Flooding risks along the Mississippi River underestimated by Army Corps of Engineers

Published 19 April 2010

Scientists argue that the U.S. Army Corps of Engineers, in an effort to correct old data on water flows in the Mississippi, may have led to underestimates of the current risk of flooding on the Mississippi between the Ohio and Missouri Rivers, and to inadequate preparations by government agencies

An attempt by the Army Corps of Engineers to correct old data on water flows in the Mississippi may have led to underestimates of the current risk of flooding along the river, scientists argue in a new study. The study argues that a change in the way water flows were measured, dating from the 1930s, mistakenly led the corps to make downward adjustments in data from the 1800s and early 1900s.

That in turn is leading to underestimates of the risk of flooding today on the Mississippi between the Ohio and Missouri Rivers, and to inadequate preparations by government agencies, said Nicholas Pinter, a geology professor at Southern Illinois University in Carbondale and the author of the new report. “If you lower the numbers for the size of past floods, then present and future risk is reduced proportionally,” he told The New York Times’s Matthew L. Wald in a telephone interview.

Gary R. Dyhouse, a hydrologist who retired from the Corps of Engineers in 1999 but still consults for that agency on Mississippi flood questions, said the adjustments in the data made by the corps were correct. Flow tests using scale models determined that actual water flows in floods occurring in 1844 and 1903 could not possibly have been as high as were estimated using instruments of the time.

For example, the contemporary estimate for the flood of 1903 was 1.3 million cubic feet per second, but later estimates put it below 800,000 cubic feet per second, he said. The estimates for a flood in 1844 had been similarly lowered. “Obviously we’ll never know what the exact discharge was in 1844,” he said.

The study was published 26 March in the journal Hydrological Processes.

Wald writes that the problem with the data arose in the mid-1930s, when responsibility for daily measurements shifted from the Army Corps of Engineers to the United States Geological Survey. The corps took the measurements from boats in the river, but the geological survey took them from a bridge.

Five tests were conducted in the 1930s and 1940s, Dyhouse said. “These measurements showed the corps was overestimating flows, in their methods, compared to the U.S.G.S.,” he said.

Pinter, though, said that after looking at additional data, he found that the older method underestimated flows, rather than overestimating them.

Disputes over the accuracy of old measurements go back more than thirty years. Some other experts, however, gave credence to Pinter’s work. Brandon J. McElroy, a research scientist at the geological survey, said that Pinter had turned to data that had long been ignored and that his conclusions were reasonable. “Based on the references that he brings forward, those are all completely justified in my view as a scientist,” McElroy said.

McElroy was a peer reviewer of the study by Pinter, and said he was not speaking as a representative of the geological survey.

Donald C. Sweeney II, associate director of the Center for Transportation Studies at the University of Missouri, St. Louis, said it was important to get the numbers right, “because the more accurately you can portray the risks of flooding to people, the more rational and well-based decisions they can make.”

Historical data should not be altered, Sweeney said, “unless there’s a darn good reason for adjusting it that you can show beyond a shadow of a doubt.”

—Read more in Nicholas Pinter, “Historical Discharge Measurements on the Middle Mississippi River, USA: No Basis for Changing History,” Hydrological Processes 24, no. 8 (26 Mar 2010): 1088-93 (sub, req.)