IE 11 is not supported. For an optimal experience visit our site on another browser.

Flood forecasts in flux

Frank Polich / Reuters
Surveyors Dick Leach and Kevin Flood measure the height of the

Mississippi River in relation to the height of the levee in Canton, Mo.

How high will the flooding go? That's been a crucial question for Midwesterners this month, and the answer requires some complex - and changeable - figuring.

Forecasting the rise of the rivers is a cross between predicting the weather and predicting a traffic jam, experts say. The good news is that this summer's flooding is something of a slow-motion phenomenon, providing time for communities to shore up their defenses. The bad news is that the lessons from the last monster flooding in the Midwest, back in 1993, have gone largely unlearned.

First, about that good news: Flood forecasts, like weather forecasts, are the province of the National Weather Service - and technological advances have generally improved the quality of the flood-warning system just as they have for other aspects of severe-weather warnings.

"The folks in the field have barely had time to catch their breath, but for the most part we haven't had negative feedback - except that everyone always wishes we had it exactly right the first time," said Noreen Schwein, manager of the hydrologic services program at the weather service's Central Region Headquarters in Kansas City, Mo. The Central Region takes in the entire area of flooding in Iowa, Wisconsin, Missouri and Illinois.

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 from his office in St. Louis.

How the system works

The weather service relies on a network of 3,790 flow gauges set up across the country by another federal agency, the U.S. Geological Survey. Those streamside gauges collect real-time data about water height and send it out in real time via satellite links or dedicated phone lines.

The weather service feeds all those readings into computer models that are based on past floods, and factors in the predicted precipation for the next 24 hours to come up with their outlook. That 24-hour time horizon is one of the system's big limitations.

"We're doing better in our weather forecasting, but there's still uncertainty there," Schwein said. That's why the weather service might say a river will crest at 12 feet above flood stage on one day, and then revise that prediction to 13 feet a day later.

The flood stage for a given location is another factor that has to be fed into the model. Defining flood stage is basically an engineering question: How high does the river have to go to threaten life or property?

"Changes in land use will cause a change in the flood stage," Schwein said.

Thus, the flood stage can vary from place to place along the river, depending on how high or low the property at risk sits on the potential flood plain. It's theoretically possible for the same river level to be far below flood stage for one location, and high above in another location.

"The equations that make up the models for river forecasting take into account the flow of the river," Schwein explained. "They attempt to take into account some of the basin features, like slope of the river. How narrow or wide the river is will change the flow. ... If you build a shopping center where there used to be a field, you'd have more runoff, and that can increase your flow and change the impact in that area."

This is where we start getting into the bad news.

Lessons unlearned

Criss and other scientists have been warning for years that increased development along the Mississippi River - ranging from mega-malls to agricultural levees to constructions aimed at improving the river's navigability - is leading to increasingly damaging floods. Some of the warnings go back 33 years, and others are as fresh as a couple of months ago.

Criss himself wrote out another warning this week for the Saint Louis Beacon, and an editorial in the St. Louis Post-Dispatch runs along the same lines today: that past warnings have gone unheeded.

"It should have been over in 1993, but what did we do? More floodplain development," Criss said.

That's not directly the fault of the weather service, Criss emphasized. But he said the weather service and other federal agencies are making things worse by characterizing this year's floods as being the type of event that won't happen again for hundreds of years.

Criss pointed to the river readings for Clarksville, Mo., where the water level is projected to go to 37.7 feet this weekend - just above the level of the 1993 flood crest and just below the defining point for a 500-year flood.

"How possible is it that twice in just the last 15 years we've had two 500-year floods? ... People have been misled into feeling confident that they can live in flood plains," Criss said.

That false sense of confidence can feed into a vicious circle of development.

"The developers love this certification, because they can say this won't happen for 500 years," Criss explained. "People are being misled, and it's a matter of privatizing the gain and socializing the loss. ... In this case, people have made a windfall by taking farmland and certifying that it's safe. Then it's not worth $1,000 an acre, it's worth a million dollars an acre. The U.S. taxpayers have to pay for these levees, and then when things fail, it's the taxpayer who has to pay again."

The addition of levees and other river engineering projects have served to narrow the Mississippi into a more confined space, Criss noted.

"If the river can't spread out, the river has got to go up. This isn't related to climate change. You force flood levels to be higher. The water is 10 feet deeper than it would be a century ago," he said.

Bad-news, good-news situation

Would removing some of the low-priority structures to give the river a wider berth solve the problem? That's a mixed bag, considering that billions of dollars worth of crops and other property are at stake.

Today's levee break at Meyer, Ill., serves as an illustration: It's a heartache for disaster officials and residents in that rural area, but it clearly takes some of the pressure off downstream. This chart shows that river levels at Quincy, just a few miles downstream, declined 2 feet just in the course of seven hours today.

Since the 1993 Midwest floods, more than 50,000 acres of Missouri River floodplains have been restored to their natural state as part of the Big Muddy National Fish and Wildlife Refuge - a project that has been held up as a model for re-creating natural escape valves for floods.

Criss doesn't recommend a wholesale reversal of existing development. He'd just like to see the warnings heeded for once. "It's always hard to go backwards," he said, "but what is incredibly bad is that all this floodplain development was done long after many, many scientists have pointed out this problem."

The Great Flood of 2008 will give scientists additional data to chew on. Even now, the weather service is tweaking its computer models to reflect the lessons learned from the flood's unprecedented heights.

"With this flood, we actually had a USGS person resident and working in our North Central forecast center up in Minneapolis," Schwein said. "It's somewhat of a theoretical process. We've never hit this level of flow before, so we don't exactly know what stage equates to a flow that we've never seen."

To monitor the floods in real time, check out the interactive map offered by the National Weather Service's Advanced Hydrologic Prediction Service. The U.S. Geological Survey has its own station-by-station map of river gauges. I clicked on the map to trace the highs and lows of the Maquoketa River, near the part of Iowa where I grew up - and it looks as if the worst is over.

You'll also want to check in with's special report on the Midwest floods, featuring an interactive map that encapsulates the news and historical data from the region's rivers.