Climate is defined as weather averaged over a period of time -- generally around 30 years. This averaging smooths out the random and unpredictable behaviour of weather.

Ever since the great mathematicians of the mid-20th Century, we've been trying to create predictable models for unpredictable systems...economics, weather, bubble chamber patterns, casino games, the stock market.

These mathematical scientists failed.  The outcome of that was Chaos theory which said that as few as 3 independent variables can product highly "intelligent" and yet unpredictable behavior.

Climate falls into that category.   For chaotic systems, the past is absolutely NOT a predictor of the future, no matter how many years of data.   There are no "regular cycles" -- yes, you may see ups and downs in a few narrow periods, but over long long times you will see ... well, craziness!  (This is what the Andrill studies show).

A fine pronouncement, but what is your evidence to support it?  And on what timescale are you talking?

There are no "regular cycles"

I think the glacial-interglacial cycles are sufficient to refute yet another unsubstantiated proclamation.

yes, you may see ups and downs in a few narrow periods, but over long long times you will see ... well, craziness!

I am at a loss as to what possible relevance this might have, if true, given that many human civilizations rise and fall in these "narrow periods".

Climate may be chaotic on million year timescales, maybe the climate is supremely self correcting on that scale, always returning to its center.  So what?  On timescales that matter, decades to centuries, all the evidence points to a system that is responsive in a generally predictable way to many kinds of forcings, including enhancing the greenhouse effect.

"What if this weren't a hypothetical question?" -- unknown

I find it baffling that people are unable to understand the difference between a next day forecast and the statistical average over 30 years. Look at baseball. Any one game a good player can be great or stink but over the course of the season he'll probably end up at his statistical 0.300 batting average. So, predicting how he'll play tomorrow is difficult. Though you can try looking at his fitness levels etc. Predicting how he'll be over the complete season is easier because you can look at his last five seasons or so...

First, how do we know that climate is really just a simple process of averaging weather over 30 years? If that was the case, how do they account for huge shifts in climate like ice-ages? As it is highly unlikely that any thirty year period is going to point to most of North America covered with a sheet of ice, clearly there are factors in play beyond merely averaging conditions.

What this ultimately comes down to is the glaring, always unmentioned hole in climate science:  

**No one has ever established that future climate is predictable with any degree of accuracy beyond random chance.

To use the baseball player analogy (predicting what a good player is likely to do over the course of a season versus any given day) there is no mathematical formula to predict what a player will do in any given year even with absolute knowledge of past results and conditions at the start of the season. You can weigh those factors and make an educated guess that increases your CHANCE of making an accurate prediction, but ultimately you are still playing with CHANCE, not actually predicting future behavior like you would with some hypothetical General Theory of Baseball.

Climate "science" works the same way. They haven't demonstrated that the climate is predictable on any scale, and yet they insist it is.  

To follow the baseball analogy to it's logical end, it's quite likely (in my opinion) that future climate isn't predictable exactly as future stats aren't. The underlying reason is ultimately the same: there are far too many variables in play for the system to be accurately "modelled" (even in principle) and ultimately you are forced to work from gross assumptions and estimations.

--Ciani

If a system is truly chaotic, however, that doesn't mean that once the system has run you can't look backward and discover patterns that have emerged. Clearly climate works this way.

[quote]"So what?  On timescales that matter, decades to centuries, all the evidence points to a system that is responsive in a generally predictable way to many kinds of forcings, including enhancing the greenhouse effect."[/quote]

Huh? What evidence points to this? If all evidence points to this, one would think they'd have actually made an ACCURATE prediction.  

If anything the evidence points the OPPOSITE way! This is just another gross assumption that is totally unfounded--what evidence points to us being able to predict the climate 100 years from now?  

Also note the reliance on some arbitrary notion of "timescales that matter"? What does that even mean? Matter to who? Equations don't care about timescales, General Relativity runs forward and backward in time from milliseconds to billions of years with the same effectiveness.  

If a system is predictable 100 years from now, it should be just as easily predictable a thousand years from now, no? The only situation wherein it wouldn't be is if we couldn't trust the 100 predictions, and that's exactly where we are at now!

[quote]"So, predicting how he'll play tomorrow is difficult. Though you can try looking at his fitness levels etc. Predicting how he'll be over the complete season is easier because you can look at his last five seasons or so... "[/quote]

This is absurd. He is revealing the weakness of his own argument with this analogy. Any future baseball prediction is going to be nothing more than an educated guess. Yes, we can increase that accuracy of that guess by taking into account past performance, but--average or no--there is no reliable mathematical way to predict the future performance of a player, exactly as there is no reliable mathematical way to predict the future of the climate.

This guy is essentially saying that climate scientists are fortune tellers that rely on past data to guess the future. So what? That's not science, it's gambling.

--Ciani

The argument that 'something that isn't predictable on 1000 years timescales therefore isn't predictable on 100 year timescales' is rubbish.

Let's go back to baseball. suppose the pitcher throws the ball and we happen to know the velocity of the ball and the bat at the moment the batsman hits it. We can then, based on classical mechanics, predict with reasonable certainty where the ball will be a second later. We cannot predict where it will be 100 years later.

That makes the point, though it in't a perfect analogy. The question is how accurately you wish to know what will happen. The climate models predict a spread of answers for the temperature rise by 2050, the minimum of which, according to the latest IPCC, is a 1 degree. Those models could be run for a thousand years but the uncertainty would become so great that it wouldn't be worth doing.

However all we really want to know is whether the chance of something terrifying happen to our planet, due to our own actions, is large enough that we ought to change our actions. The answer to this is clearly yes!

Ignoring the assertive nature of this statement, let's assume it is true. Then looking at the graph on the 'Global warming stopped in 1998' page, showing the global temps from 1850 to now, waht exactly would the prediction of the state of the climate have been in 1975? Clearly, it was stable or if anything getting cooler and so any judgement on the climate would not have bee that things were getting warmer.

Why is this? If CO2 use over the last 100 years is responsible for global warming, then this would have been true in 1975 as it is now. How can a theory depend on when it is evaluated?

And if the answer is, you need to look at the whole century, then why? Because that supports your theory?