Last night all the local TV stations interrupted regular programming to warn of an impending storm. The Weather Bug icon was flashing madly with the same warning. A severe thunderstorm was approaching Lubbock. Winds in excess of 70 mph were predicted. A wall of dust was said to be descending on the city. Hale was also predicted. It sounded like a mini Armageddon.

We had planned to go out that evening but with chaos on the way we stayed home. So guess what happened? Nothing. No wind, no hale, no lightning, no dust, no rain – nothing. It’s unusual to get the weather wrong when your time frame is only an hour, but the weather is a chaotic system (or complex if “chaos” disturbs you). The further out your forecast the more uncertain it gets. This uncertainty results from the quasi infinite input that determines the weather.

We can only guess what the weather will be like a year from today. We can be sure that it won’t snow in Texas; but we can’t tell if it will rain, if the wind will be calm, if there will be a tornado, if it will be cloudy, etc. We can give probabilities based on past observations, but that’s the best we can do.

Predicting the climate is an intrinsically uncertain business. I am not a climatologist, but I am a scientist. Accordingly, I know how hard it is to get everything right. There’s always some uncontrolled variable you forgot or that you (or sometimes anyone else) didn’t know existed. And I worked in a lab under very controlled conditions. Imagine the difficulties facing someone trying to predict the climate a century from now. Such an investigator must rely on computer models. These models in turn depend on the variables they contain and on the weight given each of them. Compounding the difficulty is predicting the effect on the environment of the climate change you expect. If your model misses one or more variables (virtually a certainty when you’re working with a century) your prediction will be way off. A variable off a little too much one way or the other will cause enormous errors. Ignoring one altogether will make the data useless, or worse harmful if they’re believed.

About the only thing one can be sure of is that all the models will be wrong. The are an infinite number of possible outcomes, but only one will happen. It will seem inevitable in retrospect. The complexity of the problem of climate change explains the vociferous debate that its study causes. There are more global warming doomsayers than there are skeptics. But science is not settled by a vote. Evidence is the arbiter. If catastrophe is a possibility prudence dictates that the putative disaster receive careful consideration, but based on the evidence currently available and its inherent uncertainty I wouldn’t bet a sizable chunk of the world’s economy on predictions of catastrophic change.

The weather bureau is again predicting meteorological Armageddon this evening. If I survive, I’ll let you know how things turn out when the power comes back on.