'Sound science' must be more than a sound bite

The president and many in Congress have recently called for the use of "sound science" to guide federal agencies in their decision-making. This is certainly a laudable goal if what they mean is the use of unbiased science to guide important matters of public policy.

There could not be a better place to require the use of unbiased science than in the government's study of flood risks along the Midwest's large rivers. Such calculations dictate where development is allowed. Mistakenly permitting additional building in the Midwest's vast flood plains would have huge implications for natural-resource management and future disaster payments from federal coffers.

The 1993 flood caused tremendous scrutiny of government policies relating to flood plains. At $15 billion in losses, the flood was one of the top five most expensive natural disasters in U.S. history. Its cost could not be overlooked even by those living on the coasts, as nearly one-third of the losses were picked up by federal taxpayers in the form of direct disaster payments or reconstruction of highways and levees.

As part of the re-examination of flood-plain policies, the U.S. Army Corps of Engineers was asked to recalculate flood probabilities along the entirety of the upper Mississippi, lower Missouri and Illinois rivers -- some 1,900 miles of river. The corps' estimates had not been updated since the early 1960s on the Missouri and the late 1970s on the Mississippi, and it was widely thought that repeated floods in recent years would cause the corps to upwardly adjust its estimated flood profiles.

Seven years and $9 million, the corps rendered its conclusion: Flooding is getting less severe in many areas, not worse. From just above St. Louis to points farther south, the corps concluded, the so-called 100-year flood had decreased in height by as much as 1.7 feet compared to its earlier estimates.

Scientists, environmentalists and many living along these rivers were amazed. How could it be that flooding has gotten less severe when Midwesterners had seemingly thrown sandbags every year for the past decade?

In fact, flooding has not gotten better, but is actually far worse today than it was a century ago. As a matter of indisputable fact, seven of the 20 highest floods ever recorded at St. Louis have occurred since 1979 (the last time the estimates were revised), including the all-time record set by the great 1993 flood.

A little further downstream, the trend is even more ominous. At Thebes, Ill., the 10 highest floods in the period of continuous record have occurred in just the past 32 years, since 1973.

A growing number of scientists, from academia and even from other agencies of the federal government, have concluded in peer-reviewed articles that large floods on the Mississippi and Missouri rivers are up to 10 feet higher today than in the early 1900s. The cause: levees and navigation structures called wing-dikes, as well as flood-plain development. Such works constrict floodwaters to a narrow channel, forcing the water to rise instead of spreading out onto the flood plain. They also destroy the natural function of the rivers, degrading aquatic and riparian habitat, or converting irreplaceable agricultural land to impervious pavement.

Something clearly went awry with the corps' complex modeling that formed the foundation of its lengthy study. Perhaps it was too much to ask in the first place that this agency -- which has built, funded or permitted most of the country's levee and channelization projects -- study trends in flooding. The corps certainly would look bad if its billions of dollars of flood-control works had actually caused flooding to get worse.

The bottom line is that the implications of this study are too large to simply ignore as another flawed corps work product. Adjusting estimated flood profiles downward by a foot or more could open up huge new swaths of flood plains to development. The Federal Emergency Management Agency uses such estimates to redraw flood-insurance rate maps, which largely dictate what portions of the flood plain can be built on. If flawed numbers are used to create these maps, hundreds if not billions of dollars could be subject to underestimated risk of flooding. Ultimately, that would cost the American taxpayer big time.

Luckily, we haven't dived off the deep end just yet. FEMA has not redrawn any flood maps for flood plains along the major Midwest rivers, and there is still time for an unbiased panel of experts to review the corps' work. Now would be a good time for the Bush administration to put some meaning in the term "sound science."

Robert E. Criss is a professor at Washington University in St. Louis who has studied and written extensively on water resource issues, including the causes of flooding. Edward J. Heisel is the executive director of the Missouri Coalition for the Environment in St. Louis.