CSIRO: warming up to five degrees by 2070 Even an La Niña event and cool weather in Australia in 2010 and 2011 haven’t reversed the overall long-term trend to a warmer globe, according to Australia’s latest State of the Climate report. The report was assembled by Australia’s peak science body, the Commonwealth Scientific and Industrial Research Organisation, and the …
"It's a computer model."
What is your point here? I keep seeing this sort of statement - on a geek site of all places!
Is it that it seems like a radical idea to model physical phenomena mathematically and to use computers to solve such models? Believe it or not people have been doing this sort of thing for decades now for all sorts of applications.
> Is it that it seems like a radical idea to model physical phenomena mathematically and to use computers to solve such models?
Perhaps you should read the 2005 paper by J N Carter from Imperial College London.
They show that there are calibrated models that have no predictive capability even when there are no modelling errors (that is: the models perfectly model the physical system.) They also show that there is no way to determine which models have predictive ability and which do not.
What is more the model they were testing only had three parameters and not the hundreds that climate models have.
> Believe it or not people have been doing this sort of thing for decades now for all sorts of applications.
Except they never rely on the computer models. A computer model might know all the physics of aerodynamics but it still can't model it properly. If it could then they would be able do away with wind tunnels and all those test flights they do to resolve the handling issues the models failed to predict. The computer model is simply one of many tools.
No. I'm saying that even a perfect model, with as few as 3 parameters, that exactly simulates the physics, might have no predictive power. It it also impossible to determine which models do and do not have predictive power. Add more parameters and it is less likely to have predictive power.
Climate models are so parameterised that they can be used to say whatever you want them to say. For example, in 2007 all but one of the models used by the IPCC claimed that the northern hemisphere would have less winter snow cover. At that time there was a small decrease in snow cover in the northern hemisphere. After several harsh winters reversed the decrease those same models, with different parameters, are now saying that the northern hemisphere will have increased snow cover.
Whether its flood, drought, cold, warmth, snow or lack of snow, in any area, the models can predict it by running the model with suitable parameters.
> Guess what? They already do this.
This is called hindcasting and it is easy to make the models match the past. To warm in 1978? Then increase to suphates released. To cold? Then decrease the sulphates. That is just one of many parameters they can manipulate to get the models to match the past.
Give me the last 20 years of UK horse racing results and I'll write a model that would have made you a profit over those 20 years. It will be pure luck if it makes a profit over the next 20 years, but don’t worry because if it doesn't I'll simply change some of the model parameters so that it would have.
Actually, it's not hindcasting, it is setting a model to run forward from a historic date and forecasting what is known to have happened to compare results. Hindcasting is to run a model backwards (something which you can't do with climate models) to get it to predict what did happen. This specifically goes to disprove your point - given the starting data you couldn't predict which horses won which races, however given the starting data for a model such as climate you can predict what was known to happen.
If the model you use is out when you've run a cohort of starting variables and observed the different outcomes, you change the parameters of the model to make it match what is known to have happened. This is called science, it's a critical part of science and not something to be suggested as cheating, which you are clearly doing.
Actually it is hindcasting. From wiki:
A hindcast is a way of testing a mathematical model. Known or closely estimated inputs for past events are entered into the model to see how well the output matches the known results. Hindcasting is also known as backtesting.
> This specifically goes to disprove your point - given the starting data you couldn't predict which horses won which races,
It doesn't disprove my point at all. When hindcasting the models wont say "it rained in London at 2:00pm on 25/12/2000", but instead give a measure of the rain over the winter season in the south of England. In a similar way to the climate model, my fictitious racing model will not say "Mr Bean in the 3:00 at plumpton on 12/11/2000" but would instead give a profit over a time period.
> If the model you use is out when you've run a cohort of starting variables and observed the different outcomes, you change the parameters of the model to make it match what is known to have happened. This is called science, it's a critical part of science and not something to be suggested as cheating, which you are clearly doing.
You literally have an infinite range of possible values for the starting variables that would give a valid hindcast. You can set these variables so that they produces any future scenario you wish. This is not called science, it is called modelling.
@AC 12:04
1) I understood hindcasting to be a different thing to you, it doesn't really matter what its called.
2) You don't seem to understand the difference between Weather and Climate and what a model will do or won't do.
3) I saw Brian Cox speaking last year - he was utterly clear: We _must_ model in order to understand what we are seeing, there is no point in experimentation or observation if we don't understand the process and that has to be by developing models. So, I'll go with what the Professor from CERN thinks about what is and isn't science rather than someone on the Internet.
> You don't seem to understand the difference between Weather and Climate and what a model will do or won't do.
My comment regarding climate models not predicting rain at 2:00pm in London, but rather predicting total rainfall over a season over a larger area perfectly demonstrates that fact the I do know the difference between climate and weather.
> I saw Brian Cox speaking last year - he was utterly clear: We _must_ model...
I agree that models need to used. But a model is just a tool to understanding, not the end result of understanding. The output from models are being treated as experimental results when they clearly are not. They are simply possibilities that they can not even assign a probability to.
It is interesting that you use a Professor from CERN as your authority figure. CERN (or the groups working at CERN) is one of the few institutions that uses computer models correctly. They use computer models to indicate where to look for the Higgs (and other things) in their experiments. They do not exclusively look where the models say they should because they know that no matter how well they understand their models they could still be wrong.
> So, I'll go with what the Professor from CERN thinks about what is and isn't science rather than someone on the Internet.
Sorry, I don't bow to authority. I'll keep on reading the scientific papers, the blogs, the news articles etc. I'll keep on working through what the data (when available), the models, the mathematics and the statistics in the papers say and from all that I will make up my own mind.
Re: Sorry I don't bow to authority.
Really? You should when they're speaking in the fields that they are authoritative in and about fields that you are not an authority in. I know it's all very fashionable to suggest that everyone is equal, but guess what? We're not. A Doctor in a subject is better than someone who isn't a doctor in that subject (except for vanishingly rare occurrences, such as "expert patients").
I go to the Doctor for an authoritative diagnosis of my health problem, I don't go to a layman.
If I had cancer, I'd go to an oncologist, rather than a gynecologist, because of the area of authority.
> I go to the Doctor for an authoritative diagnosis of my health problem, I don't go to a layman.
So do I.
When I was diagnosed with a serious illness (now fully cured) over 30 years ago I immediately went to the library and researched it. I found, through family and friends, other people who had had the disease and recovered and the family of some who hadn't. I spoke them about their experiences. This enabled me to discuss both the disease and treatment options with the doctor. What I did not do is bow to the Doctor's authority and just blindly accept his word.
Had the Doctor claimed that his computer model of the human body showed that, with my symptoms and several dozen assumptions he has made, I needed to have a kidney removed to prevent a urinary tract infection I would have found another Doctor.
"Perhaps you should read the 2005 paper by J N Carter from Imperial College London."
I have.
it makes some interesting points about curve fitting to historical data to derive a set of model parameters.
While this is a *reasonable* strategy AFAIK most (if not all) climate models actually *simulate* from first principles. How *well* they simulate, in terms of what physics is covered and frankly what is covered by some kind of fudge factor, is another matter.
I'd totally agree GCM's are unreliable and seem to still have a lot of fudge factors rather than proper physics derived elements.
But using Carter does not really cover the same *type* of model.
How do you feel about the computer models used at CERN to predict particle decay?
How do you feel about the computer models used by NASA to predict the path of asteroids (and space junk)?
How do you feel about the computer models used in Australia to predict how floodwaters will affect towns further downstream?
Do you dislike *all* computer models, or just those pertaining to climate? This enquiring mind demands to know!
I can't speak for the original poster, however while I personally like computer models I think climate computer models are a waste of time, since there are so many variables that are hard / impossible to predict, and we don't even know what all of them are / the effect they have, so it is basically just random guessing.
This is missing a key point! Those who don't live in Australia (and many Australians too) are probably unaware that the CSIRO of their childhood is nothing like the organisation of today that bears the CSIRO name.
CSIRO was once Australia's revered scientific organisation but today I'd question that. There's no doubt the CSIRO still has some excellent people but once its scientists were free to speak their minds openly, now the organisation's output is carefully crafted by media gurus long before its read by anyone outside the organisation. Often scientists regularly self-censor themselves long before the report stage to ensure that their careers won't be affected.
As with many government organisations in recent years, the CSIRO had a management takeover and everything from government policy to political correctness colour anything (reports, comments etc.) that leaves the organisation.
It's not that climate change isn't important, it is but it's what's not said or the emphasis that's put into certain aspects of reports that is of concern.
Massaging report from scientific organisations in any form is unacceptable. Even generating such a sensitive report in the first place (in the light of so much other information on the world stage), is, no doubt, politically vexatious.
Indeed, such stats make it look like research to suit an agenda. Pity that agenda isn't building state of the art nuclear power stations or taking a sensible look at how (if you believe AGW) selling shitloads of coal to one of the planet's worst polluters in order to stoke your GDP may come back to bite you on the arse. In short, that's two ways to investigate why you can't have your cake and eat it.
You have to ask who is paying off these scientists. I'm sure they'd all appreciate a supplement to their academic wages in order to fake this kind of AGW promoting nonsense. Do a little research and you'll find out just who has the money and the motive to bribe thousands of scientist...
It is the CSIRO. They are entirely dependent on Aus Gov funding, and the current Aus Gov is a bunch of socialist greenists with an insane Carbon Tax to sell.
The current head of the CSIRO is a green leaning *banker* named Simon McKeon.
Here is part of a puff piece from the left leaning paper "The Age" from when he was appointed.
"So, what does CSIRO's chairman think about one of the most polarising public policy issues we've ever confronted? This is a man who has just been appointed by a government whose leader called the issue our greatest moral challenge and then, in effect, just walked away from it when things got too politically tough.
He starts obliquely, saying he has first-hand knowledge of climate change from his activities as a world beating yachtsman. He has seen and felt the effects of climate change in the winds and water off Victoria.
''The weather systems we were getting simply moved south by about 500 nautical miles. Now, that bit is undeniable. The issue is cause, future impact, what do we do about it, is there anything we can do about it, etcetera etcetera. "
http://www.theage.com.au/opinion/society-and-culture/life-science-20100627-zbuk.html
The Age "left leaning", the government "socialist"?!?!
In what world is that remotely accurate?
Seriously, you need to understand what socialism means, it's certainly not supporting multinationals and other party donators over responsible policy choice. Which is the current model the Labor Party uses. But socialist, that is hilarious, they're a centre right party with the Liberal party moving to the extreme right.
And the Age is a seller of Real Estate advertising and football news, that's pretty much it.
Next you'll be saying Andrew Bolt is a reasonable person with valid well researched views...
It's like the cold weather here.
For years Australia has had droughts in no small part to pronounce La Nino effects and PDO. The climate chaos people have been predicting permanent drought for Australia in the same way that people here were predicting an end to snow in winter. Billions have been spent on desalination schemes.
It's raining and dams are overflowing. An inconvenient truth of a different kind.
So the story has to change. This is all about politics and nothing to do with climate or weather.
No, you are confuising El Nino and La Nina. They are opposites. It is expected to be colder and wetter now, but this is not an inconvinent truth. I This is the weather osscillations they are talking about. The CSIRO is a well respected govertment scientific organsisation, which is not privately funded. If they say its true, they will be correct. Read the 3rd paragraph of the article again.
Was trying to make the point that the predictions have to date been for more or less permanent water shortage. Hence massive spending on desal. Now there have been floods, the story has to change. If you look back at Australian weather, there's a history of alternating drought and flood. Seem to me that CSIRO are predicting no change from typical Australian weather.
I meant El Nino. Of which there have been a succession, combined with a phase of the PDO that's probably been the main cause of the lack of rain. Now, we're in a different phase of PDO, combined with La Nina and the weather is wetter...
"Global changes of this magnitude happen very rarely. They happen when asteroids strike, they happen when there's planetary volcanic activity. They're happening now because we're digging up fossil fuels and basically burning them all. And we're doing that very, very rapidly. And that transition system has a lot of unknowns in it, a lot of nasty surprises."
Why do the people feel need to write titles like that? Given the contents of the article, title "CSIRO: warming possibly by as few as 1 degree" would be about as accurate. Why not report the facts and say "CSIRO predicts warming by 3 degrees, with large uncertainty"? Even better, if you check the original report, the list of references shows that the temperature projections are based on papers from 2007 and 2008. Perhaps it would be better to base the title on something that is new in the report?
..the future looks really scary.
BUT what are these parameters..'unknown feedback factors of the positive kind' .
Ho hum. So if we assume that in addition to something we do know about - CO2 - there is also an unknown feedback factor making things worser than the worsest thing there ever was, then the curves match.
What is this unknown feedback factor? Everyone is looking for it, no one has found it.
What if this unknown feedback factor were in fact a completely different independent variable?
Well that would make concern over CO2 irrelevant, and all the politics based on it, and mean we had a different problem altogether.
One that really needs a complete investigation to it and possibly a total U-turn on policy WRT to climate change.
Why are we not looking for it?
You tell me..
http://www.clarewind.org.uk/events-1.php?event=39
.... who go confused by the various ACs replying to each other at the beginning of the thread? Who was for and who was against - I don't know!
Anyway, my pretty useless 2 pennith - in the 'against' column for computer modelling is the inaccuracies of CFD. No-one ever gets it right without some real world testing (just ask F1 teams - Virgin Racing had the only CFD-only designed car last year and they were rubbish).
The level of certainty that the boffins talk in this report would seem to at odds with the +/-100% accuracy of their results. Between 1C and 5C? I reckon the best they can say is that it will get warmer but we really have no idea by how much and we're not sure of what the effects will be.
I do take your point about F1 teams and the value of real world testing, however in the interest of taking everything into consideration I'd like to point out that 1. there is the possibility that it was the drivers who were rubbish not the car (not saying they were mind you) and 2. their computer-designed car was actually really really fast if you compare it to, say, the one I drive.
For those still reading and interested in some facts.
The March 2011 DECCW report titled NSW Ocean Water Levels, prepared by the state-owned Manly Hydraulics Laboratory, is a 347-page document that includes a specialist discussion of statistical measuring methods as well as tidal data and othere interesting stuff. The report's main findings are clear.
From 1914 until now, tides in Fort Denison have risen by an average of 0.9mm a year. But for the past 20years the rate has been just 0.4mm a year.
That is a couple of orders of magnitude less than the CSIRO pseudo bullshit generated from models (?) rather than someone actually measuring the tide levels! We have measured the tides in th western world since we became seafaring. The data is available. There are numerous academic papers which demonstrate very conclusively that the sea level rise this century is very constant and not alarming.
Dweeb
there was for removing my post - especially as the blatantly inflammatory and possibly defamatory post I responded to was not moderated.
Your bias is showing Mr Mod. Which is a pity as Richard has been doing an excellent job in showing you how to present un-biased, fact based editorial and journalism.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
