Get used to it. New research suggests Mississippi floods like this will become more common if the Midwest continues to warm.

Old Man River, Hot and Bothered

RENO, NEVADA--People living along the upper Mississippi River have experienced more than their fair share of natural disasters recently, with three huge floods in 1993, 1997, and 2001. Residents have been lamenting their bad luck and hoping for a drier future. But they might want to get a head start on filling more sandbags instead, according to new research presented here 24 July at the International Union for Quaternary Research meeting.

Predicting floods is a tricky business, particularly on a river system as large and complex as the Mississippi. So geologist James Knox of the University of Wisconsin, Madison, went looking for clues by inspecting how it has behaved in the past. First, Knox dug trenches along the river in southern Wisconsin and found traces of the three recent floods: The powerful surges of water carried and deposited more sand and larger grains than usual, and even ripped up chunks of the muddy river bottom, forming distinctive layers.

Knox then looked for similar layers from the past 7000 years. Using radiocarbon dating of charcoal and wood fragments in the sediments to determine when each flood occurred, he discovered that the large deluges tend to occur in clusters. When Knox examined climate records for the area, he found that the clusters all occur at times of rapid climate change, particularly when climates are becoming warmer and drier--which is what most models of global warming predict for the midwestern United States. Even though there may be less total rainfall, warmer winters and springs can lead to rainstorms falling on snow pack, which is a recipe for big floods, says Knox. "We may well be in line for more of these," he says.

Hydrologists have commonly assumed that floods would depend on climate, but the new study is one of the few pieces of direct evidence that backs this up, says paleohydrologist Victor Baker of the University of Arizona in Tucson. Most research money for flood prediction is spent on computer models, he says. "The only way to test them is with information from the past. You have to know the habits of the river."

Related sites
James Knox's research
Department of Hydrology and Water Resources at the University of Arizona
Information about the 1993 flood (University of Akron)
Information about the 2001 flood (USGS)