A Case Study on the Great Mississippi and Missouri Rivers Flood of 1993

Connor Keef

GEOG 3023

Dr. Scott Greene

4 May 2014

A Case Study on the Great Mississippi and Missouri Rivers Flood

of 1993

Abstract. There have been many catastrophic floods along the

Mississippi River, but one of the most devastating of these

floods was the “Great Flood of 1993.” The flood lasted six

months, from April through October, and caused more than $15

billion in damage (Larson). To date, it still ranks as the most

devastating and costly flooding disaster in United States

history.

The purpose of this project is to investigate the amount and rate

of precipitation that fell over the region to cause the rapid

water rise and to investigate the streamflow and hydrology of the

Mississippi and Missouri river basins. We will also study

changes that have been made since the event to determine the

effectiveness of new engineering in minimizing the hazards of

river floods today.

Introduction. The Great Flood of 1993 is one of the most

catastrophic and devastating floods to hit the United States of

America. Mark Twain said in his book Eruption “The Mississippi

River will always have its own way; no engineering skill can

persuade it to do otherwise…” (Johnson et. al. 1). What Twain

has said was shown to be true during the flood. Many levees had

been built along the banks of the Mississippi River and its

tributaries to regulate water levels. Hundreds of these failed

or were overtopped, causing as much as $20 billion in damages

(Johnson et. al. 2). Tens of thousands of people were forced to

evacuate their homes. Whole towns and cities were completely

underwater and millions of acres of farmland was rendered

unusable for years (Larson). So what caused this flood to be so

much more devastating than previous floods? In the following

paper, we will research the ingredients that came together to

produce a flood event of this magnitude.

Literature Review. There had been many floods of the Mississippi

River in the past, including the previous “Great Flood” in 1927.

However, they all paled in comparison to the 1993 Flood. To aid

in our analysis of the Great Flood of 1993, our group selected

several journal and government publications that investigated the

roots of the flood. Many of our sources focused on the

meteorological or climatological viewpoint, with a few others

concentrating on the hydrological aspect of the flood.

Bell and Janowiak (1994) explored the link between the El

Nino/Southern Oscillation and other atmospheric circulations and

the corresponding increased precipitation across the midwestern

United States, especially the Greater Mississippi River basin.

In the journal article “Atmospheric Circulation Associated with

the Midwest Floods of 1993,” the authors say “[the] region was

poised for potentially severe hydrologic problems prior to the

onset of excessive and focused rainfall events in June” (Bell and

Janowiak 682). An anomalously strong trough was centered over

the central North Pacific Ocean while a stronger than normal

upper level ridge was located across the western continental

United States. This, along with a dramatic intensification of

the Pacific jet stream, amplified storm tracks over the central

latitudes of the North Pacific Ocean (Bell 681). The North

Pacific Oscillation was especially strong as well (Bell 684).

They go on to say “the ‘cause’ of many of these events is simply

the temporary displacement of the atmospheric circulation from

its normal state rather than a direct response to external

forcing” (Bell 693). Multiple circulation features came together

to cause the quasi-cataclysmic flood event of 1993 (Bell 694).

Junker, Schneider, and Fauver (1999) investigated the causes

of mesoscale convective complexes and systems (MCSs and MCCs)

which resulted in the Great Flood of 1993. Over the summer of

1993, numerous sites achieved record monthly maximum

precipitation, with the estimated return period for the values at

this magnitude to be more than 100 years (Junker et. al. 701).

During that time, at least 43 different MCSs produced 24-hour

rain totals in excess of five inches (Moore et. al. 862), while

24 events had observed 24-hour rain totals at or exceeding six

inches (Junker et. al. 703). Much of their study concentrated on

analysis of upper air data in determining similarities between

the atmospheric conditions and the surface precipitation

accumulation.

While Junker et. al. investigate the causes of MCSs and MCCs

in the Great Flood of 1993, Kunkel, Changnon, and Angel (1994)

and Moore, Glass, Graves, Rochette, and Singer (2003) studied the

effects of the MCSs and MCCs on the Mississippi River basin and

why the flood was so severe. According to Moore et. al., MCSs

already account for between thirty and seventy percent of

precipitation during the warm season (April through September)

(Moore et. al. 861). Prior to the persistent rainfalls of the

warm season, the Mississippi River basin was plagued with above

normal rainfall, beginning in July, 1992 (Kunkel et. al. 813).

While a rainfall surplus is normally a good thing, it set the

stage for a historical flood which lasted as long as six months.

Rainfall totals between June 1, 1993 and August 31, 1993 were

unprecedented, with most areas in the Mississippi River basin

receiving 400 millimeters (15.75 inches) of rainfall, with many

areas collecting 600 millimeters (23.6 inches). Some areas even

received rainfall in excess of 900 millimeters (35.4 inches)

(Kunkel et. al. 813). From a historical standpoint, the

precipitation that fell has not been seen before since

observations began (Kunkel et. al. 814). For example, the June

through August rainfall amounts for Iowa were estimated as a

1000-year event using the maximum likelihood method and the

generalized extreme value distribution (Kunkel 814). From St.

Louis north to Hannibal, MO, the flood was estimated to be a 500-

year event (Kunkel 815). In St. Louis, MO, the Mississippi River

crested at 15.11 meters (49.5 feet), more than 1.8 m above the

previous record, which occurred in 1973 (Kunkel et. al. 811).

Kunkel’s research brought him to conclude that there were seven

conditions which led to the significance of the flood: record-

breaking rainfall totals, high incidence of moderate to heavy

rain events, persistence of saturated or near-saturated soils,

large-sized rain areas, the orientation of rain areas, a large

number of localized extreme rains capable of producing flash

floods, and below normal seasonal evapotranspiration (Kunkel et.

al. 818-821).

Historically and climatologically, the flood was

devastating. June experienced the greatest departure from the

average precipitation amounts (Guttman et. al. 1786). In July,

several stations recorded precipitation amounts at or above

twelve inches above average, while most sites experienced at

least six inches above average precipitation (Guttman et. al.

1786). Out of a total of 273 sites, forty-three experienced a

new maxima for three month precipitation amounts for May through

July 1993. The experienced precipitation amounts were estimated

to happen once every one-thousand years at several sites.

Precipitation amounts were approximately 200-350 percent above

average (Johnson et. al. 2).

Lins and Slack discuss the long-term trends of streamflow in

the United States by observing data from 395 stations located

around the conterminous United States (Lins and Slack 227). They

found that nationally there is an upward trend in “increasing

annual minimum streamflow” and that the upward trends actually

exceed the downtrends 4 to 1 (Lins and Slack 227). At the high

flows, however, the upward and downward trends are “roughly

equal” (Lins and Slack 228). They found that, hydrologically, the

nation as a whole is becoming wetter but not more extreme in its

events (Lins and Slack 228).

Pitlick points to this revelation, saying that “peak

discharges on the Mississippi and Missouri Rivers were

unprecedented” but the extremity of the event could be skewed

based on differing techniques (Pitlick 142). He goes on to say

that while floods were very common in the area in 1993, that most

rivers “experienced floods less than two times Qm and has return

periods of less than twenty years, and many rivers did not flood

at all” (Pitlick 143). This points toward some discrepancy, in

terms of flood frequency, in exactly how extreme this particular

event was in the Midwest.

The Mississippi River drains nearly 40 percent of the

contiguous United States as well as portions of Canada. This

drainage area is about 1.25 million square miles (Johnson et. al.

1). It should be no surprise that the river is susceptible to

fluctuations due to heavy rainfall in some areas of the basin.

But when the whole upper Mississippi River valley is inundated

with heavy rainfall for several months, the worst case scenario

can happen. And it did in 1993. At the same time, nearly 500

forecast gauges were at or above flood state (Johnson et. al. 1).

Lee Larson, the Chief at the Hydrologic Research Laboratory,

reported that 600 gauges were above flood stage at the same time.

More than one-hundred major rivers and tributaries flooded. The

flooded rivers and failed levies led to sediment being deposited

across a large portion of the Mississippi River Basin.

Data/Methods. To aid in our investigation of the Great Flood of

1993, we collected and analyzed several variables relating to

climatology and hydrology in the St. Louis metropolitan area. A

great majority of our data came from the National Oceanic and

Atmospheric Administration and the United States Geological

Survey.

Climatologically, we collected several pieces of data.

One of which was the monthly summaries for six St. Louis metro

stations: Valley Park, MO, St. Louis Science Center, MO, 7 SSW

St. Charles, MO, St. Charles, MO, Caholia, IL, and St. Louis –

Lambert International Airport, MO. This data was obtained from

the National Climatic Data Center (NCDC). Our data set was from

April 1, 1990, through October 1, 1994, which spans from three

years before the Great Flood of 1993 to one year after the event.

The data set contained total monthly precipitation and the

extreme maximum daily precipitation for that month, in tenths of

millimeters, as well as other variables such as snowfall and

temperature. We also viewed and collected information from the

NCDC on the Palmer Modified Drought Index. We also acquired data

from the United States Historical Climatology Network (USHCN) at

the Carbon Dioxide Information Analysis Center (CDIAC) website.

From a hydrological standpoint, we collected data on streamflow

and discharge from the USGS National Water Information System.

With our raw data, we created several visual aids, most

of which were in the form of graphs. First, we used the monthly

summary data to create a graph showing monthly precipitation for

all six stations. Using this data, we were able to see how much

more rain fell over the region than had the few years previous.

With the Palmer Modified Drought Index images, we created a graph

that showed what the drought situation for the area was. Knowing

this information also helped us determine how abnormal the

monthly precipitation measurements were. Using the data acquired

from the USHCN, we created a time-series graph showing

precipitation from 1878 through 2012. We also created two water

budgets: one showing the calculated average values for the 134

year period and one for 1993, the year of the event. This,

again, showed the abnormality of the event compared to normal.

Using the data from the USGS, we calculated the flood frequency

at St. Louis, and were able to determine what the frequency of

the 1993 flood event would be.

One thing that we observed is that some of the data was

missing for some of the monthly summaries. Also, the units on

the data were extremely confusing. For example, precipitation

used tenths of millimeters while snowfall measurements used

millimeters. Also, we used cubic meters per second for our

discharge data, but the data also had cubic feet per second. So,

while there were not any obvious problems or difficulties with

the data acquisition and analysis, it was important to keep a

watchful eye on our work to make sure we were using the correct

units.

Data Analysis. One of the first things that our group produced were

two water budgets. The first is a running average from 1878-

2012. This is seen in Figure 1. The average wet season begins

Figure 1: Average Water Balance for Bowling Green, MO (1878-2012)

around the first week of September and goes through the middle of

May. The dry season is from mid-May through August. This is

very typical of many places across the Central and Southern

Plains. To contrast, the dissimilarities of the water budget for

the 1993 Calendar Year, shown in Figure 2 are astonishing. The

dry season was nearly non-existent, as the precipitation and

actual

Figure 2: 1993 Water Balance for Bowling Green, MO

evapotranspiration readings are nearly the same. This is caused

by two variables. First, as Kunkel et. al. discovered in

“Climatic Aspects of the 1993 Upper Mississippi River Basin

Flood,” the nearly constant rainfall led to increased cloud cover

and below average temperatures (also shown in Figure 4), the

actual evapotranspiration was about 13 percent below average

(Kunkel et. al. 821). Additionally, the obvious second reason is

the increased precipitation. As was discussed in the literature

review section, from June through August, many areas had received

600 millimeters of precipitation (Kunkel et. al. 813). Figures 5

and 6 are time-series graphs showing precipitation and

temperatures. While neither show 1993 to be the maxima (for

Figure 3: Time-Series Graph Showing Precipitation for Bowling Green, MO

Figure 4: Time-Series Graph Showing Temperature for Bowling Green, MO

precipitation) or minima (for temperature), they do show that

they are either well above or well below average. While 1993 was

the main year of the flood, the gravity of the situation was

caused by the years previous to the event.

Many of the problems that caused the flood happened well

before the event itself. Figures 5 and 6 show the precipitation

for much of the St. Louis metropolitan area and the drought

conditions compiled by the Palmer Modified Drought Index.

Figure 5: Palmer Modified Drought Severity Index for Missouri Climate Division 2

Figure 6: Monthly Precipitation for the St. Louis, MO Area: April 1990 through October 1994

Both Figures 5 and 6 begin three years before the event and go

through the year following the event. The area was not in

drought for three whole years (and actually longer) before the

event. Much of the precipitated water ran off and caused the

flood conditions to deteriorate. The rivers, could not handle all

the runoff, and thus, caused the one of the worst floods in

United States History.

After acquiring discharge data of the Mississippi River, we

determined the flood frequency for the region. Figure 7 shows

the Flood Frequency Chart for the Mississippi River at

Figure 7: Flood Frequency by the Log Pearson III Method; Mississippi River at St. Louis, MO

St. Louis, Missouri. The 1993 Flood Discharge rate was 429,700

(USGS). This shows the severity of the flood. The 1993

Discharge Rate outranks the 200 year flood by more than 20,000

cubic meters per second.

Conclusion. It is evident why this was one of the worst floods in

United States History. With nearly 50 dead and up to $20 billion

in damages, the financial and emotional loss from this event were

astounding. What might be even more astounding, however, is the

number of things that came together to produce a flood of this

magnitude. Whether it be the water surplus in the previous years

which led to most water running off of the land, or the decreased

evapotranspiration due to cloud cover, or the repeated mesoscale

convective systems that affected the area, without these

ingredients, the flood would not have been nearly as bad. In the

wake of the flood, the federal government began to data

collection, flood control efforts, and evacuation practices for

floods nationwide. Had it not been for the occurrence of the

Great Flood of 1993, flood control and safety may not be

comparable as it is today.

Works Consulted

Bell, Gerald D., and John E. Janowiak. "Atmospheric Circulation

Associated with the Midwest Floods of 1993." Bulletin of the

American Meteorological Society 76.5 (1995): 681-95. Print.

"Great Flood of 1993." Wikipedia. Wikimedia Foundation, 05 Mar.

2014. Web. 02 May 2014.

<http://en.wikipedia.org/wiki/Great_Flood_of_1993>.

Guttman, Nathaniel B., J. R. M. Hosking, and James R. Wallis.

"The 1993 Midwest Extreme Precipitation in Historical and

Probabilistic Perspective." Bulletin of the American Meteorological

Society 75.10 (1994): 1785-792. Print.

"Historical Palmer Drought Indices." National Climatic Data Center

(NCDC). N.p., n.d. Web. 02 May 2014.

<http://www.ncdc.noaa.gov/temp-and-precip/drought/historical

-palmers.php>.

Junker, Norman W., Russell S. Schneider, and Stephanie L. Fauver.

"A Study of Heavy Rainfall Events during the Great Midwest

Flood of 1993."Weather and Forecasting 14.5 (1999): 701-12.

Print.

Johnson, Gary P., Robert R. Holmes, Jr., and Loyd A. Waite. "The

Great Flood of 1993 on the Upper Mississippi River—10 Years

Later." United States Geological Survey, May 2004. Web. 30

Apr. 2014.

<http://il.water.usgs.gov/hot/Great_Flood_of_1993.pdf>.

Kunkel, Kenneth E., Stanley A. Changnon, and James R. Angel.

"Climatic Aspects of the 1993 Upper Mississippi River Basin

Flood." Bulletin of the American Meteorological Society 75.5 (1994):

811-22. Print.

Larson, Lee W. "The Great USA Flood of 1993." The Great USA Flood of

1993. N.p., n.d. Web. 01 Feb. 2014.

<http://www.nwrfc.noaa.gov/floods/papers/oh_2/great.htm>.

Lins, Harry F., and James R. Slack. "Streamflow Trends in the

United States."Geophysical Research Letters 26.2 (1999): 227-30.

Print.

"Monthly Summaries." National Climatic Data Center (NCDC). N.p., n.d.

Web. 02 May 2014.

<http://www.ncdc.noaa.gov/cdo-web/results>.

Moore, James T., Fred H. Glass, Charles E. Graves, Scott M.

Rochette, and Marc J. Singer. "The Environment of Warm-

Season Elevated Thunderstorms Associated with Heavy Rainfall

over the Central United States." Weather and Forecasting 18.5

(2003): 861-78. Print.

"National Weather Service Weather Forecast Office." The Great Flood

of 1993. N.p., n.d. Web. 01 May 2014.

<http://www.crh.noaa.gov/lsx/?n=1993_flood>.

Pitlick, John. "A Regional Perspective of the Hydrology of the

1993 Mississippi River Basin Floods." Annals of the Association of

American Geographers 87.1 (1997): 135-51. Print.

Ponce, Victor Miguel. San Diego State University. Calculation of Flood

Frequency by the Log Pearson III Method. Web. 02 May 2014.

<http://ponce.sdsu.edu/onlinepearson.php>.

"United States Historical Climatology Network." Carbon Dioxide

Information Analysis Center. 02 May 2014

<http://cdiac.ornl.gov/epubs/ndp/ushcn/ushcn_map_interface.h

tml>.

USGS. Surface-Water Data for the Nation. Web. 02 May 2014.

<http://waterdata.usgs.gov/nwis/sw>.