much higher costs. And they have to plan for those temperatures arriving much sooner than expected.
There goes all the cost savings from not fighting global warming. Everybody underestimated the costs from global warming. Everybody discounted the costs from global warming too much by placing the damages too far in the future.
Still, if there is a future left to plan for, we are going to need some numbers to use as a basis of planning.
I suggest that we need an estimate of the total heat stress that infrastructure or people or crops or organizations will be exposed to over a continuous period of 10 days. or 20 days. And it is very non-linear. A day in 100F heat is much worse than a day in 90F heat. Three days in 90F heat is much worse than a day in 90F heat. And, if you are going to be in 100F heat for 10 days, you need to stay hydrated. And if you are building infrastructure in those kinds of places, you need to plan for the heat.
If your only basis of planning is, “Much heat!”. Then, about the best you can do is a gravel road, and adobe buildings.
There are also nonlinear responses to high temperatures in plants. http://themidwestcultivator.com/wp-content/uploads/2012/02/Chart-Photosyntesis.jpg.png
Thanks for the article, Tamino. I think those maps of 1936 vs 2010 should be the “final nail in the coffin” of the “but, but, 1936!” denier argument. But I’m sure it wont be…
It’ll be interesting to see what the 2012 map looks like, particularly in the US.
What about sample size effects in your’s and Hansen’s graphs? For example, Is there a 12 year time period between 1930 and 1980 that would produce a similar bell curve shifted to the warm side like in the 2000 to present data? Could there have been similar periods some other time during the Holocene? How do we know for sure how much we’re cooking the earth? Can statistics answer this without the passing of more years and us waiting around until it’s too late to do anything more than wring our hands?
Another good post. It is informing to see your analysis of this paper. Your clear point that the key conclusion is rapidly increasing extreme events is important. I will be interested to see how the analysis of increasing temperature variability goes in the future.
Even more frightening is the appearance of 4 and 5 sigma events. In your graph above for the US, the 4 sigma events are obviously greater than 1% (by eyeball) and 5 sigma is greater than zero. In the period of record both of these were zero. We have not even reached 2C yet, which is the current “target” for no harm. What do you think the farmers in Texas think about 2C?
Thanks for this analysis of Hansen v. Mass. The only part I’d quibble with is this: “But because of global warming, conditions today exceed expectation much more often than they used to. Much more often than we’re prepared to deal with.”
If we could somehow keep the warming to current levels, I think we’d adapt to it reasonably well. As Stuart Staniford pointed out on his blog last year (i.e. before this years drought), global food production has kept increasing despite the Russian, Texas, etc droughts.
But we’re rapidly approaching a point where this will no longer be true, and a lot of it is already “baked into the cake.” Perhaps this will be the year it starts showing up in food production as well. Depressing thought.
Australian farmers are looking forward to higher prices this year on the back of US farming difficulties. Unfortunately we are likely looking at El Nino conditions – Australian agriculture tends to suffer badly:
“Huge US corn and soybean losses just forecast by the US Department of Agriculture are helping drive US wheat prices higher.
Record US corn and soybean prices stemming from drought-ravaged yields and a 12 to 13 per cent drop in forecast production versus last year are driving the sharp wheat rise.”
I’m not sure where you are getting your data from.
Staniford used FAO data on cereal production to show a steady increase in yields over the past four decades.
[Response: I’m skeptical.]
The Russian heat wave for instance had no discernable impact on the global FAO dataset. But there is a significant lag in when the data are reported, so we can’t say anything about 2012 yet with these numbers.
Basically there are two trends on a collision course: the “green revolution” continues to spread in less developed countries, leading to more intensive and productive agriculture. And due to climate change we are seeing a rise in extreme events that can harm yields. So far the green revolution is winning: global yields have been increasing, not declining.
But at some point in the future* climate change will win out.
* Or present. If it turns out that 2012 is the year (and it COULD be, the current US drought is definitely a contender to break the trend), then we should all be even MORE scared then we already are…
Re Staniford’s graph – it’s yield, not production, and yield is based on harvested, not planted or available acreage(hectarage?). I saw a report on the news from a corn grower in Nebraska. Some of his crop is irrigated(40%, if memory serves), and is doing pretty well. Some was dependent on rain, and it’s toast. He’ll report x bushels harvested on y acres, but not (x+0)/(y+z) acres. His reported yield may not be much different from previous years, but his production will be down more than 50%. If the drought extends into next spring, he probably won’t bother planting fields too dry to sprout.
@ Douglas. Staniford has charts of yield but his interpretation of the graphs show he (initially) doesn’t really understand what yield is (I suspect he learned something in the comments section and couldn’t back down). His conclusions would be correct if his data were of production. But they’re not. So he’s not. He tries to paper over this point by pointing to an increase in USA corn and extrapolating globally but this is a MASSIVE assumption.
If he wants to make a point he need to use the right data. I’m actually not saying he is wrong – but he has not demonstrated that he’s right.
Of course the devil is in the details and I am not sure Staniford checked them.
I may not have used exactly the same date series as Staniford but my graphs suggest little or no growth since the 1970’s except in three regions of Asia–exactly where one would expect cereal production to have increased with the ‘green revolution”.
Oops, link to graph is http://www.mediafire.com/view/?ibabn51k8vv5n73
There has been substantial growth in South America too, it just looks puny relative to Asia because they’re starting from a much lower baseline. There has been a steady increase in both gross and net South American agricultural production (total and per capita) in the past 50 years.
Also, the growth rate there looks smaller when you only look at cereals…it looks more impressive looking at total ag production.
Interesting, if a bit dated, article on Brazil’s ag growth and future capacity:
Click to access 02_06.pdf
Brazil is in a huge drought right now, and there is a good chance the rainy season won’t start on time.
This video doesn’t begin to show what the true nature of drought is like there. The cacti turn brown and fall over and there is not a bit of green to be seen. All there is is red dust coating everything and a sapphire blue sky — looks like a bomb destroyed everything
This kind of drought often occurs when ENSO neutral begins to turn into El Nino.
It is my understanding (I’m a policy analyst) that the green revolution is dependent on fossil fuels (for nitrogen among other things) and mined phosphorus. Both are at or near “peak.” What then? I see little consideration of this aspect of the global environmental crisis (for that is indeed what is occurring) on most blogs and news sites. This fact is why many who study carrying capacity say we are in population overshoot. It’s not just AGW that should scare us…
“If we could somehow keep the warming to current levels, I think we’d adapt to it reasonably well.”
I’m not so sure. All you need to look at is crop failures and successes. The top 5 wheat exporting countries/groups – USA, Australia, Canada, EU, Russian Federation – occasionally have major individual or combined failures. What happens when droughts and/or floods affect 2 or more in the same growing season … in the year following a major failure in just one of them?
And if we’re talking *major* crop failure, you can be pretty sure that the country/ies in question will also fail in one or more similar crops. World reserves of grains are already pretty low. Considering the crop impacts of just Russia 2010 and USA 2012 of extreme weather, we’re literally dicing with death if we think we can keep doing this without running headlong into the statistical inevitability that this climate regime will, sooner or later, allow multiple failures in one growing year.
One important detail outside the scientific aspect is that the baseline period Hansen uses also represents the climate a lot of our infrastructure is built against.
A lot of infrastructure, rail-lines, buildings and sewers etc. is not built to cope with what was then 3+ sigma events.
Welcome to the rest of our lives, (PS: expect old shit to break).
JCH asked on your view concerning publication of the Hansen et al paper in the earlier thread on this paper. It’s quite possible (likely) that he was referring to my comments at Judith Curry’s blog. I could introduce myself as a physicist who has spent significant effort on learning about climate. I’m definitely not a “skeptic” and could describe my position as agreeing in general with IPCC WG1 when read with the eyes of a critical scientist, i.e. I consider the caveats listed important and dislike conclusions that go beyond critically read WG1.
I don’t copy my most recent comment at Climate Etc here but give only
this link where I continue to argue against the point that publishing that paper was appropriate and present also more understanding for Cliff Mass.
I had a read of this Cliff Mass critique of Hansen et al. All he is saying (repeatedly) is that AGW has increased average temperatures by 1 deg F globally (less for the region of the heat wave under question) so if the heat wave in question was 7 or 8 deg F, only 1 deg F (or a small proportion of it) can be attributed to AGW. And this, he says, demonstrates that the conclusions of Hansen et al are false and worse (although “worse” is never explained).
Cliff Mass seems to think that, at any particular place in the world, AGW will result in non-AGW temperatures being raised by a fixed amount (ie in Texas 0.5-1.0 deg F). He admits some places will be warmed more than others (eg tha Arctic) but fails to understand that this “warmed more” phenomenon will also roam around. Sometimes it will be over Moscow and sometimes it will be over Dallas.
The more frequent existance of 3-sigma weather events demonstrates such events are not just weather. They are part of AGW. Cliff Mass is entirely wrong.
The scientific question is: Is Cliff Mass right or not?
You feel certain that he is not, but can you present evidence for that?
My scientifically unproven feeling is that he is not wrong to the extent that any available data could provide statistically significant evidence for concluding that he is wrong. As far as I know nobody has provided such evidence when the paper of Hansen et al does not either contain any valid statistical tests to support the claim – or can you give a link to such an analysis.
Warming has been observed and with warming it’s to be expected that exceptionally warm episodes are warmer than comparable episodes without the warming. All that is significant, but is there scientific evidence for anything more?
Pekka Pirilä – Is Cliff Mass right or not?
No, he is not. While a 1C increase in mean temperature means that the heat extreme of a 3-sigma event will go up by 1C relative to earlier values, what he is ignoring is that previous 3-sigma events will now be closer to 2-sigma events with the shift in mean, and occur 10-20x more frequently.
If a 3-sigma event from 50 years ago was bad, 10-20x per year of that event now is much worse – which is exactly the data that Hansen puts forth.
You say “The scientific question is: Is Cliff Mass right or not?”
If this is your belief, would you confirm that you see the substance of the argument presented by Mass is as follows: – The only contribution to an extreme weather event from AGW will be no more than the local average temperature increase measured at that spot over the period for which AGW has been acting, say since 1980. Perhaps more simplistically, Mass is not unhappy with this maximum value being equal to the global average. Either way, is this what you see as the substance of his thesis?
If you consider Mass is not saying this or is saying something additional to this, do say as it would be good to know what Mass is “right or not” about before examining its veracity.
There are two different ways of comparing extreme events before and after warming.
The first way is to look at the same event, i.e. at the same temperature. In this way of looking at the events the probability changes strongly like gets perhaps multiplied by 20.
The second way is to consider extremes whose likelihood is the same before and after warming. In this case the probability is unchanged by definition but the event is different. This is the way of Cliff Mass and his expectation is that the extreme after warming deviates as much from the new average temperature as the old extreme event deviated form the lower temperature. If the new average is warmer by 1 degree then also the new extreme is warmer by 1 degree than the old extreme, if Mass is right.
I will take your silence as a ‘yes’.
Let us begin by examining Mass’s evidence – the 3-month summer average Texas temperature 1895-2011 graphed here. This he presents as a gauge of Texan temperatures to show “the long term trend … associated with global warming” His exacting analytical method (eyeballing) concludes “…perhaps we can convince ourselves there was a small upward trend since 1970 of perhaps .5-1F.” It is most geberous of him to concede the point but, hang on, where is the 7 to 8 deg F heat wave. This graph only shows 5 deg F.
So by Mass’s “correct way” of analysing the impact of AGW within the 2011 Texan heat wave, Mass is 40-60% out.
But is his “correct way” correct? Are the temperatures of a heat wave additive? Surely it is plain bonkers to consider temperatures of heat waves in this way. 1 or 2 deg above average is not a heat wave. And if you are suffering a heat wave, the last thing you need is more heat. Would it not be the straw that is likely to break the camel’s back?
And is Mass’s “correct way” (your “second way“) even going in a useful direction? Mass’s graphic demonstrating how AGW will bring “more warm temperature records” actually demonstrates a massive increase in “extreme hot weather.” That is pretty much the same as saying “It follows that we can state, with a high degree of confidence, that extreme anomalies such as those in Texas and Oklahoma in 2011 and Moscow in 2010 were a consequence of global warming because their likelihood in the absence of global warming was exceedingly small.” which are the substantive conclusions of Hansen et al, conclusions Mass insists are false and worse.
Mass is a fool and you are a fool to write “The scientific question is: Is Cliff Mass right or not?”
You backed that assertion with an argument roughly saying that you feel Mass is not wrong because there is no proof he is wrong when Hansen et al do not present a complete and proper agrument. By implication you are saying Mass’s thesis is effectively valid. That is scientifically very foolish. If there are two wrongs, they do not ever make a right.
I would add that your defence of Mass’s “correct way,” a narrower argument than his entire thesis, would perhaps be understandable if it were presented as such. But that in no way can redeem the profound failings of Mass’s thesis.
This is not the approach to discussion I have.
Some of my other recent comments discuss the issue. You may consider them as my answer to you as well although each of them was directed to someone else.
[Response: I have found all your comments to be respectful, and to address relevant issues. I urge all readers to do the same.]
Pekka Pirilä – Yes, there are two ways to compare extreme events.
Damaging events, however, are what should really concern us. Previously 3-sigma damaging events (the Russian 2010 heat wave, European 2006, 2007 heat waves, etc) are becoming more and more common.
And Cliff Mass’s analysis of peak temperature shift, while technically correct, simply ignores the frequency of damaging events. I consider that misleading.
It seems to me that it is somewhat misleading to consider an extreme-value problem in terms merely of mean and standard deviation. As I said above, in the extremes Normal isn’t normal.
I think we have to be extremely cautious in assigning a probability to a particular extreme event. The probability depends on what the tails of the distribution look like, and the tails are highly uncertain precisely because the events from them are by defintion rare.
Likewise, if we look at the moments of the distribution, we will certainly see the shift of the mean/location parameter first, followed by the standard deviation. Any shifts in the skew or kurtosis would require a lot more data. This is an inherent problem in working with extreme-value stats.
You write “Some of my other recent comments discuss the issue. You may consider them as my answer to you as well…”
So let’s play hunt the answer.
What have we got down this thread? There’s stuff about assumed unchanging variability, a major fault in Hansen et al, about changes in variability, an unsupported prediction thereon, non-gaussian tails, that analysis is required to prevent us all guessing, the absence of contradiction argued by Otto et al, stuff about policy makers, skeptics believed by incredulous people.
No. I see nothing on this thread. Perhaps it’s hiding down that link you made to a different site, where you say you “present also more understanding for Cliff Mass.” . What do we find? Criticism of Hansen et al & that it should not have been published, problems with statistical significance, etc.
Ah, ha. Could this be your answer? “The paper of Hansen et al is supposed to show the opposite of of what Cliff Mass tells, but it doesn’t because an severely erroneous analysis cannot say anything on that until it’s corrected. Tamino probably agrees on this, but does not like the way Cliff Mass states his points.”
Now that statement is plain wrong. It is Cliff Mass who is supposed to tell the opposite of the Hansen et al paper but fails to do so. Does not Cliff Mass set out with the assertion (his emphasis) “Their conclusions are demonstrably false and their characterization of the science and statistics are deceptive at best“? Yet Cliff Mass is entirely unable to demonstrate any falsehood within Hansen et al. Plenty of “let me put it another way” and “let me prove to you now” but when it comes to the crunch, Mass presents an entirely vacuous argument.
Now what I cannot understand is this – how is it that any sensible person can defend this nonsense?
For those who do not follow Climate Etc., Pekka is sincere. He is not in any way, shape, or form a member of the the fake skeptics. Never. On Climate, Etc. he consistently writes comments that further understanding, and I greatly appreciate that he slogs through that mess over there. He’s a stout physicist, and his understanding is based on his ongoing, in depth study of climate science.
I can second that opinion, he’s one of the few people over there that a sensible person doesn’t immediately skip over. He persistently slogs through the morass and gets no love as a result …
This whole difficult discussion about attribution of AGW to a single event is precisely what Hansen et. al. is trying to circumvent by inverting the argument. Hansen is looking at the global picture while Cliff Mass is again trying to do single event attribution. He’s simply missing the point.
A loaded dice will throw more sixes, right? Hansen throws a lot of dice and knows that almost all the sixes above the normal statistical mean will be a result of the loading. Cliff Mass only looks at a single thrown 6 and wonders what the loading of the dice has to do with it and then claims Hansen cannot know either.
So Hansen sees extreme heat in increasing spatial extent and frequency because he looks at all the data, thus he is confident to say that the increase is essentially present in all events. Cliff Mass is only looking at one of those events and thus has great difficulty to attribute impact of the global change for this single event.
What’s next, are we going have to discuss precisely which iceberg or km2 of ice extent melted because of the Arctic warming or simply because it’s summer?
Nice article, clear argument of the dice for extreme events being loaded by shifting up the average of the gaussian distribution.
However, your plot raises an obvious question. It would be nice to say something about the width of the distribution too. Is the recent, red, distribution in your plot still exactly as wide as the older blue one? Or is there enough statistics to say that the width of the distribution has changed?
Thanks Tamino, a really good read.
Pekka Pirila, would you please explain to me the statistical and/or physical basis for assuming that the mean temperature increase during all types of weather states equals the mean annual temperature increase? That seems to me to be an essential assumption made by Mass, and to be entirely unjustified. You might also point out where he establishes that the frequency type of weather event he ascribes responsibility of the Texas 2011 drought to has not increased since the period of climatology. Why is his critique valid when it depends on implausible and unsupported premises which, so far as I can see, beg the very question at issue?
I don’t like often the concept of null hypothesis as it’s misused so often, but here I consider it really natural to take as null hypothesis the assumption that the shape of the distribution does not change much when the average gets higher. On this point the burden of proof is on those, who claim otherwise. The paper of Hansen et al made such a claim but as discussed by Tamino there’s a major fault the analysis. There fore the analysis cannot be taken to prove anything on this point until it’s done properly and the result found to persist.
I have proposed an alternative analysis and made a speculative proposal of what the result would be if the the result of the Hansen et al paper is, indeed, spurious.
That alternative analysis would be done reversing the direction of time. We would use the most recent decade as the bas period for all data and in particular for the average temperatures of each local measurement time series. My prediction is that this reversed analysis would show that the variability was larger in early decades than it was during the last decade or two. I.e., the variability is smallest for the base period for average temperatures and the larger the further we get from that base period. Such a result is, of course, spurious both for my proposed analysis and for the analysis of Hansen et al.
Somebody with the data readily available and proper statistical tools ready for use could check whether my guess is true.
If we take the case of a lognormal, merely changing the lognormal mean would also change the width–indeed this is the case with most distributions.
People need to realize: When it comes to the tails of the distribution, Normal isn’t normal.
Forget the null hypothesis. It isn’t real. It is a purely mathematical construct that owes its existence to the fact that most measures of the strength of evidence are comparative rather than absolute.
Rather than referring to “null hypothesis” I could have referred to what’s likely to be true when average temperatures rise a little – and around 1C is a little in this sense.
It’s to be expected that the distribution of local temperature anomalies is not exactly Gaussian and that the tails are somewhat fatter than they are for a Gaussian distribution. Log-normal is not a relevant alternative. (Log of what? Absolute temperature or something else.)
All the details are something that should be answered by scientific analysis. As long as we don’t have that each of us may have personal guesses. My guess is that the distributions shift maintaining approximately the same variance.
When deciding on policies proofs are not needed but the data and other knowledge as well as their uncertainties should be described as objectively as can be done. The decision makers must learn to consider the possibility of severe outcomes and the likelihood of such outcomes. Even if they are incapable in doing that the solution is not to claim that severe outcomes are more certain than thy are. Trying to cheat by exaggerated claims will backfire – and to some extent has already done that.
the two computations are not comparable. Hansen’s shows the effects (with both a temporal and a spatial contribution) of moving away from the climate we’re adapted to, the optimum. Your computation, while formally similar, has no such interpretation (or any I can see at all).
Peka Pirilla, day temperatures, night temperatures, summer temperatures and winter temperatures are all known to have a different mean increase than the annual mean increase in temperature. Given that, the supposition that meant temperature increases for different types of weather events all equal the annual mean is not a supposition you are entitled to as a null hypothesis.
Those distinctions are irrelevant for the argument.
Pekka Pirila, it is certainly possible that “persistent pattern of mid-continental ridging and troughing on the coast” in the 2000s show a mean increase in temperature double the regional mean relative to “persistent pattern of mid-continental ridging and troughing on the coast” in the period 1951-1980. If that were the case, Mass’s argument, which assumes the contribution of global warming to temperatures is the same regardless of the specific weather pattern, would come up with false results. Indeed, applying Mass’s argument to the 2010 Moscow heat wave underestimates the impact of global warming by 40% relative to the impact as assessed by a more reasonable approach used in Otto et al, 2012
I don’t have access to full text of the Otto paper, but from the abstract (emphasis mine):
This is exactly what I have tried to explain and what applies to the argument of Cliff Mass as well.
Pekka Pirila, if you now wish to rest your case on there being “here is no substantive contradiction” between Hansen’s approach of determining the change in return interval of temperatures of a given level, and Mass’s approach of determining the contribution of global warming as merely the difference in annual mean temperature at that location, then you must also agree that Mass’s claim that Hansen 2012’s “… conclusions are demonstrably false and their characterization of the science and statistics are deceptive at best” is in fact a false claim, unsupported by the evidence he provides. You must also agree with Tamino that Mass “… utterly misses the point”. One wonders then, why you have claimed that Mass’s analysis is a telling criticism of Hansen 2012, and that Tamino is incorrect to disagree with it.
You will notice that Mass takes exception not to the subtle issue of whether or not there has been an increase of variability (a claim he merely assumes to be false in any event). Rather he takes exception to the claim that the return interval of “extreme anomaly events” has greatly decreased – a direct empirical result of Hansen 2012, and something you now claim would not be contradicted by proof that global warming has increased the temperature of those events by just 1 degree, even if Mass had provided that proof.
None of this, however, has any bearing on the fact that Mass does not establish his primary lemma, ie, that global warming contributes no more to the temperature of any weather event than the mean annual increase at that location. He merely assumes it to be true, and you simply give him a pass.
You introduced yourself to this discussion with a claim to be a scientist. Isn’t it then about time that we saw some scientific reasoning from you?
For your viewing pleasure. The reason both papers can be complimentary is because they are asking different questions – How much v increased risk.
I see there is real logic problem as to the inability to see why Mass’ critique is poor. Otto 12 should help. One should ask whether or not Mass asked the same question as Hansen 12.
Tamino, first off a heartfelt thank you for your site. I’ve been following the debate around climate change for years now. I’ve always been (I hope properly) skeptical, and really, really hoped some flaws would be found in the theory. However, your site (and others I’m sure you know) have helped me to understand how comprehensive is the science, and how serious are the challenges. Even if the mathematics is sometimes a bit beyond me. Please keep up the good work.
I think Douglas’ point on cerial production is probably fair – there has been consistent growth in production over recent decades and it could be argued over the same period the price of cerials appears to be linked to increasing wealth in developing economies as well as total crop yield. However, this doesn’t detract from the point adelady makes (and I believe Douglas would agree) that the emerging climate makes the chances of serial, systematic failure in crop growth across key exporters much more likely. The potential ramifications are scary indeed.
I can’t find the link right now but Richard Muller was recently saying the same thing as Cliff Mass. Something along the lines of 1F increase in global warming just means that what normally would have been a daily high of 100F would just 101F with global warming.
Where do guys get this shoot first (as in shoot off his mouth) and then ask questions later?
Rob, it’s the “common sense” fallacy. That is, people apply the “common sense” in their areas of expertise to other areas. Sometimes it works, often you get burned.
The Muller quote here pretty much shows he isn’t really up to speed on extreme weather. “And the recent warm spell in the United States happens to be more than offset by cooling elsewhere in the world, so its link to “global” warming is weaker than tenuous. ” NYT Op-Ed 28/07/12
Yeah, AFAIK that assertion was completely unsupported. Global anomalies were distinctly positive in June (+.63, per NCDC) and July (+.62)… any ‘offsetting’ was clearly insufficient.
The main argument of Hansen et al that the frequency of high end extremes has increased (and will continue to increase) due to climate change holds true for a pure shift of a normal distribution (ie, the argument does not depend on any change in variance).
A 1 sigma shift in a normal distribution would be expected to increase the probability of what was previously (before shifting) a high-end (> 3 sigma) event by a factor of about 20. (see this reference, for example)
This 1 sigma shift in the distribution(s) is about what has occurred over the past half century (slightly more since the 1930’s) and the factor of 20 probability increase is of just about the right magnitude to have produced the observed increase in frequency of high end extreme (>3 sigma) events.
Consider the NH land surface since the 30’s:
(analyzed by by Hansen et al in the latest discussion of their paper
The observed increase from 0.4% to about 10% (a factor of 25) is quite consistent with the theoretical expected probability increase for a shift in a (otherwise unchanged) normal distribution of roughly 1-sigma (or a little more).
That a pure shift (with no change in variance) in a normal distribution produces an increased frequency of high end extreme events is hardly a radical claim and it’s actually rather amazing that anyone would really even question this idea (or at least would be amazing in any scientific context other than a discussion of climate change)
“Trying to cheat by exaggerated claims will backfire – and to some extent has already done that.”
That’s a very big claim. For there to be an “exaggerated claim” someone had to exaggerate. What claim. Who made it. Why would you say that the person who made the claim “exaggerated”?
[Response: More important: “exaggerated claims” is one thing, “cheat” is another — requiring strong evidence of deliberate deception.]
Perhaps I should have been more precise. When I referred to claims that have already backfired I had in mind more some claims made by other people than scientists.
Cheating may be a little strong word, but what I meant is a conscious decision to present facts in a way that’s not strictly true thinking perhaps sincerely that this will lead to a more correct interpretation by the audience. The presenter tries to figure out the bias in interpretation and compensate that with an opposite bias. Even this kind of benign “cheating” will ultimately be noticed (and we have in this case a lot of people who search for such cases). The result is loss of trust.
Simplifying the presentation by bypassing caveats is a variation of the same approach that may also ultimately backfire.
There are different levels of honesty. When issues are politically controversial the scientists should choose the extreme level of honesty to maintain their higher credibility.
“The scientific debate is closing [against us] but not yet closed. There is still a window of opportunity to challenge the science,”
“Voters believe that there is no consensus about global warming within the scientific community. Should the public come to believe that the scientific issues are settled, their views about global warming will change accordingly.
“Therefore, you need to continue to make the lack of scientific certainty a primary issue in the debate.” Republican strategist Frank Luntz.
By all means, let us discuss the caveats, the uncertainties. We know CO2 is causing warming, and polar amplification – from theory and confirming observtions. The decrease in tropical-to-polar gradient will slow the progression and increase the meanders of Rossby waves, which should increase the amplitude and duration of blocking and other weather generating extreme events. What is your estimate of the probability of all the extreme weather extremes we’ve seen in the last decade – killer heat waves, crop destroying droughts, deadly floods – occurring by random chance? Can you show statistically that the next decade won’t be worse?
You agree that the tails of weather extremes aren’t normally distributed, and human political behavior in response certainly isn’t normal – food price riots were a contributor to the “Arab Spring” overthrow of oppressive governments. Is there any way to show statistically that we shouldn’t worry about “black swan” weather events triggering nasty destabilization of civil societies?
US vice presidential candidate Paul Ryan accused climatologists of a “perversion of the scientific method, where data were manipulated to support a predetermined conclusion,” in order to “intentionally mislead the public on the issue of climate change.” Do you think that a rational discussion of caveats is possible with him, or will he spin it as “an admission of uncertainty about Catastrophic Anthropogenic Global Warming Alarmism” and use it to attack the very integrity you want to preserve?
Pekka, it sounds like your claim was exaggerated.
“Even this kind of benign “cheating” will ultimately be noticed (and we have in this case a lot of people who search for such cases). The result is loss of trust.”
I don’t doubt that this is true to some extent (perhaps you have lost some trust in this way) – but we have solid evidence that most of those inclined to react as you have described had no “trust” to begin with. Most people who find something to distrust were explicitly looking to support their starting orientation of distrust. In terms of overall impact, I would highly question whether the impact of blowback from “benign cheating” is anywhere near the magnitude of impact from deliberate efforts to undermine legitimate, and well-qualified science.
Ultimately, it seems to me that the there is no reliable alternative but to evaluate the science itself. I appreciate reading your efforts to do that. IMO, assigning motivation to someone that they are “cheating benignly” (without some direct evidence) is no more well-founded than calling them a fraud. It is inherently speculative, and unscientific.
I don’t think that I have personally lost much trust. I have for long been cynical enough but at the same time valued science highly. To me it’s natural that scientists are human and I’m not eager to condemn easily the morality of individual scientists. Rather I think that the behavior is not wise and that it, indeed, backfires given more time for the development. An activist gets things started but he may turn to burden at later stages.
I do know personally people who have no general characteristics that could explain their “skepticism” but they have most certainly been influenced by the observation of exaggerated claims. They are not severely misled by active skeptic movement and false accusations raised by skeptics but noting that misleading claims are made also from the other side has influenced them strongly. The people I have in mind have academic education in engineering, medicine or something comparable and they are generally interested in environmental issues and other major issues beyond their profession.
Since I have read your posts on climate change for a while now, I feel confident that you are telling the truth about the “blowback” that you have observed.
However, my points would be that: (1) what you are describing is not a very common phenomenon as compared to people who are basically looking to confirm their starting orientation. We do have some evidence in that regard and, (2) what you are describing is small as compared to the phenomenon of opinions being shaped by overtly politicized messaging based on “science” with no validity.
Yes – scientists are humans. From that general knowledge as well as personal experience of seeing scientists of all stripes gild the lily or overegg the pudding, it only stands to reason that sometimes scientific claims will be exaggerated. But that general reality does not allow us to conclude motivation in specific situations.To do that, we need hard evidence. Lacking hard evidence, it makes better sense, IMO, to focus on the science. With that said, I hope that someone takes you up on “reversing the direction” in analyzing the data.
“When I referred to claims that have already backfired I had in mind more some claims made by other people than scientists.”
Perhaps you should name names and identify the ‘cheating’.
“Perhaps I should have been more precise. ”
In Horatio’s humble opinion, you should not have said it at all.
Such unsupported claims have no place whatsoever in this discussion and are are actually very counterproductive, particularly given the propensity of some to use such statements to paint climate scientists as exaggerators, alarmists, cheaters and even frauds.
While you may not have intended your statement to refer to scientists, it’s not hard to envision that some might construe it that way, in a direct discussion of “Hansen et al” (especially given that one of the claims made by Mass about Hansen et al was that “conclusions are demonstrably false and their characterization of the science and statistics are deceptive at best”)
And even your statement “I had in mind more some claims made by other people than scientists” and your further “explanation” (which is still very hazy with regard to just whom you are talking about) leaves the door open in that regard.
For example, immediately after referring to “compensating with opposite bias” ( “benign cheating” you call it while others might call it “spin”) and “bypassing caveats”, you go on to say that “There are different levels of honesty. When issues are politically controversial the scientists should choose the extreme level of honesty to maintain their higher credibility.”
it’s really hard to see how this is not referring to scientists. It sure seems that you believe (at the very least) that certain scientists are engaging in “benign cheating” and “bypassing caveats”.
When such claims are made without providing specific examples — and specific evidence — like it or not, it gives the reader the impression that such things are not all that uncommon (or at the very minimum, that scientists need to be reminded that they should not do such things)
If scientists have lost credibility in this whole sorry saga, it is primarily because others (wittingly and unwittingly) endlessly repeat (without evidence) claims and innuendo (on blogs, on TV and in newspapers, etc) that scientists are exaggerating, cheating, lying, spinning, holding back or even hiding information, etc, etc.
It is all very unhelpful and makes the life of scientists who (by and large are a very honest and conservative group of folks) unnecessarily difficult.
With the tools of public opinion shaping (PR, Advertising, lobbying, tankthink) a single statement that’s mistaken or just open to misinterpretation gets deliberately inflated into claims of exaggerated and false claims – by politically astute and organised climate science denial. When the initial statement is made by someone with a high public profile and it’s taken up by a mass media with an existing editorial bias or simply a taste for controversy, the impact on public opinion can be far in excess of the significance to the overall body of knowledge about how climate works. Here in Australia an out of context quote by Tim Flannery about extreme drought and dam levels continues to be raised (within the context of recent la Nina and climate change influenced record extreme rain and floods) as ‘proof’ that climate science is profoundly and fundamentally wrong. A single climate scientist being wrong even once seems to have profound public impact but a hundred false claims by that politicised opposition seem to fail to register. There is little regard for higher motivations of accuracy or truth within the political opposition to climate action and it is not the science on climate that is their real focus, just the extent of public trust and acceptance of that science.
Whether insistence of absolute accuracy in advocacy of action on climate enhances public acceptance, or, by tying it’s hands with an excess of scruples, leaves it weakened in the face of opponents who have no such inhibitions is difficult to judge; I know I personally prefer adherence to accuracy but great crises of the past saw bipartisan agreement to suppress dissent and massage the truth in order to accept necessary sacrifices. I’m not sure they were wrong to do so – those crises being both real and desperate and unable to be successfully overcome with a nation divided. I don’t think that kind of “agreed on ” propaganda could be acceptable or even possible in this age but the bipartisan agreement part is an absolute prerequisite to public acceptance of the minimum actions needed for successfully overcoming this crisis.
It’s awfully easy to make accusations that a gullible public bent on not believing our growing understanding of how reality is evolving will gobble up whole. There’s a lot of it around.
If you are not part of the problem (which this vagueness makes us think you are) then spend your energy spreading more clarity instead of making vague quotable smears that will be picked up and spread around to further obscure. Others better informed than I have provided more information here.
Meanwhile, Frank Luntz no longer believes global warming is a hoax. He appears unwilling to retract his earlier work. (If you’re interested, google might help, or better yet duckduckgo, which doesn’t infest on behalf of commercial interests as much; this found there dates from 2006).
Comparisons with the heat of the 1930s are misplaced because agricultural practices of the 30s added to the heat. We learned and modified our behavior. (O brave old world that had such people in it.) Temps this year were hotter even with the improved agricultural practices. Had we continued to plow and plant the way we did back then the Great Plains would have been a goner.
Could you elaborate on that point Jeffrey?
1930’s plowing practices increased the soil desiccation from the drought. The diminished soil moisture helped increase the air temperatures.
Yes, a big part of the problem was the plowing practices, which led to increased topsoil loss and decreased ability to of the soil to hold moisture. I hate to point it out, but US Government Action helped to stem the crisis.
Time to stop apologizing for doing good and doing right!
— Horatio’s ruination of yet another beautiful song (sorry John and Paul)
Yesterday, all those heat-waves seemed so far away
Now it looks as though they’re here to stay
Oh, I believe in yesterday.
Suddenly, I see twice the temp I used to see,
There’s a sweat-ball hanging over me.
Oh, climate change came suddenly.
Why ice had to go I don’t know it wouldn’t stay.
I sang recovery’s song, now I long for yesterday.
Yesterday, denial was such an easy game to play.
Now I need a place to hide away.
Oh, I believe in yesterday.
Nice one HA.
Man! You did it again! Cut me to the quick!
I wish there were more recordings of your wonderful ditties. I can’t imagine any apology is needed.
One of these days, maybe Horatio will do a little ditty, ’bout Mac ‘n Deny-an.
Gotta learn to play the geetar, though.
Ain’t learned to play geetar yet, but for what it’s worth (prolly not much), here’s
“Mac and Denyann”
— Horatio Algeranon’s perversification of Mellencamp’s “Jack and Diane”
Little ditty about Mac and Deny-ann
Two contrarian kids sowin’ doubt in the Heartland
Macky’s gonna be a climate audit pro
Denyann denybutante blogseat of her own show
Playin’ up chili temps inside the winter freeze
Denyann doin’ a victory lap, missin’ the forest for the trees
Macky say, “Hey Denyann let’s run off an FOI or maybe three
Tee off them climate crooks, make ’em do what we please”
And say a
Oh yeah warming goes on
Long after the thrill of denyin’ is gone
Say, Oh yeah warming goes on
Long after the thrill of denyin’ is gone, now crock on
Macky sits back reflects his thoughts for the moment
Scratches his head and does his least global mean
“Well you know Denyann, we oughtta run off to the AGU”
Denyann says “Macky, you ain’t missin’ nothin'”
Macky say a
Oh yeah warming goes on
Long after the thrill of denyin’ is gone
Say, Oh yeah warming goes on
Long after the thrill of denyin’ is gone, now crock on
So let it crock
Let it troll
Let the Hockey Team come down
And save my soul
Hold on to short trends as long as you can
Changes come around real soon
Make it warmimg again
Oh yeah warming goes on
Long after the thrill of denyin’ is gone
Say, Oh yeah warming goes on
Long after the thrill of denyin’ is gone, now crock on
Little ditty about Mac and Deny-ann
Two contrarian kids sowin’ doubt in the Heartland
No worries about the guitar for this tune, Horatio–the only really indispensible element is the hand-claps.
Kudo’s to Horatio for the inspiration, and many apologies to Roy Clark:
Yesterday when man was young
The taste of life was sweet like rain upon his tongue.
Man teased at life as if it were a foolish game,
The way the evening breeze may tease a candle flame.
The thousand dreams man dreamed, the splendid things he planned
He’d always built to last on weak and shifting sand.
Man lived by night and shunned the naked light of day
And only now he sees how the years ran away.
Yesterday when man was young
So many happy songs were waiting to be sung,
So many wild pleasures lay in store for thee
And so much pain his dazzled eyes refused to see.
Man burned so fast that oil and coal at last ran out,
He never stopped to think what life was all about
And every conversation they can now recall
Concerned itself with men and nothing else at all.
Yesterday the moon was blue
And every crazy day brought something new to do.
Man used his magic age as if it were a wand
And never saw the waste and emptiness beyond.
The game of love man played with arrogance and pride
And every flame he lit too quickly, quickly died.
The species man relied upon all seemed to die away
And only he’s left on stage to end the play.
There are so many songs of men that won’t be sung,
Man feels the bitter taste of tears upon his tongue.
The time has come for him to pay for
Yesterday when man was young…
sad but brilliant
US farmers have all moved from plow based systems to min til or no til. Much less soil disturbance and vastly improved moisture conservation. The fact that crops are there at all is a tribute to that technology change.
Since this is about extreme heat events – and so leads me to wonder about how it would be best to measure the heat, I (as a layperson) have some questions, given that science is in a very serious political war: To add to the usual graphs based on the dry-bulb temperature record, has anyone in climate science given the world such additional information as graphs showing a global wet-bulb temperature record or a global heat index record, to try to perhaps more usefully or accurately measure the enthalpy (the total amount of heat energy) in the atmosphere, this extra information perhaps eventually (if not already) showing a greater increase in the total heat energy of the system compared to what we see with the usual graphs based just on the dry-bulb temperature record?
What I’m trying to get at: Since different amounts of H2O in the air mean that equal summertime dry-bulb temperatures in places like Florida and Arizona do not yield equal enthalpy, if we were to see the long-term global dry-bulb temperature increase slow down, then could it not be the case that although the deniers would crow about this the total amount of heat energy in the atmosphere could continue upward unabated due to increasing evaporation and ensuing amounts of H2O in the air? And that this could be seen with such global wet-bulb or global heat index graphs? And that this extra information could be used to at least try to shut up such deniers or at least help educate the public?
To get a more visceral feel for what I’m trying to say, play around with some inputs at this online heat index calculator:
“Meteorological Conversions and Calculations Heat Index Calculator”
Entering 104 degrees F and 75% relative humidity for some location gives a heat index of 171 degrees F, which would kill in hours essentially every human and land animal in that location living on the surface (and of course not in air conditioning) that needs evaporative cooling to survive.
With this in mind, it might be appropriate to recall that National Academy of Sciences paper that warned about future killer – even extinction event – wet-bulb temperatures or heat indexes (they were not talking about killer dry-bulb temperatures):
“The Health Effects of Hotter Days and Nights”
Quote: “”Most people are more familiar with the heat index, or the feels-like temperature they see on the weather report. The wet-bulb temperatures we are talking about would have a feels-like, or heat-index, temperature of between 170 to 196 degrees Fahrenheit,” Huber said.
“Researchers find future temperatures could exceed livable limits”
Quote: “”The wet-bulb limit is basically the point at which one would overheat even if they were naked in the shade, soaking wet and standing in front of a large fan,” Sherwood said. “Although we are very unlikely to reach such temperatures this century, they could happen in the next.””
I find one of the leading factors in uncomfortable/unlivable ‘feels like’ temperatures to be night-time cooling – or the lack thereof. Hot, sticky days, such as you describe with wet-bulb temps face the pseudo-skeptic ‘its been hot before’ argument. But a changing trend in the number of truly oppressive nights – heat and humidity that hangs on well after dark and hits you in the face when you open the door at 6am – is news.
A nice demonstration of how this is in fact changing is at the European Climate Assessment Indices of Extremes site. For example, the index ‘TN90p’ (a proxy for the number of warm nights per decade) shows a distinct increase during the summer half of the year for the period 1979-2011; compare that to the summer half for the period 1951-1978.
I find comparing a continent-wide map that’s colored red to one that is green/blue makes a strong visual statement. However, trend graphs and data downloads are available.
Thanks for the links. One important thing to keep in mind when comparing the two: the big circles for 1951-1978 are for 12 days, the same sized big circles in 1979-2011 are for 30 days. The images downplay the visual impact.
Speak of the devil:
The shifting probability distribution of global daytime and night-time temperatures
Recent observed warming is investigated using global PDFs
PDFs systematically shift and skew towards the hotter part of the distribution
In most regions extreme temperatures increased more than mean temperatures
Markus G. Donat
Climate Change Research Centre, University of New South Wales, Sydney, New South Wales, Australia
Lisa V. Alexander
Climate Change Research Centre, University of New South Wales, Sydney, New South Wales, Australia
ARC Centre of Excellence for Climate System Science, University of New South Wales, Sydney, New South Wales, Australia
Using a global observational dataset of daily gridded maximum and minimum temperatures we investigate changes in the respective probability density functions of both variables using two 30-year periods; 1951–1980 and 1981–2010. The results indicate that the distributions of both daily maximum and minimum temperatures have significantly shifted towards higher values in the latter period compared to the earlier period in almost all regions, whereas changes in variance are spatially heterogeneous and mostly less significant. However asymmetry appears to have decreased but is altered in such a way that it has become skewed towards the hotter part of the distribution. Changes are greater for daily minimum (night-time) temperatures than for daily maximum (daytime) temperatures. As expected, these changes have had the greatest impact on the extremes of the distribution and we conclude that the distribution of global daily temperatures has indeed become “more extreme” since the middle of the 20th century.
An interesting note on extreme’s that seems relevant to the discussion. I will leave it to our moderator to supply the probability/standard deviation numbers for this (should he choose to of course).
Yesterday. Needles, CA 118F
It rained. The rain was measured at 115 F A world record.
The humidity at the time of the rain was 11% also a world record.
The previous record was set on 5 June, 2012 in Mecca, Saudi Arabia at at 109 F.
Previous to that the record was 10 July, 2010 Marrakech, Morocco on 10 July, 2010.
Considering how long weather records have been kept.
How likely would this be if Hansen was not correct?
And I still can’t get my drought work published…
The following is from
“Statistics of Extreme Events with Application to Climate”
Incidentally, the authors of that study also analysed the global mean temperature record of Jones from the1850’s through 1988 and determined that after a linear trend had been removed, the residuals looked very much a normal distribution including on the tails:
what is the best way to send you a confidential message?
[Response: Place it here. If you note that it’s confidential, then I won’t post it on the blog.]
I’m confused. The latest global temp anomaly from GISS, July, was 0.47 on this table
which uses 1951-1980 as the base. This is something like the 14th warmest (I only did a rough count), according to those anomalies, and yet, when they posted the report here
for July, they said:
“The average combined global land and ocean surface temperature for July 2012 was 0.62°C (1.12°F) above the 20th century average of 15.8°C (60.4°F). This is the fourth warmest July since records began in 1880.”
So, using 51-80 as the base period makes it NOT the fourth warmest July since 1880?
I’m not a mathematician (I play a Philosopher on the internet), but it seems to me that if it was the fourth warmest July since records began, then it would be the fourth warmest July no matter what you picked for your base period.
[Response: Yes. However, GISS and NCDC are different data sets from different organizations. Still, I’m surprised they’re that different. I’ll take a look at the data.]
Dur. Sorry about that. But thanks for checking, as I’d be interested to know what accounts for the big difference.
Nick Stokes has a 4-year comparison. Scroll down to see it.
And I’m confused. The July 2012 GISS temperature presented down the link given by Scott Supak (@ssupak) is ***** ie yet to be published.
Has the figure been ‘unpublished’ or was the link given persumptive?
(Recent NCDC/GISS differences graphically – (Usually 2 clicks to ‘download your attachment’) Over the last year the largest difference was about 0.1 deg C.)
Al, Nick has them all, July, except HadCrut3 on his site:
HadCrut – NA
GISS – .47
NOAA – .619
UAH – .28
RSS – .292
Thanks for the nod. July is now up on the GISS site, all official-like. I just wondered/worried that with the link to the official site given but with no July posted whether there was something up.
Re the difference between GISS & NCDC.
I did do a quick plot of NCDC minus GISS to see how much variation there was between them. It appears a 0.15°C difference is not at all uncommon. Temporary graph (Usually 2 clicks to ‘download your attachment’)
Somebody speculated that the GISS July drop could be related to the amount of energy going into the melting of ice in the arctic. Is that plausible?
I’m surprised NCDC has a July Highlights published at all, given how much coverage they are still missing at the moment:
They are missing most of Africa, most of Canada and the Arctic, large chunks of South America, they’ve got about half of Asia, ….
The highlights look premature to me. It may be NCDC has been doing the highlights with this little coverage all the time, but I’d rather wait a month or two for more data. Do we have any idea on what data coverage the GISS data for July is based?
I could not find what the actual temp of one standard deviation is for the period of 1951-1980. I could not find it reading through the Hansen article. If the standard deviation for the global temp over this time period is small, then 3 sigma above this would be expected with a 1 F increase in global temps since this reference frame.
On the GISS page Hansen did write that a 1C temp difference in the Pacific ocean would be a 3 sigma event because of the low variability of the data.
Talking about “benign cheating”–anyone who has participated in the peer review system knows that statements not well supported by the data and even minor exaggerations are shot down very quickly. Journalists and environmental groups often exaggerate and journalist often show “misplaced skepticism.” However, it’s unfortunate if opinions against science are based on how it is reported by journalists and environmentalists.
What do you mean by “environmentalist”? It seems you’ve fallen for the PR talking point that attempts to silence those of us who have been around forever, promoting the idea that stewardship rather than exploitation is a better model for survival. Voices are silenced for fear of sounding “extreme”. It’s a tactic rather than a truism.
I don’t claim that no environmentalist ever espouses illogic, witness the anti-vaccine thing and some other excesses, but if they could have silenced “Silent Spring” with accusations of extremism, they would have. Oh wait … they did, and still are.
Gol danged, exaggeratin’, cheatin’, tree huggin’, polar bear loviin’, syphilization threatnin’ envarmintalists!!
You forgot commie free love pinko, if I remember correctly.
And it’s 1 2 3 what’re we fightin’ for
Don’t ask me I don’t give a dam’
Next stop is Vietnam
5 6 7 open up the pearly gates
ain’t no time to wonder why
whoopie we’re all goin’ to die
JCH: I’m quite convinced that it is attributable solely to the remarkably negative temperature anomaly over Antarctica (south of 55°S – except the Antarctic Peninsula). You may check the spatial distribution of the monthly GISS anomalies here: data.giss.nasa.gov/gistemp
This might also explain the difference between NOAA and GISS …
In an earlier post on Rabett Run, what was most important to Eli was the increasing asymmetry to the hotter side (warmer would be wrong in this case). That means more outliers even beyond 3 sigma. We are indeed skewing ourselves.
Given that the shape of the density curve is determined by the standard deviation of the sample (assuming normal distribution), wouldn’t it be more informative to plot how stdev evolves over the historical record?
When I do this for 11yr moving periods on USA48 detrended anomalies, the recent summertime variability looks rather ordinary. The periods centered between 1960 thru to 1980 look persistently low. The post-1980 periods have an upward trend, but look normal compared to the pre-1960 periods.
I accept that there has been a shifting in the mean, but arguing increased variability using selected periods’ density curves is vulnerable to charges of selection bias.
Hansen et Al 2012 is historical, and does not project into the future.
We are seeing a rapid loss of Arctic sea ice with resultant rapid changes in NH albedo and atmospheric water vapor. Any expectation of the future, must include what is going on in the Arctic now.
These sudden and extreme changes in the Arctic will cascade southward and result in extremes in our weather that are not discussed in the paper.
Well, you termed my comments “utter foolishness” so I don’t have much to lose continuing the thread. I am NOT saying the global warming has “little to do” with recent records…but that natural variability is more important at this point. Is it not clear that a warming trend will greatly increase the number of 3 sigma occurrences if you use the same, earlier base period? Yes, there are a LOT more 3 sigma events. But without global warming we would have still have big events…perhaps 2.5 or 2.7 sigma…you know what I mean. So the dominant cause of the anomalies is natural, with an assist from GW. Why is this interpretation wrong? Can someone please explain my error?
Did you even read Dr. Hansen’s paper, or at least the second one on this topic? Do you not see that the curve is skewed to the right? Do you know what skewed means? What part of 10 times more likely do you not get?
How many sigma events covering what percentage of the globe?
And yes, the numbers do not change using a larger baseline period. It was a good question tho.
Click to access 20120811_DiceDataDiscussion.pdf
Am I right in saying that if the occurence of “more extreme events” is due to a shift in the probability distribution, then the actual supplementary events (those which shouldn’t have beenn classified as “extreme” if things would have remained the same, but that have crossed the border because of the shift) are entirely concentrated in the interval between the “old” border and the “new” border with the same relative probability (for instance the “old” three sigma border and the “new” three sigma one) ? because all the events above the “new” three sigma events would have been statistically as probable as the “old” three sigmas events, so would have been classified as extreme as well, albeit with a smaller amplitude ?
thus the real extra events are actually not the worst ones, but those “just above” the limit of what we consider as extreme events, i.e. the smallest amplitude extreme events.
Sounds ‘overthought’ to me; if we adopt a new definition of ‘extreme events’ as the distribution shifts, then there won’t be more!–or at least, not necessarily. The corollary to that would be that we are implicitly staying with the old definition of ‘extreme.’
Kevin, that’s exactly what I meant. If you shift the limit of “extreme events” , keeping for instance the “above 3 sigma limit” with the NEW sigma, the number of extreme events remains the same (much like the number of poor people cannot decrease if they are always defined as the ten percent with the lowest income). So measuring an “increase” of extreme events can only be done if you keep the old limit constant – what Hansen did if I understood correctly. But then the “extra” events are concentrated in the lowest part of the distribution, between the old and the new limit. Note that when records are beaten, it’s almost always by a “very tiny amount”, in accordance with what I’m saying.
Well, sure–that’s ‘normal’ (as in ‘normal distribution’–if I may be forgiven a little wordplay.)
Hah! And turning to Gavin’s RC post on Hansen et al (2012), what do I read but this?
Let’s try a different approach.
Heat waves have an origin. They are the superposition of a background field and the effects of the upper atmospheric wave pattern. Generally heat waves, like the ones in the Midwest this year and last, are usually associated with high amplitude and persistent ridging. Now it is clear that GW can increase the background field—but the amplitude so far in the midlatitudes is modest (0 to 2 F perhaps, considering where you are). More in the Arctic. The background does not explain the amplitude Midwest heat waves. But there was clearly high amplitude and persistent ridge during the recent events. Now the question then becomes—is the excessive ridging the last two summers the result of GW? Was the large scale flow pattern somehow enhanced by GW? There is no reason to expect this…either from modeling or theoretical results. So the conclusion we must draw is that these events were mainly from natural variability.
[Response: On the contrary, there is theoretical reason to expect it. Arctic amplification and the loss of Arctic sea ice (both rather undeniable) decrease the equator-to-pole temperature gradient, which affects at least the persistence of heat waves. The presentation by Jennifer Francis is enlightening on this point.
The observed fact is that extreme events are much more prevalent than before. It strikes me as naive to contend that global warming is cause of the increased frequency but not of any event.
I’m preparing another post on the issue (look for it within a few days), which may be a good forum for more detailed discussion.]
Please watch the Francis presentation linked by tamino, and then tell us if it changes what you think about Hansen’s statistical analysis. The recent paper on the subject is at http://scholar.googleusercontent.com/scholar?q=cache:8Qh4niPATxsJ:scholar.google.com/&hl=en&as_sdt=0,48 (abstract and access to pdf through pay wall at http://www.agu.org/pubs/crossref/2012/2012GL051000.shtml)
I don’t understand why you are so committed to a position that doesn’t appear to 1) address the same question as Hansen is asking, or 2) incorporate facts easily found, e.g., work by Francis et al. I’m not a scientist (just a policy analyst) but that’s the impression I get from reading the back and forth here, at RC, etc. In short, after all the foregoing dialogue and follow up, do you agree with Mr. Watt’s (WUWT) conclusion: “Their (Hansen et al 2012) conclusions are demonstrably false and their characterization of the science and statistics are deceptive at best.”
To those of you quoting the Francis and Varvis paper, I should say that I believe that work has substantial problems. It only considers lower/mid tropospheric flow, and there really is no indication that the main jet…centered considerable higher has slowed down…in fact, because of upper tropospheric warming in the tropics, it probably will not…or even strengthen! In addition, GCM experiments in general do not show increased blocking under GW. The bottom line is that there is no reason to expect more blocking under global warming. Perhaps new results will change that story…but at this point, that is where we are. That quote is not Watts…it is mine. Hansen’s conclusions ARE deceptive at best and that is coming from someone who believe the case for human-induced GW is a strong one….cliff
Nice Ad-Hom, Cliffy.
If all you have is a hammer, everything looks like a nail.
The conclusions of the Francis and Varvis paper start (emphasis mine)
The next paragraph tells:
Should we take seriously their reservations or dismiss them and conclude that the scientists do not dare to tell what they really think?
I do seriously believe that the problem in getting stronger climate policies implemented and maintained in future cannot be solved by overstating the certainties and risking the possibility that many of the particular claims turn out to be erroneous. The basic case for accepting that CO2 causes significant warming and that this is a serious concern is strong. The policy arguments must be based on the most solid facts. It’s better to argue more strongly that certainty is not needed for action than to present unjustified claims on the level of certainty.
Stating that the recent extreme events are of the nature that many climate scientists expect to become more common is alright but claiming that we have scientific evidence that variability has increased is not alright until the evidence passes all requirements generally set for science and statistical significance.
It’s a real dilemma that obtaining statistically significant evidence may take too long but if that is the case, it cannot be solved by erroneous analysis or unsubstantiated claims.
A commenter on Cliff’s blog asked if Francis, Vavrus 2012, http://marine.rutgers.edu/~francis/pres/Francis_Vavrus_2012GL051000_pub.pdf, didn’t address a potential mechanism. Cliff did not respond to this commenter but when I pressed him on it in an e-mail he stated;
“Francis and Varvus has a major flaw….they only considered lower and middle tropospheric flow. The jet stream above is not weakening and thus their underlying mechanism does not make sense.”
I queried Dr. Francis on this. She did not have time for a thorough reply but did say that “. . . we are not the only ones who have demonstrated a weakening of the upper-level flow — true, we looked at 500 hPa flows, but others have looked at higher levels and found a weakening in fall and winter that was attributed to a weaker poleward temperature gradient. It should also be noted that we discussed only the zonal (west to east) component — it could very well be that the jet overall is not weakening, and in some areas strengthening, as the land heats faster than the ocean and increases the thickness gradient in the east-west direction (leading to stronger meridional winds).”
She also pointed me to “Explaining Extreme Events of 2011 from a Climate Perspective” Peterson, Stott and Herring 2011, http://journals.ametsoc.org/doi/abs/10.1175/BAMS-D-12-00021.1
I haven’t had a chance to absorb this yet but it seems pertinent to the current discussion.
You will find that Mass has already been through that collection of papers you link to & called out Ropp et al for being “flawed … scary … sexy” and “most worrisome”. His counter argument is rather familiar. Yawn.
— by Horatio Algeranon
“Hansen’s not deceptive
Only his conclusions.
This is not invective
Only word confusions.”
That’s not Watts. That’s Mass’s own wording
Due to events in the Arctic, and work around that, I’ve been too busy to read this paper. But I did read the whitepaper on which it seems to be based. Frankly I think this work represents the most frightening observationally based work on AGW I’ve yet seen.
FWIW here were my thoughts back in November:
I note with regret that Prof. Mass continues to characterize Hansen as deceptive. As for his claim that the signal is swamped by ‘natural variability’, clearly he has not seen or not understood Fig. 4 in Hansen, or Fig. 4 in the post publication discussion where the analysis is extended to the 1931-2011 period. For crying out loud, approximate Gaussian has shifted 1 sigma to the right.
Given his continued vilification of Hansen, I feel no compunction in posing the question: Is Prof. Mass a liar or a moron ? A perusal of his work indicates at least minimal competency, so I incline to the former.
[Response: I agree that his continued characterization of Hansen as deceptive, is out of line. But I think the “liar or moron?” statement is also. Let’s not burn bridges until we’re certain we won’t want to reach the other side.]
I am disappointed the moderator would allow such name calling. I have surely seen Fig. 4 and of course there is a shift to the right of the Gaussian. So what? I have not called Hansen a deceptive person…but that his statistics were deceptive. There is a big difference folks. Anyway, I think I will end my participation on this blog…cliff mas
[Response: I called him out on his name-calling. I also called you out, because your many repetitions of “deceptive” (not just here but in many places) are downright ugly. Your miserable appeal to “his statistics” rather than “Hansen” is nothing but sophistry. Have the courage to back up your accusation — running away from your own words is cowardly.]
Um, no. Your words were:
Goalpost shift/misdirection ploy by you, sir. Hoist by thy own petard, eh?
And that is coming from someone who believe the case for human-induced GW is a very strong. Strong enough to withstand agenda-driven character assassinations…
He’s a bit thin skinned, that Cliff Mass.
And I would suggest there is a distinct whiff of hypocracy here. Okay, I have not seen him using the ‘moron’ word, but he has used the ‘L’ word. This was posted on his blog 29 August and aimed at those who ‘hype’ the “relatively small impacts during the past few decades that they can motivate people to act. A lie for a good cause.”
Mass does not accuse Hansen et al of ‘hyping,’ rather of presenting “exagerations” and “much worse.” And if that does not constitute a trail of proof, in a post of 15 July Mass does directly accuse Rupp et al 2012 of “hyping” and thus ‘lying.’
So now I’m looking for a word to describe someone who thinks we wouldn’t notice him calling folk ‘liar,’ a word beginning with M perhaps?
Cliff, I am curious why you would think that standard deviation is “deceptive” while average is not. Both are summary statistics, are they not? As such, unless we know the actual distribution, they can tell only a small part of the story. Also, it is the extremes that are most damaging (e.g. the Missisippi at record lows halting barge traffic), so the 3 sigma and above events are quite relevant.
“The Normal and the Deviant”
— by Horatio Algeranon
An average is the normal
Which normally won’t deceive
But radical deviations
Are not to be believed.
That’s deceptive, at best (and a potential source of mass confusion)
If we cut through all the tosh, Cliff Mass is really making three points although none are as clearly made as they could be.
Firstly, the natural variations of weather are large but AGW (however large they are at present) will become larger. I would suggest that is not good news. Whether or not the impacts of AGW can be discerned within the ‘noise’ of weather, however tenuous (or obvious) the evidence, that nature’s worst tantrums will eventually be bested by those induced by AGW only puts Texas 2011 (& other extended +3 deg C events) into context.
Secondly, if the variance of temperature remains constant, an assertion for which Mass contends that there is no evidence to prove otherwise, it would mean that the severity of future extreme weather events will only increase by the average level of regional/global warming. Given Mass’s first point, a compelling conclusion is required here, not a flippant one. Mass is wrong to assume ha has found a trump card in this matter. He has not. Even if today extreme weather events are only being increased by average AGW, will this remain so in future as the weather is pushed into new configurations unseen for millions of years? Will this be so even in the short term?
Thirdly, the method by which an extreme weather event can be created can be explained in terms of phenomena that appear as natural weather. In basic terms, this is a surely a matter of physicis. If the natural mechanisms of weather were so compromised that this were not so, if new phenomena weather were discernable, the case for severe AGW would be rediculously obvious and this discussion thread would not need to exist.
OK, here’s something I don’t understand. If there is another side in this debate, Cliff Mass probably represents the best of it. He doesn’t deny science, but rather argues that the science doesn’t point to as bleak a future as others contend.
So why oh why, when he comes to comment on this blog do you insult him (and yes, accusations of dishonesty, hypocrisy, etc. are insulting) rather than engaging him constructively?
It seems many of you guys are much more willing to spend your time on the freakshow over at WTFUWT than you are to engage in thoughtful discussion with an intelligent professional who happens to disagree with you. I wonder why that might be.
[Response: I wanted to engage in exactly that manner. That’s why I called out “sidd” for his insulting comment.
But let’s not pretend that Cliff Mass is innocent. He has repeatedly, in many places, referred to Hansen’s work as “deceptive,” and his protest that this is not an insult rings hollow. In my opinion, his harping on that point even after its insulting nature has been pointed out, is strong evidence that it’s intentionally insulting. His statements about the substance of the argument I find unpersuasive, and he has a habit of responding to evidence with a hand-waving “such-and-such is deeply flawed” but neither evidence nor logic to back it up.
After his repeated insults of James Hansen, when Mass is insulted he flounces away indignantly. Looks like petulant hypocrisy to me.
You want dialogue? That takes two. It seems Cliff Mass is more willing to spend his time insulting James Hansen on the freakshow over at WTFUWT than to engage in thoughtful discussion with us. I wonder why that might be.]
Tamino, Don’t misinterpret what I said. I am not defending Cliff. Rather, I am lamenting that the kittens let him escape before the mother cat(s) had a chance to play with its prey.
[Response: I agree that that strictly sticking to the subject would have increased the chance of a productive discussion.
I’m no stranger to snark, but we shouldn’t let it sabotage genuine dialogue. So to everybody — don’t burn your bridges until and unless you’re certain there’s no prospect to reach the other side.
That still leaves plenty of bridges to burn.]
Having read Mass’ blog, and seeing the repetitive assignations there and here by Mass regarding the supposed “deception”, it becomes quite clear that Mass begins his analysis of Hansen with the presumption of malfeasance and then proceeds to prove it (at least in his own mind). Like others have said, I find his arguments unpersuasive, even notwithstanding the pervasive character assassination he perpetuates.
Had he come here in good faith with an open mind then I would have readily accorded him full measure and opportunity to convince me.
Another Texas sharp-shooter, huh?
Although I have a strong suspicion that someone’s already made that observation…
Snarkates (& Tamino),
I don’t think any mummy cat would have been that happy being woken up just for a play with Cliff Mass. IMHO, his contribution here in 4 comments was very thin & rather uninteresting.
(A) Small mean dT shifts brings lots more extremes but not great shift in Tmax. ‘Is there something I’ve missed?’
(B) The weather behind these extreme events has no reported AGW linkage so the conclusion must be drawn.
(C) F&V’s linkage isn’t valid but perhaps new data will change the conclusion. Hansen’s claims are deceptive.
(D) Hansen is not a liar. He is fooled by his own statistics.
Mass tells us he is no climate denier (although I note references of his to UHI & AMO that suggest otherwise). If he is truly concerned by the media running so rich with unscientific nonsense spoken by climatologists or (more narrowly) taken from their work/official testemony, would you not expect the search word “Lindzen” to return more than zero hits on his blog?
The problem, Dear Al, is that you’ve demonstrated nothing other than your hostility and inability to argue based on the facts. Yes, Cliff is wrong. Yes, he could use a few sessions of charm school. The problem is you did not school him, and you deprived the rest of us of the opportunity to do so.
So, now, Cliff can go back to his buddies at WTFUWT and brag about how he held his own in the lion’s den and how rude we are, and how we had no substantive arguments and resorted to insults and ad hominem. You let him win because you lost your cool. Don’t do that. Punish him with facts, with reasoned argument and relentless insistence on the literature. That is how it is done.
Cliff Mass -“But without global warming we would have still have big events…perhaps 2.5 or 2.7 sigma…you know what I mean. So the dominant cause of the anomalies is natural, with an assist from GW. Why is this interpretation wrong? Can someone please explain my error?”
Tamino pointed out your glaring error in the post above – the frequency of extreme temperatures increase dramatically in a warming climate. Whereas in a climate with no long-term warming or cooling (i.e stationarity) the probability of record-breaking in a time series decreases with time – the 1/n rule. Your comment demonstrates complete ignorance of these facts, and the supporting peer-reviewed literature. But then, you do not seem to have understood what Hansen (2012) was actually about, so that’s not exactly surprising.
I think it would be a further error on your part to think that the majority of readers here have not noticed your bluster, rhetoric, mimophancy, and lack of any substantive argument. And you are going to appear increasingly foolish as extreme temperatures afflict the planet in the coming years. But hey, it’s your reputation.
Well, I am not sure if I’ve got the whole story right. So here is what I currently think about it and will be very glad if someone can confirm or correct me. First of all, I do understand that Hansen did his study for whole North Hemisphere. I understand that N hemisphere average temperature is a weighted average of temperatures at stations, spread around it (I’ll omit weighted in further writing just to simplify things for my torching head). As I understood Tamino’s critics, there is no detectable difference in standard deviation of noise (after removing trends). However, as the trends are different for various stations, some of them warm more and some less and that change would be normally distributed around some mean for given period. So across the hemisphere, this actually means the widening of normal distribution and thus observed increase in 3-sigma events.
Now, there is another two questions which I want to ask. First: Let say we compare events from the past to the current and we get what is predicted. But, if we’d compute temperature anomalies from current baseline, the past should show us increase in 3 sigma negative events comparing to the current temps (just thinking of symmetry here).
Second: Regardless of this symmetry, it is actually false argument as the base period of the past is the actual environment ecosystems are adapted to (20 deg C in arctic may be good for crocodiles, but not for the actual ecosystem) ?
I know I’ve made some simplications, but this is how in my layman viewpoint I see things now and I am asking, if this is close to correct view.
“Mathematical expectation” (used in risk analysis to estimate expected cost and in gambling to estimate expected return) would seem to be the logical way to assign a “global warming contribution” to an extreme temperature “event” like the Russian heat wave.
Multiply the probability that the weather “event” was due to climate change times the observed deviation above the long term average (eg, for the particular month ).
If one did that for the Russian heat wave for the month of July, for example (using the 80% probability estimated with Monte carlo by Rahmstorf and Coumou and using 5C as the approximate increase above the July mean) one would get an expected “contribution from climate change” of about (0.8) (5C) = 4 degree C.
This method yields an estimated “contribution from climate change” that is considerably larger than one gets by assuming that the contribution is merely equal to the shift in the mean of the temperature distribution (the latter being the “expectation” that some folks might have, using the term in a non-mathematical way)
Horatio, I don’t think that’s really an expectation. I think your probability is actually a confidence level–not the same thing at all.
Perhaps Horatio misinterprets what Rahmstorf himself has said on the issue, but it sure seems like he was talking about a “probability” (not a confidence level) in “Russian heat Wave statistically linked to Climate Change” for example.