Rethinking, Then Rebuilding New Orleans
This time around, science should contribute to a systemic long-term plan that will better accommodate the natural forces that shape the Mississippi Delta.
New Orleans will certainly be rebuilt. But looking at the recent flooding as a problem that can be fixed by simply strengthening levees will squander the enormous economic investment required and, worse, put people back in harm’s way. Rather, planners should look to science to guide the rebuilding, and scientists now advise that the most sensible strategy is to work with the forces of nature rather than trying to overpower them. This approach will mean letting the Mississippi River shift most of its flow to a route that the river really wants to take; protecting the highest parts of the city from flooding and hurricane-generated storm surges while retreating from the lowest parts; and building a new port city on higher ground that the Mississippi is already forming through natural processes. The long-term benefits—economically and in terms of human lives—may well be considerable.
To understand the risks that New Orleans faces, three sources need to be considered. They are the Atlantic Ocean, where hurricanes form that eventually batter coastal areas with high winds, heavy rains, and storm surge; the Gulf of Mexico, which provides the water vapor that periodically turns to devastatingly heavy rain over the Mississippi basin; and the Mississippi River, which carries a massive quantity of water from the center of the continent and can be a source of destruction when the water overflows its banks. It also is necessary to understand the geologic region in which the city is located: the Mississippi Delta.
The Mississippi Delta is the roughly triangular plain whose apex is the head of the Atchafalaya River and whose broad curved base is the Gulf coastline. The Atchafalaya is the upstream-most distributary of the Mississippi that discharges to the Gulf of Mexico. The straight-line distance from the apex to the Atchafalaya Bay is about 112 miles, whereas the straight-line distance from the apex to the mouth of the Mississippi is twice as long, about 225 miles. (These distances will prove important.) The Delta includes the large cities of Baton Rouge and New Orleans on the Mississippi River, and smaller communities, such as Morgan City, on the Atchafalaya. (Although residents along the Mississippi River at many places considerably to the north of New Orleans commonly refer to their floodplain lands as “the Delta,” the smaller rivers and streams here empty directly into the Mississippi River, not the Gulf of Mexico, and hence geologists more properly call this region the alluvial plain of the Mississippi River.)
The Mississippi River builds, then abandons, portions (called “lobes”) of the Delta in an orderly cycle: six lobes in the past 8,000 years (Fig. 1). A lobe is built through the process of sediment deposition where the river meets the sea. During seasonal floods, the river spreads over the active lobe, depositing sediment and building the land higher than sea level. But this process cannot continue indefinitely. As the lobe extends further into the sea, the river channel also lengthens. A longer path to the sea means a more gradual slope and a reduced capacity to carry water and sediment. Eventually, the river finds a new, shorter path to the sea, usually down a branch off the old channel. The final switching of most of the water and sediment from the old to the new channel may be triggered by a major flood that scours and widens the new channel. Once the switch occurs, the new lobe gains ground while the old lobe gradually recedes because the sediment supply is insufficient to counteract sea level rise, subsidence of the land, and wave-generated coastal erosion.
Mississippi Delta switching. The succession of different river channels and delta lobes during the past 5,000 years are numbered from oldest (1) to youngest (7). (Meade, 1995 U.S. Geological Survey Circular 1133, fig. 4C; also see Törnqvist et al., 1996, for an updated chronology.)
Geologist Harold Fisk predicted in 1951 that sometime in the 1970s, the Mississippi River would switch much of its water and sediment from its present course past New Orleans to its major branch, the Atchafalaya River. In order to maintain New Orleans as a deepwater port, the U.S. Army Corps of Engineers in the late 1950s constructed the Old River Control Structure, a dam with gates that essentially meters about 30% of the Mississippi River water down the Atchafalaya and keeps the remainder flowing in the old channel downstream toward New Orleans. Trying to meter the flow carries its own risks. During the 1973 flood on the Mississippi, the torrent of water scoured the channel and damaged the foundation of the Old River Control Structure. If the structure had failed, then the flood of 1973 would have been the event that switched the Mississippi into its new outlet—the Atchafalaya River—to the Gulf. The Corps repaired the structure and built a new Auxiliary Structure, completed in 1985, to take some of the pressure off the Old River Control Structure. The Mississippi kept rolling along.
Still, the fact remains that the “new” Atchafalaya lobe is actively building, despite receiving only one-third of the Mississippi water and sediment, while the old lobe south of New Orleans is regressing, leaving less and less of a coastal buffer between the city and hurricane surges from the Gulf. This situation has major implications.
At one time, the major supplier of sediment to the Mississippi Delta was the Missouri River, the longest tributary of the Mississippi River. However, the big reservoirs constructed in the 1950s on the upper Missouri River now trap much of the sediment, with the result that the lower Mississippi now carries about 50% less sediment (Fig. 2). It is ironic that the reservoirs on the Missouri, whose purposes include flood storage to protect downstream areas, entrap the sediments needed to maintain the Delta above sea level and flood level. Much less fine sediment (silt and clay) flows downstream to build up the Delta during seasonal floods, and much of this sediment is confined between human-made levees all the way to the Gulf, where it spills into deep water. Coarser sediment (sand) trapped in upstream reservoirs or dropped into deep water likewise cannot carry out its usual ecological role of contributing to the maintenance of the islands and beaches along the Gulf, and beaches can gradually erode away because the supply of sand no longer equals the loss to along-shore currents and to deeper water.
If Hurricane Katrina, which in 2005 pounded New Orleans and the Delta with surge and heavy rainfall, had followed the same path over the Gulf 50 years ago, the damage would have been less, because more barrier islands and coastal marshes were available then to buffer the city. Early settlers on the barrier islands offshore of the Delta built their homes well back from the beach, and they allowed driftwood to accumulate where it would be covered by sand and beach grasses, forming protective dunes. The beach grasses were essential because they helped stabilize the shores against wind and waves and continued to grow up through additional layers of sand. In contrast to a cement wall, the grasses would recolonize and repair a breach in the dune. (A similar lesson can be taken from the tsunami-damaged areas of the Indian Ocean. Damage was less severe where mangrove forests buffered the shorelines than where the land had been cleared and developed to the shoreline.) Vegetation offers resistance to the flow of water, so the more vegetation a surge encounters before it reaches a city, the greater the damping effect on surge height. The greatest resistance is offered by tall trees intergrown with shrubs; next are shorter trees intergrown with shrubs; then shrubs; followed by supple seedlings or grasses; and finally, mud, sand, gravel, or rock with no vegetation.
One of the major factors determining vegetation type and stature is land elevation. In general, marsh grasses occur at lower elevations because they tolerate frequent flooding. Trees occur at higher elevations (usually 8 to 10 feet above sea level) because they are less tolerant of flooding. Before European settlement, trees occurred on the natural levees created by the Mississippi River and its distributaries, and on beach ridges called “cheniers” (from the French word for oak) formed on the mudflats along the Gulf coast. The cheniers were usually about 10 feet high and 100 or so feet wide, but extended for miles, paralleling the coast. Two management implications can be derived from this relationship between elevation and vegetation: Existing vegetation provides valuable wind, wave, and surge protection and should be maintained; and the lines of woody vegetation might be restored by allowing the Mississippi and its distributaries to build or rebuild natural levees during overbank flows and by using dredge spoil to maintain the chenier ridges and then planting or allowing plant recolonization to occur.
The sediment loads carried by the Mississippi River to the Gulf of Mexico have decreased by half since 1700, so less sediment is available to build up the Delta and counteract subsidence and sea level rise. The greatest decrease occurred after 1950, when large reservoirs constructed trapped most of the sediment entering them. Part of the water and sediment from the Mississippi River below Vicksburg is now diverted through the Corps of Engineers’ Old River Outflow Channel and the Atchafalaya River. Without the controlling works, the Mississippi would have shifted most of its water and sediment from its present course to the Atchafalaya, as part of the natural delta switching process. The widths of the rivers in the diagram are proportional to the estimated (1700) or measured (1980–1990) suspended sediment loads (in millions of metric tons per year). (Meade, 1995 U.S. Geological Survey Millions of metric tons of Circular 1133, fig. 6A)
Of course, the vegetation has its limits: Hurricanes uproot trees and the surge of salt or brackish water can kill salt-intolerant vegetation. Barrier islands, dunes, and shorelines can all be leveled or completely washed away by waves and currents, leaving no place for vegetation to grow. The canals cut into the Delta for navigation and to float oil-drilling platforms out to the Gulf disrupted the native vegetation by enabling salt or brackish water to penetrate deep into freshwater marshes. The initial cuts have widened as vegetation dies back and shorelines erode without the plant roots to hold the soil and plant leaves to dampen wind- or boat-generated waves. The ecological and geological sciences can help determine to what extent the natural system can be put back together, perhaps by selective filling of some of the canals and by controlled flooding and sediment deposition on portions of the Delta through gates inserted in the levees.
The Mississippi River typically floods once a year, when snowmelt and runoff from spring rains are delivered to the mainstem river by the major tributaries. Before extensive human alterations of the watersheds and the rivers, these moderate seasonal floods had many beneficial effects, including providing access to floodplain resources for fishes that spawned and reared their young on the floodplains and supporting migratory waterfowl that fed in flooded forests and marshes. The deposition of nutrient-rich sediments on the floodplain encouraged the growth of valuable bottomland hardwood trees, and the floodwaters dispersed their seeds.
Human developments in the tributary watersheds and regulation of the rivers have altered the natural flood patterns. In the Upper Mississippi Basin, which includes much of the nation’s corn belt, 80 to 90% of the wetlands were drained for agriculture; undersoil drain tubes were installed and streams were channelized, to move water off the fields as quickly as possible so that farmers could plant as early as possible. Impervious surfaces in cities and suburbs likewise speed water into storm drains that empty into channelized streams. The end result is unnaturally rapid delivery of water into the Upper Mississippi and more frequent small and moderate floods than in the past. In the arid western lands drained by the Missouri River, the problem is shortage of water; it is this phenomenon that led to the construction of the huge reservoirs to store floodwaters and use them for irrigating crops in the Dakotas, while also lowering flood crest levels in the downstream states of Nebraska, Iowa, and Missouri.
In all of the tributaries of the Mississippi, the floodplains have been leveed to various degrees, so there is less capacity to store or convey floods (as well as less fish and wildlife habitat), and the same volume of water in the rivers now causes higher floods than in the past. On tributaries with flood storage reservoirs, the heights of the moderate floods that occur can be controlled. On other tributaries, flood heights could be reduced by restoring some of the wetlands in the watersheds; constructing “green roofs” that incorporate vegetation to trap rainfall; adopting permeable paving; building stormwater detention basins in urban and suburban areas; and reconnecting some floodplains with their rivers. Between Clinton, Iowa, and the mouth of the Ohio River, 50 to 80% of the floodplain has been leveed and drained, primarily for dry-land agriculture. On the lower Mississippi River, from the Ohio River downstream and including the Delta, more than 90% has been leveed and drained. Ironically, levees in some critical areas back up water on other levees. In such areas, building levees higher is fruitless—it simply sets off a “levee race” that leaves no one better off. In the Delta, the additional weight of higher, thicker levees themselves can cause further compaction and subsidence of the underlying sediments.
The occasional great floods on the Mississippi are on a different scale than the more regular moderate floods. It takes exceptional amounts of rain and snowmelt occurring simultaneously in several or all of the major tributary basins of the Mississippi (the Missouri, upper Mississippi, Ohio, Arkansas, and Red Rivers) to produce an extreme flood, such as the one that occurred in 1927. That flood broke levees from Illinois south to the Gulf of Mexico, flooding an area equal in size to Massachusetts, Connecticut, New Hampshire, and Vermont combined, and forcing nearly a million people from their homes. With so much rain and snowmelt, wetlands, urban detention ponds, and even the flood control reservoirs are likely to fill up before the rains stop.
In order to protect New Orleans from such great floods, the Corps of Engineers plans to divert some floodwater upstream of the city. Floodwater would be diverted through both the Old River Control Structure and the Morganza floodway to the Atchafalaya River; and through the Bonnet Carré Spillway (30 miles upstream of New Orleans) into Lake Ponchartrain, which opens to the Gulf. All of these structures and operating plans are designed to safely convey a flood 11% greater in volume than the 1927 flood around and past New Orleans.
But what is the risk that an even greater flood might occur? How does one assess the risk of flooding and determine whether it makes more sense to move away than to rebuild?
The Corps of Engineers estimates flood frequencies based on existing river-gauging networks and specifies levee designs (placement, height, thickness) accordingly. The resulting estimates and flood protection designs are therefore based on hydrologic records that cover only one to two centuries, at most. Yet public officials may ask for levees and flood walls that provide protection against “100-year” or even “1,000-year” floods—time spans that are well beyond most existing records. Because floods, unlike trains, do not arrive on a schedule, these terms are better understood as estimates of probabilities. A 100-year levee is designed to protect against a flood that would occur, if averaged over a sufficiently long period (say 1,000 years), once in 100 years. This means that in any given year, the risk that this levee will fail is estimated, not guaranteed, to be 1% (1 in 100). If 99 years have passed since the last 100-year flood, the risk of flooding in year 100 is still 1%. In contrast to other natural hazards, such as earthquakes, the probability of occurrence does not increase with time since the last event. (Earthquakes that release strain that builds gradually along fault lines do have an increased probability of occurrence as time passes and strain increases.)
In essence, engineers assume that the climate in the future will be the same as in the recently observed past. This may be the only approach possible until scientists learn more about global and regional climate mechanisms and can make better predictions about precipitation, runoff, and river flows. However, the period of record can be greatly extended, thereby making estimates of the frequency of major floods much more accurate than by extrapolating from 200-year records of daily river levels. Sediment cores from the Mississippi River and the Gulf of Mexico record several episodes of “megafloods” along the Mississippi River during the past 10,000 years. These megafloods were equivalent to what today are regarded as 500-year or greater floods, but they recurred much more frequently during the flood-prone episodes recorded in the sediment cores. The two most recent episodes occurred at approximately 1000 BC and from about 1250 to 1450 AD. There is independent archeological evidence that floods during these episodes caused disruptions in the cultures of people living along the Mississippi, according to archeologist T. R. Kidder.
These flood episodes occurred much more recently than the recession of the last ice sheet, and therefore they were not caused by melting of the ice or by catastrophic failures of the glacial moraines that acted as natural dams for meltwater. They were most likely caused by periods of heavy rainfall over all or portions of the Mississippi Basin. Until more is known about climate mechanisms, it is prudent to assume that such megafloods will happen again. Thus, this possibility must be taken into account in designing flood protection for New Orleans, especially if public officials are serious about their expressed desire to protect the city against 1,000-year floods.
Building a “new” New Orleans
If New Orleans is to be protected against both hurricane-generated storm surges from the sea and flooding from the Mississippi River, are there alternative cost-effective approaches other than just building levees higher, diverting floods around New Orleans, and continuing the struggle to keep the Mississippi River from taking its preferred course to the sea? Yes, as people in other parts of the world have demonstrated.
The Romans used the natural and free supply of sediment from rivers to build up tidal lands in England (and probably also in the southern Netherlands) that they wished to use for agriculture. People living along the lower Humber River in England developed this to a high art in the 18th century, in a practice called “warping.” They had the same problems with subsidence as Louisiana, but they encouraged the sediment-laden Humber River to flood their levee districts (called “polders”) when the river was observed to be most turbid and therefore carrying its maximum sediment load. The river at this maximum stage was referred to as a “fat river” and inches of soil could be added in just one flood event. People of this time also recognized the benefits of marsh cordgrass, Spartina, in slowing the water flow, thereby encouraging sedimentation, and subsequently in anchoring the new deposits against resuspension by wind-generated waves or currents.
Could the same approach be taken in the Delta, in the new Atchafalaya lobe? Advocates for rebuilding New Orleans in its current location point to the 1,000-year+ levees and storm surge gates that the Dutch have built. But the Netherlands is one of the most densely populated countries in Europe, with 1,000 people per square mile, so the enormous cost of building such levees is proportional to the value of the dense infrastructure and human population there. The same is not true in Louisiana, where there are approximately 100 people per square mile, concentrated in relatively small parcels of the Delta. This low population density provides the luxury of using Delta lands as a buffer for the relatively small areas that must be protected.
However, the Dutch should be imitated in several regards. First, planners addressing the future of New Orleans should take a lesson from the long-term deliberate planning and project construction undertaken by the Dutch after their disastrous flood of 1953. These efforts have provided new lands and increased flood protection along their coasts and restored floodplains along the major rivers. Some of these projects are just now being realized, so the planning horizon was at least 50 years.
The old parts of New Orleans, including the French Quarter, were built on the natural levees created by the Mississippi River (red areas along the river in the figure), well above sea level. In contrast, much of the newer city lies below sea level (dark areas). Flooding of the city occurred when the storm surge from Hurricane Katrina entered Lakes Pontchartrain and Borgne and backed up the Gulf Intracoastal Waterway (GIWW), the Mississippi River Gulf Outlet (MRGO), and several industrial and drainage canals. The LAKE walls of the canals were either overtopped or failed in several places, allowing water to flood into the city.
(Courtesy of Center for the Study of Public Aspects of Hurricanes, as modified by Hayes, 2005.)
Planners focusing on New Orleans also would be wise to emulate Dutch efforts to understand and work with nature. Specifically, they should seek and adopt ways to speed the natural growth and increase the elevation of the new Atchafalaya lobe and to redirect sediment onto the Delta south of New Orleans to provide protection from storm waves and surges. A key question for the Federal Emergency Management Agency (FEMA), the FEMA equivalents at the state level, planners and zoning officials, banks and insurance companies, and the Corps of Engineers is whether it is more sustainable to rebuild the entire city and a higher levee system in the original locations or to build a “new” New Orleans somewhere else, perhaps on the Atchafalaya lobe.
Under this natural option, “old” New Orleans would remain a national historic and cultural treasure, and continue to be a tourist destination and convention city. Its highest grounds would continue to be protected by a series of strengthened levees and other flood-control measures. City planners and the government agencies (including FEMA) that provide funding for rebuilding must ensure that not all of the high ground is simply usurped for developments with the highest revenue return, such as convention centers, hotels, and casinos. The high ground also should include housing for the service workers and their families, so they are not consigned again to the lowest-lying, flood-prone areas. The flood-prone areas below sea level should be converted to parks and planted with flood-tolerant vegetation. If necessary, these areas would be allowed to flood temporarily during storms.
Work already is under way that might aid such rebuilding efforts and help protect the city during hurricanes. The Corps of Engineers, in its West Bay sediment diversion project, plans to redirect the Mississippi River sediment, which currently is lost to the deep waters of the Gulf, to the south of the city and use it to create, nourish, and maintain approximately 9,800 acres of marsh that will buffer storm waves and surges.
At the same time, the Corps, in consultation with state officials, should guide and accelerate sediment deposition in the new Atchafalaya lobe, under a 50- to 100-year plan to provide a permanent foundation for a new commercial and port city. If old New Orleans did not need to be maintained as a deepwater port, then more of the water and sediment in the Mississippi could be allowed to flow down the Atchafalaya, further accelerating the land-building. The new city could be developed in stages, much as the Dutch have gradually increased their polders. The port would have access to the Mississippi River via an existing lock (constructed in 1963) that connects the Atchafalaya and the Mississippi, just downstream of the Old River Control Structure.
This plan will no longer force the Mississippi River to go down a channel it “wants” to abandon. The shorter, steeper path to the sea via the Atchafalaya might require less dredging than the Mississippi route, because the current would tend to keep the channel scoured. Because the Mississippi route is now artificially long and much less steep, accumulating sediments must be constantly dredged, at substantial cost. Traditional river engineering techniques that maintain the capacity of the Atchafalaya to bypass floodwater that would otherwise inundate New Orleans also might be needed to maintain depths required for navigation. These techniques include bank stabilization with revetments and wing dikes that keep the main flow in the center of the channel where it will scour sediment.
The new city would have a life expectancy of about 1,000 years—at which time it would be an historic old city— before the Mississippi once again switched. The two-city option might prove less expensive than rebuilding the lowest parts of the old city, because the latter approach probably would require building flood gates in Lake Ponchartrain and new levees that are high enough and strong enough to withstand 500- or 1,000-year floods. In both scenarios, flood protection will need to be enhanced through a continual program of wetland restoration.
In evaluating these options, the Corps of Engineers should place greater emphasis on the 9,000 years of geological and archaeological data related to the recurrence of large floods along the Mississippi River. Shortly before the recent hurricanes hit the region, the Corps had completed a revised flood frequency analysis for the upper Mississippi, based solely on river gauge data from the past 100 to 200 years. Unless the Corps considers the prehistoric data, it probably will continue to underestimate the magnitude and frequency of large floods. If the Corps does take these data into account in determining how high levees need to be and what additional flood control works will be needed to prevent flooding in New Orleans and elsewhere, then the actual costs of the “traditional” approach are likely to be much higher than currently estimated. The higher costs will make the “working with nature” option even more attractive and economically feasible.
The Corps also should include in its assessments the gradual loss of storage capacity (due to sedimentation) in existing flood control reservoirs in the upstream Mississippi Basin, as well as the costs and benefits associated with proposed sediment bypass projects in these reservoirs. For example, the Corps undertook preliminary studies of a sediment-bypass project in the Lewis and Clark Reservoir on the upper Missouri River in South Dakota and Nebraska because the reservoir is predicted to completely fill with sediment by 2175, and most of its storage capacity will be lost well before then. By starting to bypass sediments within the next few years, the remaining water storage capacity could be prolonged, perhaps indefinitely. But studies showed that the costs exceeded the expected benefits. In these studies, however, the only benefits considered were the maintenance of water storage capacity and its beneficial uses, not the benefits of restoring the natural sediment supply to places as far downstream as the Delta. It is possible that the additional sediment would significantly accelerate foundation-building for “new” New Orleans and the rebuilding of protective wetlands for the old city. Over the long term, the diminishing capacities of such upstream storage reservoirs also will add to the attractiveness of more natural options, including bypassing sediments now being trapped in upstream reservoirs, utilizing the sediments downstream on floodplains and the Delta, and restoring flood conveyance capacity on floodplains that are now disconnected from their rivers by levees.
Action to capitalize on the natural option should begin immediately. The attention of the public and policymakers will be focused on New Orleans and the other Gulf cities for a few more months. The window of opportunity to plan a safer, more sustainable New Orleans, as well as better flood management policy for the Mississippi and its tributaries, is briefly open. Without action, a new New Orleans— a combination of an old city that retains many of its historic charms and a new city better suited to serve as a major international port—will go unrealized. And the people who would return to a New Orleans rebuilt as before, but with higher levees and certain other conventional flood control works, will remain unduly subject to the wrath of hurricanes and devastating floods. No one in the Big Easy should rest easy with this future.
Center for the Study of Public Health Aspects of Hurricanes, Louisiana State University, Baton Rouge (http://hurricane.lsu.edu/floodprediction/NewOrleans/New_Orlean_Elevation2.jpg).
Brian Hayes, “Natural and Unnatural Disasters,” American Scientist 93, no. 6 (2005): 496–499.
Robert H. Meade, ed., Contaminants in the Mississippi River, 1987-92 (U.S. Geological Survey Circular 1133, U.S. Government Printing Office, Washington, DC, 1995).
Torbjörn E. Törnqvist, Tristram R. Kidder, Whitney J. Autin, Klaas van der Borg, Arie F. M. de Jong, Cornelis J. W. Klerks, Els M. A. Snijders, Joep E. A. Storms, Remke L. van Dam, and Michael C. Wiemann,“A Revised Chronology for Mississippi River Subdeltas,” Science 273 (1996): 1693–1696.
Richard E. Sparks (firstname.lastname@example.org) is director of research at the National Great Rivers Research and Education Center in Godfrey, Illinois.