How can we predict climate change when we cannot even predict the weather?

Answer 1

Climate is not weather, though this is a misconception.

Weather deals with the stochastic fluctuations occurring throughout the atmosphere generally on a day-to-day basis. These are predicted using weather models governed by equations that can effectively describe the atmosphere at these scales. Meteorologists are quite good at predicting the weather a few days in advance, but not so much beyond a week, as you know.

Climatologists are not interested in making a prediction for one particular day 10 years in the future; you will never see that. Climatologists are only interested in average weather over much longer time scales, and the key difference here is that at these scales, the atmosphere is governed by very different processes. Climatologists are not interested in small-scale eddies in the atmosphere, diurnal fluctuations, local-scale convection, etc., and meteorologists are. These are not processes that can be modeled and predicted very far in advance anyway.

Of primary importance to us is the Earth's global radiation budget, which means looking at variables such as how much energy is coming into the system from the sun, how much is escaping back to space, and how that radiation behaves when it passes through the atmosphere at its different wavelengths. How much is reflected, transmitted, scattered, and absorbed? The precision to which we must know this is not the same as that of the atmospheric scientist, where small differences will balloon into huge errors in only a few days time.

The oceans are also very important to climate, whereas they are not nearly as important in weather forecasting (depending on where you are of course). And oceans have "memories" that last thousands of years. Same with ice sheets, which change very slowly. But at our timescales, these changes are very important and are also capable of being incorporated into our models. Meteorologists do not care about melting ice because it occurs far too slowly to make a difference.

Another example is with greenhouse gases and the cycles they participate in between the land surface and the atmosphere. With an understanding of exactly how each greenhouse gas behaves in the atmosphere as far as how much extra radiation it absorbs, we can tell predict how much different levels in the atmosphere in the future will affect future temperatures. The trickier part is predicting this as a part of the cycles in which they participate and the feedbacks that are involved. Again, meteorologists have no interest in this because the timescale of change is too long.

So given our understanding of things like the radiation balance, the ocean circulation, the ice sheets, and biogeochemical cycles and greenhouse gases, we can input these variables into our climate models and predict generally how the Earth will respond well into the future. It's true that we don't have a perfect understanding of how all of these processes will come together, and we wish we had more observations and more computing power to model this on a smaller grid, but the models are pretty good given all these circumstances.