GIS-based crisis communication: A platform for authorities to communicate with the public during wildfire

(1)

IN

DEGREE PROJECT THE BUILT ENVIRONMENT, SECOND CYCLE, 30 CREDITS

,

STOCKHOLM SWEDEN 2019

GIS-based crisis

communication

A platform for authorities to communicate with

the public during wildfire

FELIX ALTHÉN BERGMAN

EVELINA ÖSTBLOM

KTH ROYAL INSTITUTE OF TECHNOLOGY

(2)

Preface

By finalizing our education with this thesis, at the Transport and Geoinformation Technology master’s program at the Royal Institute of Technology (KTH) in Stockholm, we were able to combine our main interests into something that could provide inspiration to further develop one of the most important functions of our society. Our main interests include communication and the behaviour of people combined with the power of geographic data and geographic information systems. These lead us not only to apply for the master’s program in Transport and Geoinformation Technology, but also to choose the topic of crisis communication using maps for the thesis.

Throughout the thesis work we have had a lot of support from Agima Management AB, where the work has been conducted. Therefore, we would like to thank everyone there for the support and ideas they provided. A special thanks to our supervisor Mikael Grönkvist for swift and accurate feedback as well as guidance when we needed it.

Besides that, Susanne Ingvander at MSB deserves a sincere thanks for taking her time providing us with information, feedback and data. Further, we would like to express our gratitude to our supervisor at KTH, Andrea Nascetti, for taking his time and helping us solve technical errors during the development of the prototype.

Finally, thanks to our supervisor Gyözö Gidofalvi and examiner Prof. Yifang Ban for supervising and examining the thesis.

Felix Althén Bergman & Evelina Östblom Stockholm, Sweden

(3)

Abstract

Today, people are used to having technology as a constant aid. This also sets expectations that information should always be available. This, together with ongoing climate change that has led to more natural disasters, has laid the foundation for the need to change the methodology for how geographical data is collected, compiled and visualized when used for crisis communication. This study explores how authorities, at present, communicate with the public during a crisis and how this can be done in an easier and more comprehensible way, with the help of Geographical Information Systems (GIS). The goal is to present a new way of collecting, compiling and visualizing geographical data in order to communicate, as an authority, with the public during a crisis. This has been done using a case study with focus on wildfires. Therefore, most of the work consisted of the creation of a prototype, CMAP – Crisis Management and Planning, that visualizes fire-related data.

The basic work of the prototype consisted of determining what data that exists and is necessary for the information to be complete and easily understood together with how the data is best implemented. The existing data was retrieved online or via a scheduled API request. Event-related data, which is often created in connection with the event itself, was given a common structure and an automatic implementation into the prototype using Google Fusion Tables. In the prototype, data was visualized in two interactive map-based sections. These sections focused on providing the user with the information that might be needed if one fears that they are within an affected location or providing the user with general preparatory information in different counties. Finally, a non-map-based section was created that allowed the public to help authorities and each other via crowdsource data. This was collected in a digital form which was then directly visualized in the prototype’s map-based sections.

The result of this showed, among other things, that automatic data flows are a good alternative for avoiding manual data handling and thus enabling a more frequent update of the data. Furthermore, it also showed the importance of having a common structure for which data to be included and collected in order to create a communication platform. Finally, by visualizing of dynamic polygon data in an interactive environment a development in crisis communication that can benefit the public’s understanding of the situation is achieved.

This thesis is limited to the functionality and layout provided by the Google platform, including Google Earth Engine, Google Forms, Google Fusion Tables etc.

(4)

Sammanfattning

I dagens samhälle är människan van vid teknik som ett ständigt hjälpmedel. Detta sätter också förväntningar på att information alltid ska vara tillgänglig och uppdaterad. Detta tillsammans med pågående klimatförändringar som lett till fler och svårare naturkatastrofer har lagt grunden till att det finns ett behov av att förändra hur man samlar in, sammanställer och visualiserar geografiska data som används för kommunikation i en krissituation.

Denna studie utforskar hur myndigheter, i dagsläget, kommunicerar med allmänheten vid en krissituation och hur detta kan göras på ett enklare och mer givande sätt med hjälp av GIS. Målet är att visa ett nytt sätt att samla in, sammanställa och visualisera geografiska data för att, som myndighet, kommunicera med allmänheten under en kris. Detta har gjorts som i en fallstudie med fokus på skogs- och gräsbränder. Merparten av arbetet bestod därför av framtagande av en prototyp, CMAP – Crisis Management and Planning som visualiserar brandrelaterade data. Grundarbetet till prototypen bestod av att fastställa vilken data som finns och är nödvändig för att informationen skulle bli lättförstådd och komplett samt hur denna bäst implementeras. Den existerande data som implementerades hämtades online eller via ett schemalagt anrop av APIer. Händelserelaterade data skapas ofta i samband med själva händelsen och därför skapades en gemensam struktur och direktimplementation till prototypen för denna data med hjälp av Google Fusion Tables. I prototypen visualiserades data i två interaktiva kartbaserade sektioner. Dessa sektioner fokuserade kring att förse användaren med den information som kan behövas om man befarar att man befinner sig på en drabbad plats eller att förse användaren med allmän förberedande information inom olika län. Slutligen skapades även en icke kartbaserad sektion som möjliggjorde att allmänheten kan hjälpa myndigheter och varandra genom ”crowdsource” data. Denna samlades in i ett digitalt formulär som sedan direkt visualiserades i prototypens kartbaserade delar.

Resultatet av detta visade bland annat att automatiska dataflöden är ett bra alternativ för att slippa manuell hantering av data och därmed möjliggöra en mer frekvent uppdatering. Vidare visade det även på vikten av att ha en gemensam struktur för vilken data som ska inkluderas och samlas in för att skapa en kommunikationsplattform. Slutligen är visualisering av dynamiska polygondata i en interaktiv miljö en utveckling av kriskommunikation som kan gynna förståelsen för situationen hos allmänheten.

Studien är begränsad till att skapa en plattform baserad på den inbyggda funktionaliteten och designen som erbjuds i Googles plattform, detta inkluderat Google Earth Engine, Google Formulär, Google Fusion Tables etc.

(5)

List of figures

Figure 1. Progress report of wildfires in Gävleborg County, 2018-07-18 20:00. ... 9

Figure 2. The front page of Krisinformation.se, 2019-05-16. ...11

Figure 3. The different pages of the application provided by Krisinformation.se, 2019-06-05. .... 12

Figure 4. Schematic description of the research strategy for the thesis. ... 16

Figure 5. The compiled fictive information regarding the evacuated areas presented in the prototype. ... 19

Figure 6. The compiled fictive information regarding the estimated fire areas presented in the prototype. ... 20

Figure 7. The compiled fictive information regarding the suspended lakes presented in the prototype. ... 20

Figure 8. Parts of the fire ban data compiled in a fusion table used in the prototype. ... 21

Figure 9. The initial page of the prototype, 2019-05-13 11:00. ... 29

Figure 10. The prototype’s main menu 2019-05-13 11:00. ... 30

Figure 11. The location-based information section of the prototype 2019-05-13 11:00. ... 31

Figure 12. Showing the drop-down selector and how STA announcements are presented, 2019-05-13 11:00. ... 32

Figure 13. The visualization if no event is in vicinity to the chosen location, 2019-05-13 11:00. ... 33

Figure 14. Information presented when chosen location is in vicinity to an estimated fire, 2019-05-13 11.00. ... 34

Figure 15. Information presented when chosen location is within estimated fire area 2019-05-13 11:00. ... 35

Figure 16. Information presented when chosen location is within safe distance to a fire, 2019-05-13 11:00. ... 36

Figure 17. The initial area-based information section of the prototype, showing the national information, 2019-05-13 11:00... 37

Figure 18. Information displayed about two current events, 2019-05-13 11.00. ... 38

Figure 19. The area-based information section for Stockholm county, without any information visualized in the map, 2019-05-13 11.00. ... 39

Figure 20. The area-based information section for Stockholm county, with all information displayed in the map, 2019-05-13 11:00. ... 40

Figure 21. The map showing the fire, suspended lake, evacuated area, possible evacuation zones and an available evacuation accommodation, 2019-05-13 11.00. ... 41

Figure 22. The map showing the fictive fire, suspended lake and evacuated area, 2019-05-13 11.00. ... 42

Figure 23. The map showing the fictive fire and the Possible future evacuation zones, 2019-05-13 11.00. ... 43

Figure 24. The general information in the Stockholm county, valid 2019-05-13 11.00. ... 44

Figure 25. Post event information, 2019-05-13 11:00. ... 44

Figure 26. The help menu of the prototype, 2019-05-13 11:00. ... 45

(6)

List of tables

Table 1. Data used in the location-based information section of the prototype. ... 27 Table 2. Data used in the area-based information section of the prototype. ... 28

(7)

List of Abbreviations

API Application Program Interface

CEP Civil Emergency Planning

GCS Google Cloud Storage

GEE Google Earth Engine

GIS Geographic Information System

GUI Graphical User Interface

ID Identifier

KML Key Markup Language

KTH Kungliga Tekniska Högskolan (Royal Institute of Technology)

MSB Myndigheten för Samhällsskydd och Beredskap (Swedish Civil Contingencies Agency)

ODC Open Data Cube

SMHI Sveriges Metrologiska och Hydrologiska Institut (Swedish Metrological and Hydrological Institute)

STA Trafikverket (Swedish Traffic Agency)

(8)

1 Introduction

1.1 Background

Societies have always been changing over time, people learn from and develop each other as well as the structure and technique of the society. Today's society is no exception since the digitization influences almost all organizations. Hospitals are dependent on a working IT-system, more stores are available online than on the streets and whole professions might soon be exchanged for a robot or online solution. This does not only change the way companies and authorities' function, but also what people expect and demand from them. When customers can track deliveries that they have ordered from shipping to arrival in real time, do a whole week’s worth of grocery shopping from a mobile phone or meet a doctor online at any time of the day, the expectation of correct and personalized information, in real time, is bigger than ever before and still growing. This kind of expectation persists, and most likely when it is of most importance, in times of crisis. When an incident occurs, which puts the society and its citizens in direct or indirect danger, the need for relevant and accurate information on what is happening, what will happen and what there is to do in order to be safe is as most crucial.

There are countless number of causes of a national crisis. Some of them being man made and others being natural disasters such as flooding, tsunamis, forest fires. Today the natural disasters are a common cause for a national or even international crisis, e.g., the flooding caused by the monsoon in the south of Asia in 2017, the 2018 Attica wildfires in Greece and the “Camp Fire” which was the deadliest forest fire in Canadas history (Baldassari, 2018).

Due to the climate changes and thereby the global warming, changing of winds and rain, both the frequency and severity of the natural disasters are to increase globally (Flannigan, Stocks and Wotton, 2000).

In Sweden, when a national crisis, local crisis, or bigger incident that could lead to a crisis occurs the authorities are responsible to inform the people affected. The responsible authorities are both the effected counties, the Swedish Civil Contingencies Agency (MSB) and other involved authorities, which could be the police, fire department etc. The information is today communicated using several different channels to make sure that as many people as possible can take part of the information as quickly as possible. The initial, acute information called Viktigt Meddelande till Allmänheten (VMA), in English Important Announcement to the Public, is always communicated through the local radio, TV and from Krisinformation.se through push notices in the application that MSB provides. The VMAs can also be communicated through social media or other applications. Depending on your location, VMAs can be received through an outdoor alarm system called Hesa Fredrik, as well as through text or voice messages to phones (Swedish Civil Contingencies Agency, 2018b).

According to a survey carried out by MSB about people's general attitude about the authorities during a crisis, based on previous experiences, the public was positive to how the initial instructions were communicated and could quickly act on the recommendations from the authorities. Despite this, the myth that the public's reaction to a crisis would be to act irrationally

(11)

and cause panic still lingers (Coombs and Holladay, 2010). This could lead to the authorities withholding parts of the information to avoid amplifying the panic that is, due to the myth, thought to occur in conjunction with a crisis. This withholding of the information could be the explanation of why the survey participants issued that the information communicated during the crisis did not meet all expectations. The expectations that were not met were those of the quality of the information, which was considered poor and the lack of compilation of the information which was published on different authorities’ websites. This yielded that the information was hard to understand and problematic to place geographically. However, an expectation that was met was the update frequency. Hence, the frequency of which the information is published is not an issue for the public, but rather the content, quality, compilation and presentation of the information (Coombs and Holladay, 2010).

1.2 Purpose and research question

The purpose of this thesis is to find new ways to create, manage, compile and visualize geographical data presented by the authorities to the public during a crisis, in order to ensure effective crisis communication. This includes analysing the public’s information need, proposing new workflow structures for the authorities and inclusion of new data presented in a prototype of a communication platform. This is to make sure that as many people as possible receive the information necessary, that the information is easy to interpret and understand, and that the handling of the crisis will be as simple as possible for all parties involved. Hence, the research question for this thesis is

How can authorities’ geographical data be collected, compiled and visualized using GIS to assure an effective crisis communication to the public?

1.3 Limitations

To be able to make a detailed analysis of what information is needed during a crisis and how that information is best collected, compiled and presented, the thesis is limited to a specific type of crisis. Hence, only larger fires, such as forest fires and grass fires will be handled in the prototype. Another limitation is that only the communication between authority and the public is handled, thereby excluding the internal communication between authorities. Lastly, the chosen platform to build the prototype limits some functionality since only the predetermined functions and tools in the prototype can be applied when building the prototype.

1.4 Target group

In line with the purpose for the study, the target group for the thesis is indirectly the general public, since they will be the group that hopefully will benefit from the suggested communication platform that meets their expectations regarding crisis communication. More directly, the target group is MSB that can use the results of the needs analysis as well as the suggested communication platform as a proposition on how a future, more simplified and efficient platform can be structured.

(12)

3

2 Literature study

The literature study provides an overview of the previous studies in the field of crisis communication together with an overview of today's situation regarding the communication between authorities and the public during a national crisis in Sweden. Lastly an overview of the requested changes in the communication structure is briefly mentioned.

2.1 Related work

2.1.1 Crisis communication

Crisis management is a well analysed topic that has been studied in countless number of articles, studies and theses. Since progression in this field could lead to a palpable abatement in casualties, deaths and properties affected. There are several aspects to consider when studying crisis management and of the most important ones is crisis communication, which have also proved to be one of the biggest challenges in crisis management. Therefore, much of the research that studies crisis management handles crisis communication in different ways (Dransch, Rotzoll and Poser, 2010).

There are several studies that purely handle communication with the public during a crisis, such as “The Handbook of Crisis Communication”. This report defines crisis communication to consist of collecting, processing and distributing the information that is needed to address an ongoing crisis. Besides that, it describes different attitudes and behaviours within the public that can be generated by how and when information is distributed, where the importance of consistency of the information was emphasized. This means that information should be disseminated through the same communication channels, at approximately the same time and with a similar content for the public's confidence in the authority to remain. As an extension, a study on how the communication is commonly implemented was performed and the importance of informing the public on how they should behave was highlighted. It was stressed that the public should be informed with instructions and adjusting information that is easy to comprehend (Coombs and Holladay, 2010).

2.1.2 Alternative prototype platforms

2.1.2.1 Open Data Cube

Open Data Cube (ODC) is an open source solution where one can access, manage and analyse geographical data. It provides a foundation of multiple data architecture solutions such as Africa Regional Data Cube, Digital Earth Australia and more. ODC is developed mainly for analysis of temporally-rich earth observation raster data. However, one can also analyse data such as elevation models, geophysical grids, model outputs and interpolated surfaces that can be loaded in to the platform. The platform provides the service of cataloguing large amounts of geographical data. ODC also enables Python based API for the user to query and access data. One potential flaw of the ODC platform is that it solely provides raster data analytics, hence, vector data analyses are not available (Open Data Cube, 2019).

(13)

2.1.2.2 Mapbox

Mapbox is a platform where the user can create their own interactive maps using HTML, CSS, JavaScript and their own WebGL-based JavaScript library. With Mapbox, one can either create one’s own data or add them as an API. Mapbox provides the possibility of creating interactive Web Applications, however, this feature only supports vector datasets. However, raster datasets can be uploaded as tiles but doing so, they are not editable. Mapbox Studio is a feature in Mapbox which lets the user design and style its map to a great extent. This includes for example custom fonts and styles depending on zoom level. Map box also provides the opportunity for the user to analyse data using Turf.js which is an open source JavaScript library. One downside of Mapbox is that it is not free of charge (Mapbox, 2019).

2.1.2.3 Not using a predefined platform

One solution when building a platform is to not rely on an existing platform, instead developing it from scratch. Doing so, one must build the design for a web application, create the necessary server functionalities, crowdsource implementation and collect all data needed for presenting the information. This gives the developer greater freedom when it comes to visualization and custom functionalities, however it requires more extensive work since there are no predefined functions or basemaps etc. provided.

In the development of a platform, client-side and server-side applications has to be set up. Using pre-existing client-side libraries, e.g., Bootstrap, JQuery and AngularJS, one can develop the front-end of the application. Bootstrap, for instance, is a HTML, CSS and JS framework for the development of the design and functionalities of the application. For Bootstrap to function, JQuery is required. It is a JavaScript (JS) library which simplifies the usage of coding in JS (Bootstrap, 2019).

The client-side is the interactive part of the application which needs to be linked to the server-side in order to function. This can be done using Node.js which uses JavaScript to let the developer manage the server. Using Node.js, server data can be created, opened, read, deleted, modified etc. (W3Schools, 2019).

2.1.3 Data

A sign that the traditional way of communicating using maps is changing is the increasing usage of crowdsourced data. As the usage of smartphones equipped with GPS receivers increases, the possibility to utilize their position during a crisis increases as well. This does not only, as highlighted in the article “Volunteered Geographic Information for Disaster Management with Application to Rapid Flood Damage Estimation”, yield a larger amount of data to use as a foundation for decision making when managing a crisis but also provides a possibility for the public to interact during a crisis (Poser and Dransch, 2010). Crowdsourced data can complete

(14)

5

data that other sensors cannot collect and could be used in a similar fashion as the data that today is collected manually by the forces in the field1_.

Fu et al. addresses the complexity of national meta data standards due to their need of describing all properties of geospatial resources. The complexity is explained as an obstacle for compiling with metadata standards and sharing information, where lack of time and funds are not uncommon reasons. Fu et al. emphasize that a solution to this is that the software developer provides processes that simplify, or even automate, the creation of standardized metadata (Fu and Sun, 2010).

Blåljuskollen is a checklist developed by Lantmäteriet and the Swedish Traffic Agency for Sweden’s municipalities. Its purpose is to ensure that the municipalities provide and update quality assured geographical data which is necessary for the emergency personnel to find the right directions when on call-out. Blåljuskollen does not directly relate to crisis communication, however, it handles an important part of this thesis, which is data management. Lantmäteriet mentions that the chain of collection and the provision of geographical information consists of each party's efforts and that the chain is not stronger than the weakest link. Thus, each party must understand the importance of their mission to ensure that the information is properly provided and thus enable the end product to function properly. To assure that the chain does not break, Lantmäteriet and the Swedish Traffic Agency has understood that a structured and standardized data management is important. Blåljuskollen instructs on which datasets that should be provided, how to structure the data and set up attribute values (e.g., spell correctly). Documentation of the data collection processes and responsibility distributions are especially emphasized, so that the update rate and quality have no risk of falling through the cracks (Lantmäteriet, 2019a).

2.1.4 Visualization in crisis communication

That an information broadcast during a crisis is easy to comprehend is of great importance. Siricharoen Waralak wrote, in his study “Infographics: The New Communication Tools in Digital Age”, that the human brain easier understands relationships and patterns if it is visualised. Here infographics is defined as “interactive visual representations of data to amplify cognition” and exemplified as images and symbols. The aim with infographics is stated as to communicate information in a clear and effective way using graphical means. This applies to all types of data, but it is also stressed that maps are considered the best way of distributing geographically based information (Siricharoen, 2013).This theory is reinforced by the conclusions in S. Fitrianie and L.J.M. Rothkrantz’s article “Communication in Crisis Situations Using Icon Language”. This article focuses on communication and decision making during a crisis and the importance of simplifying the interpretation of the information received by a person. It is stated that subjectivity and ambiguity in the information would be reduced by communicating with maps

(15)

and icons. It is advocated that this kind of communication would reduce of faulty decisions when managing the crisis (Fitrianie and Rothkrantz, 2005).

The advantages with visualisation, specifically using maps, is highlighted in yet another study where “Cartography and Visualisation” declares that visualised data is an important part of how knowledge is constructed in the human brain (Orford, Dorling and Harris, 2003). The focus of the article is on visualization of spatial data using maps and differentiates the usage of maps to visual thinking and communication. Here, visual thinking implies an analytical work with the visualised information in an interactive way, whereas communication is done using a static map. However, it is also stated that the interactivity might change with any further technical development (Orford, Dorling and Harris, 2003).

Although the previously mentioned studies state the fact that maps are suitable for communicating spatial information, there are still risks with communicating information visually instead of via text, sound etc. The risks, which are highlighted by M.J. Kraak and F. Ormeling in the book Cartography Visualization of Spatial Data, lie in the interpretation of the visualized information. There is no guarantee that the information is interpreted as the creator of the map intended it. However, since it is a well-known risk, it is prevented using cartography that defines standards for how information should be communicated (Kraak and Ormeling, 2013).

Despite the extensive research on the field of communication, in general, using maps and during a crisis, not all aspects have been covered. This due to that most of the research is categorised similarly to the article “Cartography and Visualisation”, but only focuses on the communication between the individuals or teams working with mitigating the crisis, either between authorities or from the public using crowdsourced data. This includes experts, decision-makers, forces in the field and volunteers. However, one salient stakeholder that is seldomly included in the studies is the public. This is mentioned as a dearth in the research in the article “The contribution of maps to the challenges of risk communication to the public”(Dransch, Rotzoll and Poser, 2010). In this article it is analysed how the communication from the authorities to the public could be simplified and improved using visualization of spatial data in an interactive format. Initially the article analyses the challenges with informing the public. Some of the main challenges mentioned are how the public perceive the information, how the information is presented and disseminated. Regarding the public's perception, normalcy bias is mentioned as a challenge. However, regarding the dissemination and presentation of information, the choice of target group, vividness and interactivity of the information is mentioned as the biggest challenges (Dransch, Rotzoll and Poser, 2010). Later in the article, the optimal dissemination and presentation of the information is discussed in the context of the public's perception. The article’s detailed analysis advocated the usage of spatial data, interactive platforms and GIS for enhanced understanding of information. Unfortunately, it focuses mainly on distributing information that could prevent a crisis.

As mentioned, an extensive amount of research has been made on the individual subjects that form the concept of communication during a crisis. This includes communication and

(16)

7

management, spatial data, GIS and new techniques such as crowdsourcing. It also emphasizes the absence of research that compiles all those subjects with focus on communication between the authority and public during an ongoing crisis.

2.1.5 International examples of fire crisis communication

In U.S.A, the state of California is heavily affected by wildfires due to their dry and hot climate. The Department of Forestry and Fire Protection, also called Cal Fire, do not only protect the people and property of California, they also provide a web page service where citizens can access information on the ongoing situation of fires. The application is called California Statewide Fire Map and is embedded in the Cal Fire website. The application is map-based and interactive in the sense that the user can zoom in and out in the map and click on one of the objects to retrieve more information. The objects are in point shape and either as a red flame of fire (active fire), a black flame of fire (contained fire) or red flame of fire with a lightning (lightning complex). The additional information to the map feature is a name of the fire and a descriptive text explaining the situation of an object, e.g., close to what road, estimation of size, the date and time when the information was added as well as updated and how much of the fire that is contained.

Cal Fire also provide a mobile application called CAL FIRE Ready for Wildfire. This application contains four sections. One which serves to help people prepare for a wildfire with a checklist. Another which is an interactive fire map with features like the map on the website. A third which is a list of the displayed fires in the map and the fourth is providing fire news which is information directly from the Cal Fire twitter account (Cal Fire, 2019).

This kind of application is similar to the one provided by MSB today. The difference is the information content in the application, which is specifically focused on wildfires. This functionality could be implemented in the application Krisinformation.se as well. However, the application CAL FIRE Ready for Wildfire still has some shortcomings, which include using only point data instead of polygon and line data.

2.2 Situation analysis

In Sweden today, there is a civil emergency planning, CEP, system which is based on the society’s ability to prevent and manage crises. The CEP of Sweden is based on three principles which lay the foundation for how a crisis should be managed for normal functions of the society to be kept during a crisis (Swedish Civil Contingencies Agency, 2018a).

• Principle of responsibility – The authorities which has a responsibility normally also has a responsibility when there is a crisis. This also includes supporting and coordinating with other authorities.

• Principle of proximity – Sweden is divided into counties and on a lower level; municipalities with a large amount of autonomy. This autonomy entails a responsibility within CEP, thus, if there is an incident, the responsibility of management is associated to the municipality where it occurred. Support and assistance are provided from the appurtenant county.

(17)

• Principle of similarity – When an incident happens that disturbs the functions of the society, the authorities are not to change their organization. Thus, they should work as they would under less extraordinary conditions.

This structure of crisis management is also applied on the responsibility of alerting and communicating with the public in times of crisis. Thus, each county shall, with its containing municipalities, communicate information concerning an incident or crisis (SFS 2015:1052, 2015). During the summer of 2018 in Sweden, high temperatures and minimal rainfall resulted in many wildfires spreading throughout the country whereas Gävleborg County was one of the most affected counties. In an early stage of the summer, 9500 acres of forest fires were reported in Ljusdal municipality, situated in Gävleborg County (SOU 2019:7, 2019). To grasp the spread of 9500 acres of wildfire in Ljusdal, it can be compared to approximately 13 500 football fields burning at that time. The communication from Gävleborg County to the public during the fires of 2018 was with high frequency. Up to five information updates per day could be received through their website during the prevailed crisis of one month (Länsstyrelsen Gävleborg, 2018b). The information that could be accessed was via text and complementary maps uploaded as pdf files. Information such as the location of the fire, the upcoming weather forecast, which measures were taken by the different authorities and recommendations to the public could be accessed through the daily progress reports. Figure 1 shows a complementary map of the situation on the 18th of July. It is a “dead map”, i.e., non-interactive, nonetheless giving an overview of the situation in both text and visualization.

The maps that Gävleborg County provided were produced through helicopter flights over the burning area whereas maps were drawn by hand. With complement from the coastguard’s IR scans over the fire, maps could be digitalized and then uploaded on the website2_.

(18)

9

Figure 1. Progress report of wildfires in Gävleborg County, 2018-07-18 20:00.

Gävleborg County used the wildfires' spatial attributes in order to present the prevailed situation through maps. However, this was not the case for all counties and municipalities. Dalarna and Jämtland County, the adjacent counties of Gävleborg were also highly affected by the fires. They did not produce maps to communicate to the public. In their case, solely text was used in order to describe the situation.

Due to the many counties, as well as the municipalities within each county, there is a big stream of information that is published in times of a crisis. Since, due to the principles, every organization or authority close to the occurrence must communicate using their primary channels. That yields a flow of information in many different channels. Regularly, information is distributed using regional radio and TV. It is also published on the authorities' web pages, mainly the municipalities’ and counties’ web pages, which is also shared on their social media pages. Besides that, the commercial media publishes news about the happenings as well.

In order for the public to be able to receive relevant and accurate information in times of crisis, MSB has the Government's mandate to coordinate information from all different channels to the public (SFS 2008:1002, 2008).

During the wildfires of 2018 in Sweden, MSB worked as a coordinator of the information published by the different authorities to the public. Through press conferences with the media, they gave statements on the extent of the fires as well as progress reports on the measures taken

(19)

to handle them. Since many people use the media as a source of information when in crisis, this was a vital part in communicating information about the prevailing situation.

2.2.1 Krisinformation.se

MSB is also handling a communication platform called Krisinformation.se. It is available both as a website and as a mobile application. This is their main platform where they coordinate all relevant information to the public. However, Krisinformation is not a provider of information in the sense that they supply their own information. They are solely publishing confirmed information from responsible actors; authorities, organizations and companies/networks with responsibilities of essential services for the society. They also share the news published on the web page on social media such as Facebook, Instagram and Twitter in order to reach out to as many as possible3_.

The content published on the website and the application of Krisinformation.se differs. The website’s content is more extensive in the sense that it contains a wider range of information in terms of pre, during as well as post crisis. Here, the public can access information about what threats there are against the society, how to prepare for them as well as historical events and how one can learn from them. During a crisis or events that affects the society, information, VMA’s and news from authorities are published on the front page of the website, see Figure 2. These posts are normally summaries of the main information for the reader to get a quick grasp on the most relevant information. If the reader wants more detailed information, the link to the original source is always attached4_.

3_{Ingvander, Susanne, Lägervik, Amanda, Ullén, Jeanna. Administrator in GIS and web editors at MSB. Interview. 2019-02-13} 4_ibid.

(20)

11

Figure 2. The front page of Krisinformation.se, 2019-05-16.

Through a list of affected counties and municipalities published on MSB’s website, the reader can access their websites to get more knowledge about the fires in that specific area (Swedish Civil Contingencies Agency, 2018-08-24). There were no map features available directly on Krisinformation’s website. The only information presented as maps was from original sources such as SMHI (Swedish Metrological and Hydrological Institute) and Gävleborg County which could be accessed through their associated websites.

During an interview made with representatives from MSB, Susanne Ingvander, GIS administrator at MSB, emphasizes in an interview the difficulties with publishing progress reports as dynamic maps in times of a crisis such as the fires of 2018. Due to the dynamic nature of the crisis; areas that were burning in the morning could be put out in the evening and vice versa, creates hardships in ensuring the accuracy of the map. This factor can result in inaccurate information

(21)

updates which in turn can lead to casualties. Due to this issue, maps are provided statically, “dead maps”, with time stamps which refers to how the situation was at that certain time5_.

In comparison to the website, the application of Krisinformation is map based in its foundation. When the app is opened, the user is faced with posts of current warnings and information updates in form of a list. By switching to the map feature, these posts are visualized as points on a map. Since the map is interactive, the user can zoom in to its area of interest, click on a point and receive a short summary of the incident with a link to the source of the information. The user also has the possibility of showing events based on county regions. The application and its map are also interactive in the sense that there are settings to choose from in the menu. The user has the possibility of turning push-notices on for VMAs and the map feature can either be shown in hybrid, normal, satellite or terrain. For Android users the possibility of showing Google’s traffic information is also provided.

Figure 3. The different pages of the application provided by Krisinformation.se, 2019-06-05.

When it comes to the data sources, the app is providing four sources of information which you can toggle between on the interactive map to show their associated content. The sources used in the app are provided from the Swedish Transport Administration (STA), SMHI and Krisinformation as well as the mobile software’s own information of the location of hospitals and police departments. The information updates from the STA and SMHI are provided through APIs. The benefit is that the information is provided and published instantly when the STA or SMHI supply them. However, this system has its limitations. The format which the source creates its data establishes strict boundaries on how the data can be visualized in the app. If a point layer is provided, the data will only be able to be presented as points. Further and as one can see in

(22)

13

the Figure 3 above, there is no distinction between the different sources in the symbology of points6_.

During the wildfires of 2018, SOS Alarm provided information and data which was published in the application of Krisinformation. This is, however, not a service which is normally provided. The reason is that their data is highly identity protected, thus not able to be provided as an API. For SOS Alarm to be able to provide data, segments of the data must be changed to protect the identity of the persons it concerns. Consequently, the data was provided three times a day during the fire crisis as an Excel sheet whereas some columns of information was deleted7_.

Jeanna Ullén, web editor at Krisinformation, states in the mentioned interview with MSB representatives, that the dream situation of data sources would be if all counties and municipalities would provide their data to Krisinformation. However, she emphasizes the importance of a standardized data format where they would provide the information and thereby realize the dream. Jeanna also mentions the police and fire department as possible future data sources8_.

The symbology used when visualizing the spatial attribute of the events in the application is solely points. When placing the points on the map, the editorial staff of Krisinformation uses a hashtag function where they write the concerned region, e.g., Gävleborg County, as a hashtag which then automatically places the point in the centre of the region. If there is knowledge about the exact position of the event and it is of importance, the point can be placed manually but normally rapid publication of the information is prioritized. The current app does not support visualization in form of polylines or polygons. That is, if a fire is to be reported as an event, the point shape is used. This applies also when there has been a traffic accident between two traffic interchanges. In that case, the event is visualized as a point on the road segment with complementary text which describes the two traffic interchanges that are affected9_{. One can}

argue that this type of data management and presentation do not fully utilize the full potential of geospatial information.

2.3 Needs analysis

Despite regulations and technical development specialised on quick and accurate information distribution, the public’s attitude regarding the communication structure today is not unambiguously positive. Studies of the Swedish public's attitude regarding the authority's communication during crisis shows that less than half of the people have a positive attitude towards how the information is distributed today. Peoples’ attitude differs based on the magnitude of the crisis and which authority that is responsible for the communication. People

6_{Ingvander, Susanne, Lägervik, Amanda, Ullén, Jeanna. Administrator in GIS and web editors at MSB. Interview. 2019-02-13} 7_ibid.

8_ibid. 9_ibid.

(23)

are in general more pleased with the communication from the authorities if there is a national crisis rather than a local crisis, where only about one third is positive (Enkätfabriken, 2019). There are two main reasons the communication is perceived as negative by the public. Firstly, there is a perception that the information is insufficient (Swedish Civil Contingencies Agency, 2013). Secondly, that the information is posted by several different sources, through several different channels. This interferes with the possibility to create a broad picture of the situation and leads to a concern within the public, since important information could have been missed (Swedish Civil Contingencies Agency, 2014).

The fact that people find the information confusing and request a web page and application that gathers the information shows that Krisinformation needs to be better advertised and that how information is provided on the pages needs to be improved. Today only one out of four people are familiar with Krisinformation.se and that only one out of one hundred have, if they were informed about an incident, gotten that information from Krisinformation.se (Enkätfabriken, 2019).

Despite this, 63% of people state that they would use a web page, online search engine or application to search for information during a crisis. This shows that the usage of a web page and application is suitable, but that the information presented is not enough. The kind of information that is requested is local information and information that could help prepare for an incident or act quickly after a VMA. The importance of information being objective, clear, easy to find and understand is highlighted (Enkätfabriken, 2019).

(24)

15

3 Methodology

This section firstly addresses the general strategy on how the thesis project was carried out. Secondly, an introduction to the case study is made, followed by definitions of technical terms that are recurring in the thesis. An elaborate description of the data collection method is then brought up. The datasets used in the development of the prototype are listed in Section 3.3 with a description on their source, format and what they were used for in the prototype. Further, the method of processing data is then explained. Additionally, the theory behind how the user interface was designed is clarified. Lastly, the procedure for visualizing data in the prototype is described.

3.1 Research strategy

The thesis main goal is, as mentioned, to create a prototype for communication between authorities and the public during a national or local crisis caused by a wildfire. The application should contain all necessary information and present it so that the information is easily understood even during the stressful situation a wildfire could generate. To ensure this, a list of requirements was set as a base for the prototype. The requirements were based on an analysis of how the public wishes to be communicated with and how the communication is performed today together with what possibilities today’s technology might provide if used in a different way than previously. The requirements are set both for what information, and thereby data, that is needed, and the functionality of the user interface in the platform. Based on the requirements, the prototype is created.

Google Earth Engine (GEE) was the platform used to develop the prototype. With its already implemented map interface, the need of developing a platform from scratch could be avoided. Also, it has an extensive catalogue of datasets which can be implemented using GEE’s existing analysis algorithms. Further, since it is linked to Google's other services, such as Google Fusion Table, Google Cloud Storage and Google Sheets, data can easily be structured and incorporated into the platform. Consequently, Google Earth Engine was considered an appropriate development platform for the prototype.

In an early stage of the development process of the prototype, the frame for the user interface is created and the data collected. The main types of data for the prototype will be dynamic data and static data, i.e., data that changes over time and data that seldomly or never change, see Section 3.3. Lastly the prototype will be tested, with focus on functionality and content. See Figure 4 for the schematic description of the research strategy.

(25)

Figure 4. Schematic description of the research strategy for the thesis.

3.2 Case study

Since the wildfires of 2018 are close in time when writing this thesis and there is documentation on how the general public perceived the authorities’ communication of crisis information, it is of relevance to apply the wildfires of 2018 as a case study for this thesis. Hence, the prototype of the communication platform will be developed for scenarios of national wildfires in Sweden. This means that the data collection and presentation, the design of the interface and the functions within the prototype are developed to satisfy the communication of information from the authorities to the public when in times of a crisis caused by wildfire.

(26)

17

3.3 Definitions

Crisis can be defined on a broad spectrum whereas it differs depending on the nature of the situation. The Handbook of Crisis Communication concludes that there is not a universally accepted definition of crisis, however, the conceptual similarities such as unexpected, negative and overwhelming permeate them all. The Handbook of Crisis Communication mainly treats the definition of crisis in an organizational perspective. However, they highlight the importance of not limiting crisis to only have impact on organizations, instead the harm (physical, financial or psychological) on stakeholders should be rated as the most significant negative outcome of a crisis. This angle of the definition is in line with the focus of this study; to develop an aid for the stakeholders, i.e., the public, to be able to manage a crisis and stay out of harm. To conclude, in this study, a crisis is defined as a serious, unexpected and overwhelming event which can lead to direct or indirect negative outcomes on the stakeholders, i.e., the general public (Coombs and Holladay, 2010).

Anaconda is a program that is used to connect a python environment to GEE, using the Earth Engine API. This is needed to ingest data provided in API form.

API, Application Programming Interface, is the interface which lets software applications communicate with each other. By using an API request to a data server, a response containing the requested data can be ingested in an application.

Crisis communication is defined in The Handbook of Crisis Communication as the collection and processing of information along with the creation and dissemination of crisis messages to people outside of the crisis management team. As clarification, since the word communication can be interpreted as bidirectional, crisis communication is referred in this study as the unidirectional communication from authorities to the public (Coombs and Holladay, 2010). Crontab is a program that enables a user to schedule runs of scripts or commands on a computer based on the Unix operating system, i.e., Linux or Mac. It uses crons, or daemons, that only needs to be set up once. After that it will be dormant until it is scheduled to run. It does not consider day saving time and will not run if the computer or server is turned off (Wallen, 2011).

Forest fires are uncontrolled fires in forests and other land. It is mainly caused by negligence when lighting fires or by the strike of lightning (Nationalencyklopedin, 2019).

Fusion table is a web application hosted by Google, used to gather, share and visualize data tables. It is in the prototype created in this thesis used to gather manually created data from municipalities, counties and other authorities to make the information distribution standardized.

Google Cloud Storage (GCS) is a cloud server solution provided by Google. It provides easy storage and distribution to other Google products and can be handled using script that bases on googles own version of java script (Google, 2019).

(27)

Google Earth Engine (GEE) is providing the tools and satellite data to make scientific analyses of large geospatial datasets and visualising them in an application. The applications are programmed using a version of the programming language Java, with Googles own predefined functions. Any data not included in the platform can be added as an asset or by using Fusion Tables and Sheets (Gorelick et al., 2017).

Python is used as the programming language for handling the data requested by API. Due to it being, using Anaconda, compatible with the Google environment. Python version 2.7 is used due to achieve the best combability.

3.4 Choice of methods for data collection

3.4.1 Static data

The static data is rarely modified and will work as base features for the map interface. Emergency rooms, evacuation shelters and county borders are examples of static data. These datasets are created or collected and implemented beforehand with the prototype.

3.4.2 Dynamic data

Dynamic data, i.e., data that is rapidly changing, is yet to be used for crisis communication in an interactive platform. This is mainly due to the frequency of which the occurrence changes rarely being the same as the frequency of which the information could be updated. This yields a risk of presenting inaccurate information10_{. In conclusion, today’s way of presenting the situation of a}

fire event is done by using dead maps with a time stamp for that specific time. However, with meticulous and structured collection of the manually created data together with usage of automated data flows, the dynamic data can be presented accurately in the prototype.

3.4.2.1 Data from online sources

The dynamic data that was collected online is weather warnings from SMHI, traffic warnings from the STA and VMA from MSB. This was achieved using a Python script that collects, converts and uploads the data. It was collected from each publisher's respective database using an API request. In the API request, only the necessary information was filtered out and collected. The response to the API request was then converted into a suitable file format for geographical data, in this case the Shapefile format, since it is compatible with GEE’s asset upload unlike the API request. The data was then uploaded to an online data storage, GCS, from where it was ingested into GEE as an asset. The function to upload data to GCS and ingest data into GEE from a local python script was enabled by connecting a python environment to the GEE API. The connection enabled the usage of all functions and tools that GEE provides in the online code editor in a local python script together with the possibility to upload data from a computer or server to the GEE platform, via GCS. The connection was done using Anaconda.

(28)

19

To make sure that the data in the prototype is frequently updated, so that the user can get near real time information, the script must be run frequently. This was automated to run hourly on the computer that the script was stored on. The automatization is created using Crontab, with the specific cron that, as mentioned, schedules a run of a specific command or script and executes it on set schedule.

3.4.2.2 Coordinate data from online authorities

Today, through the principle of responsibility, each authority must provide data and information concerning an event. This is commonly done locally, for instance within a municipality or county. Thus, relevant data exists but is distributed on different websites and platforms. Beyond that, different data structures are common. In order to coordinate the data from the different authorities, a standardized data structure is vital for the ingestion of data into the prototype to work properly. By using Google Fusion Tables with predetermined structures, several representatives can use the same Fusion Table to upload their specific information, without having to collaborate with each other to determine what information should be included. Due to the possibility that any representative from the authority can manage and create spatial data on another platform, store it as a KML file and directly upload it into an existing Fusion Table. When the data is added to the Fusion Table it is directly updated in the prototype thanks to the import of the data using a link request. In the prototype, several event related data sources are compiled in fusion tables, see Specification of data and Figure 5, Figure 6 and Figure 7. Here a common structure including dates and times for timestamps, area identification IDs, descriptions and geometries are included as a standard in the structure.

(29)

Figure 6. The compiled fictive information regarding the estimated fire areas presented in the prototype.

(30)

21

An example of non-event related information that would benefit of this kind of coordination is fire ban information. Due to them being determined by the municipalities and as of today only issued on each municipality's web page. Further, since it is also very useful for information connected to a crisis such as a forest fire, it is compiled in a fusion table for the usage in the prototype, see Figure 8.

Figure 8. Parts of the fire ban data compiled in a fusion table used in the prototype. 3.4.2.3 Crowdsource data from the public

Another functionality in the prototype that has yet to be fully provided in today’s crisis management, is the opportunity for the public to help each other. During the crisis of 2018 Gävleborg county compiled contact information to people that could provide evacuation accommodation for farm animals but only locally on the web page (Länsstyrelsen Gävleborg, 2018a). This is in the prototype further developed to a crowdsource function for providing evacuation accommodation. The municipality is responsible for providing accommodations for people that needs to evacuate. The accommodations are often located in larger facilities such as schools etc. The crowd source functionality would provide a possibility for people to evacuate together with their pets or farm animals or for evacuation of only farm animals. The functionality is enabled by the usage of the Google platform, since all Google’s products can be connected. The crowdsourced data is collected using a Google Form. There, the user enters the location for

(31)

and information about the accommodation together with their contact information. The answers collected in the form is compiled in a Google Spreadsheet which in turn is connected to a Google Fusion Table. The connection is created in G Suite Developer Hub, i.e., Googles developer platform, using a pre-existing script provided by Lee Clontz on the community GitHub (Clontz, 2014) and the built-in trigger functions. This yields that every time a user submits the form, the information is stored in the Google Spreadsheet, the location is geocoded into coordinates, the Google Fusion Table is updated, and the information is directly ingested into the prototype using a link request.

3.5 Specification of data

Available evacuation accommodation

Source: Crowdsourced using Google Forms. Format: Point data.

Refresh rate: Dynamic, updated when a new entry is made. Insertion to the prototype: Fusion table.

Function in the application: Provide location and information about evacuation accommodation.

Attributes: Through the Google Form, the user enters the needed information to establish an evacuation accommodation for those in need. Attributes such as the location of the accommodation, the time span when the accommodation is available, how many people, pets and farm animals there are room for as well as the user’s contact information, name and telephone number are included.

Counties

Source: Lantmäteriet (Lantmäteriet, 2019b) Format: Polygon data.

Refresh rate: Static.

Insertion to the prototype: Manually.

Function in the application: Visualize the Swedish counties to enable choosing areas to get information within.

Attributes: The name of the counties in Sweden.

Emergency rooms

Source: Manually created based on Vårdguiden (Vårdguiden, 2019). Format: Point data.

Insertion to the prototype: Fusion table.

Function in the application: Provide location and information about the emergency rooms.

Attributes: The name of the emergency room, the associated telephone number and a link to the emergency room’s website.

Estimated fire area

Source: Authority coordinator. Format: Polygon data.

Refresh rate: Dynamic, manually updated when changed. Insertion to the prototype: Fusion table.

(32)

23

Attributes: Name of the affected area which works as an ID, description of what type of fire there is and other important information, the area of the fire as well as the time for when the dataset was first uploaded and last updated.

Evacuated areas

Source: Authority coordinator Format: Polygon data.

Refresh rate: Dynamic.

Insertion to the prototype: Fusion table.

Function in the application: Visualizes areas that are evacuated.

Attributes: The name of the evacuated area which could be a nearby town or other landmark to work as an ID, the date when the evacuation was done and the date when it is safe to return to the area.

Fire ban

Source: The Swedish municipalities Format: Polygon data

Refresh rate: Dynamic, manually updated when changed Insertion to the prototype: Fusion table

Function in the application: Visualize which counties where it is prohibited to start an open fire outdoors

Attributes: Name of the concerned municipality, a description of the fire ban and the time for when the fire ban was last updated.

General information

Source: SOS Alarm (SOS Alarm, 2019a; SOS Alarm, 2019b), Vårdguiden (Vildbird, 2018), Krisinformation.se (Engström and Looström, 2014; Krisinformation.se, 2018a), SVT, MSB (Swedish Civil Contingencies Agency, 2019), the Swedish Red Cross society (Röda korset, 2019).

Format: Text and links Refresh rate: Static

Insertion to the prototype: Manually

Function in the application: Inform the user of ways to prepare for and handle a crisis.

Municipalities

Source: Lantmäteriet (Lantmäteriet, 2019b). Format: Polygon data.

Function in the application: Visualize the Swedish municipalities to facilitate understanding of the map.

Attributes: The name of the municipality.

National and local traffic warnings

Source: The Swedish Traffic Agency (Trafikverket, 2019). Format: Point data.

Refresh rate: Dynamic, updated every hour.

Insertion to the prototype: Automatically using script on server.

(33)

Attributes: Location which works as an ID, the announcement that informs if there, for example, has been an accident or if the road is closed, severity of the event, explanation of the event and its impact, the time when the announcement was last updated as well as the link to the Swedish Traffic Agency’s website.

National and local weather warnings Source: SMHI (SMHI, 2019).

Format: Polygon data.

Refresh rate: Dynamic, updated every hour.

Insertion to the prototype: Automatically using script on server.

Function in the application: Visualize national and local weather warnings.

Attributes: The districts that SMHI has divided the country and surrounding oceans into, the announcement that provides the information about the weather warning, the time when the announcement was last updated as well as the link to the SMHI’s website.

Possible future evacuation areas – Zone 1 and Zone 2 Source: Computed in the prototype

Format: Polygon data Refresh rate: Dynamic

Insertion to the prototype: Included in the platform

Function in the application: Prepare people for a possible evacuation prior to the occurrence of an evacuation

Shelters

Source: MSB (Swedish Civil Contingencies Agency, 2012). Format: Point data

Refresh rate: Static

Function in the application: Provide information about shelter rooms close to the users chosen location

Attributes: The address where the shelter is located, which works as an ID, in which city the shelter is located, how many people the shelter holds and link to MSBs webpage that provides more information about the shelters.

Suspended lakes