USGS - science for a changing world

Kansas Water Science Center

Kansas Big Water - the Flood of 1951

Kansas Big Water Presented by the U.S. Geological 

Introduction of the Presentation

This presentation is about the flood of 1951 in Kansas and Missouri and its effects on people and flood policy.

Included are photos of flooding in the Kansas, Marais des Cygnes, and Neosho River Basins.

Methods of reducing flood losses are discussed.

A description of the collection of streamflow data and how it can be analyzed to determine the 100-year flood is included.

And, finally, a look into the future to see if there will ever be another flood like '51.

How do you remember 1951?

Collage of old photographs from 1951 of Yogi 
Berra, Joe DiMaggio, Miss America 1951, a 1951 Cadillac, Koren Conflict, Hank Williams, Sr., 
and a child in a Hop-a-Long Cassidy outfit.

The Yankees Yogi Berra and Joe DiMaggio win another World Series (this was the last year for Joe, first year for the rookie Mickey Mantle.

1951 Miss America, Yolanda Betbeze, (Bet-BEEZ) refused to be a Catalina Swimsuit model. Catalina withdrew its support from the Miss America Pageant and started the Miss USA Pageant.

Class was a 1951 Cadillac series 62.

The Korean Conflict continued.

Hank Williams (senior of course) was singing "Hey Good Lookin' whatcha got cookin'."

And every kid wanted a Hop-a-Long Cassidy outfit, complete with six-shooters.

Photo of women and children evacuated from North 
Topeka being carried ashore at the north end of Topeka Avenue bridge by rescue workers 
(photograph courtesy of Topeka Capital Journal.

But in Topeka, and in many cities in eastern Kansas, people were moving out!

They were fleeing the rising waters of the Kansas River.

Those trapped by the floodwaters were evacuated by boat as the streets and highways were closed to traffic.

Map showing rainfall amounts over eastern Kansas 
and locations of USGS streamflow-gaging stations in operation in 1951.

The flood was the result of a deluge of epic proportions. It had been wet all spring and early summer in Kansas, and the rivers were running full. A series of monster storms, beginning July 9th and ending July 13th, left more than a foot of rainfall over much of eastern Kansas. On this map, light-blue areas received greater than 8 inches, medium blue received greater than 12 inches, and darkest blue received greater than 16 inches.

The triangles are locations of USGS streamflow-gaging stations that were in operation in 1951. The red triangles are those stations that experienced the highest flood since streamflow records began.

The flooding could have been much worse on the Kansas River had the center of the storm been located 50 miles farther north. The huge volume of water from the 3 days of torrential rainstorms was divided between the Kansas, the Marias des Cygnes, and the Neosho River Basins. On July 13th, these three rivers had a combined flow of over 1.1 million cubic feet per second! This discharge was slightly more than the flow of the Mississippi River at St. Louis during the flood of 1993. However, it came from less than one-tenth the area. Flooding also occurred along the Verdigris River.

Photograph of North Lawrence, Kansas, July 13, 

The resulting surge of water filled many river valleys from bluff to bluff. Here we see the Kansas River at Lawrence.

North Lawrence was totally inundated.

Photograph of Kansas City Industrial District, 
July 13, 1951.

Probably the greatest catastrophe occurred at Kansas City where the Kansas River surged over the levees and virtually destroyed all residential and industrial interests in the flood plain. Note that the levees surrounding the Kansas City Airport were just high enough to protect that area.

Drowned livestock from the Kansas City Stockyards (between the Kansas and Missouri Rivers) and from farms upstream mingled with oil, sewage, sediment, and flotsam from thousands of destroyed homes.

Peak discharge at Kansas City was 510,000 cubic feet per second.

Photograph of Marais des Cygnes River at Ottawa, 
Kansas, July 11, 1951.

Floodwater from the Marias des Cygnes River overwhelmed the flood defenses of Ottawa and many surrounding communities.

Peak flow at Ottawa was 142,000 cubic feet per second at a stage of 42.5 feet.

Photograph of Neosho River at Parsons, Kansas, 

The flood on the Neosho River was nearly as large as that on the Kansas River.

This photo shows the Neosho Valley just east of Parsons, Kansas, on U.S. Highway 160 during the flood of 1943, which was a "garden variety" type flood of a mere 67,000 cubic feet per second. During the 1951 flood, the water was 11 feet deeper at this location, and the flow was 410,000 cubic feet per second!

The flow 83 miles upstream at Iola, Kansas was even greater at 436,000 cubic feet per second on July 13th.

As the floodwave moved downstream away from the center of greatest precipitation, the flood spread out, and the peak discharge became lower.

Graph showing the flood peaks in Kansas River 
Basin, July 1951.

This was not the case in the Kansas River Basin. The excessive precipitation of July 9-13th occurred along the axis of the Kansas River and all locations along the river had their highest flow at the same time.

This chart shows the timing of the peak flows in the Kansas River Basin. The solid lines show rivers that crested on or before the 13th of July. The dashed lines are for the rivers that crested after the 13th.

The Kansas River, with gaging stations near Junction City, Manhattan, Wamego, Topeka, Lecompton, DeSoto, and Kansas City, all had their highest stage and discharge on July 13th, (white triangles and line). The Saline River near Tescott (pink triangle and line) also peaked on July 13th, and the Big Blue River at Manhattan peaked the day before on the 12th.

However, the peak on the Smoky Hill River at Enterprise (green triangle and line) and on the Solomon River at Niles (yellow triangle and line) was on July 14th, a day after the highest flows occurred farther downstream.

Photographs showing transportation across the 
State at a standstill.

Transportation across the State came to a standstill.

Many airports were closed, including Kansas City, Lawrence, Topeka, Manhattan, Fort Riley, and Salina.

Trains were washed from their tracks, and hundreds of miles of rail were either closed or destroyed.

Hundreds of miles of streets and roads were impassible or damaged.

Photograph of a fire in a flooded area.

As homes and businesses were destroyed by the floodwaters, fires broke out from ruptured gas lines and oil tanks. Unable to dispatch fire trucks to the blazes because of impassable roads, the firefighters could only watch as fires burned to the water line.

Photograph of a woman returning home after the 
floodwaters had gone down.

Everyone had to deal with the mud and all that was in it. "All that was in it" included dead animals, bottles, ruined furniture, lumber, trees, and the deadly typhoid bacteria.

Photograph of a section of U.S. Highway 24 between 
Lawrence and Kansas City that was washed out by the flood.

It took workers over 2 weeks to repair U.S. Highway 24 between Lawrence and Kansas City, and months to repair or replace damaged or destroyed bridges.

In Kansas, 33 water-supply systems were shut down, requiring water be brought to the affected communities by tank trucks.

Photograph of prime farmland and the mounds of 
coarse sand that was left behind.

Prime farmland along the flooding rivers was scoured, and mounds of coarse sand was left behind. Some acreage was lost permanently as the river's main channel changed location.

Photograph of farmers pondering the fate of a 
tractor buried up to its hood by sand.

Farmers ponder the fate of a tractor buried by sand in addition to their way of life along the river.

Graph comparing the flood depths of 1993 with the 
floods of 1951, 1844, 1903, and 1785 on the Kansas and Missouri Rivers from Ogden, Kansas, to 
St. Louis, Missouri.

This graph compares the depth of the 1993 flood with three other large floods on the Kansas and Missouri Rivers from Ogden, Kansas, to St. Louis, Missouri.

The 1951 flood along the Kansas River was about 5 feet deeper than the 1993 flood from Ogden to Lecompton. However, at Kansas City and St. Louis, the Missouri River was from 10 to 7 feet higher in 1993. The reason, tall levees protected much of the urban areas from inundation, but forced the water to much higher elevations immediately upstream by squeezing the flow together.

Even though the 1951 flood was the most costly flood of record for Kansas, the one with the greatest flow and depth was the Great Flood of 1844. From Ogden to Lecompton, Kansas, the depth of this great flood was 4 to 5 feet DEEPER than the 1951 flood and nearly 10 feet deeper than the 1993 flood.

A major flood that occurred in 1903 had been the costliest until the flood of 1951. By the time it reached Lecompton, flood depths were only 2 feet lower than the '51 flood.

There is also evidence of a great flood in 1785 at St Louis, but no information of its severity in Kansas exists.

Map showing the major route of travel in 1844 was 
the Santa Fe Trail, not along the Kansas River flood plain..

Even though the Great Flood of 1844 had a much higher discharge and depth, there was virtually no development in the flood plain. The major route of travel in 1844 was the Santa Fe Trail. As seen on this map, there was much more interest in the smaller rivers south of the Kansas River over which the trail passed.

However, there is record of docks along the Missouri River near Independence being destroyed during the flood of 1844.

Photograph of Lawrence, Kansas, 1903, the 
Massachusetts Street bridge had just washed out.

The flood of 1903 was an eye opener for all those living along the Kansas River. Occurring some 59 years after 1844, rapid and extensive development within the flood plain was destroyed by the raging waters in 1903. This photo was taken in Lawrence looking northward over the Bowersock Mills and Power generator. The north span of the Massachusetts Street bridge had just collapsed.

Reducing Future Flood Losses

Graphic showing the ways to reduce flood losses.

Another flood of the magnitude of '51 could cause a tremendous economic loss. How can we reduce flood losses? Flood losses can be reduced by utilizing a combination of four methods.


  1. Building levees which protect high loss areas,
  2. Building flood-control reservoirs, which store floodwater for later release, thus lowering peak flows,
  3. Requiring businesses and homes in flood-prone areas to have flood insurance, which limit flood-plain development, and
  4. Utilizing flood warnings and flood forecasting, which allow the timely evacuation of people and property from the flood plain.

Photo showing how levees protect some areas but 
make flooding deeper in others.

Levees can protect some areas of the flood plain but only at the cost of raising flood levels at others. This aerial photo shows parts of Kansas City, Kansas, during the 1993 flood which were protected by the levee (right side). On the left side of the photo is an area without a levee, and the water was quite deep and flooding was extensive.

Photograph of Tuttle Creek Lake, July 1993, a 
flood-control reservoir.

Flood-control reservoirs like Tuttle Creek Lake north of Manhattan reduce flooding downstream. During the flood of 1993, the flood-control reservoir system in Kansas and Nebraska reduced the flow at Kansas City from 266,000 cubic feet per second to 172,000 cubic feet per second. Had the reservoirs not been in operation, flooding would have been much worse along the Kansas River, and the levee system at Kansas City would certainly have been overtopped.

Photograph showing the Rocky Ford Dam and power 
plant during the 1951 flood on the Big Blue River.

The Big Blue River is a major tributary that nearly doubles the flow of the Kansas River at Manhattan. This photo shows the Rocky Ford Dam and power plant during the 1951 flood on the Big Blue River, which had a peak flow of 93,400 cubic feet per second. Without the reservoir, the flow in 1993 would have been 107,000 cubic feet per second, instead of the observed 60,000.

Federal Emergency Management Agency, Flood Insurance Studies keep flood-plain development to a minimum

Photograph showing a flood insurance study and 

FEMA Flood Insurance Studies are performed for flood-prone communities in which the risk of flooding is determined by hydrologic analysis and areas that could be flooded are mapped. Anyone wishing to have a residence or business within a flood-prone area must purchase flood insurance and/or must build according to the projected 100-year flood depths. All new structures must be at least 1 foot above the 100-year flood levels.

This keeps flood plain development at a minimum.

Photograph of an example of flash-flood warnings 
and flood-crest predictions generated by the National Weather Service.

Flood Warnings. The National Weather Service is responsible for issuing flash-flood warnings and flood-crest predictions to warn persons to evacuate flood-prone areas. These warnings are disseminated through radio, television, and other warning systems.

All of these previous methods of reducing flood losses 
require hydrologic data. The U.S. Geological Survey is the premier agency for the collection 
of hydrologic data and the analysis of that data.

Collage of photographs showing how streamflow measurements 
are collected and a sample of a streamflow rating.

The U.S. Geological Survey has been in the business of supplying hydrologic data to the public for over 100 years. Measurements of floods on the Kansas River during the 1903 flood were made by USGS hydrographers using current meters and sounding weights and measurements are still being made using similar techniques, but under safer conditions today. Flow measurements of velocity and cross sectional area of the river result in a determination of volume per unit time or cubic feet per second.

At the time of the flow measurement, the depth of the water in the stream above some arbitrary datum is measured.

Measurements under various flow conditions are plotted on a graph, which becomes the streamflow rating for a particular site. Once the rating is developed, only the stream depth must be known in order to determine the flow in cubic feet per second.

1 cubic foot per second is 7.5 gallons per second or 450 gallons per minute.

Photograh showing a streamflow-gaging station in 
the field, the data that is received from satellite, and a map that is automatically generated 
for the Web in near real-time, there are 150 critical locations throughout Kansas.

The depth or stage of the water in the stream is recorded by the gaging station, and the stage data are transmitted by satellite link from the river bank to the USGS office. During flood conditions, the information is downloaded every 15 minutes.

By utilizing the streamflow rating for that site, a continuous record of flow or discharge becomes available. The data are automatically placed on the Web and are available for anyone. The top graph shows the discharge data for the Kansas River at DeSoto (solid line) in relation to the median or middle value (diamonds) for each day.

There are 150 locations (circles) throughout Kansas where this important information can be accessed.

Map Explanation
Black- highest flow on record for that day
Blue- highest 10%
Green - normal flow
Orange -lowest 10%
Red - lowest flow on record for that day

Photograph of a flood-frequency analysis 

A good example of how streamflow data are used can be seen in the flood-frequency analysis.

Annual peak discharges compiled over a period of record (say 10 or more years) can be used to estimate the statistical probability of occurrence of a certain magnitude of flood for 1-year period. This is called a flood-frequency analysis.

The probability value is determined by ranking the annual floods from greatest to least. The rank number of each flood is divided by the period of record plus 1 year. The result is the probability of magnitude. Each magnitude flood has a certain probability of occurring next year, and all probabilities add up to 100%. It is possible to get several 20-year floods in the same decade.

In flipping a coin, there is a 50% probability of getting a head and 50% probability of getting a tail. Question: What is the probability of getting a head after flipping four tails in a row? Answer: 50%. It works the same for floods.

The 100-year flood does NOT occur once every 100 years. Instead, that magnitude of flood has a 1% chance of occurring in any 1 year.

[1993 flood at DeSoto 10-50 year flood (0.1 to 0.02 probability)]
[1951 flood 100-500 year flood (0.01 to 0.002 probability)]

Will the Big Waters come again?.

History says that they will--

But when?

Research hydrologists are currently working on this question by analyzing streamflow data and its relation to various climatic indicators.

Graph showing 3-year moving averages of mean 
annual flow for the period of record for Neosho River at Iola and Kansas River at DeSoto, and 
precipitation for Eastern Kansas.

In addition to performing a frequency analysis on peak flows to determine the probability of a flood of certain magnitude happening next year, USGS research hydrologists have been investigating the time distribution of floods.

This graph shows a 3-year moving average of annual precipitation (light-blue line in the middle, with the 3-year moving average of streamflow for two large rivers in the State. There is a good relation between annual peak flows (floods) and annual average streamflow.

In this case we are looking at the Neosho River at Iola (top) and the Kansas River at DeSoto (bottom).

Large floods have occurred in the years that have a excessive precipitation and high annual streamflow--for example, 1993, 1973, 1951, 1927, and 1903. The average time between major flow periods is near 22 years, which is a common cycle in the Midwest for both floods and droughts. The 22-year cycle is quite similar to another well known cycle, the 22-year Hale solar mmagnetic cycle. (Magnetic poles of the Sun switch approximately every 11 years.)

But how does the magnetic properties of the Sun affect the flow of rivers?

Graph showing which years are at high risk for 
major flood using solar irradiance variations.

A Kansas researcher has taken this idea a step farther and may be able to predict when the next big flood on the Kansas and Neosho Rivers will occur.

If the precipitation and streamflow data from the previous slide are examined in greater detail, they also reveal a quasi-11-year cycle. These variations match quite well with the solar irradiance variations. Solar irradiance is the amount of energy that the Earth receives from the Sun. This value changes from year to year. The best relation exists when the solar data are moved ahead or lagged 5 years. The 5-year lag is consistent for rivers throughout the Central US.

The lag of 5 years gives us an opportunity to forecast streamflow up to 5 years into the future. Generally, solar irradiance variations can be estimated 10 years in advance, which gives us a 15-year look into the future.

If the relationships seen in the past continue, we would expect an increase in the risk of major flooding on the Kansas and Neosho Rivers sometime during the years 2003 to 2005 and again during the period 2013 to 2015.

The mission of the USGS. Hydrologic data, 
analysis, and research.

Even though the risk of flooding for any 1 year may be low, floods CAN and WILL occur in any year, any month of the year, and any time of day. The USGS is prepared to collect the best hydrologic data, under the worst of conditions, and make that information available to the public as soon as possible to protect life, property, and our way of life.

Hydrologic Data, Analysis and Research--That's the U.S. Geological Survey!

Science for a changing world

For information on hydrologic data, USGS 
publications, current projects and programs, and research in the Kansas District, visit our 
Web site at URL:

For information on hydrologic data, USGS publications, current projects and programs, and research visit our Web site at

A fact sheet titled "The 1951 Floods in Kansas Revisited" can be viewed online at: