Learning from the storm - Analysing FEMA’s learning-from-disasters process between 2005 and 2013

(1)

Master Thesis

Crisis and Security Management

Public Administration

Supervisor: Drs. G. M van Buuren Second reader: Prof.dr. E. Bakker

March 2016

LEARNING

FROM THE

STORM

Analysing FEMA’s

learning-from-disasters process

between 2005 and 2013

(2)

“By seeking and blundering we learn.”

(3)

Preface

In 2005 the world was shocked to find out that the United States could fail so miserably in the response to Hurricane Katrina. Eleven years later, it is still hard to envision that a modern, Western country did not learn from its earlier mistakes and experiences with crisis management. Perhaps it is this age of terrorism threats and cyber-attacks – that makes citizens and governments alike forget about the destructions that nature can wreak. Even living in the Netherlands, where the threat of high sea levels should be tangible and retrievable, people do not worry much about rising water or storms. It is therefore even more important that governmental emergency response organizations are ready for whatever threat comes their way.

Learning from previous experiences remains a crucial factor in the preparation for the next disaster. Crises may be unexpected – but they should never come as a surprise.

(4)

List of figures and tables

Figures

Figure 1 – Research Model ... 27

Figure 2 - National Hurricane Centre forecast August 26, 2005 - 5PM EDT ... 33

Figure 3 - Breaching of New Orleans levees ... 35

Figure 4 - Main Findings Select Bipartisan Committee ... 39

Figure 5 - Policy Coordination Framework for Response ... 43

Figure 6 National Emergency Communication Plan - Goals ... 49

Figure 7 - Sandy Track Map ... 61

Tables

Table 1 - Internal Information and communication ... 97

Table 2 - Coordination from the responsible government level ... 98

Table 3 - Cooperation between emergency response organizations ... 100

(7)

Introduction

Until recently natural disasters were seen as acts of God that could not be avoided. Once in every while a hurricane or flood would take a country by surprise, leaving a trail of destruction (Farazmand, 2007). Nowadays – the surprise of disasters is downplayed andthe belief that the effects of these kind of incidents can be mitigated prevails. Instead, the focus of governments seems to lie with man-made disasters like armed conflicts.

Unfortunately, there is ample evidence of the destructiveness and the risk that natural disasters still pose. The 2004 tsunami in the Indian ocean, Hurricane Katrina in 2005, the 2009 Australian bushfires, the Haiti earthquake and Pakistan floods – both in 2010 – and the earthquake that shook Japan in 2011 are all examples of natural disasters that have immediate and “incredibly far-reaching consequences for the safety and wellbeing of individuals and communities” (Futamura et al, 2011). Natural disasters do not distinguish between poor or rich countries: even the most wealthy countries can be severely hit by a hurricane, bushfire or earthquake.

This became evident in August 2005, when several countries in the Mexican Gulf region fell victim to the emergence of a destructive hurricane named Katrina. The United States was hit hard: the hurricane had ramifications in over 25 US states, ranging from extreme rainfall and flooding to power cuts (Wise, 2006). The government’s crisis management actions before and during this disaster, both on the federal and on the individual state level, were found to be seriously lacking (Farazmand, 2007; Wise, 2006). Different evaluations pointed to several lessons that needed to be learned. Eventually, Katrina led to a reorganization and change in strategy of the federal organization responsible for crisis management activities in response to overwhelming disasters: the Federal Emergency Management Agency (from now on: FEMA) (Roberts, 2006).

In 2012 the US was hit by another natural disaster, called ‘superstorm Sandy’. At first sight the crisis management activities surrounding ‘superstorm Sandy’ seem satisfactory, which could imply that FEMA has learned the lessons it took away from Katrina. But is this really the case? Evaluations often portray events in a specific light. The evaluations and reports regarding FEMA’s emergency response to Sandy might not reflect reality, but instead display the federal emergency response in too positive a light - regardless of the actual success or failure of FEMA’s efforts.

(8)

Learning from a disaster is interesting from both a practical- and a theoretical point of view. Failures in the preparation for or response to a disaster can lead to very serious situations, involving constant public scrutiny and a lot of pressure on emergency managers. Disasters are therefore seen as either catalyzing learning – or impeding it. On the one hand, the public attention and pressure could offer a ‘window of opportunity’ for a better crisis management strategy. On the other hand, this pressure could lead to hasty decision making. Politicians want to seem engaged and ‘take action’ – sometimes regardless of the possible negative effects of such impromptu decisions (Birkland, 2009).

There is still not much empirical research on the implementation of ‘lessons learned’ after a disaster. Current research mainly focuses on the crisis management of one specific natural disaster. This leaves out a very important feat of learning: lessons that are taken from an disruptive event also need to be implemented. If they are not implemented, these evaluations and lessons learned documents are more to the likes of what Birkland called ‘fantasy documents’, than the beginning of actual learning.

Instead of focusing on the event itself, researching the learning-process in the aftermath of a disaster can provide new information regarding the complete learning-from-disasters process. This knowledge might shed light on the phenomenon of mistakes that have been identified – but still recur at the next disaster. Researching the learning from disasters process could show the underlying mechanisms responsible for actual learning.

A case study research of the implementation of lessons learned after a big disaster could offer interesting new insights in the field of disaster management. If it can be shown at what point in time or with regard to which subjects lessons are implemented and therefore learned, this could provide pointers to the factors that lead to – or impede successful learning from a disaster. As no rational organization would want to make the same mistakes twice, this information could greatly benefit both governmental- and corporate emergency managers.

Research Question

This thesis will research the learning process within FEMA, starting from the evaluation of Hurricane Katrina in 2005. This research requires a multi stage research design. First it needs to be established if there was a shift in policy concerning the management of

(9)

natural disasters in the United States that can be attributed to the failed response to Katrina. Second, the emergency response to Sandy will be researched, to see if the ‘lessons learned’ from Katrina were actually implemented. This leads to the following research question:

To what extent have lessons been learned by the United States’ Federal Emergency Management Agency after their failed response to Hurricane Katrina in 2005 - as shown in

their response to Superstorm Sandy in 2012?

To answer this research question the following sub-questions will be answered. The first two sub-questions will be answered in the theoretical framework:

1. What does an emergency response entail? 2. How can an organization learn from disasters?

The remaining sub-questions analyze the different steps in the learning-from-disasters process, and will be answered in the chapters that portray the gathered data:

3. What were the main issues in the emergency response to Hurricane Katrina? 4. What changes in crisis management policy and policy practice can be observed

in the United States after Hurricane Katrina?

5. Which lessons were implemented and therefore learned by the U.S. Federal

Emergency Management Agency before and during Superstorm Sandy in 2012?

Structure of the thesis

In the second chapter, the first two sub-questions will be answered. In order to provide a theoretical background, it provides an overview of the current state of crisis management literature. It explains what crisis management entails and which policy actions are expected based on these theories. This is followed by an overview of the existing literature on organizational learning, focusing on learning from incidents and learning from disasters.

Chapter three will outline the research design. This involves a single case-study of FEMA – the agency responsible for emergency response in the United States. The analytical approach that is chosen suits the research purpose, as it allows for within-case inferences about a theorized mechanism. To test whether this mechanism is present, the

(10)

different steps that together form the learning process will be operationalized. Furthermore, this chapter describes the process of data collection and analysis.

Chapter four analyzes the crisis management activities during Katrina. The gathered data on Katrina provide information about the skills that FEMA possessed at the time of the emergency response to Hurricane Katrina. This will answer the third sub question.

The next chapter discusses the adjustments made in this regard after Katrina. It portrays the proposed policy changes, and the implementation of these changes in the period before Superstorm Sandy hit in 2012. This answers the fourth sub question.

In the following chapter, the FEMA response to hurricane Sandy will be researched. This will shed light on the implementation of lessons learned that were identified after Katrina – answering the fifth sub question.

In the next chapter the findings on both Katrina and Sandy will be analyzed. The final, concluding chapter reflects on the entire research and answers the research question. It offers policy recommendations regarding the managing of natural disasters like hurricanes and suggestions for further research.

(11)

Theoretical Background

This chapter presents an overview of the most important schools of thought regarding crisis management and how these theories are entrenched in the broader theoretical debates of crisis management theory. This is followed by an overview of the literature on organizational learning, focusing on the learning-from-incidents literature. These two schools of thought combined offer the opportunity to research the ‘learning from disasters’ process.

Defining disasters

Ontologically speaking, defining the term ‘crisis’ is not possible: “(t)he core of crisis is precisely the fact that an event, a dynamic, does not fit into the conventional references, formats and codes – and moreover, is threatening to destroy those very references, formats and codes.” (Topper & Lagadec, 2013: 8). Moreover, “(c)rises and disasters are ‘inconceivable threats come true’ – they tax our imagination and outstrip available resources.” (Boin & ‘t Hart, 2010: 358). The crisis management literature therefore often describes this as a ‘lack of definition’, instead of an impossibility (Roux-Dufort, 2007; Parker et al, 2009). To identify the phenomenon under scrutiny, it is however useful to use a working definition.

A crisis can be seen as “an interruption in the reproduction of economic, cultural, social and/or political life” (Johnston, 2002 in Hagar, 2010). This covers a broad array of events, from natural disasters, to human-made disasters like terrorist attacks or financial crises. The focus of this thesis lies on the concept of a natural disaster. Disasters are “largely unexpected and unavoidable events.” (Putnam, 2002). Also described as “a sudden calamitous event bringing great damage, loss, or destruction” (Meriam Webster, n.d.), their suddenness is described as an important feature.

A crisis or disaster is often defined as a singular event. This may however not do justice to the concept. Theorizing about the subject is difficult, as “(t)his position gives it a singularity and a contingency that distance it from more structuring models and that make replication and generalization of research results difficult.” (Roux-Dufort, 2007: 106). To understand the reasons and mechanisms of a crisis, approaching it exclusively from the angle of the incident is not enough. Most definitions of a crisis make a

(12)

distinction between ‘crisis’ and ‘normal situation’ - but have little attention for the grey area in between. Crises are seen as unexpected, unpredictable and surprising events. But the “intermediate states between normality and crisis (...) would allow us to approach a crisis as a process of accumulation of deficiencies and weaknesses rather than as a sudden and extraordinary irruption.” (ibid: 107).

It can be said that organizations cause their own crises, for example by a lack of preparation. It is therefore necessary to identify which processes put organizations in a vulnerable position on a regular basis. Therefore the occurrence of a hurricane is not a crisis in itself, and does not cause a crisis. The hurricane is the event that brings the underlying crisis to light. This impacts the way that crises are researched: the triggering event, like a hurricane, should not be the starting point of an investigation. To gain a complete picture, the (lack of) preparation beforehand should also be considered.

The question of the singularity of a disaster impacts the surprise regarding this phenomenon. Surprises can be defined as “unexpected events, which appear unknowable, unpredictable and unmanageable.” (Aradau, 2014: 79). Natural disasters used to be seen as acts of God or the gods, that could not be avoided (Farazmand, 2007; Topper & Lagadec, 2013). This shifted to a concept of crisis and disaster where preparing for these phenomena has become part of our lives (Farazmand, 2007). Although a disaster is defined as an unexpected event - it’s existence is not as surprising. Even when the scale and the exact timing remain to be seen - natural disasters are usually within the scope of our imagination and preparatory actions can be taken. A crisis or disaster is not a surprise per se: they can be avoided, eliminated, or controlled (Aradau, 2014).

But even when they are somehow expected, natural disasters can still lead to sudden policy failures. The surprise literature that covers this subject sees several underlying reasons for this kind of failure: surprise, unpreparedness and the failure of advanced warning. “The existence of one or more of these factors can leave one susceptible to an acute policy failure.” (Parker et al., 2009: 207). Overcoming warning-response problems is therefore of the utmost importance to prevent sudden policy failures. “Falling prey to a warning-response failure generally indicates a malfunction in one or more links along a complex chain of policy, intelligence, warning, and response.” (ibid.). Even if the warning-response problem can’t be eradicated and natural disasters

(13)

remain a huge vulnerability - preparation and mitigation can still make a huge difference with regard to their impact and destructiveness.

The ability of a community to cope during this event is also a defining feature of a disaster. When a community cannot cope with the event - the hazard becomes a disaster. The International Federation of the Red Cross defines a disaster as “a sudden, calamitous event that seriously disrupts the functioning of a community or society and causes human, material, and economic or environmental losses that exceed the community’s or society’s ability to cope using its own resources” (IFRC, n.d. a). This definition leaves out the question whether the disaster was a surprise or could have been prevented, instead focusing on the fact that the event harms, and overwhelms a specific population. It is this definition that will be adopted in this thesis, as it fits the research goal of exploring how to learn from these events. In this thesis a disaster will be consider as a ‘given’: even if it could have been prevented – the event did happen, invoking crisis management activities.

Crisis management

The unexpectedness of a natural disaster plays into the preparation for it. Technological improvements enable the gathering of more information about a gathering storm, which improves the possibilities for early warning. Whether or not crisis managers heed those warnings depends on the warning-response gap: the discrepancy between information and action.

Crisis management can be defined as “the discipline dealing with risk and risk avoidance.” (Haddow et al., 2013: 1). As shown above, disasters are the cause of widespread harm. Crisis management aims to either prevent this harm, or to minimize its impact and restore the order. Systemised strategies are developed, but learning from previous experience proves difficult: “the infrequency with which disasters occur makes it hard for responders to test and improve their strategies, to ensure that they can be counted on to mitigate threats and hazards predictably and to resolve their consequences effectively.” (Donahue & Tuohy, 2006: 3). Crisis management used to be viewed from a rather technical perspective. It would focus on predictable faults and failures. The solution to these ‘crises’ would simply be to “train technicians to intervene in a timely and effective manner, according to predefined procedures and based on time-proven expertise.” (Boin & Lagadec, 2000:186). But modern crises have too many

(14)

converging factors, and require more than just these technical solutions. Nowadays, crises (can) involve “large systems consisting of entangled networks with a hitherto unknown complexity, the immediate mediatisation of incidents, abrupt changes in collective perceptions and social demands suddenly condemning what was hitherto tolerated.”(Boin & Lagadec, 2000: 186).

The Department of Homeland Security (DHS) of the United States defines emergency management as “the coordination and integration of all activities necessary to build, sustain, and improve the capability to prepare for, protect against, respond to, recover from, or mitigate against threatened or actual natural disasters, acts of terrorism, or other manmade disasters.” (DHS, 2008 in Baird, 2010). This definition mentions several distinct phases in crisis-management (or emergency management) - each with its own particular (policy) actions. The different phases are contested: different organizations and researchers distinguish a different number of phases or call them by different names (Baird, 2010).

In this thesis, four phases will be distinguished: mitigation, preparation, response and recovery. These four stages are widely recognized in the field of crisis management (Baird, 2010; Schneider, 1995; Rosenthal and Kouzmin, 1993). They are also mentioned in the United States government’s disaster response process, which mentions four objectives “(1) mitigating a disaster or preventing one from occurring in the first place, (2) preparing areas for potential emergency situations, (3) providing immediate relief after a disaster strikes, and (4) helping individuals and communities recover from the effects of natural disasters.”(McLoughlin, 1985 in Schneider, 1995). A model with four distinct phases does not reflect the exact reality of crisis management, as these four phases do not always occur in succeeding order: “while responding to a crisis, chief executive officers may be particularly keen on measures mitigating the impact of future crises. Moreover, each phase may lend itself to various specifications.”(Rosenthal & Kouzmin, 1993: 6). However, the theoretical distinction of these phases enables the academic study of crisis management.

Preparation

As shown above, crises tend to be unexpected and unique. Preparing well for the unknown seems impossible and like a contradiction in terms - but this is exactly what crisis management entails (Boin & McConnell, 2007: 53). Both foreseeable and

(15)

unforeseeable disaster situations may occur - and societies need to be prepared for all of them. Preparing for a disaster “is the process of identifying the personnel, training, and equipment needed for a wide range of potential incidents, and developing jurisdiction-specific plans for delivering capabilities when needed for an incident.” (DHS, 2008 in Baird, 2010).

Planning for a crisis is by definition limited, as crises are unique. Standardized solutions may not work, as effective crisis responses “are necessarily improvised, flexible and networked (rather than planned, standardized and centrally led). They are driven by the initiative of operational leaders and the strength of the pre-existing ties between the teams and organizations they represent. Any attempt on the part of strategic decision-makers to plan and command each and every aspect of crisis response impedes flexibility and local initiative” (Boin & ‘t Hart, 2010: 362). Therefore, preparation requires assessing future risks. This assessment influences the kind of precautions that are taken and the scale of these preparatory actions.

Response

Inter-organizational relations are valuable to create systemic resilience. But during the response phase of crisis management trust issues and troubles with the sharing of information are, unfortunately, still common (Boin & McConnell, 2007; Chen et al., 2013). As the response phase is characterized by “immediate actions to save lives, protect property and the environment, and meet basic human needs” (DHS, 2008 in Baird, 2010: 18), there is a focus on speed and efficiency. This may cause “high levels of stress and uncertainty and disruptions in the operations of critical communications and management systems.” (Chen et al., 2013: 132). Response activities include evacuation, offering shelter and first aid - all of which should have been foreseen in emergency plans. Importantly: it requires mobilization and transfer of resources to the impact area.

There is some discussion in disaster literature regarding the exact starting- and ending point of response efforts. The line between response and preparation or recovery activities remains rather vague. Haddow (2008) states that “the response function is classified as the immediate actions to save lives, protect property, and meet basic human needs. The recovery function (...) often begins in the initial hours and days following a disaster event and can continue for months and, in some cases, years, depending on the severity of the event.” (Haddow, 2008 in Baird, 2010: 23-24). Baird

(16)

notes that the beginning and ending of phases cannot be expressed in ‘clock time’, but should be determined by the performed activities. The boundaries of the stricken area, or impact area, differ both per storm and per definition. Areas and communities can be affected either directly or indirectly. Palen and Liu (2007) discern the impact area which is hit directly by the disaster; the fringe area that is indirectly hit; and the filter area where resources like equipment and information are staged (Palen and Liu, 2007: 730-731).

Then there is the community aid zone and the regional aid zone. These are the stage for community efforts and government services like shelters. This is important, because the event will be known to a much wider public than just the people directly affected. As natural disasters are often highly visible, they are “exactly the type of event that the media prefer to spotlight. They are unusual, different, spectacular, and at times even horrific.” (Schneider, 1995:15). This media attention allows for more public scrutiny. This makes the boundaries of a natural disaster an important point of attention for (political) decision makers. After hurricane Katrina, evacuations caused the scattering of the population across the United States. “The community zone of effect (...) was therefore extended into other US states, where resources like housing and school education were heavily taxed. The regions of community and regional aid were blurred and extended effectively across the entire reach of the US.” (Palen & Liu, 2007: 731). This resulted in a different experience of the different stages of crisis management across communities.

Public private partnerships

Building resilience does not necessarily involve just the government and citizens. To create systemic disaster resilience, it can be beneficial to create public-private partnerships, as governments lack the resources and capabilities to handle wide-spread natural hazards (Boin & McConnell, 2007; Chen et al., 2013). Different phases of crisis-management know different actions with regard to both citizen involvement and the type of public-private partnerships (from now on: PPP’s) that are at play. Chen et al. distinguish contractual and non-contractual partnerships. The first is usually seen in relation to large construction projects when rebuilding after a natural disaster. Non-contractual partnerships often facilitate the coordination of inter-sectoral resilience-building - generally seen in the preparation phase of crisis management - and “provide

(17)

overarching institutional structures for information sharing and inter- and intra-sectoral policy dialogues.” (Chen et al., 2013: 134).

Information logistics

During crisis management activities, the gathering, storing and sharing of information is of the uttermost importance (Putnam, 2002; Hagar, 2010). To coordinate different actors, real-time, accurate situational awareness is crucial. Ideally, coordination needs a steady flow of information between different government levels - and between government organizations and private actors. This applies to both search-and-rescue missions and the marshalling of resources and capabilities, like medicine or aid workers. Unfortunately a crisis does not offer the ideal circumstances: there are power cuts, telephone lines may be broken by fallen trees, (public) transport might have come to a stop when the roads and tunnels are flooded. Preparation and mitigation need to either prevent- or deal with these challenges.

While basic connectedness or operability can already prove challenging during crises, the introduction of new technologies is often accompanied with the problem of interoperability. New communication devices as simple as telephones might not be compatible with the devices used across the state- or county-border. And even when all the technology is functioning, information or misinformation will still circulate and propagate. A decision made by a leader at the top of the decision chain can be completely misinterpreted “leading to incoherent, conflicting and sometimes absurd action.” (Topper & Lagadec, 2013: 13). This is in part why verifying information is of the utmost importance within emergency response organizations.

Recovery

After a natural disaster, the recovery phase starts. The actual beginning of this phase is sometimes referred to as ‘after the response phase’. This signifies an important feat: the activities in the recovery phase are similar to those in the response phase. However, the difference does not just lie in the sequence of events: this is the phase that is aimed at returning ‘back’ to a normal state of affairs (Schneider, 1995). This phase includes a broad range of activities, from developing restoration plans and reinstating government services, to starting housing programs and providing long-term care for affected persons (DHS, 2008 in Baird, 2010). Overall, the response phase is defined more broadly and

(18)

abstract: “(t)he descriptions of other phases rely on active terms such as “activities,” “tasks,” and “actions,” but the recovery phase is described more in terms of goals and desired content.” (Baird, 2010: 15). Also, it could be said that in the response phase “the sense of urgency is diminishing.” (Chen et al, 2013: 132).

The goal of returning to a normal state of affairs after a disaster has struck, is neither undisputed nor easy. The destruction of entire neighborhoods by a natural disaster offers opportunities to several involved actors. Most emergencies cast a long shadow in terms of the politics of crisis management: recovery activities often take place in an hostile climate of trauma and recrimination (Boin & ‘t Hart, 2010; Chen et al, 2013). Questions about accountability start to rise and the process of account-giving starts: “managing the process of expert, media, legislative and judicial inquiry and debate (…) in such a way that responsibilities are clarified and accepted, destructive blame games are avoided and a degree of catharsis is achieved” (Boin & ‘t Hart, 2010: 359). This is also the time for evaluation of the crisis management performance.

Mitigation

Mitigation, the act of preparing for possible (natural) disasters in the long run, is a continuous process. An example of mitigation is raising awareness of natural hazards in schools, or planting mangrove trees to reduce the risk posed by tidal surges (IFRC, n.d. b). The difference between mitigation and preparedness, is that mitigation tries to limit the impact of future hazards in general - where preparation covers a specific (bound-to-happen) threat. This is also shown in the definition the DHS uses for mitigation, which states that mitigation involves “activities providing a critical foundation in the effort to reduce the loss of life and property from natural and/or manmade disasters by avoiding or lessening the impact of a disaster and providing value to the public by creating safer communities. (...) These activities or actions, in most cases, will have a long-term sustained effect.” (DHS, 2008 in Baird, 2010: 16). Engineering techniques and hazard-resistant construction also fall under mitigation, as well as improved environmental policies and public awareness of possible hazards (UNISDR, n.d.).

Preparation and mitigation can work within different epistemic regimes. From a ‘risk and uncertainty’ perspective, risks can be modeled and risk assessments are based on statistical information about the frequencies of a class of events (Aradau. 2014: 77). This framework accepts that risk cannot predict an individual event: “Contingency is

(19)

tamed through the move from the individual to the multiple and from uncertainty to risk.” (ibid.). On the other hand, the epistemic regime of ‘surprise and novelty’ assumes that an event - in this case a disaster - is always a possibility: “Preparedness and resilience are the answers to the surprising event and its emergent novelty. Contingency is not tamed, but incorporated, literally lived with.”(ibid). Instead of trying to estimate the risk of something happening - this framework prepares for the event because it is always a possibility. “Prevention requires that one knows the source and dynamics of threats, but the literature shows that this is impossible for most if not all organisations.” (Boin & McConnell, 2007: 52). This is where resilience comes into play.

The term resilience refers to groups and societies capability to ‘bounce back’ after a disastrous event: humans have an astounding capacity to reorganize and adapt. Although resilience means that citizens can take care of themselves - Boin and Lagadec do see the necessity of organizing for resilience. They assert that preparation (or mitigation) activities should include resilience-building, as this allows for a “rapid, flexible, innovative and effective response when a future crisis presents itself.” (Boin & Lagadec, 2000:188). Especially during the immediate aftermath of large natural disasters, an effective response critically depends on the resilience of citizens, first-line responders, and operational commanders (Boin & McConnell, 2007: 54). Mitigation can help build resilience amongst communities and individual citizens. An example is schooling programs about hazards, that will help the public to more readily understand a warning when it is issued (Mileti, 1995: 2). Resilience aims for groups or individuals to be continuously prepared for a disastrous event.

Another form of resilience is systemic resilience, which refers to the crisis response capacity of a society. Boin & ‘t Hart (2010) argue that this resilience is dependent on the quality of inter-organization relations, as major crises are almost invariably tackled by networks of organizations. These networks cut across disciplinary, jurisdictional and public-private sector boundaries (Boin & ‘t Hart, 2010: 365-366). Coordinating a network of organizations comes with its own set of problems: pivotal actors accord it low priority, the divide between fulltime and volunteer crisis management workers can prove difficult, and there are differences in lines of accountability and decision making cycles (Boin & ‘t Hart, 2010; Boin & McConnell, 2007). As systemic resilience is dependent on inter-organizational relations, mitigation

(20)

can be of use in this process. The ties in ‘crisis networks’ need to be strengthened on a regular basis.

Resilience is a relatively new term in security and disaster literature (Aradau, 2014). The term has long been seen as answering the problem of vulnerability or “the problematization of future events as surprises” (ibid.:75). Aradau argues that the term resilience enables a shift in focus “away from the quantitative availability of resources, and towards the scope of available response options” (Cole & Nightingale in Aradau, 2014: 76). The resilience perspective thus changes the problematisation of a possible disaster: seeing disasters as inevitable presupposes that prevention will not always be an option. In other words from a government perspective “the promise of security that underpins the liberal state is subtly rephrased – ‘we may not be able to protect people’.” (Aradau, 2014: 82). Building resilience “requires long-term commitment to safety in the absence of immediate salient threats and a willingness to invest in resources” (Chen et al., 2013: 132). This necessitates mitigation activities.

Learning from disasters

Because crises and disasters are highly visible events they attract a lot of attention. Questions of accountability and liability arise, and the experience is generally expected to result in ‘learning’. As mentioned above, this phase offers much potential for another ‘crisis’ and should not be underestimated. Different stakeholders have much to gain or lose from the experience, functioning either as accelerator or as a barrier for learning (Boin & ‘t Hart, 2010, McConnell, 2011). Lovell (1985) argues that organizational lessons are “formulated through a process of negotiation or bargaining (…) the product of an organizational or political dynamic, rather than as the products of the application of logic and pure reason to the past” (Lovell, 1985 in Stern, 1997: 78). This makes it hard to capitalize on the learning potential that is inherent to experiencing crisis situations, and impedes learning.

Following this line of reasoning, it becomes clear that evaluations need to be regarded with a critical eye to the implicit messages they may carry. Although official inquiries “do perform the roles of fact-finder and auditor of crisis episodes, (…) such roles often overlap and conflict with other roles such as insulating elites from blame and controlling the parameters of scrutiny focused on dominant policy systems and sub-systems” (McConnell, 2011: 64). Non-official evaluations can also cause problems. This

(21)

depends largely on the source: political adversaries might publish their skewed views on events, just as lobby groups aim to focus the attention on successes that relate to their own interests (ibid.).

Aside from the political plays that are often involved, crises also offer an opportunity to learn and reform (Chen et al, 2013: 132). Although change and learning are not guaranteed – crises are “are one of the few ways in which established policies, procedures, cultures and legitimacies change course (Baumgartner and Jones, 1993 in Boin & McConnell, 2007: 57). Natural disasters are what Kingdon calls focusing events – events that open a window of opportunity for (immediate) policy change, or “improved understanding of the social or the natural forces that lead to a disaster” (Kingdon, 1995 in Birkland, 2009: 147).

One form of evaluation is a ‘lessons learned’ document, which aims to “prove that some authoritative actor has ‘learned its lessons’ about a disaster and that, given this learning, will not replicate its errors” (Birkland, 2009: 147). Birkland is critical of lessons learned reports after natural disasters, even calling them ‘fantasy documents’. He states that the lessons described in these documents are often not implemented or ‘learned’ at all – even when the writers think they are.

Organizational learning

Much research has been done about the process of learning, both in a scientific setting and in the practice of the emergency management fields (Chen et al., 2013). In a paper comparing research on organizational learning with the research on learning from incidents (or LFI), Drupsteen & Guldenmund define LFI as “an organizational learning process of putting lessons learned from incidents into practice to prevent future incidents” (Drupsteen & Guldenmund, 2014: 82). Organizational learning involves the collection of information, processing information and storing information (Argyris & Schön, 1996 in Drupsteen & Guldenmund, 2014). Incidents are thereby defined as “unwanted and unexpected events within the organization with an effect on safety” (Drupsteen & Guldenmund, 2014: 81). The three main processes in LFI are similar to those of organizational learning, namely “the analysis of events, the use of lessons learned and sharing and storing information” (ibid: 92). There are important differences between the evaluation of an emergency response operation and incident- or accident analysis described above: “accident analysis focuses primarily on what led to the

(22)

accident, while we are primarily interested in the management of the situation given that an accident has already taken place, or some other kind of extreme event has affected society” (Abrahamsson et al., 2010: 16). However, this thesis follows Abrahamsson et al. in their conclusion that a similar model might be used in the evaluation of emergency response operations.

Learning from incidents is influenced by its circumstances. An important factor is the scale and impact of an incident determine the attention it gets. This is true both for external (media) attention, and the attention that an organization subsequently pays to learning processes (Birkland, 2009; Drupsteen & Guldenmund, 2014). Organizations feel a greater need to learn from incidents that have severe consequences. Disasters receive overwhelming publicity, which impacts the learning process in several ways. On the one hand, publicity will motivate organizations to invest in the learning process: disasters can function as focusing events, which “increase the changes that arguments about policy failure revealed by the event will gain attention (Birkland, 2006: 174). The underlying assumption is that this increased attention for policy failure leads to efforts to learn, resulting in efforts to change policy. This is why post-disaster space offers opportunities for catalyzed learning (Chen et al, 2013: 139). On the other hand, this media-attention increases the pressure on policymakers to ‘do something’. Birkland states that under these circumstances “given the haste of the decisions made in the wake of these events, the risk of superstitious learning – that is, learning without some sort of attempt to analyze the underlying problem – is greatest.” (Birkland, 2009: 148). This is particularly relevant when “failing to act would seem to reflect a lack of compassion for the victims of a disaster” (Birkland, 2006: 175). An example is the case where the media portray people who suffer damages from a natural disaster as ‘victims’ – even though these people did not prepare in any way for a well-known risk, or build their houses in an unsafe area like the beach. This portrayal ‘forces’ politicians to offer help to these people.

The opposite problem is when the incentive to learn is lacking because an incident (like a natural disaster) is perceived as rare. Whether or not an incident is thought likely to occur, is in the eye of the beholder: “(e)vents are more likely to be considered “rare” when individuals or organizations that observe or directly experience these events see them as unusual—in the sense that they depart from ordinary experience with the same type of event, or are unique in the sense of having no close

(23)

parallel” (Lampel et al, 2009: 836). This corresponds with the conclusions of the surprise literature regarding the definition of a crisis, and the research on biases. Unfortunately, this factor negatively influences the willingness to learn from these events.

The Implementation of ‘lessons learned’

It should be noted that learning does not happen in a vacuum: the social surroundings are important to its success. Crisis situations offer vastly different surroundings for policy change or policy learning than regular surroundings (May, 1992; Stern, 1997; Lagadec, 1997; Birkland, 2009). As this thesis analyses the changes in emergency response of a governmental organization caused by learning, it is necessary to focus on the definition of learning that fits this subject. Peter May (1992) distinguishes between three types of learning in this regard: instrumental policy learning, social policy learning and political learning (May, 1992). With instrumental policy learning, lessons are learned about the effectiveness and viability of policy tools with regard to a problem. Social policy learning involves learning about the social construction of policy problems and “the interaction of policies with the targets of policies.” (May, 1992 in Birkland, 2009: 147). The third form of learning is political learning. Political learning “entails lessons about policy processes and prospects. Policy advocates become more sophisticated in advancing problems and ideas by learning how to enhance the political feasibility of policy proposals.” (May, 1992: 332).

Based on this distinction, Birkland researched policy learning after crises. He states that “there is prima facie evidence of learning if policy changes in a way that is reasonably likely to mitigate the problem revealed by the focusing event” (Birkland, 2009: 150). This definition presupposes that learning involves the acquisition of new skills, or competence: “the changes in the level of skill exhibited by the actor in question between two or more points in time becomes a key indicator of the degree of learning which has taken place” (Stern, 1997: 71). The development or distribution of material resources to accommodate this, can also be taken into account. Similarly, Moynihan (2008) states that “learning refers to the identification and the embedding of practices and behaviors by the network to improve crisis response.” (Moynihan, 2008: 351).

In order to have conclusive evidence of learning, it is necessary to discern this learning from other kinds of political or policy change that could have taken place in this

(24)

timeframe. A consequence-based approach therefore asks “whether the change in question has increased or decreased utility.” (Stern, 1997: 72). This entails both normatively positive and negative learning: “The question is whether the new information or skills have enabled the actors to achieve their purposes better, regardless of whether the observer likes those purposes or not.”(ibid.). In this thesis, the definition of Birkland is used. Following this definition, policy learning is the implementation of lessons learned in a way that improves the utility of the actors involved.

It is clear then, that after the acquisition of information and the identification of lessons learned, these lessons also need to be implemented. In order to speak of actual learning, it is necessary to create practical recommendations and actions. However, research of both Birkland (2009) and Drupsteen and Guldenmund (2014) conclude that the implementation of lessons learned is seriously lacking. Drupsteen and Guldenmund state that “when lessons are learned, follow-up steps to use the lessons for prevention are necessary, which are often neglected” (ibid.: 94). Birkland asserts that often, learning processes are not ‘serious’: “many of these documents and the processes that create them are mere reﬂections of a group’s or interest’s preferred social construction of a problem and its ‘target populations’” (Birkland, 2009: 154). This results in resisting investigation of disasters; denying or even ignoring its lessons. This is acknowledged by Stern (1997: 78), who names two important obstacles to learning from crises: defensiveness and opportunism. Defensiveness is the tendency to “avoid or suppress information which suggests (...) performance failure” (ibid). The second obstacle Stern discerns is opportunism, which refers to actors who deliberately exaggerate their contribution for strategic reasons (Stern, 1997: 78). This could paint a distorted picture of the chain of events that passed – which in turn impedes information gathering and therefore the learning process.

The literature on organizational learning, policy learning and LFI demonstrates that learning has a prominent place in crisis management. On the one hand, crises offer a major opportunity for learning. On the other hand, the surroundings of a crises can form obstacles to learning. Although learning is commonly featured in disaster literature, it has also become clear that there is not much empirical research on the implementation of lessons learned. By analyzing evaluations and lessons learned documents, this thesis hopes to contribute to this field. It explores what lessons were identified and implemented by the US government after hurricane Katrina.

(25)

Research Design

In this chapter the research design and case selection will be explained. This is a single-case study, researching the learning process of an emergency response organization in the social surroundings of a natural disaster. Furthermore, the data exploitation and data analysis will be discussed.

Case study research

The goal of this thesis is to see whether ‘learning’ from previous emergency response operations takes place. To research the causal mechanism hypothesized in the theoretical background chapter, the Federal Emergency Management Agency (FEMA) will be analyzed. This agency is responsible for the emergency response to overwhelming disasters in the United States. As this involves a single research unit, a qualitative research method is most suitable. In the field of disaster studies, there is not much empirical research on the implementation of lessons learned after a disaster (Drupsteen & Guldenmund, 2014). A case study offers the opportunity to explore the developments in this field with flexibility and open-mindedness (Stebbins, 2001: 9). It is a qualitative research method that provides the opportunity to identify important factors that contribute to a particular outcome (Landman, 2008: 249). Case studies allow for conceptual linkages of incidents and describe the socio-historical context of a case as an explanatory framework” (Goulding, 2002: 18).

The case of FEMA is chosen based on theoretical sampling, “whereby groups or institutions are selected on the basis of their theoretical relevance” (Babbie, 2004: 292). From a theoretical point of view, this agency is interesting because it acts when a specific community is overwhelmed. FEMA’s mission is “to support our citizens and first responders to ensure that as a nation we work together to build, sustain and improve our capability to prepare for, protect against, respond to, recover from and mitigate all hazards.” (FEMA, n.d.). The organization is responsible for coordinating government-wide relief in the United States. Created by President Carter in 1979, it nowadays follows the procedure first established in the 1988 Stafford Act. This legislation designates FEMA to come into action and offer financial and physical assistance after a presidential disaster declaration. An emergency is declared, when the president deems federal

(26)

assistance to State and local efforts necessary to “to save lives and to protect property, public health and safety, or to lessen or avert the threat of a disaster” (Bazan, 2005). In other words, FEMA acts when local and State efforts are overwhelmed.

This operational jurisdiction fits well with the definition of a disaster employed in this thesis. FEMA operates in a very large area with very diverse (natural) disasters: its function is well served with learning from the disasters it has managed previously. Learning literature points out that learning has taken place when an actor has identified and implemented ‘lessons learned’ in a way that improves the intended utility (Stern, 1997: 72).

When conducting a case study, it should be kept in mind that a case study is always an instance of a broader set of cases (Gerring, 2006: 13). Governments have to prepare for and respond to natural- and man-made disasters – knowing that failure in such a test “has brought down governments and triggered regime-changing revolutions” (Farazmand, 2007: 149). With an increase of effectiveness and efficiency in mind, government agencies tasked with emergency response evaluate their actions to extract lessons they need to learn.

This thesis focuses on the process of learning from disasters. With the loss of more than 1800 lives and the displacement of more than 700.000 people, hurricane Katrina was one of the costliest and deadliest natural disaster that ever hit the United States (OIG, 2006). The choice for the case of FEMA learning from hurricane Katrina, is based on this extreme value of the variable ‘impact and scale of the disaster’. Although the definition of a disaster presupposes an overwhelmed community, the costs that were made and lives that were lost due to Hurricane Katrina are yet to be surpassed. This makes the emergency response to hurricane Katrina an extreme case study.

From a methodological point of view, extreme cases are interesting because “the variation that one wishes to explore as a clue to causal relationships is encapsulated in these cases” (Gerring, 2006: 102). FEMA learning from a disaster of the scale of Katrina can be seen as prototypical of the phenomenon of learning from disasters. Keeping in mind that disasters are always overwhelming, the learning process of an emergency organization from an extreme situation could provide new insights in the factors that influence the learning from disasters process.

The learning from disasters literature is still divided in this regard. On the one hand, the enormous scale and impact of Hurricane Katrina could offer the opportunity

(27)

for catalyzed learning. In his 2006 book on learning from disasters, Birkland confirms this choice, stating that

“only the very largest storms (…) get serious attention and have the greatest potential influence on learning. Smaller incidents do not get attention because they are often successfully addressed by existing organizations and policies; Hurricane Katrina got more attention than did all four of the hurricanes that struck Florida in 2004 because the response to the Florida hurricanes was generally perceived as adequate, and because no individual storm was catastrophic, whereas Katrina was a catastrophe that overwhelmed the national emergency management system.” (Birkland, 2006: 19)

On the other hand, the extreme media-attention put the politicians involved under a lot of pressure – increasing the risk of superstitious learning without any attempt to analyze the underlying problem or problems. Legislation that followed shortly after Katrina – the Post-Katrina Emergency Management Reform Act (or PKEMRA) – could be a prime example of this phenomenon.

Research model

In this thesis, the policy actions of FEMA over time will be analyzed, requiring a multi-stage research approach. In the first part, the skill of FEMA will be determined based on the crisis-management activities surrounding hurricane Katrina in 2005. Several key lessons that need to be learned will be identified, and used during the second part to analyze the skill of FEMA in their response to superstorm Sandy. Following the learning from incidents and organizational learning theory, this leads to the following

preliminary research model:

Establishing ‘learning’

To establish whether or not an organization has learned from a natural disaster, an in-depth study of their activities at more than one point in time is necessary. First the organization’s level of skills at the initial point in time needs to be identified. Second, it

Failed emergency response

Collecting

information Processing information

Embedding proposed changes

Implemen-tation Emergency response Figure 1 – Research Model

(28)

needs to be established whether these skills have changed over time. To distinguish the change in skill as a result of learning from other external processes, it should be possible to trace back these changes to an active pursuit of change. After the gathering of information and the processing of information, the lessons learned that are gathered should be embedded in new policies and procedures aimed at improving the utility of FEMA. These policies should acknowledge and reflect the lessons of the experience of the ‘failed’ disaster response (Stern, 1997: 81).

To explore whether FEMA learns from the preparation for- and response to a disaster, a longitudinal single case study will be conducted. This study will focus on the emergency response of FEMA to natural disasters, with two hurricanes as units of analysis: Hurricane Katrina in 2005, and Superstorm Sandy in 2012. The choice for superstorm Sandy is based on the same variable that directed the choice for hurricane Katrina: scale and impact of the disaster. This variable of interest needs to be similar to enable a valid comparison, which is the case with these natural disasters. In the emergency response to Katrina, FEMA needed its entire skill set. Comparing this incident to a much smaller disaster would offer no conclusive evidence regarding the improvement of these skills.

In the period after 2005, there were several disasters that required federal assistance. Superstorm Sandy - like Katrina - was an unusual natural disaster in terms of scale and impact. Where the stronger category 1 Hurricane Katrina impacted 15 million people in 16 states – the much larger post-tropical storm Sandy impacted 19.4 million people in 17 states (Mesenbourg, 2012). Both hurricanes impacted large metropolitan areas, located by the sea: New Orleans, Louisiana in 2005 and New York City, New York in 2012. More importantly: both natural disasters involved presidential emergency declarations that required the involvement of FEMA in many states.

Katrina – collecting and processing information

The first part of this thesis employs both secondary research (desktop research) and primary research (document analysis of primary sources). The key skills that were lacking in the crisis-management of Katrina are identified based on the findings of an independent evaluation of the federal response; supplemented with evaluations and ‘lessons-learned’ documents of other (governmental) organizations involved.

(29)

Evaluations and lessons-learned documents can be categorized as both primary- and secondary sources. The goal of this thesis is to research to what extent FEMA has learned from Hurricane Katrina. Therefore, the skills of FEMA during Katrina, and in what way these skills changed during the preparation for- and response to Sandy need to be established. Evaluations supply information for the narrative, but are also primary sources of information regarding what policy actions were taken.

The independence of the FEMA evaluation is important, as the conclusions of evaluations often differ per evaluating actor (McConnell, 2011). This independence will be established based on the composition of the investigation committee and on the way it is received by other interested parties and different media outlets. In the US bi-partisan political climate, this is necessary to rule out party politics. A typical example of this, is the federal evaluation of the response to Katrina, dubbed a ‘fantasy document’ (Birkland, 2009) and a ‘whitewash’ (Washington Post, 2006), as it exempts government officials that made mistakes with dire consequences. Another factor that weighs in, is that this evaluation fails to mention that there were more than 1800 Katrina-related deaths. Omissions like this show an effort to portray the actions of the evaluating party in a more positive light, which points in the direction of a public relations activity instead of an honest attempt to learn from the experience (Birkland, 2009: 154).

This thesis follows the recommendations of Birkland (2009) to look into the range of ‘lessons’ documents produced after a disaster: “These include anything from changes to standard operating procedures to major statutory changes, as well as internal reports and analyses.” (Birkland, 2009: 154). Evaluations from several government agencies will be used as primary sources for information regarding the policy actions of different federal actors, as will federal policy documents and the reports of congressional hearings regarding this matter. As the timeframe of this research does not allow for the analysis of every single evaluation of hurricane Katrina, the ‘lessons learned’ that are distilled from an independent evaluation will be compared to other evaluations until no new insights come forward. As this is a subjective choice , it is possible that some information regarding the (lack of) skills of FEMA will be missed. However, triangulation of this information with other sources should yield insight into the most prevalent lessons learned. Both national and local (online) newspaper articles covering the preparation for and response to Katrina will be consulted to triangulate the information from these documents.

(30)

The skills of FEMA are determined negatively: instead of assessing all the skills that FEMA possessed during the emergency response to Katrina, the lessons learned documents and evaluations will be probed for the skills that they missed at the time. This method ensures a more complete picture, as the skills that are present are less studied: both evaluations and lessons-learned documents focus on the skills that an organization is lacking.

In order to determine the key factors in the policy actions surrounding Katrina, the recommendations of several Katrina evaluations will be used. The key factors will not be determined beforehand, but identified using coding. First, the significant amount of data on Katrina is reduced to feasible blocks of information. This will be done using open coding on the collected data, analyzing bits of data and then referring a code to these bits. This means that data collection and analysis are conducted jointly – with constant comparison to previously coded text. In this analysis, the data will be coded for ‘lessons that need to be learned’. The coding categories are used to “summarize the data and tell the story” (Phillips, 2014: 550).

Embedding and implementing ‘lessons learned’

The evaluation process includes the gathering of information and processing this information. After the publication of lessons learned documents and action plans, the learning process requires that these recommendations are embedded and implemented. This thesis considers a recommendation to be embedded if it is translated into policies and/or changes in the operating procedures or organizational structure of the involved organization(s). Then, in order to rule out so-called ‘paper changes’, there should be proof that these policies and organizational changes are implemented.

Operationalization of the implementation of a policy depends on the problems it aims to tackle and the proposed changes. Policies proposed after hurricane Katrina that aim to improve the utility of FEMA will be gathered from secondary sources like academic papers and news articles. Information on (organizational) changes will be gathered from both internal and external reports on the functioning of FEMA. The focus of this research will be on policies that can be traced to the lacking skills identified during Katrina. However, as this thesis also intends to uncover factors that possibly influence the learning from disasters process - serendipity should not be excluded. Therefore –and to sketch a complete picture of the political surroundings - policies that

(31)

cannot be linked back to the identified failures but do intend to solve a problem identified after hurricane Katrina will be described briefly, too.

Sandy

In order to conclude whether learning has taken place and what factors influenced the learning process, the findings from Katrina will be compared to the data collected on the emergency response to superstorm Sandy by the federal government of the United States, in 2012. First, the implementation of the proposed policy changes that can be traced to the missing skils during Katrina, will be researched. This is followed by researching the key indicators of the skills of FEMA. Both the timeline of events leading up to the storm, and the policy activities of involved government agencies will be extracted from relevant policy documents and evaluations of the emergency response.

Evaluations of different (federal) organizations involved will be scanned, based on the key indicators that are identified in the previous chapter on hurricane Katrina. The choice of these evaluations will be based upon availability at the time of research. The information offered by evaluations that concerns other organizations will be compared by this organization’s evaluation of the same events in order to prevent a biased view of events and policy actions.

Research scope

The time-frame of this thesis will comprise of the scope of the bi-partisan evaluation of Katrina, with a focus on the preparation and response phases. The scope of this evaluation covers the time-period between August 23 and September 30, 2005. The analysis will focus on data concerning the lessons learned and main recommendations.

(32)

Hurricane Katrina

In this chapter the events leading up to hurricane Katrina will be outlined. Then, the chief lessons learned during the crisis management surrounding hurricane Katrina will be distilled from the bi-partisan congressional evaluation. These recommendations will then be rewritten to pose as indicators for the analysis of crisis management activities surrounding superstorm Sandy.

Timeline hurricane Katrina

The Alabama (AL) National Guard headquarters begins to monitor the tropical depression that would eventually become hurricane Katrina on Tuesday, August 23 2005. It is the beginning of the hurricane season and FEMA starts a large training effort to enhance capability (DHS, 2006: 87).

On Wednesday, August 24 at 1200 UTC, the tropical depression that had formed over the Atlantic evolved into a cyclone and was named Katrina. At that point in time it was centered over the Bahamas, later turning westward toward southern Florida (Knabb et al, 2005). In anticipation of this cyclone making landfall, FEMA’s Hurricane Liaison Team (consisting of FEMA, the National Weather Service, and state and local emergency management officials) was activated and deployed to the National Hurricane Center (Select Bipartisan Committee, 2006: 59). The Joint Operations Center at the National Guard Bureau (NGB) was set up as the operation center for the Katrina response, and the NGB sent Liaison Teams to Alabama, Mississippi and Louisiana (ibid: 66).

During the afternoon of Thursday, August 25 Katrina strengthened further, reaching hurricane status at approximately 2100 UTC “less than two hours before its center made landfall on the southeastern coast of Florida.” (Knabb et al, 2005: 2). At the time of landfall - 2230 UTC, 1830h local time - Katrina was classified as a Category 1 hurricane on the Saffir-Simpson Hurricane Scale, with maximum wind speeds of 130 km/h. It stayed on land for 6 hours, resulting in minor damage and several casualties.

Overseas, Katrina weakened to a tropical storm, with wind speeds dropping to 110 km/h. FEMA activated the National Disaster Medical System, which was meant to

(33)

transfer patients out of the impact zone (Select Bipartisan Committee, 2006: 297). The Centre for Disease Control activated their Emergency Operations Center.

On Friday, August 26 the storm gathered in the Gulf. In the 24 hours after 0600 UTC, the maximum sustained winds increased from 120 km/h to 175 km/h - resulting in a Category 3 Hurricane. At 1700 hours local time, the National Hurricane Centre indicates that Katrina’s track has shifted significantly. Where they first predicted a second landfall in Florida on the 28th or 29th of August, they now predict a second landfall near eastern Louisiana and the Mississippi Coast (DHS, 2006: 145). In Alabama, the Alabama Emergency Operations Center is activated (ibid.). Governor Riley “personally visited all of the counties in the Gulf, holding numerous press conferences to urge local residents to evacuate pursuant to the mandatory evacuation orders” (Select Bipartisan Committee, 2006: 62). In Mississippi (MS), Governor Haley Barbour activated the state’s National Guard, initiating military preparation for the hurricane (ibid: 61). In Louisiana (LA), Governor Blanco issues a State of Emergency and directs the execution of the State Emergency plan.

On the morning of Saturday, August 27 the Governor of Mississippi declares a State of Emergency and the Mississippi Emergency Operations Center is activated. The