|
Post by tacoman25 on Apr 17, 2009 18:37:12 GMT
And therin lies the difference between true science and dogma. But dogma will bring in more money. I suspect dogma leads people to pick the graph they want to make their point. The Pplot above is ludicrous as it picks a single point, 2002, as an end point. Well yesterday was warmer than today, so I expect we will be hitting absolute zero in Devon by September! Here's another plot from rankexploits which is a bit fairer. But I should add that the average of a multimodel ensemble will hide the variability in the climate. I'll just put the link as the image is huge: rankexploits.com/musings/wp-content/uploads/2009/04/20yeartrends.jpgAt face value, then, the models are trending 0.005-0.01C above the observations for the past few years, and have diverged a bit in the past 4 years. This person has attempted to put error bars on it which shows that the models and observations are not out of line till about 2006. Do you believe then that 3 years of diversion are meaningless?
|
|
|
Post by gridley on Apr 17, 2009 21:28:15 GMT
Do you believe then that 3 years of diversion are meaningless? Actually, in that chart I see 5 years of divergence (1999-2001 & 2006-2009) and 5 years of correlation (2001-2006). Hey, 5 out of 10 isn't that bad! In school it would only be... oh wait, that's a failing grade.
|
|
|
Post by nautonnier on Apr 18, 2009 2:09:11 GMT
And if your life or livelihood depended on the forecast (sorry 'modeled trend' ) correctly matching reality - would you still claim that the graphs 'match'?
Well - many people's livelihoods DO depend on decisions made based on claims that those graphs are matching. And 'scientists' repeatedly saying that the forecasts from AR4 are totally valid and in fact greatly understate the problem - are setting things up to have politicians destroy many people's livelihoods. We have the President of the U.S.A. standing up and saying that the Arctic and the Antarctic are catastrophically melting - when they most definitely are not - and based on that and the 'scientific advice' wanting to 'tax power companies into bankruptcy'.
Look at those graphs again - do they match? If you accept that there is divergence - WHY is there divergence - what reduced the heat input or increased its output in 2003 onward?
But of course you are in agreement with furrycatherder that those graphs match. There is no divergence and the temperatures are rising precisely as 'forecast' by AR4 - correct? Not a hint from either of you of a Bayesian lack of confidence?
|
|
|
Post by jimg on Apr 18, 2009 5:03:38 GMT
The diversion wasn't my issue.
The models show an accelerating rate of change, not the decelerating rate of change that the measurements show.
|
|
|
Post by steve on Apr 19, 2009 17:16:17 GMT
The diversion wasn't my issue. The models show an accelerating rate of change, not the decelerating rate of change that the measurements show. The divergence is obviously related to the "weather". If you individually showed each of the 16 models in the ensemble, each would show much variability with divergences and convergences. The lower trend is more interesting. Perhaps we'll get 3C warming this century rather than 4.5C.
|
|
|
Post by ron on Apr 19, 2009 17:28:49 GMT
Hmmm, I have a question: If the models are only trends and not forecasts, why to they show variability rather than straight line?
|
|
|
Post by annav on Apr 19, 2009 17:50:47 GMT
To have a divergence between model trends and data trends one has to believe in the error bars. I do not believe in the error bars for the models. They are not error bars. They are an attempt to simulate the chaotic behavior of climate by using many fits to known temperatures with different initial conditions, not a variance of the many parameters' errors that enter into their calculations., nor the systematic errors. Only the albedo has a systematic error of www.leif.org/research/albedo.png the order of 8%( if the albedo is 0.31 and 8% change is .024). The data shown are recently published and certainly are not in the IPCC models. I think if the true experimental errors on the parameters used in the models were used the error bars of the models would be off scale, making any comparison meaningless.
|
|
|
Post by jimg on Apr 20, 2009 5:16:15 GMT
The diversion wasn't my issue. The models show an accelerating rate of change, not the decelerating rate of change that the measurements show. The divergence is obviously related to the "weather". If you individually showed each of the 16 models in the ensemble, each would show much variability with divergences and convergences. The lower trend is more interesting. Perhaps we'll get 3C warming this century rather than 4.5C. So what you're saying then is that the models can't replicate the past behavior, can't replicate current behaviour, but because it has the "trend" correct, we should trust that the trend will continue indefinitely. Even a broken clock is correct twice a day.
|
|
|
Post by tacoman25 on Apr 20, 2009 5:35:38 GMT
The diversion wasn't my issue. The models show an accelerating rate of change, not the decelerating rate of change that the measurements show. The divergence is obviously related to the "weather". If you individually showed each of the 16 models in the ensemble, each would show much variability with divergences and convergences. The lower trend is more interesting. Perhaps we'll get 3C warming this century rather than 4.5C. Well, if you were to simply extrapolate based on the rate of warming increase from 1900-2000, this century will have similar warming to the past century (about .6C). The warming from 1980-2000 was very comparable to the warming from 1920-1940, and the warming from 1900-1950 was very similar to the warming from 1950-2000. So far, there is no sign of accelerating warming...and that is a problem for popular AGW theory.
|
|
|
Post by steve on Apr 20, 2009 10:08:54 GMT
The divergence is obviously related to the "weather". If you individually showed each of the 16 models in the ensemble, each would show much variability with divergences and convergences. The lower trend is more interesting. Perhaps we'll get 3C warming this century rather than 4.5C. So what you're saying then is that the models can't replicate the past behavior, can't replicate current behaviour, but because it has the "trend" correct, we should trust that the trend will continue indefinitely. Even a broken clock is correct twice a day. You are failing to understand what the models have been asked to do, are capable of doing, or what they are expected to do. None of the models in the ensemble have been started off with conditions that perfectly match the real earth. None of the models perfectly replicate the earth system. Even if you commissioned the Magratheans to build 16 almost identical earths with closely matching initial conditions, their temperatures would vary around their own trends in a similar way, and by definition, 5% of the time their temperatures would be outside the 95% confidence interval for the expected trend. But all 17 "earths" show a warming trend.
|
|
|
Post by ron on Apr 20, 2009 14:08:04 GMT
But the 95% confidence level would be a heckuva lot tighter, and would be real as opposed to the 95% over-confident guess level in these man-made computer simulations. These bear as much resemblance to the real world as early Mickey Mouse cartoons - crude at best.
|
|
|
Post by trbixler on Apr 20, 2009 14:36:26 GMT
Steve Are you suggesting that you understand? 'You are failing to understand what the models have been asked to do, are capable of doing, or what they are expected to do. ' 43 days and counting is not in the models. But please note that solar variability has existed long before there was life on our planet.
|
|
|
Post by steve on Apr 20, 2009 18:08:49 GMT
Steve Are you suggesting that you understand? 'You are failing to understand what the models have been asked to do, are capable of doing, or what they are expected to do. ' This post calls for a sarcastic response. No....you don't say! Then the models must be wrong. I lose!
|
|
|
Post by icefisher on Apr 21, 2009 4:08:31 GMT
But all 17 "earths" show a warming trend. Most likely that problem is attributable email. Obviously if you build all 17 models using the same assumptions they will come up with the same results. Don't confuse science with popularity. I can assure you that in the financial sector there were thousands of models all in essential agreement regarding derivatives varying only by a small percentage, but even thousands of model earths won't provide assurance. . . .ask Alan Greenspan about that.
|
|
|
Post by jimg on Apr 21, 2009 5:47:21 GMT
Steve, I think the problem lies in this:
From a scientific point of view, the models show correlation to current (or past) trends.
When the models deviate from current observations, it means that there is a physical phenomena that is not being represented adequetely in the model.
As we learn the nuances of the physics involved and those that we don't, they should (will) be added to the model to create a better simulation.
Models are intended to increase our understanding.
---------------------------------------------
What has happened:
Is the model has been substituded for the real. It has become its own being, a simulacra of the real.
We rely more on the model to tell us what is real, even though the model does not reproduce current phenomena, it kindof sortof matches the trend at times. (Except when it doesn't.)
The argument is this: Do the models have a sufficient understand of the underlying physical processes to accurately forecast future weather?
Then it is a matter of confidence. I personally do not have the confidence to enact draconian measures, that will coincidentally enrich a number of individuals who manufacture nothing, process nothing, and distribute nothing. The only thing that these individuals are doing is inserting themselves in between the producers and the consumers to take a share of the pie, by force.
|
|