PROTECTING SOCIETY FROM FLOOD DAMAGE
A CASE STUDY FROM THE 1993
UPPER MISSISSIPPI RIVER FLOOD
James T. Lovelace, P.E.
Chief, Hydrologic and Hydraulics Branch, St. Louis District
U.S. Army Corps of Engineers
Claude N. Strauser, P.E., L.S.
Chief, Potamology Section, St. Louis District
U.S. Army Corps of Engineers
Protecting society from flood damage, or more correctly minimizing such damage, requires a solid flood management philosophy. There are three key elements to large scale flood management; a good structural flood protection system, the ability to make sound decisions during the flood and post flood evaluations leading to positive improvements to the overall flood management plan. The authors will use the 1993 and other recent Upper Mississippi River floods to help illustrate these element
DESCRIPTION OF EVENT
The 1993 flood affected most of the Upper Mississippi River Basin. The basin drains all of part of 13 states and several Canadian provinces and encompasses 723,700 square miles above the confluence of the Mississippi and Ohio Rivers at Cairo, Illinois. The upper basin includes the Mississippi River from its source in Minnesota and the principal tributary, the Missouri River, which drains 529,300 square miles above its mouth in St. Louis, Missouri, including 9,700 square miles in Canada. Other major tributaries include Minnesota, Wisconsin, Iowa, Des Moines, and Illinois Rivers (See Figure 1.) It is the basin where the deluge of rain and consequent record flooding occurred during the spring, summer and fall of 1993.
The conditions that produced the flood began in the summer of 1992. July, September and November, 1992 were much wetter than normal in the Upper Mississippi River Basin. Winter precipitation was near normal, but a wet spring followed. The period from April to June 1993 was the wettest observed in the upper basin in the last 99 years. As a result, soils were saturated, and many streams were flowing well above normal levels when summer rains began.
A persistent atmospheric pattern during the summer of 1993 caused excessive rainfall across much of the Upper Mississippi River Basin. Major flooding resulted primarily from a series of heavy rainfall events over the Upper Mississippi Basin from May to August 1993 which were unmatched in the historical records of the Central United States. During the June-August 1993 period, rainfall totals surpassed 12 inches across the eastern Dakotas, southern Minnesota, eastern Nebraska, and most of Wisconsin, Kansas, Iowa, Missouri, Illinois, and Indiana. Over 24 inches of rain fell on central and northeastern Kansas, northern and central Missouri, most of Iowa, southern Minnesota and southwestern Nebraska. Up to 38.4 inches fell in east-central Iowa.
Wet antecedent soil, swollen river conditions, and record rainfall resulted in 1993 flood levels that ranged from below the 100-year up to the 500-year recurrence interval magnitude at many locations. For example, the 1993 flood stage at Louisiana, Missouri (about 100 miles above St. Louis, Missouri), is estimated to have had a recurrence interval of nearly 500 years. At St. Louis, Missouri, the recurrence interval was about 175 years and at Chester, Illinois (about 70 miles below St. Louis, Missouri), the resources interval was about 100 years.
The Flood of 1993 in the Midwestern United States was a hydrometeorological event without precedent in modern times. In terms of precipitation amounts, record river levels, flood duration, area of flooding, and economic losses, it surpassed all previous floods in the United States. Floods greater than the 1993 flood will happen in the future and it would be prudent to prepare for such future floods. U.S. Army Corps of Engineers (1995); U.S. Government Accounting Office (1995); Report of the Interagency Flood Plain Management Review Committee to the Administration Flood Plain Management Task Force (1994); U.S. Army Corps of Engineers (April 1994); Gaffney, Richard M. (1996); St. Louis District Corps of Engineers (March 1994).
THE UPPER MISSISSIPPI RIVER FLOOD PROTECTION SYSTEM
During the past 150 years or so the Mississippi River Basin has undergone extensive development by mankind. About 12 million people live along the mainstem and millions more along it's tributaries. The river annually generates about 7 billion dollars in agriculture and forest products and 429 billion dollars in manufactured goods. Over the years structural flood protection both public and private has been built to protect this population area and the associated economic investment.
The flood control system for the Upper Mississippi is made up of three components: flood control reservoirs, urban levees/floodwalls, and agricultural levees. There are about 60 Federal flood control reservoirs above St. Louis and about 1,600 levees. About 95% of these levees are agricultural levees (most are privately owned) providing relatively low levels of flood protection to millions of acres of cropland against floods of 10 to 50 years frequency. The remaining 5% are urban levees/floodwalls (mostly Federal) built to a very high level to protect cities and towns against floods of great magnitude. When the performance of a flood control system like this is evaluated, all three components must be considered as a whole and not evaluated as separate features.
During the 1993 flood, the Federal flood control reservoir system stored over 17 million acre feet of flood water. None of this flood water reached St. Louis until after the crest in August 1993. These reservoirs are credited with reducing flood levels at St. Louis by about three feet.
During the 1993 flood, all the levees/floodwalls built to urban design standards withstood the onslaught. No urban levee or floodwall was overtopped and the densely populated areas they protected were not flooded by the river.
As might be expected, most of the agricultural levees were overtopped in 1993 due to their relatively low level of protection. When very large floods occur, the agricultural levees overtop and this serves to reduce pressure on urban levees/floodwalls.
The flood control system on the Upper Mississippi River is a "patchwork" compared to the comprehensive system on the Lower Mississippi River, but it is still a substantial system. The Upper Mississippi River Flood Control System could not prevent all damages caused by a flood like the one in 1993, because it was not designed to do so. About 15 billion dollars of flood damage actually occurred in 1993. It is estimated that about 10 billion dollars of these damages occurred in the flood plain, the rest on upland agricultural land. The structural flood control system, however, did an excellent job and prevented 19 billion dollars of additional flood damage and averted an even worse disaster. Based on these estimates the total flood plain damages from the 1993 Upper Mississippi River flood could have been as high as 29 billion dollars had the structural flood protection system not existed. This indicates that the flood protection system in existence in 1993 reduced potential flood damages by well over 50%.
In April 1994 very heavy precipitation fell in western Missouri and eastern Kansas over part of the Missouri River basin about 250 miles west of St. Louis. During a relatively short period of time that April, as much as 14-18 inches of rain fell in that area. More than 10 inches of rain also fell in northern Missouri over parts of the Missouri and Mississippi River basins. It has been estimated that this amount of rain could have produced a flood stage of 47.4 feet (17.4 feet above flood stage) at St. Louis. This would have been the second highest flood at St. Louis exceeded only by the Great Flood of 1993. This great flood did not happen and the stage at St. Louis reached only 36.6 feet (6.6 feet above flood stage). This disastrous flood just a little less than one year after the Great Flood of 1993 did not occur because of the performance of two Corps of Engineers flood control reservoirs. (See Figure 2) Truman Lake in southwest Missouri (about 200 miles west of St. Louis) on the Osage River utilized almost half its available flood control storage and managed to absorb inflows of almost 400,000 cfs while holding releases to 500 cfs. Mark Twain Lake in northern Missouri (about 100 miles northwest of St. Louis) on the Salt River used almost 40% of its available flood control storage while absorbing peak inflows of about 75,000 cfs but limiting releases to only 50 cfs. The performance of these two flood control reservoirs reduced the stage at St. Louis by almost 11 feet and truly made this the "1994 flood disaster that didn't happen". Report of Interagency Flood Plain Management Review Committee to the Administration Flood Plain Management Task Force (1994); U.S. Army Corps of Engineers (April 1994); Busse, Lovelace and Strauser , "The Great Flood of 1994, The Disaster That Didn't Happen", Proceedings of the USCID Flood Management Seminar, St. Louis Missouri, (March 1995); Craig and Delaney, "Longitudinal Changes in the Mississippi Floodplain Structure", Status Report of the Upper Mississippi River Long Term Resource Monitoring Program, U.S. Geological Survey, (April 1997).
FLOOD MANAGEMENT DECISIONS DURING THE FLOOD
During the 1993 flood the many Corps flood control reservoirs were very effective in preventing much downstream flooding. While these reservoirs are not rigidly regulated as part of a basin wide flood control system they do each have water control plans that guides daily regulation during times of flood. The 1993 flood caused special problems because of the duration (about 6 months) and the multiple flood crests leading up to the record crest in July and August 1993. For most of the summer of 1993 the various Midwest offices of the Corps and National Weather Service conducted daily conferences by telephone to help regulate the individual reservoirs in the system in the most efficient manner possible. Mark Twain Lake (see Figure 2), while a small part of the overall 60 reservoir flood control system, provides a good example of how these reservoirs were regulated in 1993.
This reservoir was placed in operation in 1984. It controls a drainage area of 2318 square miles and is regulated for flood control, hydro-electric power, water supply, fish and wildlife conservation, recreation and water quality enhancement. 884,000 acre-feet is reserved for the storage of flood water. The maximum controlled release of water during floods is 12,000 cfs. In 1993 fourteen major storms occurred over the watershed above the dam during the period of April through September. During this period the Mark Twain Lake flood control pool was filled and emptied 3.5 times. Average daily inflows were as high as 92,000 cfs but releases from the reservoir never exceeded 12,000 cfs and during critical flood periods downstream releases were maintained as low as 50 cfs. This required very close coordination with other Corps Districts, National Weather Service, Hydro-power authorities and downstream flood fighting groups. The goal of Mark Twain regulation during this period was to assure that flood water was released at the maximum rate of 12,000 cfs whenever possible in order to keep the flood control pool as low as possible but cut back releases to minimum levels when flood crests were near at downstream locations. This intense coordination took place every day for almost six months and this reservoir prevented several devastating floods on the Salt River below the dam and routinely reduced Mississippi River flood stages by several feet. As an incidental benefit much of the flood release was made through the hydro-power turbines and very large amounts of hydro-power were generated. When you consider that similar activities went on in relation to all other Corps reservoirs in the Midwest the number of daily reservoir flood management decisions made in 1993 reaches staggering proportions.
Flood management decisions did not end with the crest of the 1993 flood. After the flood peaked at St. Louis in August 1993 two critical but potentially conflicting issues existed. Most of the federal flood control reservoirs were full, and prudent water control practice called for maximum releases to restore the flood control capacity of these reservoirs. Such high releases, however, could hold up to the fall of the Mississippi River and prolong it's closure to navigation (the river was closed to navigation for 36 days in 1993). The estimated value of lost navigation in 1993 was over 100 million dollars. At one time over 2000 barges were tied up waiting for the river to open for navigation.
Resumption of navigation at too high a level was particularly dangerous since the surviving levees were saturated and unstable. The Corps of Engineers, Coast Guard and towing industry established a joint Traffic Control Center in mid-July to evaluate requests for emergency vessel movements and develop implementation procedures for the timely and orderly reopening of the waterway to commercial navigation. When soils and foundation engineers determined levee conditions were no longer critical, the joint Control Center initiated a program of "test tows" and invited levee commissioners and rivertown mayors to witness and comment upon the wave wash effects of these large tows. Initially, tow speed and tow size restrictions were necessary to reduce risks of damage but navigation returned to normal within a matter of weeks.
By carefully monitoring weather forecasts, rainfall and rainfall runoff computations; releases from the flood control reservoirs were adjusted to adequately evacuate the flood water while allowing the Mississippi River to fall at a rate appropriate for levee safety. This procedure was successful and the river was reopened to navigation at the earliest possible time, the flood control capacity of the reservoirs was restored and no levees were endangered by premature resumption of navigation. Busse and Kopsky, "Lake Regulation During Record Breaking Conditions", ASCE Conference, San Antonio, Texas (1995); Lovelace and Strauser, "Perception and Reality Concerning the 1993 Mississippi River Flood: An Engineer's Perspective", Procedures of the International Workshop on Floodplain Risk Management, Hiroshima, Japan (1996); Manor, R. "Losses are Mounting by the Hour". article in the St. Louis Post Dispatch, (July 1993)
Mississippi River Mainstem Levee Overtopping.
The Columbia Drainage and Levee District (D&LD) is one of the mainstem Federal levees along the Mississippi River below St. Louis. Columbia D&LD, like most of the other levee districts, had a constant problem with seepage and sandboils caused by the extreme high water against their levees for extended periods of time. The Corps of Engineers awarded three emergency contracts to help stabilize the seepage problems, and levee district personnel built sandbag levees around numerous boils. However, between 0800 and 0900 on 1 August 1993, the levee overtopped in several locations, quickly flooding 14,000 acres of farmland including many homes. The area filled at the rate of approximately 1 foot per hour. Columbia D&LD was the first of the big mainstem levees to overtop. After the interior filled, water overtopped the lower flank levee along Fountain Creek at approximately 2330 on 1 August 1993 causing another break. The force of water was so great at this site that it continued across Fountain Creek overtopping and breaking through the northern flank of Harrisonville D&LD. Fountain Creek has levees on both sides extending from the hills to the east to the Mississippi river on the west, separating the Columbia and Harrisonville D&LD's (See Figure 3). The city of Valmeyer, located just south of the Harrisonville breach, soon began to flood forcing all remaining residents to evacuate immediately,
Harrisonville D&LD, Stringtown D&LD #4, Fort Chartres and Ivy Landing D&LD #5 comprise one continuous levee system, protecting some 46,500 acres of farmland including many homes. As the water began to fill the system, it was feared that the northern flank levee along Prairie Du Rocher Creek would overtop, and cause the southern flank levee along Prairie Du Rocher Creek to overtop and break in a manner similar to that which happened following overtopping and break in the Columbia D&LD. This would cause flooding in the city of Prairie Du Rocher. A massive sandbagging effort was initiated to raise the northern flank levee to prevent its overtopping. The morning of 2 August, St. Louis District sent a team of experts to the area to make an assessment. The team consisted of the Chief of the Geotechnical Branch - George Postol, the Chief of the Potamology Section - Claude Strauser, and the East-side Flood Fight Area Engineer - Dave Mueller. Options and recommendations were given to the levee commissioners of the three districts, including the levee commissioners for the Prairie Du Rocher and Modoc Levee and Drainage District (the next levee district south), as follows:
1. Let the flow of the Mississippi River come down the system at its own rate and take the chance of the lower flank levee being overtopped, which then could break the Prairie Du Rocher Creek southern flank levee, similar to what happened at the northern flank levee of Harrisonville.
2. Open the triple 72" drainage structures on the lower end of the Fort Chartres riverfront levee to build up the water in the lower end of the system to cushion the force of the coming flow.
3. If opening the triple 72's was not enough, then make a cut to a section of the lower end of the riverfront levee to increase the inflow of water at the lower end of the system. This water would fill the interior of the lower end of the levee system and create a "cushion" for the on-rushing water coming down from the upper end of the levee system. This would also provide an outlet for the water after the system filled, hoping to prevent overtopping of the southern flank levee.
The local levee commissioners made a decision to try and open the triple 72's. Two were opened and one could not be opened. The commissioners then decided at 1800 hours, 2 August, to request the Corps of Engineers to degrade a section of riverfront levee. (See Figure 3) A local contractor working in the area was awarded a contract on 3 August to remove as large a section of levee as he could safely remove within a short window of opportunity. A 400' section of levee 4-5 ft deep was removed by mechanical means. During this excavation flood water was flowing into the levee system.(See Figure 4) Local groups suggested blasting more outlets to speed the flow of water out of the levee district. The Corps recommended against this because of the possibility that blast vibrations would cause damage to saturated levees in the area. In the predawn hours of 4 August, local citizens set off explosive charges in two locations, contrary to Corps advice in an attempt to enlarge the openings in the levee and provide an outlet for the water entering the levee system. A few hours after the blasts, the Corps contractor used his equipment to enlarge the two openings to a depth of about 5 ft. By this time, flood water was flowing out of levee system into the river. (See Figure 3 & 5) Even though water overtopped the sandbagged flank levee of Prairie Du Rocher Creek in some isolated locations, the levee held and the city of Prairie Du Rocher was saved.
This innovative approach was successful and is an excellent example of good flood management decision making. These decisions are often made in a very short time frame and rely mainly on the knowledge and experience of the decision making engineers.
As demonstrated by these case studies, good flood management decisions made during a flood the size and duration of the Great Flood of 1993 saved many millions of dollars in additional flood damages and prevented much personal hardship. U.S. Army Corps of Engineers, "After Action Report - The Great Flood of '93",(1994).
POST FLOOD ACTIONS
After the Great Flood of 1993 many hundreds of both private and federal levees were badly damaged and in need of repair. In the St. Louis District alone (300 miles of the Mississippi River and part of the lower Illinois River) 47 levee districts qualified for federal assistance in making repairs. The repairs included repairing levee damage due to overtopping as well a repairs to relief wells and pump stations. The total repair bill just for the St. Louis District was abut 82 million dollars. In the case of many private levees some were repaired and some were abandoned.
The post-flood repairs on some of these levees were speeded up considerably by using a new state-of-the-art hydrographic survey vessel, the M.V. Simpson. This vessel carries a "sweep hydrographic survey system". Two long arms extend out from each side of the vessel and allow a very wide swath of river bottom elevations to be obtained in a short period of time. (See Figure 6) This vessel actually entered the degraded portions of the levee and surveyed the depth of the degration while the water was still well above flood stage. Design work to repair the levee could then begin immediately instead of waiting many weeks for conventional survey methods to obtain this data. Many weeks were saved in repairing overtopped levees by using this technique.
In the aftermath of the 1993 Flood , many different groups were motivated to formulate opinions on how to prepare for the next flood. The following four reports demonstrate the wide range of opinions on this subject:
The American Rivers Report. American Rivers, a private river conservation organization, took the position that the role of the Federal Government in flood management should be reduced and the resulting savings in funds should be channeled to state and local governments to help relocate existing development out of flood plain and to restrict any future flood plain development. This group took the position that existing levees and reservoirs were part of the flooding problem, not part of the solution. They basically supported a non-structural approach to flood management in the future.
The DELFT HYDRAULICS Report. The Upper Mississippi, Illinois and Missouri Rivers Association, a private group concerned with management of these watersheds commissioned DELFT HYDRAULIC in the Netherlands to develop an "outsiders" view of how the resources of these rivers were being managed. DELFT's view was that, on the whole, the resources of these rivers were generally under utilized. In terms of the Upper Mississippi River levee system they thought it needed to be upgraded and in some areas significantly raised. They pointed out that along the Rhine River in the Netherlands, the flood plains are protected against a 1250 year flood. They suggested that leveed agricultural areas be designed for deliberate flooding during extreme events to relieve pressure on urban areas. DELFT supported restoration of selected flood plain areas to a wetlands condition for ecological reasons.
Flood Plain Management Assessment Report. This was a Congressionally authorized study carried out by representatives of many Federal and state agencies. The aim was to assess flood control and flood plain management in the Upper Mississippi River Basin. In general this report confirmed that the structural flood control system was an important part of flood management. This report expressed the view that flood plains would be best managed in the long term through a combination of structural and non-structural measures applied in a reasonable manner that recognizes the risk of occupying flood prone areas.
Report of the Interagency Floodplain Management Review Committee. The President of The United States requested a complete report on the Great Flood of 1993. A group of federal experts developed this report as a part of that effort. This report is generally referred to as the "Galloway Report" after Brigadier General Galloway who was Executive Director of this effort. The report recognized the need for structural flood control measures like levees, especially for urban communities and infrastructure (roads, sewage plants, etc.) but stressed avoiding the risks of the floodplain by relocating existing development outside the floodplain where appropriate and restricting new development. This report also stressed restoration of wetlands where appropriate. The overall concepts of this report are shown in Figure 7, a schematic drawing from the report. Faber, S., "The Real Choices Report - The Failure of America's Flood Control Policies" (Jan 1995); DELFT HYDRAULICS, "A Balanced Management Plan for the Upper Mississippi, Illinois and Missouri Rivers" (Jan 1997); U.S. Army Corps of Engineers, "Floodplain Management Assessment of the Upper Mississippi, Illinois and Lower Missouri Rivers and Their Tributaries" (1995); Report of the Interagency Floodplain Management Review Committee to the Administration Floodplain Management Task Force, "Sharing the Challenge: Floodplain Management Into the 21st Century" (1994); Williams, R., "Flood of '93 - Surveys Provide Crucial Support", article in P.O.B., Volume19, Number1 (Oct-Nov 1993).
In the four years since the flood of 1993 many efforts have been made to help reduce future flood damages; many levees have been repaired, several small communities have been moved out of the floodplain, many thousands of individual buildings located in the floodplain have been purchased and removed, work is progressing on levees and or floodwalls for several unprotected urban areas, thousands of acres of land previously protected by levees are being purchased from willing sellers for restoration to wetland conditions. In many cases infrastructure such as roads, bridges, sewage treatment plants, etc. have been raised or protected from future flooding. The Corps of Engineers has developed a new computer model (UNET) to help analyze any proposed changes to the floodplain and to also provide an improved forecast model for future flood events. The proper application of flood insurance and floodplain land use restrictions have been stressed to local governments.
So far, adjustments to Upper Mississippi River flood management have been varied but seem to be progressing in a careful, deliberate manner using all the tools available; both structural and non-structural approaches are being used to bring about a reduction in flood damages in the future.
1. U.S. Army Corps of Engineers (1995), "Flood Plain Management Assessment of the Upper Mississippi and Lower Missouri Rivers and Their Tributaries".
2. U.S. Government Accounting Office (1995), "Midwest Flood - Information on the Performance, Effects, and Control of Levees".
3. Report of the Interagency Flood Plain Management Review Committee to the Administration Flood Plain Management Task Force (1994). "Sharing the Challenge: Flood Plain Management Into the 21st Century".
4. U.S. Army Corps of Engineers (April 1994), "U.S. Army Corps of Engineers Annual Flood Report to Congress for Fiscal year 1993".
5. Gaffney, Richard M. (1996), "Flood Report Analysis, "Missouri Department of Natural Resources, Report #54, 1996.
6. St. Louis District, U.S. Army Corps of Engineers (March, 1994), "The Great Flood of 1993 - After Action Report".
7. Busse, Lovelace and Strauser, (March 1995), "The Great Flood of 1994, the Disaster That Didn't Happen",
8. Craig and Delaney, (April 1997), "Longitudinal Changes in the Mississippi River Floodplain Structure", Status Report of the Upper Mississippi River Long Term Resource Monitoring Program, U.S. Geological Survey.
9. Busse and Kopsky, (1995), "Lake Regulation During Record Breaking Conditions", ASCE Conference, San Antonio, Texas.
10. Lovelace and Strauser, (1996), "Perception and Reality Concerning the 1993 Mississippi River Flood: An Engineer's Perspective", Proceedings of the International Workshop on Floodplain Risk Management, Hiroshima, Japan.
11. Manor, R. (July 1993), "Losses are Mounting by the Hour", article in the St. Louis Post Dispatch.
12. Faber, S., (Jan 1995), "The Real Choices Report - The Failure of America's Flood Control Policies".
13. DELFT HYDRAULICS, (Jan 1997), "A Balanced Management Plan for the Upper Mississippi, Illinois and Missouri Rivers".
14. Williams, R., (October - November 1993), "Flood of '93 - Surveys Provide Crucial Support", article in P.O.B., Volume 19, Number 1.