Catch-22, is a logical paradox arising from a situation in which an individual needs something that can only be acquired with an action that will lead him to that very situation he is already in; therefore, the acquisition of this thing becomes logically impossible.Wikipedia
To prompt worldwide concerted action to severely reduce CO2 emissions we need to have such clear evidence of a real and present danger that we are already committed to dangerous climate change. At the stage we have observations of massive climate disruption that is clearly outside of previous natural variability it will be increasingly hard for people to continue to turn a blind-eye to the problem, putting action off until evidence of a disaster is at hand will no longer be an option.
Thus it is a case of Catch-22: To prove the danger of climate change we need evidence, not only that climate change is underway and is likely to be disastrous, but conclusive evidence that climate change is a disaster. However to gain that evidence we would need to leave climate change to proceed to a point where we are experiencing the disaster. The problem with leaving climate change to proceed until we have clear and irrefutable evidence of a catastrophe is that by that stage it will be too late to avert further global warming and an intensification of the problem.
Realclimate have recently posted an important piece by Professor Ray Pierrehumbert on the Athabasca Oil Sands being opened up due to the starting of a pipeline: Keystone XL. The article argues that the opening up of the Athabasca Oil Sands makes avoiding 2degC warming very unlikely. As James Hansen put it with regards avoiding dangerous climate change "If the tar sands are thrown into the mix it is essentially game over [to avert 2degC warming]."
In the comments Prof Pierrehumbert referred to a 2010 GRL paper by Kyle Armour and Gerard Roe this being "Climate commitment in an uncertain world." PDF here. There's also a related set of slides from Gerard Roe on the matter of committed forcing here.
From Gerard Roe's slides there's a useful graphic that shows how relatively well-constrained climate sensitivity actually is. Climate sensitivity is a measure of how much warming you'd get for a change in climate forcing, such as a doubling of CO2.
This shows how different strands of evidence reveal that climate sensitivity is most likely in the range 2 to 3 degrees C. However there is the long upward tail which suggests high climate sensitivity cannot be ruled out. This tail is mainly because of uncertainty about aerosol forcing. Because aerosol forcing is negative any uncertainty in this can only be reflected in how much higher climate sensitivity may be, not how much lower.
Aerosols are formed by natural and human driven processes, essentially they're droplets of water, however their source determines their nature and this can effect how they act on the climate. Aerosol forcing due to humans is definitely negative because the main action of aerosols on climate is to reflect sunlight back into space, hence cooling the Earth. The following graphic from Gerard Roe's slides shows the climate forcings of various key factors.
Note how Ocean Heat Storage, CO2 and non-CO2 greenhouse gasses are tall thin bell curves, this is because their values are fairly well defined. However the bell-curve for the direct (reflection of sunlight) and indirect (impact on clouds) effects of aerosols is spread-out because it's not clear what the real value is.
Take the factors in the above graph and add them and you get the overall current climate forcing, the black bell-curve.
Now what happens if you remove the forcing due to aerosols is that you move the black bell-curve to a more positive position. This reflects the warming that you'd get if those aerosols were to cease.
If human burning of fossil fuels were to cease or massively reduce there would not be a cooling, there would be a warming. This is because the CO2 itself is only slowly removed from the atmosphere due to the chemical processes involved. The initial fall would take centuries and would leave a long tail of just under half the peak atmospheric concentration, that long tail would take around 100,000 years to dissipate. However the aerosols would leave the atmosphere within weeks to a couple of years. So upon cessation of fossil fuel burning the full warming of CO2, which is currently hidden by the unknown cooling of aerosols, would be manifest.
Humanity has a tendency, both on the individual and collective levels, to leave problems until they become serious. Only then are options for action considered and implemented. In the case of climate change this is a foolhardy and dangerous policy. As more of the world becomes industrialised and usage of fossil fuels increases the committed warming the policy of 'wait and see' becomes more and more dangerous.
By the time the majority decide action is needed on climate change it may be too late for action to avert the edge of the precipice.