The first suggestion that CO2 might drive climate change was made by Tyndall in 1859.

The first crude but usefully predictive model was created by Arrhenius in 1896.

That's 1896.

How much more evidence does it take?

That's true. CO2 photo-thermal effects have been known for a long time.

I have no dog in this hunt - but I believe that short- and medium- term projections that start diverging from actuals almost on the first day are a problem for getting broad support for the hypothesis.

And it's a very interesting question, don't you think? Why aren't temperatures going up more? Is the earth modulating it somehow? The biosphere appears to attempt to maintain some kind of thermal homeostasis, and appears to be better at cooling the hot than warming the cold. How does that work?

Homeostasis could be achieved by a balance of waver vapor and cloud cover which are both strong factors on temperature but acting in the opposite directions.

"Thus the possible positive and negative feedbacks associated with increased water vapor and cloud formation can cancel one another out and complicate matters. The actual balance between them is an active area of climate science research."

http://www.acs.org/content/acs/en/climatescience/climatescie...

At the macro-level, sure. That's one effect. At the micro (?) level, there's an interesting effect where stands of trees (and rain forests) will emit cooling aerosols directly into their canopy. Pretty cool!

As tosseraccount wrote, "Temperature changes that are accurately predicted by models..."

Why would an accurate model have anything to do with causation? If humans are causing climate change it will happen if we have a good model or not and vice versa.

Does it matter? I'd like to see some solutions to the "problem" that amount to more than putting lipstick on a pig, and project to be more effective at fixing the issue than that method of porcine beautification.

People have been forecasting various millenarian crises for millenia. It seems to be what we're good at... Nowadays, we're beyond thinking that the moon demon is going to eat the sun god, or that the four horsemen of the apocalypse will sweep down and bring disease, war, famine and death, Malthusian over-population will have us eating our children, the Y2K bug will result in chaos and rioting, Peak Oil will have us wearing ass-less chaps, hockey masks, and killing each other Mad Max style, nuclear Armageddon will give rise to a race of atomic spiders, and now anthropocentric global warming will... something something, rising sea levels. All of these scares came and went.

You do realise that this is not an argument against anything being true. Either humans are responsible for global change or they are not. What we do about it is a totally different question.

How would you establish causation?

How do you establish causation with any experiment?

Generically? Elementary school scientific method: By formulating a hypothesis, then testing the hypothesis.

Hypothesis: The recent rise in global average temperature is caused by man-made increase in emissions of CO2.

Now, how do you test?

Well, we could just stop making CO2, wait 50 years and see if the temperature goes down. But then that's only one data point, and you haven't controlled other potential factors. One big one: The sun doesn't always put out the same amount of energy.

So now you need to make sure the sun's heat output remains constant in order to test if a 50 year elimination of CO2 output stops the global temperature rise.

But wait...what if the sun's output has been increasing over the last 50 years? That could have caused the temperature rise you formed your hypothesis about. If you stop the increase of the sun's output, and you stop the CO2 output, now you have two variables that you've changed, and you can't say, with any sort of certainty, which one caused the temperature to quit rising.

Neither of these options are really practical, so, instead, we try to create a mathematical model of the entire system, feed it the data that we do have, and try to model the expected results. Models fail routinely. Lets take hurricane models for example:

They're right, sometimes, their kinda right, sometimes, and their wrong, a lot. Why do you think that is? Their inaccurate and don't have sufficient data points. A perfect model would take every single variable in an entire system, know with certainty their effect on the system as a whole, and it would never be wrong. Some systems we have that luxury. Earth's atmosphere is not one of those.

We do not know even know what every single variable is. We don't know with certainty what effect the variable has on the system, so we can only model it with a best-guess approach.

If someone in 1980 looked at this[1] chart, they would conclude with a great deal of certainty that solar activity is the cause of temperature fluctuations over the last 100 years. After 1980, that no longer appears to be true. But global carbon emission have been sharply increasing since 1880[2].

So what logic concludes that CO2 can go up for 100 years, without causing an increase in global temperature, but as soon as the global temperature stops tracking solar output, assign the cause to CO2?

I'm not saying CO2 isn't the cause. I'm certainly not qualified to make that determination, and a qualified person probably wouldn't make that determination because they wouldn't have enough information to do so. But it certainly wouldn't hurt to eliminate our reliance on burning hydrocarbons for energy. This has more benefits than just reducing CO2 output.

We need better data and a better model, since we can't actually experiment.

[1] http://icons.wxug.com/metgraphics/climate/facts/Solar_vs_Tem...

[2] http://cdiac.ornl.gov/trends/emis/graphics/global.total.gif