Something to say? Say it here.
I’m still wrangling climate models, trying to figure out exactly how they can take quality input and yet still manage to turn it into complete garbage. I’ll try to get a preliminary report out soon.
All welcome in the sandbox as long as you all play nice. Any biting and name-calling will land offenders in the troll bin as I swoop by. New posters trapped in moderation, please be patient, someone will attend your comments as soon as practicable.
Have fun. Normal services will be resumed as I get time.
Well, here’s a random thought based on the taming of smog and other industrial environmental and health concerns. It seems that over my lifetime, the severity of any proposed fix is inversely proportional to the severity of the problem. Or – you’ll know its junk science when …
Originally, the goal was to reduce the product of combustion to harmless carbon dioxide – as taught at my college by physics professors in 1981. Now, the goal is to eliminate harmless CO2, trace levels of mercury lower than that found naturally, and dust/PM2.5 by eliminating combustion.
Originally, the goal was to reduce water shortages by increasing the supply of fresh water. Now, the goal is to eliminate an indeterminant threat to the delta smelt by eliminating fresh water for consumption, agriculture and energy.
Originally, there was DDT ….
And the severity exponential multiplier is that they now always want these things done by governments – which are in turn governed by the UN. And governments, in need of crusades to increase their growth, are listening.
Coach, I remember those same lectures. The ultimate goal was that combustion would actually result in CO2 and H2O, and nothing else. Of course, that’s an impossible result, but we’ve come amazingly close. So, now the target has been moved, and we can’t have CO2 any longer, and of course all those other impossible to remove traces of “stuff”, mercury, small particulates, et al. are the new bogeymen. The shame is that we didn’t see it coming. We, those of us who engineer and study science to do the right things, took them at their word, and worked our a$$es off to solve what they claimed were the problems. Now they claim the solutions are the problems. We were duped and hoodwinked.
The advantage is that now we know that they were dishonest from the git go. We won’t be fooled a second time.
Anything touched by government is always in grave danger of the Sadim touch – you know, the reverse Midas touch where everything turns to crap. The IPCC, the EPA, et al, are iconic purveyors of the syndrome.
Let me add that some act as if nothing were done in the last 45 years. They preach “pollution” like it’s 1967.
Me thinks we wasted trillions. Some was well spent, but the GDP of 50 countries was spent on nothing.
“I’m still wrangling climate models, trying to figure out exactly how they can take quality input and yet still manage to turn it into complete garbage. ”
That’s self evident. Any model that depends on self modifying parameters (i.e. mimics feedback) is going to fail in exponentially with the number of parameters added.
Modeling a building in an earthquake is pretty simple, modeling the movement of each leaf in a tree in a storm is impossible.
While true Petrossa and well-recognized among adherents to the scientific method models are not generally seen as implausible if not impossible predictive tools by the general population and media.
What we need is to educate the public (and media) and to do this we must prove CGCMs mess up specific calculations by X margin because Y and blunt statements of the obvious will not do.
NCAR’s RRTM (rapid radiative transfer model) actually delivers sensible results – they claim within 1.5W/m2 of observations but they might be working with better precision than I am, I seem to get boundary issues of up to 3 W/m2. The major issues must be in the integration or the climate sensitivity parameter because while sensible forcings (∆F) are fed to CGCMs absolute garbage comes out in the form of surface response (∆Ts), which is how we get the absurd warming “storylines”.
Problem is though that that represents a tremendous number of lines of code to examine and prove for a single GCM coupling and there are 18 CGCMs in the suite feeding results to the IPCC. Then there are the varied integrations of dynamic, slab or static oceans, aerosol and particulate simulators…
Then there’s the problem of eye-glaze – that’s where reporters and the general public who have not followed the arcane details of climate science find such details overwhelming and retreat to the default position of “Oh well, the experts says it’s right…”.
The IPCC’s documented climate sensitivity parameter is wrong. Dead flat wrong and that may be the best approach at the end of the day – the IPCC’s stated use of a false value for the parameter and a highly simplified calculation demonstrating it to be impossible.
I don’t think this is something which can be explained to the general public nor to non IT versed scientists alike. The level of abstraction is just to high. You can compare it to fMRI. 1000′s of medical specialists use it regularly without having the faintest clue it’s not a ‘real’ picture they see but a virtual one which may or may not reflect reality.
I had several fierce arguments in fact with some fMRI users trying to explain how it couldn’t really work. And those where highly intelligent, well trained professional scientists.
Since compared to a climate model, fMRI is very simple, it’ll be impossible to explain even to the writers of climate models to get them realize it’s a lost case.
Attacking models is a Don Quichote action i guess. The only way IMHO is to keep on hammering down the point that none of them can even predict the past without feeding them actual real recorded data. And even then it’s not that hot.
Barry Quixote, that’s me
I’m afraid Petrossa that I wrongly believed that demonstrating failure um postdicting the past would show people the lack of prognostic value in CGCMs but I overlooked one simple thing. We are enormously outgunned and as fast as we could highlight a failure entire paid web sites would spring up with pretty pictures and rent-an-expert figureheads airily claiming we were wrong because of some “sciency-sounding” gibberish. There’d be a flood of press releases and placed media articles and we’d get hate mail as agents of whatever Big-Evil was in vogue that week.
This is what has driven me to seek one simple commonality in climate models where we can demonstrate raw error.
Editor, you might try reading an undergraduate level process control textbook about feed-forward (predictive) control. It might give you a good notion about how to present this to people with a marginally technical background.
The idea is simple, you look at your inputs and determine how to control your process. However, it’s useless without feedback (reactive) control as well because your model is inherently flawed. If you don’t have feedback (where you look at your outputs and adjust parameters accordingly), you will at best be sub-optimal and at worst, lead your process into a destructive state.
I like it Ben of Houston, although I fear it assumes audience skills not in evidence.
That’s probably really harsh and a product of being extremely tired. I’ll browse the bookshelves on process control because you are right, thanks for the suggestion.
The model is designed to fit the agenda. Truth in – lies out. It’s all pre programmed.
That’s perhaps a little too harsh, scizzorbill. I know some researchers using climate models whom I trust implicitly. The problem is that they are presented with a black box model and told that it works. They are not given access to code nor processing windows in which to run boundary testing.
With the added problem that the creators of the black box themselves are incapable of determining if and how the black box works.
At a certain point the complexity of the code involved (most of it written by computers) surpasses the capacity of one person to encompass it all.
So the model is computer tested by another program to see if it works.
In the end no one knows what it does in detail or how it does it. So even if given the code you’d get nowhere fast.
Even the code for the spaceshuttle, which by now dates about 20 years, is of such a complexity it can’t be comprehended by humans.
Windows is totally beyond control. Not even 10% of it is actually handcoded.
True again Petrossa, in a lot of cases although most requested GCM and OACGCM model code has been provided on request. Most helpfully on digital media and a couple in turnkey format ready for off-the-shelf compilation. That said several were provided as hard-copy printouts, roughly 30 file boxes of them and a few had been grepped and every comment stripped. They complied with the letter of the law, at least but 300K lines of mixed language code is not possible to navigate without programmer comments and declarations of intent so even the modules are useless for the purpose.
Yep. That’s how the Dutch government treats FOI requests. You have to pay 10 cents for each fotocopy and they send you 30.000 somewhat related documents with a bill. That way you learn quickly not do that again
But on the upside:
http://www.jpost.com/Sci-Tech/Article.aspx?ID=273385&R=R1
More will follow.
XML Parsing Error: not well-formed
Location: http://www.jpost.com/Sci-Tech/Article.aspx?ID=273385&R=R1
Line Number 40, Column 46: if(document.body.clientWidth < 1360)
—————————–
Works fine with me but here is a tinyurl: http://tinyurl.com/d2vxp8v
tinyurl worked. Thx.
‘Israel not committed to renewable energy future’
By SHARON UDASIN
06/11/2012 02:02
“We must learn from solar industry mistakes of other countries,” experts tell Herzliya conference sponsored by private firms.