The Source

by Martin Doyle

  The flooding didn’t stop in 1927. Starting in 1928 in Florida and continuing for the next decade throughout the United States, floods pummeled just about every region of the country. In the late summer of 1928, floods in central and south Florida killed 2,500 people, mostly poor blacks in the Everglades.5 Then, beginning in January of 1935, fatal floods hit the state of Washington, then the James River in Virginia and the Kanawha River in West Virginia, killing dozens of people and causing millions in damage. Later that spring, floods in Kansas and Texas in turn drew the attention to the middle of the country; then floods wracked New York and parts of the Midwest along the Ohio River. All told, more than two hundred people died in the floods of 1935 and damages topped $130 million. Less than a year later, floods began wreaking havoc yet again throughout the Northeastern United States. Rivers there reached levels previously thought unimaginable: the Connecticut River crested almost nine feet higher than any flood recorded since Europeans had settled there in 1639. Throughout Pennsylvania, 82,000 buildings were destroyed, and Pittsburgh sat under 16 feet of water on Saint Patrick’s Day with 47 dead and 67,000 homeless; Pennsylvania would later estimate total damages at a shade over $210 million. Mother Nature seemed to drive home the point by flooding Washington, D.C., for good measure: the Potomac River rose almost 20 feet above flood level, submerging the capital’s riverfront parks and requiring thousands of sandbag levees to protect the Lincoln and Washington Monuments.6

  These floods ended the era of state-centered federalism for flood control. With the Flood Control Act of 1928, the federal government got into the business of flood control. The 1928 act was staggering: It authorized spending $325 million ($4.5 billion in 2010 dollars) on flood control on the Mississippi and Sacramento Rivers, making it the largest public works project ever considered—exceeding even the $310 million authorized for building the Panama Canal. This act was focused on two specific rivers that had some connection with interstate commerce, thus allowing Congress to keep flood control under the umbrella of navigation. But when Congress passed the subsequent Flood Control Act of 1936, everything changed: not only did Congress authorize $310 million to fund the work, but Section 1 of the act opened with a “Declaration of Policy” stating that “it is the sense of Congress that flood control is a proper activity of the Federal Government.” That is, the federal government was now fully in the business of flood control. And the Corps was the arm of the federal government doing the work, becoming the nationwide planning and construction agency for flood-prone rivers.7

  Whereas the 1928 Flood Control Act had focused on the Mississippi River as well as some work on the Sacramento, the 1936 act was a truly national bill in that it sought to address flooding across the United States. More importantly, it also wholeheartedly adopted reservoirs, floodways, and cutoffs as valuable methods of flood control in addition to levees. Indeed, almost two-thirds of the funds the bill provided were earmarked for reservoir construction in New England, New York, California, and tributaries of the Mississippi River including the Ohio River (and its tributaries) and the Arkansas River. Every major river in every region was going to have a reservoir for flood control.

  That so many floods hit during the Great Depression was an essential element of changing the federalist calculus of flood control. The 1930s brought two essential changes in mind-set. First, the enormous costs for the infrastructure were to be federalized in the name of absorbing labor, and flood control projects were ideal due to the sheer scale of the projects and the number of jobs they absorbed. The building of reservoirs across expansive watersheds meant that many different regions of the nation got projects, and workers across the United States got jobs. Because much of the work was simple manual labor, poor workers could be drawn from the rural, impoverished regions that needed jobs. Second, the 1930s were the apex of the Progressive Era, when leaders were enamored with systems planning, optimization, and engineering. Charles Ellet’s vision of the flood problem as a watershed, system-wide problem dovetailed with the Progressive ideology of the time: the system had to be optimized, which meant that a central agency—the Corps of Engineers—had to be in charge of planning, engineering, and coordinating.

  Large river basins quickly became peppered with Corps-designed reservoirs and floodways or bypasses. Levees and floodwalls would remain critical parts of the system—and be expanded—but headwater reservoirs would store spring floods and then release the flows gradually during the summer, thus subsidizing low flows for navigation, irrigation, or hydropower. And when needed, floodways would be activated to divert some flow out of the main river. River systems became highly engineered, optimized hydraulic machines. Early twentieth-century floods gave the motivation; the progressives gave the ideology, and the New Deal provided the resources.

  Seventy-five years after the 1936 act, in 2011, Kent Parrish at the Corps and Peter Nimrod at the Mississippi Levee Board managed their parts of this system while faced with a 1927-scale flood. Each organization had its own roles and responsibilities that drew on the advantages of each level of government. The flood of 2011 demonstrated how the system worked when it worked very, very well.

  In the decades since the federal government first moved into flood control, it has implemented Ellet’s vision across the Lower Mississippi River. The system that emerged converted the Mississippi’s natural flows and variability into a conceptualized pipeline with valves and tanks—bypasses and reservoirs—to divert and store the worst-case scenario of the projected flood. Under that scenario, the various tributaries would be delivering their maximum or near maximum flows at the same time as other tributaries. To manage the flows, reservoirs would store water while bypasses diverted water in precise flows at precise locations.

  This is the system of river management that now exists, and it is the one Kent Parrish thinks about when he joins conference calls with Corps personnel up and down the Mississippi and Ohio Rivers, as well as with hydrologists from the National Weather Service. In the late nineteenth and early twentieth century, levee engineers had to guess how high the river would get each year based on haphazard telegrams from upstream sources and their own gut instincts acquired through years of living along the river. Federalizing the river systems, as is the case in the twenty-first century, dramatically improved the availability of information. Through their complex system of weather stations, flow gauges, and computer models, the combination of federal agencies—the U.S. Geological Survey, the National Weather Service, and the Corps of Engineers—now forecast with remarkable accuracy how high the river will rise at every city along the river.

  In spring of 2011, rain had been coming down in the Ohio River basin for weeks and weeks, as well as in the Upper Mississippi. Recalling the events two years later, Parrish remembered the atmosphere in the conference room when he and his engineers sat down for the interagency call on Good Friday in 2011. That was the call when the National Weather Service dropped the bomb of their forecasted river stages (water-level heights): 59 feet at Greenville and 53.5 feet on May 13 at Vicksburg. “When they said that number,” Parrish remembered, “My jaw dropped.” Someone else in the room simply let out a long whistle. A flood stage measuring in the mid-50s put this potential flood beyond the floods of 1873 or 1973 and into the orbit of 1927. It turns out that they underestimated; the Mississippi River eventually crested at Vicksburg on May 19 at just over 57 feet.

  But more important than accuracy of the information itself, the federal government now had infrastructure that was all working as a system in a kind of managed, organized way. Or, as Parrish put it, “You gotta coordinate this stuff.” Given predictions of the upstream flood wave as it made its way downstream, Corps reservoirs on major tributaries such as the Cumberland and Tennessee Rivers were being cleared of water ahead of time. The idea was to reduce their discharge for later when the Ohio was predicted to crest. Thus, as Ellet had suggested over a century earlier, the coordinating reservoirs successfully reduced the total load on the system.

  Ne
vertheless, the water level was still rising in the Lower Mississippi. Once the forecasting and reservoir operations and bypasses had been used, flood control came down to the local entities that bear the brunt of fighting floods. And this is where levee districts come in.

  The Mississippi River levees were built up dramatically between 1882 and 1972.

  The work of the local flood districts always seems to involve sandbags. Before levees on the Mississippi River were extended to their current height, engineers lived in chronic fear that the river would overtop the levee, thus destabilizing it and causing it to fail. To avoid overtopping, sandbags were used to add a few more feet to the height of levees. However, now that the levees are impressively high, the chances of overtopping are smaller, and sand boils have become the greatest threat to levees. When rivers are at flood stage, the river itself is dozens of feet higher than the land on the other side of the levee. This creates enormous hydraulic pressure on the levee itself: the pressure exerted by the Mississippi River on the levee at flood stage is about the same as that exerted by a school bus with all its weight pressing on a single square yard of area. Because levees have been built and rebuilt to resist this pressure, water instead finds that going through the soil beneath the levee is the path of least resistance. If the water flows too quickly under the levee, it can pull and flush soil—often sand—along with it. This creates a sand boil on the landward side of the levee because sub-levee water flow is effectively eroding sand beneath the levee. If the sand boil continues, along with the soil erosion underneath, the levee will collapse under its own weight.8

  The way to fight sand boils is to increase water pressure on the landward side to counteract the pressure of the water flowing underneath. This is done by encircling the sand boil with sandbags. As the water rises within the ring of sandbags, it begins to look like a swimming pool; once it gets deep enough, the pressure equalizes and slows the flow of water beneath the levee.

  Sandbagging sand boils is the work of local levee districts. It is the last line of defense—trench warfare—against a flood. During the 2011 flood, levee districts in Louisiana alone placed more than a million sandbags around boils. On the Mississippi side, Peter Nimrod and others from the Mississippi Levee Board continuously checked every foot of the 212 miles of their levees in search of boils or other weaknesses. When sand boils emerge, the levee district has to find the labor to take care of them immediately. Mobilizing sandbag crews, often in the middle of the night, is a large part of Nimrod’s job at the levee district.

  Most people, if they ever handle a sandbag, do it when the sand is dry. But flood fighting is wet, noted Nimrod: “Sandbaggin’ in a flood is moving sandbags wet; imagine how heavy they are; about 65 to 70 pounds.” Each time a sand boil appears, the levee crews—often including prisoners who work in exchange for time off their sentences—dash to it and ring it with sandbags. They often do this backbreaking work the entire night while the river looms four stories above them on the other side of the levee. Standing at the remnants of a sand boil that they fought during the 2011 flood, Bobby Thompson from the levee board described how he felt during the many days and nights of the flood fight: “I wasn’t scared. I was just tired. I never been in combat, but I don’t know how real fightin’ can be more tirin’ than flood fightin’.”

  A sand boil near Greenville during the Mississippi River flood of 1927.

  This mundane, on-the-ground, backbreaking work is part of twenty-first-century flood control. It involves a staggering array of governments and responsibilities: NASA satellites report precipitation, and the National Weather Service forecasts flood stages; Corps of Engineers generals zip around by helicopter to optimize reservoir levels and operate bypasses and floodways; state governments organize evacuations of the people living in areas that might be flooded; and local levee districts up and down the river do the nitty-gritty work of ensuring the structural integrity of the levees. This system is all the result of a layered approach to federalism that in many ways has worked. On May 8, 2011, the Mississippi River gauge at Greenville, Mississippi, topped 60 feet for the first time since 1927; yet the levees held while the upstream reservoirs peaked at their maximum storage capacity. Indeed, the discharge of the Mississippi River from Cairo, Illinois, to Baton Rouge, Louisiana, was higher in 2011 than it had been in 1927, yet flooding caused minimal damage.9

  Despite the levee managers’ enormous successes in 2011, when I visited two years later, their attention was already focused on their next potential project—a gargantuan pump. This pump, known as the Yazoo Backwater Project, was envisioned to send enormous amounts of water out of the Yazoo River area of the delta, over a levee, and into the Mississippi River. While much of the project had been constructed, including the levees, the final piece was the not-yet-constructed pumps. Both Kent Parrish and Peter Nimrod spoke of the pump topic repeatedly as we toured levees of the region. The men were frustrated by not being able to get money appropriated from the federal government to build this project that they felt was so plainly needed. As they pointed out, Congress had authorized the project in the mid-twentieth century; but lawmakers hadn’t yet appropriated funds to pay for the work. Parrish and Nimrod were frustrated that the federal government would not fund another part of this clearly successful infrastructure project that had been part of the broader projects protecting the Mississippi delta. Perhaps more cynically, they were trying to bring home some political pork as part of protecting their region from floods. And they were not alone.

  Although the need to control floods had brought the federal government into the business of dams and levees, the need to absorb labor during the Depression had justified the federal government to spend freely on flood control infrastructure. When states and levee districts were economically crippled, the federal government stepped in and paid for these projects, largely to provide jobs but also to get large projects built. But like many other aspects of river infrastructure during this era, projects didn’t end once the Depression did. The pendulum of federalism had swung dramatically toward the federal government being the locus of power, and there was little desire to go back to the way things were.

  Flood control infrastructure projects became a favorite flavor of political pork for the next half century. Congressional leaders found they could deliver local benefits at federal expense, and the Corps provided an unending list of projects for Congress to fund. The 1936 Flood Control Act kick-started the flurry of activity and led to the Corps constructing hundreds of flood control reservoirs along with tens of thousands of miles of levees. Eight years later the 1944 Flood Control Act allowed for the construction of mammoth dams, along with their accompanying reservoirs, on the Missouri River: Fort Peck, Garrison, Oahe, Big Bend, Fort Randall, and Gavins Point together inundated over 750 miles of river and stored almost 74 million acre-feet of water in a string of human-made lakes that is longer than Lake Michigan and Lake Superior together. Subsequent flood control acts followed rapidly and approved projects in every corner of the United States that had not gotten a project from the 1936 act. From the founding of the United States until 1928, local governments had borne the bulk of flood control costs; but from 1928 to the close of the twentieth century, the federal government took up spending. This was the era of minimum federal restraint for flood control.10

  And then, just as dramatically as it had begun, the pendulum of federalism swung back. For flood control infrastructure, the pendulum started its dramatic reversal on November 17, 1986, at the pinnacle of the Reagan Revolution and the president’s efforts to decrease the federal government’s role in almost every aspect of society. For flood control, this meant putting the impetus back on local and state governments to spend their own money on projects. The Water Resources Development Act of 1986, passed in November, required local sponsors—local districts or state governments—to pay half the costs of projects. Proposed projects, even those already authorized by Congress, came to a screeching halt. Local levee districts and even state gover
nments found they didn’t desire many flood control projects nearly as much when they had to pay 50 percent of the often exorbitant costs. Indeed, this was the case with the Yazoo Backwater Project in Mississippi; the project had been authorized in 1941, and some components had subsequently been constructed. But after the 1986 act, construction of the final pump station was postponed indefinitely because the local district could not afford its share of the costs. Projects across the United States faced similar fates. From 1986 on, the era of massive flood control infrastructure was largely over; federal restraint on spending and shared local responsibility proved enough to stanch the hunger for constant construction of flood control mega-infrastructure.11

  Infrastructure was the first flavor of political pork for flood control. But with federal coffers shut for infrastructure, pork seekers began looking for other sources of federal funds, and they found a vast supply in the form of disaster relief. Just as building flood control infrastructure is not necessarily greedy or poor governance, providing funds for disaster relief can create similarly perverse incentives when not disciplined by federalism. And just as Charles Ellet anticipated the problems of the flood control system in a government report, the emerging problems with disaster relief would be exposed in another scientific report on flooding—this time written by a brigadier general from the Corps of Engineers.