I’m quite happy to publish other people’s criticisms of topical and semi-topical papers. Judith Curry has sent in the following comments on Klotzbach et al, (mentioned on roadmap), abridged by her from a previous post on the tropical listserv, responding to specific talking points on the Klotzbach paper that were posted on Gray’s website.
Judith’s Comments
If the increases in TC activity found by Emanuel [2005] over the past 30 years (based on data from 1975-2004) and Webster et al. [2005] over the past 35 years (based on data from 1970-2004) are robust, one would expect to see similar trends over the shorter time span evaluated in this paper (1986-2005), especially since SST increases have accelerated in the past twenty years.
This is flawed logic, fallacy of distribution of the divisional type, whereby you cannot assume that what is true of the class is true of its members. You cannot dice up the 35 year period and expect the same statistical relationships to be present in each segment. 35 years is marginally short to identify a statistically significant trend (people who criticized our study because the length of the data record is too short raised a legitimate point). 20 years is definitely too short. The reason for this is that both the atlantic and pacific have large multidecadal modes. if you pick a period that is too short, what you are seeing is one piece of the mode. The Pacific has a big multidecadal mode that peaked around 1990, so most of the data outside the N. Atlantic is biased by this particular sampling.
There is considerable disagreement about the data quality before the middle 1980s. Best track datasets for the Western North Pacific, the North Indian Ocean and the Southern Hemisphere before 1985 should be “used with great caution” according to the authors of the best track dataset.
No rigorous uncertainty analysis has been conducted to date. the data in the western North Pacific, which has 40% of global hurricanes/ typhoons, actually DECREASES in quality after 1987, when aircraft reconnaissance flights were discontinued in this region. so the arguments about choosing only the data since 1985 owing to better data quality is substantially flawed
With regards to ACE, there has been a large increase in ACE in the North Atlantic basin since 1986. There has been a large decrease in ACE in the Northeast Pacific basin since 1986. All other basins show small upward or downward trends. Globally, there has been a slight increasing trend from 1986-2005; however, if only the past sixteen years are evaluated (1990-2005), there has actually been a slight decreasing trend.
the ACE confounds the number of hurricanes with the intensity (ACE effectively includes both). There are several ocean basins where the number of cyclones is actually decreasing, which would lead to a decrease in ACE (while at the same time showing an increase in intensity). See also #1, re problems of just looking at a 20 yr time period.
With regards to the number of Category 4-5 hurricanes, there has been a large increase in North Atlantic storms but also a large decrease in Northeast Pacific storms. When these two regions are summed together, there has been virtually no increase in Category 4-5 hurricanes (i.e., 47 Cat. 4-5 hurricanes from 1986-1995 and 48 Cat. 4-5 hurricanes from 1996-2005). For the globe, there has been an approximate 10% increase in Category 4-5 storms from 1986-1995 to 1996-2005; however, most of this increase occurred from the late 1980s to the early part of the 1990s in the Southern Hemisphere where some data quality issues may have still been present. There has been very little change in the number of Category 4-5 hurricanes since 1990, which is an agreement with Figure 4, panel A from Webster et al. [2005].
The more relevant metric would the percentage of cat 4-5 hurricanes (which is what Webster et al. focused on). If the total number of hurricanes is decreasing outside the N. Atlantic, then you would expect some decrease in Cat 4-5. The fact that you still get an increase in cat 4-5 hurricanes implies a substantial increase in the % of Cat 4-5. See also #1, re problems in looking at just a 20 year time period.
There is a positive correlation (significant at the 99% level) between SSTs and ACE values and Category 4-5 hurricanes for both the Atlantic and the Northeast Pacific basins. However, correlations between SSTs and ACE values and Category 4-5 hurricanes for all other basins (i.e., Northwest Pacific, North Indian, South Indian and South Pacific) are not significant.
This argument was debunked by the Hoyos et al. paper recently published in science. While other factors such as wind shear etc are important determinants of the intensity of individual storms and even for seasonal average intensity (owing to factors such as El Nino), there is no trend in wind shear, humidity, etc. The analysis of Hoyos et al. clearly shows that the global increase in intensity shares information with the global trend in tropical sea surface temperature (and not wind shear, etc.)
Climate Audit 
119 Comments
link to the K paper pdf or at least a formal citation?
David Smith initially posted a link (which I could not find and reminds that this site needs a better search tool) so I found it directly here.
Dr. Curry said:
And then from the paper under review we have:
Which contains the claim that bender continues to throw out for comment by Dr. Curry and the GT students.
It would appear that you finally got your answer, bender, if not directly: “fallacy of distribution of the divisional type”.
Good observations, Ken. On both counts. The statistician’s term for “fallacy of distribution of divisional type” in time-series is “nonstationarity”.
Steve M to increase legibilty can you please:
1) Put a link to the “gray website”
2) clearly mark which part of the text is written by Curry
the current layout is a bit confusing.
Re #5 I am assuming everything after “Judith’s comments” are her exact words.
Ref: #2
I believe that the paper I linked in Comment #2 is the one that Dr. Curry reviewed but I remember reading a paper that David Smith had linked that I confused with this one in commenting previously or could have been combining information from 2 different papers. David Smith could you please list all the papers on tropical storms that you have diligently linked previously? I promise to bookmark them all this time.
What I find most interesting about reading these wide array of papers on TS activity is that those that present the raw data without filters or spectral analysis and in unadulterated form would appear to present a picture that trends of TS activity vary greatly with basin of occurence and globally do not appear to show significanlty increasing trends.
The raw data from all these papers taken together would appear to indicate that while SST can explain some of the TS variance, and particularly the number of Category 4-5 hurricanes, there are other factors that together account for most of it. That’s a rank amatuer view of it and I stand ready to defer to the more expert who would care to show my misinterpretations of the gist of these papers taken together.
A clarification and a question:
Re: #6 The text following the **** are Judy’s comments in response to Phil’s text which appears before the ****’d paragraphs.
Phil’s summary from which the text is pulled can be found here:
Click to access klotzbach2006_talkingpoints.pdf
A stats question:
You roll a 5-sided die with sides numbered 1-5, of unknown fairness and stationarity, in 30 sets, with each set each containing 90 rolls. You observe the following outcomes of fours and fives between the first 15 sets and second 15 sets:
*171 4s and 5s in the first set (out of 15*90 = 1350 rolls)
*259 4s and 5s in the second set (out of 1350 rolls)
1. Easy question: Not knowing whether the die is fair or not, is there evidence that the weights for 4s and 5s changed from one period to the next?
2. Hard question: Having in hand the answer to #1, what are the odds that the last 10 sets would have the same number of rolls as the middle 10 sets?
Close obsevers will see that this is not quite analogous to Curry v. Klotzbach as her data is 1975-2004 and his 1986-2005. (He also found 164 v. 180 for his two ten-year periods.)
It seems to me that question #2 is answerable in principle and in the abstract, simply as a matter of statistics.
I shoud add a disclaimer here — I don’t think any of the above is policy relevant, but I do find it scientifically quite interesting!
First, I’m grateful that Dr. Curry is taking time from her considerable workload to participate here. There is no doubt that she, like most of us, have much to do, especially in a dynamic field like hers. It’s a rare opportunity for us to learn directly from one of the experts in the field. Much appreciated!
Now, I have some comments on Klotzbach (surprise, surprise) but I’d like to start with some “outsider” comments on the state of the science. Then, I offer a big-picture point of view, to try to put some of this discussion in context for casual readers.
My state-of-science comments are controversial but I think that Dr. Curry may agree with much of what I write about the big-picture part. Of course, she can correct me where I err!
First, four comments on tropical cyclone science:
1. Science understands some things about hurricanes, but there is much that is not known. It is still a young science. If you think boreholes and tree rings are problematic, well, hurricanes are part of that same company.
2. Accurate storm data is hard to come by, even today, because hurricanes are relatively small-scale events in a poorly-sampled part of the world.
3. Historical storm data, especially pre-1950, is full of problems. It’s better than nothing, but in some cases, not by much. (Ironically, I think that some of the people who generated this historical data may have overstated its quality. They now are eating some crow and having to acknowledge that, while they made a best effort, the quality is limited.)
4. So, when you read these hurricane threads, realize that this is a battle of ideas among smart experts who are lightly-armed with facts. (This is not intended as an insult, as much of science is in the same boat, but rather just an indicator of why these threads sometimes get hung up on the nuances of data and why proper statistics are so important in this effort.)
The bigger picture:
5. The world has warmed in recent times, including the oceans. These are facts. All sides of the discussion agree about this.
6. The big question is, how will that global warming affect hurricanes? If the warming creates more, and more-powerful, storms, then that is important for people to know. If the US Gulf Coast is going to be hit by three or five category 4 hurricanes a year, on average, that is important to policy makers. (Now, dummies like me who live only a few miles from the Gulf may say, “I’ll adapt”, but that’s for another thread and day.) All sides agree on the importance of this question.
7. The “Grays” (Gray, Landsea, most NHC people) think that modestly-warmer seas (say, 0.5C) will lead to modestly stronger hurricanes and (possibly) a modestly-greater storm count. They believe that, while SST play an important role in storm intensity and frequency, other factors, notably wind shear, play a bigger role.
8. The “Greens” (Emanuel, Curry, Webster, other mostly University and Government people) believe that they see something surprising and far more troubling. What they see is a strong increase in the number of severe hurricanes. The surprise comes from the observation that the increase is considerably greater than what Emanuel’s thermodynamic model predicts. The Greens don’t have a clear physical explanation for why higher SST causes such a large increase in intensity, but “facts are facts” and the physics will follow the data.
End of Part 1!
To make her point about “flawed logic” and a “fallacy of distribution of the divisional type, whereby you cannot assume that what is true of the class is true of its members,” Judith Curry seems to ignore an important justification for Klotzbach’s statement that one would expect to see similar trends over the shorter time span evaluated in his paper (1986-2005): “especially since SST increases have accelerated in the past twenty years.”
Since the thesis of Curry and her colleagues is that increasing SSTs are responsible for the increase in hurricane intensity that they found, the last 20 years should have should have shown it even more strongly because SST’s increased more then than they had before. That didn’t happen.
From Dr. Curry:
Why is this measurement more relevant? And why is cat 4-5 the magic metric? Why not 3-4-5 or just 5? What is the reasoning for picking percentage of cat 4-5 hurricanes as the prime dimension with explanatory power?
The First Green Papers:
9. The first key Green paper was by Emanuel, who released a paper in May, 2005. This paper noted that sea surface temperatures (SSTs) have increased in the hurricane regions and also noted that the storms have become more powerful. Emanuel attributed the stonger storms to the higher sea temperatures. He showed the relationship in a dramatic graph, which made the newspapers and is still in use today.
Emanuel uses a thermodynamic equation to physically relate SST with storm power. (One interesting problem with his equation is that it predicts a rather modest increase in storm power due to the observed SST rises, but the storm power has increased much more than the equation predicted. His paper explains part of the discrepancy but not all of it.)
Emanuel states a strong statistical relationship between SST and Atlantic storm power. The period examined was circa 1950 to 2003. The SST used is that of a “box” in the eastern Atlantic.
10. The second key Green paper is Webster et al, found at
http://www.sciencemag.org/cgi/content/full/309/5742/1844 , released in September, 2005.
Webster examined storm trends in the various basins, using the best data available.
Webster found that the annual number of hurricanes worldwide has stayed more or less steady.
Webster found that the duration of hurricanes (“storm-days”) has somewhat decreased.
Webster found that the average maximum hurricane windspeed has remained more or less unchanged.
Webster found that, while the number of worldwide hurricanes has remained “constant”, there has been a noticeable shift towards intense (categories 4 and 5) hurricanes.
Webster noted that SST have increased in all storms basins.
Webster suggested that the increase in intense hurricanes is related to the higher SSTS, and used Emanuel’s thermodynamic argument as the physical basis for the connection.
Webster et al uses no particular statistical methods, but rather is mostly a standard time-series presentation and discussion of data.
11. Note that these papers were released around the time of Katrina et al. Imagine the storm of publicity they triggered (personally, what I remember is the science writer for my local paper doing a “The science is in, and yes indeed, we’re all going Straight to Hell and it’s Your Fault, Smith!” story, with Emanuel’s famous graph attached.)
Next: The Grays Strike Back
#9 David, you say:
My comment: I agree that the world has warmed. But then you follow this up with: how WILL that warming affect hurricanes. So you seem to implicitly assume that the warming will continue (or that the effect on hurricanes has some lag, but since you try to correlate past warming with past hurricanes, I don’t think that’s what you have in mind). So you already assume that we know what caused the warming. In other words, you implicitly assume that it’s all due to anthropogenic emissions of GHG. Am I right?
Then you say that this is all important for policy makers. How do you know? What policy are you implicitly assuming here? My point here is that you seem to be making a number of implicit assumptions. You even claim that everyone agrees on these assumptions. I don’t think that things are this straightforward.
Personnally, I don’t think that hurricanes are that relevant to policy makers. They are relevant to the people who live in specific areas, and to insurance companies. If I lived in such an area, it wouldn’t matter that much whether there are 3 or 4 hurricanes per year. The mere possibility that there can be one cat5 hurricane is enough to determine what sort of house I should build, and in what particular area. If you think (I don’t know, I just ask…) that capping CO2 emissions worldwide is an important an valid policy because it will reduce the number of hurricanes, I beg to disagree…
Fortunately, where I live, I only have to worry about snow storms…
I’ve changed the formatting to improve comprehension. It’s all I’m good for.
“The world has warmed in recent times, including the oceans. These are facts. All sides of the discussion agree about this.”
Actually if S&C’s recent satellite temperature graphs are accurate, then the statement should be ammended to read:
“The northern hemisphere has warmed in recent times, including the oceans. These are facts. All sides of the discussion agree about this.”
Since hurricanes are a climate mechanism for moving heat from the tropics to the temperate regions, the disproportionate distribution of warming to the northern latitudes versus the tropics might very well lead to a decrease in cyclone frequency and intensity. Grossly oversimplified, less thermal disequilibrium equals less heat transport equals fewer/weaker cyclones. Perhaps this explains the NE Pacific outcomes.
As for the time frame commentaries, what meaningful knowledge can be gained analyzing statistically irrelevant time horizons in these studies? The shorter the period studied the more impact anomalies like ’98 El Nino and 2005 Gulf Hurricane Season will have on outcomes. We see this all the time in financial markets reporting, as start/stop dates are much more important than portfolio managers in determining return. The entire climate enterprise suffers from a lack of agreed conventions and standards, even on simple stuff like how to properly measure ground level temperature with a thermometer (convention failures for interpretation of satellite temperature data and statistical methods are worse). In this hall of mirrors world, you can pick the dataset that best suits your purposes and run with it. Its not “cherry-picking”, merely another go round about which methodology is more meritorious. If you are lucky, you might even get a real answer or two.
Re #12: David, I would characterize the papers in the context of a spectrum (no attempt to develop a pdf). Consider a scale of 1-10, and for simplicity i will refer to the authors rather than the papers in an attempt to integrate both their published (in some cases multiple papers) and their public statements on the subject. This is my take on this anyways.
1 Emanuel (all of the increase is AGW, existence of AMO is questioned)
1.5 Mann (ditto, but does not question existence of AMO, just its influence on hurricanes )
2 Greg Holland (coauthor of WHCC and former Gray student, strong pro AGW statements)
2.5 Trenberth (mostly AGW, a little AMO)
3.5 Webster (clear AGW signal, but we cannot project forward without better understanding)
3.5 Curry (recent presentations: about 2/3 AGW, 1/3 AMO + 20 yr cycle)
4 Knutson (AGW, but the effect should be smaller than that inferred by Emanuel and WHCC)
4.5 Pat Michaels GRL paper (the paper is in stark contrast to his public statements, which rate 8)
4.9-5.1 RP Jr (trying to carve out a scientifically neutral position; Roger pls clarify)
6 Chelliah and Bell (AMO, but something different has happened since 1995 that may be externally forced)
8 Landsea (AGW will cause a small increase, but nothing detected yet)
8 Pat Michaels (public statements, in contrast to his published paper)
9 Klotzbach
10 JJ Obrien (nothing published, but a presentation on his website has really been making the rounds; no AGW)
(15) Bill Gray (arguably off the spectrum, he expects global cooling and diminishing hurricane activity in 3-8 years)
To me, the more interesting thing is to look at 2.5-8, the people/papers that are saying both, but with varying emphases (rather than focusing on the extremes).
in terms of statistical activities, can the position at 1 or 10 be falsified?
Re #16 Howard, the following is a false assumption:
Since hurricanes are a climate mechanism for moving heat from the tropics to the temperate regions, the disproportionate distribution of warming to the northern latitudes versus the tropics might very well lead to a decrease in cyclone frequency and intensity. Grossly oversimplified, less thermal disequilibrium equals less heat transport equals fewer/weaker cyclones.
Hurricanes occur in the summer hemisphere, where the equator to pole temperature gradient is weak. There are regions in the winter hemisphere where the SST is high enough, but no tropical cyclones. We don’t understand why, and we don’t understand what controls the total number of TCs (why 90 per year, and not 9 or 900?) The exact role that hurricanes play in the general circulation of the atmosphere/ocean remains unclear. Low resolution climate models (that cant resolve hurricanes) credibly simulate the global climate and the various heat exchanges, so hurricanes do not seem to be of first order importance to the general circulation. Webster, Emanuel, and Trenberth are all addressing aspects of the “big picture” issues of the role of hurricanes in the global climate system, each with different ideas and approaches. Until we understand these issues, we cannot credibly make projections into the future as to how hurricanes would behave in a greenhouse warmed world. This statement was actually made explicitly in WHCC.
Re #13 Francois, thanks for pointing to this.
I don’t mean to imply that we know why the world has warmed, or whether it is continuing. A better word would have been “has” instead of “will”.
If the US is going to be hit by Ivan and Katrina-type storms routinely, then it indeed is important to politicians here. I think that both the Greens and Grays agree with this. Hurricanes also have an impact on US public awareness which the other possible impacts, like melting permafrost, do not. Also, hurricanes affect the southern US, which provides much political support for the “skeptics”, and if that support fades, then the US policy on CO2 may change. Hurricanes are a big deal here.
Personally, I’m a “lukewarmer”, in that I believe that manmade CO2 makes the world warmer than it would otherwise be, but whether that effect is big or small, or important or not, is unknown to everyone. My suspicion is that the CO2 impact is small and not particularly important, but I keep reading and asking questions, in the hope of learning whether my suspicion is valid or not.
Thanks again,
David
Re #17 thanks. Pat Michaels is interesting. I read his book “Satanic Gases” some years ago and always saw him as acknowledging that AGW exists but that the likely impact is moderate.
Re #15 Fair enough, thanks. Perhaps a better statement is to say that the tropical oceans have warmed (both N and S hemispheres).
Re #18 Curry said: “There are regions in the winter hemisphere where the SST is high enough, but no tropical cyclones. We don’t understand why”
Perhaps because the convergence zone is in the other emisphere or close to the Equator so that the tropical ocean during winter is under the influence of the dry descending air, high shear environment, associated with the southermost position of the subtropical jet? Tropical oceans are quite in wintertime not only regarding ciclones, but even for all kind of convection.
Sorry, in #21 “southermost” means “closer to Equator”
I eagerly await the next episode. I find your accounts entertaining and informative. You should continue to give us your views on these papers. With all the information being presented it becomes easier to form our own conclusions on them — and even for the rank amateurs like myself.
The specific paper that you linked a while back and I cannot now locate (I confused it with Klotzbach) was the one that looked at tropical storms from the perspective of partial correlations between PDI, SST and GT, as I recall, and concluded that GT explained little of PDI and seemed to be indicating that GT had an influence on the troposphere that could counter balance the positive effect that SST had on TS and PDI. It was so different from the papers discussed in more detail here that I wanted to hear some expert comments on it. Can you, David, or anyone help me?
Some amateur confusion:
The meteorologist’s argument appears inconsistent to me. On the one hand they argue for the natural periodicity of hurricane intensity, possibly related to the AMO. On the other we have Dr. O’Brien, presumably on this side of the debate, poo-pooing the idea that a small change in SST, from say 90 to 90.2 degrees in the Gulf of Mexico, could account for a rough doubling of Cat 4 and 5 storms. See this interview with him here.
Off the top of my head I would tend to agree with Dr. O’Brien. Yet, we seem to have an apparent correlation between AMO SSTs and major storm frequency. Maybe its just the old warning, correlation does not imply causation?
BTW, I’ve been cruising around on Obrien’s departmental web site and do not know what presentaton Dr. Curry was referring to above.
PS, to Dr. Curry: Greetings from just up the road in “runaway bride” Duluth, Georgia 🙂
I’ve really enjoyed your participation here.
Hi Joel, Jim Obrien’s presentation can be found at http://www.marshall.org/article.php?id=364
Good afternoon (or morning), Ken. I’ve been trying to remember the paper to which you refer, and I am drawing a blank. I’ll keep trying, though.
There’s an interesting paper by Elsner, Tsonis et al on the relationship between global air temperature and hurricane PDI, and I’ll look for a link to it.
David
Re #27, here’s a map of the relationship. This compares Atlantic PDI with SAT for the months of Aug-Oct. Note that the main correlations are negative, rather than positive.
w.
Re 27, 28: I just found the Elsner, Tsonis paper here (link button not working, by the way). After much analysis, they come to the same conclusion that I found in a couple minutes by creating the map above, viz:
The site that makes the maps is here. The beauty of the site is that, in addition to creating maps for built-in variables, it can also use a file containing your own variables. To make the map, I used a file I created of the month-by-month Atlantic PDI.
w.
Re #29 Remarkable chart! I hope you and bender are moving ahead with your project.
Don’t know much about hurricanes and don’t have time to learn much. But I gotta ask a question: Is it too naieve and simplistic to suggest that cyclones/hurricanes are essentially negative feedbacks to warming and serve to rid the Earth of excess heat, by pumping it into space? One of the ways Earth has of stabilizing climate and preventing some type of “tipping point?” Maybe one of the reasons there weren’t many in the Atlantic this year is because last year’s hurricanes blew so much energy to space?
R: #29
Willis E and David Smith the Elsner/Tsonis paper linked in #29 is the one I was looking for. Thanks much. It has now and forever been properly bookmarked.
(Dr. Curry noted in #19 that there is a spectrum of thought on the hurricane/global warming question. I accept that but, for brevity, I need to stay with groupings. However, I will create a middle category which is neither Green nor Gray and name it after the green-gray color “Sage”.)
The Grays Strike Back
The Emanuel and Webster papers arrived at the same time as Katrina and Rita. They all were soon joined by AP, CNN, BBC and the NYT. A storm of controversy ensued.
12. Dr. Gray, the Dean Emeritus of Sturmleute, issued a paper which touched on a number of issues:
* the quality of historical storm intensity data is fair-to-poor, for a whole host of reasons
* it is especially difficult to compare data across time, because measurement techniques changed
* decent intensity data does exist since about 1980, thanks largely to satellites and improved techniques
* thatpost-1980 data show no global rise in storm intensities, despite higher SSTs
* the physical basis (Emanuel’s thermodynamic considerations) offered by Emanuel is weak
13. Dr. Landsea, of the US National Hurricane Center, noted that
* the quality of historical storm intensity data is fair-to-poor, for a whole host of reasons
* in particular, there is a problem with a bias-correction factor used by Emanuel on Atlantic data
* Emanuel’s statistical approach has dubious aspects, like the unsmoothed endpoints (and Emanuel’s key 2005 statistical approach may have recently been “busted” by the CA Sages “bender” and “willis”).
* PDI for US landfalling hurricanes, for which somewhat decent records exist, show no upward trend over the last 100years
* there is a strong need to reanalyze and correct, to the extent possible, historical storm data
14. The Sage Dr. Pielke, Jr noted that there is no upward trend in hurricane destructiveness in the US, when adjusted for inflation and population. That is inconsistent with the rising-PDI position of Emanuel.
15. The Sage Dr. Michaels et al noted that there is a relationship between SST and maximum windspeed, but that it is modest and may involve a “magic temperature” of 28.25C. (Michaels looked at the SST beneath each individual storm, whereas Emanuel looked at a SST “box” through which the storms may or may not have traveled.)
Have to go, bedtime for my son.
Next: Replies, Hoyos, Klotzbach
I did a LexisNexis search on “Judith Curry” to see how her science is represented in the non-journal press. Among other things, I found Judith’s Congressional Testimony before the House Committee on Congressional Reform, dated Thursday, July 20, 2006.
There, Judith recounted her conversion from skeptic to advocator as regards AGW: “As a result of SHEBA, and my increasing awareness of the impacts of the warming in the Arctic (see ACIA 2004), I became convinced that greenhouse warming was having a substantial impact in the Arctic.”
SHEBA is the “Surface Heat Budget of the Arctic Ocean” study.
What Judith was apparently saying here is that she became convinced of greenhouse warming as a consequence of empirical observations, and not from a causal connection established by a falsifiable physical theory. That is, the conclusion of greenhouse causation was achieved by jump and not by deduction from theory. I recall that jumping to conclusions involves flawed logic.
Judith then went on, “a series of national and international assessments undertaken by the Intergovernmental Panel on Climate Change (IPCC), the U.S. National Academies, and the U.S. Climate Change Science have made it very difficult to maintain a credible position of scientific skepticism regarding the influence of humans on global warming. The past year has seen striking resolutions to two controversies involving the data record of climate change that support anthropogenic greenhouse warming: the synthesis report on the surface temperature reconstructions over the past two millennia the (NAS, 2006) and the synthesis and assessment report on temperature trends in the lower atmosphere (CCSP, 2006)”.
“Further, the draft IPCC 4th Assessment Report presents climate model simulations that are far more sophisticated and accurate than were available in prior assessments, substantially increasing the credibility of such simulations and the associated projections.” (Bolding added)
I added the bolding to show the evidence Judith considered definitive for AGW.
So, what is that evidence? The NAS panel report that actually showed proxy reconstructions prior to 1600 are scientifically unsustainable, nevertheless apparently supports the position that 2-millennium-long proxy reconstructions demonstrate unusual 20th C warming.
The CCSP 2006 report on lower atmospheric temperature trends, available here, that shows (Figure 3.1) moderate surface warming after 1980 (similar in magnitude to surface warming between 1910 and 1940).
I’d like to know how a positive trend in atmospheric temperature ineluctably entails a conclusion of greenhouse warming.
And finally, the sophisticated climate models (unarguably sophisticated) demonstrate that 0.6 C in net 20th C warming is surely from Athro-CO2, despite the fact that no uncertainty metric has ever been published and tens of W/m^2 remain absent from GCM models.
Line of evidence #1, the millennium-length proxy reconstruction, is wrong. Line of evidence #2, the recent surface temperature trend, is indeterminate. Line of evidence #3, projections by GCMs, remains unevaluated as regards uncertainty bounds, coarse in microphysical resolution, and incomplete as regards overall physics.
It seems to me that the evidence above makes it very difficult to maintain a credible position of scientific advocacy regarding the influence of humans on global warming.
Re: 35 — I neglected to link Judith’s July 20 testimony before the House Committee on Government Reform. It’s here, along with testimony given by others during that meeting, including Roger Pielke, jr.
“The chairman’s introductary essay on that page shows a bit of droll cynicism, by the way. He wrote, “It is our job to ask whether we’re responding appropriately where there is scientific consensus, and whether we’re facilitating the research and ensuring an unbiased review where there is not. … We were looking forward to hearing from Dr. Jim Hansen, NASA’s preeminent climate change scientist. But we learned just days ago that he was no longer available to testify. Let the record show he was not muzzled, not by this Committee at least.
Nor will we be hearing from Vice President Gore, who has spoken often of Congress’s and the Administration’s “blinding lack of awareness” about this “planetary emergency” and whose spokesperson told the L.A. Times the Vice President would “go anywhere and talk to any audience that wants to learn about climate change and how to solve it.” The Committee asked the Vice President to pick any date in June or July, but apparently ours was not one of the “audiences” he had in mind. While Mr. Waxman and I are disappointed, we understand that movie screenings and book signings are time consuming, and we hope his book signing in Northern Virginia went well yesterday.“
Re #34: I think you need to be a bit pickier about material that has not passed through peer review, David. Michaels’ stuff is nearly as shaky, for reasons you should know.
Re #35: Pat, it’s very clear that no amount of evidence will convince you.
Dr. Curry, thanks for your response. Does anyone have a handle on how much heat is actually moved north by tropical cyclone activity versus other transfer mechanisms?
As a New Englander, we see a fairly regular diet of warm core type hybrid type noreasters (rogues), characterized by rapid intensification (Walter Drag’s ‘bombogenesis’)and thunder snow. They are more prevelant in years where local SST is in a warm phase (like this year with a Manatee and tropical fish in Narragansett Bay). As the northern branch seasonally intensifies and buckles south, baroclinic forces obviously predominate, but winter storms that form “in the box” (40/70)regularly tap heat content of tropical origin. GFS in particular seems to miss this dynamic with some frequency as it apparently under-estimates cyclonic heat content and overwhelms intensification and storm track with a bias towards rushing cold fronts off the coast (I am repeating hearsay here, for what its worth, but it seems consistent with what I have observed).
I swear, I don’t get it. Looking at annual averages for the period 1949-2005, we find:
There is no statistically significant trend in the PDI of hurricanes (p=0.72, cannot reject the null hypothesis by a long, long ways).
There is no statistically significant trend in the number of landfalling hurricane (p=0.36).
There is no statistically significant trend in the PDI of landfalling hurricanes (p=0.72).
There is no statistically significant trend in the number of total days of tropical storm wind or greater in landfalling storms (p = 0.06)
There is no statistically significant trend in the number of total days of hurricane wind in landfalling storms (p = 0.49)
There is no statistically significant trend in the number of total days of tropical storm wind or greater in all storms (p = 0.06)
There is no statistically significant trend in the number of total days of hurricane wind in all storms (p = 0.61)
There is a small but statistically significant trend in the sea surface temperature (August-October) of the Atlantic Cyclogenesis Region (trend = 0.06°C/decade, p=0.03)
Where is the evidence for increased hurricanes? Where is the evidence for increased hurricane strength? Where is the evidence for increased hurricane duration? Where is the evidence for increased hurricane destructiveness?
The sea in the ACR has warmed very slightly. Hurricanes have not increased in number, duration, or destructiveness.
What is all of the furor about?
w.
STATISTICAL COMMENTS:
Trends are calculated using the method of Theil-Sen. The T-S trend is the median of the trends of all possibile pairs of data points. This method is robust overall, and also works with missing data. In addition, it is robust regarding outliers and endpoints. In a well-behaved dataset, it is almost identical to the trend obtained by linear regression. However, it performs better than linear regression in non-normal datasets, including autocorrelated datasets.
Significance is calculated in the presence of autocorrelation by the method of Yue, as detailed in The influence of autocorrelation on the ability to detect trend in hydrological series, Sheng Yue et al. To quote from the paper, the basis of the method is:
#37 — “Re #35: Pat, it’s very clear that no amount of evidence will convince you.
Steve B., it’s very clear you don’t understand the basics of scientific thought. In the absence of a falsifiable physical theory, there is no evidence.
You’re wrong in any case. You, or anyone else, can convince me by showing me a GCM with uncertainty limits the size of, or smaller than, the CO2 forcing, in W/m^2.
A lurker based in environmental science here.
I simply want to thank Judith Curry for her continuing contributions at climateaudit.org.
Based upon my knowledge of the literature, her input is almost invariably direct, astute, and stimulatingly written – even on the few occasions where I deign to differ. Thanks again!
Re #35 Pat, AGW has been elevated to a theory. It has survived attacks by skeptics for several decades now. It represents the best available explanation of the temperature record for the last 100 years (both physical and statistical). It has also shown predictive capability.
A theory does not mean a “law”. Even Sir Isaac’s Laws of Motion took a hit from Einstein. Skeptics do and should continue to search for flaws in theories (and even Laws). It is becoming increasingly difficult to maintain a credible position of skepticism on the validity of the theory, but many uncertainties remain particularly owing to the complexity of system under consideration. My reasoning for accepting the AGW theory was not a single empirical observation, but rather from a decade long examination of what was going on with the Arctic sea ice ( I have published about 20 papers on this topic)
Falsifying a theory is very different from identifying an uncertainty or flaw in someone’s analysis method. I have yet to see an alternative physical or statistical explanation that comes close to AGW in credibility (cosmic rays or something else that we don’t really understand could conceivably emerge, but this is very very very far from providing a better physical or statistical explanation). While the “tunings” of climate models certainly bring into question the credibility of forward projections that are far outside the current climate regime (say > 1C), they have established themselves as being able to credibly simulate the current climate (this is not to say the simulate every aspect of the climate in every region). Note, in a model with an enormous number of degrees of freedom, a few tuning parameters do not dominate the solution. Continued progress in climate modelling is reducing the number of tuning knobs.
WIth regards to the hockey stick, MBH and the hockey stick reflect a hypothesis (not a theory). MM pointed out errors in the statistical analysis and interpretation of some of the data sets. MM’s analysis has been essentially vindicated by the Wegman and North reports. Clearly more work needs to be done, and since 98/99 MBH, considerable work has been done. North’s statement “but it doesn’t really matter”, correctly infers that the MBH hypothesis has not been falsified by MM.
This is my summary of the state of the science. Climate is an incredibly complex system, there is much work to be done to further our understanding and modelling capabilities. We have a viable theory of AGW, but future projections of AGW are quite uncertain (this is where it gets interesting in terms of policy). Decision makers need to accept the uncertainty and get on with their decision making process, factoring in the uncertainty.
#39 Willis, owing to the large multidecadal variability in the Atlantic and our understanding of the global external forcings during the period 1949-2005, attempting a linear trend through this period isn’t very useful (although this is arguably what Emanuel was doing in his first paper). This is why people have been focusing on the period since 1970 (when the signal from AGW starts to be clearly seen in the climate model attribution studies). This also why Webster says you need to look globally so that all of the large multidecadal cycles in individual basins get averaged out, and why Mann/Emanuel and Trenberth say that in the NATL, you need to sort out what is AMO and what is AGW. These are the challenges in our current thinking on the subject.
David, I am really enjoying your analysis. I support your selection of material. Re what is peer reviewed and what is not. The extreme positions of Gray and OBrien are not published (Gray’s stuff has been rejected, but i don’t think Obrien even tried to publish his), but i think they are worth considering as the tails of the spectrum (it is indeed unfortunate that the unpublished views have held such sway in the media). Also, the simple plots in my testimony (unpublished) I think are very useful but are probabably unpublishable unless something “fancy” is done with all this. I suggest that we look at the arguments and data, and not be totally constrained by what is published.
Judith,
That’s a pretty big call to make around here. What would say was the MBH hypothesis that “has not been falsified”?
My opinon is that MBH 98/99 is worthless (and that is a considered opinion), and I’m not convinced of the merit of subsequent paleo reconstructions.
#45/Judith: I also consider that statement simply asthonishing. And add two more questions:
1) What do refer to as you say “since 98/99 MBH, considerable work has been done.”?
2) For which hypothesis do you consider MBH9X giving support, and what are the facts in MBH9X supporting that hypothesis?
This is true, and we should try to remember that MBH9X was “groundbreaking” because it was supposed to falsify the prevailing theory at the time (the prevailing theory, laid down – mainly subjectively – in journals by scientists such as Lamb etc., was that the earth was warmer during the medieval warm period than it is today).
Steve M demonstrated that MBH9X offered no statistical significance in terms of falsifying the prevailing theory; therefore, had failed to accomplish what it had set out to do. Furthermore, most of the subsequent paleo reconstructions suffer from similar problems that prevent meaningful estimation of temperature prior to 1600 (as outlined in the NAS panel report).
The only conclusion we can draw from the NAS report is that we simply do not know whether it was warmer in the last 1000 years than today or not. The confidence intervals are unknown prior to 1600. The subjective view carries equal weight to a graph with confidence intervals that are “off the page”.
Correct. And that protects us from the individual tinkerer who focuses on a narrow set of tunable parameters in order to confirm his hypothesis. However this fact does not protect us from “the blind watchmaker” – the team of tinkerers who, over the course of 30 years of tinkering, evolve a complex tuning where the solution is not dominated by any one parameter, but rather by teams of parameters conspiring with one another to produce a directed result.
The immovable mass of a model is no guarantee it isn’t junk. Junk comes in all sizes.
Re Hockey Stick, i have stated many times that I am not going to get drawn into a detailed discussion on this, and particularly not on the hurricane thread. I am trying to take the big picture view of credibility of the AGW theory, which is central to our arguments about hurricanes and global warming. I do not view the hockey stick debate as all that central (which is one of many reasons i don’t want to spend much time on it). Falsifying the MBH hypothesis would require identifying unambigous evidence of a global warming during MWP. Alternative hypotheses with greater evidence may eventually carry the day, but they will have to go through the entire process of surviving attacks by skeptics, etc. At this point, MBH has survived as a hypothesis (but has nowhere near the support required for a theory), and in my assessment the MM critique of MBH, while valid, has not been a major blow to the AGW theory. This is my assessment. I realize that there is a broad spectrum of opinions and assessments on this issue. The scientific process will eventually sort this out. Back to hurricanes (at least on this thread).
Re #48 Bender, yes you can tune to get the “correct” space/time average, but you cannot tune a model with so many degrees of freedom to get the correct space/time variability. The performance of climate models is quite credible in terms of simulating the major features of space/time variability, and for every little discrepancy pls see RP Sr blog, this is what he focuses on. The critical issue in terms of projections is that the tunings almost certainly result in incorrect sensitivities once you move very far beyond the average conditions for which the model was tuned. This is why continued efforts are being made in the climate modeling community to add more degrees of freedom associated with resolution and physical processes so that the model is increasingly constrained by physics (not the tuning knobs) and can provide more realistic sensitivity to relative large external forcings. I’ve posted on this before somewhere, but the complexity of climate models in terms of the number of degrees of freedom is way beyond the size of the system that dynamical systems physicists and mathematicians can handle with their current theories and methods. This does mean that climate models are not physics, only that climate models are of incredible complexity and pushing the limits of our understanding of how interpret and evaluate these models as dynamical systems.
Judith, I appreciate your desire not to be drawn into the hockeystick debate, but I guess I’d like to put on the record that “warmest in 1000 years” (or variations) is not really an MBH hypothesis, it’s a result.
The closest I can come to a MBH hypothesis is “we can reconstruct NH tempertature for the past 1000 years”, and that one is definitely “busted”.
Judith, you asked for it by editorializing on the hockey stick. If you don’t wish to justify such comments, then you shouldn’t make them. There are a lot of knowledgeable readers here about all aspects of matters hockey stick and editorializing that passes muster in Georgia Tech seminars won’t pass muster here simply because people here are better informed on this issue. Let me make one editorial comment in reply and then we can leave the topic alone here.
As my editorial comment, what we stated was that that MBH could not claim 20th century uniqueness based on their data and methods. MBH was the study primarily cited by IPCC on this issue, MBH98/99 have been cited in about 1000 academic publications according to Google Scholar and the leading study in the field. We did not preclude the possiblity that other data and methods could establish the conclusion. But that would not vindicate MBH.
However, in addition, I deny that any of the other Hockey Team studies are equal to this challenge. There has been much discussion of these "other" studies on the blog. Working these discussions up to journal articles is one of a dozen things that I need to do. But regardless of whether I do this or not, to my knowledge, other than what I’ve done here, there has been no detailed analysis of these studies. At a minimum, I observe that many of these studies use bristlecones (which the NAS panel said shouldn’t be used) and use the cherry-picked Yamal series instead of the Polar Urals Update.
Again, to others, let’s leave this topic alone on this thread.
Judy, would you agree that there are competing hypotheses, both viable at this time?
#52 Steve, I agree that I opened myself up to at least some discussion on the hockey stick owing to my public statements in testimony, where I referred to the debate and the the North report in passing. Strictly in the context of the arguments that I am making re hurricanes and global warming, I am looking at the overall picture. The 1000 year record is part of the overall picture certainly. But my main point is that in my opinion the AGW theory has survived in spite of MM’s valid critique of MBH.
I would love to see a “hypothesis” paper on the 1000 year reconstruction analogous to my BAMS article, where the hypotheses, arguments, and uncertainties are laid out. My BAMS article certainly was “biased” by my own perspectives and even selections, but i think even a personal perspective has the advantage of clarifying issues such as what the hypothesis actually is can actually focus the critiques and help evaluate their significance in supporting or refuting the hypothesis.
Steve, I don’t see why we should take seriously your criticisms of other studies, given that they are unpublished. Selective snippets of criticism on this blog are inadequate for understanding or contesting broader analysis of a paper or several of them. To your point, that the recons have not been audited, so what? You haven’t audited them either. Because there’s no publication. That means there was no audit.
#55. TCO, you’re confounding two points as you do so often. I didn’t say that my criticisms of these other studies should be accepted based on comments at a blog. I said that, other than what I’d done, there had been no detailed analysis of these other studies. If such other studies exist, please identify them for me. It’s not my personal obligation to perform due diligence on all these studies.
Part III: Emanuel Replies; Hoyos; Klotzbach
16. Emanuel replies to Pielke and Landsea
* to Pielke, the database of landfalling hurricanes is too small to allow meaningful conclusions about PDI trends
* to Landsea, you’re right about my questionable technique of mixing smoothed and unsmoothed points on my plots, and correcting that will make some difference in the Atlantic, but the correction has little impact on the more-important Pacific chart. More importantly, my SST/PDI high correlation still stands and is what counts.
* I used a bias-correction method from Landsea himself which he says I misapplied. It’s his method, what the hey, and I accept his point
* besides the SST/PDI correlation, there is a strong correlation between SST and Atlantic storm count over the last 130 years. So, the hurricane/SST connection is ‘everywhere’
17. Hoyos-the-Green (April, 2006)
* the paper looks at factors (humidity, wind shear, stretching deformation) other than SST to see if a connection can be made between them and the increase in severe hurricanes
* agrees that these factors affect individual storms, and even seasons, but do they play a role in the longer-term intense hurricane trend
* no global connection is found between humidity trends and severe hurricane trends
* no global connection is found between wind shear trends and severe hurricane trends
* no global connection is found between stretching deformation trends and severe hurricane trends
* short=term connections yes, but not trends
* I saw a word I like and have added to my vocabulary: “deconvolution”
Klotzbach-the-Gray (May, 2006)
* looks at the recent 20-year period (1986-2005) only
* says the 20-year limit is chosen because intensity data before 1986 has substantial quality problems
* figures that this short period is good-enough because SST have risen substantially and a SST/ACE (similar to SST/PDI) trend should be evident
* looks at each of the world’s storm regions
* finds no global trend in ACE (similar to PDI)
* finds no global trend in the number of severe hurricanes
* notes that these findings contradict Emanuel’s and Webster’s
* suggests that classical factors (wind shear, etc) rather than small SST changes remain the key considerations in storm intensities
Next: Common Grounds and Battle Grounds – the Key Issues
Then: David vs Goliaths
Then: And the Winner Is…
Steve, consider a para break to have been added halfway, if that helps you. I wanted to address both points specifically as they are related and as you made a remark “other then what I’ve done here”.
WRT examination: The studies themselves are examinations and there are multiple studies. If you think they are wrong, it is incumbent on you to describe how, in publication. If you don’t bother, then the studies will stand. Despite the “firing from a protected location” going on at this blog.
BTW, I liked the Emanuel responses to comments. Where he could agree, he agreed. Where there was a remaining difference, he noted it. Where he wants to make clear that the comment is correct, but that the implication is minor, he explains that. [snip – you’re getting to be a bore with offtopic comments.]
#59. TCO, Emanuel said that he “neglected” to remove the endpoints in his filtering. Show me an algorithm that requires manual removal of endpoints after filtering. Emanuel’s endpoint pinning requires deliberate intervention in R.
Judith Curry invite us to see Pielke Sr weblog for “little discrepancy” because, her words, “the performance of climate models is quite credible in terms of simulating the major features of space/time variability”
I do not believe that Roger Pielke Sr agree with her!
Pielke states:” Global and regional climate models have not demonstrated skill at predicting climate change and variability on multi-decadal time scales.”
Really just a “little discrepancy”!
http://climatesci.atmos.colostate.edu/main-conclusions/
60. [snip – no more thread invasion]
#54 — Judith wrote, “But my main point is that in my opinion the AGW theory has survived in spite of MM’s valid critique of MBH.”
Judith, the AGW claim requires that there be a physics-based model capable of falsification at the level of the observable of interest. That is, the theory must make predictions at the level of the desired test. In terms of AGW, that means GCMs must make predictions at the level of 2.7 W/m^2 forcing. They cannot now do that. The fact that GCMs produce realistic-looking temperature profiles of an Earth-like climate is no validation or vindication of AGW because the GCM outputs reflect forcings that are at least an order of magnitude larger than the CO2 forcing. And even so, the predictions are poor compared to observations, at the level of 20-40 W/m^2. The climate model intercomparison project has shown that.
You’re essentially claiming that AGW, which is a conclusion and not a hypothesis, remains viable because GCMs capture Earth-like climates in the large scale. This is scientifically incorrect. So long as GCMs cannot resolve a forcing at a 2.7-4 W.m^2 level, AGW is no more than unfounded speculation. To suppose otherwise is to claim to know despite the absence of knowledge. Precocious knowledge, otherwise called revelation, is the province of another enterprise.
James Lane in #51 put his finger right on the central point of the other issue. I’ll respect Steve M.’s request to desist on that subject in this thread.
Re #57: You’re missing some important papers, Trenberth for starters.
Re #61: Paolo, understanding Macropielkean requires use of a filter. In this case, consider how long the models have been around in something like their current form and contrast that with the term “multi-decadal.”
Re #62: “…the way of the Clinton.” Interesting that we would have to reach back so far for a suitable example. But no politics on this blog!
Steve B, I’ll read Trenberth
#42 — Judith also wrote, “My reasoning for accepting the AGW theory was not a single empirical observation, but rather from a decade long examination of what was going on with the Arctic sea ice ( I have published about 20 papers on this topic).”
Judith, you could have made 100 years worth of empirical observations in the Arctic, and that cumulated body of factual information would not have warranted a conclusion of Anthro-CO2-driven climate warming. The only way to establish causality in science is by way of a theory that links cause and effect. Inductive conclusion-jumping is never a substitute for that, not even when large bodies of related facts are available. The process you describe — fact to conclusion absent theory — is no different from begging the question.
Let me put it this way. Twenty years of insistence on AGW has not falsified the possibility that the current warming is effectively causeless — yet one more pseudo-period of a climatological strange attractor; the orbital wandering of a chaotic system.
AGW is not a unique solution to the ‘problem’ of climate warming. Given the inability of the GCMS to resolve 4 W/m^2, in actuality AGW is no solution at all.
Re #64
Here is a link for Trenberth, which looks like it summarizes his thoughts on hurricanes in a warming environment. His points flow well and his article is the only one I’ve read who talks about some important details like atmospheric stabilization.
However, I don’t see any big or surprising statements or claims. Now, he’s quick to attribute tropical warming to CO2, but that is not my focus. If I’m missing something, please point it out.
Thanks
Re #67: Go here for the three 2006 papers. You can pick up the others you’re missing by checking the references of several of the most recent papers.
Re #68
Steve B., I went to the site and encountered some kind of PDF problem when I tried to read the paper. So, could you summarize the relevant points?
Thanks
Re #69: That would be rather time-consuming. You couldn’t read any of the three papers? Try again; but if you still can’t see them I’ll email them to Steve M. so he can send them to you.
Re 61, Steve you are missing a fact: the performance of a model has no connection with its age. Dont’you know that a month old model can be run from initial conditions as old as you want?
Re #53 TCO, re competing hypotheses, i would say that there are three viable hypotheses:
1. AGW only
2. AGW + natural variability
3. natural variability
I would say that Gray’s global cooling hypothesis is not viable
The natural variability hypothesis is weak (mainly rests on Goldenberg 2001), mainly resting on the inadequacies in the AGW arguments.
Almost everyone agrees that there should be “some” element of AGW; the real issue is the magnitude of the AGW signal.
#66 Pat, you are missing the point. My little attempt in the 1990’s to “falsify” the AGW hypothesis failed. Based upon my personal expertise and interest, I looked at the “polar amplification” and the arctic warming that the models were predicting. I thought this was way overblown, and without arctic amplification you wouldn’t have much warming. I spent a decade to trying to find “negative feedbacks” that would slow the warming down. I didn’t find anything, just more feedbacks. So AGW survived my little “skeptical attack”. As for the overall AGW theory, as a scientist i cannot personally take on every aspect of it. My personal conclusions about AGW have been formed by a complex process; after all it is a complex issue and people can obviously come to different conclusions on this.
Re: #30
From the Elsner, Tsonis and Jagger paper we see them showing the effect of SST at high frequencies on hurricane power dissipation in the NATL basin is positive, but when they remove the effect of SST, the correlation between GT (surface global warming) and PDI is negative. It is almost as if the took an opposite approach to that of Emanuel and Emanuel and Mann and got nearly an opposite conclusion.
The authors also note the lower frequency (decadal) dependence of PDI on SST as noted by Emanuel but they also see dependence at high frequencies. This to me brings forth the issue that I do not recall being addressed here and that being how does one describe the physics that would explain why PDI should depend on low frequency SST in the context of Emanuel and Mann’s approaches. The authors here use filtering that appears to me similar to that used by Emanuel 2005 to arrive at the same high correlation between NATL PDI and SST. They then go on in an apparent attempt to explain the effect by way of the resulting low frequency thermohaline circulation. I am guessing here that any weaknesses found in the statistical approaches applied in the Emanuel 2005 paper are going to apply to this paper or at least this part of it.
The remainder of the authors’ conclusions are less clear cut to me but include:
That last conclusion must be considered controversial in the climate science world.
Re: #74
Just noticed the Elsner:PDI thread after posting my comment here. Sorry about that. You people are moving too fast for TOM.
Re: 72
I would refer everyone here for a discussion on attribution, if you havent already done so.
re: #73
By “polar amplification” do you mean something like albedo changes and thermohyline processes?
Did you consider changes in cloud cover and if so how did you eliminate them? And what figure do you accept for the CO2 doubling forcing (not counting feedbacks)?
That there is some CO2 forcing isn’t an issue for me. What I doubt is that there’s any substantial net water vapor feedback and that therefore a total warming for doubling CO2 is probably less than 2 deg. C. Not that I doubt that water vapor by itself will cause a forcing, but that there are lots of negative feedbacks involving H2O which are unlikely to be overwhelmed by a doubling of CO2.
Judy, have you read Jan Veizer’s paper (or any of the papers referenced therein)? I think solar energy/cosmic rays/magnetic fields provide a very very very better physical AND statistical explanation. Relative to whether or not the MWP was warmer than today, see all the studies summarized here. What I am interested in are some studies (which have not been discredited, like the multiproxy HS studies) that show that the MWP was NOT as warm or warmer. The CO2-AGW hypothesis (it’s not yet a theory in my mind) really does not have much going for it from a physical standpoint, considering the demonstrated multi-century lags between temperature rise and CO2 rise in the past (i.e., CO2 rise follows temperature rise, not vice-versa). I also don’t think the CO2/AGW hypothesis is falsifiable with the information currently available.
Re 73, Judith, as always your comments are interesting. You say:
Did your research include Polyakov’s study, Observationally based assessment of polar amplification
of global warming, on the lack of observational data supporting the “polar amplification” theory? Dr. Polyakov is with the International Arctic Research Center at the University of Alaska in Fairbanks. The abstract says (emphasis mine):
Also, while the North Pole is generally warmer, the South Pole shows no such change. The physics of CO2 warming are such that the response should be proportional to how low the temperatures are, so both should show warming … but isn’t.
w.
#73 — Judith, you wrote, “after all it [AGW] is a complex issue and people can obviously come to different conclusions on this.”
So, do you agree that the scientific case leaves plenty of room for serious doubt that Anthro-CO2 is the prime cause of current climate warming?
If not, then how do you reconcile that demurral with GCMs that cannot resolve the effect? And that therefore cannot support your position? Consider: You say that you, “spent a decade to trying to find “negative feedbacks” that would slow the warming down. I didn’t find anything, just more feedbacks.” But if the feedbacks are not properly or fully specified in a physical theory, how would you know where to look, or know what to expect? That is, in the absence of a theory that is at least descriptively (if not quantitatively) complete, how would you know your search was either appropriate or exhaustive?
Common Ground and Battleground
It’s a truism that, to reach any meaningful conclusion, one has to have meaningful data. The effort to find, or disprove, a SST/ACE/cat45 connection in the historical record continues to stumble over poor-quality historical data, especially on intensity. Regardless, the effort plows ahead because the question is important:
* There is consensus that historical intensity data is of questionable accuracy, especially pre-aircraft and pre-satellite, and also outside the Atlantic and North Pacific.
* There is consensus that, to the extent possible, the historical data needs reanalysis.
* Expect a battle over the reanalysis methodology, once it is completed and the results released, depending on whose ox gets gored by the results.
* Suggestion: make public the reanalysis methodology so that all parties can review it, the sooner the better. If this has been done, great. If not, then do so. Involve statistical people, it can’t hurt.
* There is debate over the quality of early-satellite data. The period circa 1972 to 1984, which covers some critical years, is in question.
* There is consensus (excepting Emanuel) that storm count data from the 19’th century is of dubious quality.
* There is consensus (excepting Emanuel) that the preferred method for describing a storm’s magnitude is ACE rather than PDI.
* There is battle over how to separate natural SST oscillations from any AGW effect. Statistical help needed.
* There is battle over what, if anything, can be concluded from timeframes (20 years, 35 years, 50 years) that are shorter than the longest natural oscillations. Statistical help needed.
* There is battle over the use, and quality, of landfalling storm data and whether it can be considered to be representative of all storms. Statistical help needed.
* There is battle over the validity of Emanuel’s SST/PDI “correlations”. Statistical coup de grace needed.
* A battle is badly needed over the physical explanations offered about how SST and PDI/ACE are connected.
Next: David vs. the Goliaths
Hmmm… link
#42: Judith, when you write in #42:
it sounds perilously close to the “What else could it be?” argument one hears in discussions of Intelligent Design.
Re: # 81
Do not want to start disagreements, but I believe Dr. Curry has previously expressed some concerns about the use of ACE in that it over emphasized the number of TSs.
Your episodes give a summary of the discussion in the thread and are entertaining. I do not know what others here think, but it is just what the doctor would have ordered for this rank amateur. I think they expose a side of your personality that I do not see in some of your other posts.
#81 David, excellent analysis. I’ve seen a recent (as yet unpublished) study that also questions even the satellite data (at least the analysis of it outside the NATL) in the more recent years. Problems with data quality (outside the NATL) could swamp any nuances in its statistical analysis.
#83 Ken, my problem with PDI and ACE is that they don’t just depend on wind speed, but also on the number of storms and the number of storm days. I’ve seen several studies (as yet unpublished) that show that PDI is dominated by N rather than intensity. Since the main hypothesis that we are working with involves the relationship of intensity (wind speed) and not N, it doesn’t make sense to me to use PDI as a proxy for intensity (wind speed). Re PDI vs ACE, ACE is the more often used variable and is more directly related to the physices (kinetic energy). PDI and its relation to damage is not exact, not clear which power of the windspeed actually relates to damage.
Dave, your posts are very nice. What is your background?
RE #83 Ken, I try to take some edge off my comments,as it’s all too easy to come across as too serious here. It’s fun to try, too.
Slightly off-topic, but I’d like to share a hurricane story. There are probably few people worldwide who know this in any detail.
Last year, had Hurricane Rita hit Houston instead of smaller towns to the east, the US (and probably others) would have likely suffered an economic recession and experienced shortages unlike anything experienced in modern times, for a number of months.
The petroleum and chemical industries are massed near the Gulf Coast. Those industries are highly interconnected: plant “A” supplies plant “B” and plant “B” supplies plant “A”. Mutual dependence. If “A” shuts down because of storm damage, then “B” can’t run, unless it can get a supply from another source.
In 2005, Katrina damaged a lot of petrochemical facilities near New Orleans. The result was that all of industries’ extra capacities had to be used to compensate for the loss of those New Orleans supplies. The system lost its redundancy – all reserves were in use.
Had Rita hit Houston, and done major damage there, the damaged facilities would have taken months to repair (due to a shortage of repair people and parts), which is bad. Even worse, the untold story is that many undamaged facilities would also have to shut down, due to a lack of supplies from Houston’s damaged plants. Then, their customers shut down, and so on. It’s a chain reaction.
The US petrochemical industry would have been destabilized and on its knees for months, and there would have been shortages of many materials, not just fuel.
Rita didn’t hit Houston, but the petrochemical industries shut down as a precaution. When they went to restart, the interconnectedness bacame apparent, and the entire industry destabilized and began to wobble. It was like watching a power grid lose stability and seeing the lights dim and brighten, dim and brighten and almost go out. And that was from a near-miss.
Judy, how can you know the variability of the climate system without evidence of previous extent of variability. If MBH is found without basis (not because of counter-evidence, but because it lacks statistical significance), doesn’t that leave things open?
Engineer by training, TCO, but I’ve been doing other things for many years. Actually lived in Steve M.s Canada for a while (loved it). And you?
Regarding ACE, it is apparent that it has shortcomings if viewed as an energy index, because it takes no account of the physical dimensions of a system, which can vary widely – see this image: http://en.wikipedia.org/wiki/Image:Typhoonsizes.jpg
Until indexes can also take the dimensions of a system into account along with the intensity, they will have little real value as a physical energy index, as a very large cyclone clearly has many times the energy of a much smaller cyclone of similar intensity, and a large but less intense cyclone may have greater energy than a smaller but more intense cyclone.
David, I just wanted to know if you were an amateur or a professional. The details of my upbringing are quite inconsequential.
Amateur, of course.
Re: #91
TCO, it is rather obvious that a general classification like amateur has little predictive value of a poster’s knowledge and that is why I apply the term rank amateur to describe myself.
The other problem with these general descriptors is evident when applying them to yourself where you, as a self-described amateur from a technical prospective, are yet so wise in all other ways that you are eminently qualified to advise those more technical than yourself on all issues, technical and otherwise.
There seems to be much talk about “multi-decadal variation”. There are also well known 100k+ year variations. I can’t help thinking that it’s more than concievable that there are variations/oscillations in climate that cover 100, 1000, 10,000 (and so on) years, and that they may not be in phase, thus creating an extremely complex “map” of what constitutes non-anthro affected terrestrial climate.
A thought on severe hurricanes, which is a highlighted part of Webster et al:
While stumbling through the internet I found a fact about severe (categories 4 and 5) hurricanes. Severe hurricanes have a least one short period in which they undergo rapid, almost “explosive”, intensification. According to this link, almost all category 4 and 5 hurricanes have a period where their pressure drops 25 to 30mb in a day. That’s a big-time drop.
I scanned the East Pacific database and it looks like their cat 4 and 5s also showed “explosive” periods wherein a weak hurricane became a severe one. There is no gradual transformation from weak to severe – it is explosive.
I wonder if this is a clue that a severe hurricane is a “different animal” from its weaker cousins, with some kind of fundamental difference in structure. The explosive phase is actually a transformation into a different structure, one that is more efficient.
If cat 4 and 5 are different animals from 1,2,3s, then I struggle even more with making a connection between small SST changes and increases in cat 4 and 5 storms. I see SST increases as as “higher octane fuel”, at least from what I’ve read of Emanuel, and not a transformer of storm structure.
One comment I left out of #94:
It not widely realized that tropical storms and hurricanes have structural differences. Tropical storms have centers of rotation but no eye. To become a hurricane, a tropical storm has to form an eye, which is a complicated thermal structure that involves descending air and some amount of deep convection adjacent to the eye (the eyewall).
Structural differences exist in tropical cyclones and they are important.
(There are some exceptions to the above, but those are ones which likely had some other transitory or special aspect that allowed them to “break the rules”.)
#94 Hi David
I too am struggling with the small SST changes and how those could effectively double the number of storms. The Grayers are inconsistent here, as they claim cycles such as AMO are causal, yet they dismiss the small dSSTs as a factor.
Michaels’ paper IMO was especially good (caveat: I am an amateur). They presented pretty clear evidence that a temperature threshold (28.25o) was required for major storms (cat 3 +) but that once that temperature was reached, other factors were of equal or greater importance.
They noted the doubling of storms from the 1982-1994 period (AMO neg.) to the 1995-2005 period (AMO pos.), but then rather weakly cited an older study stating that shear or whatever was the cause. I can’t buy that just yet.
It strikes me that if there was 0.5o warming (AMO or AGW) then you’d have a few storms that crossed Michaels’ temperature threshold that wouldn’t have otherwise. Some 20-35% of those would become majors. Yet, that number would be small (not the doubling we’ve seen supposedly by AMO or AGW).
What I’d like to see are areal maps showing a 28.25 degree contour (in the appropriate region) representative of the AMO +/- phases and whether the encompassed area grew substantially in the warmer AMO cycle. Storm tracks would be nice to see on that map, too. Also, I’d like to check if summer SSTs in the western Carribbean and Gulf of Mexico are above the 28.25, even in colder AMO phases. If so this would pretty much rule out AMO and AGW influences, at least for landfalling U.S. hurricanes. (Both easy to do — maybe this weekend.)
#93
Demetris Koutsoyiannis (here) has thought a lot about this characteristic of hydrological and climatological processes and what it might imply about the underlying stochastic processes. Incidentally, one consequence of this “extremely complex ‘map'” is that standard trend tests, particularly those designed for iid data (e.g. Mann-Kendall), tend to greatly overstate the statistical significance of observed trends (e.g. here).
RE #93 Neil, I agree. There are oscillations on top of oscillations on top of oscillations. Some positive, some negative, some big, some small, some high frequency, some low frequency. Identifying and untangling those are Herculean tasks. It is possible that the recent warming is AGW, and it is possible that it is an overlap of natural oscillations, and it is possible that it is a combination of AGW and natural. My belief is the combo, but I have no idea of the ratio.
Joel, I bet you’ll find that the Western Caribbean and Gulf exceed 28.25C even in the cooler years.
My gut guess is that the cat 4 and cat 5 storms have special upper-air features which improve the ventilation. I think there is a point where the upper anticyclone becomes a “pump” instead of simply a chimney. Being at the top of the troposphere, the anticyclone may be affected by all sorts of odd things, things like gravity waves and events in the lower stratosphere. For me it is a fascinating subject.
I look forward to hearing what you spot on the maps.
David
I just finished reading a fascinating talk by Bill Gray, the renowned hurricane expert, which I have posted up here. (Link button still not working … grrr.)
It is a fascinating, clear, and complete discussion of hurricanes, models and global warming. I recommend it highly to anyone interested in the subject, and would welcome anyone’s comments on Dr. Gray’s speech.
w.
Explosive deepening in tropical cyclones is currently poorly understood, although things are improving in this area. Two key factors are high sea surface temperatures (SSTs) feeding the convective engine from below, and low wind shear to allow a near symmetrical vertical vortex structure to develop.
According to current theory a tropical cyclone cannot develop over SSTs of less than about 26.5C (I’m aware of exceptions), and explosive deepening of a cyclone into a very destructive system requires SSTs about 28.5C or higher (there have been exceptions to this rule too) – the key to the exceptions seems to be abnormally cold upper level temperatures that enhance the vertical temperature gradient (speculative).
Another dramatic change can happen with eyewall replacement cycles, where a very powerful Cat 4 or 5 cyclone can quickly morph into a much weakened system over a few hours, before commencing to rebuild intensity again. This fascinating process usually starts in a system with a small tight eye with deep convection wrapped in a wide band around it, and a secondary convective band begins to build and wraps around the central core, restricting the flow of wind into the core and causing it to collapse. The secondary band then takes over as the eyewall of a much larger eye which begins to contract as the sytem rebuilds intensity. The full eyewall replacement cycle often takes about 12 hours.
Oops re #101 – minor correction – 28.5 should read 28.25.
Re #102 My impression, too, is that explosive deepening into cat 4,5 and the nature of severe hurricanes are poorly understood. Even the hypotheses seem to be lacking.
The weakening during an eyewall replacement is usually temporary. That makes me wonder if there are structural aspects to a severe hurricane that allows it to regain severe status after an eyewall replacement. Just as there is usually a measurable pressure drop when a system forms an eye feature, which is a structure, I wonder if the explosive drop of a severe hurricane indicates the development of a yet-undefined atmospheric structure.
Your point on the importance of a near-vertical structure is excellent. One of the problems with Hoyos et al is that they look at 850-200mb shear only. They should have looked at other, intermediate levels too, before saying that there is no trend in shear that could explain higher storm intensity.
Unfortunately, one of the complicating factors is that shear data, even at the 200mb level, is shaky, and accurate shear data for, say, 300 or 400mb may be unavailable. That may have impacted Hoyos’ choice of shear.
Personally, I’m ready to “move on” with regards to hurricane trends, with the exception of looking at Emanuel’s and Gray’s work. Emanuel continues to impress me as slippery and Gray’s non-hurricane work (my initial looks at it) looks like it has some soft reasoning.
Our intense-hurricane question is plagued by poor historical data and weak hypotheses. There is a reanalysis of some historical data underway. I hope the reanalysis is high-quality.
If we get good reanalysis data, and it shows something, then I expect scientists to either (a) present a good, detailed hypothesis that explains the cause of the results or (b) clearly say, “I can’t explain these results but, wow, this looks really interesting”.
The bottom line is that we won’t know whether hurricanes are getting worse until the reanalysis is completed. In a way, this hurricane question should be put on the shelf for a couple of years.
Meanwhile, though, I have developed an inclination, which is not as strong as an opinion. It is based on preponderance of evidence, imperfect though each piece of evidence might be. I realize that preponderances of evidence can be wrong, but it’s all that is available.
I will change my inclination to an opinion, either way it might go, when given good data and a good hypothesis. We are not there yet.
The key question is whether something has changed which is making hurricanes worse (max windspeed, duration, count). Plausible physical explanationas are important to my inclination, especially with all the flaws in data:
1. Per Emanuel, Atlantic PDI has trended significantly higher since 1950 (however, CA results show a trendless pattern)
2. Per Emanuel, western Pacific PDI has trended significantly higher since 1955 (however, CA results show a trendless pattern)
3. Per Webster, the percentages of global hurricanes that are cat4,5 have increased in all basins (1975-90 v 1990-2005) (historical data is questionable)
4. Per Webster, the total global count of tropical cyclones (hurricanes + tropical storms) is trendless
(historical data is questionable but, if anything, earlier storms were undercounted)
5. Per Webster, the total global count of hurricanes is trendless (historical data is questionable but, if anything, earlier storms were undercounted)
6. Per Webster, the average windspeed of hurricanes is unchanged (historical data is questionable)
7. Per Webster, the global count of hurricane-days (one hurricane times one day equals one hurricane-day) is trendless to slightly down (same caveat about historical data quality)
8. Per Pielke, Jr, the count of US landfalling hurricanes is trendless
9. Per Webster, global humidity in the tropical regions is trendless (But Webster combined 925 and 850-500 humidities which is apples and oranges)
10. Per Smith charts, tropical precipitable water is trendless
11. Per Webster, wind shear (850-200mb levels) is trendless (need to look at other, intermediate levels)
12. Per Webster, stretching deformation is trendless
13. Per Klotzbach, ACE in the western hemisphere (Atlantic + eastern Pacific) is trendless over the last 20 years (20 years is extremely short compared to natural cycle frequency.)
14. Per Klotzbach, ACE in the rest of the world is trendless is trendless over the last 20 years (20 years is extremely short compared to nayural cycle frequency.)
15. Emanuel mentions that global tropical count is trendless
My inclination is to believe that little has changed in the world of tropical cyclones in the last 70 years. There may be modest upward trends in intensity, consistent with the higher SST, but that would be about it. My inclination is to think that Webster is the victim of bad historical storm classification data.
David
re # 101
Willis your link to the Bill Gray speech contained this paragraph
If skillful GCM climate forecasts were possible, we would be eager to track their skill. Currently, GCMs do not make seasonal or yearly forecasts. They dare not issue these forecasts because they know they are not skillful. GCM climate forecasts cannot compete with empirical climate forecast schemes. How can we trust GCM climate forecasts 50 and 100 years into the future (that cannot be verified in our lifetime) when these same models are not able to demonstrate shorter range forecast skill of a season or a year?
Not too much loopiness there!
Re #106 Climate models do make seasonal and annual forecasts. I note here that to make a credible seasonal forecast, you must include the ocean. US NWS seasonal forecasts do not include the ocean (European seasonal forecasts do include the ocean). For season-year forecasts in the U.S. I refer you to
International Research Institute for Climate and Society http://iri.columbia.edu/pred/productlist.html
COLA http://www.iges.org/dsp/
Bill Gray is not exactly up to date on the current climate modelling situation or the literature
I note in Judith’s links, none had predicted the development of an El Nino this July 2006.
In fact, one of the links had no updates after 1997 (why did you link to a dead site.)
In fact, the one link from columbia edu still had low probability of El Nino developing (25%) in their July 2006 update. They didn’t go over 50% until the September 20th update (when it was already clear from the buoys that El Nino conditions had developed.)
http://www.cdc.noaa.gov/map/clim/sst_olr/sst_anim.shtml
COLA has moved — some pages from the IGES homepage were 404, but at least one of their forecast pages has this pointer:
All of the weather maps have been relocated to http://wxmaps.org
That site works and has current info
Columbia’s pages appear to me to be current, when I looked, a link led here:
http://iri.columbia.edu/climate/forecast/sst/
Interesting exchange today between Pielke Sr. and Judy Curry at
Sorry for the break — at Climate Science
Sorry for the break — at Climate Science
#111 HFL,
this subject came up OT in the Dan Hughes Software Validation topic.
Now I am really confused [or somebody is lol] Thanks for the info!
#116 I apologize Rocks. Had I been paying attention I would have commented in the DHSV thread.
HFL
Hey no apology needed, it was off topic over there and here it is not! 🙂 I was only confused about the subject of ocean temps, not where on the blog the discussion was taking place.
It’s really interesting to google “ocean cooling” and “ocean warming” and see all the different information that comes up.
Dr. Curry,
Even if Dr Gray was not in touch with the recent computer modeled seasonal forecasts, what has that to do with the reality that most of those forecasts were once again wrong?
Judy Curry, 2006 on Klotzbach’s analysis of trends finding no changes:
Judy Curry today, on the new Elsner et al. Nature study out today finding increasing trends in intensity, based on 25 years of data:
http://www.nature.com/news/2008/080903/full/news.2008.1079.html
Roger, you took the words out of my mouth.
Also Pat Michael’s GRL (2006) paper on SST reads differently to me today. This GRL paper also was vociferously commented upon my Kerry Emanuel.