crisis management in katrina’s immediate aftermath...
TRANSCRIPT
1
Crisis Management in Katrina’s Immediate Aftermath: Lessons for a National Response to a Super
Disaster
Arjen Boin, Leiden University & Louisiana State University
Jim Richardson, Louisiana State University
25 August 2015
Katrina was one of the worst disasters in US history. It is also the best-studied disaster in history.
Many lessons were learned by a host of academics and professional inquiries. But these lessons are
quite contradictory when viewed in concert; more importantly, some of these lessons are plain wrong.
Most importantly, we have missed crucial lessons that would help to strengthen our resilience in the
face of future super disasters. We still do not understand why certain things went wrong in the
response to Katrina and why some things went surprisingly well. Building on an extensive review of
reports and inquiries, and drawing on insights from crisis and disaster management studies, this
paper identifies critical factors that determine the success and failures of a societal response to super
disasters. The paper offers a combined focus on the local level (New Orleans), the state level
(Louisiana) and the federal level.
PRESENTED AT THE KATRINA@10 CONFERENCE, LOUISIANA STATE UNIVERSITY
FRIDAY, 28 AUGUST 2015
DRAFT PAPER – NOT FOR CITATION – COMMENTS WELCOME
Contact author:
Arjen Boin, Ph.D.
Department of Political Science
Leiden University
The Netherlands
2
What went wrong? Just about everything [..] the basic narrative is becoming clearer: hesitancy,
bureaucratic rivalries, failures of leadership from City Hall to the White House and epically bad luck
(Newsweek, 12/9/2005, p. 45)
1. Introduction: The Shame of Katrina
In the summer of 2005, Hurricane Katrina produced a “mega-disaster”, the largest in US history.
Katrina devastated a major, troubled city and the entire region around it. We might say that Katrina
was a natural disaster, caused by a killer hurricane. But it was also, and perhaps foremost, a man-
made disaster. Katrina could wreak havoc because of a woefully inadequate protective structure. The
failure to protect the Crescent City has been scrutinized, and rightly so.
But it was the response to the event that turned Katrina into a national disaster. Public and political
assessments declared government failure: government officials failed, the president failed, FEMA
failed, Brownie failed, the system failed. One report summarized its findings in terms of “a litany of
mistakes, misjudgments, lapses, and absurdities all cascading together” (rep-x). The response was
considered so bad that “Katrina” has become shorthand for shameful performance of governments
anywhere.
This is a powerful indictment. It is an indictment that does not take into account the circumstances
under which government agencies were asked to respond. This was the first time that government
agencies were confronted with a major US city that had almost completely flooded. There were no
plans or routines for such a “difficult disaster” (Brinkley 249).
It is also an ill-informed accusation of failure, for it does not take into account the many things that
went right before, during and after Katrina. To be sure, the response to Katrina was not as good as
one might have hoped or expected. Mistakes were made. Some actors failed, some failed miserably.
But it is easily forgotten that many things actually went remarkably well, especially given the
circumstances. For instance, the evacuation of New Orleans was a clear if underappreciated success.
The evacuation of a large city is very hard and rarely done (and rarely done well).1 The timely
evacuation of New Orleans likely prevented a disaster of biblical proportions (the Hurricane Pam
scenario, which we will discuss below, predicted over 50,000 deaths for a Katrina-like disaster).
After the city flooded, a flotilla of heroic rescuers – both volunteers and professionals – saved many
lives in their search and rescue efforts (Derthick, 2007). The Federal government sent more resources
to Louisiana in the first two weeks after Hurricane Katrina than it had sent to Florida for all of the
previous year’s hurricanes combined (FR44). The southern states and the federal government
worked together to provide medical assistance to the injured, distraught and displaced.
1 When Hurricane Rita forced the evacuation of Houston (the fourth city of the US), the traffic was a nightmare
and over a hundred people died in traffic accidents (CB269).
3
If we take a long-term perspective, the joint performance of local, state and federal government
aimed at bringing the Gulf Coast area back can also be qualified as highly effective. The federal
government spent over $110 billion to assist the states and in particular New Orleans (and this is to
say nothing about the $40 billion in private insurance spending). While these efforts were, of course,
not flawless, it is simply remarkable to walk around today in a bustling New Orleans.
This paper aims to do two things. First, it offers a more balanced approach in assessing the response
to Katrina. We note what went wrong, but we also look at what went right. Unlike other
assessments, we make use of an explicit framework that guides our assessment and analysis of the
response. We describe the quality of the response in terms of four functions that citizens expect
from their government before, during and after a disaster: preparation and detection, sense-making,
and coordination. Second, we identify the factors that affected the capacity of government to fulfill
these functions. We conclude with the “real” lessons of Katrina. We look for lessons that are specific,
evidence based, and replicable.2
2. Studying a Mega-disaster: How to assess success and failure
This brings us to the central puzzle of this paper: how can we fairly assess the performance of a
cobbled-together response network that must perform under dire conditions? If government failed
massively, as the critics assert, how can we account for the successes? If government agencies
performed well on some tasks, how come they failed at other tasks for which they had more time
and resources?
The assessment of crisis and disaster management is a subjective affair (Bovens and ‘t Hart, 1996).
One reason is that we often do not have all the information needed to pass judgment. Another
reason is that we do not have a thorough understanding of the causal relations between actions and
outcomes. What, exactly, did Mayor Giuliani do in the aftermath of the 9/11 attacks that made him a
celebrated crisis manager? And why was President Bush’s crisis management after 9/11 assessed in a
much more favorable light than his crisis management efforts in response to Katrina? Why was the
English response to the London bombing attacks widely viewed as exemplary?
But perhaps the most important reason why the responses to crises and disasters can be viewed so
differently is that we do not have a widely shared normative framework that specifies what we may
expect from our leaders and government organizations in times of crisis. Our approach combines a
set of clearly explicated expectations of government performance with a keen understanding of the
“impossible” constraints that crises and disasters tend to impose on such performance. This
approach is based on years of crisis research (Rosenthal, Charles and ‘t Hart, 1989; Rosenthal, Boin
and Comfort, 2001; Boin et al 2005, 2008, 2013). Building on empirical findings, we assert that if
government authorities engage in a selected set of crisis management tasks, they are more likely to
minimize the effects of a large-scale crisis or disaster. These strategic tasks are:
2 We draw on official reports, academic articles, books and media accounts.
4
Preparation and detection: work together to prepare for known and unknown threats;
organize to detect credible signals of emerging threats.
Sense-making: organize to collect, interpret and disseminate critical information that
enables a shared understanding of unfolding crisis events.
Coordination and critical decision-making: enable vertical and horizontal collaboration
between all parties of the response network; identify and make critical decisions that
must be made at the strategic level.
Meaning-making: formulate and communicate a convincing and enabling narrative that
explains what has happened and what is being done to minimize the consequences of
the crisis.
Enhance resilience: bring together all the resources needed to restore the stricken
society in such a way that it emerges stronger and ready for the next disaster. This task
complements a focus on the short term (stop the crisis) with a long-term perspective
(restore trust, rebuild).
But an assessment of government performance during crisis must take account of the conditions in
which strategic crisis management plays out. We must consider what is actually possible in the case
of a super disaster. In a super-disaster such as Katrina, there are clear limits to what crisis
management can achieve. A super disaster not only creates huge challenges, it also renders coping
capacities useless. A super disaster tends to happen in vulnerable places, where we find weak
institutions and a less affluent population. All this was true for Katrina: the floods destroyed
prepositioned goods and hit one of the most underprivileged regions of the United States, one not
blessed with the strongest public institutions.
We therefore define successful crisis management in terms of “doing the best that can be expected
given the circumstances”. Viewed from this perspective, we can speak of a successful crisis response
if government writ large makes an honest attempt to fulfill these strategic tasks in a legitimate way.
We can now start out to formulate explicit expectations with regard to the joint performance of
local, state and federal government organizations before and during Hurricane Katrina. Given what
was known (or could have realistically been known) before the disaster and given the immediate
impact of this super disaster, we formulate the following expectations with regard to the strategic
tasks outlined above:
Detection and preparation: governments at all levels must take the threat seriously and
act as best as they can to prepare and protect the population. Finding: government
agencies did take this hurricane seriously and prepared as well as could be expected.
5
Sense-making: governments at all levels should cooperate to share and analyze
information; they should share their emerging picture(s) of the situation. Finding:
isolated successes, no system to put it all together. The City of New Orleans did not
provide the one thing we might have expected even from a weak city: local sense-making.
Coordination and decision-making: governments at all levels should work together to
make sure that critical tasks are being performed by those who are best placed to
perform them (and decisions should be made at the appropriate levels). Finding: The
federal disaster structure proved very complex and not geared towards managing a
catastrophic event.
Meaning-making: governments at all levels should cooperate to formulate and
communicate a shared frame to survivors and the general public. Finding: the politics of
crisis management played out in a vicious and dysfunctional way, undermining and
reversing the positive dynamics that mark the initial phases of most super disasters.
Resilience: governments at all levels should collaborate to facilitate the revival of local
communities and regional economies. Finding: the Gulf Coast in general is back and
that’s simply amazing given the devastation caused by Katrina.3
This paper discusses the first three of these executive functions (the other two will be discussed in
the book manuscript under construction from which this paper is drawn).
3. Preparation and Detection: Why didn’t they see it coming?
After a crisis,, a complex and dynamic event is often boiled down to a simple and evocative narrative.
In the narrative, the causes are clear if not self-evident. After-action reports and political inquiries
uncover evidence that someone had forewarned but was subsequently ignored. Whether we talk
about Pearl Harbor, the explosion of space shuttle Challenger, the attacks of 9/11, or the financial
crisis – these events, in hindsight, might appear knowable and thus preventable (even though this
would require no errors in analysis and judgement).
This always prompts the question: Why did they not see it coming? If it was foreseeable and thus
preventable, someone clearly did not do his job. Or worse, someone gambled that known risks would
not materialize, thus endangering the lives of citizens.
The post-Katrina narrative fits this mold perfectly. Collective wisdom has it that Katrina was
foreseeable and it was foreseen; this disaster should have been prevented; it’s a shame this
happened. Those in charge were incompetent.
3 Certain areas in New Orleans and St. Bernard Parish are still below pre-Katrina levels. Bringing these areas
back is an ongoing struggle.
6
The problem with this narrative is that the very characteristics of a so-called “Black Swan” event tend
to outwit collective imagination. They are long-tail possibilities – events that never happened before
but are theoretically possible. Few people can envision a future they have not seen before. As the risks
are not calculable and the consequences unimaginable to many, they tend to be ignored. But when
the risk actually materializes, it is clear that it could have been known.
So why did they not see it coming? Once we take the character of Black Swans into account, we open
ourselves to the possibility of different conclusions. We can then entertain the possibility that the
authorities could not foresee this disaster, but nevertheless were as prepared as one might expect
them to be for a “normal” disaster. It allows for the conclusion that not much could have been done
better – even if it was not nearly enough for the super disaster that Katrina turned out to be.
Surprise!
On Wednesday August 31st, America woke up to a super-disaster. Devastation had set in on Monday,
of course, but the initial message that dominated the national news was that New Orleans had
“dodged the bullet.” It slowly became clear on Tuesday that no bullet had been dodged. The New
York Times reported on Wednesday that “the scope of the catastrophe caught New Orleans by
surprise.” In truth, the entire nation had been caught by surprise.
But that’s not the way most people remember it. Months after Katrina, no one talked about surprise.
“Perhaps the single most important question the Select Committee has struggled to answer is why
the federal response did not adequately anticipate the consequences of Katrina striking New Orleans
and, prior to landfall” (rep-137). Or as Chairman Tom Davis stated during the Select Committee
Hearing, December 14, 2005: “That’s probably the most painful thing about Katrina, and the tragic
loss of life: the foreseeability of it all” (rep-80). The House report asserts that “this crisis was not only
predictable, it was predicted [..] government failed because it did not learn from past experiences”
(rep-xi).
But research tells us that prediction of these super-disasters is impossible (Clarke, 1999; Tetlock,
2005). The distinction between known risks and unknown risks helps to understand why this is the
case. Known risks are threats that materialize with some sort of regularity and play out in more or
less similar ways. Examples include river floods, earthquakes, epidemics, financial crises and
hurricanes. Yet, the exact occurrence of known risks is hard to predict. So while we know that
hurricanes are a likely occurrence in the summer, we have no idea which state (if any) will be
affected this summer and when.
Unknown risks are in a different category altogether. They are unique events for which no statistical
base rate exists. Examples include the Mad Cow disease, the 9/11 attacks, the Iceland ash cloud
(2010) and the Fukushima disaster. It is simply impossible to predict such Black Swans, because they
occur so rarely (or may have never occurred). Importantly, it is impossible to adequately plan for
such unique events (Clarke, 1999). Some of these unknown risks (or Black Swans) flow from known
risks: “normal” crises that play out in unsuspected and devastating ways. The recent financial crisis
was such a crisis. The flooding of New Orleans after hurricane Katrina was another Black Swan.
7
In the case of a Black Swan, the best we can hope for is timely detection: recognizing that something
out of the ordinary is happening. But early recognition turns out to be difficult as well, both for
individuals and organizations. People are simply not very well equipped to make sense of
uncertainties; they tend to make use of mental shortcuts (psychologists call these “heuristics”) that
are not very effective when it comes to sniffing out an unfamiliar threat (they do better when it
comes to known threats).4
We might expect organizations to compensate for the cognitive limitations of their employees. Alas,
that is not what seems to happen (Turner, 1978; Perrow, 1984, 1986; Catino, 2013). It turns out that
organizations fail often and easily in puzzling together pieces of information that in hindsight turn out
to be critical. The research identifies many factors (institutionalized mindsets, organizational
cultures, limited capacities), but the upshot is very simple: organizations are not very good at
detecting the unimaginable. Organizations may become well versed at collecting and analyzing the
information that has been shown to matter; they don’t know to recognize the signals of impending
anomalies.
This means that the recognition of a Black Swan is not a fair evaluation standard for organizational
performance. If we cannot demand the impossible (predicting a disaster) and if we take seriously the
research lessons that point out how hard early detection of a Black Swan is, the following question
emerges: If a disaster cannot be predicted and a Black Swan is hard to recognize, what may we fairly
expect from those government agencies that operate in the domain of crisis and disaster
management?
We propose three “fair” expectations to help us assess how well government organizations handled
the pre-response to Katrina; we offer these expectations in terms of three evaluative questions:
1. Did the responsible organizations and/or figures of authority willfully ignore clear and
unambiguous signals of an impending disaster?
2. Did they take adequate preparatory measures in light of what could have reasonably been
foreseen or expected?
3. Did they share available information with those in the path of the disaster and warn people
of what they knew?
Did the responsible organizations and/or figures of authority willfully ignore clear and unambiguous
signals of an impending disaster?
A pervasive story line explains how local, state and federal authorities ignored clear signals that
Katrina would produce a super-disaster for New Orleans. In media accounts, Congressional reports,
4 See Kahneman’s (2011) Thinking fast thinking slow on the inaccuracies of these heuristics. See the fascinating
debate between Kahneman and Klein (2009) on the differences between recognizing known and unknown
threats.
8
and academic studies it is presented as a “case closed” story. The evidence can indeed seem
irrefutable and overwhelming, the assessment damning as a result.
A typical story about “missed signals” would include a reference to a study by University of New
Orleans researchers who predicted in 2004 that even a moderate hurricane would devastate the city
(CB13). It would mention the “plenty of general warnings” in the years previously (Brinkley, 2006:
14). It would cite LSU researcher Van Heerden (2006; Brinkley, 2006: 81) who had predicted
thousands and thousands of casualties in case of a hurricane. And it would bring out the smoking
gun: Hurricane Pam.
Hurricane Pam never really happened. Pam was the center piece of a fictitious disaster scenario
designed by a Baton Rouge disaster management consultancy (IEM). In the summer of 2004, local
and state authorities in Louisiana participated with FEMA in the scenario exercise. The aim was to
assess the state of preparedness for a hurricane strike. To be sure, the scenario bore some
resemblances to Katrina. Pam was a big hurricane, just like Katrina. It even followed the same path.
So when it became known after Katrina that officials had practiced on a similar scenario, the
reactions were understandably incredulous. How could the authorities not have been prepared?
Critics charged that public authorities, especially FEMA, had not learned or did not act on the lessons
learned from the Pam scenario.
That’s too simple, however.
First of all, Pam was not a simulation but a planning exercise. This is more than a semantic difference.
Officials typically use interactive disaster simulations to practice decision-making, cooperation and
coordination under stress. They can then test their plans, skills and capacities. But that is not the
purpose that Hurricane Pam served.
The purpose of the exercise was to help officials develop joint response plans for a catastrophic
hurricane in Louisiana (rep-81). The scenario exercise was “designed to be the first step toward
producing a comprehensive hurricane response plan [..] to provide general guidance, a sort of “to do
list” for state and localities (rep-82). Many actors worked for days on the scenario, thinking through
their aims and what they would need to accomplish their aims. They did not “practice.” They were
trying to improve their plans.
The aim of the exercise thus was to explore what issues of preparation and cooperation might
emerge during a hurricane. The participants in the workshop focused on issues ranging from search
and rescue and temporary sheltering to unwatering, debris removal, and medical care. These issues
were then compared to the available plans (identifying weaknesses and strengths). As a bonus,
participants got to know their strategic partners a little better (CB16).
Second, Hurricane Pam as a disaster was very different than Hurricane Katrina. In fact, the Pam
scenario was considered a bit unrealistic at the time (Brinkley, 94). Hurricane Pam did not breach any
levees; there were only the usual overtoppings (CB15). Yet, in the scenario, about 175,000 people
9
were injured, 200,000 became sick, and more than 60,000 were killed (rep-81). Looking back, it is
hard to see how such a scenario could possibly cause so many casualties.
Some issues were not covered at all. For instance, the scenario did not feature law-and-order or
communication problems (CB21). The CEO of IEM, Ms Beriwal, later explained that while issues
related to security and communications were on the agenda, the development of a plan to
coordinate the displacement of school children took precedence (rep-82).
Third, many lessons were learned. Numerous action plans ranging from debris removal, to sheltering,
to search and rescue were developed. For example, state transportation officials took the lessons
learned from the Pam exercise and previous hurricanes and revised the state’s contraflow plan (rep-
82). The contingency plan for the medical component was almost complete when Katrina made
landfall. Officials said although the plan was not yet finalized, it proved invaluable to the response
effort (rep-83). The celebrated performance of Louisiana Department of Wildlife and Fisheries
officials was at least partially due to the Pam exercise: rescue teams applied a model developed in
the Hurricane Pam exercise, bringing hurricane victims to high ground, where they were supposed to
receive food, water, medical attention, and transport to shelters (SS8).
Although participants may have failed to generate a comprehensive, integrated, and actionable plan
in time for Hurricane Katrina, these workshops did have positive impacts. To quote one official: “the
workshops and planning process— knowledge of inter-jurisdictional relationships and capabilities,
identification of issues, and rudimentary concepts for handling the consequences—have been
beneficial to all involved in the hurricane response”. (FR25)
To be sure, the exercise was not perfect. The search and rescue group developed a transportation
plan for retrieving and evacuating stranded residents, but this plan apparently did not work (or was
never applied) (rep-82). Moreover, FEMA officials made promises they did or could not keep (CB20).
But if Pam is to be taken seriously as a predictor, as critics claim, we can only conclude that Katrina
was handled much better than the fictitious Pam. If anything, Pam was an elementary to providing
the cases of effective, improvised response.
Did anyone actually foresee the Katrina disaster?
The responsible authorities did not foresee the flood disaster that followed hurricane Katrina, as
many critics have correctly pointed out. At the same time, there is little evidence that anyone saw
this disaster coming (and offered a specific warning to that effect). As Hurricane Katrina approached
Louisiana, Governor Blanco was understandably concerned “that many people would play a familiar
game of ‘hurricane roulette’—tempting fate and staying home in a gamble that this storm would be
no worse than the last one they weathered in their home” (FR26). One of the most vocal hindsight
critics of the “relative ignorance” of government people, Douglas Brinkley, did not evacuate himself.
He booked a hotel room in downtown New Orleans to ride out the storm.
Experts did not do much better. What politicians and public administrators expected and feared most
was wind damage. Few imagined that the levees would break and the city would be flooded (CB32).
10
To be sure, there was plenty of talk about the possible topping of levees (Brinkley, 2006: 98). The
NHC issued advisories that warned the levees in New Orleans could be overtopped by Lake
Pontchartrain and that significant destruction would likely be experienced far away from the
hurricane’s center. (FR28). It was reported that NHC director Mayfield had cautioned the levees
would be breached, but no such warning was issued. “What I indicated in my briefings to emergency
managers and to the media was the possibility that some levees in the greater New Orleans area
could be overtopped, depending on the details of Katrina’s track and intensity,” Mayfield later
explained (rep-70).
Intriguingly, a DHS “fast analysis report” predicted early Sunday evening that the storm would breach
the levees and leave at least 100,000 poverty stricken people stranded on roof tops; the doomsday
scenario was based on a computer simulation, but it came very late in the game. The White House
got a copy at 2 am (CB122-3; 277, 278). Around that time, hurricane watchers who studied Katrina up
close in a plane just began to understand how bad it was when they flew into the massive hurricane
on Sunday morning 4am (73).
Did authorities take adequate preparatory measures in light of what could have reasonably been
foreseen or expected?
Even though authorities did not recognize a super-disaster in the making, they did take Hurricane
Katrina very seriously. There was no downplaying, or ignoring, the potential effects of the hurricane.
Quite the contrary: authorities warned that a very dangerous storm was coming and they prepared
accordingly.
On Wednesday August 23, Katrina was threatening Florida. Louisiana was not on the list of states
likely to be affected by Katrina, but Governor Blanco activated the emergency center, activated the
National Guard and canceled a trip (CB98). The EOC in Baton Rouge conducted communications
checks with all the state agencies and parishes (rep-64).
On Friday, Louisiana Governor Kathleen Blanco and Mississippi Governor Haley Barbour declared
states of emergency for their respective States (FR24). Louisiana had the EOC up and running with its
full staff complement by Friday afternoon. A direct hit on Louisiana, specifically New Orleans, had
become increasingly likely. That afternoon at 17.00, Governor Blanco and Mayor Nagin held a press
conference in New Orleans at which they urged New Orleanians to evacuate. On Sunday, Nagin
declared a mandatory evacuation (which had never happened before). Hours before landfall, the city
was ready for a hurricane: all institutions – universities, the Audubon Zoo, the Aquarium, the D-Day
Museum (Brinkley, 2006: 41) – had been closed down. Those who had not evacuated, were filing into
the Superdome.
In anticipation of the storm, emergency responders were standing by to begin search and rescue as
soon as it was safe to proceed (FR35). Louisiana’s Department of Fish and Wildlife (LDWF)
coordinated with the Louisiana National Guard to get boats placed on trailers and pre-positioned at
Jackson Barracks in New Orleans (rep-64). LDWF had 200 agents with boats in a ring around New
Orleans (Brinkley, 2006: 121, 116).
11
The Adjutant General of the Louisiana National Guard, Landreneau, had mobilized 2000 soldiers
(Brinkley, 2006: 116).5 An additional 120 soldiers were dispersed through New Orleans (120) in
Jefferson Parish, St Bernard and Plaquemines (Brinkley, 2006: 121). The Louisiana National Guard
deployed liaison officers to the thirteen southernmost parishes projected to suffer the greatest
impact from the storm (FR26). “Louisiana had a plan”, said Marsha Evans, then head of the Red Cross
(Brinkley, 2006: 116).
After the storm, the local preparation was severely criticized. But New Orleans did organize.
Unfortunately, the city prepared for a hurricane, not for a flooded city.
Contrary to what became common knowledge after the storm, federal authorities (FEMA in
particular) were “leaning forward” in the preparatory phase. On Friday, FEMA supervisor Leo Bosner
recognized Katrina as a “nightmare scenario” (CB99). He collected information about the storm and
sent a report around Saturday morning, 5:30 am, with a clear warning that this might be big (CB100).
FEMA director Brown understood the potential of Katrina (CB101; Brinkley, 2006: 37). During a noon
videoconference with FEMA regional staff and EOCs, Brown said that “My gut hurts on this one [..]
We need to take this one very, very seriously”. He admonished his staff to “lean forward as much as
possible [as] this is our chance to really show what we can do” (CB 101-2; FR28).6 In that same
conference, Joe Hagin, the White House deputy chief of staff, listened in. He thought the planning
was “in good shape” (CB 102).
Saturday evening, FEMA’s William Lokey arrived in Baton Rouge and was appointed Federal
Coordinating Officer (FCO). As the senior Federal official in charge of supporting the State of
Louisiana, he began coordinating efforts with the Louisiana Office of Homeland Security and
Emergency Preparedness (FR27).
After the storm, some people would claim that they did not think that FEMA was doing enough in the
face of “the Big One.” There is some truth to this observation, as FEMA was operating on routine
mode – routine for a large-scale hurricane, that is. FEMA had been lauded for precisely similar
actions that same summer and the summer before. In the face of Katrina, the Agency positioned an
unprecedented number of resources in affected areas. In fact, FEMA’s efforts far exceeded any
previous operation in the agency’s history (rep-59; Brinkley, 2006: 131).
On a Sunday noon video conference, Louisiana’s William Doriant called Katrina “catastrophic”. His
colleague, Jeff Smith (the director of the State’s Department of Emergency Preparedness) “said the
state was happy enough with the supplies FEMA had en route to the region” (CB 114). In response,
5 It is interesting to note that Landreneau wore two hats, as head of both the National Guard and the Louisiana
Office of Homeland Security and Emergency Preparedness (LOHSEP) (rep-67).
6 Brown apparently also talked off-line to the governors, “to make sure the governors weren’t going to tell me
something privately that maybe they didn’t want to share publicly, and they seemed satisfied at that point with
the help they were getting.” (FR29)
12
FEMA director Brown urged his people to “jam up the supply chain” – “just keep jamming those lines
full as much as you can with commodities” (CB114).
Excerpts Sunday noon video conference:
COLONEL SMITH: …I can tell you that our Governor is very concerned about the potential loss of life
here with our citizens, and she is very appreciative of the federal resources that have come into the
state and the willingness to give us everything you’ve got, because, again, we’re very concerned with
this.
COLONEL (BILL) DORIANT: The Emergency Operations Center is at a Level 1, which is the highest state
of readiness. We’ve currently got 11 parishes with evacuations, and climbing. (….) Evacuations are
underway currently. We’re planning for a catastrophic event, which we have been planning for,
thanks to the help of FEMA, when we did the Hurricane Pam exercises. So we’re way ahead of the
game here. (HR432)
COLONEL DORIANT (cont.): We’re also taking a look at our sheltering needs, long-term sheltering
needs, looking at sites to start bringing in the temporary housing. So we’re not only fighting the
current battle, managing expectations here with our local parishes, but we are also working with
FEMA and our other federal partners to have the most effective response and recovery that we
possibly can during this time. (…) I think that at this point in time our coordination is as good as it can
be. (HR433)
COLONEL SMITH: (inaudible) Resources that are en route, and it looks like those resources that are
en route are going to – to be a good first shot. Naturally, once we get into this thing, you know, neck
deep here, unfortunately, or deeper, I’m sure that things are going to come up that maybe some of
even our best planners hadn’t even thought about. So I think flexibility is going to be the key. (…) We
appreciate your comments. I think they were to lean as far, far as you possibly can, you know,
without falling, and your people here are doing that. And that’s the type of attitude we need in an
event like this. (HR433)
MIKE BROWN: Any questions? (Missing) on the commodities that I want to see that supply chain
jammed up just as much as possible. I mean, I want stuff (missing) than we need. Just keep jamming
those lines full as much as you can with commodities. (HR434)
Did authorities share available information with those in the path of the disaster and warn people of
what they knew?
After Katrina, many people claimed that they were not warned. Bu the authorities did make a
concerted effort. Initially, Katrina did not pose a threat to Louisiana and Mississippi. But by Friday, it
was clear Katrina was headed towards New Orleans. When the authorities understood that Katrina
was targeting their cities, they began their usual warning activities. In addition, there were grassroot
13
efforts. In an effort to reach as many citizens as possible, Governor Blanco and her staff contacted
clergy throughout Saturday night and early Sunday morning to ask them to urge their parishioners to
evacuate immediately. (FR26)
That does not mean people acted upon those warnings. As is often the case in the Gulf Coast states,
many people seemed unconcerned about the impending storm (FR25). Saturday night witnessed the
usual hurricane parties (Brinkley, 2006: 60). While most people left, thousands stayed behind.
Mayor Nagin initially appeared a bit more cautious. On Saturday afternoon, Mayor Ray Nagin hosted
a press conference, during which he recommended evacuations of Algiers, the Lower Ninth Ward,
and low-lying areas of the City. (FR26). He did not order a mandatory evacuation until Sunday
morning (after much pressure from Blanco, Bush and Mayfield).
The director of the National Hurricane Center, Max Mayfield, had been alarmed by the complacency
in New Orleans. He made “frantic calls” to both Blanco and Nagin (Brinkley, 2006: 57). In a Sunday
morning conference call with Bush, Chertoff, Brown and other senior staff members, Mayfield “laid it
on thick and straight” (but he spoke of “minimal flooding” and never mentioned the possibility that
the hurricane would breach the city’s levees, Brinkley, 2006: 112-3). Sunday morning (10:11 am), the
NHC released Advisory 23, which was formulated to provide a “horrible glimpse of the future”
(Brinkley, 2006: 79). It warned of a “most powerful hurricane with unprecedented strength”,
predicting “incredible human suffering”. According to Brinkley (2006: 80-81), “there was very little
science in it, only savage imagery”. The text was so incendiary that the staff of NBC’s Nightly News
worried that it was not real.
By Sunday, officials were crystal clear in their messaging. Nagin called Katrina “a once-in-a-lifetime
event” (NYT 29/8). Joseph Fein, NO (NYT 29/8) said it was “the most threatening we have seen”.
Blanco (NYT, 29/8): This storm is bigger than anything we have dealt with before.” President Bush in
a televised speech urged people to leave.
4. Understanding the Unimaginable: Why Collective Sense-making Failed
One of the most perplexing features of this disaster was how uninformed public authorities appeared
during that first week: it almost took two days to understand that New Orleans had flooded,
authorities did not seem to know the difference between the Superdome and the Convention Center,
they had no clue about the deplorable situation in many hospitals, the desperation in outlying areas,
or the actual level of violence in the city. Even though much information was available, government
organizations at all levels could not put the pieces together and thus did not understand what was
happening on the ground.
In their detailed description of the federal response in response to Katrina, Cooper and Block
(2006:xiv) concluded that “pertinent, accurate and real-time information flowed in great waves
through government agencies.” They were right: key officials at all levels possessed detailed
information about the flooding and the situation in the city. In theory, information should flow
through the system. In practice, I did not quite work that way.
14
In that first week, few people had a full picture of the devastation in the south. It took DHS nearly 36
hours to conclude that the levees had been breached in New Orleans and that the city was under
water (Brinkley, 2006: 133). It took even longer for government agencies to discover the near-
complete devastation in St Bernard Parish (this did apparently not happen until 47 Canadian
Mounties arrived on Wednesday, Brinkley, 2006: 181). Authorities were also slow to grasp the
desperation of survivors in and near the Superdome, the Convention Center, or on the elevated
highways. Television viewers across the world found out where survivors were heading and
congregating, and how badly they suffered, before policymakers did. The latter never knew about
the plight of the hospitals or fully comprehended the actual level of violence in the city (which was
much lower than widely believed).
The official inquiries found all government organizations wanting. At DHS “early situational
awareness was poor, a problem that should have been corrected following identical damage
assessment challenges during Hurricane Andrew” (rep224). The military faced similar problems: the
biggest challenge for Northern Command was “gaining and maintaining situational awareness as to
the catastrophic disaster.”188 The same could be said for the authorities in New Orleans, the state
authorities in Baton Rouge, and officials working at FEMA and the President’s office: they were all
struggling to understand what was going on “down there”.
How is that possible? Before we answer that question, we should note that in hindsight it is always
possible to reconstruct what happened when (and who knew, or could have known, about it). This
“hindsight bias” is known to affect the assessment of crisis researchers; as a result, they tend to
underestimate the sense-making challenge as experienced by decision-makers.
This explanation of complete organizational failure is both too simple and implausible. It somehow
suggests that these organizations, with all their investments in sense-making capacities, massively
failed at an essential facet of their job description. More importantly, perhaps, it cannot explain why
some aspects of the response went really well. Such a binary performance – abject failure v.
expected success – does not adequately describe the problem.
Crises are characterized by deep uncertainty. And as they evolve, “there are frequently additional
negative surprises” (Leonard and Howitt, 2009). When a situation is new and baffling, crisis managers
will try to get information, analyze it, establish a picture of the situation, share that picture, and
update it as new information becomes available (Weick, 1995). The challenge is to bring information
from many different sources together and create a joint picture of the situation.
Institutionalized information processing mechanisms no longer suffice, as they are geared towards
routine processes.7 Normal processes of information collection and communication often fall apart;
organizational chains fragment. Key information (How many people died? How many wounded?) is
simply hard to come by. It takes precious time to survey a disaster site, collect critical information,
7 Media are the exception: their routine processes are particularly well geared towards crisis sense-making
(Goidel and Miller, 2009; Miller et al 2014).
15
summarize it in an understandable way, and get that information to the right person in the chain.
Moreover, for first responders, the collection of this information is rarely a priority.
A better starting point is to expect collective sense-making failures in the initial phase of a complex
and catastrophic crisis like Katrina. Many crises – ranging from 9/11 to Katrina, from Sandy to Boston
– have shown how hard it is to make sense of a fast-moving threat that defies plans and challenges
experience. Virtually every inquiry report on large-scale crises and disasters asks why it takes so long
for authorities to figure out what is going on. All these reports show that relevant information was
indeed available, but that authorities nevertheless did not understand how bad the situation was.
This is the rule rather than the exception.
We must therefore begin by noting that the “accurate information flowing in great waves through
government agencies” is not always recognizable as such. It must be culled from a tsunami of
irrelevant, ambiguous or false information. Moreover, on some critical events there may be no
information available whatsoever (no news is not always good news). For instance, authorities did
not learn about the plight of the hospitals and nursing rooms until well after the disaster.
It is fair to expect improvement over time, however. That did not happen. The authorities did not
manage to “up their game” during that first week. This explains the ongoing confusion at all levels of
government during that first week.
The literature offers three types of explanation that we will explore:
Limited capacity to collect and verify information: technical and institutional factors
A failure of collective imagination (Clarke): the inability to understand and appreciate
unsuspected events
A breakdown in the inter-organizational communication chain
Capacity to collect: Communication breakdown and weak institutions put to the test
During that first week, sense-making problems began at the local level. In New Orleans, there was no
collective sense-making effort. There were individual nodes, but the information from those nodes
was not brought together. There were two reasons for this: a breakdown of communication means
and the malfunctioning of local institutions.
One of the simplest explanations for the slow sense-making is found in the failure of communication
means. As happens so often in times of disaster, people quickly discovered that they could not
communicate through the normal channels. There was “massive inoperability” due to failed,
destroyed, or incompatible communications systems (rep-163). Louisiana State Senator Robert
Barham, chairman of the State Senate's homeland security committee, summed up the situation in
the state: “People could not communicate. It got to the point that people were literally writing
messages on paper, putting them in bottles and dropping them from helicopters to other people on
the ground.” (FR37)
16
The lack of communication means made it hard to gather basic facts. And when people had critical
information, it was hard to communicate that information. Reporters of the Times-Picayune, at their
first editorial meeting on Monday evening began to realize that the city was becoming inundated
(Brinkley, 2006: 190).8 But it was hard to move around the city and get an idea of how far-spread the
flooding was. A New York Times photographer had an illustrated story by Tuesday morning (he had
hitched a ride on a FEMA helicopter surveying the damage). But the reporter could not get the story
to the NYT offices in New York (Brinkley, 2006: 235). So even in New Orleans (where the rising water
was everywhere), it took professional people whose job it was to make sense of unfolding events a
relatively long time to understand the scope of the unfolding disaster.
The breakdown of local institutions did not help. With its long history of hurricanes, one would
expect New Orleans to “have it down” and be prepared to play a key role in the response. One would
expect that local experts with extensive knowledge of the city would be available to provide situation
reports to assist those coming from the outside to help. One would expect a local hub, where
information from around the city is brought together.
While the New Orleans EOC could have been the information hub during Katrina, at least two
problems prevented it from becoming so. One problem was that the ultimate head of the local
response organization, Mayor Nagin, was not present at the EOC. Nagin chose to ride out the storm
at the Hyatt Regency on Poydres Street. He did remain active and communicated widely, but he
created an additional information hub that was not integrated with the EOC. For instance, he
received very specific information about breaking levees (Brinkley, 2006: 147, 243), but it is not clear
if and how that information made its way into the EOC.
A second, more serious problem was the performance of the NOPD. One would expect them to
deliver a stream of situational reports, being the “eyes and ears” of the city administration. They
never became an information node. By all accounts, the NOPD simply fell apart (Brinkley, 2006: 205).
“As an institution… the New Orleans Police Department disintegrated with the first drop of
floodwater (rep-246).
It is important to note that the NOPD did not fail in its entirety. Many officers performed heroically.
And we should take note just how hard the force was hit. Many officers were stuck in flooded areas
and had lost their homes.
Yet, the NOPD did not play a key role in assembling and transmitting situational information. Its
history was marred by incidents of police brutality. (HR43) As a result, residents had little trust in the
NOPD (49). This distrust was reinforced by the behavior of “dozens” of NOPD officers, some engaging
in violent (Brinkley, 2006: 475) or racist behavior (384-5), others who sank to looting (203; 361ff).
While it is hard to assess just how many officers misbehaved, the reputation of the NOPD was deeply
hurt when MSNBC provided live coverage of looters, including police officers, ransacking a local Wal-
8 The paper moved its headquarters to Houma soon thereafter.
17
Mart in New Orleans (rep-247). The result was organizational isolation. Other organizations in the
response network quickly shied away from the NOPD, which hurt its information position.9
To make things worse, Police Chief Eddie Compass became a source of wild but false rumors (224,
282, 365). It is clear in hindsight that Compass simply did not have reliable information and was
repeating what he thought he knew. His capacities may well have been impaired by sleep deprivation
and stress (Brinkley, 2006: 388; see also Spike Lee movie When the Levees Broke). To his credit, he
never left his post. But the rumors he helped validate would have a terrible impact on the bigger
picture that was emerging nation-wide.
A flywheel was set in motion. The rumor mill quickly created a picture of mayhem. Across the nation,
and indeed the world, it appeared that Katrina had turned New Orleans into a living hell. Many of the
stories that gave rise to this story later turned out to be exaggerated or false. In reality, there hardly
was a problem of law and order. It was the perception of a law-and-order problem that created a
tense situation.
The picture was reinforced by political and administrative actions, which, in turn, were shaped by the
emerging (but inaccurate) picture. Mayor Ragin declared martial law on Tuesday night (Brinkley,
369). New Orleanians began to wear arms to protect themselves. Some hired private security firms.
The governor issued a “shoot-to-kill” order. Law enforcement agencies operated in battle gear. All
this reinforced the emerging national picture, which, in turn, prompted local reactions.
The failure of local institutions to provide adequate information about the evolving situation was
especially jarring because federal institutions had no boots on the ground during those first days.
FEMA had only a few people in the city who had limited means of contacting colleagues outside. The
Corps of Engineers had no idea about the situation (it did not even have a representative on the
ground). When the Corps began to report to DHS on Monday afternoon, the crucial information was
buried in other stuff (Brinkley, 2006: 142).
Failure of collective imagination
While local capacities to gather and analyze critical information were impaired, there was accurate
information available. It was readily available, if only the right people would have understood what
they had. But this often a critical shortcoming: people cannot appreciate the importance of
information that deals with events that they are unprepared to entertain. To understand a situation,
one needs information. But information never materializes in handy packages, delivered to the door.
Information needs to be gathered, interpreted, analyzed, verified and tested. To do this, one needs
to have an idea what the information should look like (what is it that we want information about?).
In simple (and perhaps simplified) terms, what is needed is a degree of imagination.
There are quite a few reasons why many people “miss” essential information that is right in front of
them (cf. Clarke, 1999). One reason is that people are not very good at thinking about things they do
9 FEMA refused to work with NOPD says Riley (Brinkley, 2006: 509).
18
not expect to happen. If you do not have at least a vague idea, however rudimentary, about the
situation one seeks to understand, it becomes hard to actively seek information. Psychological
research has shown time and again that you don’t see what you don’t expect to see and you are
more likely to see what you expect to see.
A puzzling limitation of our mind, as Kahneman (2011: 14) explains, is “our excessive confidence in
what we believe we know, and our apparent inability to acknowledge the full extent of our ignorance
and the uncertainty of the world we live in.” When people think they know what is going on, or
cannot conceive of what might be going on, they are very likely to believe arguments that appear to
support what they think they know, even when those arguments are unsound (Kahneman, 2011: 45).
“Contrary to the rules of science who advise testing hypotheses by trying to refute them, people seek
data that are likely to be compatible with the beliefs they currently hold” (Kahneman 2011: 81).
Another reason is that people find it very hard to cope with uncertainty (Kahneman, 2011). The brain
likes to eliminate uncertainties by suggesting a variety of shortcuts or what psychologists call
“heuristics.” These heuristics work quite well, as they help us to make sense of situations we do not
understand right away. Unfortunately, these heuristics do not necessarily lead to a correct picture of
the situation.
Moreover, the brain’s sense-making capacities quickly deteriorate under stress. When we get tired,
many routine tasks become much harder to accomplish (Coates, 2012). It becomes harder to make
accurate assessments, switch tasks, and gauge risks; we are more likely to make selfish choices, use
improper language and make superficial judgments.
Yet another reason is that group interaction often introduces additional barriers to a correct and
shared picture of the situation. The particular way in which a group shares and discusses pieces of
information (even the order in which it is presented), the relations between group members (trust,
dislike etc), the setting in which a group convenes, and the behavior of the group leader all affect the
outcome of group deliberations. Seemingly small factors may have disproportionate effects on a
group’s sense-making abilities.
All those factors were at play that first week. When the levees gave way that early Monday morning,
few people in the various crisis centers had a mental picture of a submerged city or anticipated the
magnitude of the destruction (Brinkley, 2006: 172). Even those who should know best, because they
were actually there and knew the city well, initially failed to grasp the extent of the disaster. For
example, the iconic Robinelle of WWL radio, one of the few stations working throughout the ordeal,
initially told his listeners that New Orleans had dodged the bullet (Brinkley, 2006: 134).
“When that storm came by, a lot of people said we dodged a bullet. When that storm came
through at first, people said, Whew. There was a sense of relaxation, and that’s what I was
referring to. And I, myself, thought we had dodged a bullet. You know why? Because I was
listening to people, probably over the [airwaves], say, the bullet has been dodged. And that
was what I was referring to [..] There was a sense of relaxation in the moment, a critical
moment.” (SA12)
19
Editor Jim Amoss of the Times-Picayune was “chipper” Monday morning (Brinkley, 2006: 181). At
2pm, a few editors of the paper went out to survey the city. While the disaster was rapidly unfolding,
they did not immediately recognize it until they biked around the city (Brinkley 2006: 186). It was
only Monday evening that Amoss noticed water (downtown) (185). Fox reporters drove all over
town, but still missed the big story (200). Professional people looking for news to report, who were in
the middle of it, did not realize they were witnessing a super-disaster unfolding.
It is no wonder, then, that those in crisis centers far removed from the scene needed more time to
grasp the enormity of events. We know that information about breached levees did, in fact, reach
the various centers. Yet, in Monday’s videoconference at 11 am, in which many centers were
represented, levee breaches were not a topic of conversation. The Corps downplayed alarming
information about the levees, Mayfield – the Cassandra of previous days – commented that the
levees were unlikely to have been breached.10 Blanco also discounted a report of breaching (Brinkley,
2006: 138). On that first Monday, FEMA director Brown was one of the few federal officials who
seemed to have a grasp of the scale of the disaster: he urged his staff to resist the tendency to think
they had dodged the bullet (138).
Another official who did have a fairly comprehensive view of the situation early on was Marty
Bahamonde, FEMA’s lone representative in New Orleans. Bahamonde took in the scale of the
disaster during two brief helicopter rides (the first one at Monday, 5:15pm), taking pictures (Brinkley,
2006: 144-5). When Bahamonde called the FEMA Public Affairs office (where he worked), his
assessment of the situation was questioned (147).
The failure of imagination likely impaired initial sense-making at the White House. At 2:20 p.m. on
August 29, a HSOC report stated some Louisiana parishes had eight to 10 feet of water and an
unspecified number of Louisiana and Mississippi residents were stranded in flooded areas. In a 6:00
p.m. HSOC report, the White House was advised extensive flooding in New Orleans could take
months to reverse through the dewatering process. Even when, at 12:02 a.m. on August 30, the
White House received the Bahamonde spot report in which the quartermile breach in the levee near
the 17th Street Canal was reported (rep-142), White House officials did not believe they had
confirmation of any levee breaches. An earlier Army Corps of Engineers’ report had not confirmed
them and because “this was just Marty’s observation, and it’s difficult to distinguish between a
[levee] overtopping and a breach” (rep-142).
But the enormity of Katrina was not fully understood by the White House until at least Tuesday,
August 30 (rep-143). White House officials did not consider the breaches confirmed until roughly
6:30 a.m. the next morning, upon receipt of an updated situation report from DHS.11 As far as the
10 We should note that there was widespread confusion and misuse of the terms ‘breach’ and ‘overtopping’ by
observers and reporters who did not fully understand the distinction between the two terms, or whose
observations were not sufficient to enable differentiation of one from the other. (FR35)
11 The importance of the slowness in understanding was later downplayed: “Confirmation of a full breach would not have changed anything we would have done,” Rapuano said. “We weren’t going to repair the levees overnight, and search and rescue was already operating in full gear, regardless.” (rep-142). But confirmation of the breach of the
20
president knew, this was a normal hurricane and everybody was handling it (Brinkley, 2006: 139). It
was not until later in the day – a full day and a half after landfall – when Michael Brown informed
President Bush, Vice-President Cheney, Secretary Chertoff, and Deputy Chief of Staff Karl Rove in a
telephone call that at least 90% of New Orleans’s population had been displaced and that responders
“needed military assets; this was the big one.” He added that FEMA “needed the help of the entire
cabinet… DOD and HHS and everybody else” (Brinkley, 2006: 160-1). Brown later testified that this
was the turning point in the President’s comprehension of the catastrophe:
And as I recall my first statement to him was, you know, Mr. President, I estimate
right now that 90 percent of the population of New Orleans has been displaced.
And he is like, My God you mean it is that bad? Yes, sir, it was that bad.
Choking points in the information chain
Organization theorists have pointed out that many barriers within and between organizations often
prevent the necessary flow of information and the sharing of perceptions. To understand a large-
scale crisis typically requires a large number of actors, operating at different levels of the system, to
share and compare their picture of the situation. The more actors and the more variety in
organizational stripes and feathers, the harder it becomes to establish a shared picture of a dynamic
situation.
In his classic Man-made Disasters, Barry Turner (1978) explains how specialization and division of
labor in modern organizations create and institutionalize different ways of seeing, which, in turn, can
create collective “blind corners” (ways of not seeing). Organizations are “blindsided” by strong
cultures that emphasize a focus on routine events.
The problem of blind-sided organizations plays out across the response network. Sense-making takes
place in different units, at different organizational levels, across organizations; this leads to multiple,
conflicting interpretations, all of which may be plausible. Such a variety of perspectives makes it hard
to collectively puzzle together available information into a complete picture of a dynamic situation.
The deepest divide often opens between the worlds of first responders and strategic crisis managers
(Boin and Renaud, 2013). These actors operate at different levels (and in different worlds). The
strategic layer of decision-makers typically gathers at an Emergency Operation Center (ECO) and
manages a crisis from that location. The operational level is at the heart of the crisis: where the
explosion occurs, the shooting happens, or the levees break. It is where ranking officers from all
involved disciplines first respond to the scene.
At both levels, individuals are trying to order and understand the situation, drawing from a
cacophony of voices and images, comparing notes in an atmosphere marked by stress and chaos. The
levees could have had practical implications for White House involvement in the response. Flooding from breaches and flooding from overtopping have different consequences. Key point, but nobody understood it at the time (rep-142).
21
sense-making is done “on the side,” as most people in the chain of command have other activities
they must engage in: making decisions, calling partners, dealing with media and coping with things
that do not work. The perceptions of both worlds must all add up somehow, quickly and correctly.
In normal times, institutional mechanisms such as deliberation, specification and verification
(characteristic of the bureaucratic process) compensate for sense-making shortcomings. In the
absence of such mechanisms, an “appreciative gap” can rapidly emerge and divide the strategic from
the operational level.
The US had a system in place to do just overcome these pitfalls. In theory, the division of labor is
quite simple. The states always have the primary responsibility for managing a disaster that occurs
on their territory. If a state cannot cope and requires assistance, it can ask help from the federal
government. Information systems in the federal response system follow a similar logic. The response
organizations close to the disaster site report upwards in a chain of command. Information from the
states comes together in DC, where all federal organizations have crisis rooms. A special center of
DHS creates a picture of the situation. This happens again in the White House. And when this gives
rise to questions, requests are sent down the line again.
While information flowed upwards, the response system could not produce an accurate and timely
picture of the situation. In fact, the very systems that were designed to allow for smooth
information-processing would help blind authorities from noticing what quickly became obvious to
most anybody else.
One source of problems was found at the federal level, where critical information remained bottled
up in an especially designed information room. As stated earlier, waves of information (not all of it
accurate) were reaching DC. Various centers were collecting, interpreting, analyzing, summarizing
and sharing their information through bureaucratic layers. The Department of Homeland Security
had a center designed to perform this task of collective sense-making. A year before Katrina, the
Homeland Security Operations Center (HSOC) became operational. The HSOC was intended to
‘connect the dots’ during a disaster. It was blessed with a $70 million annual budget, and a staff of
300 people. With hundreds of trained people paying close attention to an emerging disaster, one
might expect collective sense-making to be effective.
HSOC became a major choking point in the information chain. On Monday evening the HSOC failed to
conclude that levees had breached in New Orleans (rep-140). That was strange, as HSOC had
information that indicated the opposite. Mid-afternoon on August 29, the U.S. Army Corps of
Engineers (USACE) notified DHS of a reported levee overtopping in St. Bernard’s Parish, a reported
levee breach in the West Bank, and a small breach in Orleans Parish reported by local firefighters
(FR36). However, as late as 6:00 PM EDT that day, the DHS Homeland Security Operations Center
(HSOC) reported to senior DHS and White House officials that, “Preliminary reports indicate the
levees in New Orleans have not been breached, however an assessment is still pending.” (FR36)
The HSOC director, Matt Broderick, played a key role here. Broderick reportedly did not read his
emails or watched TV. When Broderick on Tuesday morning learned through the radio on his drive to
22
work how bad the situation was in New Orleans he could not believe his ears (Brinkley, 2006: 155/6).
It was not until Wednesday that Broderick was finally convinced that the situation was really bad.
This lack of sense-making capacity was of self-inflicted. A former military official, Broderick insisted
on personally “making sense” of the incoming data. He had learned in Vietnam that in the fog of war
it was rarely immediately clear what the “hard facts” were (Brinkley, 2006: 156). He did not want to
commit the error of relaying mere impressions; he only wanted to pass on verified facts. Much of the
information he received during those first days did not meet his rigorous standards. Reports of
breaking levees were treated as mere rumors or “impressions” until verification was achieved
(Brinkley, 2006: 132/3).
Broderick turned HSOC into a fact-checking machine (Brinkley, 2006: 157). But HSOC did not have the
capacity to check facts. DHS and FEMA had only a few boots on the ground (and those officials did
not know the city (183). FEMA did little to improve its information position during those first days.
When James Lee Witt arrived in Baton Rouge on Friday (247), he was surprised to see the FEMA
communication truck in the EOC parking lot (CB 189-209). It is no surprise, then, that Broderick again
missed the story of the day on Thursday: the people at the Convention Center. When he began
reporting the “facts” of the disaster, his information no longer added any value. To make matters
worse, Broderick then stepped out of his role and began to micromanage rescue efforts (Brinkley,
2006: 181). A House report would later conclude that the HSOC “failed to provide valuable
situational information to the White House and key operational officials” (rep-3). This explains, at
least partially, why the White House and DHS chief Chertoff were “flying blind” those first days
(Brinkley, 2006: 158; 177).
The White House also had a small center to make sense of crisis. The HSC commenced 24- hour
operations the morning Katrina hit New Orleans (rep-133). It appears the White House took several
steps to improve the flow of information and strategic advice into the President. For example, HSC
staff solicited regular situation reports from almost every federal agency for the White House
situation room. A House report would later conclude that “the White House failed to de-conflict
varying damage assessments and discounted information that ultimately proved accurate (rep-3).
5. Horizontal and vertical coordination in response to a catastrophic event
What does a full-fledged response to a mega-disaster require? First and foremost, the immediate
mobilization of resources: for search and rescue, tending to people, feeding them, and moving them
out of harm’s way. This, in turn, requires smooth cooperation between many actors – at different
levels of governance and across sectors – who must work together to assure an effective response.
Some of that cooperation will just happen. But at least some of it will have to be organized. The
combination of intricate multi-level interactions, swift action, and massive resources – all delivered
without political infighting – can only happen if there is coordination. This is one of the biggest
challenges governments face in times of crisis.
Most governments have a plan for a coordinated response. So did the US before Katrina (described in
the National Response Plan). The formal structure of the disaster response was built on the idea of a
23
bottom-up response, recognizing the key position of governors. The plan sketched a clear division of
labor between local, state and federal actors. The adoption of NIMS and ICS introduced a shared
language that could connect the many actors in a response operation.
It did not work as expected or as hoped. Even though many organizations and individuals performed
to the best of their abilities, their efforts did not add up to an effective and timely response. Two
factors explain this outcome. First, key actors were late in recognizing that their efforts would not be
sufficient in the face of the extraordinary scale of the disaster. Second, the plan for dealing with a
super-disasters simply was not good enough. As a result, the stricken areas in Louisiana and
Mississippi had to wait a long time for outside help.
We define coordination as the set of activities aimed at orchestrating the collaboration between key
actors in a disaster response (Boin and Bynander 2015). This collaboration has a vertical and a
horizontal dimension, which implies that coordination has both a facilitative and a more directive
character. Whereas vertical collaboration may be susceptible to steering efforts, these horizontal
forms of collaboration are much harder to govern.
The vertical dimension in our story runs from New Orleans, through Baton Rouge, to Washington
D.C. While there were many actors active on this vertical dimension, we will concentrate here on a
defined set of key actors: the state of Louisiana, FEMA, DHS, President, and the Pentagon. The city
administration of New Orleans was essentially non-functioning, which means the vertical axis was
not “anchored” in the disaster setting. This would prove a source of trouble for vertical cooperation
and coordination.
The horizontal dimension pertains to the collaboration between partners who do not stand in a
hierarchical relation to each other. The emerging collaborations varied far and wide: from
collaborating citizens (New Orleanians and outside help) to states assisting each other; from private
companies bringing in help to citizens bringing in private security firms. Horizontal collaboration thus
played out at different levels (local, state, national and international), crossing public-private borders
as well. Intriguingly, horizontal coordination seemed to work better than vertical coordination,
certainly in the initial phase of the response.
The system in place: A National Response Plan for “normal” and “catastrophic” disasters
The Robert T. Stafford Disaster Relief and Emergency Assistance Act (the Stafford Act) sets out how
the Federal government may assist state and local governments “in carrying out their responsibilities
to alleviate the suffering and damage” caused by disasters. The Act authorizes the President to direct
federal agencies to support state and local response efforts, but it makes clear that the Federal
government must respect State sovereignty. A state has to request assistance from the federal
government before federal agencies can offer help or coordinate assistance. The Stafford Act
establishes a detailed process for State governors to request assistance from the Federal government
when an incident overwhelms State and local resources. (FR12)
24
The Stafford Act thus creates an inherent tension within the US response system between President
and Governor. The 9/11 attacks gave this tension much more weight: when an urgent response is
required, there may be little time to follow legal procedures.
Under the Stafford Act, the President can designate an incident as an “emergency” or a “major
disaster;” the National Response Plan (see below) adds a third category: a “catastrophic event.” The
differences between these categories are not absolute, as becomes clear from these definitions:
Emergency: Can be declared unilaterally by the President; authorizes the Federal government to provide essential assistance to meet immediate threats to life and property; assistance is limited in scope and may not exceed $5 million without Presidential approval and notification to Congress.
Major disaster: Can only be declared by the President after the Governor has requested it and has
certified that the state and local governments are overwhelmed; authorizes the Federal government
to provide essential assistance to meet immediate threats to life and property (same); the full
complement of Stafford Act programs can be authorized, including long-term assistance (e.g. public
infrastructure recovery, consequence management)
Catastrophic event: any natural or manmade incident, including terrorism, that results in
extraordinary levels of mass casualties, damage,, or disruption severely affecting the population,
infrastructure, environment, national morale, and/or government functions (National Response Plan,
2004:43).
With the catastrophic event, the NRP introduced an exception to the normal procedure: if a disaster
occurs that outstrips state capacities to respond, the federal government can jump in without such a
request (“proactive response”). The underlying idea was simple, its origins in the wake of 9/11 are
clear: When a catastrophic incident occurs, regardless of whether the catastrophe has been a warned
or is a surprise event, the Federal government should not rely on the traditional layered approach
and instead should proactively provide, or “push,” its capabilities and assistance directly to those in
need. (FR19) This Annex promised “an overarching strategy for implementing and coordinating an
accelerated, proactive national response to a catastrophic incident.” But it was not clear what this
strategy was and how it differed from a “normal” response to a large-scale disaster. It did not specify
which actions should be taken and what components should be utilized under the NRP had a
catastrophic incident been declared. (FR15)
This may work well when it is immediately clear that an incident is catastrophic (think of a large
earthquake or terrorist event), but the plan did not take into account that some events may develop
into a catastrophic event. The Catastrophic Incident Supplement (an annex to the annex, published in
April 2005) states that the procedure is “designed to address a no-notice or short-notice incident of
catastrophic magnitude” (p. 5). If an event is recognized as such, the supplement promises
accelerated delivery and an “aggressive concept of operations” – just what was needed, of course.
This amounted to “pushing” predesignated resources to a staging area, something FEMA was already
doing. The “push” idea is great, but it was predicated on the immediate recognition and declaration
of a catastrophic event.
25
The declaration of an Incident of National Significance – disaster or catastrophe – opened the door to
involving the Department of Defence (DoD) (100). A fundamental principle of DOD’s approach to civil
support holds that “[I]t is generally a resource of last resort” (rep-39). Only when civil authorities
cannot handle it, the DoD steps in (rep-203). But the DoD would not be governed by any other
federal organization. This structure almost guaranteed two characteristics of a response to a
catastrophe. First, the system almost ensures that DoD assistance will be too late, as it takes time to
understand that the system is overwhelmed. Second, it guarantees coordination tensions between
DoD (must “take over”) and civil authorities (must have failed).
FEMA: A de-institutionalized agency
FEMA and its director (“Brownie”) would become the national scapegoat for the federal response. In
hindsight, this verdict is not a fair one. FEMA did not have the means or the authority to orchestrate
a proper response to a catastrophic event like Katrina.
The post-Katrina discussion of FEMA’s performance (or lack thereof) suffers from a widespread
confusion about FEMA’s role in a disaster. FEMA is not responsible for providing such a response.
FEMA does not, for instance, provide mass care or transportation after a disaster (FR17). That was (in
this case) first and foremost the responsibility of the states (McCreight, 2015). 12 FEMA exists
primarily to coordinate other Federal agencies and departments during emergency response and
recovery—acting as an honest broker between departments and agencies, providing a command
structure, and serving as the single point of entry for State and local officials into the Federal
government. (FR17).
But it is true that FEMA had seen better times. FEMA was an organization in a state of de-
institutionalization, overstretched by a string of hurricanes (Lu, 2014). In March 2005, the Mitre
corporation wrote a report about FEMA, tearing it apart (CB91). FEMA’s budget had been cut (CB84).
Its professional ranks had declined [..] since 2002, a number of its top disaster specialists, senior
leaders, and experienced personnel had left, described as “FEMA brain drain” (rep-152).
After the 9/11 attacks, FEMA was merged into the new Department of Homeland Security.13 The
reasons for this unprecedented merger included enhanced coordination and a more effective
response to domestic crises – especially terrorism (Kettl, 2003)). Academics and disaster policy
makers criticized the merger, as they feared that the national preoccupation with terrorism would
crowd out the attention for federal disaster response (Tierney, 2006). Natural disasters were much
more likely to occur than terrorism, they argued. Similar sentiments prevailed within FEMA, as
Michael Brown’s book makes clear. With the merger, the FEMA director lost the cabinet-level
position held under the Clinton administration.
12 Critics could only see it one way: when it was pointed out that FEMA performed very well in 2004, this was
suddenly due to Florida’s excellent preparation CB86-7.
13 The Homeland Security Act was passed by Congress in November 2002. DHS opened its doors in March 2003.
26
Much was later made of FEMA’s demoralized state coming into this disaster, but is hard to see how
this state of “institutional depletion” could explain the flawed response to Katrina. Only recently,
after all, FEMA had performed very well in response to a range of disasters.
There was simply was no structure and process in place to deal with a super-disaster. More
specifically, there was no mechanism to smoothly “scale up” if and when FEMA could no longer fulfil
its mandate.
Vertical coordination
Setting the stage: local “dis-anchoring”
Vertical coordination is not the same as command and control. The concept of vertical coordination
refers to activities aimed at orchestrating the actions of actors that have their own responsibilities
and mandates; they cannot be ordered to do something (or it would be highly ineffective and
inefficient to do so even if it were possible). Decades of public administration and crisis management
research findings suggest that it is not easy to effectively and legitimately orchestrate a crisis
response from the top down (Boin and Bynander, 2015).
This coordination challenge is not made any easier when local actors – the backbone of an effective
response – are overwhelmed and cannot perform their envisioned task. That is exactly what
happened in New Orleans. The immediate consequence was that there were few resources or actors
featuring in the formal plans to be coordinated. Local responders did not receive the support they
needed - a Louisiana firefighter stated, “the command structure broke down—we were literally left
to our own devices.” (FR37). Many requests for assistance could not be answered - Members of the
Hammond (Louisiana) Fire Department reported receiving “a lot of ‘I don’t knows’ from [local]
government officials”. The national response had lost its local footing from the start. The challenge
thus immediately shifted to coordinating and facilitating local and incoming actors that had never
collaborated before under these circumstances.
The National Guard filled the vacuum, at least to a degree. The National Guard HQ in New Orleans
(Jackson Barracks) was wiped out by a flash flood early Monday morning. “For a crucial 24 hours
after landfall”, The New York Times reported, Guard officers “were preoccupied with protecting their
nerve center from the waves topping the windows at Jackson Barracks and rescuing soldiers who
could not swim”.14 This was just bad luck: Jackson Barracks was not located in a flood-prone area. In
fact, Jackson Barracks had weathered all manner of floods and hurricanes over the years. The Guard
moved its headquarters (consisting of 375 guardsmen) to the Superdome by boat and helicopter the
next morning.15 It would become a local hub for coordination, unrecognized by the “civil” line of
command.
14 Scott Shane and Thom Shanker. “When storm hit, National Guard was deluged too”. The New York Times, 28
September 2005.
15 Shane & Shanker.
27
The Louisiana Guard immediately contacted the National Guard Bureau in Washington and asked for
more help. By noon on Monday, 52 state guard commanders were going through a laundry list of
local needs and started to organize assistance. As a result, “helicopter search and rescue teams
began arriving late Monday from as far away as Wisconsin”.16
The Guard maintained order at the Superdome and fed the thousands of evacuees.17 From the
Superdome, the Guard also coordinated efforts of the police, firefighters and volunteers.18 The Guard
had 64 helicopters running rescue sorties. Local police and firefighters hooked up with the Guard,
making use of their communication means. The Guard was “controlling more than 200 boats, most of
which were run by mixed crews of Guardsmen, police, firefighters and officers of the Louisiana
Department of Wildlife and Fisheries”.19
Accomplishments Louisiana National Guard:
- Conducted security and screening at the Superdome on Aug. 28 (Source: http://www.defense.gov/news/newsarticle.aspx?id=16778)
- Traffic control and security (Ibid) - Transporting & distributing food, water, and ice (Ibid) - Conducting search & rescue (Ibid) - Provide generator support (Ibid) - Setting up shelters (Ibid) - Helping state police with evacuations (Ibid) - 10,244 sorties flown
(http://www.realclearpolitics.com/articles/2006/05/katrina_what_the_media_missed.html)
- 17,411 saves by air (Ibid) - 88,181 passengers moved (Ibid) - 18,834 cargo tons hauled (Ibid) - Had 150 NG aircraft running by the end of the week (Ibid) - Brought in 300,000 MREs and 397,000 liter bottles of water
Yet, despite all these good efforts, the governor and her staff rapidly lost control over the situation.
But things still looked under control on Monday. This explains why governor Blanco asked for help
but did not “scream loud enough” (Brinkley, 2006: 266) when President Bush called her on Monday
late afternoon. This makes sense: the level of devastation had yet to become fully clear. This also
explains why Bush did not do much: he was told by Brown that everything was under control (266-
267). At this point, federal-state coordination was not troubled by tension.
16 Dolinar.
17 Dolinar
18 Lou Dolinar, “Katrina: What the Media Missed”. RealClearPolitics. May 23, 2006.
19 Dolinar
28
Vertical coordination between the state of Louisiana and the federal government had started well.
Governor Blanco requested federal assistance on Saturday, August 27, asking that President Bush
declare an emergency. Later that same day, President Bush declared an emergency for the state of
Louisiana. William Lokey was named Federal Coordinating Officer. On Sunday, August 28, in
recognition of the potential catastrophic impact of Hurricane Katrina, Blanco asked President Bush to
“declare an expedited major disaster for the State of Louisiana as Hurricane Katrina. President Bush
declared a major disaster for Louisiana (rep-63).
FEMA worked closely with GOHSEP during the weekend. During the Sunday noon video conference
with Brown, Bush, Chertoff and state officials (Brinkley, 2006: 97) there was no sign of coordination
confusion. The Louisiana EOC reported that evacuations were going well, that it had no unmet needs,
and that FEMA was “leaning forward” as far as possible. (FR29). At this point in time, the system
appeared well prepared (and arguably was well prepared) to deal with a major hurricane.
On Tuesday, the governor and FEMA director spent quite a bit of time together. They took a chopper
ride (with Lt. governor Landrieu and Senator Vitter, Brinkley 2006: 279). They visited the Superdome
(with Nagin and Compass 282). The atmosphere was described as “good” (283). Priorities were
agreed on (sandbags, evacuating the Superdome). Ragin handed a detailed wish list for Brown (161),
which he passed on to incoming the FEMA site chief, Philip E Parr (162). It was understood that FEMA
could not deliver resources out of thin air. Dr. Walter Maestri, the Jefferson Parish Director of
Emergency Management explained it well: he understood that FEMA may not provide help until 48-
72 hours later—but then he expected help (rep-83). By Wednesday (48 hours after landfall), the
relation had begun to deteriorate as FEMA could not deliver what it had promised.
But this was not a “normal” disaster. After landfall, it quickly became clear that Louisiana was
overwhelmed. New Orleans institutions did not function. Many individuals performed well if not
heroically, but the breakdown of local institutions made it hard to coordinate from the bottom up.
The New Orleans Police Department fell apart. Other city departments did not step up.
The state simply did not possess the means to deal with the sheer devastation and suffering that
soon would become apparent. The Louisiana Department of Wildlife & Fisheries (W&F) performed
heroically during the search and rescue operation (all praise would later go to the Coast Guard).
There were 4,000 Louisiana guardsmen on duty when Katrina hit; by the next day all available 5,700
Guardsmen were on duty.20 Even though a portion of the state’s most important resource had been
washed away by a flash flood in New Orleans, the National Guard remained active and highly
effective.21 From the Superdome, the National Guard’s main command organized troops, helicopters,
shelters and “a triage and medical center that handled 5,000 patients (and delivered 7 babies)”.22 The
Guard coordinated the efforts of police, firefighters and volunteers, as well as coordinated the efforts
20 Shane and Shanker.
21 Shane and Shanker.
22 (http://www.realclearpolitics.com/articles/2006/05/katrina_what_the_media_missed.html
29
of incoming Guard forces from other states.23 Ultimately, more than 50.000 National Guard troops
from all 50 states came to Louisiana to aid in the response (but reinforcements would not arrive in
large numbers before Thursday).24
When the existing structure does not work: The challenge of upscaling
It took some time for federal administrators to discover that the system in place was not suited to
coordinate the response to a super-disaster. The mindset in Washington D.C. immediately before
Katrina made landfall in Louisiana could be characterized in simple terms as “we’ve got this covered”.
FEMA had responded well to a range of hurricanes (both in the previous summer and in the weeks
before Katrina). President Bush could vacation in Texas in the sure knowledge that Chertoff had
things under control; Chertoff could take a relaxed stance as he relied on FEMA to manage the
situation. FEMA director Brown did not nothing to disturb this confident approach, as he was
convinced that FEMA was working well with the states under threat – which was true and confirmed
by the leaders in the local EOCs. Days into the disaster, Bush could rightly assume that Michael
Brown was “doing a heck of a job.”
Sometime in the middle of that first week, it became clear on the ground that FEMA could not deliver
on its promises. Some FEMA officials blamed Louisiana for the delays in the relief efforts.25 Brown
would later call the state of Louisiana ‘dysfunctional’ and said that they did not have unity of
command (rep187). Wells accused the state of Louisiana of being too occupied with evacuation
efforts to participate with the federal government’s pre-landfall planning of search and rescue, rapid
assessment teams, medical evacuation, sheltering and temporary housing, and debris removal. Wells
also claimed that the state bypassed FEMA for federal assistance, and then later complained that
FEMA did not know what was going on and that FEMA could no coordinate the federal effort.
This criticism overlooks the fact that the very reason for FEMA’s existence was (and is) to assist
overwhelmed states. FEMA came into this disaster well-prepared and well positioned. But at some
point that week, FEMA itself was being overwhelmed, and it could not deliver what was promised
and expected. Looking back, we can sum up FEMA’s problems as follows:
FEMA did not have “eyes and ears” in the disaster zone. This made it hard to understand
what was needed (and where); it made it impossible for FEMA to orchestrate the
logistics of disaster goods being moved to the impact zone.
23 (http://www.defense.gov/news/newsarticle.aspx?id=14897;
http://usacac.army.mil/cac2/cgsc/carl/download/csipubs/wombwell.pdf, p.70)
24 Shane and Shanker.
25 Note that not all FEMA officials (i.a. Lokey) agree with this assessment.
30
FEMA could not get what was needed on the ground: FEMA was well prepared for a
“normal” disaster, but not for Katrina. When FEMA realized that the agency could not
meet expectations, it passed the baton – but then got cut from the loop.
FEMA did not manage expectations. It made promises that could not be kept and failed
to explain its limitations. This helped to maintain unrealistic demands and expectations
at the local level.
No eyes and ears
As FEMA is not a disaster manager, it had no boots on the ground in New Orleans apart from the
well-informed Marty Bahamonde and Strickland (but they apparently did not know each other –
Brinkley, 2006: 237).26 The agency relied on functioning local institutions. This had always worked
well. Not this disaster. As Bill Lokey, the FEMA Federal Coordinating Officer in Louisiana, explained:
“The locals were overwhelmed. We were going to be overwhelmed. There was no way, with my
experience and what I had to bring to the table, I was taking a knife to a gunfight.” (SA33)
FEMA could not coordinate from the ground up. It had nobody qualified to do this in the city. FEMA
people found it hard to reach the city as they did not understand the geography of the submerged
city (258). Communication failed, so Lokey could not talk directly with his advance team leader in
New Orleans, Parr. Lokey and his staff in the EOC did not know another FEMA official, Marty
Bahamonde, was in New Orleans during and immediately after landfall until they were informed by
FEMA headquarters on late Monday, August 29. Before that time, they did not even know
Bahamonde was there or what his function was. (rep-190).
Unable to deliver
FEMA’s failure to deliver both on expectations and promises did not stem from a lack of trying. In
fact, in at least some respects, FEMA’s response was greater than ever before (rep-151). FEMA had
pre-positioned three US&R task forces (one Type I and two Type II) and the Blue IST (?) at Barksdale
Air Force Base in Shreveport.27 Two of its five Mobile Emergency Response Support (MERS)
detachments were prepositioned and quickly moved to the affected areas in Louisiana and
Mississippi soon after landfall. (FR43). FEMA had 8 medical teams on standby (Brinkley, 2006: 69 -
check).
After landfall, FEMA brought in more resources: 50 FEMA medical assistance teams, 25 FEMA search-and-rescue task forces, and Approximately 1,700 trucks of ice, water, and MREs.28 FEMA had 70 people in DC calling around for supplies 335. FEMA had delivered over $1 billion in assistance to
26 James Strickland was a member of FEMA’s Search and Rescue team (sen325). He was there to partake in the
S&R mission set up by FEMA.
27 (Hearing Lokey Committee on Homeland Security and Governmental Affairs http://www.gpo.gov/fdsys/pkg/CHRG-109shrg26751/html/CHRG-109shrg26751.htm). 28 Bourget, P. (2005), p.16 http://www.gwu.edu/~icdrm/publications/PDF/EMSE334_Katrina.pdf
31
evacuees in all fifty States and the District of Columbia by September 17—less than three weeks after landfall. (FR49). The FEMA US&R teams performed well, ultimately rescuing over 6,500 people. (FR38). When FEMA gained access to several helicopters, FEMA began ferrying food and water to people stranded on high ground even though there was no formal request by the state to perform this function. In addition, FEMA contracted with over 100 ambulances to transport hospital evacuees. This mission was not requested by the state, but FEMA responded proactively rep-139. From August 31 to September 4, FEMA also deployed ten Disaster Mortuary Operational Response Teams (DMORTs) and both of its Disaster Portable Morgue Units (DPMU) to help State and local personnel identify and process bodies at those collection points. (FR48)
It was not enough. Within days, FEMA could not deliver what was requested. FEMA simply could not
procure enough resources to match the rate at which commodities were being consumed. (FR44);
Carwile stated, “We never had sufficient personnel to meet requirements.” Scott Wells, Deputy FCO
for Louisiana, would later complain that a 90-person FEMA regional office “is woefully inadequate”
to perform its two primary disaster functions, operating a regional response coordination center and
deploying people to staff emergency response teams in the field (rep-157).
FEMA officials tried to arrange requested goods and services. FEMA does not own any of these
resources, they merely coordinate the allocation and use of them. This means that FEMA has to
request these resources and manage the allocation process. Many actors – not least the media – did
not seem to grasp the extent of FEMA’s role. FEMA, in turn, failed to explain its role and
competences. This created a persistent performance gap, at least in the eyes of those in desperate
need of those requested goods and services.
Louisiana officials thus had reasons to complain. FEMA was not delivering on its promises (Brinkley,
2006: 190). As GOHSEP officials expected FEMA to arrange buses (as promised) to evacuate people
from the Superdome, they stopped their own efforts to organize buses (184). By Friday, the governor
had become so frustrated with the federal response that she brought in James Lee Witt, a consultant
and former FEMA chief under President Clinton (247).
When the coordinator is overwhelmed
The midweek breakdown of FEMA came as a surprise in Washington D.C. The dominant mindset at
FEMA and DHS was that “if there is anything that FEMA does and has done over the last twenty
years, it’s been hurricanes” (Chertoff) (CB240). During his testimony before the Select Committee,
Chertoff freely admitted that he did not have much experience with disasters and he did not consider
himself a hurricane expert (rep-132). The mind-set at DHS long remained one of “normal disaster”
that FEMA could handle (131-2). As the scale of the disaster was not predicted or foreseen by anyone
in the federal government, Chertoff quite rightly refused to pre-emptively scale up to “catastrophic
organization”. In consistent fashion, DHS did not really get involved before the end of Tuesday
August 30, when Secretary Chertoff declared Hurricane Katrina to be an Incident of National
Significance (INS) (FR41; Brinkley, 2006: 368).
32
By then, the organization of the response was coming apart at the seams. But when FEMA tried to
“scale up” the response, the federal network rapidly dissolved into separate hierarchies. The
response, in other words, became fragmented.
Halfway through that first week, FEMA began to seek assistance. The deteriorating relations between
DHS and FEMA undermined a coordinated response. On several occasions, Chertoff could not locate
Brown (who later said he simply ignored calls from HSOC); when a livid Chertoff would finally reach
Brown he told the FEMA director to sit tight in BTR (CB170).
The underlying problem was that nobody seemed to understand what a catastrophic event was and
what should happen when the response “scaled up” to deal with such an event. The idea of a
catastrophic event is that it requires a qualitatively different type of response network than a
“normal” disaster. The NRP-CIA was specifically written for a disaster such as Katrina. But the authors
of the NRP-CIA did not anticipate that it might be unclear when an event takes on catastrophic
proportions (rep-137). It is clear [in hindsight, yes] the consequences of Hurricane Katrina exceeded
all of these criteria and required a proactive response.
Things became even more complicated when FEMA approached DOD about taking over the logistics
mission on Thursday, September 1, according to staff interviews with senior FEMA officials (Brinkley,
2006: 100/186).
The involvement of DoD creates the immediate question: Who’s in charge? In addition, it prompts
the question how different organizational hierarchies (state, FEMA, DHS, DoD) should work together.
There is no question that the DoD played an important role in the response to Katrina. On Tuesday,
Deputy Secretary of Defense Gordon England had authorized U.S. Northern Command
(USNORTHCOM) and the Joint Chiefs of Staff to take all appropriate measures to plan and conduct
disaster relief operations in support of FEMA. (FR42). DoD set up Joint Task Force Katrina. It would
become “the largest military deployment within the United States since the Civil War.” (rep-201).29
By September 1, JTF-Katrina included approximately 3,000 active duty personnel in the disaster area;
within four days, that number climbed to 14,232 active duty personnel.30
But it took time before the DoD could become truly effective. While the DOD’s principle of non-
involvement with civil affairs initially withheld the DoD from action, the DOD leadership actively
encouraged a culture-switch to a ‘can do’ approach which allowed the DOD to step in and act
effectively.31 Although skilled and trained in war-fighting missions abroad, conducting joint missions
within the US, quickly and under tremendous public pressure, posed integration challenges (rep-
225).
29 National Guard forces represented more than 70% of the military force for Hurricane Katrina (rep-202). 30 Two C-130 firefighting aircraft and seven helicopters supported firefighting operations in New Orleans.
(FR43) 31 Moynihan, D. P. (2012). A theory of culture-switching: Leadership and red-tape during Hurricane Katrina.
Public Administration, 90(4), 851-868.
33
The Joint Task Force was commanded by LTG Honoré, a Louisiana native. Honoré arrived on Thursday
in New Orleans (Brinkley, 2006: 524), without troops but reportedly with CNN reporters in tow. He
was extremely effective in the “demilitarization” of the response, ordering troops to adopt a less
aggressive stance towards the local population. The New Orleanians were understandably relieved to
see one of their own in uniform, taking command. His arrival on scene coincided with a turning point
in the response – help was finally arriving.
Louisiana officials were initially critical about DoD’s process for receiving, approving, and executing
missions, which they considered bureaucratic. The process for requesting DOD active duty forces has
several layers of review and is understandably not well understood or familiar to state officials who
rarely would need to request DOD support (rep-204). FEMA was still supposed to request specific
assistance from the military. But FEMA has lost control. Louisiana’s Adjutant General created a work
around and made the request directly of General Russel L. Honoré — without coordinating the
request through FEMA — the established process to request all federal assistance (rep-204, 5). The
Defense Department “actually drafted its own requests for assistance and sent them to FEMA, which
copied them and sent them back to the Department of Defense for action” (rep-204).
As a key manager of the federal response, Honoré received criticism as well. Hull noted that as
Honoré made command decisions away from his headquarters and that his staff was not always
informed. “We track General Honoré’s location by watching CNN,” JTF Katrina staff said (rep-225).
The National Guard 38th Infantry Division, composed of smaller Guard units from many states,
reported they never formally coordinated with Northern Command (rep-219).
Eventually, over 50,000 National Guard members from fifty-four States, Territories, and the District
of Columbia deployed to the Gulf Coast, providing critical response assistance during this week of
crisis. (FR43). Once forces arrived in the Joint Operations Area, they fell under separate command
structures, rather than one single command. (FR43). The standard National Guard deployment
coordination between State Adjutants General (TAGs) was effective during the initial response but
was insufficient for such a large-scale and sustained operation. (…) A fragmented deployment system
and lack of integrated command structure for both active duty and National Guard forced
exacerbated communications and coordination issues during the initial response. (FR43)
An official report summarized the criticism in the following terms:
Joint Doctrine was largely ignored. In the melee of the first few days where lives literally hung
in the balance, perhaps this was a necessary course of action. However, as the Active Duty
Force began to develop, the JTF Katrina headquarters never transitioned from the very
tactical mindset of life saving to the operational mindset of sustaining and enabling a Joint
Force. Since the Forward Command Element (General Honoré) was unable to communicate,
they became embroiled and distracted with the tactical and were unable to focus on even
the most basic of operational issues [..] Since the JTF did not establish a commander for all
land components, 1st Army, 5th Army, and the Marine Corps were unclear on JTF Katrina
34
expectations, causing confusion and lack of coordination between land forces in New Orleans
(rep-225).
It was not clear to local officials who was in charge in Washington DC. Louisiana’s Smith stated
“[a]nyone who was there, anyone who chose to look, would realize that there were literally three
separate Federal commands (rep-189). Although the DHS Secretary designated a PFO to be the
deferral government’s representative under the NRP structure and to coordinate the federal
response, the efforts of all federal agencies involved in the response remained disjointed because the
PFO’s leadership role was unclear. In the absence of timely and decisive action and clear leadership
responsibility and accountability, there were multiple chains of command, a myriad of approaches
and processes for requesting and providing assistance, and confusion about who should be advised
of requests and what resources would be provided within specific time frames. (HR420) The Select
Committee reported that it “found ample evidence supporting the view that the federal government
did not have a unified command” (rep-189).
Federal response officials in the field eventually made the difficult decisions to bypass established
procedures and provide assistance without waiting for appropriate requests from the states or for
clear direction from Washington. These decisions to switch from a “pull” to a “push” system were
made individually, over several days, and in an uncoordinated fashion as circumstances
required.(rep-132). The response to Katrina may have “evolved into an ad hoc push system”(rep-
138), but it did deliver: In addition to ground operations, a joint DHS, DOT, and DOD airlift
successfully evacuated over 24,000 people, constituting the largest domestic civilian airlift on U.S.
soil in history. (FR40)
On September 5, Secretary Chertoff appointed Vice Admiral (VADM) Thad Allen to the position of
Deputy PFO. (FR47). Allen was appointed the FCO for Louisiana, Mississippi, and Alabama in addition
to PFO (rep-136), which was unprecedented. The Secretary was reportedly confused about the role
and authority of the PFO. The apparent confusion over the authority and role of the PFO does not
seem to have been recognized until almost two weeks after Chertoff selected Allen to replace Brown
as PFO. Confusion or not, Allen’s appointments ultimately proved critical for energizing the JFO and
the entire Federal response to Hurricane Katrina. (FR47)
Horizontal coordination
Horizontal coordination seemed to work reasonably well on the ground. Federal search and rescue
assets from the Coast Guard, FEMA Urban Search and Rescue (US&R) Task Forces, the Department of
Defense, and other Federal agencies worked in concert with State and local responders to rescue
tens of thousands of people (FR38). The National Guard was particularly successful in coordinating
the rescue efforts of many local and incoming groups.
One of the key successes of horizontal coordination was found in the EMAC procedure. The states
had in place a formalized agreement for mutual cooperation in cases of disaster. The cooperation
between the southern states was hampered by the scale of the disaster: neighboring states
Louisiana, Mississippi and Alabama could not help each other, as they were all overwhelmed. But
35
Florida (which had just been hit by Katrina) sent troops to Mississippi and Texas helped Louisiana
(190). By all accounts, the cooperation worked well and provided much needed resources.32
But in Washington, D.C. cooperation and coordination were in demand.
FEMA found it hard to coordinate horizontally within its organization. FEMA was working with a dual
structure. Because Hurricane Katrina was advancing toward Louisiana (Region VI), and Florida,
Mississippi, and Alabama (Region IV), both FEMA regions conducted response and recovery
operations. (FR17). While this formal division of labor had always worked before, it created internal
coordination challenges in the case of Katrina. The disaster was so big, it involved multiple regions
(which normally does not happen).
FEMA did work with other federal agencies like the U.S. Forest Service and city firefighters from
across the country to staff FEMA positions in the state (rep-157). But FEMA found it hard to work
with some other agencies. For instance, National Disaster Medical System (NDMS) teams also formed
an integral component of the medical response to Hurricane Katrina, collectively treating over
100,000 patients. (FR46). Several agencies assigned responsibilities in the NRP under ESF-8, Public
Health and Medical Services, sent liaisons to the HHS Operations Center in Washington, D.C., and the
HHS Secretary’s Emergency Response Teams (SERTs) in the affected States. (FR46) HHS struggled in
its NRP role as coordinating agency for ESF-8. HHS was criticized for lacking control over vital medical
assets, over-relying on departmental routines, and not having adequate disaster plans. FEMA
compounded HHS coordination difficulties. FEMA deployed NDMS teams without HHS’s oversight or
knowledge. FEMA administrative delays in issuing mission assignments exacerbated the lack of
coordination within ESF-8 and created additional inefficiencies. In order to respond swiftly, HHS felt
compelled to take emergency response actions without mission assignments, bypassing FEMA. While
this may have pushed additional assets to the region, it also had a deleterious effect on the Federal
government’s situational awareness of its deployed assets. (FR47)
There were many complaints about FEMA. The agency turned away out of state rescuers (Brinkley
441). On Monday, Brown directed all outside emergency workers to stay home (254). FEMA stopped
the activities of a California rescue time because they had no license (Brinkley 2006: 537, 554). State
and local officials expressed frustration that requests for assistance were not processed because they
did not follow the formal request process (rep-139). For example, the American Bus Association
spent an entire day trying to find a point of contact at FEMA to coordinate bus deployment without
success. (FR45).
On Tuesday, DHS began efforts to strengthen horizontal coordination at the federal level. DHS
initiated a virtual National Joint Information Center (JIC)141 and conducted the first of what would
become daily National Incident Communications Conference Line (NICCL) calls with other Federal
32 New Mexico Governor Richardson complained that his 200 Guard members were not requested
while offered. This illustrates the complexity of the situation.
36
departments and agencies. (FR42). At a Wednesday 1 pm press conference, Chertoff reported that
he was extremely pleased with the help offered by federal departments (Brinkley, 2006: 440).
The relations between DHS and other federal organizations would soon suffer from bureau-politics,
however33. DHS had trouble working with the FBI (232). Congressional researchers reported about “a
pointless “turf war” between DHS and DOJ” over which agency was in the lead. (SS13). Moreover, the
relation between DHS and DoD was not clear.
FEMA could neither efficiently accept nor manage the deluge of charitable donations (Brinkley, 2006:
188). FEMA did not know where the supplies were and when they would arrive (200). Brown turned
may aid offers down (250). Absent an implementation plan for the management of foreign material
assistance, valuable resources often went unused, which frustrated many donor countries. (FR45)
Private sector companies encountered problems when attempting to donate their goods and
services to FEMA for Hurricane Katrina response efforts. (FR45) The performance of private
corporations has been widely celebrated after Katrina. Large-scale corporations such as Walmart and
Home Depot organized and distributed much needed resources across disaster-stricken areas.34 But
the same corporations were frustrated by the lack of coordination at both the state and federal level.
The Louisiana crisis center could not match the large number of incoming requests for resources with
the list of offered resources.35 At the federal level, there was no venue for corporations to offer their
assistance. When a Walmart representative called, a mid-level official acted flexibly to bring Walmart
on board (but he reportedly got punished later for violating rules in doing so) (Brinkley, 2006: 260 ff).
6. Conclusion: Do we know how to manage a super disaster?
Ten years after, the response to Katrina is widely viewed as a qualified failure. Critics assert that the
response was too little, too late. This assessment is too bleak in at least two ways. First, it overlooks
all the things that actually went very well (particularly the search and rescue efforts). Second, the
critics do not explain what we could have reasonably expected. What is a reasonable time period
between the onslaught of a super disaster and the response?
Looking back, the criticism on the response appears to concentrate on the evacuation of those who
stayed behind in New Orleans. Many people suffered, there is no doubt. But that is part of a disaster.
Given the circumstances imposed by a super disaster, some things went reasonably well here, too.
People were fed and protected at the Superdome, and evacuated four days after the disaster. We
33 Add sources on crisis management and bureau-politics
34 Add sources – see article on Wafle House in IJPE special issue
35 GOHSEP brought in a professor of the University of Louisiana, Dr Ramesh Koluru, who worked with his team
to create software that would allow for quick matching of requested and offered assistance. This effort would
later inform the need for a unique initiative to facilitate cooperation with the private sector during disasters:
the SDMI Business Preparedness Center, situated at the Louisiana State University South Campus.
37
might argue if this was way too late, but it simply takes a minimal amount of time to organize an
evacuation from a city under water.
Some of the really bad stories (the bridge, the looting) had very little to do with the shortcomings of
a federal response but were due to the performance of local organizations. At least some of these
stories were wildly exaggerated in media accounts. And some of the really bad stories (the hospitals)
did not become known until well after the city had been emptied.
This brings us back to the aim of this paper. How should we assess the response and what lessons
can we learn for future disasters. We employed a simple yet effective framework of executive crisis
tasks to organize our investigation. Let’s revisit our preliminary findings (as our research is still
ongoing).
How much preparation is enough for a super disaster?
Given the evidence, it would be hard to argue that the authorities did not take Katrina seriously, or
that they failed to prepare and warn the local populations. These preparatory efforts included nearly
all the activities one would expect. There was a well-executed evacuation for car owners; FEMA had
pre-staged resources; the state of Louisiana and the federal Coast Guard had boats ready; and
shelters were organized. As a federal investigation concluded: “Given what authorities thought they
knew, we can agree with the assessment that there was exceptional preparatory effort at all levels”
(FR 21).
In other words, the real lesson here – lost to many – is that the pre-landfall preparation saved lives.
We can only imagine what would have happened if the authorities had been as unprepared,
uninterested and uncaring as they were made out to be in hindsight. There would not have been any
warnings, no evacuation, no pre-staging of boats and medical teams. The 60,000 deaths of the
Hurricane Pam scenario may well have become reality.
In hindsight, we know that the preparatory actions were not sufficient in light of the immense scale
of destruction caused by Katrina, which we have qualified as a Black Swan event. Research tells us
that Black Swans will happen and cause surprise, even if you stare them in the face. The question,
then, is: what can we expect when something truly unexpected happens?
The second lesson here is that authorities may be well prepared, have days advance notice, and still
be surprised and overwhelmed. In other words, we will have to get used to the idea that “normal”
disasters can develop into super disasters. This means that detection and sense-making become
important conditions for a timely and effective response.
Detection and sense-making
Once a Black Swan materializes, effective sense-making is both critical and very hard to organize.
Katrina simply confirmed what research has been telling us. On the day of landfall, authoritative
reporting from the field was extremely difficult to obtain because of the widespread destruction of
38
communications infrastructure, the incapacitation of many State and local responders, and the lack
of Federal representatives in the city. Even the Department of Defense lacked situational awareness
of post-landfall conditions (Rep-4).
As a result, local, State, and Federal officials were forced to depend on a variety of conflicting reports
from a combination of media, government and private sources, many of which continued to provide
inaccurate or incomplete information throughout the week, further clouding the understanding of
what was occurring in New Orleans. Collective sense-making failed.
Some organizations did manage to generate a fairly accurate picture of the situation (think of the
Coast Guard and the informal National Guard hub at the Superdome). These emerging information
nodes did not depend on the system and its sense-making capacities; they occurred in isolated
pockets where sense-making was decoupled from the system.
What can be done to improve collective sense-making? The post-disaster inquiries understandably
pointed towards improving sense-making centers. One of the recommendations (FR36), for instance,
was to “establish a National Operations Center to coordinate the National response and provide
situational awareness and a common operating picture for the entire Federal government”. This new
Center would “combine and co-locate the situational awareness mission of the Homeland Security
Operations Center (HSOC), the operational mission of the National Response Coordination Center
(NRCC), and the strategic role currently assigned to the Interagency Incident Management Group
(IIMG)” (FR69).
Our research suggests a different approach. We agree that specialized sense-making units are part of
the solution. But it is critical that such centers are able to detect and connect to emerging
information hubs. It is, in other words, critical that these centers do not depend on bureaucratically
organized streams of information, which are likely to malfunction during a disaster.
Fault lines in coordination
Well after the crisis, the scathing criticism would focus on the lack of coordination and the results
flowing from that absence. The conceptualization of coordination in these critical reports is rather
crude: it does not distinguish between horizontal and vertical coordination, nor does it differentiate
between orchestration and collaboration. In addition, the analysis of causal factors is simplistic at
best, blaming individuals (“Brownie”), organizations (FEMA/DHS) or structures (“the federal
response”). These factors are then juxtaposed against heroes (General Honore) or heroic
organizations (the US Coast Guard).
We advocate a more fine-grained and theory-based approach. By differentiating between
orchestration and collaboration, we identified where the “pain” was in plans that aimed to facilitate
a coordinated response. By separating horizontal from vertical coordination, we may encounter
problems but also strengths.
39
Immediately after Katrina, when new hurricanes threatened the south, the key lesson learned
emphasized the importance of centralization. Chertoff created a raft of new positions not envisioned
in the National Response Plan (CB298). The centralization reflex manifested itself in the micro-
management of the states, which was not appreciated by the governors (CB271). Even after time for
reflection, which allowed for a study of the lessons learned by disaster researchers, political reports
reflected a desire for centralization in the apparent belief that a concentration of authority will make
for a more effective response.
This will not work, as research findings have made clear time and again. The literature advocates a
bottom-up approach, which means that those closest to the disaster formulate requests for
assistance, which are organized and provided by those managing the network. Academics agree that
a top-down approach – where a manager directs all interactions between multiple parties – cannot
and will not work.
What also does not work is differentiating between a “normal” disaster and a “catastrophic” disaster.
Creating different types of responses for different types of disasters is asking for confusion and
bureaupolitics. Different plans and different approaches require that people understand the
difference and can recognize when one type has evolved into another. If that does not happen,
different hierarchies will exist next to each other, which, in turn, requires additional coordination.
The lesson here is not simply “you get what you pay for” (rep-158), as the mantra in the disaster
community had it. The lesson is that the US did not have a proper structure to coordinate, both
horizontally and vertically, a large-scale response network. A better approach is to search for
coordination nodes – formal and informal – that work and then build coordination structures around
those working nodes (Boin and Bynander, 2015). This is a very different approach, which requires a
flexibility to deviate from bureaucratically layered processes.
40
References
Boin, A. and Bynander, F. (2015) “Success and failure in crisis coordination” Geografiska Annaler, 97,
1: 123-135.
Boin, A. and Renaud, C. (2013). Orchestrating Joint Sensemaking Across Government Levels:
Challenges and Requirements for Crisis Leadership. Journal of Leadership Studies. 7(3): 41-46.
Boin, A., ‘t Hart, P., Stern, E. & Sundelius, B. (2005). The politics of crisis management: Public
leadership under pressure. Cambridge: Cambridge University Press.
Boin, A., McConnell, A., & Hart, P. T. (Eds.). (2008). Governing after crisis: the politics of investigation,
accountability and learning. Cambridge: Cambridge University Press.
Boin, A., Kuipers, S. and Overdijk, W. (2013). Leadership in Times of Crisis: A Framewrok for
Assessment. International Review of Public Administration. 18(1): 79-91.
Bovens, Mark and Paul ‘t Hart 1996. Understanding Policy Fiascoes. New Brunswick: Transaction.
Brinkley, D. (2006) The Deluge.
Catino, M (2013) Organizational Myopia: Problems of Rationality and Foresight in Organizations.
Cambridge: Cambridge University Press.
Clarke, L. B. (1999). Mission Improbable: Using Fantasy Documents to Tame Disasters. Chicago:
University of Chicago Press.
Coates, J. (2012) The Hour Between Dog and Wolf. London: Penguin Press.
(CB) Cooper and Block (2006) Disaster: Hurricane Katrina and the Failure of Homeland Security.
Derthick, M. (2007) “Where federalism didn’t fail”, Public Administration Review, 67 (IS): 36-47.
FR – Federal Report
Goidel, Robert and Andrea Miller 2009. “News Organizations and Information During a Natural
Disaster: Lessons from Hurricane Katrina.” Journal of Contingencies and Crisis Management 17 (4):
266-273.
Kahneman, D. (2011) Thinking fast and slow.
Kahneman, D., & Klein, G. (2009). Conditions for intuitive expertise: a failure to disagree. American
Psychologist, 64(6), 515. DOI: 10.1037/a0016755
Kettl, D. F. (2003). Contingent Coordination Practical and Theoretical Puzzles for Homeland Security.
The American Review of Public Administration, 33(3), 253-277. DOI: 10.1177/0275074003254472
41
Leonard, Herman B. and Arnold Howitt 2009. Managing Crises Responses To Large-Scale
Emergencies. Washington, DC: CQ Press.
Lu, X. (2014) Managing Uncertainty in Crisis (dissertation Utrecht School of Governance).
McCreight, R. (2015) “Constitutional authority in crisis: Examing a state governor’s emergency
powers”, Homeland Security & Emergency Management.
Miller, A., Roberts, S. and LaPoe, V. (2014) Oil and Water: Media lessons from Hurricane Katrina and
Deepwater Horizon. University Press of Mississippi.
Perrow, Charles 1984. Normal Accidents: Living With High Risk Technologies. Princeton, NJ: Princeton
University Press.
Rep – House of Representatives, A Failure of Initiative
Rosenthal, Uriel, Arjen Boin, and Louise Comfort eds. 2001. Managing Crises: Threats, Dilemmas,
Opportunities. Springfield, IL: Charles C. Thomas.
Rosenthal, U Charles, M. T., & 't Hart, P. (Eds.) (1989). Coping with crises: The management of
disasters, riots and terrorism. Springfield, Il: Charles C. Thomas.
SS – Senate Report, Hurricane Katrina: A Nation Still Unprepared
Tetlock, Philip 2005. Expert Political Judgment: How Good is It? How Can We Know? Princeton, N.J.:
Princeton University Press.
Tierney, K. (2006) “The Red Pill”, Understanding Katrina: Perspectives from the Social Sciences.
Retrieved from: http://understandingkatrina.ssrc.org/Tierney/
Turner, B.A. (1978) Man-made Disasters. London: Wykeham.
Van Heerden, I. (2006) The Storm.
Weick, Karl E. (1995) Sensemaking in Organizations. Sage