PERSPECTIVE

Reconstruction of New Orleans after Hurricane Katrina:

A research perspective
R. W. Kates*†, C. E. Colten‡, S. Laska§, and S. P. Leatherman¶
*33 Popple Point, Trenton, ME 04605; ‡Department of Geography and Anthropology, Louisiana State University,
Baton Rouge, LA 70803; §Center for Hazards Assessment, Response, and Technology, University of New Orleans,
New Orleans, LA 70148; and ¶International Hurricane Research Center, Florida International University, Miami, FL 33199

Edited by William C. Clark, Harvard University, Cambridge, MA, and approved August 29, 2006 (received for review July 18, 2006)

Four propositions drawn from 60 years of natural hazard and reconstruction research provide a comparative and historical perspec-
tive on the reconstruction of New Orleans after Hurricane Katrina. Decisions taken over its 288-year history that have made New Or-
leans so vulnerable to Katrina reflect a long-term pattern of societal response to hazard events—reducing consequences to relatively
frequent events, and increasing vulnerability to very large and rare events. Thus Katrina’s consequences for New Orleans were truly
catastrophic—accounting for most of the estimated 1,570 deaths of Louisiana residents and $40 –50 billion in monetary losses. A
comparative sequence and timing of recovery provides a calendar of historical experience against which to gauge progress in recon-
struction. Using this calendar, the emergency postdisaster period appears to be longer in duration than that of any other studied
disaster. The restoration period, the time taken to restore urban services for the smaller population, is in keeping with or ahead of
historical experience. The effort to reconstruct the physical environment and urban infrastructure is likely to take 8 –11 years. Con-
flicting policy goals for reconstruction of rapid recovery, safety, betterment, and equity are already evident. Actions taken demon-
strate the rush to rebuild the familiar in contrast to planning efforts that emphasize betterment. Because disasters tend to accelerate
existing economic, social, and political trends, the large losses in housing, population, and employment after Katrina are likely to
persist and, at best, only partly recover. However, the possibility of breaking free of this gloomy trajectory is feasible and has some
historical precedent.

F
our propositions drawn from 60 text of climate fluctuations (8), the re- tive that helps explain the vulnerability
years of research on natural haz- duction of mortality over time, as in of New Orleans to Katrina.
ards (1–5) and reconstruction death rates and, in some cases, in aggre-
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

after disasters (6) and 288 years gate deaths, has been found in longitu- New Orleans Flood and Hurricane History
of environmental history (7) provide dinal case studies, including drought in Before Katrina. For three centuries, New
perspective on the vulnerability of the the United States and Africa (8), tropi- Orleans sought to lessen the impacts of
city (parish) of New Orleans to Hurri- cal cyclones in Bangladesh (2), and its recurrent floods and hurricanes by
cane Katrina and its prospects for re- floods and hurricanes in the United providing marginal increases in safety.
construction. The first addresses the States (2, 4). A reduction in property However, in doing so, they laid the
paradox of the human ability to reduce damage is less clear because aggregate groundwork for the next catastrophic
over time the consequences of hazards property damages have risen along with failure. In its 288-year history, New Or-
while increasing their catastrophic po- increases in the population, material leans has had 27 major river or hurri-
tential. The second describes the se- wealth, and development in hazardous cane-induced disasters at a rate of one
quence and timing of reconstruction areas (9). At the same time, the individ- about every 11 years (14, 15). A pattern
phases found in a number of long-term ual losses for hazard victims have been of three responses runs through that
studies of reconstruction. The third reduced as the larger society absorbs a history. After each event, the city rebuilt
considers the conflicting goals and be- portion of their losses through disaster and often expanded, small differences in
haviors for reconstruction that arise in relief and insurance (2–4). elevation determined the location of the
recovery efforts after disaster. The Catastrophes from rare events can be well-to-do and the poor, and levees were
fourth proposition examines how disas- attributed to the sheer magnitude of rebuilt and often raised. River floods in
ters accelerate preexisting demographic, such events. Thus, the linkage between the years after the city’s founding in
economic, social, and political trends reducing consequences to relatively fre- 1718 did not deter its French founders
and lead to very different trajectories of quent events and increasing catastrophic from pressing forward with building the
recovery. potential needs to specify the mecha- colony’s capital at this strategic location,
nisms involved. For example, the suc- nor were they deterred in 1722 and 1723
Reduction in Consequences, Increase cessful prevention or rapid suppression when hurricanes destroyed the incipient
in Catastrophic Potential of forest fires leads to a buildup of com- city. Inequity in the location of neigh-
Over the long term, societies reduce bustible material that increases the cata- borhoods and in the distribution of
consequences to relatively frequent haz- strophic potential of fires that escape flooding burdens also appears early.
ard events (e.g., return periods of 100 rapid suppression (10, 11). Most rele-
years or less) through improved technol- vant to New Orleans is the so-called
ogy and social organization. However, ‘‘levee effect,’’ in which construction of Author contributions: R.W.K. designed research; R.W.K.,
C.E.C., S.L., and S.P.L. performed research; R.W.K., C.E.C.,
the reduction in risk to relatively fre- levees induces additional development S.L., and S.P.L. analyzed data; and R.W.K., C.E.C., S.L., and
quent events may increase vulnerability leading to much larger losses when the S.P.L. wrote the paper.
to major hazard events (e.g., return pe- levee is eventually overtopped (12). A The authors declare no conflict of interest.
riods of ⬎100 years) resulting in catas- more general statement of this proposi- This paper was submitted directly (Track II) to the PNAS
trophes characterized by large loss of tion is found in the safe development office.
life or property, major population loss, paradox in which increased safety in- †Towhom correspondence should be addressed. E-mail:
and out-migration, and even societal duces increased development leading to [email protected].
collapse. First noted in 1979 in the con- increased losses (13). It is this perspec- © 2006 by The National Academy of Sciences of the USA

www.pnas.org兾cgi兾doi兾10.1073兾pnas.0605726103 PNAS 兩 October 3, 2006 兩 vol. 103 兩 no. 40 兩 14653–14660

 When levees failed in 1816 and again in
1849, high water drove many of the
city’s poor, found in the lowest loca-
tions, from their homes for up to a
month (16, 17).
The response to riverine and hurri-
cane-induced floods in the Louisiana
colony was to build levees. By 1728, it
was mandatory for all land owners to do
so along their riparian frontage. Later,
levee heights were increased to 1 foot
higher than the last high-water stage.
Even as responsibility for levee con-
struction gradually shifted from land
owners, to the state, and ultimately to
the federal government, designed pro-
tection was based on the last storm.
Each increase in storm severity thus led
to a succession of catastrophic failures.
This has continued to this day with the
exception of 1927 when the great Missis-
sippi flood threatened New Orleans.
State and federal authorities responded
by dynamiting a breach in the levee that
flooded St. Bernard Parish down-river
to the detriment of its residents (18).
Further improvement of the levees, Fig. 1. Levee construction, subsequent development, and Katrina flood area in New Orleans, 1900 –2005.
aided by two major floodways that di-
vert high waters, has appeared to make
the city safer from river floods but not and new baby-boomer suburbs (Fig. 1). Its estimated pre-Katrina population of
from hurricanes. In a political culture This expansion followed the last round 437,186 (21) lived in a bowl, half located
of levee construction and reconstruction below sea level, between the natural
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

that often rewarded development and

patronage at the expense of safety and after Hurricane Betsy (1965) when levees of the Mississippi River and the
efficiency, completion of an effective ⬎300,000 residents were displaced and built levees (pierced by canals) along
hurricane protection system suffered 27,000 houses destroyed (14). The im- Lake Pontchartrain. In the 4 years pre-
from misplaced priorities (19). proved and much expanded levee system ceding Katrina, there were extensive and
Exemplifying the safe development led to a doubling of the protected area. repeated warnings from both scientists
paradox, improved drainage techniques Expected benefits from new develop- and the media that the ‘‘big one’’ would
enabled expanded development behind ment within that protected area were eventually hit the city. These included
the levees. In the 20th century, the city used to justify the project (7, 13). New specific concerns for the evacuation of
expanded in two major movements off massive drainage systems accompanied an estimated 130,000 residents without
the natural levee, across the Metairie the levees. When a spate of intense rain- vehicles, homebound, or in hospitals and
Ridge toward the vulnerable wetlands storms exposed the inadequacies of the in-care facilities (22–25).
near Lake Pontchartrain (Fig. 1). In the pumps and canals in the late 20th cen- Beginning on the morning of August
first period (1900–1950), early suburbs tury, the U.S. Army Corps of Engineers 29th, 2005, Katrina brought severe but
developed assisted by a municipal drain- participated in a major overhaul of these not catastrophic winds, record rainfalls
age system that helped dry out the elements of the system to protect the (up to 14 inches in 24 h), and stormwa-
mucky soils north of the city. These ar- development that crowded within the ter damage as the city’s pumping system
eas felt the impact of a severe hurricane levees. failed to keep up with the rain. Then,
in 1915 that damaged some 25,000 As in previous episodes of urban ex- within hours of the initial impact, major
buildings. A local levee district was cre- pansion, those with means moved to floodwalls along the 17th Street Canal,
ated in 1930 to enable residential devel- new suburbs, and the poor remained London Avenue Canal, and Inner Har-
opment. The state collaborated in the within the core city, generally in the bor Navigation Canal (Industrial Canal)
effort to fill in the lakeside from the low-lying locations (7, 19, 20). Conflicts failed, allowing water to surge into
natural beach and build a massive 9-foot between local and federal authorities ⬇80% of the city and essentially fill the
concrete seawall that in turn further en- over the final form of the hurricane bowl to depths ranging from 5 cm to
couraged the city’s lakeward expansion. protection system greatly delayed its 5 m (26). Days later, parts of New Or-
In general, the poor remained in the completion and exposed everyone to leans would be reflooded from intensive
city and often occupied low areas va- heightened risk. Thus, this most recent rains accompanying Hurricane Rita.
cated by those leaving for the newer round of levee construction and its con- As many as a million residents in the
suburbs. In 1947, hurricane storm surge sequences would lead to America’s metropolitan area may have responded
and waves overtopped lakefront levees greatest natural hazard catastrophe. to public calls for evacuation on August
and produced severe impacts to these 27th and 28th, leaving an estimated one-
suburban neighborhoods in Orleans and The Catastrophe of Katrina. In keeping quarter of New Orleans residents unable
Jefferson Parishes. with other disasters, this long history of or unwilling to leave. These residents
The second, post-World War II ex- marginal increases in safety that encour- took refuge in the Superdome, the Con-
pansion (1965–2000) accommodated aged new development made New Or- vention Center, in hospitals and nursing
population growth with public housing leans a catastrophe waiting to happen. homes, in upper stories of their homes,

14654 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0605726103 Kates et al.

 or on elevated highways, or died during
the week before full poststorm evacua-
tions could be completed. The evacu-
ated residents traveled or were moved
to other cities, and within a month, ref-
ugees from New Orleans could be found
in every state. Extensive media coverage
shared the failure of complete evacu-
ation, the plight of those remaining
in the city, and the subsequent out-
migration with a global audience. The
burden of these failures fell heaviest on
the African-American, poor, aged, and
infirm members of the population. Four
months after Katrina, the population
was estimated at 158,353, only 37% of
the pre-Katrina number (21).
The full death toll is still not known,
and out-of-state deaths in the month
after Katrina are still being reviewed.
But the estimated death toll for Louisi-
ana is 1,570, most of which were New
Orleans residents (27). As in all other
disasters, where the costs of hazards are
often hidden and underestimated and a Fig. 2. The sequence and timing of reconstruction after Katrina in New Orleans with actual experience
(solid lines) and sample indicators for the first year along a logarithmic time line of weeks after the disaster.
consistent set of methods and databases
The long-term projections (dashed lines) are based on an emergency period of 6 weeks, a restoration
do not exist, the true costs of Katrina in period of 45 weeks, and a 10-fold historical experience for reconstruction.
New Orleans will never be known (4, 5,
28). Limited estimates of damage to the
built environment, losses to the econ- draining of floodwaters. Before this pe- munity (40). For example, San Francisco
omy, and the costs of emergency and riod ends, the restoration period is was somewhat analogous to New Or-
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

reconstruction assistance are available, started, where the repairable essentials leans with its 1906 population of 400,000
along with some observations of social of urban life are restored. And well and its catastrophic losses (550 dead,
and environmental consequences (29– before this stage is over, replacement 220,000 homeless, 55% of its housing
35). Putting together these scattered reconstruction begins to provide the in- units destroyed, and 300,000 evacuees).
data, we estimate an aggregate mone- frastructure, housing, and jobs for the The 1906 emergency period lasted 4
tary loss of around $40–50 billion in destroyed city and predisaster popula- weeks, the restoration period 40 weeks,
Orleans Parish including direct property tion, followed often by a commemora- functional reconstruction 9 years, and
losses ($20–22 billion), still ongoing eco- tive or betterment reconstruction. the commemorative reconstruction even
nomic losses ($4–8 billion), and emer- Studies of earthquakes in Italy (37), longer (36). But even in this early study,
gency assistance ($16–20 billion). The Nicaragua (6), and the United States major differences were observed in the
human and social disruption has also (6, 38), and floods in the United States recovery between social, economic, and
been extraordinary given these losses,
(8, 39), have found that the second and ethnic groups. For example, a sample of
the out-migration, the trauma of experi-
third periods last approximately 10 times residents selected from city directories
ences, and the breakup of the commu-
the interval of the previous period. The showed that 1 year after the earthquake,
nity. Only the environmental losses have
use of this sequence and time scale of 74% of unskilled workers had disap-
been somewhat less than expected as
reconstruction can serve two important peared from the area compared with
high levels of toxic materials found in
the environment were primarily prod- purposes: to underscore the length of 40% of white-collar workers (40). Using
ucts of industrial development before time required for reconstruction and to this sequence and timing, how does the
Katrina (33). provide a calendar of historical experi- reconstruction of New Orleans compare
ence against which to gauge progress in with other large and rare disasters?
The Timing of Reconstruction the four periods.
Critics rightly note that the sequence Eleven Months of Restoration, Eight
A review of the limited set of long-term
of recovery processes can be uneven, to Eleven Years of Reconstruction
case studies of reconstruction after di-
sasters tells us that reconstruction takes that phases can overlap, and, most im- Fig. 2 shows a plot of the reconstruction
a long time. Reconstruction is part of a portantly, that different social groups, experience for 1 year after Katrina and
sequence of four identifiable postdisas- even within the same community, can projects future reconstruction activity by
ter periods: emergency, restoration, re- experience the sequence quite differ- using the four periods of historical expe-
construction, and commemorative or ently (5). Such differences can be partly rience. But applying the historical expe-
betterment reconstruction. These four captured by the initial length of the rience to New Orleans is complicated by
periods were first examined in a retro- emergency period, which serves as an the magnitude of the damage and fail-
spective study of San Francisco after the overall measure of both the magnitude ures in response, the massive forced out-
earthquake and fire of 1906 (36). The of damage and the response capacity of migration, and the external aid available
emergency period is characterized by different communities subject to the for restoration and some reconstruction.
search and rescue, emergency shelter same hazard event (38). These Because of the magnitude of damage
and feeding, the establishment of order, differences can also be evaluated sepa- and failures in response that character-
the clearing of major arteries, and the rately for varied groups within a com- ized Katrina, the emergency period ex-

Kates et al. PNAS 兩 October 3, 2006 兩 vol. 103 兩 no. 40 兩 14655

 tended over 6 weeks. This endpoint is leans plays in African-American politics, smaller city of 250,000 as a ‘‘sustainable,
determined by the ‘‘dewatering’’ of New culture, and education quickly overcame environmentally safe, socially equitable
Orleans, defined as the point when early commentary that perhaps New community with a vibrant economy. Its
flood waters were pumped and drained Orleans should not be rebuilt. But cities neighborhoods would be planned with
from the city. However, because of the and regions seeking to reconstruct after its citizens and connect to jobs and the
extraordinary damage and dispersal of a disaster seem to simultaneously pursue region. Each will preserve and celebrate
the population, an alternative length for goals to rapidly recover the familiar and the heritage of culture, landscape, and
the emergency period could be as long aspire to reconstruct in safer, better, architecture’’ (45).
as 14 weeks, when the end of emergency and sometimes more equitable ways. But 1 year after Katrina, the unified
shelter on December 3, 2005, is used as Conflict arises between groups or insti- neighborhood planning process, envi-
its conclusion. tutions and even individuals pursuing sioned by the city and the state, has barely
Restoration—repairing what is repair- these different goals because they can- begun. Some neighborhoods had begun
able in the infrastructure of urban life— not be given equal attention in time, their own planning process; other neigh-
began in the second week. But the resources, and values. In addition, in borhoods had professional assistance pro-
forced out-migration and low rate of accomplishing one goal, another may be vided separately by the mayor and the city
return complicates the calculation of less achievable. For example, compare council. It has taken 10 months for the
restoration. The result is that much re- achieving both rapidity of recovery and mayor, city council, and civic leaders to
pairable, but population-dependent, in- safer reconstruction, or betterment for agree on a unified planning process with
frastructure has not been restored or some segments of the population and professional assistance for 73 neighbor-
used. Most services for which there are equity for others. hoods and on the preparation of a city-
data available (electricity, gas, public For three centuries, New Orleans has wide infrastructure plan (46). Underlying
transportation, schools, hospitals, and had the recurrent opportunities found in the fits and starts in neighborhood plan-
food stores) are functioning at less than other disasters to rebuild the familiar in ning has been the reconstruction ap-
half of pre-Katrina capacity (34). safer, better, and more equitable ways. proach of various planning consultants to
The emergency period appears to be It essentially rebuilt the familiar, ex- rebuild the ‘‘high ground first—damaged
longer in duration than any other of the panded between disasters, and provided areas maybe’’ and its conflict with the
studied disasters and can be readily ex- marginal increases in safety but laid the most important equity issue—the rebuild-
plained by the evident failures in the groundwork for the next catastrophic ing of pre-Katrina African-American
initial evacuation and response. Using failure with major burdens falling on the neighborhoods.
the 6-week emergency period, the ex- poor. Now, 1 year after Katrina, the The major planning documents (43–
pected restoration period should be 60 planning effort and the actions taken to 45) reflect the contrasting planning
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

weeks in duration based on historical reconstruct New Orleans clearly reflect goals. As seen in other reconstructions,
experience; however, the actual restora- the pattern of conflicting reconstruction betterment emerges as the major plan-
tion period is 40 weeks when the near goals. Planning for reconstruction is di- ning goal. Perhaps to compensate for
restoration of the preexisting levee sys- vided between city, state, and federal the ‘‘high ground first—damaged areas
tem is used as the key indicator for its government, each assisted by outside maybe,’’ they do give prominence to is-
conclusion. This shorter restoration advisors and contractors, with distinctive sues of equity and citizen participation
period can be explained by the major but often overlapping responsibilities in the planning process. Safety, with one
commitment of funding, resources, and and intentions. At the federal level, ex- exception, is presented with few details
leadership to rebuild the levees and to tensive safety-related studies have been but calls on the federal government to
overcome the clear failures of past con- undertaken by the U.S. Army Corps of provide protection against category 5
struction. An alternative explanation Engineers (26) and a related planning hurricanes. Rapid recovery was not a
might be that because failures in emer- effort by the Federal Emergency Man- focus of the plans.
gency response extended the emergency agement Agency (FEMA) for the pro- In striking contrast to the reconstruc-
period 2 weeks beyond what the logistics duction of new 100-year flood elevation tion plans, the actual decisions and re-
of the hazard event might have re- maps taking into account Katrina’s building undertaken 10 months after
quired, (e.g., 4 weeks in San Francisco), flooding, the subsidence of benchmark Katrina—the so-called ‘‘facts on the
the shorter 40-week restoration period levels, and protection from rebuilt ground’’—clearly demonstrate the rush
better reflects the historical scale of levees (42). to rebuild the familiar found after all
experience. At both the state and city level, paral- disasters. Proposals for a building mora-
Ideas for reconstruction began to cir- lel and competing planning processes torium were almost universally rejected
culate even within the emergency pe- were launched by the Louisiana Recov- by residents. Federal government grants
riod, and serious planning efforts began ery Authority, appointed by the gover- to the state and payments of flood in-
10 weeks after Katrina while restoration nor, and the Bring New Orleans Back surance will now provide significant, but
was underway. Again, the historical ex- Commission, appointed by the mayor of not sufficient, funds for rebuilding (47).
perience argues for an extended period New Orleans. The initial plan was a More than 38,000 building permits have
of reconstruction, between 8 and 11 brief ‘‘starting point’’ plan developed by been issued for rebuilding to residents,
years, depending on the restoration pe- the Louisiana Recovery and Rebuilding ostensibly with ⬍50% damage (34).
riod used. Conference (43) and included features Many homeowners succeeded in having
that were similar to the city plans devel- their damage estimate reduced to below
Conflicting Goals and Differential oped by the Urban Land Institute with that key benchmark to enable rebuilding
Outcomes of Reconstruction proposals for category 5 flood protec- without elevation of the structure. New
The long history of urban experience tion; light rail, parks, and playgrounds; maps, to be used in testing eligibility for
(41) has few examples of cities failing to and selective neighborhood rebuilding flood insurance, have not been com-
rebuild in some fashion. For New Or- (44). A more detailed set of reconstruc- pleted because rebuilt levees could not
leans, the desire to overcome the fail- tion plans came from the Bring New be certified as protecting at the level
ures of the emergency response and to Orleans Back Commission, whose Urban previously protected. In their absence,
maintain the distinctive role New Or- Planning Committee envisioned a FEMA requires ground elevation of up

14656 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0605726103 Kates et al.

 to 3 feet (42). Given the depth of flood- building permits issued before the as they appeared, because although
ing experienced, this appears to be a FEMA advisory, as well as the FEMA 75% of the damaged-area population
modest requirement. advisory itself, make rebuilding possible was African-American and 29% poor,
with no (or minimal) elevation. As to areas with little or no f looding had
A Safer City? Further facilitating a return wind damage, New Orleans had adopted 46% African-American and 25% poor
to the familiar are the completed safety the model international building codes (58). A little over half of the f lood
improvements and further work under- before Katrina. Yet even such highly deaths were African-Americans, and
way. New Orleans will be somewhat touted buildings as the Superdome, used deaths occurred primarily among the
safer, but not so safe as it could be. It as a shelter of last resort, showed con- infirm and aged (27).
will surely be flooded again in the fu- siderable wind-induced exfoliation, and There were clearer racial and class
ture reflecting the threat of even greater expert inspections showed that many differences in the ability to cope with
or more threatening storms from the buildings failed because the code stan- the flood, to return, and to rebuild (59).
multidecadal cyclical period of high hur- dards were not sufficient, enforced, or Those with personal transport were able
ricane frequency (48). Furthermore, the applied to older buildings (52). to seek refuge with family, with friends,
intensities of these storms are probably In early June, a city plan was an- or in public shelters of their choosing
being exacerbated by global warming nounced for the complete evacuation out of the storm’s path. After the storm,
(49–51) and by sea level increase and of the city including households without many evacuees who had to rely on
continued subsidence of the land. Haz- cars, the disabled, and the infirm. It spe- emergency transport out of the city were
ard research offers five major types of cifically rejects ‘‘shelters of last resort’’ scattered to totally unfamiliar locations
adaptation that could be used to lessen within the city (53). But with hurricane with some family members taken to sep-
such risk (1, 2). Adaptive actions taken season underway, many aspects of the arate locations. Half of New Orleans
or planned to make New Orleans safer plan (e.g., assistance for elderly in get- residents lived in rental housing. Most
address three of these: rebuilding of the ting to the pick-up points, evacuations public housing remains boarded up, and
levees, a limited effort to make build- sites, use of trains) have not been de- four major housing developments will
ings flood and wind resistant, and prep- fined or tested through preparedness not be rebuilt. Temporary housing has
aration of a new evacuation plan. No drills. Although much discussed, no new been slowed by ‘‘not in my backyard’’
actions have been taken to change land action has been taken to change land objections.
use or even to restore wetlands. use or restore wetlands. All current The far-flung poor are also less able
The U.S. Army Corps of Engineers plans recommend creating parks, open to participate in postflood deliberations,
nearly fulfilled their promise to rebuild space, or restored wetlands in some of although many did return temporarily in
and strengthen the current levee system the lowest areas for amenity and beauti- sufficient numbers to reelect Mayor Ray
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

by June, the beginning of the 2006 hurri- fication, as an appropriate use for land Nagin. Middle class and especially pro-
cane season. This follows in the historical that is not rebuilt, and most importantly, fessional residents, both black and white,
tradition of rebuilding the levee system as internal stormwater and flood reten- are core members of the reconstruction
equal to or slightly higher than the most tion basins. But maps showing such planning committees. The new unified
recent flood. Thus, at a cost of $4.5 bil- parks and open space in badly flooded effort may provide new opportunity for
lion, the levees that failed have been re- neighborhoods were seen by many resi- more equitable participation in the plan-
built roughly to the nominal 5-m elevation dents as predecessors to the loss of their ning effort (46).
previously authorized but with improved property or neighborhoods.
earth materials, better anchored flood Many environmental scientists also A Better City? As is often the case in di-
walls, and armoring to permit the levees argued that wetlands in the delta below sasters, in the immediate aftermath of
to survive overtopping. Supplementing the city serve as a buffer zone that Katrina, hopeful boosters and politicians
these improvements are the installation of dampens the storm surge (54). Before proclaimed that reconstructed New
gates to close off three of the canals and Katrina, a $14 billion marsh restoration Orleans will be ‘‘bigger and better.’’ Al-
improved pumps and energy supplies for plan, known as Louisiana Coast 2050, though bigger is not likely, what consti-
management of interior stormwater or had been proposed but not funded (55). tutes better will be the focus of much
flooding. However, these features may not In the months after Katrina, strong sup- debate. Three major but overlapping
be sufficient. They will not be fully in port emerged for river diversion to the concepts of betterment have emerged.
place until 2007 and may still leave the west above the Bird’s Foot Delta (at the A new urbanism envisions a smaller but
city at risk from heavy rains (26). mouth of the river) permitting a release carefully planned city with revitalized
Protecting individual structures by of sediment to enhance the barrier older neighborhoods and restored por-
making them flood- and wind-resistant shoreline and thus protect the marshes tions of badly flooded neighborhoods
has been a major feature of modern that protect the city (56, 57). selected by residents. All are equipped
hazard research. The most common with new schools, parks, walks to stores
form of flood resistance is elevation A More Equitable City? Extreme events and services, and sustainable architec-
above some expected flood level. In reveal the extreme differences in the ture built along a backbone of light rail
New Orleans some individual home and way we live and die, cope, and rebuild. public transport bringing people to both
business owners had sought their own Historical reconstruction experiences, downtown and suburban jobs. The im-
protection by elevating structures on as well as New Orleans history, consis- proved city focuses on reversing the past
piers, often using the space below the tently report on inequitable patterns of by creating a new and advanced school
structure for open or enclosed garages social vulnerability and outcomes of system, an honest city government that
(e.g., Times-Picayune newspaper), addi- reconstruction. New Orleans was a is an efficient provider of services and
tional storage, or shady workspaces. predominantly black city (68%), and protection, a more multiracial and inte-
Nonetheless, there is considerable reluc- media coverage would easily suggest grated city that can reverse population
tance to elevate damaged buildings, that poor African-Americans were the loss, and a city safer from crime as well
because of appearance, cost, and the prime victims of the f lood, the botched as disaster. The investment city focuses
technical limits for elevating concrete evacuation, and the inadequate shelter. on the new economy, creatively using
slab homes. For most residences, city But the distinctions were not as sharp significant public and private funds to

Kates et al. PNAS 兩 October 3, 2006 兩 vol. 103 兩 no. 40 兩 14657

 rebuild and invest in previous areas of or comfort. In addition, some previous failures of response to Katrina by all
strength: tourism, culture, medicine, ed- land uses are diverted to commemora- levels of government. Another possibil-
ucation, and the ports. tive or betterment reconstruction. Pro- ity to reverse decline is the potential
Some modest progress can be found posed denser development could house impact of upgrading the hurricane pro-
for each of these visions, but none cur- a larger population, but proposed parks, tective system to a category 4 or 5,
rently makes a compelling case for their open space, and flood detention basins which could encourage a new round of
realization. The delayed neighborhood would remove significant acreage. Fu- major development while ironically in-
planning, seen as the heart of a new ur- ture improvements of existing levees creasing the future catastrophic poten-
banism, is finally underway. State super- and new internal levees would also re- tial. It is also possible to envision a
vision of schools has replaced the local quire additional area. Moreover, some fortuitous and rapid growth in some of
system. Charter schools, church schools, of the city may end up as brownfields the four economic and employment ar-
and newly restored and reorganized requiring long-term cleanup before eas identified by most planners: culture,
public schools will create a more diverse development. health, education, and port economy.
system, albeit with fewer pupils. A few However, the past is not necessarily a All four of these are growing rapidly in
new services have been provided, such prologue to the future. There are inter- many other cities fueled by larger na-
as public, free WiFi internet connections esting examples of reconstructions that tional and global trends. Culture econo-
in the central business district. A few have broken the trends of their predisas- mies are replacing retail and office
decisions have been made to restore or ter trajectories. In these cases, the city’s functions in many central cities. The
rebuild key hospitals and education reconstruction is assisted by some larger aging of the boomer generation will only
facilities. contextual changes that overcome the add to the rapid expansion of health
local situation. Three types of such con- education and provision of health care
Trajectories of Recovery textual changes have been found after services. Globalization will spur educa-
There is historical and comparative other disasters: external aid that is suffi- tional opportunities to maintain eco-
evidence that recovery after disaster ciently large to actually spur develop- nomic and technological competitiveness
generally follows the predisaster trajec- ment, spillover effects from larger and draw more international students. It
tory with the disaster even accelerating regions on a different trajectory, or also creates a demand for an expanded
previous trends. For cities with grow- spillover effects of the disaster itself. and specialized port economy.
ing population and economies, the di- The first two changes are illustrated in Finally, a possible but remote change
saster may accelerate that growth; the successful reconstruction of the de- is the type of political change seen
whereas for cities in economic and so- caying industrial city Tangshen, China, after other great disasters, particularly
cial decline, it may hasten decline (5, after the 1976 earthquake that killed in Latin America (64). In such a
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

6, 36 –39, 60). This is not encouraging over half a million people. The Chinese change in trajectory, New Orleans
for New Orleans, whose population government refused external assistance would benefit from a new national ini-
had declined by 31% from a peak 1960 from outside China but mobilized excep- tiative to address issues of race and
Census estimate of 627,525 to the esti- tional amounts of internal assistance to poverty deeply embedded in the soci-
mated July 2005 pre-Katrina popula- not only rebuild but to spur develop- ety and for which Katrina served as a
tion of 437,186 (21). Five months later, ment. In addition, the city also benefited metaphor and call to action. But an
the Census estimated the population at from the major economic changes and entrenched local and national political
158,253, a loss of 64% from the al- growth in the Chinese economy that culture does not bode well for such a
ready lower July 2005 number (21). coincided with the completion of its major trajectory change.
One future projection to 2008 foresees functional reconstruction (62). A more
a population of 279,000 or 60% of the recent example is the growth of Home- Sustaining New Orleans
pre-Katrina population (61). stead, FL, after it was destroyed by Hur- From the extensive research on natural
Economic projections are similar. One ricane Andrew in 1992 and suffered a hazards and the smaller body of re-
foresees two possible levels of economic decade as a depressed local economy. search on reconstruction after disasters,
growth for the metropolitan area, as mea- The rising economic tide of downtown we selected four key propositions to ex-
sured by employment. The metro New Miami and soaring real estate values plain the catastrophic vulnerability of
Orleans area recovers 41% of jobs by the have made the open farm lands of New Orleans to Katrina; to observe the
year 2008 in the moderate scenario. In the Homestead into the new suburbia (63). pace, process, and progress of recon-
high-growth scenario, 66% of the 190,000 Disasters themselves have strong spill- struction; and to consider its trajectory
jobs lost from pre-Katrina levels are re- over effects. Thus, the Great Plains for recovery. From this perspective, we
covered, but only if an unprecedented droughts of the 1890s (8), the Mississippi argue that the broad sequence of deci-
capacity for house construction can be floods of 1927 (18), and the dust bowl sions, made during New Orleans’ history
created (32). In a different projection for droughts of the 1930s (8) all led to mas- and resulting in an increased vulnerabil-
a five-parish region that has lost 40% of sive out-migrations and reshaping of the ity to Katrina, reflect a long-term pat-
its population and 13% of its pre-Katrina social and political landscape of the na- tern of societal response to hazard
employment, the optimistic economic re- tion. And in Latin America, the failures of events: reducing consequences to rela-
covery scenario projects only 73% of its the governments in power to rescue, shel- tively frequent events while increasing
population and 93% of its employment 5 ter, and rebuild the areas affected by the vulnerability to very large and rare
years after Katrina (33). Managua earthquake of 1972 and the events.
There will also be less space to sup- Mexico City earthquake of 1985 led to The sequence and timing of recovery
port reconstruction because historical profound political changes (64). is somewhat controversial, but it pro-
experience indicates that reconstruction These examples suggest contextual vides a calendar of comparative histori-
always requires more land, sometimes changes that might reverse the current cal experience against which to gauge
two to four times the previous area (6). limited prospects for New Orleans. The progress in the necessary phases of
To replace previous housing and infra- first, after the Tangshen experience, reconstruction. Using this calendar,
structure, rebuilding must conform to would be an extraordinary national ef- the emergency period appears to be
new standards of activity, construction, fort inspired by the desire to reverse the somewhat longer in duration than any

14658 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0605726103 Kates et al.

 other of the studied disasters, but the process that although rational, seems tives are needed to enable all home-
restoration period is in keeping with to threaten the recovery of some poor owners to rebuild to forthcoming
or ahead of historical experience. The and African-American neighborhoods. FEMA elevations. Experience from
effort to reconstruct the physical envi- Because disasters tend to accelerate developing countries would argue for a
ronment and urban infrastructure is existing economic, social, and political redundant system of neighborhood
likely to take a decade, and no com- trends, the trajectory for full recovery sanctuaries in the form of public build-
memorative reconstruction is in sight. (preexisting population, economy, and ings—schools or community centers—
The conf licting policy goals of rapid infrastructure) is not promising. The with upper f loors that are able to
recovery, safety, betterment, and eq- large losses in population and employ- withstand f looding, maintain power
uity and their relative strengths and ment after Katrina are an accelerated and water, and be converted easily
weaknesses largely ref lect experience continuation of its 45-year-long decline from everyday use to shelters. Better-
with large disasters in other places in population and economic growth, ment reconstruction could use the
and times. The actual decisions and now compounded by the major losses opportunity of reconstruction to make
rebuilding undertaken to date, the so- in housing stock. But the bleak pros- New Orleans a sustainable city, and
some locally produced plans are al-
called ‘‘facts on the ground,’’ clearly pect offered by the accelerated trends
ready available (65). Some equity
demonstrate the rush by the residents proposition can be yet altered by larger
would be achieved if all evacuees who
themselves to rebuild the familiar. contextual changes found in other
want to return have the help needed to
This trend is found after all disasters, reconstruction experiences. return by voucher-supported housing,
whereas those involved in the planning However, drawing upon the exten- reconstructed public housing, new de-
process do not share the same urgency. sive research perspective, even the velopments, or reconstituted neighbor-
The effort to reconstruct the failed smaller New Orleans of the future can hoods. All Americans have a stake in
levees to their existing height is also in achieve a better balance in its recon- the rapid, safe, better, and just recon-
keeping with historical action. Plan- struction efforts. Missing from rapid struction of New Orleans.
ning, as in all other studied reconstruc- recovery has been adequate attention
tions, strongly emphasizes betterment. to the needs of evacuees who lived in We thank Dr. Keqi Zhang and Mr. Quin
The considerable emphasis on equity rental housing, especially public hous- Robertson of Florida International University
in planning conferences and documents ing. The range of safety actions needs and Clifford Duplechin of Louisiana State
differs from previous experience and to be more redundant than simply re- University for preparing the graphics; E.
may ref lect a greater formal sensitivity storing or improving the surrounding Rovai for unpublished material on California
earthquakes reconstruction; J. K. Mitchell for
to minority concerns. The obvious in- levee system. Critical areas within the material on the Tangshen, China, reconstruc-
equities in risks from f looding and in city can be hardened by using second- tion; and the U.S. Army Corps of Engineers
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

the failures of evacuation result in an ary protection both by elevation and by Office of History for funding a portion of the
inherent conf lict with a reconstruction a set of inner levees. Financial incen- historical research.

1. White GF (1945) Human Adjustment to Floods: A 12. Burton I (1962) Types of Agricultural Occupance of tem (US Army Corps of Engineers, Washington,
Geographical Approach to the Flood Problems of Flood Plains in the United States (Univ of Chicago DC), Vol 1. Available at https:兾兾ipet.wes.army.
the United States (Univ of Chicago Press, Chicago). Press, Chicago). mil兾welcome91.htm.
2. Burton I, Kates RW, White GF (1993) The Envi- 13. Burby RJ (2006) Ann Am Acad Political Social Sci 27. Louisiana Department of Health and Hospitals
ronment as Hazard (Guilford, New York). 604:171–191. (2006) Reports of Missing and Deceased. Available
3. Mileti DS (1999) Disasters by Design: A Reassess- 14. US Army Corps of Engineers, New Orleans Dis- at www.dhh.louisiana.gov兾offices兾page.asp?ID⫽
ment of Natural Hazards in the United States trict (1972) History of Hurricane Occurrences along 192&Detail⫽5248. Accessed July 7, 2006.
(Henry, Washington, DC). Coastal Louisiana (US Army Corps of Engineers, 28. The H John Heinz III Center for Science, Eco-
4. Cutter SL ed (2000) American Hazardscapes: The New Orleans). nomics, and the Environment (2000) The Hidden
Regionalization of Hazards and Disasters (Henry, 15. Elliott DO (1932) The Improvement of the Lower Costs of Coastal Hazards (Island, Washington,
Washington, DC). Mississippi River for Flood Control and Navigation DC).
5. National Research Council, Committee on Disas- (US Waterways Experiment Station, Vicksburg, 29. Risk Management Solutions (2005) RMS Combines
ter Research in the Social Sciences: Future Chal- MS). Real-time Reconnaissance with Risk Models to Esti-
lenges and Opportunities (2006) Facing Hazards 16. Fenner E (1849) South Med Rep 1:56–62. mate Katrina Losses (news release). Available at
and Disasters: Understanding Human Dimensions 17. Forshay C (1849) South Med Rep 1:63–70. www.rms.com兾newspress兾pr㛭091905㛭hukatrina㛭
(Natl Acad Press, Washington, DC). 18. Barry JM (1997) Rising Tide: The Great Mississippi
lossmethodology.asp.
6. Haas JE, Kates RW, Bowden MJ eds (1977) Flood of 1927 and How It Changed America (Simon
30. AIR Worldwide (2005) AIR Worldwide Estimates
Reconstruction Following Disaster (MIT Press, & Schuster, New York).
Total Property Damage from Hurricane Katrina’s
Cambridge, MA). 19. Houck O (2006) Tulane Environ Law Rev 19:1–68.
Storm Surge and Flood at $44 Billion (press re-
7. Colten CE (2005) An Unnatural Metropolis: Wrest- 20. Lewis P (2003) New Orleans: The Making of an
lease). Available at www.iso.com兾press㛭releases兾
ing New Orleans from Nature (LSU Press, Baton Urban Landscape (Center for American Places,
2005兾09㛭29㛭05.html.
Rouge, LA). Santa Fe, NM).
8. Bowden MJ, Kates RW, Ray PA, Riebsame WE, 21. Frey WH, Singer A (2006) Katrina and Rita Im- 31. US Department of Housing and Urban Develop-
Warrick RA, Johnson DL, Gould HA, Wiener D pacts on Gulf Coast Populations: First Census Find- ment, Office of Policy Development and Research
(1981) in Climate and History: Studies in Past ings (Brookings Institution, Washington, DC). (2006) Current Housing Unit Damage Estimates:
Climates and Their Impact on Man, eds Wigley Available at www.brookings.edu兾metro兾pubs兾 Hurricanes Katrina, Rita, and Wilma. Available at
TML, Farmer G (Cambridge Univ Press, Cam- 20060607㛭hurricanes.pdf. www.dhs.gov兾interweb兾assetlibrary兾GulfCoast㛭
bridge, UK), pp 479–513. 22. Fischetti M (2001) Sci Am 285:77–85. HousingDamageEstimates㛭021206.pdf.
9. Cutter SL, Emrich C (2005) EOS Trans Am Geo- 23. McQuaid J, Mark S (June 23–27, 2002) The Times- 32. Inman News (March 27, 2006) Housing is Key to
phys Union 86:381, 388–389. Picayune. Economic Growth in New Orleans. Available at
10. Agee JK, Skinner CN (2005) For Ecol Manage 24. Laska S (2004) Nat Hazards Observer 29:4–6. www.inman.com兾member兾specialreports兾405hi兾
211:83–96. 25. Federal Emergency Management Agency (2004) story.aspx?ID⫽50713.
11. Folke C, Carpenter S, Elmqvist T, Gunderson L, Hurricane Pam Exercise Concludes (news release). 33. Interagency Performance Evaluation Task Force
Holling CS, Walker B, Bengtsson J, Berkes F, Available at www.fema.gov兾news兾newsrelease. (2006) Performance Evaluation of the New Orleans
Colding J, Danell K, et al. (2002) Resilience and fema?id⫽13051. and Southeast Louisiana Hurricane Protection Sys-
Sustainable Development (ICSU, Paris). Avail- 26. Interagency Performance Evaluation Task Force tem (US Army Corps of Engineers, Washington,
able at www.icsu.org兾Gestion兾img兾ICSU㛭DOC㛭 (2006) Performance Evaluation of the New Orleans DC), Vol 7. Available at https:兾兾ipet.wes.army.
DOWNLOAD兾64㛭DD㛭FILE㛭Vol3.pdf. and Southeast Louisiana Hurricane Protection Sys- mil兾welcome91.htm.

Kates et al. PNAS 兩 October 3, 2006 兩 vol. 103 兩 no. 40 兩 14659

 34. Liu A, Fellowes M, Mabanta M (2006) Special 45. Bring New Orleans Back Commission, Urban 57. National Research Council Committee on the
Edition of the Katrina Index: A One-Year Review of Planning Committee (2006) Action Plan for Restoration and Protection of Coastal Louisiana
Key Indicators of Recovery in Post-Storm New New Orleans: The New American City. Avail- (2006) Drawing Louisiana’s New Map: Addressing
Orleans (Brookings Institution, Washington, DC). able at www.bringneworleansback.org兾Portals兾 Land Loss in Coastal Louisiana (Natl Acad Press,
Available at www.brook.edu兾metro兾pubs兾 BringNewOrleansBack兾Resources兾Urban%20 Washington, DC).
2006㛭KatrinaIndex.pdf. Planning%20Action%20Plan%20Final%20Report. 58. Logan JR (2006) The Impact of Katrina: Race and
35. Center for Philanthropy, Indiana University pdf. Class in Storm-Damaged Neighborhoods (Brown
(2006) Gulf Coast Hurricane Relief Donations. 46. City of New Orleans (2006) Mayor, City Council, Univ, Providence RI). Available at www.s4.
Available at www.philanthropy.iupui.edu兾 Civic Leaders Agree on ‘‘Unified New Orleans brown.edu兾Katrina兾report.pdf.
Hurricane㛭Katrina.html. Neighborhood Plan’’ (press release). Available 59. Pastor M, Bullard RD, Boyce JK, Fothergill A,
36. Kates RW, Pijawka D (1977) in Reconstruction at www.rockfound.org兾AboutUs兾Foundation Morello-Frosch R, Wright B (2006) In the Wake
Following Disaster, eds Haas JE, Kates RW, Bow- Announcement兾167. of the Storm: Environment, Disaster, and Race
den MJ (MIT Press, Cambridge, MA), pp 1–23. 47. Louisiana Recovery Authority (2006) Summary of
37. Geipel R (1991) Long-Term Consequences of Di- After Katrina (Russell Sage Foundation, New
the Proposed Action Plan Amendment (Amendment York). Available at www.russellsage.org兾news兾
sasters: The Reconstruction of Friuli, Italy, in Its 1) for FY 2006 CDBG Disaster Recovery Funds.
International Context, 1976–1988 (Springer, New 060515.528528.
Available at www.lra.louisiana.gov兾assets兾roadhome兾 60. Bates FL, Peacock WG (1993) Living Conditions,
York). HousingSummary050206.pdf.
38. Rovai E (1994) Yearb Assoc Pac Coast Geogr Disasters and Development: An Approach to Cross-
48. Goldenberg SB, Landsea CW, Mestas-Nuñez CW,
56:49–74. Cultural Comparisons (Univ of Georgia Press,
Gray WM (2001) Science 293:474–479.
39. Stephen S (1983) Disasters 7:194–201. Athens).
49. Emanuel K (2005) Nature 436:686–688.
40. Bowden M, Pijawka D, Roboff GS, Gelman KJ, 61. McCarthy K, Peterson DJ, Sastry N, Pollard M
50. Webster PJ, Holland GJ, Curry JA, Chang HR
Amaral D (1977) in Reconstruction Following Di- (2005) The Repopulation of New Orleans After
(2005) Science 309:1844–1846.
saster, eds Haas JE, Kates RW, Bowden MJ (MIT Hurricane Katrina (Rand, Santa Monica, CA).
51. Mann ME, Emanuel KA (2006) EOS 87:233–244.
Press, Cambridge, MA), pp 69–145. Available at www.rand.org兾pubs兾technical㛭
52. National Institute of Standards and Technology
41. Chandler T, Fox G (1974) 3000 Years of Urban reports兾2006兾RAND㛭TR369.pdf.
Growth (Academic, New York). (2006) Performance of Physical Structures In Hur-
ricane Katrina and Hurricane Rita: A Reconnais- 62. Mitchell JK (2004) in New Zealand Recovery
42. Federal Emergency Management Agency (2006) Symposium Proceedings, ed Norman S (Ministry
Flood Recovery Guidance: Advisory, Base Flood El- sance Report (National Institute of Standards and
Technology, Gaithersburg, MD). of Civil Defence and Emergency Manage-
evations for Orleans Parish, Louisiana. Available at ment, Wellington, New Zealand), pp
www.fema.gov兾pdf兾hazard兾f lood兾recoverydata兾 53. City of New Orleans (2006) 2006 Emergency
Preparedness Plan. Available at www.cityofno. 47– 68.
orleans㛭parish04–12-06.pdf.
com兾Portals兾Portal46兾portal.aspx?portal⫽46& 63. Iuspa-Abbott P (July 4, 2003) South Florida Busi-
43. American Institute of Architects (2006) Starting
tabid⫽38. ness Journal. Available at http:兾兾southflorida.
Point: Report from the Louisiana Recovery and Re-
54. Bohannon J, Enserink M (2005) Science bizjournals.com兾southf lorida兾stories兾2003兾07兾
building Conference. Available at http:兾兾lrrc.aia.org兾
SiteObjects兾files兾lrrc㛭startingpoint㛭crf.pdf. 309:1808–1809. 07兾story1.html.
44. Urban Land Institute (2005) Executive Summary of 55. Louisiana Coastal Wetlands Conservation and 64. Drury AC, Olson RS (1998) J Conting Crisis
Key Recommendations—A Strategy for Rebuilding Restoration Task Force and the Wetlands Con- Manage 6:151–161.
Downloaded from https://www.pnas.org by 84.232.175.114 on November 23, 2023 from IP address 84.232.175.114.

New Orleans, Louisiana, November 12–18, 2005 (Ur- servation and Restoration Authority (1998) Coast 65. Bring New Orleans Back Commission, Sus-
ban Land Institute, Washington, DC). Available at 2050: Toward a Sustainable Coastal Louisiana tainability Subcommittee (2005) Sustainable
http:兾兾www.uli.org兾Content兾NavigationMenu兾 (Louisiana Department of Natural Resources, Architecture and New Orleans Katrina Re-
ProgramsServices兾AdvisoryServices兾KatrinaPanel兾 Baton Rouge). covery www.bringneworleansback.org兾Portals兾
exec㛭summary.pdf. 56. Stokstad E (2005) Science 310:1264–1266. BringNewOrleansBack兾portal.aspx?tabid⫽101.

14660 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0605726103 Kates et al.