Bishop Hill Reinhard Böhm
Aug 1, 2010 Climate: MWP This is a translation of the interview with Reinhard Bohm that I mentioned a few days back. Many thanks to the reader who provided it. There are a couple of places where the meaning of the German original is unclear.
It is a great pleasure for me to be able to introduce to you one of the most well know Austrian climate scientists, Dr Reinhard Böhm from ZAMT(the Central Institute for Meteorology and Geodynamics). Dr Böhm is regarded as a well respected expert in the field of climate research, in particular for climate modelling of the Alpine region. (You can dowload his CV here with a selection of his publications). Dr. Böhm is a much sought-after guest commentator and interviewee for the mainstream media, (among others, Wiener Zeitung, Der Standard).
It was of particular interest for me to interview Dr. Böhm, since I regard his statements with regards to his scientific work, as being truly compliant with the scientific ethic. This would be better expressed through metaphor: Imagine Diogenes, the Greek cynic - in a metaphorical sense - at a congress of climate scientists appearing with a lamp vainly searching for a climatologist and loudly insisting, that someone should point out one to him. One would point unhesitatingly to Dr. Böhm. As one of the few scientists in his field, he incessantly lobbies, in true aristotolean tradition, for a workable middle ground between climate alarmism and skepticism, in order not to loose all sense of reason.
William of Baskerville: Dear Dr. Böhm, I regard myself as honored that you could find the time to give me an interview. A heartfelt thank you for that!
Dr. Böhm: Thank you very much for your "being honored" - I think of it as self-evident to be able to find time for discussions about climate change. If nothing else, the length of your questions convinced me. I'm no friend of inappropriate summaries, which only in exceptional cases are able to do a complex subject justice.
Please allow me to offer a small reappraisal of the praise you've heaped upon me. I believe I am one of the few in my field who stand for reason in the debate about climate. Maybe too few do that in public. Many due to a lack of time (which i can understand, since the "climate debate" has taken us completely off guard), many through bad experiences, many perhaps due to uncertainty, of which the scientific expert is always wary.This is nevertheless construed negatively in the public arena.
Referring to your image previously from antiquity, had Herr Diogenes used a microphone rather than a lamp at the climate conference, he would have picked up many rational and critical conversations - scientific in essence.
William of Baskerville: Everyone is talking about climate science. Only a few years ago climatology probably ranked, in terms of status, alongside an underwater basket-weaving course and the well educated lay person would only have heard of Dr. Hubert Lamb, the then doyen of climate science, by chance. From this condition of scientific understanding, hardly appreciable to the outside world, emerged the science of climatology at the very latest by the beginning of the 90s. The well known German climatologist Hans von Storch wrote:
Up until the 1980s the dynamics of climate were the main focus of climate research, but since the 1990s it has revolved around the threat of climate catastrophe and climate protection. (Link)
Von Storch attempts to capture this change of state - from a purely internal scientific dialogue, devoted to discerning the truth, to a new condition which appropriates and intermingles with extraneous political subject matter - with the term "postnormal".
Personally I see here the effects of the application of the postmodern body of thought ("panfictional disappearers of reality" - [see explanation in the comments]) to the scientific method. This "change in paradigm" can probably be viewed as being partly responsible for the well renowned and widely accepted phenomenon known as the medieval warm period, first described by Dr. Hubert Lamb in 1965 and confirmed a hundred times and even expanded upon by the IPCC, being expunged from our collective memories through a handful of studies (Mann et al. 1998, 1999). Deming said
Decades of work was overturned by one journal article. The MWP had been reinterpreted out of existence. (Deming, D.: Global Warming, the Politicization of Science, and Michael Crichton's State of Fear, in: Journal of Scientific Explorations, Vol. 19, No. 2, 249, 2005).
Today, more than 10 years afterwards, we see climatology split in to several camps. On the one side those who believe they recognize a looming apocalypse (Polar bears becoming extinct, coastal towns being flooded, melting poles etc.) and, on the other side, those who view any statements about the warming of the earth as being bereft of any scientific evidence. In the middle are those scientists, who however, because of the politically charged atmosphere, find hardly anyone who will listen to them.
Could you convey to my readers the situation as you see it, from your position as an expert and as a simple citizen. Where do you see the causes for this situation, the aberrations within climatology and in particular, where would you start to correct the problem.
Dr. Böhm: When one knows one's own area of speciality inside out, the only perspective remaining is the one "inside looking out". I don't believe anyone who ever claims to be able to totally divorce himself from his/her own perspective and see the world from an entirely impersonal point of view - and thereby objectively. The simple citizen that you ask for, who views everything impartially, I will only claim to be when the discussion revolves around things of which i have no expert knowledge. For example, how one extracts greenhouse gases from the atmosphere, whether human society in the next hundred years will come together, or whether it will become more heterogenous, what influence the climate had on the journeys the Vikings made to Greenland and Labrador and similar things.
And at this point we have already come to the very core of the matter,that which the whole climate debate suffers from: On both sides of the apparently unreconcilable positions of extremity we find so-called experts writing and talking about things in which they are not expert. They should be allowed to do so of course, otherwise there would be no public debate and I consider it to be one of the highest ideals of a democratic and free society. But they should do so on a level playing field and not from a position of being a self proclaimed expert on the subject.
I know how hard it is, not least because the boundaries are not sharply defined but remain nebulous. I start with myself and in truth I probably have a superior level of knowledge about modeling the climate of the Alpine region, but my expertise does not lie in climate modeling, but in quality control and data analysis. That however fits in well with your specific questions relating to the MWP.
William of Baskerville: Herr Dr Böhm, let's talk about the subject of my blog. I deal explicitly with the question of what scientific evidence there is for the claim that a global medieval warming period existed. In the light of my research, I have come across a lot of "peer-reviewed-papers", in which evidence for the existence of a MWP for the Alpine region is cited. At the same time diverse proxy data has been collated, in part to form "multiproxy-reconstructions".
To name but a few examples - Holzhauser et al.: Glacier and lake-level variations in west-central Europe over the last 3500 years, in: The Holocene 15, 6, 2005.(abruf-/ downloadbar [pdf-Format, 1 MB] unter: Glacier and lake-level variations in west-central Europe ...)
- Giraudi: Late Holocene glacial and periglacial evolution in the upper Orco Valley, in: Quaternary Research, 71, 2009.(abruf-/ downloadbar [pdf-Format, 1,1 MB] unter: Late Holocene glacial and periglacial evolution ...)
- Mangini et al.: Reconstruction of temperature in the Central Alps during the past 2000 yr froma y180 stalagmite record, in: Earth and Planetary Science Letters 235, 2005.(abruf-/ downloadbar [pdf-Format, ca. 190 KB] unter: Reconstruction of temperature in the Central Alps ...)
- Büntgen et al.: Summer Temperature Variations in the European Alps, A.D. 755-2004, in: Journal of Climate, Volume 19, 2006.(abruf-/ downloadbar [pdf-Format, ca. 1,6 MB] unter: Summer Temperature Variations ...)
With reference to the above I would like to talk about one of your projects, in which, as described in the prospectus, for the first time a reconstruction of the climate of the last thousand years has been made. It concerns the project ALP-IMP (see http://www.zamg.ac.at/forschung/klimatologie/klimawandel/alp-imp/
Would you like to tell us more exactly about how this project started and in due course the results of this research project with respect to the evidence for the existence of a medieval warm period in the Alpine region?
Dr. Böhm: In the EU project ALP-IMP (Overarching program, http://www.zamg.ac.at/alp-imp), I was lucky enough to have the opportunity to lead a research project, in which European institutes , ranging from renowned to highly prestigious, participated. The main objective was the climate reconstruction of the last millennium, the designated region being the Alpine region. The sought after techniques were instrumental capture of data (I), climate modeling (M) and indirect climate analysis. (Proxies and hence the P from IMP). The project ran for three years (2003-2006) and brought a "harvest" of 50 peer reviewed publications. Very good and professional results were achieved(and that was really "unique" as the prospectus describes) through a co-analysis of thousands of tree rings sets from the region. From the temperature sensitive trees along the treelike , the project partners of the WSL-Institute in Birmensdorf in Zurich(Kerstin Treydte, Jan Esper, Ulf Buentgen, David Frank), the university of Innsbruck (Kurt Nicolussi) and the BOKU-Vienna (Michael Grabner, Sofia Leal) were able to form a reconstruction of the summer temperatures in the Alps right back to the year 755 AD and down to a temporal resolution of one year. Our project colleagues from the CRU of the university of Norwich (Dimitrios Efthymidais, Phil Jones, Keith Briffa) were able to show, through a combination of using their own world-wide encompassing CRU-dataseries and our collated data(HISTALP- dataseries of the ZAMG), that the temperature curve from the high regions in the Alps is representative of large parts of continental Europe and that the the temperature curves for winter or the entire year could not be extrapolated from the measurements made, that were limited to the summer growing months.
I have tried - perhaps this is somewhat boring for the reader - to be precise , in order to render the following statement more accurately:
Yes , we have been able to demonstrate the existence of a medieval warm period through large regions of middle europe for the summer and early autumn periods, that ( and this was new) clearly existed in two particularly warm phases in the 10th and 12th centuries that were split by a cooler period, and which then subsided. The following centuries brought a gradual cooling to a small ice age, whose core period, from around the late 16th until the early 18th century, was 2 degrees C colder than both of the warm centuries of the MWP. A clear warming up began in the second half of the 19th century that culminated in a maximum in the middle of the 20th century, that achieved the same temperature levels seen in the warmest centuries of the MWP. The entire recent warming trend of the 80s and 90s has at least attained the same level of warming as the warmest decades of the MWP. The minor differences between the warmest decades of the MWP and the current warm periods lie within the bounds of uncertainty for both indirect and direct climate data. (large interannual variability, homogeneity problems, non-climatic influences on the proxies etc.)
William of Baskerville: Dr. Lamb spoke not only of middle europe in his publications, but his statements on this subject extended also to Europe and the northern hemisphere, more explicitly to Greenland. Numerous studies of past days, speak, what I find to be, a clear language. I can cite more than 50 studies for Europe (2003-2010). For the northern hemisphere Greenland and Iceland are highlighted in the following studies.
- Dahl-Jensen et al.: Past Temperature Directly from the Greenland Ice Sheet, in: Science, Vol. 282, 9. October 1998.(abruf-/ downloadbar [pdf-Format, ca. 620 KB] unter: Past Temperature ...)
- Vinther et al.: Climatic signals in multiple highly resolved stable isotope records from Greenland, in: Quaternary Science Reviews, Volume 29, Issues 3-4, February 2010.(Abstract abrufbar unter: Climatic signals in multiply highly ... )
- Paul et al.: Diatom-inferred climatic and environmental changes over the last ~9000 years from a low Arctic (Nunavut, Canada) tundra lake, in: Palaoegeography, Palaeoclimatology, Palaeoecology, Volume 291, Issues 3-4, 15. May 2010.(Abstact abrufbar unter: Diatom-inferred climatic ...)
Schließlich eine Studie zu Island: - Ran et al.: Diatom-based reconstruction of palaeoceanographic changes on the North Icelandic shelf during the last millennium, in: Palaoegeography, Palaeoclimatology, Palaeoecology, Article in Press, Corrected Proof. (Abstract abrufbar unter: Diatom-based reconstruction ...)
Do you regard Lambs statements with regards to the northern hemisphere as having been verified? What is the current scientific thinking, is there a consensus?
Dr. Böhm: Yes, there are a large number of well conducted analysis, that confirm the existence of the MWP over a large area. The variation in amplitude between the MWP-LIA is smaller, the greater the measuring area becomes - a matter of course, that arises from the diversity of regional aberrations from the average measured over this area.
Since the Hockeystick-reconstruction by Mike Mann et al (1999), a large number of further reconstructions have been attempted that had the entire northern hemisphere as their objective. As is typical in science, the more exact an object is analyzed, the more the results diverge from one another. We have now arrived at the "Spaghetti curves" from the dogma of the "Hockeystick" , where the amplitude MWP-LIA lies between 0.3 and 1.0 degrees C. In order to enliven this dry description, I have added a diagram of the Hockeystick and the "Spaghettis".[4] To be more precise it relates to one of more than 80 short contributions, that our group made available on the website of the ZAMG (http://www.zamg,ac,at), forming part of our hopefully most extensively rational and objective "morsel of information" relating to the climate debate. Take the "infobits" over the climate of the last millennium as a kind of sneak preview.
William of Baskerville: When we go beyond Lambs statements, although Lamb actually cited evidence "from the arctiic to New Zealand" (Cf., Lamb: The early medieval warm epoch and its sequel, in: Palaoegeography, Palaeoclimatology, Palaeoecology, 1: 14ff.), the question still remains, whether the MWP was a global phenomena. Once again dozens of studies appear to be unequivocal:
Several have been selected , but not representatively:
- Tandong et al.: Temperature and methane records over the last 2ka in Dasuopu ice core, in: Science in China, Vol. 45, No. 12, Dezember 2002.(Abstract abrufbar unter: Temperature and methane records ...)
- Demezhko und Golovanova: Climatic changes in the Urals over the past millennium - an alnalysis of geothermal and meteorological data, in: Climate Past, 3, 2007.(abruf-/ downloadbar [pdf-Format, 8 MB] unter: Climatic changes in the Urals ... - ich rate jedoch, den Download von nachstehender Seite, mit "speichern unter", vorzunehmen: http://www.clim-past.net/3/237/2007/cp-3-237-2007.html)
- Cook et al.: Evidence for a 'Medieval Warm Period' in a 1,100 year tree-ring reconstruction of past austral summer temperatures in New Zealand, in: Geophysical Research Letters, Vol. 29, No. 14, 2002. (abruf-/ downloadbar [pdf-Format, 370 KB] unter: Evidence for a 'Medieval Warm Period' ...)
- Driese et al.: Possible Late Holocene equatorial palaeoclimate record based upon soils spanning the Medieval Warm Period and Little Ice Age, Loboi Plain, Kenya, in: Palaeography, Palaeoclimatology, Palaeoecology, Volume 213, Issues 3-4, 21. October 2004.(Abstract abrufbar unter: Possible Late Holocene ...)
Also interesting is one of the very recent eclectic works by Ljungqvist "Temperature proxy records covering the last two millennia: a tabular and visual overview" (der Abstract abruf-/ downloadbar [txt-Format, ca. 1,3 MB] unter: Temperature proxy records covering the last two millennia), in which he arrives at the conclusion, after an initial evaluation of 71 "proxy-series", that the MWP can be classified, based on solid evidence, as a global phenomena. From the abstract:
Here, the first systematic survey is presented, with graphic representations of most quantitative temperature proxy data records covering the last two millennia that have been published in the peer-reviewed literature. In total, 71 series are presented together with basic essential information on each record. ... Both the Medieval Warm Period, the Little Ice Age and the 20th century warming are clearly visible in most records, whereas the Roman Warm Period and the Dark Age Cold Period are less clearly discernible
When one considers that it revolves around a multitude of different proxies, starting with treerings and ending with (Ice)coredata and the Lambian period (ca 900-1300 AD) …??..then it appears to me that there remains no more doubt, that there is strong scientific evidence the MWP was a global phenomena - although strongly differentiated temporally and spatially. What is your assessment with regards to this?
Dr. Böhm: When you call it "strong scientific evidence for the existence of a global MWP' , then I agree with that. There is no doubt about it. If you had said, like the extremists on one side happily state, that the MWP has clearly exceeded the current temperatures, then I would have had to contradict you. That is not "state of the art". The fundamentalist fraction on the other side talks down the MWP with glee ……[?? the rest of this sentence is really not intelligible]
The reason why it is possible to disagree about the global extent of the MWP and the global "amplitude" of the MWP-LIA is because unfortunately we don't have the requisite density of proxies, despite a number of available proxy sets, because many of the proxies are only seasonal, and this seasonality for natural proxies is biased towards summer periods, and because historical documents in the first half of the century are very thin on the ground. These shortcomings in global proxies that reach back to the MWP have to be dealt with using sophisticated statistical techniques, about which there is still no real consensus.
Last but not least a comprehensive physical understanding of climate variability over a thousand years and more is dependent upon external climatic forcings being documented beyond reproach (which unfortunately is not yet clear as evidenced by the furious debate over the climatic forcing via the Sun). Only then can climate simulations over a period of more than 1000 years back in to the past be calculated on powerful supercomputers, through which questions such as the cause-and-effect mechanisms of the MWP, LIA and the current warm period can be rationally explained.
I don't want to put people off till the cows come home. The future has already begun. Just recently in the climate hosting centre in Hamburg a 7000 year long simulation with the model ECHO-G has been carried out and the anlaysis has just begun...
Reader Comments (33)
I like that, and thanks to the translator!
To explain the obscure bit: "panfitionialistisch, weltverschwindlerischem" is a misspelling of "panfiktionalistische Wirklichkeitsverschwindler" from the Closing Remarks section of "Grundprobleme der Erkenntnistheorie" (see Google Books) p191 and means "pan-fictional disappearers of reality", that is, people who make reality disappear by claiming that it is a fiction. The context is modern fashions in the theory of knowledge, which (apparently) reject both the idea of a knowing subject and a knowable object:
Concluding Remarks
The starting point of our journey through the historical landscape of the problems of the modern theory of knowledge was the understanding of cognition as a relation between a knowing subject and a known object. On several occasions there was an opportunity to rethink this understanding. Above all, we had cause to criticise the attempt to establish knowledge as the knowledge of a lonely solipsistic subject. The modern subject does not however coincide with the subject of Descartes' methodological solipsism. In our postmodern present time we are confronted us with exaggerations that link legitimate criticism of the Cartesian subject to the announcement of a general "disappearance of the subject". And while one is at it, one can also allow the correlative, reality as the object of knowledge, to disappear as well. What has actually disappeared, or ought to disappear, is the idea that we could see reality as it is in itself, beyond the world as phenomenon, but this disappearance was already introduced by Kant more than 200 years ago. If the pan-fictional disappearers of reality now want us to believe that reality is merely a "simulation", they must have lost the distinction between appearance [Erscheinung] and illusion [Schein] [tr. note - a Kantian distinction]. The proposition that excessive skepticism is only the reverse side of a disappointed dogmatism proves once again to be well founded. It seems that many have not yet finished mourning for the metaphysical loss of absolute truth and of the knowability of the thing in itself. There can be no talk of the overcoming of modernity, chiefly because it has not been understood. The first step to change this is a critical reading of their classics.
I'll have to remember "panfiktionalistische Wirklichkeitsverschwindler" when responding to leftists who deny a critical piece of inconvienent history or science.
Well I who have least right to disagree with a reputable scientist... do not like it.
Obviously it is nice to have someone who has a good reputation validate your beliefs but I dont like lots of what he says:
His studies included shed loads of tree ring data series??
and then:
Last but not least a comprehensive physical understanding of climate variability over a thousand years and more, is dependent upon external climatic forcings being documented beyond reproach(which unfortunately is not yet clear as evidenced by the furious debate over the climatic forcing via the Sun). Only then can climate simulations over a period of more than 1000 years back in to the past be calculated on powerful supercomputers, through which questions such as the cause and effect mechanisms of the MWP, LIA and the current warm period can then be rationally explain.
His final solution is a computer model?
We only need to understand the sun and the a computer model will do the rest?
Many apologies to all those with greater scientific knowledge but just as I would not like Mann to agree with me, so I dont like this guy being on my side.
Reconstructions are needed when the truth is not available from a better source. We do not have photographic evidence or live witnesses available to tell us what happened in the Napoleonic wars. However we are all fairly confident that Wellington kicked Napolean's Butt.
How do we know this? People wrote it down, lots of people wrote it down.
Well people have been writing things down for thousands of years and they wrote about the climate.
The history books tell us that there was a Medieval Warm Period any scientist who then denies or agrees with that on the basis of a proxy reconstruction is an idiot. (forgive me)
Mr. Dung you don't have to ask for forgivness. If we needed any evidence that we are in an intellectual cul de sac climatology is that eveidence. Science it is not. Just arguments about statistical niceties, lawyerly discussions about what is is and untold amounts of money to wash it down with. Ultimate decdence.
Dung and Steiner:
I whole-heartedly agree with your comments. Moreover, mimicking the past is no proof of a model's accuracy; only when a model accurately predicts future behaviour of the phenomenon under study and over a sufficiently long time period, can the model be deemed reliable. As attributed to Yogi Berra, baseball catcher, "The future ain't what it used to be."
Morley Sutter
Hmmm I realise that I need to explain that I am not attacking McIntyre or our gracious host here. McIntyre was intellectually stimulated to question a leaflet pushed through his door which stated that 1998 was the warmest year in the last 1000 years.
McIntyre was a state prize winning mathematician and he investigated this using the tools he was most comfortable with.
Andrew Montford told the story of McIntyre's efforts in a way never done before and rightly deserves massive credit for his skillful writing and for moving our position forward.
However I am suggesting that common sense and logic can be at least as useful as science and computer models.
The science is straightforward and settled: if you at CO2 to the atmosphere you always get warming.
Well actually no you dont. From 1940 to 1970 we didnt get warming and from 1998 up today we have had no warming.
In the second half of every interclacial in the current ice age, CO2 rises and temperature falls.
Today's warming is unprecidented. Well actually no it is not: Medieval Warm Period, Holocene Maximum and the last 5 interglacials were all warmer than this one according to the IPPC graphs in AR4.
You can dig deep into the science or you can just say "That does not make sense".
Thanks for the translation work.
Bohm makes some interesting points but again we come back to a goddamn model. How are they going to check the numbers that come out of this model? If the model claims that the average temp in August in Saxony in 1284 was X degrees c then how do they check this?
Without a check then the model is like those jokey emails that prove that Santa Claus has to travel at mach 5 to visit 2 billion children in 200 countries in 24 hours on Christmas Eve.
Morley Sutter: " Moreover, mimicking the past is no proof of a model's accuracy; only when a model accurately predicts future behaviour of the phenomenon under study and over a sufficiently long time period, can the model be deemed reliable. "
That's right, even though the Met Office claims that ability to replicate the past has validated their models.
In the 1970's, when researchers were struggling to make reliable pattern recognition systems [machine reading of hand-written characters, automatic analysis of microscope images and so on] it slowly dawned on them why the systems did not work in the real world, whereas they had seemed very promising when under development in the lab.
The fundamental error they were making became known as "testing on the training data". Climate modellers seem to be fooling themselves in exactly the sam way.
"The "climate debate" has taken us [reasonable climate scientists] completely off guard" - Bohm
Really?
When the public image of the field for the last twenty years has been and is dominated by the dubious science and obvious advocacy of the IPCC?
When the field has enjoyed a gigantic influx of funding unprecedented since the end of the Cold War, due to the claims of the IPCC and its media flunkies?
When environmentalists and politicians are now clamouring to turn off the power plants and knock Western populations back to the living conditions of the nineteenth century by 2050 or sooner?
And he's surprised there's acrimonious debate?
I trained in science and married a scientist. I am well aware that it is almost impossible to overestimate the political naivety and tunnel vision of working scientists. But this takes the biscuit.
To avoid duplication of the considerable effort involved and to pool expertise, a good way to deal with translation projects like this is to make them collaborative. I've set up a translation wiki at http://mwpblog.wikispaces.com. where you can see a translation of William von Baskerville's most recent (1 August) blog entry, along with the version of the Boehm interview to which I linked a couple of threads back. It could be a good place to put translations of other posts from the MWP blog, at least initially. Anybody who wants to join in the fun just needs to register on Wikispaces as a user and then to send me (Wikispaces user mwpblog) a message so that I can give them editing rights.
[BH adds: good idea - I've put a head post up]
All the time I've spent trying to follow this tree-ring-circus, it seems more about confirming the ringmasters predetermined conclusions than it does about science. Does anybody besides the people involved in studying tree rings actually think they're useful for anything?
Beside the Bohm-fire :)
"The "climate debate" has taken us [reasonable climate scientists] completely off guard" - Bohm
Really?
The original German is better rendered "we've all been rather overwhelmed by the climate debate", the implication being the amount of work involved rather than surprise. In this kind of translation it's best to steer close to the original even if it makes for rather wooden-sounding English :)
I also like the 'tree-ring-circus' :)
I like it because:
The first part touches on post-normal vs normal science, the role for scientists as 'honest brokers' vs appeals to authority and the levels of certainty in paleoclimatology. That touches on current debate at Pielke Jnr's and collide-a-scape and also what we learned from Climategate. Those emails showed disagreement between Briffa et al and Mann et al regarding the degree of certainty for the 1,000 year reconstruction, with Briffa urging a more cautious statement, then getting leaned on to support a statement the evidence did not support. Reconstructions show the existence of the MWP and LIA, as a global phenomena but based on those reconstructions, amplitude is less certain. Yet as Dr. Böhm puts it
doh. pesky quotes :)
<rant>
Give me 1000 points of data as a function of time and I will fit an almost perfect model to it. It is call polynomial regression in the form of Y = (X**N)+ f(X**N-1)+ f(X**N-2) +f(X**N-3) .... where f are scalar coefficients.
What does it mean -- probably not a goddamn thing, which is one reason it is not used very much. However, it will fit the data beautifully.
And because the fitted curve will match the experimental data, the correlation coefficient (which is nothing more that the cosine of two vectors) will be close to 1.
And the big test is will it predict the future? Possible for the sort term, but not very likely unless you do have a multi-cyclic phenomenon.
So much for modeling. A waste of time. If someone does have a model that "works", then have him tell me what the weather will be like for the next five years and then let's use wait and see. "Predicting" the data you already have is merely a matter of calculation.
If Social Psychology (which depends on Factor Analysis, which is almost the same as Principal Component Analysis) is a "soft science", Climate Science is "diarrhea science."
And in case you are wondering, I spend several years at the Cornell University computing center helping candidates in the "social sciences" create a thesis based of a random Eigenvector spat out by the computer.
</rant>
AH:" I also think it's the same with climate models. They're known for limitations, but are improving. They can help test what we know and don't know, but it's like Martin A says about testing on the training data. The training data for climate science should be the historical data. If they can use that to model the past accurately, predicting the future should be more robust. If they can't, then it's just the usual GIGO problem."
The problem is that the past records are the training data - ie what the Met Office et al use to tweak their parameters. They then go and say "look, our model correctly predicts what happened in the past". This amount to testing the system on the same data used to train it.
A model can be a very false representation of physical reality, yet, with enough fiddle factors to be twiddled, it can replicate the data it has been tuned to replicate with great precision.
Unless you have a very precise physical model for the system whose behaviour you want to predict (eg such as exists for the dynamics of a satellite in orbit for example), your tweaked fiddle factors will fail you as soon as you move out of the zone over which it has been tuned to give the right results.
One reviewer of the Hockey Stick Illusion on Amazon (who gave the book 1 point, and admitted that he had not actually read the book) stated "(...) this concept global warming deniers have that climate models are worthless is extremely irritating. Climate models are extremely sophisticated (run on the fastest supercomputers in the world) and have proven to be highly accurate." He did not mention what was the proof of their high accuracy.
My experience of the extreme difficulty of validating models, even for relatively simple and well understood engineering systems, convinces me that climate models are inherently incapable of being validated to the extent that they can predict future climate. I think people who believe in their predictions, like the Amazon reviewer, are quite simply profoundly deluded.
I'm sorry for the Amazon reviewer but his irritation is probably going to continue...
Don Pablo de la Sierra "climate science is ..."
Haha.. I'll stop laughing soon. I will. Hahah....
"A model can be a very false representation of physical reality, yet, with enough fiddle factors to be twiddled, it can replicate the data it has been tuned to replicate with great precision."
The basic limitation I see with modelling is we don't understand clouds at all. There probably is a simple parameterisation that exists that generates the correct response for water vapour and clouds in response to an increase/decrease in energy in the system. With this I reckon a fairly simple model could predict fairly accurately the correct feedback and forcing to something like increased co2. The problem is that it is a circular argument and your results are critically sensitive to the parameterisation you picked in the first place and rest of the model wouldn't be that important. You will very likely know from the parameters you put in which direction the feedback will be.
Basic intuition say the water cycle acts as a negative feedback. ie dry deserts are very hot in the day and cold at night (say 0 - 50 degrees C). Very humid tropical climates (say Singapore - say 23 - 35 degrees C) live in a narrow zone of temperatures, being very hot and humid in the day and hot and humid at night. This to me says water moderates temperature changes, so I would need to be convinced that it acted as a positive feedback.
To argue that computer climate models are getting better all the time is pointless if it only means that they fit past data better all the time. What's important about them is their predictive ability of which so much has been made and which appears dubious.
The point about computer climate models is that they are infinitely and subtly tweakable and it's easy to present the results they churn out as having a ring of authority, and treat it as evidence. An ideal propaganda tool that few are in a position to question.
This reminds me of a comment made about enthusiasts of the far-seeing prophecies of Nostradamus, expressed as cryptic verses, which could mean anything and probably mean nothing.
"They're great at predicting things which have already happened, but not so hot when it comes to the future".
Man always dwells upon what he knows and starts to work out how the world works on the basis of that knowledge.
Look back 1000 years at what man knew then, the gulf between then and now is huge beyond description. To all intents and purposes man knew absolutely nothing 1000 years ago.
However look forward 1000 years (if we make it that far) and the gulf will be greater by far.
Scientists will look back to this time and laugh themselves silly about the fact that man thought he could predict and control the Earth's climate.
One reviewer of the Hockey Stick Illusion on Amazon (who gave the book 1 point, and admitted that he had not actually read the book) stated "(...) this concept global warming deniers have that climate models are worthless is extremely irritating. Climate models are extremely sophisticated (run on the fastest supercomputers in the world) and have proven to be highly accurate." He did not mention what was the proof of their high accuracy.
That would be last Winters record cold in Scotland which the Met Office accurately predicted (Sacasm off ;) )
JohnH That would be last Winters record cold in Scotland which the Met Office accurately predicted (Sacasm off ;) )
I was surprise to read (in their own publicity somewhere) that the Met Office use the same model for long term weather forecasting that they use for climate prediction. But I thought it did explain why their long term weather forecasts are biassed in the direction of barbecue summers and mild winters when what we actually get are mild summers and deep-freeze winters.
The Met Office is funded to the extent of £170M annually for producing climate change propaganda, 'climate change advice' etc. Any Met Office programmer, responsible for tuning their climate model, who tuned it to predict no Global Warming would simply be tuning themself out of any prospect of a bonus/promotion/salary increase. As someone said "follow the money".
Last winter's cold snap did generate a good deal of heat for the Met Office. Perhaps that is what he was thinking about.
Gimme a break, Dr Reinhard Böhm ---- the Met wasn't even CLOSE with their model..
The in scheme of things, who is John M. Scully? He is the most recent 1-star review of HSI on Amazon and indicates that he was involved in some of the GW debate back in the 90's. I would be interested to have someone more knowledgeable than I read his posts and comment on them.
The benefit of Amazon is that it is also read by an interested public who might not otherwise get involved with the dialogue, and the back and forth there has been interesting.
C Johnson The in scheme of things, who is John M. Scully?
The "review" in Amazon appears to be the same as one posted on Realclimate by someone under the pseudo "Rattus Norvegicus" - apparently a frequent poster on the site for "Climate science from climate scientists".
The torrent of hatred posted on Realclimate and directed at Judith Curry, following her comments on HSI have to be seen to be believed. If anyone doubted it's a religion, the torrent of invective from The Faithful leaves no further room for doubt.
Partly it's a new religion, but without a doubt there's an aspect of followers of an old religion, Marxism, coalescing around a new nucleus. The scope this offers for social control must be absolutely mouthwatering for them.
Martin A: "The problem is that the past records are the training data - ie what the Met Office et al use to tweak their parameters. They then go and say "look, our model correctly predicts what happened in the past". This amount to testing the system on the same data used to train it."
I don't necessarily agree that's a problem. The problem to me is the predictive powers of models have been very much oversold to the public. That isn't a new problem.. 20 years ago someone wanted a 3-D terrain model of the whole Earth surface at 1m resolution for mission planning, ie taking the output and loading it into navigation systems. We explained that was rather ambitious. Climate models don't work at anywhere near that resolution, yet we get told they can predict at 1km or sometimes even sub 1km accuracy. Yet many climate models are just atmospheric models. Some are ocean-atmospheric models. Some may include some land variables. But most exclude the hard to model variables like clouds, aerosols, biological feedbacks and random events like volcanoes. They don't model clouds well, yet we're told they can model precipitation. They're supposedly 'robust' in their physics modelling, yet people are still debating if cloud or water feedbacks are positive or negative. Given the level of current uncertainty, why are they being used to plan trillion dollar policy decisions?
So far, I think the fiddle factors are being abused to 'prove' their predictive ability when they're best used to prove uncertainties. If the historical data is questionable, they may not even be proving that.
Training data may also be why we're expected to gloss over the CRU's problems with their data quality. As I understand it, many models are trained against climate observation data from the IPCC's DCC, which are 11 variables from 1901-2000 supplied by the motley CRU. If that data isn't 'robust', then it's again back to GIGO. I think there's also a problem with model inbreeding given training 'data' for predictions is model generated (ie Worldclim) so errors get amplified.
I still don't think we should abandon models as being too expensive or too difficult given we still need accurate forecasting, but climate science and especially it's publicists do need to be more honest and open about their current capabilities and certainties.
Also curious if anyone's been investigating any merit in combining analogue and digital computing for climate modelling given the world's mostly analogue, and analogue may model physical processes more accurately.
Atomic Hairdryer
There is the old saying "garbage in, garbage out".
As for the modeling going on, none of it appears to consider that some 93% of the total free carbon (not locked up in rock and other minerals) is dissolved in the oceans of the world, and that the carbon dioxide moves back and forth into and out of the sea water all the time. In sort, the oceans of the world are a gigantic buffer for CO2, and when dissolved in the oceans, it can not possibly have an effect on air temperature. Reference Here
In short, they don't even have their hands around the basic physics, the fundamental mechanisms, or even how to collect temperature data. And they are going to build models?
As for an analogue computer, they are pretty much dead as the dodo bird. Still I would like to see such a beast. :)
Don Pablo de la Sierra
"In short, they don't even have their hands around the basic physics, the fundamental mechanisms, or even how to collect temperature data. And they are going to build models?"
Modelling of CO2 exchange between the atmosphere and the oceans is another scandal waiting to surface. Invented physical effects eg "evasion factor /buffer factor /Revelle factor" are taken as proof that CO2 injected into the atmosphere remains there for centuries - with a clear AGW agenda.
(Despite the clear physical evidence from the impulse of C14 injected into the atmosphere by the last atmospheric nuke tests in 1961/2 that CO2 injected into the atmosphere declines as a simple exponential with a time constant of a couple of decades. See http://cdiac.esd.ornl.gov/trends/co2/welling.html and the numerical data it presents for C14. Put it in a spreadsheet and plot the log of the C14 measurements - you get a dead straight line.)
For a summary of the baloney see http://folk.uio.no/tomvs/esef/ESEF3VO2.htm
Atomic Hairdryer - I think that you and I may be talking about two different sorts of model, hence the differences in view.
I think you may be talking about what I would call static models - eg a model for the altitude of every point on the earth's surface (points at 1m intervals, say). I call these static models because they represent things that do not change with time.
I am talking about models that represent how things evolve with time - the dynamics of a system. Constructing models dynamic models is hard for anything more than the very simplest systems. Validating the models is harder still.
If you don't understand the dynamic behaviour of the system you are modelling, your efforts to produce a model to predict its future behaviour will fail. And the dynamics of the atmosphere, not to mention the dynamics of the ocean (and the interactions between the two) are far from understood in the detail needed to produce useful simulations capable of predicting future behaviour. As the Met Office makes clear with each "barbecue summer" prediction.
I am aware of the fallacy of saying "I cannot imagine xxx, therefore xxx is impossible".
All the same, my view is that "climate scientists" attempting to model the evolution of the climate have simply not grasped the difficulty of the problem. This is illustrated, to me, by the Met Office's confident statements that heir models been validated. If they believe that, they are truly deluded. The apparent belief that more powerful supercomputers will solve any remaining difficulties is just another facet of their self-delusion.
By the way, I don't think analogue computers are the answer. They were used for simulating dynamic systems described (or approximated) by ordinary differential equations. Everyone heaved a sigh of relief when digital computers with the right software arrived in the sixties and seventies and the analogue computers went to the rubbish dump.
The accuracy of an analogue computer was - at best - about the same as you'd get from a digital computer using fixed point arithmetic and a wordlength of around 12 bits. It sounds paradoxical but an analogue system can be simulated with immensely greater precision by using a digital computer than with an analogue computer.
But, in any case, it's the lack of correspondence between the model and the physical reality that is the problem, not the limitations of the computer used to get numbers from the model.
Martin A
It is far worse than you say. I said 93% of the "FREE" CO2 was dissolved in the waters of the world. That is CO2 in the form of trees and plants and animals, atmospheric CO2 and other forms that can be easily converted one to another. Most of the CO2 in this world is locked up in ROCK -- limestone and marble, with quite a bit of chalk thrown in as well. I don't have those numbers yet, but I am looking for it.
Now when it comes to O2 +C <--> CO2, you do need green plants, but that should include much of the algae in the oceans as well.
There does need to be a lot of honest research done on these issues, but I haven't found it yet. It is mostly biased to the AGW mantra.
And analogue computers aren't dead -- just hidden. Look at you power meter on the side of your house. True the newer ones are all digital, but I still have mine until next week. Then I join the "Smart Grid"
Still, you are quite right about those that we use to use to "calculate". Almost as accurate as a slide rule.