Monday
Apr162012
by 
Bishop Hill
Bishop Hill Quote of the day
Apr 16, 2012 Climate: Models One climate modeller we interviewed explained that the climate is a ‘heterogeneous system with many ways of moving energy around from system to system’ which makes the theoretical system being modelled ‘tolerant to the inclusion of bugs'.
From the Pipitone and Easterbrook paper on validating climate model software, currently the subject of a guest post at Judith Curry's.
Reader Comments (129)
And tolerant of being wrong as well presumably.
‘heterogeneous system with many ways of moving energy around from system to system’
None of which we properly understand, but we've modelled it anyway...
Re: Breath of Fresh Air
Tsk Tsk. Climate models are never wrong. Its the world that needs adjusting.
Nothing is wrong. Everything needs adjusting (that is certainly true). Sic transit merde (pardonnez-moi pour mon Francais).
The mechanism by which energy moves around the system is the nutrient distribution for phytoplankton growth. That in turn controls the local cloud area and albedo hence solar energy in, particularly the short wavelength, high energy stuff that penetrates deeply into the ocean.
So, it's a bit like a patchwork quilt as the phytoplankton, the dominant organism, control their environment, even to the extent of influence ocean currents.This is far to complex to model because it is in essence an organic computer anyway.
CO2 was thought up in the mid 1970s as a pathway into a revival of eugenics, Get on the wrong side of the phytoplankton and they'll do the job far more effectively that the next highest organism!
A complex dynamic system is one of three or more mutually interacting variables. In global atmospheric physics, major factors include solar radiation, thermodynamic coupling, above all plate tectonics-- the latter conventionally immune to climatological scrutiny as too long-term.
Any such complex system will by mathematical-physical definition exhibit random-recursive, chaotic-fractal behavior absolutely indeterminate in detail due to Edward Lorenz's celebrated "butterfly effect" (1960). Researchers professing otherwise ipso facto disqualify themselves from serious consideration as ignoramuses refusing to acknowledge ineluctable principles underlying the very nature of their disciplines.
Since "climate studies" by definition have no empirical foundation --no-one "experiments" with long-term weather patterns-- the specialty is a classificatory exercise akin to botany. Where formulaic patterns exist only in the eye of a beholder, so-called models are but circular-reasoned reflections of investigators' pseudo-scientific prejudices, no more respectable than Aristotle's "impetus" or Ptolemaic epicycles.
To the extent your garden-variety Luddite academic opposes post-Enlightenment industrial industrial/technological civilization root-and-branch, just so will extreme-radical reactionaries such as Ehrlich, Holdren, Mann, and Hans Joachim Schellnhuber call for totalitarian World Government with themselves as arbiters of every humane interaction. Faugh.
Didn't the writer Ray Bradbury invent the "butterfly effect" in his 1950s story A Sound of Thunder? I'm sure Lorentz worked out a mathematical rationale behind it, obviously.
'tolerant to the inclusion of bugs') is peculiar English, and needs some speculation to make some sense of it.
My take is it means that the software involved is so poorly understood that it can function well enough to fool some of the people most of the time even when there are appreciable defects in it. Now those defects could be in poor code, poor concepts, poor structure, poor assumptions, poor inputs, and even in poor outputs if intermediate processing is done of the results before they get sent to the printers, so to speak.
Not the stuff to base squillion dollar decisions on, nor to frighten the children and horses with. It belongs under intensive care in the research lab, where the pecularities of each run will motivate the modfications for the next one into the indefinite future - all the time with defects doing their dastardly deeds.
I think it is worth pasting in the conclusion of Judith's anonymous reviewer:
Source - http://judithcurry.com/2012/04/15/assessing-climate-model-software-quality/
It's also worth reading the comment by Robin. Latimer Alder is on form too. It would appear that climate models are about as fit for purpose as the proverbial chocolate tea-pot.
I have absolutely no idea what this post is about
Read George Orwell's essay on the English language:
http://www.resort.com/~prime8/Orwell/patee.html
A friend asked me recently to edit his thesis on a market research paper he is writing. Part of his paper was a quote from an academic which I just could not understand - it made no sense.
I get the same sense here.
OK. I have got to learn more about climate science!
@sankara
No need to learn any more about climate science. I think you have incisively got to the root of the matter already! A lot of it is academic gibberish.
"...tolerant to the inclusion of bugs." Remarkable.
I translate that as being: "Hey - it's all total crap. But it's good enough for "climate science".
Latimer Alder
I recently read the HSI and our host is a model of clear writing
Why can't the English learn how to speak!
(Apologies to G. B. Shaw for taking him out of context!
Richard Black - thinks a very sophisticated lobby exploits doubt -
quoted here:
http://finchannel.com/Main_News/B_Schools/107319_The_media_and_climate_change/
Richard said: “What we saw after that was a number of editors convinced climate change was a ‘scam’. One member of the press talks about being cold-shouldered by editors and being accused of wasting time and resources to get on air. We were back at the end of the bulletin, if at all.”
Richard said the subject often suffers from appearing “boring”, and spoke of the challenges facing the traditional media as staff numbers are cut. According to the University of Liverpool, but he encouraged researchers and academics to come up with innovative information to keep the issue on the agenda.
Richard added: “It is intrinsically difficult to try and get people interested in climate change because many of the most significant impacts are in the future, and exist in other countries. There is a lot of room for doubt, and a very sophisticated lobby exploits that doubt.”
Jimmy Haigh, assuming that the crappy software has no built-in bug bias, and given the overall complexity of the problem, it is reasonable to assume that the bugs "cancel out" so overall have no impact.
Now all I have to do is convince my QA manager that this works for the software I'm currently developing, and the test cycle should be dramatically reduced!
Lapogus quoting the anonymous reviewer:
Not so much a nail in the coffin as a stake through the heart for the Pipitone and Easterbrook paper. But encouraging that someone somewhere thinks GCMs need a defence. Enough people have heard that this foundation of the global warming scare is questionable to make that necessary.
Mdgnn:
The mechanism? Please treat as a rhetorical question. Reviewing the Pipitone and Easterbrook paper and thus GCMs is surely the point here. But mdgnn has a habit of taking a position much more foundational in the science than the subject in question and making cryptic claims that would deserve one of the following from the host here (or any climate blog):
1. A thread all to themselves
2. To be ignored until they have been expressed in a full paper.
I haven't spotted any interaction from mdgnn with Jonathan Jones and Simon Abingdon's thread which put some of his earlier claims under the spotlight, which had been expressed in vast numbers of threads until that careful critique was made. As Jorge said an hour ago:
But it seems mdgnn has moved on. I write to point out this behaviour. May I emphasize that I don't want this thread to become a place to discuss mdgnn's latest claims or indeed his behaviour in general. But what does one do?
Alternatively...
Given the overall complexity of GCMs, the fact that they give results somewhat in-line with expectations shows that any bugs present must be very minor in impact, as any non-trivial bug would swing the outputs so far that it would be laughable.
Apr 16, 2012 at 3:58 PM | steveta
They already are laughable.
Steveta; the models are fiddled by using twice low level cloud optical depth and a variable net AIE currently exactly the same but opposite sign 'forcing' to the claimed AGW.
Because the latter is based on incorrect aerosol optical physics [I have written the paper explaining why], no climate model can predict climate and to persist in this farrago is a waste of money and effort.
PS, I have contributed to the discussion thread with an authoritative explanation of the 'two disc problem'. 'Back radiation' does not exist, cannot exist. This whole exercise of climate modelling must be restarted with professional oversight.
Mydogsgotnonose has begun to reply to Jonathan Jones on the Radiative Transfer thread in Discussion. That was exactly 15 minutes after I pointed out here (still in moderation) that he hadn't done so. I'm not sure I fully grasp the answer. But consider my comment in moderation corrected on that point. Game on. And sorry for the interruption :)
PPS. those who are trying to support the claim that the 'two disc' problem proves 'back radiation' exists have forgotten the most basic part of radiation physics, which is that Kirchhoff's Law of Radiation only applies at thermal equilibrium, and passing heat to space, an infinite sink for radiation, is as far from equilibrium as possible.
Because of this I am working out a revised way of looking at conduction of heat through the atmosphere, which is presumably the reason for the two-disc model in the first place. Now I don't know what passes for erudite academic discussion nowadays, but since my time it must have deteriorated quite a bit to judge from the comments on that thread.
Did Jonathan Swift write about a similar failure of scientific analysis?
Now now, dog. If you really have proved something everyone else has missed, being smug and obnoxious about it is not the way to win people over to examining your proof (which I've glanced at and don't understand)
Hi Richard; my answer is fairly obvious; there is no thermal impedance in the direction heater to final radiation to the enclosure so no temperature drop.
And because there is infinite thermal impedance in the reverse direction there can be no 'back radiation'.
Game, set and match methinks........;o)
[Sorry, but I was taught by some very good people how to really solve problems; you do it by simplifying them to the essentials.]
I am afraid that I got into trouble for the comment below on a previous thread. However (at the risk of being snipped) it is probably more appropriate here:
Unverified and unvalidated computer models can never provide useful information on the state of a system. Numerical simulations containing a large and uncertain number of dependant and independant vairiables, acting within a non-linear and sometimes chaotic system, can have no predictive power at all.
Mydog, I will say this, that naive assumptions of thermal equilibrium dog climate science as far as I can see. As for the rest, I await the academics of whom you speak so highly and lovingly. You are obviously seeking to wind them up. But you took four days to respond to one of Jonathan Jones's posts yourself. I would suggest that the others take at least one deep breath before trying to express themselves on how to make head or tail of what you've written.
BigYin: I hope to have presented things as tongue in cheek but what everyone must understand that science is right or wrong and sometimes to get to the right answer requires a tremendous intellectual effort to pose the problem in its simplest way.
My emphasis has been to emphasise this very important bit of physics not myself. Apologies for giving the wrong impression.
Hi Richard. I was busy elsewhere. However, what I perceive from this argument where it seems I was confronted with a party trick carefully calculated to 'prove back radiation ' exists, is that there has been a serious failure of physics' education for so many to have been misled, Never assume anything. In radiation problems, emissivity and absorptivity may not be equal.
It helps having pioneered the small particle part of nanotechnology because the interaction of radiation with small particles also gives surprising results at times.
My dog
Please could you stop posting about your theories on every thread. Please now keep this to the discussion forum.
Thanks
One stumbling block for those who try to model climate is that temperature doesn't tell the whole story. Heat is the important variable. Worse yet, temperature is a poor measure of transient heat. High temperatures in ENSO and other phenomena represent heat shedding mechanisms. Measuring global temperature, as currently done, is like attempting to estimate the size of the crowd inside a soccer stadium by counting the people coming out of a few exits.
@jorgekafkazar
'Measuring global temperature, as currently done, is like attempting to estimate the size of the crowd inside a soccer stadium by counting the people coming out of a few exits'
Tough. It is (supposedly) temperature that is the thing that will bring down all the woes of the world upon our shoulders if we don't modify our evil lifestyles by three weeks come next Michaelmas threefarthing...or whatever the dreadful threat is this week.
And we all pay climate modellers good money with lots of toys for them to play with to get it right.
So far their track record of getting anywhere even near what actually happens is about as good as I could do with graph paper and a copy of Arrhenius's second greenhouse paper. And at far less cost. (*)
Since there are no theoretical breakthroughs even on the radar - let alone arrived, it doesn't seem likely that climate modelling will ever be a fruitful line of enquiry. We could save a lot of money by closing down 90% of it immediately.
(*) On reflection, Arrhenius outperforms any of today's models. He wasn't seduced down the primrose path of 'feedbacks'.
"One climate modeller we interviewed explained that the climate is a ‘heterogeneous system with many ways of moving energy around from system to system’ which makes the theoretical system being modelled ‘tolerant to the inclusion of bugs'.
I do hope the "tolerant" cove doesn't get a job with Boeing, he seems to be giving considerable latitude to the notion that software should be, if possible, bug free.
OK BH......
This gave me flashbacks of having to learn Z notation. If I had to do that as an engineer, climate modellers should be forced to as well. No development funding without Z specification!
I have just read the first few pages of that paper. And then skimmed through the rest.
I am just dumbfounded.
I have spent half my working life testing software.
I just do not know where to begin with that piece of garbage. I do not think I will even waste my time.
The list of things that are just not done in climatology grows:
-Revision control
-Code reviews
-Validation
-Peer review
-Statistics
-Application of the scientific method
-Debugging
(I managed to truncate the list before getting to 'knowing how to plot a graph in Excel')
climate is a ‘heterogeneous system with many ways of moving energy around from system to system’ which makes the theoretical system being modelled ‘tolerant to the inclusion of bugs'.
We have a few ideas as to what components might be involved. How it might all fit together is a mystery and we have absolutely no idea as to what weight we might assign to anything.
Homo erectus ... meet A380 schematic . (and for doubting 'the field' I'm a 'flat earther'?)
3*2,
The difference is that if A380 safety critcal software in the multiplexed guidance, navigation and control computers fails due to "bugs" - people die! Not so with "climate models".
Oh, hang on a minute......
Mischievous obfuscation- beloved by politicians, climate activists and used car salemen. The English language can be massaged pulled and pushed to suit, just like climate graphs and models. They are orchestrated together into a symphony of spin.
Take the humble word 'sanction'
(def from the FreeDictionary) - Occasionally, a word can have contradictory meanings. Such a case is represented by sanction, which can mean both "to allow, encourage" and "to punish so as to deter." ... English has a few other words that can refer to opposites, such as the verbs dust (meaning both "to remove dust from" and "to put dust on") and trim (meaning both "to cut something away" and "to add something as an ornament").
egregious - both shocking or surpassing
Anyone with a knowledge of exponential error accumulation would know that.
2 weeks is all that's needed to turn any recursive prediction into a random number generator.
jiminy c....well at least help out the anonymous commentator at climate etc to get the most precise refutation possible of the paper.
There's 2 levels of Error.
1 Static code analysis should reveal sections of code error.
2 The computer model does not correspond with the real world.
I guess both errors are present.
The Bishop has just been accused on Twitter of quote-mining. Too bad this is what follows in the original: The combination of both factors means that code defects are either made obvious (and so immediately fixed) or made irrelevant by the nature of climate models themselves and therefore never reported as defects..
And Pipitone and Easterbrook approve of that most supremely idiotic statement.. In fact immediately before the quote they write: At the same time, we have evidence that climate model behaviour is robust.
"It is often impossible to provide complete requirements for the software upfront, and in fact, the requirements are expected to emerge and change over the lifetime of the project."
This sounds like every development project I've ever worked on. There are well proven techniques that help with this, in particular those associated with Agile methodologies.
Ironically, one thing we Agilistas are keen on is Test Driven / Test Assisted Development.
omnologos - those quotes are amazing - it is like watching a WC Fields film - we have got everything wrong but the models are robust. The walls are collapsing but our models are robust.
let us see how the warmistas will spin it...
the fact that our software was shite made no difference to the result. The World is warming beyond control.
many other researchers used our shite software and produced the same results. Hengist loves it.
MANN software...9 out of 10 idealogically obsessed climate researchers use it
What is a bug in the GCM software? Anybody knows?
Richard Drake wrote:
Indeed. Although it is possible that Pipitone and Easterbrook were influenced by the sage advice recently offered by Richard Black (as Barry Woods noted above):
In the absence of any definition of "innovative information", one might reasonably infer that Black would applaud the efforts of Pipitone and Eastwood; the creative writing exercise in which they appear to elevate the mediocrity of the models might well constitute "innovative information" ;-)