Information Sharing in Times of Crisis

1

Information Sharing in Times of Crisis

Analysis of five crises on an inter-regional level

Charley Meijer, Bsc S1338501

Charley.meijer@hotmail.com Master Crisis and Security Management

Leiden University - Institute for Security and Global Affairs Thesis Supervisor: Dr. S.L. Kuipers

Second Reader: Dr. W.G. Broekema 03-03-2019

2

Index

2 Factors information sharing………....9

2.1 Technical failures/incompatibilities……….9

2.2 Organizational barriers………...9

2.3 Conflict of Jurisdiction……….11

2.4 Information of private organizations……….11

2.5 Information overload………12 2.6 Different interpretations………...12 2.7 Cultural incompatibilities………12 2.8 Existing relationships………...13 2.9 Wait-and-see attitude………13 2.10 Relevance of information………...13 2.11 Expectations………..14

3 Crisis management in the Netherlands……….14

4 Research design and method………16

4.1 Research Design………16

4.2 Data Collection and Sources………...17

4.3 Operationalization……….19 4.3.1 Dependent Variable………..19 4.3.2 Independent Variables………..20 4.4 Data Analysis………..22 4.5 Limitations………...24 5 Results………...25 5.1 Technical failures/incompatibilities………25

5.1.1 Malfunction of vital communication networks………..25

5.1.2 The insufficiency of technical alternatives………26

5.2 Organizational and policy structures……….27

3

5.2.1.1 Positions unclear/unfulfilled………28

5.2.1.2 Disagreement among equals………28

5.2.1.3 Clarification among equals………..29

5.2.1.4 Agreement among equals………..29

5.2.2 Vertical structure issues……….30

5.2.2.1 Overarching body………..……….30

5.2.2.2 Liaisons………31

5.2.2.3 Equals………..32

5.3 Conflict of Jurisdiction……….……….32

5.4 Information of private organizations………..…………32

5.5 Information overload………..…………34 5.6 Different interpretations………..…………..34 5.7 Cultural incompatibilities……….………...35 5.8 Existing relationships………...……..……...35 5.9 Wait-and-See Attitude……….………36 5.10 Relevance of information……….37

5.10.1 Not sharing information………..37

5.10.2 Not knowing which information is needed………...38

5.11 Other obstacles to information sharing………....39

5.11.1 Unable to reach………...39

5.11.2 Inexperienced………...39

5.12 Comparing crises………..40

6 Discussion, Conclusion & Future Research Options……….41

6.1 Discussion……….41

6.2 Conclusion………....43

6.3 Limitations of the Research……….43

6.4 Implications for the crisis response……...……….………...44

4

List of abbreviations / translations

CoPI Commando Place Incident Team

ERC Emergency Response Center

GBT Policy Team of the Municipality (Gemeentelijk

Beleidsteam)

GRIP Coordinated Response Plan

ICCB Interdepartementale Commissie Crisis Beheersing

LCMS het Landelijk Crisis Management Systeem

LOCC Landelijk Operationeel Coördinatiecentrum

Nationaal Noodnet National Emergency Network

NCC Nationaal Crisis Centrum

NCV Noodcommunicatievoorziening

Nederlandse Voedsel Dutch food and drug administration en WarenAuthoriteit

ROT Regional Operational Team

Trans receiver Portofoon

ChemPack the crisis at Chemie Pack Moerdijk

MalTel the crisis of the Malfunctioning Telecommunication

Network KPN

Shell the crisis at Shell Moerdijk

PowOut the crisis of the Power Outage

5

1. Introduction

1.1 Crisis

Within seconds, an incident can evolve into a chaotic and complex situation, in other words, a crisis (Yang et al. 2010). Such ‘a crisis’ is defined by Rosenthal et al. (1989) as “a serious threat to the basic structures or fundamental values and norms of a social system, which, under conditions of time pressure and very uncertain circumstances, demands critical decision making” (p.10). Within this research, this definition is adopted.

When a crisis occurs, multiple organizations from different disciplines are sent to respond to the crisis area. For the reason that, when an extreme event happens, not a single organization has the equipment to alleviate all the effects (Bui et al, 2000, Boin & Bynander, 2015). Think of first responders such as the police, fire brigade, and medical responders. In addition, because a crisis has an enormous impact on basis structures of society, multiple organizations, other than first responders, become involved. Such organizations could be electricity companies, municipal departments, the airport authorities, housing depots, or the public prosecution office. All these organizations have their own responsibilities and tasks to complete, which makes it necessary to establish coordination.

While responding to a crisis, communication and coordination problems are likely to occur (De Bruijn, 2006, Boin & Bynander, 2015). This makes it more likely for all

organizations involved to fail to respond promptly. If they do so, the economical, as well as, the political and social consequences are immense (Gawronksi & Olson, 2000; Carley & Harrald, 1997). “Public agencies bear legal responsibility for the protection of lives, property and continuity of operations, and local agencies bear the brunt of first response” (Comfort & Kapucu, 2006, p.312).

Despite having gathered a lot of knowledge on information sharing during a crisis, and on coordinating in an effective way, this topic remains at the top of the research agenda (Bharosa, Lee & Janssen, 2010). The intriguing question here is; why does it remain so difficult for crisis response organizations to share information and coordinate their actions? What challenges and obstacles of sharing information and coordination are encountered during crisis response? This research will attempt to find the answer to these questions by examining the following research question: What obstacles explain (in)adequate information

sharing in five cases of Dutch crisis management on an inter-regional level in the period of 2010-2018?

6

As a relatively simple answer to this question, there is the assumption that it probably has to do with the dynamic, unpredictable and complex environment in which these

organizations work together (Bharosa, et al, 2010). Besides that, all organizations involved have their differences, for example in culture and organizational structure, which impedes effective collaboration (Mendonça, Jefferson & Harrald, 2007). However, the question about the challenges and obstacles of sharing information and coordinating effectively remains (Ren et al, 2008 in Bharosa et al, 2010, p.50), making this study highly relevant to the academic field. In addition, this study has a societal purpose. By understanding the challenges of information sharing during the crisis response, the performance during a crisis will improve. Chapter 3 discussed further relevance of this study.

First at section 1.2, the coordination of crisis response will be looked at, followed by elaboration on the multiple organizations that are involved in section 1.3. Section 1.4 will clarify how information sharing connects to coordination. In chapter 2, different factors influencing information sharing will be elaborated on. Then, chapter 3 explains the Dutch crisis management. After that, chapter 4 explains the research design and method, chapter 5 includes all results, which are elaborated on in chapter 6 ‘Discussion & Conclusion’. This study will conclude with limitations and recommendations for further research.

1.2 Coordination

It is undisputed that there is a need for coordination in times of crisis. The lack of coordination can lead to possible failures and magnifying damage (Bharosa et al, 2010). Coordination can be seen as both the problem and the answer in a crisis situation (Quarantelli, 1988). On one hand, coordination often shows to be difficult to establish, but on the other hand, if successful, a crisis can be handled effectively, limiting the damage.

It is not always clear what is understood by ‘coordination’. Koop & Lodge (2014) define it as “the adjustment of actions and decisions among interdependent actors to achieve specified goals” (p.1313). What can be derived from this definition is that ‘coordination’ has to be established between different organizations, for each of them to complete their own task. They cannot function on their own. They need each other’s expertise to reach their own goal. For example, a medic can only treat a victim of a fire, if the fire brigade rescues the victim. To rescue the victim, the fire brigade needs to know in what area the medic is located so they can deliver the victim to that area. In addition, the medic needs to know whether the victim has been in contact with fire and smoke. This is important information to treat the victim medically, which can only come from the fire brigade.

7 1.3 Multiple organizations involved

The example of the medic and fire brigade needing to work together shows that multiple organizations are involved in crisis response. As Dantas and Seville (2006) describe it: “The challenges involved in coordinating an effective response to large scale emergency events are compounded by the number and variety of organizations involved” (p.43). What makes coordination difficult is the fact that all the organizations involved have their own goals to reach. Even more difficult is integrating all these organizations and their jurisdictions into an inter-organizational, inter-jurisdictional response system, while the chaotic conditions of a crisis are unfolding (Comfort & Kapucu, 2006).

This research will adopt a top-down perspective on coordination, as opposed to a bottom-up perspective. The following section explains the bottom-up perspective. The top-down perspective argues for a form of guidance that makes sure that all people, units and organizations, work together. If all the different organizations want to coordinate their actions, it is unlikely that this will arise in an efficient and timely manner. Therefor there is need for a formal mechanism that specifies how different organizations relate to each other (Thompson, 1967). By defining who is in charge and who the responsible parties are, it should be clearer to all organizations what their goal is. In addition to this, it is important to include the collective process of planning and training into this perspective. This helps prepare all organizations, “by putting anticipated needs in context, by clarifying mutual requirements on e.g. communications and logistics, and by clarifying the strengths and weaknesses of network partners” (Boin & Bynander, 2015). The top-down perspective is relevant for this research, for it shows resemblance with the Dutch Emergency System. Chapter 3 describes this system.

In contrast stands the bottom-up perspective, which is about how groups and

organizations work together in an effective way without being directed from the top. Essential to this principle is the assumption that in a situation where people face an unforeseen, hard to solve problem, they tend to collaborate to come to a solution (Boin and Bynander, 2015). It remains unclear in which situations such collaboration emerges. This, and the fact that the Dutch Emergency System follows the top-down perspective, makes the bottom-up perspective irrelevant for this research.

1.4. Information sharing

Information is a difficult element in a complex situation, such as a crisis, for there is an operational picture to put together, formed by many little pieces. The complexity of a crisis arises from a variety of elements, processes, systems, and actors. Which makes it hard to get a

8

clear overview of the crisis (McEntire, 2002). The best way to get a clear picture of a crisis is free sharing of accurate information. However, with all different organizations involved, information sharing can be a daunting task (Bui, Cho, Sankaran & Sovereign, 2000).

When a crisis occurs, all organizations involved need to collect and communicate data and information that are necessary to minimize a crisis’ impact (Dantas & Seville, 2006). This data consists of the characterizations of a crisis, such as the location of the crisis, the amount of damage, the amount of casualties, and the availability of physical and human resources. However, as described by Bharosa et al (2010), response teams find themselves facing a number of obstacles. Bui et al. (2000) make an even stronger statement: “… operations that require the participation of multiple agencies are prone to problems related to information exchange…” (p.428).

The obstacles that Bharosa et al (2010) identify when it comes to information sharing a coordinating crisis response, is the theoretical framework adopted in this research. In their research, Bharosa et al (2010) identify a large number of obstacles and challenges

encountered by crisis response workers. This means that there is not a single factor

responsible for impeding or facilitating information sharing and coordination. Since there are various factors identified, they are subdivided in three primary categories: 1) Procedures, 2) Responsibilities and 3) Values, Norms and Priorities.

Firstly, ‘procedures’ is defined as ‘an established or official way of doing something’ by the Oxford Dictionary. This research will add the phrase ‘in certain situations’, referring to a crisis. Therefore, the definition of ‘procedures’ in this research is set as ‘an established or official way of doing something in a certain situation’. The factors falling within this category are: ‘Technical Failures/Incompatibilities’ and ‘Organizational and Policy Barriers’.

Secondly, ‘Responsibilities’ is ‘the state or fact of having a duty to deal with something or of having control over someone’ according to the Oxford Dictionary. This research will include this exact definition. Derived from Bharosa et al (2010) the following factors are included: ‘Conflict of Jurisdiction’ and ‘Information of Private Organizations’.

Thirdly, ‘Norms, Values and Priorities’. The Oxford Dictionary defines a ‘norm’ as ‘something that is usual, typical or standard’. A ‘value’ is defined as ‘the regard that

something is held to deserve; the importance, worth, or usefulness of something’. A ‘priority’ is defined as ‘the fact or condition of being regarded or treated as more important than others’. Combining these definitions, the following interpretation of this category is used: ‘the way in which a situation should be handled, based on the standard and importance of the given fact’. The following factors are covered with this category: ‘Information Overload’, ‘Different

9

Interpretations’, ‘Cultural Incompatibilities’, ‘Existing Relationships’, ‘Wait-and-See Attitude’ and ‘Relevance of Information’.

Knowing what is understood by a crisis, having a definition of coordination, knowing the importance of information sharing, and understanding how the involvement of multiple organizations make it all complicated, makes it all easier to find an answer to the research question. Chapter 2 will set out the factors of Bharosa et al (2010) as a theoretical framework.

2. Factors information sharing

Bharosa, Lee & Janssen (2010) identified several different aspects that “may influence inter-organizational information sharing” (p.57). As described before, these aspects divide in three categories, namely 1) Procedures, 2) Responsibilities, and 3) Norms, Values and Priorities. The following chapter will describe these categories, and clarify the differences between the factors that fall within these categories.

2.1 Procedures – Technical Failures/Incompatibilities

Communication technologies are used for better crisis handling, “particularly

disseminating information relating to mitigation, preparedness, warning, response, recovery or learning.” (Garnett & Kouzmin, 2007, P.178). Communication technologies for the

distribution of information to all organizations involved in the crisis response. However, communication technologies have many vulnerabilities, and even more when it comes to the use of them in crises (Garnett & Kouzmin, 2007; Heegaard & Trivedi, 2009). For example, water or wind damage, or power outages are causes of technological failures.

Furthermore, interoperability of communications technologies between different organizations causes problems in information sharing, impeding crisis response (Garnet & Kouzmin, 2007). When the technical equipment of one organization does not link with the technical equipment of the other, the data cannot be shared (Bui et al, 2000). Both nature and humans can cause technological failures (Garnett & Kouzmin, 2007).

2.2 Procedures – Organizational Structures

Crisis response organizations mostly have a traditional organizational structure. Crisis management research shows that these traditional structures are either unavailable or

inflexible for urgent information sharing (Yang et al, 2010). This is because these traditional structures often have had a long time to develop, and solve problems in a certain way. These

10

problems often occur in routine circumstances, when there is time to identify problems, train personnel and plan actions. However, in a crisis, such fixed ways are always too inflexible to cope with the stream of information that needs to be processed and shared (Dawes, Cresswell & Cahan, 2004, De Bruijn, 2006; Comfort & Kapucu, 2006). Therefore, the way in which crisis response organizations are structured can pose a problem to information sharing.

Having different layers of responsibilities and different organizational structures, adds to this problem. For example, the level of clearance might pose a problem. When two persons communicate with each other about a problem. The situation might occur where one person knows the information the other person needs, however this information is confidential and the other person does not have the clearance level to know this information. When this happens within an organization, this might pose a problem, which it will be even more when it happens between two organizations.

Another problem that occurs when it comes to information sharing, is the sharing of sensitive information. First, the sharing of information comes with risks. Professionals should always consider whether the costs outweigh the benefits of sharing the information (De Bruijn, 2006). The difficult thing here is that sharing information could ensure a better performance of an organization, but it may also harm the organization or worse, threaten the national security.

To illustrate, in some situations information needs to be shared to make sure the safety of the responding organizations and their employees can be guaranteed (Bui et al, 2000). However, by doing so, the sharing of sensitive information might occur. In this situation, it is important that only people with a certain degree of clearance have access to the information. However, how does anyone know what the clearance level of an employee of another organization is? How will they get in contact? On what platform can they share the

information, without people with a lower clearance level having access? These insecurities make that the costs-benefit consideration will conclude negatively and thus the sharing of information will not happen. A solution to this problem is an access control mechanism that could be put in place, however, the existing access control mechanisms that are in use are too inflexible to function in a crisis situation (Yang, Yao, Garnett & Muller, 2010).

Elaborating on the subject of sensitive information, not only technical barriers of access are the problem, most organizations handling sensitive information have

confidentiality restrictions in place, preventing them from even considering sharing this information in the first place (Dawes, Cresswell & Cahan, 2004).

11 2.3 Responsibilities – Conflict of Jurisdiction

Another obstacle encountered is differences in jurisdiction. Every organizations that carries the burden to respond to a crisis, has its own specific goals. This can be a problem when there needs to be coordination. One on hand this is a problem, for a crisis can cover several territorial jurisdictions (Dantas & Seville, 2006). On the other hand, this is a problem, for a crisis can cover several types of jurisdictions. For example, a fire releasing toxic gases in the air. The fire brigade reacts to the crisis, for they are in charge of handling the fire. Medics react to the crisis, for they are responsible for handling all victims. The police reacts to the crisis, for they are responsible for keeping curious people out. In addition, several other companies react to the crisis, for they need to measure the amount of toxic gasses in the air, warn people about the toxic gasses and see that they affect no one.

All these organizations have their own type of jurisdiction, with their own goals. Now the situation might occur where for example, the medics need to enter a specific area to treat victims, but the fire department will not allow it, for that area poses a threat to the wellbeing of the crisis responders. The medics and the fire brigade will then discuss who has jurisdiction and gets to decide what happens.

2.4 Responsibilities – Information of Private Organizations

Crisis do not only affect governmental organizations, but also private organizations. Some of those private organizations suffer physically or economically from a crisis. Other private organizations help in the crisis response (Comfort & Kapucu, 2006). For example, in a car accident where electrical wiring broke, which is endangering the response team. The electricity company needs to cooperate to make sure safety is guaranteed for all actors involved.

In some crises, some private organizations have critical information that is necessary in managing the crisis, which can be inaccessible to other responding organizations (Dawes, Cresswell & Cahan, 2004). For example, if a power outage occurs there will be several incidents, which need the police, fire brigade and medical assistance. However, they need information about the power outage itself from the electricity company. The electricity company might withhold information, because they are uncertain about the situation and are not ready to take responsibility.

12 2.5 Norms, Values and Priorities – Information Overload

There has to be a selection on what to share and what not, otherwise an enormous overload of information is the result. Information overload is a serious problem when it comes to information sharing during a crisis (Bharosa et al, 2010; Wolbers & Boersma, 2013). If a crisis responder receives too much information about the situation, this will mostly be

meaningless information (De Bruijn, 2006). This means that most information shared between crisis responders, is not relevant to every single crisis responder. Adding to that, the

information overload may lead to focusing only on selected sources. If a crisis responder wants to get a clear picture from all information that is being shared, the situation might occur where the search for relevant information is overruled by al the irrelevant information. With the disastrous consequence of missing valuable information.

2.6 Norms, Values and Priorities – Different Interpretations

The moment information is shared, the actor sharing it loses control over it (De Bruijn, 2006). Information can mean one thing to someone, and something completely different to the other. Professional terminologies, for example, are a barrier for successful information

sharing. If the information transfers to another organization, but this organization does not understand the jargon of the other organization, there is a language barrier (Dantas & Seville, 2006; Comfort & Kapucu, 2006; Wolbers & Boersma, 2013). Insufficient evaluation or validation of the information causes confusion over information (Rake & Nja, 2009).

Furthermore, the situation might occur where a dominant narrative develops, which does not leave room for different interpretations of information (Wolbers & Boersma, 2013). For example, this might be the case when there is one person in charge, who sees only one solution to a problem. However, others might see the situation in a different way, with different solutions. If those others do not get the change to explain their ideas, only one solution is explored, while it may not be the best solution possible.

2.7 Norms, Values and Priorities – Cultural Incompatibilities

Cultural differences also complicate information sharing between organizations. Organizations with conflicting cultures, might misunderstand each other, or are less willing to share their information with each other, for example because they do not care about the safety of the other organization (Bui et al, 2000; Dantas & Seville, 2006; Comfort & Kapucu, 2006; Wolbers & Boersma, 2013).

13

Furthermore, (intelligence) organizations often have the value of information

monopolization and not cooperating with others (De Bruijn, 2006). Obviously, this makes it harder to come to a cooperation during a crisis.

2.8 Norms, Values and Priorities – Existing Relationships

Earlier literature concludes that the existence of a network of relationships among response workers, strongly affects the quality of communication and coordination (both organizational and personal) (Dawes, Cresswell & Cahan, 2004). In the situation where crisis responders know each other, for example because they have worked with each other before, the information sharing and coordination of the crisis response seems to come across far less obstacles.

2.9 Norms, Values and Priorities – Wait-and-See attitude

According to de Bruin (2006), a wait-and-see attitude is seen in organizations, when they get to much autonomy. This results in not sharing information with organizations in need of that information. As Waugh and Streib (2006) explain: “Encouraging the willingness to share information is believed to support collaboration…” (Wolbers & Boersma, 2013, p.187). This means emphasizing that information needs sharing. Either by authorities, or

organizations themselves (Bui et al, 2000).

2.10 Norms, Values and Priorities – Relevance of Information

It is not always clear beforehand which information is relevant to other crisis response organizations (De Bruijn, 2006). The cause of this problem can come from the fact that each organization has its own goal and organizations are not familiar with each other’s goals.

Furthermore, it seems that crisis response organizations only engage in gathering information, instead of sharing it (Bharosa et al, 2007). Representatives of different

organizations involved in crisis response, often ‘forget’ to share their information with other organizations, because they tend to focus only on their own goal (Dearstyne, 2007; Wolbers & Boersma, 2013). This also has to do with the fact that often, there is a limited

understanding what the consequences will be of information on the actions of other professionals (Wolbers & Boersma, 2013).

Contradicting information can be rather frustrating for rescue workers, just as late information is. This way, it is unclear what information is correct or if information is missing.

14

Not having a complete overview of a crisis makes crisis response work very difficult (Carlson & Davis, 1998; Bui et al, 2000).

2.11 Expectation

With the description of the factors derived from Bharosa et al (2010), the theoretical section is almost finished. Before setting out the structure of crisis management in the

Netherlands, this section will explain the expectation of this study. The aim of this study is to see whether the factors of Bharosa et al (2010) are still relevant 9 years later. It is expected that the Dutch crisis response system improved over time, meaning that the analysis encounters the factors of Bharosa et al (2010) less as obstacles and more as stimulation to information sharing over time.

3. Crisis management in the Netherlands

Since this research will be conducted studying crises that happened in the Netherlands, it is important to give an overview of the Dutch crisis response system. In the year of 2010, a new law came into effect, named ‘the law on safety regions’. Among other things, the law pursues an efficient and high-quality organization of crisis management. In addition, the law mainly contains conditions under which the organization of the crisis response of safety regions needs to operate. One of the pillars of this law is the regulation of crisis response on an

inter-organizational level. This is in line with the observation of Comfort & Kapucu (2006) that inter-organizational collaboration has become more and more part of the functioning of organizations.

The organization of the crisis response in the Netherlands has the following structure. A system points a level of severity to a crisis called Coordinated Response Plan (GRIP in Dutch). From level 1 to level 4 (being most severe). Level one and two are for relatively small crisis, where the command structure remains quite clear, for there are not (that) many

organizations involved. The main reason to up a crisis to level three or four is the clear administrative hierarchy that is activated. In such a situation “the commanding officers of the relief agencies meet on two levels of decision-making: The Regional Operational Team (ROT) and the Commando [‘commando’ in Dutch means command, in the military sense of the word] Place Incident Team (COPI)”. In addition, GRIP5 joined into the crisis structure in 2013. This new level was to get some clearance about what to do when multiple safety

15

regions are included in a crisis. When multiple safety regions are affected by a crisis and there is an administrative need, GRIP5 comes into play. Furthermore, it is essential to know that the Emergency Response Center (ERC) only has a facilitating role and is not part of the command hierarchy (Bharosa et al, 2007, p.54). Besides the ROT and COPI who operate on a safety region level, there is also a team established in the municipality that is affected by a crisis. This is called the GBT (Gemeentelijk Beleids Team) or the Policy Team of the Municipality. The law on safety regions has determined that in case of a GRIP3 or GRIP4 the chain of command is as follows: COPI-ROT-GBT.

The Dutch crisis response organizations use an integrated information system called C2000. It is a replacement for the almost hundred analog communications systems previously used by the emergency services. C2000 is always available for the police, the fire brigade, medics and some sections of the army. C2000 helps emergency workers communicate with each other and the ERC (Expertgroep C2000, 2009). The NCV (NoodCommunicatie

Voorziening) is a telecom network used by government and crisis responders during a crisis, when the regular network (C2000) is overloaded or completely down.

Since the implementation of the Law on safety regions, the Inspectorate of Safety and Justice has done three evaluations about the progress of the implementation of Law on safety regions (2010, 2013, and 2016). This so-called ‘state of the emergency response’ shows a growth of the safety regions. It turns out that most safety regions are well prepared for a crisis. In addition, the collaboration with other partners is getting better. However, most safety regions need to improve their performance when an actual crisis is unfolding (Inspectie Veiligheid en Justitie, 2016).

However, these studies looked at the independent safety regions and how the crisis response is organized within their borders. The current study will explore the crisis response on an inter-regional level, analyzing which improvements need to be made between different safety regions. This is the first study to conduct such an analysis, which is highly relevant for society, for the Dutch crisis management has not implemented procedures about inter-regional crisis response.

With this overview of the Dutch crisis response system it should be clear how crisis responders approached the crises included in this research. The following chapter will describe the research design and method, and includes a description of all crises included in this research. This description includes some other abbreviations that are necessary to understand the situations that were encountered during the crises.

16

4. Research Design and Method

The following section (4.1) discusses the research design of this study. Section 4.2 discusses the data collection and sources. Section 4.3 explains the operationalization of the variables. Followed by section 4.4, which elaborates on the data analysis. And finally, section 4.5 discusses the limitations relevant to this analysis.

4.1 Research Design

This research conducts a comparative case study to answer the research question. Case study research is an in-depth examination of phenomena in real-life context (Yin, 2009). As this research aims to gain insight about what factors of information sharing influence an ineffective coordination, a case study design is suited. The different cases of inter-regional crises makes for a good representation of real-life events. Thus, making it possible to do an in-depth examination of these crises. To elaborate, this is a comparative case study.

‘Comparative’ means that similarities and differences will be explored systematically (Mills, Durepos & Wiebe, 2010). To be able to compare, several cases are selected.

The cases selection bases on a most similar system design (MSSD). It is essential to select cases that are similar regarding their background variables, except the outcome is different (Seawright & Gerring, 2008). This method is useful to explain the variances in outcomes, while you control for background conditions and important alternative explanations (George & Bennet, 2005). This research has selected cases that happened in the Netherlands, which means that the structure of the crisis response is similarly in each case. Furthermore, all cases happened on an inter-regional level. However, the crisis differ on the dependent

variable, for they are either adequate or inadequate regarding information sharing. The

criterion for (in-) adequate information sharing is set out in chapter 4.3.1. The variation of the independent variables, will explain for the difference in outcome (Seawright & Gerring, 2008). A disadvantage to the MSSD is that finding cases that are exactly alike is almost impossible (George & Bennet, 2015). This study is not able to completely avoid this flaw, for it included five different types of crises: a fire, a telecommunication network malfunction, an explosion, a power outage and a weir collision. This is a deliberate decision, for it increases the external validity of the study, as described in section 4.5.

A within-case analysis completes the comparative case study. This means looking for either causal mechanisms and causal processes, or both (Rohlfing, 2012). This research will be looking into situations that describe inadequate information sharing, or situations that

17

describe adequate information sharing. By looking into these situations, the cause of (in-) adequate information is explored. Thus, analyzing which independent variables explain for the outcome of the dependent variable.

4.2 Data collection and sources

The data for this study is retrieved from secondary sources. They include assessments from the Inspectorate of Justice and Security and the Dutch Safety Board. From both the Inspectorate of Justice and the Dutch Safety Board we conducted a list of incidents and crises between 2010 and 2018. From this list, all crises meeting the definition of Rosenthal et al. (1989) were marked. The five crises studied in this research remained, after filtering out all crises affecting multiple safety regions.

From the five crises included in this research, three assessments are from the

Inspectorate of Justice and Security, and two assessments are from the Dutch Safety Board. None of the crises are investigated by both institutions. This means that this research analyzes on report of each crisis. The reports are similar in the way that both institutions have the same goal in conducting research to a crisis. Namely, learning and improving the way of handling a crisis, and improving the safety in the Netherlands. However, it is important to pay attention to the fact that the Inspectorate and the Safety Board are different organizations. The

Inspectorate of Justice and Security is an independent institution. They decide whether they want to investigate a case or not. The Dutch Safety Board, on the other hand, works on behalf of the Dutch Ministry of Justice and Security. This research considers this difference;

however, it will not make a difference. Both the assessments of the Inspectorate and the Safety Board include recommendations on how to better the crisis response. These

recommendations are the criterion to determine whether a crisis was (in-) adequate when it comes to information sharing.

The first crisis that is included in this research is the fire at Chemie-Pack Moerdijk, “ChemPack” from now on (Onderzoeksraad voor de Veiligheid, 2012). On January 5 of 2011, a fire broke out at an industrial area where chemicals were processed. This is the safety region of Midden- en West-Brabant. The fire caused a large plume of smoke, which was carried north by the wind. The smoke plume covered parts of the safety region Zuid-Holland Zuid, which made it an inter-regional crisis. During the crisis, the GRIP levels of both safety

regions differed for almost four hours. Zuid-Holland Zuid leveled up to GRIP4 at 15.43, while it took Midden- en West Brabant until 19.30 to level up to GRIP4 after several parties

18

insisted. This caused a lot of (unnecessary) confusion and discussion between the two safety regions.

Besides the regular crisis responding organizations that got involved, Chemie-Pack itself was involved in the crisis response. Because of the development of toxic gasses, the GAGS (Gezondheidskundig Adviseur Gevaarlijke Stoffen) got involved. This is a health consultant of hazardous substances. The GAGS works together with the MPL

(MeetPlanleider) which is a coordinator. Together they decided to focus on measuring certain toxic substances to check if the public would be in danger.

The second crisis included in this research is the Malfunction of a Telecommunication Network, “MalTel” from now on (Inspectie Justitie en Veiligheid, 2012). In the night of July 26 to 27, 2011, KPN maintains one of their telecom nodes when a technical failure occurs. KPN is a Dutch telecom provider. The failure of connections has a major impact on the vital infrastructure of the safety regions Rotterdam-Rijnmond, Zuid-Holland Zuid, Hollands Midden and Zeeland. The safety regions Rotterdam-Rijnmond and Zuid-Holland Zuid level up to GRIP2, and Rotterdam-Rijnmond even to GRIP4 the moment that daily life starts. Hollands Midden and Zeeland deem it not necessary to level up to a certain GRIP-level.

The third crisis included in this research is the explosion and fire at Shell Moerdijk on June 3, 2014, “Shell” from now on (Onderzoeksraad voor de Veiligheid, 2015). Two severe explosions caused a big fire in the safety region Midden- en West-Brabant. This fire also caused a large plume of smoke, which affected the safety region Zuid-Holland Zuid. The use of the GRIP levels happened on different times in the two safety regions. Safety region Midden- en West-Brabant leveled up to GRIP3, while safety region Zuid-Holland Zuid leveled up to GRIP2. The majors in charge of the safety regions consulted each other and agreed to the differentiated use of GRIP.

This crisis was chosen to be included in this research, for it creates the unique opportunity to see whether the crisis response had learned anything from the same situation three years earlier at Chemie-Pack Moerdijk. The crises are most similar to each other, opposed to the rest of the crises that are included.

The fourth crisis included in this research is a Power Outage that hit on March 27, 2015, “PowOut” from now on (Inspectie Justitie en Veiligheid, 2016). About one million households lost power, traffic lights do not work, public transport does not work, the mobile phone network does not work, and several organizations must switch to their emergency power supply. The safety regions of Amsterdam-Amstelland, Gooi en Vechtstreek,

19

The safety regions Gooi en Vechtstreek, Kennemerland, and Zaanstreek-Waterland level up to GRIP4, while Amsterdam-Amstelland and Noord-Holland Noord level up to GRIP2 and Zeeland does not use the GRIP-structure. During the crisis, there is a lot of confusion on the accuracy of information. The safety region Amsterdam-Amsterdam, in which lies the source of the problem, gets the role to initiate interregional coordination, but does not fulfil this role properly.

The fifth crisis included in this research is a Weir Collision on December 29, 2016, “WeirCol” from now on (Onderzoeksraad voor de Veiligheid, 2018). A ship loaded with 2,000 tons of benzene hits a closed weir. The skipper did not feel well, as well as the weather circumstance of dense fog, are the causes of this crisis. During the collision, the ship made a large hole in the weir through which large amounts of water flows. This caused problems with the water levels in a large area. The safety regions of Brabant-Noord and Gelderland-Zuid are directly affected by the weir collision. They both level up to GRIP2, shortly after levelling up to GRIP1. The reason for this is the fast drop in the water level.

Furthermore, two regions of Rijkswaterstaat were involved; Oost-Nederland and Zuid-Nederland. Rijkswaterstaat “is responsible for the design, construction, management and maintenance of the main infrastructure facilities in the Netherlands” according to their website. Furthermore, the Netherlands has a Water Authority, which in this crisis included two regions (Rivierenland and Aa en Maas). The administrative complexity makes this crisis interesting to include in this research.

4.3 Operationalization

In the following paragraph operationalizes the dependent and independent variables of the research question.

4.3.1 Dependent Variable

‘Information sharing’ is the dependent variable of this research. This research will test what factors (which are the independent variables described in the next paragraph) cause information sharing to be either ‘adequate’ or ‘inadequate’, in responding to a crisis. The following definition of ‘inadequate information sharing’ is used: ‘the situation in which information is either not shared or not shared in a clear or coherent way’. Opposed by adequate information sharing: the situation in which information is shared in ‘a clear or cogent way’ (Mortensen & Ayres, 1997).

20

During the analysis of the reports, information sharing is inadequate when three of the recommendations in a report are about improving the information sharing. ‘Three

recommendations’ are chosen as a measurement, because the amount of three makes it clear too much did go wrong, which led to inadequate information sharing. The amount of

recommendations about information sharing per crisis are in table 1.

An important additional statement to make is that analyzing these documents is done with an objective view of the researcher. This means that the text that is written down in the reports is interpreted the way it was meant to be. Using the recommendations made in the reports, helps assessing whether information sharing was (in-) adequate in an objective way.

Table 1. Number of recommendations on information sharing per crisis.

Chemie-Pack Moerdijk Malfunction Telecom network Shell Moerdijk Power Outage Weir Collision Number of recommendations 4 2 1 1 4

3 recommendations or more = inadequate information sharing.

4.3.2 Independent variables

The factors influencing information sharing during a crisis are the independent

variables, which are derived from the work of Bharosa et al (2010). As described in chapter 1, these factors divide into three categories, namely 1) Procedures, 2) Responsibilities and 3) Norms, Values and Priorities. Table 2 shows an overview of this division. Table 2 also shows the indicators that represent a factor in the crisis reports.

What should be taken into consideration is that the factors are indicated in a negative way. This research studies what factors influences information sharing in a negative way. This is to come to recommendations on how to better the crisis response. However, when a factor is encountered in a positive way, this is noted to. The code sheet marks these positive

21 Table 2. Codebook Content Analysis

Category Definition Factor Indicators Co

de Procedures An established or official way of doing something in a Technical Failures/Incompatibilities

Description of the failure of technologies or failure of technology to connect to other technology. A1 certain situation. Organizational and Policy Structures Description of when organizational structures turn out to be inadequate. Description of when it is unclear wo carries responsibility. Description of sensitive information hardening the sharing of it or the

description of confidentiality procedures.

A2

Responsibilities The state or fact of having

Conflict of jurisdiction Description of conflict of goals. B1 a duty to deal with something or of having control over someone. Information of Private Organizations Description of information in the hands of a private organization.

Description of information not shared by a private organization. B2 Norms, Values and Priorities The way in which a situation should be handled, based

Information Overload Description of too much information being shared. Description of the loss of information within this overload. C1 on the standard and importance of the given fact.

Different Interpretations Description of not understanding information. Description of insufficient validation of information. Description of a dominant narrative. Description of misunderstanding. C2

Cultural incompatibilities Description of conflicting cultures. Descriptions of keeping information within the organization.

22

Existing Relationships Description of earlier developed relationships between responders. Description of failure in knowing who does what.

C4

Wait-and-see Attitude Description of not promoting the sharing of information.

Description of

organizations not acting first, but waiting for others to take initiative.

C5

Relevance of Information Description of not knowing which

information is relevant to others.

Description of an organization being obtained with gathering information, rather than sharing information. Description of an incomplete overview of the crisis, while it turns out that information was known to others.

C6

4.4 Data Analysis

Dividing the factors into three categories is useful for the content analysis of the selected crisis reports. “Content analysis is a research technique for making replicable and valid inferences from texts (or other meaningful matter) to the contexts of their use”

(Krippendorf, 2004, p.18). This is a relevant technique to use in this research, for inferences about information sharing are made from the crisis reports.

Distinguishing these categories in the crisis reports establishes an overview of all factors influencing the sharing of information. This overview is in the form of a coding scheme; figure 2. This means that during this analysis, the text in the reports is coded. Coding is “the process whereby raw data are systematically transformed and aggregated into units which permit precise description of relevant content characteristics” (Holsti, 1969, p.94). By filling in which section of the text represents which category, a systematic overview of the data is given. Followed by an analysis of the factors found in the sources per case and their relation to the outcome of the case.

23 Figure 2. Coding scheme.

24 4.5 Limitations

This section considers several limitations of the research method of this study. For example, one could argue that this research bases its findings on reports that only describe failure in the information sharing between actors in crisis management. Considering that the reports focus on analyzing the crises on what can be done better, all positive notes are left out. Leaving out everything that went well in sharing information, could give an inaccurate picture of the status of the crisis response. Only describing the negative encounters in information sharing would suggest that the crisis response learned nothing over the years. Therefor this study is biased and not objective. However, the opposite is true. Reading all case reports show that there are positive notes on information sharing. For example, in several situations the reports describe how different people agree on following a certain strategy.

Furthermore, it is important for a study to be reliable, for it makes the results trustworthy. If another researcher, when conducting the same research will find the same results as this research, this research is replicable and has a higher value of reliability. This research establishes this by providing a detailed explanation on how the research is

conducted, including an overview of the categories, factors and indicators that are used to identify the results. In addition, the sources used in this research are highly reliable, for the Dutch Safety Board and Inspectorate of Justice and Security are respectable independent investigative organizations. However, there is one weakness that influences the reliability negatively. Only one researcher conducted the analysis. Even if this researcher worked as precise as possible, mistakes will always be made.

Besides reliability, it is very important for a study to be valid. “A valid study is one that has properly collected and interpreted its data, so that the conclusions accurately reflect and represent the real world that was studied” (Yin, 2011, p.78). This study might not have a high internal validity, for multiple independent variables are studied. This makes it impossible to state direct causes of inadequate information sharing. However, making statements of the causes of inadequate information sharing is not the purpose of this study. This study has the goal of finding out whether the Dutch crisis response improved their information sharing over time. Furthermore, external validity needs to be on point. External validity is the extent to which the analysis of this study can be generalized to the real world (Gerring, 2011). This study has a high external validity. With the implementation of different kind of crises, namely a fire, a telecommunication network malfunction, an explosion, a power outage and a weir collision, the results will represent a wide range of crises. Therefore, the factors that seem to

25

influence the adequacy of information sharing, are more likely to apply to a random picked crisis. If only a fire was analyzed, the external validity would be much lower.

Now that all data is analyzed, the results can be set out. This will be done per factor in the following chapter. At the end of the chapter 5, all results will be summarized in table 3.

5. Results

The following sections analyses which factors of Bharosa (2010) pose an obstacle to

information sharing in five crises. In addition, some of the factors present themselves not to be an obstacle, but encourage the sharing of information.

5.1 Technical failures/incompatibilities

This analysis encountered two major technical failures/incompatibilities. The first is the malfunctioning of vital communication networks. This occurred in four out of five crises.

5.1.1 Malfunction of vital communication networks

The most severe case occurred during the MalTel crisis. During this crisis, several connections and facilities failed, including C2000, P2000, and Nationaal Noodnet.

Furthermore, the safety regions of Rotterdam-Rijnmond, Zuid-Holland Zuid, Hollands Midden and Zeeland could not reach their central alarm room (Inspectorate for Safety and Justice, 2012, p.53). The malfunction of these systems were due to the failure of the KPN telecommunication network (idem, p.52). These communication systems are crucial

connections the network of crisis responders and had a negative effect on the functioning of the safety and security organizations in several safety regions (idem, p.18).

The crisis of Shell is the second crisis in which the malfunctioning of a vital

communications network occurred. During this crisis, the LCMS caused perplexity in several occasions. The Dutch Safety Board reports five major problems with the LCMS in the Shell case. One of those problems is a technical failure, namely some officials could not log in the LCMS system (Dutch Safety Board, 2015, p.81). This made them unable to share and collect information, thus, not coming to an interregional overall picture. (The following sections discuss the other four problems with LCMS during this crisis, for they do not fall within the realm of technical failure/incompatibility)

26

During the WeirCol crisis, the LCMS is also the vital communication network causing problems. It is impossible for the two CoPI’s (from the safety regions of Brabant-Noord and Gelderland-Zuid) to merge their pages together in LCMS (Dutch Safety Board, 2018, p.70). This resulted in two separate crisis pictures, instead of one common picture. However, the LCMS helped create a common picture later on. The LOCC solved the problems by merging the different pages together (idem, p.71). In addition, the safety region of Gelderland-Zuid had a LCMS page of their own (idem, p.68), informing other parties. This helped the sharing of information between crisis responders by creating an overall picture of the crisis.

Another malfunction of a vital communication network happened during the crisis of PowOut. With the loss of power, the NCV does not work. All safety regions have emergency power installed, so their NCV should still be working. However, not all municipalities have this feature. The curious thing here is that this service is especially for times of crisis when the regular services malfunction. This implies that municipalities should have emergency power installed, so they too can use the NCV. Furthermore, the NCV turns out to be unknown by most of the Dutch crisis responders. The use of the NCV stands so far from daily practice that the question is whether it can be used effectively (p.31). This last point will be elaborated on in the section ‘other obstacles’. Elaborating on this crisis, it stands out that the C2000 kept working. The emergency generators kicked in immediately after the power outage, which kept the C2000 working (Inspectorate for Safety and Justice, 2016, p.30). However, not all people involved in responding to a crisis have a connection to C2000. For example, mayors, public officials and communication advisors mostly rely on their telephone to share information (idem, p.30). During the power outage phones did not work as they were supposed to do, which means that information sharing was hard with this vital communication network being offline.

5.1.2 The insufficiency of technical alternatives

The second technical failure/incompatibility that stood out conducting this analysis was the insufficiency of technical alternatives. The failure of NCV described in the last section, partly falls within this kind of technical failure, however, it categorizes under the malfunctioning of vital communications networks, for it is supposed to replace the “normal” communication networks when they malfunction. The alternatives falling within this section are alternatives whose purpose is not necessarily to replace the “normal” communication networks.

27

The use of trans receivers turned out to be insufficient. Trans receivers were used during the crises MalTel and WeirCol (Inspectie Veiligheid en Justitie, 2012; Onderzoeksraad voor de Veiligheid, 2018). During the crisis of MalTel, the trans receivers replaced the use of phones when the KPN telecommunication network failed (Inspectorate for Safety and Justice, 2012, p.53). During the crisis of WeirCol they were an addition to the usual communication techniques (Dutch Safety Board, 2018, p.70). In both crises the trans receivers malfunctioned because of their scope. The scope of trans receivers is not wide (Inspectorate for Safety and Justice, 2012, 53). This means that two trans receivers need to be relatively close to each other, in order for them to receive each other’s messages. Apparently, these two crises were on a larger scale than the trans receivers had connection, which made it harder to share information.

Furthermore, during the Shell crisis, WhatsApp was part of the crisis response. Different crisis teams made use of so-called ‘groups’ in which they shared information. However, WhatsApp has no connection to any vital communication network, so the information shared in these groups could not contribute to an overall picture of the crisis (Dutch Safety Board, 2014, p.82). The Dutch Safety Board (2014) names three possible disadvantages for the use of WhatsApp: “unintentional exclusion of crisis partners, the missing out on information shared in a ‘group’, and information not reaching the formal resources like LCMS” (p.89).

5.2 Organizational and Policy Structures

Conducting this analysis, two overarching structural issues stood out. There is a distinction between horizontal structure issues and vertical structure issues. The difference in this distinction lies in the way crisis responders relate to each other. Are they on the same level or is one superior to the other, hierarchy wise.

5.2.1 Horizontal structure issues

The issues encountered on a horizontal structural level, are very diverse. There are four subcategories. Two issues pose problems for the sharing of information, and two issues in which horizontal structures are used to the advantage of information sharing. 1) Positions unclear/unfulfilled, 2) Disagreement among equals, 3) Clarification among equals, and 4) Agreement among equals. The following sections discuss these subcategories separately.

28 5.2.1.1 Positions unclear/unfulfilled

The crisis of ChemPack describes the situation in which it is unclear to others what the role of ROT members is. The ROT of the safety region of Zuid-Holland Zuid has leadership over the communications strategy and content. In addition, staff members of the local

municipality supported the ROT in their job. Which lead to the situation in which information coming from the ROT was send from emails with a municipality email address. This led to confusion about the divisions of tasks by those on the receiving end of the information (Dutch Safety Board, 2012, p.86)

Another problem encountered, was the situation in which no one was available to carry out a certain position within the chain of crisis response. This happened during the crisis of MalTel with the use of LCMS. Not all crisis teams had an information manager available during deployment. Consequently, not all teams used the LCMS actively (Dutch Safety board, 2015, p.81). The second crisis, in which such a situation occurred, was WeirCol. The safety region Gelderland-Zuid was dependent of the safety region Brabant-Noord. However, the safety region Brabant-Noord did not have a calamity coordinator, which would have been able to provide safety region Gelderland-Zuid with the information they needed.

Consequently, the leader of the CoPI of Gelderland-Zuid could not come to a clear operational picture (Dutch Safety Board, 2018, p.68).

5.2.1.2 Disagreement among equals

Another situation that posed a problem to information sharing during the crisis of ChemPack was the disagreement on GRIP levels between the head of the safety region Zuid-Holland Zuid and the major of Moerdijk. The major of Moerdijk does not see any reason to level up from GRIP3 to GRIP4, while the head of the safety region Zuid-Holland Zuid asks him to. The major of Moerdijk does not see the necessity to level up, after consultancy of the fire department, who states that the fire is under control. Only after insistence of several other crisis responders, and the realization that he has no idea what is going on in the adjacent safety region, there is agreement on levelling up to GRIP4 (Dutch Safety Board, 2012, p.87).

Other equals that did not cooperate smoothly were the public fire department and the local fire department of Shell Moerdijk. The choice to make the officer of the fire department of Shell the link between the officer of the public fire department and the specialists of the fire department of Shell, contributed to less efficient and at times incomplete information sharing. (Dutch Safety Board, 2015, p.88). This implicates that a direct link between the officer of the public fire department and the specialists of Shell, would have prevented inefficient and

29

incomplete information sharing. Despite these shortcomings, it is important to note that the moment the public fire department arrived, the transfer of information and responsibilities from the fire department of Shell Moerdijk went smoothly (idem, p.77).

5.2.1.3 Clarification among equals

The ‘clarification among equals’ describes situations in which crisis response teams are doing their job, but are unaware of other crisis teams they need to cooperate with. Two situations like this came up in the analysis. Although these situations start out as examples of a problem for information sharing, they end as situations in which information sharing is used to their advantage.

First, during the Crisis ChemPack the fire releases toxic gasses into the air. Both safety regions involved in this crisis send teams to measure these gasses and report whether they posed any danger. At first, those two teams did not know of each other’s existence. With intervention of the central alarm room, the teams fine-tuned their measures, so that they would measure the same gasses (Dutch Safety Board, 2012, p.90).

The second situation played during the crisis WeirCol. Both safety regions,

Gelderland-Zuid and Brabant-Noord, installed a CoPI, each on their own side of the river “de Maas”. However, they did not know of each other’s existence. This caused confusion among several crisis responders. For example, when calling someone on the other side of the river, that person stated that ‘he did not have time, for he was almost entering the CoPI’ after which he hung up the phone. The caller got confused, for he did not see a CoPI anywhere. It took until 01:00 hours before they found out that two CoPI’s were taking place, while both CoPI’s started around 23:00 hours (Dutch Safety Board, 2018, p.69-70).

These situations show the importance of knowing which parties respond to a crisis. However, there is a wide range of crisis response organizations. The situations described in the former sections also show the relevance of an organizational structure being in place. The situation at ChemPack was resolved with the help of the central alarm room, who knew how to help with regard to the organizational structures that are in place. Moreover, the WeirCol crisis shows the consequences of organizational structures not being in place.

5.2.1.4 Agreement among equals

The Shell crisis encounters a situation that shows there is also agreement among equals. In other words, a situation in which horizontal structures are used to the advantage of crisis responders. The levelling of the GRIP structure of each involved safety region happened

30

in a differentiated way. The major of Moerdijk, levelled up to GRIP3 in the safety region Midden- en West- Brabant. He aligned this decision with the major of Dordrecht, who also happen to be the head of the safety region Zuid-Holland Zuid. Furthermore, the heads of safety region Midden- en West-Brabant and Zuid-Holland Zuid agreed with the major of Moerdijk not to level up to GRIP 4 or GRIP5. It would not add to the efficiency of the crisis response (Dutch Safety Board, 2015, p.80-81).

5.2.2 Vertical structure issues

The second kind of structural issues, are characterized by a vertical structure. This means that vertical structures pose a problem for information sharing during a crisis, or that the encounter of a vertical structure is used to the advantage of information sharing. The situations relevant divide into three categories: 1) Overarching body, 2) Liaisons, and 3) Equals.

5.2.2.1 Overarching body

This category describes situations in which a crisis response organization/team, who is higher ranking in the crisis response structure, poses a problem in information sharing, or who is of a positive influence on information sharing.

Starting with the PowOut crisis, the LOCC has a positive influence on the sharing of information. During the crisis, the two safety regions Brabant-Noord and Gelderland-Zuid have different interpretations of the operational picture. The LOCC notices this and makes sure both safety regions come to a common operational picture (Inspectorate for Safety and Justice, 2018, p.25). This is an example of a higher-ranking organization positively

influencing the sharing of information.

The ChemPack crisis shows an example of a higher-ranking crisis response

organization posing a problem to information sharing. At first, the ICCB shows a positive influence on information sharing, for they insist that the safety regions communicate with each other and align their information to the public. However, the ICCB also decided on operational situations (such as measurement and extinguishing the fire) and decision making (for example, drafting an action perspective for the safety regions) without acknowledging the safety regions (Dutch Safety Board, 2012, p.117). The Dutch Safety Board (2012) explains that it seems that government services played a major role during the ChemPack crisis, while GRIP4 was the highest level activated of the crisis management system during the crisis. GRIP4 is the highest level, to manage a crisis within a safety region, meaning governmental

31

organizations do not regulate the crisis. This led to a lack of clarity about the division of duties and powers of each crisis response organization, which led to the sharing of confusing information (idem, p.117).

During the crisis ChemPack, the major of Moerdijk decided to level up to GRIP3 so the structure of CoPI-ROT-GBT was clear and official. This made the role of each team clear and should promote information sharing. However, the GBT did not receive any information from the field. Thus, the team at the top of the chain did not receive the information they needed, making this situation a representation of a vertical structure being problematic in the sharing of information. There are no reasons given as to why. Adding a positive note to that, the ambiguous relationship between the CoPI-ROT and the GBT improved when the head of the safety region became a liaison at the GBT (Dutch Safety Board, 2012, p.86).

5.2.2.2 Liaisons

Which brings us at the second category of vertical structures, liaisons. The crisis MalTech describes two situations in which a liaison is added to one of the teams, with the goal of promoting information sharing and establishing an overview of the crisis. This subcategory contributes to the category of vertical structures, because a liaison adds to the sharing of information between crisis responding teams and not the decision-making within the team. Therefore, they are subordinate to the team, which makes it a vertical structure.

One example shows a positive influence of a liaison on information sharing. A liaison of the ‘Rotterdam port company’ took place in the ROT of safety region

Rotterdam-Rijnmond. This gave him the ability to warn his co-worker on a ship in the harbour that the telecommunication network would soon fail, information coming from the safety region. This led to the shutting down of the harbour, making sure the situation was safe (Inspectorate for Safety and Justice, 2012, p.58).

On the other hand, the Inspectorate for Safety and Justice (2012) describes how a liaison of KPN joined the ROT of safety region Rotterdam-Rijnmond, who could not give any information about the malfunctioning telecommunication network (p.54). This surprises the safety region, because KPN often joins the safety region in their crisis practices (p.39). This situation describes how a structure that should promote information sharing (a liaison) can also cause problems for the sharing of information. More on this situation in section 5.4.

The WeirCol crisis positively encounters a liaison situation. The ROT of Brabant-Noord adds a liaison of Rijkswaterstaat to their team, and a liaison of the water authority ‘Aa en Maas’. Both share relevant information about the situation. For example, the liaison of

32

Rijkswaterstaat reports that concerned parties along the Maas are already alarmed (Dutch Safety Board, 2018, p.73).

5.2.2.3 Equals

Another interpretation of vertical structures is the situation in which there are several equals, who expect one of them to take the lead. This is the case in the PowOut crisis. The power outage hits six safety regions. Five of them expect that the safety region Amsterdam-Amstelland take the lead, because the power outage started in that region. Which it did not. The Inspectorate for Security and Justice (2016) notes that both the source- and effect regions have not been able to initiate interregional coordination. This connects to ambiguity about levelling up to GRIP4 and GRIP5 whenever safety regions have to collaborate simultaneously (p.22). With the confusion on the role of each safety region, caused by an ambiguous system of crisis management, there was no common operational picture established. Because of this unclear common operational picture, it remains unclear how to arrange the alignment of crisis communication with other safety regions, vital partners and other network partners, such as the airport, public transport agencies, and organizations affected by the power outage (p.29).

5.3 Conflict of Jurisdiction

Conflict of jurisdiction is the situation in which one party claims to have/or has the right to execute certain authorization about a place or task. The analysis did not encounter that situation. However, another form of conflict of jurisdiction did influence the sharing of

information in one situation, during the Shell crisis. The safety region of Midden- en West-Brabant had sent a text message through the national alert system. However, after

consultation, it appeared that the text message was not sent to the inhabitant of the other safety region, due to the policy that one safety region is not allowed to send such text messages in another safety region. The safety regions corrected this flaw in jurisdiction, and within an hour, everyone who needed the information received a new text message (Dutch Safety Board, 2015, p.84-85).

5.4 Information of Private Organizations

During some crises, the information of a private organization is crucial in managing the consequences of a crisis. This was the case during two crises: MalTel and WeirCol. First, the crisis caused by the malfunction of the telecommunication network of KPN. The