"Slo Mo" Disaster

Alan Boyle has an interesting story on flood prediction. Well it is to me, anyway.

Robert Criss, a professor of earth and planetary sciences at Washington University in St. Louis, agreed that the forecasts have been "remarkably accurate" - within the limits of the system, that is. He noted that the flood wave is working its way down the Mississippi River at about walking speed, giving the forecasters time to analyze the water's course, and giviing emergency officials time to react.

"It's like a traffic jam. The cars move slowly through the jam, and this big stuff is coming our way slowly and inexorably," Criss said.

The damage will be in the billions. And of course, some will say that this is a sign of climate change. But the real reason that the cost of these disasters is increasing is not because the weather is any different than it has been in the past but rather because people foolishly build in flood plains, because they don't understand the nature of statistics. There is no such thing as a "hundred year flood," at least in the sense that you can expect that there will be one per century, and after you've had one, you're safe for another hundred years. All it means is that statistically, one would expect one to occur that often, on average. Having one does not inoculate you from having another the next year (or even the next month), any more than chances that the next coin flip will be heads is increased by a previous tail. It's fifty-fifty every flip, and it's one in a hundred every year (assuming that the estimate is correct). This is the same kind of thinking as the guy who always carried a bomb on the plane with him, on the logic that the chances that there would be an airplane with two bombs on it were minuscule.

A perfect example is the 2004 hurricane season, which I drove over from California in early September to enjoy. I arrived in Florida just in time to put up shutters and batten down the hatches in our new house, when Frances hit us.

It was the first time a major storm had hit the area in many years, and most of the people who had lived here, even long-time residents, had gotten complacent. In fact, I recall sitting next to someone on a plane to LA earlier that summer, shortly after we'd bought the house, but before the storms. He was a real estate agent in Palm Beach County, and I mentioned that one of the things I didn't like about moving to south Florida was the hurricanes. He waved it aside, saying, "we don't get hurricanes here." I just shook my head.

Anyway, three weeks later, just as we were getting power back on and cleaned up from Frances, we got hit by Jeanne, which made landfall in almost exactly the same place (up around Fort Pierce). So this was not only a "hundred year" (or perhaps a "thirty year") hurricane, but we had two of them within a month. And of course, the cost of hurricanes will continue to grow, not because hurricanes are getting worse, but because, as in the midwest, and partly out of statistical ignorance, we continue to provide them with ever more, and ever more expensive targets.

[Update a couple hours later]

Jeff Masters thinks that climate change is causing 500-year floods to become more frequent. I don't think we have enough data to know that for sure (particularly since things have actually been cooling down in the last few years), but as he points out, another anthropogenic effect is the draining of wetlands for farming and building of levees to protect them. Levees work fine (until they suddenly don't) but they intensify effects down stream.