Read an Excerpt
A Watershed Year Anatomy of the Iowa Floods of 2008
UNIVERSITY OF IOWA PRESS Copyright © 2010 University of Iowa Press
All right reserved.
Chapter One What Causes Floods in Iowa?
A. Allen Bradley, Jr.
All rivers flood. In fact, they flood with surprising regularity-almost every year or two (Leopold et al. 1964). Some floods are harmless and almost go unnoticed, with water barely spilling out of the river's banks. Others are natural disasters that draw national attention.
Iowa experienced such a disaster not long ago. Floodwaters destroyed homes and businesses, shut down city services, disrupted travel in the region, and damaged farms and cropland. That was 1993. In its aftermath, many Iowans probably assumed they had witnessed the biggest flood that would occur in a very, very long time. And then came 2008, a second major flood disaster in fifteen years.
What do these floods tell us about the nature of flooding in Iowa? This chapter explores some general truths about floods-when they typically occur in Iowa and why. It also examines the common traits that big floods share.
The meteorological and hydrological causes of floods in the Midwest are well known. Floods occur whenever more water runs off the landscape than the river can hold within its banks. Too much runoff is caused by excessive rainfall or melting snow and ice-water amounts far greater than can soak into the ground. If theground is saturated or frozen and cannot hold more water, it makes flooding that much easier.
Are there certain times of the year when a flood is more likely? The flood history of Iowa (see figure 1-1) shows that spring and summer define the flood season; about 90 percent of all floods occur from March through August. In contrast, floods rarely occur in the fall and winter.
Iowa's climate plays a dominant role in the timing of floods. Fall floods are rare because soils are so dry; by fall, the summer sun and growing plants have depleted the water stored in the ground. Although big rainstorms still do occur, the ground can typically soak up a lot of the rainwater. Winter floods are rare because precipitation is often in the form of snow. While it can rain in winter, and rain on winter's frozen ground can produce flooding, this is uncommon because rainfall intensity is usually much less than at other times of the year. Still, rain and snow during the fall and winter help replenish the moisture in the ground, leaving soils much wetter-and less able to soak up water-at the start of the flood season in spring.
Spring floods occur in a variety of ways. The accumulation of snow during the winter months, and its melt over a relatively short period (of days and weeks), produces many spring floods (Soenksen et al. 1991). Spring weather, characterized by the clash of warm and cold air masses, marks the return of heavy rain with strong low-pressure systems. If widespread rain falls shortly after melting snow has saturated the ground, or if melting is caused by a warm period with rain on snow, a flood is often the result. Such events are responsible for the springtime (March) peak in Iowa floods (see figure 1-1).
Late spring and summer floods are caused by thunderstorms (Soenksen et al. 1991). It is both the heavy rainfall that thunderstorms produce and the location and timing of the rain that determine the magnitude and extent of flooding (see chapter 2). Indeed, most thunderstorms pass too quickly to produce a flood. But certain weather conditions-like a stationary front between warmer and cooler air masses-allow thunderstorms to repeatedly develop and dump heavy rainfall over the same region (Chappell 1986). With the slowly changing weather patterns typical of summer, thunderstorms developing over the span of several days can produce large rainfall accumulations. Such events are responsible for the late spring/early summer (June) peak in Iowa floods (see figure 1-1).
The drainage area of a river (the land area that contributes water to its flow) also plays a significant role in the timing of floods. Springtime flooding is more common on larger rivers, while summertime flooding is more common on smaller rivers and streams (see figure 1-1).
It is a bit of an oversimplification, but in general, larger rivers flood due to a high volume of runoff-widespread runoff from snowmelt or rainstorms over an extended period of time. As water runs off the landscape from a large drainage area, it accumulates and concentrates in the river. In contrast, smaller rivers flood due to a high rate of runoff-intense rainfall rates from individual thunderstorms can sometimes be enough. Water reaches swales and streams so quickly, and the river rises so rapidly, that flash flooding occurs within hours or a day of the storm.
Although these generalizations about floods describe their common traits, big floods have their own timetable. Iowa's biggest floods are concentrated in the late spring and summer (see figure 1-2); most of the biggest floods have occurred during the months of June and July. The dominance of big floods caused by summertime thunderstorm systems is dramatic for rivers large and small. Although springtime snowmelt and rain floods are very common in Iowa, they simply do not produce the biggest floods we see. Put simply, big floods are different.
Each big flood has its own story, but the floods in 1993 and 2008 stand out in Iowa's flood history. For large and medium-size Iowa rivers, about one in three experienced their biggest flood in 1993 or 2008; for small rivers, about one in five experienced their biggest flood then. No other Iowa floods match their widespread impact.
What was so unique about the weather in 1993 and 2008 that created such monster floods in Iowa? In many ways, the precursors to the 1993 and 2008 floods were eerily similar. Both floods occurred after a wet winter and spring. For the eastern third of Iowa, 2008 ranks as the wettest winter in the 114 years that weather records have been kept for the state; this was followed by the second wettest spring on record. Although the winter and spring precipitation were less in 1993, they were still well above average.
Then the rains intensified in the summer. The summer of 1993 stands out as the wettest on record. The precipitation accumulation was a whopping 2.3 times the summer average-an amount only 5.5 inches shy of the average precipitation for a year! Summer 2008 was rainy, with storms producing flooding in June, but precipitation diminished somewhat afterwards; still, it ranks as the ninth wettest summer on record. Looking at precipitation totals from the fall preceding the floods through the following summer, 1993 was the wettest year on record in eastern Iowa, and 2008 was the second wettest.
In both 1993 and 2008, the excessive precipitation over the year produced a large amount of water running off the landscape. Both years were extraordinary in terms of the sheer volume of water coursing through Iowa's rivers. In 1993, the annual discharge volume was the largest seen on record, about three to four times the long-term average at most streamgage sites in the Cedar and Iowa River basins. In 2008, the annual discharge volume was two to three times the average, the second largest on record. Coincidentally, both 1993 and 2008 were preceded by a wetter-than-average year, suggesting that soils may have held more moisture than usual as winter arrived.
Perhaps the most striking difference between the 1993 and 2008 floods was the duration of their summer rainy period. In essence, the 1993 flood was a summer-long event. An unusual and persistent weather pattern caused storms to develop and repeatedly track over the Midwest for nearly two months, from mid-June until mid-August (Wahl et al. 1993). Heavy rainfall accumulations in Iowa (see figure 1-3) resulted from a series of heavy thunderstorms throughout the summer, and affected most of the state at some point. By contrast, the 2008 flood was short-lived; the unusually wet weather persisted for only a few weeks, first dumping heavy rainfall in the northern and northeastern parts of the state, and later encompassing regions farther to the south (see figure 1-3). After heavy rains in late May and early June, Iowa remained relatively dry in the weeks that followed.
The longer rainy period in 1993 produced a much longer flood duration and higher summer runoff volumes, which were clearly felt at the Coralville Reservoir on the Iowa River just upstream from Iowa City (see figure 1-4). In both 1993 and 2008, the Coralville Reservoir filled during the wet spring, as reservoir releases were managed (according to its operating guidelines) to reduce flooding downstream. In June 2008, the rising water in the reservoir began overflowing the emergency spillway, but reservoir levels quickly lowered with the drier weather after the flood crest. In July 1993, after water began overflowing the emergency spillway, reservoir levels remained near capacity until September.
The differences in flood duration are also evident in the highest river levels seen at different locations (see figure 1-5). In 2008, the floods crested throughout Iowa from late May to mid-June. Afterward, they receded through the remainder of the summer. But in 1993, rivers crested over a span of five months and remained high for much of the summer.
The two flood events also differed significantly in the size of the area they affected. Overall, the impact of the 1993 flooding was significantly more widespread. Flooding affected not only Iowa but also portions of Illinois, Kansas, Minnesota, Missouri, Nebraska, North Dakota, South Dakota, and Wisconsin (see figure 1-6). That flood ranks among the biggest in both the upper Mississippi and lower Missouri River valleys. In contrast, the impact of the 2008 flooding was most severe in Iowa and Wisconsin, although the flooding also affected portions of Illinois, Indiana, Minnesota, Missouri, Nebraska, and South Dakota. On the Mississippi River, the 2008 flood approached 1993 levels only at the streamgage in Keokuk, Iowa; magnitudes diminished rapidly downstream.
Do these two floods provide clues as to what the next big flood may be like? They do suggest that it most likely would be a late spring or summertime flood, since this is the time when persistent and widespread thunderstorms can produce large rainfall accumulations (Hirschboeck 1991). If 1993 and 2008 are any indication, what happens in the winter and spring is important in setting the stage for a big flood. In both those years, after a wet winter, snowmelt and spring rains on then-saturated soils conspired to push rivers out of their banks in early spring throughout much of the state. As summer approached, the landscape could not soak up much more water, and rivers were primed to flood again when heavy rains ensued. In the future, a wet and active spring flood season should send a signal that the worst may not be over.
Are there any lessons that we should learn from witnessing two major floods in 15 years? One such lesson is that flood disasters are to be expected, and some may be even bigger than the ones we have already seen. We do indeed need to be very concerned about future flooding. Just because we have experienced a flood does not mean that the risk of another is less in subsequent years. Floods are random events; as any gambler knows, a run of bad luck can (and does) occur by random chance.
Unfortunately, many first perceived the 1993 flood as a once-in-a-lifetime event, an unsurpassable benchmark against which all other floods would be measured-a perception fueled by its long-lived nature, its widespread impact, and the media attention surrounding the disaster. The 1993 flood was unprecedented in many ways, but it was not extraordinary at all locations in the Midwest where flooding occurred (Parrett et al. 1993). Regrettably, the mistaken belief that 1993 was a benchmark may have shaped attitudes about the need to prepare for future big floods, and the response (or lack of response) to the rising floodwaters in 2008. Will we make the same mistake in the aftermath of the Iowa floods of 2008?
Chapter Two Why Were the 2008 Floods So Large?
Witold F. Krajewski Ricardo Mantilla
In June 2008, eastern Iowa experienced some of the worst flooding ever recorded. Floods, commonly defined as river waters overflowing their banks, devastated cities and the countryside alike. Some 1.2 million acres of Iowa's agricultural land was affected by floodwaters. From a plane, it was difficult to decipher the main channel of many rivers. The confluence of the Iowa and Mississippi Rivers seemed like a sea dotted with silos that protruded from the water's surface. In Cedar Rapids, a six-foot-tall man standing on the west bank levees would have had water flowing six feet above his head.
The peak flow on the Cedar River at Cedar Rapids reached 140,000 cubic feet per second (cfs) on June 13. This extremely large flow was a full five times as large as the river's average annual peak flow at this site. On the Iowa River at Marengo, the peak flow reached 51,000 cfs on June 12, four times as large as the average annual peak flow at this site. Downstream from Marengo, the Coralville Reservoir offered some protection, but in Iowa City (with a downtown area about 8 river miles below the dam), the Iowa River still reached a peak flow of 41,100 cfs and flooded significant areas of the University of Iowa campus. Peak flow there was three times the average annual peak flow (USGS 2009).
The sense of the 2008 floods' tremendous magnitude also is portrayed in figures 2-1 and 2-2, hydrographs that plot May and June discharge (river flow) at several streamgages on the Cedar and Iowa Rivers and their tributaries. Note how in most instances, especially for rivers draining large areas, the 2008 peak greatly exceeded the average annual peak flow, a level that roughly corresponds to a river filling its banks to the brim.
What was the genesis of these floods? Why were they so large? Many have pondered these questions, trying to grasp unique features of weather and landscape that might provide answers. Here we consider three contributing factors: the severe winter that preceded the floods, the high-intensity rainstorms of late May and early June, and the possibility of a perfect storm-not necessarily extremely large, but one in which precipitation was perfectly timed and located to raise the flow in river drainage networks to extraordinary levels.
First, consider the preceding season. The region experienced one of the snowiest winters in recent memory. Average snow depth across Iowa reached 11 inches by late February (NOHRSC 2008). The snowmelt saturated the ground, which remained wet well into May and significantly delayed the planting of crops in many parts of the state. With cool temperatures and croplands remaining bare, little vegetation was available to dry the soil through transpiration (the pulling of moisture from plant roots through leaves into the air).
Could these winter conditions help explain the June flooding? Examine the graphs in figure 2-3 (also plate 14). Panel A shows the daily variation of snow cover, snow depth, and snow water equivalent (the amount of water that would result from melting the snow) for December 2007 through March 2008, averaged over the state of Iowa (NOHRSC 2008). Note that while average snow depth reached over 11 inches by the end of February, the snow was virtually gone by the end of March, with snowmelt hastened by March temperatures rising above freezing (panel B; IEM 2008). The effect of this snowmelt on March flooding is clearly seen in panels D and E. That month, river discharges at both Cedar Rapids and Marengo rose in a single wave, reaching nearly 40,000 cfs and 18,000 cfs respectively, but then dropped back to normal levels by the end of March (USGS 2009). Thus, the heavy winter snows were not directly responsible for later flooding.
The snows did, however, lead to flood-prone conditions. Some of the melting snow remained in the fields as soil moisture. Thus, when the April rainstorms arrived (panel C; IEM 2009), the soils could absorb little of the spring rainfall. Instead, the wet ground produced significant runoff and additional high discharge waves in the rivers (note April discharges in panels D and E). Flooding from these storms peaked around the end of April.
For most of May, Iowa received relatively little rainfall. River levels throughout eastern Iowa returned to normal by May 25, as if providing the proverbial calm before the storm. Farmers planted crops late in May, and air temperatures reached typical levels for this time of the year.
Excerpted from A Watershed Year Copyright © 2010 by University of Iowa Press. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.