Applications of the Distributed Flood Forecasting System
Project Personnel (University of Iowa): Dr. Ricardo Mantilla (Assistant Research Engineer), Dr. Witold Krajewski (Professor)
One of the first applications of the distributed flood forecasting system was to investigate why floods were so severe on the Iowa and Cedar Rivers in June 2008. Our conclusion can be summarized as the “perfect storm,” where the timing and location of rain events conspired to maximize flood intensity at the hardest hit locations. The figure above, known as a hydrograph separation, shows the contribution of different storms to the peak flooding.
To understand this concept, one needs to follow the flow of water from where it falls on the landscape to a specific point on the river. River basins (or watersheds) are defined by all the points in a landscape that drain into a common place on a river. Every basin may include hundreds or thousands of streams and creeks that form a complex river network draining the landscape. The speed of water flow through the interconnected basin depends on several factors, most importantly land cover, soil type, level of soil saturation, proximity to a drainage channel, and the slope and friction drag along the channel. Once in a channel, water flows rapidly, but rarely faster than five feet per second in Iowa rivers.
Figure 2. Travel time for water in the Cedar River to reach Cedar Rapids and water in the Iowa River to reach Marengo.
For any location in the basin, we can roughly estimate the time that it takes for water in the river network to reach the basin outlet. Figure 2 shows the Cedar and Iowa River basins above Cedar Rapids and the Coralville Reservoir respectively, with their zones of water travel time. You can imagine that the water you watch at one point on a river is the sum of the runoff generated hours or days earlier from multiple upstream locations. When a single storm hits Charles City, for example, Cedar Rapids residents will observe its storm water flow about four days later. If another rain event hits Waterloo a day or two after the Charles City storm, when that storm water is flowing through Waterloo, the water from both events will combine to arrive in Cedar Rapids about the same time. This creates a “water traffic jam” analogous to rush hour traffic jams.
Figure 3. Location and timing of rain events on the Cedar and Iowa Rivers in mid June 2008.
Figure 3 gives an unprecedented picture of the location and timing of the daily rainfall accumulations on key dates in June 2008. The lower band of the June 8 intense rainfall reached Cedar Rapids a day or two after that storm, creating the first rise of the big flood (top Figure). The second, more northern band of June 8 rainfall required a longer travel time, about five days, and arrived at Cedar Rapids together with the massive rainfall that fell just north of Cedar Rapids on June 12, causing the city’s peak flood flows on June 13. Thus consecutive rain events were compounded to produce an unexpectedly rapid river rise and a single well-defined and extremely large peak flow in Cedar Rapids.
This same concept also explains why small basins experienced only mild flooding. These basins typically drain storm waters in a single day, so consecutive events do not usually cause a “water traffic jam.”
Understanding this concept and combining maps of Iowa’s river basin boundaries with up-to-the-minute precipitation data will help the IFC to improve Iowa’s flood prediction capabilities. This information may be combined with stream and river stage information from the IFC’s river stage sensors (see River Stage Sensors) to further improve our ability to monitor and predict flooding in Iowa.
Figure 4. Iowa Flood Center sensors placed at Iowa bridges, shown here as truck symbols, can be used to help monitor flood water traveling through a system and serve as a warning system when a bridge or road is flooded.
Funding Sources for this Project: State of Iowa, Iowa Flood Center and The National Science Foundation
