Case Study · Major River Flood · 1993
From May through September 1993, the upper Mississippi and Missouri Rivers flooded nine states for five months. More than 1,000 levees failed or were overtopped. 50 people died. $15 billion in damage. The rivers had been leveed to prevent this. The lesson that followed: levees don't eliminate flood risk. They move it, concentrate it, and then release it all at once when they fail — on communities that built right up to the wall because they believed the wall would hold.
Nine Midwestern States · May–September 1993
The spring of 1993 arrived in the upper Midwest with saturated soils from an unusually wet preceding year. When an exceptional rainfall pattern developed across the upper Mississippi and Missouri River basins — above-average precipitation from May through July across nine states — the rivers had nowhere to send the water. Groundwater was already high. Soils were already saturated. The rivers swelled. And swelled. And kept swelling, through June, through July, through August. The NOAA National Weather Service analysis of the 1993 flood describes conditions unprecedented in U.S. recorded history: approximately 600 river forecast points above flood stage simultaneously. Nearly 150 major rivers and tributaries affected. The longest sustained major flooding event ever measured in the United States.
What gave the 1993 flood its extraordinary consequence was not just its scale but what happened to the levees. The U.S. Army Corps of Engineers data, documented in the NOAA flood paper, records 40 of 226 federal levees as having failed or been overtopped — and 1,043 of 1,345 non-federal levees. In the Kansas City Corps district, the failure count for non-federal levees was 810 of 810. Every single non-federal levee in that district was overtopped or failed. The STLPR analysis of the flood quotes river engineer Gerald E. Galloway's post-1993 finding directly: "There is no such thing as absolute protection from floods." And: "There is no such thing as a levee that can't be overtopped." These were not cautionary abstractions. The 1993 flood had just proven them, at scale, across nine states.
May–Sep 1993
Duration
50
Deaths
$15B
Damage
1,000+
Levees Failed
9 States
Affected
The infrastructure statistics from the LiveScience account of the flood convey what it means for a flood to last five months across a major national transportation corridor: barge traffic on both mighty rivers halted for almost two months; Mississippi River bridges out from Davenport, Iowa to St. Louis, Missouri; ten commercial airports flooded; all railroad traffic in the Midwest halted; 1993 and 1994 crop harvests lost; numerous sewage treatment plants destroyed. 50,000 homes damaged or destroyed. 54,000 people evacuated — some never to return. The Mississippi River reached 49.58 feet at St. Louis on August 1, 1993 — a record that had stood for more than a century, broken by a full foot. The flood was not just a water event. It was a full-scale collapse of the infrastructure that modern Midwestern life depended on.
The Science
Think of a river in flood as water looking for the path of least resistance across a landscape. In a natural, unleveed floodplain, flood water spreads out over a wide area — the floodplain absorbs the excess flow, reducing the peak height of the flood and storing water that re-enters the river slowly. A levee forces the river to stay in its channel, preventing it from spreading into the floodplain. This raises the peak flood height — the water that would have spread out is confined to the channel and rises higher instead. The GAO's analysis of the 1993 flood noted that while the Army Corps' levees provided significant protection (saving $7.4 billion in damage), they also "caused floodwater to rise higher than it normally would because they confine a flood to a portion of a floodplain." The 1993 flood didn't just overwhelm levees. The levee system itself contributed to the extreme crest heights that overwhelmed them.
The STLPR analysis of the 1993 flood's legacy makes the problem explicit: after the disaster, the levee breaches at Chesterfield, Missouri, led officials to build a massive new "Monarch Levee" — which "has opened the way for extensive floodplain development." The lesson from 1993 — that no levee provides absolute protection — was acknowledged in policy documents and then effectively reversed in practice by building a bigger levee that invited more development behind it. Galloway's warning, cited in STLPR, is precise: "But it's going to be a problem if a perfect storm strikes the Midwest." The cycle of flood control infrastructure enabling floodplain development enabling greater exposure when flood control fails is the central structural problem in U.S. flood management that 1993 documented most clearly.
The Big Thompson Canyon flood in 1976 lasted hours. The 1993 Midwest flood lasted five months. The difference is not just scale — it is the type of problem. A flash flood is a life-safety emergency where the immediate decision determines survival. A multi-month major river flood is a sustained infrastructure and economic crisis where the cumulative damage from weeks of inundation — waterlogged soils, flooded basements, damaged foundations, contaminated wells, crop losses, transportation shutdowns — exceeds the damage from any single flood peak. The NOAA NWS flood analysis notes that "tens of thousands of people were evacuated, some never to return to their homes." The loss is not just property but community — neighborhoods that went under water and never came back together as functional places.
Timeline
01
Spring 1993: Saturated soils and above-normal snowmelt set up the basin for any further rainfall to produce flooding immediately. May–June: Persistent, above-average rainfall across nine states. Soils cannot absorb more water. Rivers begin rising. 600 river forecast points eventually above flood stage simultaneously. Nearly 150 major rivers and tributaries affected. The duration is unlike anything in recorded U.S. history.
02
Summer 1993: 1,000+ levees fail or are overtopped. In the Kansas City Corps district: 810 of 810 non-federal levees fail — 100%. Federal levees hold better but 40 of 226 still fail or are overtopped. The GAO later finds that while federal levees saved $7.4 billion in damage, they also raised flood heights above what they would have been in an unleveed floodplain. Communities behind failed levees have no time to prepare.
03
August 1, 1993: Mississippi River crests at 49.58 feet at St. Louis — a record that stood for over a century, broken by a full foot. Barge traffic halted on both rivers for almost 2 months. Bridges out from Davenport, IA to St. Louis. 10 commercial airports flooded. All Midwest railroad traffic halted. 50 deaths. 50,000 homes damaged or destroyed. 54,000 evacuated. $15B in damage. The 1993 and 1994 crop harvests are lost.
04
1993–present: STLPR documents post-flood policy evolution: "There is no such thing as absolute protection from floods." But in practice, the response in many areas was to build bigger levees — which then attracted more floodplain development behind them. LiveScience (2008): "Some residents who moved into the floodplain since 1993, with assurances from FEMA and other officials that they would be safe, have seen their towns submerged... in many cases have no flood insurance." History repeated itself, as documented.
Human Decisions
What the flood produced
The GAO's analysis of the 1993 flood documented that 157 of 193 Army Corps of Engineers levees in affected areas successfully prevented flooding and reduced damage — saving an estimated $7.4 billion. The federal levee system performed as designed in the majority of its structures. The distinction the GAO makes is important: it was the non-federal levees — built with little or no engineering oversight, often by local communities or agricultural interests — that failed at catastrophic rates. 1,043 of 1,345 non-federal levees failed.
STLPR's retrospective documents that the 1993 flood was "a watershed event for floodplain management, with some of the lessons learned from the deluge exerting influence far beyond the Midwest." Policy emphasis shifted from engineering-only solutions to risk assessment and land use management. The Association of State Floodplain Managers notes increased recognition of state and local roles in floodplain decisions. Wetland restoration was recognized as a flood mitigation measure. These policy shifts were genuine responses to what 1993 proved about the limits of engineering solutions alone.
The structural problem 1993 documented
The STLPR account of post-1993 Chesterfield, Missouri is the clearest illustration: levee failures in 1993 prompted officials to build the Monarch Levee, which "has opened the way for extensive floodplain development." The Galloway quote the article uses is precise: "There is no such thing as a levee that can't be overtopped." Yet the response to levee failure was to build a bigger levee — behind which more development moved. The cycle is not a failure of engineering. It is an incentive structure: a bigger levee reduces insurance costs and increases land values, which attracts investment, which creates pressure to maintain and expand flood control infrastructure, which gives the appearance of permanent protection that no levee can actually provide.
The NOAA NWS paper's table of levee failures documents the Kansas City Corps district's non-federal levee performance: 810 of 810 failed. Not one held. These levees protected agricultural land, residential communities, and businesses that had been built behind them under the assumption that they would hold under all but extraordinary flood conditions. The 1993 flood was extraordinary. Every single levee in the district failed. Communities that believed themselves protected were instead exposed to flooding they had no preparation for because they had no reason to expect it.
The LiveScience account documents the most consequential downstream effect of the levee-false-security cycle: "Some residents who moved into the floodplain since 1993, with assurances from FEMA and other officials that they would be safe, have seen their towns submerged... creating personal financial disasters for people who in many cases have no flood insurance." The assurances were technically accurate within the design parameters of the rebuilt levees. They were also structurally identical to the assurances given before 1993. And the result, when another flood came, was structurally identical to 1993.
The cascade lesson
The 1993 Midwest flood is the definitive American case study on the relationship between flood control infrastructure and flood risk exposure. The levees that protected communities throughout the normal range of flood events created the conditions for catastrophic exposure when a flood exceeded their design capacity. Development that moved in behind levees — because levees reduced flood insurance requirements, raised land values, and created the perception of protection — became exposed when the levees failed. The policy response was to build bigger levees, which attracted more development, which created more exposure for the next flood that exceeds the new levees' capacity. STLPR's Galloway quote has not been disproved in the three decades since: "There is no such thing as absolute protection from floods." For individual households, the lesson is that living behind a levee or in a floodplain with flood control infrastructure is not the same as being safe from flooding. It means being safe from flooding within the levee's design parameters — and potentially more exposed when those parameters are exceeded.
What You Can Do Now
The 1993 flood's lessons apply to any household that lives near a river with levee protection, in a floodplain that has been developed, or in an area that has been told it's safe from flooding because of engineered flood control.
Every levee has a design capacity: the height of flood it was built to hold. Floods that exceed that height overtop or fail the levee, releasing water that may have been accumulating behind it. Ask your local flood control district or Army Corps of Engineers office: what is the flood level that would overtop the levee protecting your area? What would happen in the protected zone if it failed? This information exists and is public — and knowing it transforms the levee from a security guarantee into a risk parameter with specific limits.
Find your local flood resourcesThe LiveScience account documents that residents who moved into the floodplain after 1993 with FEMA assurances and no flood insurance experienced financial catastrophe when subsequent floods arrived. Being in a zone designated as protected by a levee may reduce your flood insurance premium — but it does not mean you cannot flood. Flood insurance outside high-risk zones is less expensive than in them, and it covers the scenario the 1993 flood documented: levee failure producing flood in areas that were not supposed to be flood-prone.
Insurance resilience guideThe 1993 flood halted barge traffic for two months, flooded ten commercial airports, stopped all Midwest railroad traffic, and took out Mississippi River bridges for hundreds of miles. These are the supply chain arteries that deliver food, fuel, and goods to Midwestern communities. A two-week food and water buffer — stored at home and rotated regularly — provides independence from the supply chain during the acute disruption phase of a major flood event.
Two-week preparedness guideSome people evacuated in 1993 "never returned to their homes." A rapid evacuation during a major flood event may be permanent or semi-permanent. Having a go-bag with originals or copies of essential documents — insurance policies, property records, identification, financial records, medical information — means you can function during a displacement that may last weeks or months. Scan and store documents digitally in cloud storage as a backup.
Document preparedness guideThese are not the same statement. "Protected by a levee" means "protected from flooding up to the flood level the levee was designed for." It does not mean "protected from all flooding." FEMA's flood maps in levee-protected areas may designate your zone as reduced-risk — but they also typically indicate the residual risk if the levee fails. Ask your local floodplain administrator what the residual risk designation means for your property. Knowing the distinction between "lower flood insurance premium" and "not a flood risk" is the most important conceptual takeaway from 1993.
Flood risk guideFlood case study series
1927 Mississippi covers federal disaster response origins. Big Thompson 1976 covers flash flood survival. Buffalo Creek 1972 covers industrial dam failures. Grand Forks 1997 covers forecast error and compound disasters. Each teaches a distinct lesson about what floods actually destroy and how.
Full flood case study seriesSources