After floods: restoring ecosystems - Special Section: America Under Water - Cover Story

USA Today (Society for the Advancement of Education), July, 1994 by Richard E. Sparks, Ruth Sparks

Decaying riverfronts have been turned into public parks and tourist attractions, and farmland has been revitalized by nutrient deposits.

The flood of 1993 was the result of an unwitting alignment between the forces of nature and the civil works of humans. While it was an economic disaster, it was a boon to many kinds of native plants and animals that not only are adapted to flooding, but depend upon it. Restoration of the natural services of floodplains - including production of fish and wildlife and conveyance and storage of floods - should be evaluated as an alternative to rebuilding agricultural levees, especially at a time when farmers are paid not to grow crops so as to avoid overproduction and costly surpluses.

The flood of 1993 really began in the fall of 1992, when heavy rains left soils saturated throughout the winter, so they could not absorb more moisture when the normal spring rains began in June. Once the rains began, they did not quit. Instead, unusual weather patterns kept pumping warm, moist Gulf air over the Midwest, where the moisture was wrung out by cooler than normal air masses from Canada. Not until the first week in August, 1993, did this pattern begin to break up. By then, portions of the lower Missouri and lower Illinois rivers, as well as much of the Upper Mississippi River, had experienced three flood crests, including the highest one in late July, which reached 49.58' on the St. Louis gauge, uncomfortably close to the 52' design capacity of the floodwall. One more smaller crest arrived in late September before the water finally dropped to the low flow level. One of the cruelties of this flood was that some people began repairing their homes following one crest, only to be hit by another. Not everyone fared as well as St. Louis. Neighboring St. Charles, Mo., for example, almost completely was inundated. Of the 1,576 levees in the Upper Mississippi River Basin, 1,082 were breached or overtopped.

Was the flood an act of God? Certainly in terms of rainfall, this was a very unusual event. Eight of 10 selected National Weather Service stations throughout the upper Midwest received more than 200% of the average July rainfall for the 30-year period 1961-90; three received more than 400%. Should this amount of rain have caused an estimated 12,000,000,000 in flood damage and have any lessons been learned that can be applied in preparation for future floods?

The extraordinary amounts of rain could be expected to produce unusually high river levels, and indeed, peak flood stages at 64 of 154 river gauging stations in the Upper Mississippi River Basin exceeded previously recorded levels. What is both surprising and disturbing is that, at 28 of the 64 stations, the record stages were caused by flows that were below previous records. How could record stages be produced by below-record flows?

River constriction. A look at the historical record of river stages and flows at gauging stations near St. Louis is instructive. Although the 1993 flow was about 20% less than the estimated record flow of 1844, the 1993 crest was 20% greater (about eight feet). There is some uncertainty about the 1844 flow because measurement techniques have changed, but the trend of increasing flood heights for a given river flow is documented by more recent measurements as well. For instance, the 1903 peak flow was just slightly less than 1993's (by about one percent), yet the latter's crest was 11.47' (30%) greater. The 1973 flood stage at St. Louis has been downgraded from a 100-year or even a 200-year event to only a 30-year one, meaning that great floods now occur with increasing frequency. Analyses of the 1973 flood by hydrologist Charles Belt of St. Louis University and a team of river engineers from Colorado State University indicated a significant man-made contribution, caused by constriction of the main river channels by wing dams for navigation and of the floodplains by levees for agriculture.

Earlier engineering studies on the Illinois River reached much the same conclusions, following major floods in 1844, 1904, 1913, 1922, 1926, and 1927. The 1983 Missouri River Flood Plain Study concluded that bank stabilization and navigation structures, excessive accretion of land along the channel, and floodplain constriction by levees, development projects, and highway and railway embankments all contributed to flood damage from Sioux City to St. Louis in 1951, 1952, and 1973. The same area was damaged in 1984, 1986, and 1993. The recommendations were consistent - set the levees back and use some of the levee districts to reduce flood stages. Flood damages would be lessened through buyouts or flood easements on levee districts that have broken repeatedly in the past (some actually are below the normal low water elevation of the river and have broken as many as 12 times since 1935), while flood heights against the remaining levees would be reduced.

Upland drainage. Man-made changes in the upland watersheds and tributaries, as well as mainstem rivers and floodplains, probably have increased flood heights. Well-intentioned water management projects were designed to keep water off the land or get it off the land as quickly as possible. Uplands have been tilled, drained, and paved over to speed water off the farms and urban areas; most midwestern streams are channelized to flush water quickly into the rivers; and 85-89% of the original wetlands in Illinois, Missouri, and Iowa-the three states hit hardest by the 1993 flood-have been drained. The importance of wetlands in reducing flood crests by soaking up rain and releasing it slowly was demonstrated in Illinois, where the ratio of peak stream flow to average rainfall decreased by 3.7% for every one percent increase in wetland area within a watershed.