Daniel Sempere-Torres is Professor at Universitat Politècnica de Catalunya, Barcelona, Spain. Anastasios Karakostas is Director at DRAXIS Environmental S.A. and former Senior Researcher at the Centre for Research and Technology Hellas, Thessaloniki, Greece. Claudio Rossi is Program Manager and Senior Researcher at LINKS Foundation, Turin, Italy. Philippe Quevauviller is Former Professor at Vrije Universiteit Brussel, Brussels, Belgium and Research Programming and Policy Officer at the European Commission.

• List of Contributors xiiSeries Preface xvi1 The ANYWHERE Paradigm Shift in Responding to Weather and Climate Emergencies: Impact Forecasting, Dynamic Vulnerability and the Need for Citizen's Involvement 1Daniel Sempere- Torres and Marc Berenguer1.1 Disaster Risk Management in Times of Climate Change: The Need of a Proactive Approach 11.2 Adapting Risk Management to the 'New Normality': The Case of Flood Risk Management 21.3 Changing the Paradigm: Impact- Based Multi- Hazard Early Warning Systems to Move from Reactive to Pro- Active Emergency Response Strategies 41.3.1 From Reactive to Proactive Emergency Response Strategies 51.3.2 The ANYWHERE MH- IEWS 91.4 The New Paradigm: Dynamic Vulnerability 131.5 Future Work: From Multi- Hazards to Multi- Risk IEWS 16Notes 17References 182 Hydrometeorological Drought Forecasts: Lessons Learned from ANYWHERE and Next Steps to Improve Drought Management 23Samuel J. Sutanto and Henny A.J. Van Lanen2.1 Introduction 232.2 Method for Forecasting Hydrometeorological Droughts 252.2.1 The Climate (ECMWF SEAS5) and Hydrological (LISFLOOD) Models 252.2.2 The Drought Indices 262.2.3 The Drought Forecast Algorithms 282.3 Hydrometeorological Drought Forecasts 302.3.1 Meteorological Drought Forecasts 302.3.2 Hydrological Drought Forecasts 312.4 Drought Forecast Performance 332.4.1 The Origin of Seasonal Drought Forecast Skill 332.4.2 Examples of Assessment of Seasonal Drought Forecast Performance 342.5 Importance of Catchment Memory 382.6 Outlook and Future Improvements 402.6.1 Drought Impact Forecasts 412.6.2 Compound and Cascading (CC) Dry Hazards 43References 443 Experiences and Lessons Learnt in Wildfire Management with PROPAGATOR, an Operational Cellular- Automata- Based Wildfire Simulator 49Andrea Trucchia, Mirko D'Andrea, Francesco Baghino, Nicolò Perello, Nicola Rebora, and Paolo Fiorucci3.1 Introduction 493.1.1 Mathematical Models for Wildfire Management 503.2 Synopsis of Propagator Development: More than a Decade of Wildfire Simulations 523.3 Propagator Model 553.4 Case Studies 623.4.1 Data Retrieval 623.5 Results and Discussion 653.5.1 Performance Indicators 653.5.2 Performances of Test Cases 703.5.3 An Example of Continuous Improvement and Operational Deployment: Implementation in Ireland 713.6 Conclusions 71References 734 Building an Operational Decision Support System for Multiple Weather- Induced Health Hazards: ANYWHERE Developments and Future Applications 77Claudia Di Napoli4.1 Introduction 774.2 Heatwave Prediction in ANYWHERE 794.2.1 The Universal Thermal Climate Index 804.2.2 Forecasting Algorithms 804.2.3 Heatwave Products 814.2.4 Integration in the MH- EWS 814.2.5 Temperature Products 814.3 Air Pollution Prediction in ANYWHERE 834.3.1 Air Quality 834.3.2 Forecasting Algorithms 854.3.3 Air Quality Products 854.3.4 Integration in the MH- EWS 854.4 ANYWHERE MH- EWS in Action: The European 2017 Heatwave 864.5 Implementation at Pilot Sites 874.5.1 Integration of Local Heatwave and Air Pollution Products 904.5.2 Evaluation at Pilot Sites 924.6 Future Applications 934.6.1 Impact- Based Warnings 934.6.2 Multi- Hazard Forecasting 954.6.3 Cold Spells as a Health Hazard 974.6.4 Social Sensing 974.6.5 Protecting the Vulnerable 984.7 Conclusions 98Funding 99Acknowledgements 99Notes 99References 995 The EUMETNET OPERA Radar Network – European- Wide Precipitation Composites Supporting Rainfall- Induced Flash Flood Emergency Management 105Shinju Park, Marc Berenguer, Daniel Sempere- Torres, and Annakaisa Von Lerber5.1 Introduction 1055.2 The EUMETNET OPERA Radar Precipitation Composites 1065.3 Monitoring the Quality of the Opera Precipitation Composites 1085.4 Application of Opera Precipitation Composites for Flash Flood Hazard Nowcasting 1105.5 Conclusions and Outlook 113References 1166 Towards Impact- Based Communication During Climate Emergencies: A Community- Based Approach to Improve Flood Early Warning Systems 119Erika Meléndez- Landaverde, Daniel Sempere- Torres, and Shinju Park6.1 Introduction 1196.2 Impact- Based Early Warning Systems (IB- EWS) for Actionable Decisions: Key Aspects 1216.2.1 Partnerships for an Effective Co- Design IB- EWS 1226.2.2 End Users: Identifying Needs for Emergency Response 1236.2.3 Risk Identification and Impact Data Collection 1246.2.4 Evaluation of IB- EWSs 1256.3 The Next Step for Community- Based EWS: The Site- Specific EWS Framework (SS- EWS) 1256.3.1 The Site- Specific Early Warning System Framework (SS- EWS) 1266.4 The SS- EWS in Catalonia, NE Spain: Experiences and Lessons Learned 1286.4.1 Community- Based Sessions in Terrassa: The Co- Design Process and Experiences 1296.4.2 Community- Based Emergency Response: SS- EWS Real- Time Application in Terrassa 1326.4.3 The Site- Specific Warnings (SSWs): Their Influence on the Risk Perception and Understanding of Users in Blanes 1326.4.4 A4alerts: Mobile Application for Emergency Communication 1346.5 An Outlook on Future Community and Impact- Based Communication Tools for Floods 135Notes 137References 1377 Challenges for a Better Use of Crowdsourcing Information in Climate Emergency Situational Awareness and Early Warning Systems 141Milan Kalas, Joy Ommer, Amin Shakya, Saša Vraníc, Denys Kolokol, and Tommaso Sabattini7.1 Introduction 1417.2 Crowd- Generated Content to Support Emergency Management and Early Warning 1437.2.1 Examples of the Citizen Science in Disaster Risk Management 1437.2.2 Tools 1447.2.3 Challenges in the Integration and Application of Citizen- Generated Content in DRM 1457.3 ANYWHERE Applications and Their Lessons Learnt 1467.3.1 Crowd Mapping to Support Real- Time Risk Assessment 1477.3.2 Social Media Streaming to Increase Emergency Situational Awareness 1477.3.3 A Crowdsourcing Solution for Collecting Information on the Magnitude and Impact of Disasters 1537.3.4 Towards a Holistic System 1557.3.5 Facilitating Communication Between Actors in Emergency Management 1577.4 Conclusion 158Note 159References 1598 Co- Evaluation: How to Measure Achievements in Complex Co- Production Projects? ANYWHERE's Contribution to Enhance Emergency Management of Weather and Climate Events 163Oliver Gebhardt and Christian Kuhlicke8.1 Introduction 1638.2 Application of the ANYWHERE Co- Evaluation Framework 1658.2.1 Step 1: Context Analysis 1668.2.2 Step 2: Description of Baseline Scenario and ANYWHERE Scenario 1668.2.3 Step 3: Selection of Suitable and Feasible Criteria 1668.2.4 Step 4: Selection of Appropriate Co- Evaluation Method 1678.2.5 Step 5: Data Collection 1678.2.6 Step 6: Data Aggregation and Analysis 1688.3 Discussion of Co- Evaluation Results 1688.4 Discussion 1768.5 Conclusion 177Notes 177References 1789 Using Artificial Intelligence to Manage Extreme Weather Events: The Impact of the beAWARE Solution 181Anastasios Karakostas, Stefanos Vrochidis, and Ioannis Kompatsiaris9.1 Introduction 1819.2 Overall Objectives of the Project 1829.3 The Impact of beAWARE 1889.3.1 Scientific and Innovation Impact 1889.3.2 Economic Impact 1919.3.3 Safety Impact 1919.3.4 Training Impact 1919.3.5 Policymakers 1939.3.6 First Responders 1949.3.7 General Public (Citizens) 1959.4 Conclusion 196Acknowledgement 197References 19710 Innovative Visual Analysis Solutions to Support Disaster Management 199Emmanouil Michail, Panagiotis Giannakeris, Ilias Koulalis, Stefanos Vrochidis, and Ioannis Kompatsiaris10.1 Introduction 19910.2 Related Work 20010.3 Methodology 20310.3.1 Disaster Detection 20410.3.2 Object Detection 20510.3.3 River Level Monitoring 20610.3.4 Drone Analysis 20610.3.5 Traffic Analysis and Management 20910.4 System Evaluation 21110.4.1 Disaster Detection 21210.4.2 Object Detection and Tracking 21310.4.3 River Level Monitoring 21510.4.4 Drone Analysis 21710.4.5 Traffic Analysis and Management 21910.5 Conclusions 221References 22111 Social Media Monitoring for Disaster Management 224Stelios Andreadis, Ilias Gialampoukidis, Stefanos Vrochidis, and Ioannis Kompatsiaris11.1 Introduction 22411.2 Social Media Analysis 22511.2.1 Framework Overview 22511.2.2 Data Collection from Twitter 22611.2.3 Analysis of Social Media Data 22711.2.4 Data Representation 23211.3 Social Media Clustering 23411.3.1 Evaluation of Spatial Clustering Techniques 23411.3.2 The Proposed Spatiotemporal Clustering 23611.4 Visualizations 23711.4.1 Annotation Tool 23711.4.2 Demonstration Tool 23911.5 Conclusion 240Notes 241References 24112 Human- Centred Public Warnings 243Claudio Rossi and Antonella Frisiello12.1 Introduction 24312.2 Risk Communication 24512.2.1 Risk Communication Key Aspects 24612.2.2 United Nation Guidelines 24912.3 Technical Standards and Recommendations 25012.3.1 Standards and Requirements for Public Warning Systems Implementation 25012.3.2 The Common Alerting Protocol 25112.3.3 Recommended System Architecture 25212.3.4 Use of Technical Standards 25712.3.5 Media Adaptation and Usability of Alerts 26012.4 Future Outlooks in Public Warning and Risk Communication 26712.4.1 Crowdsourcing Approaches 26712.4.2 Organizational Best Practices 269Note 271References 27213 A DRM Solution for Professionals and Citizens 275Claudio Rossi, Antonella Frisiello, Gianluca Marucco, and Marco Pini13.1 A Novel Mobile Application for DRR 27513.2 The I- REACT Co- Design Approach 27613.2.1 The Co- Design Process in the I- REACT Project 27713.2.2 From Data to Specifications: The Results of I- REACT Co- Design Activities 28013.3 The Development and Implementation of the I- REACT Mobile Solution 28513.4 Gamified Crowdsourcing for Disaster Risk Management 29013.5 The I- REACT Wearable Solution for First Responders 29313.5.1 Ad- hoc Positioning Wearable Device for Enhanced Localization 29413.5.2 Operational Scenario 29513.5.3 Device Operating Modes 29713.5.4 Communication Flow 29913.5.5 Wearable Device Implementation and Prototyping Cycles 29913.5.6 Wearable Device Performance Validation 30113.6 Improved Positioning of First Responders Using EGNSS Technologies 30213.6.1 A Service- Oriented Cloud- Based Architecture for Mobile Geolocated Emergency Services (EGNOS in the Cloud) 30413.6.2 EDAS Service Selector, Decoder and Storage 30613.6.3 Augmented PVT and Integrity Computation 30713.6.4 Implementation of the Architecture of the Cloud Software Module 30813.6.5 Performance Evaluation of the Implementation 30913.6.6 Positioning Integrity Computation for Consumer- Grade GNSS Receivers 313References 32314 Transforming Data Coming from Social Media Streams into Disaster- Related Information 326Claudio Rossi, Edoardo Arnaudo, Dario Salza, Giacomo Blanco, and Lorenzo Bongiovanni14.1 Introduction 32614.2 Natural Language Processing Methods for Emergency- Related Text Processing 33114.2.1 Document Representation 33214.2.2 Document Classification 33314.2.3 Named Entity Recognition 33414.3 Model Architecture 33514.4 Classification Results 33614.4.1 Bag of Words with SVM 33614.4.2 CNN with Multilingual Word Embeddings 33714.4.3 CNN with XML- T Contextual Word Embeddings 33814.5 Image Filtering and Classification for Contextual Awareness 33914.5.1 Filtering Unwanted Images 33914.5.2 Methodology for NSFW Classification 34014.5.3 Classifying Relevant Images 34114.5.4 Methodology for Image Classification 34314.6 Event Detection 34514.6.1 Related Work 34614.6.2 Methodology 34914.6.3 Evaluation of the Event Detection Pipeline 35114.7 Impact Extraction 35414.7.1 Related Work 35414.7.2 Methodology 35614.7.3 Aggregating the Information 35714.7.4 Evaluation Results 35814.8 Annex 1: Definition of Yara Rules for Impact Estimation 360Funding 362Notes 362References 36215 Conclusions and Perspectives 368Philippe Quevauviller15.1 Introduction 36815.2 Policy Background 36915.2.1 Civil Protection Policies 37015.2.2 EU Strategy on Adaptation to Climate Change 37215.2.3 Water Framework and Marine Policies 37315.2.4 Links with Projects Subject to this Book 37415.3 Actor's Interactions and Community Building 37515.3.1 Who are the Actors? 37515.3.2 Community Building 37715.4 Research Trends Related to Disaster Risks (Including Climate Extremes) in the Security Research Area 37915.4.1 Societal Resilience 37915.4.2 Tools for Integrated Risk Reduction for Extreme Climate Events 38115.5 Conclusions, Gaps and Recommendations 383Notes 384References 384Index 386

Daniel Sempere-Torres is Professor at Universitat Politècnica de Catalunya, Barcelona, Spain. Anastasios Karakostas is Director at DRAXIS Environmental S.A. and former Senior Researcher at the Centre for Research and Technology Hellas, Thessaloniki, Greece. Claudio Rossi is Program Manager and Senior Researcher at LINKS Foundation, Turin, Italy. Philippe Quevauviller is Former Professor at Vrije Universiteit Brussel, Brussels, Belgium and Research Programming and Policy Officer at the European Commission.

• List of Contributors xiiSeries Preface xvi1 The ANYWHERE Paradigm Shift in Responding to Weather and Climate Emergencies: Impact Forecasting, Dynamic Vulnerability and the Need for Citizen's Involvement 1Daniel Sempere- Torres and Marc Berenguer1.1 Disaster Risk Management in Times of Climate Change: The Need of a Proactive Approach 11.2 Adapting Risk Management to the 'New Normality': The Case of Flood Risk Management 21.3 Changing the Paradigm: Impact- Based Multi- Hazard Early Warning Systems to Move from Reactive to Pro- Active Emergency Response Strategies 41.3.1 From Reactive to Proactive Emergency Response Strategies 51.3.2 The ANYWHERE MH- IEWS 91.4 The New Paradigm: Dynamic Vulnerability 131.5 Future Work: From Multi- Hazards to Multi- Risk IEWS 16Notes 17References 182 Hydrometeorological Drought Forecasts: Lessons Learned from ANYWHERE and Next Steps to Improve Drought Management 23Samuel J. Sutanto and Henny A.J. Van Lanen2.1 Introduction 232.2 Method for Forecasting Hydrometeorological Droughts 252.2.1 The Climate (ECMWF SEAS5) and Hydrological (LISFLOOD) Models 252.2.2 The Drought Indices 262.2.3 The Drought Forecast Algorithms 282.3 Hydrometeorological Drought Forecasts 302.3.1 Meteorological Drought Forecasts 302.3.2 Hydrological Drought Forecasts 312.4 Drought Forecast Performance 332.4.1 The Origin of Seasonal Drought Forecast Skill 332.4.2 Examples of Assessment of Seasonal Drought Forecast Performance 342.5 Importance of Catchment Memory 382.6 Outlook and Future Improvements 402.6.1 Drought Impact Forecasts 412.6.2 Compound and Cascading (CC) Dry Hazards 43References 443 Experiences and Lessons Learnt in Wildfire Management with PROPAGATOR, an Operational Cellular- Automata- Based Wildfire Simulator 49Andrea Trucchia, Mirko D'Andrea, Francesco Baghino, Nicolò Perello, Nicola Rebora, and Paolo Fiorucci3.1 Introduction 493.1.1 Mathematical Models for Wildfire Management 503.2 Synopsis of Propagator Development: More than a Decade of Wildfire Simulations 523.3 Propagator Model 553.4 Case Studies 623.4.1 Data Retrieval 623.5 Results and Discussion 653.5.1 Performance Indicators 653.5.2 Performances of Test Cases 703.5.3 An Example of Continuous Improvement and Operational Deployment: Implementation in Ireland 713.6 Conclusions 71References 734 Building an Operational Decision Support System for Multiple Weather- Induced Health Hazards: ANYWHERE Developments and Future Applications 77Claudia Di Napoli4.1 Introduction 774.2 Heatwave Prediction in ANYWHERE 794.2.1 The Universal Thermal Climate Index 804.2.2 Forecasting Algorithms 804.2.3 Heatwave Products 814.2.4 Integration in the MH- EWS 814.2.5 Temperature Products 814.3 Air Pollution Prediction in ANYWHERE 834.3.1 Air Quality 834.3.2 Forecasting Algorithms 854.3.3 Air Quality Products 854.3.4 Integration in the MH- EWS 854.4 ANYWHERE MH- EWS in Action: The European 2017 Heatwave 864.5 Implementation at Pilot Sites 874.5.1 Integration of Local Heatwave and Air Pollution Products 904.5.2 Evaluation at Pilot Sites 924.6 Future Applications 934.6.1 Impact- Based Warnings 934.6.2 Multi- Hazard Forecasting 954.6.3 Cold Spells as a Health Hazard 974.6.4 Social Sensing 974.6.5 Protecting the Vulnerable 984.7 Conclusions 98Funding 99Acknowledgements 99Notes 99References 995 The EUMETNET OPERA Radar Network – European- Wide Precipitation Composites Supporting Rainfall- Induced Flash Flood Emergency Management 105Shinju Park, Marc Berenguer, Daniel Sempere- Torres, and Annakaisa Von Lerber5.1 Introduction 1055.2 The EUMETNET OPERA Radar Precipitation Composites 1065.3 Monitoring the Quality of the Opera Precipitation Composites 1085.4 Application of Opera Precipitation Composites for Flash Flood Hazard Nowcasting 1105.5 Conclusions and Outlook 113References 1166 Towards Impact- Based Communication During Climate Emergencies: A Community- Based Approach to Improve Flood Early Warning Systems 119Erika Meléndez- Landaverde, Daniel Sempere- Torres, and Shinju Park6.1 Introduction 1196.2 Impact- Based Early Warning Systems (IB- EWS) for Actionable Decisions: Key Aspects 1216.2.1 Partnerships for an Effective Co- Design IB- EWS 1226.2.2 End Users: Identifying Needs for Emergency Response 1236.2.3 Risk Identification and Impact Data Collection 1246.2.4 Evaluation of IB- EWSs 1256.3 The Next Step for Community- Based EWS: The Site- Specific EWS Framework (SS- EWS) 1256.3.1 The Site- Specific Early Warning System Framework (SS- EWS) 1266.4 The SS- EWS in Catalonia, NE Spain: Experiences and Lessons Learned 1286.4.1 Community- Based Sessions in Terrassa: The Co- Design Process and Experiences 1296.4.2 Community- Based Emergency Response: SS- EWS Real- Time Application in Terrassa 1326.4.3 The Site- Specific Warnings (SSWs): Their Influence on the Risk Perception and Understanding of Users in Blanes 1326.4.4 A4alerts: Mobile Application for Emergency Communication 1346.5 An Outlook on Future Community and Impact- Based Communication Tools for Floods 135Notes 137References 1377 Challenges for a Better Use of Crowdsourcing Information in Climate Emergency Situational Awareness and Early Warning Systems 141Milan Kalas, Joy Ommer, Amin Shakya, Saša Vraníc, Denys Kolokol, and Tommaso Sabattini7.1 Introduction 1417.2 Crowd- Generated Content to Support Emergency Management and Early Warning 1437.2.1 Examples of the Citizen Science in Disaster Risk Management 1437.2.2 Tools 1447.2.3 Challenges in the Integration and Application of Citizen- Generated Content in DRM 1457.3 ANYWHERE Applications and Their Lessons Learnt 1467.3.1 Crowd Mapping to Support Real- Time Risk Assessment 1477.3.2 Social Media Streaming to Increase Emergency Situational Awareness 1477.3.3 A Crowdsourcing Solution for Collecting Information on the Magnitude and Impact of Disasters 1537.3.4 Towards a Holistic System 1557.3.5 Facilitating Communication Between Actors in Emergency Management 1577.4 Conclusion 158Note 159References 1598 Co- Evaluation: How to Measure Achievements in Complex Co- Production Projects? ANYWHERE's Contribution to Enhance Emergency Management of Weather and Climate Events 163Oliver Gebhardt and Christian Kuhlicke8.1 Introduction 1638.2 Application of the ANYWHERE Co- Evaluation Framework 1658.2.1 Step 1: Context Analysis 1668.2.2 Step 2: Description of Baseline Scenario and ANYWHERE Scenario 1668.2.3 Step 3: Selection of Suitable and Feasible Criteria 1668.2.4 Step 4: Selection of Appropriate Co- Evaluation Method 1678.2.5 Step 5: Data Collection 1678.2.6 Step 6: Data Aggregation and Analysis 1688.3 Discussion of Co- Evaluation Results 1688.4 Discussion 1768.5 Conclusion 177Notes 177References 1789 Using Artificial Intelligence to Manage Extreme Weather Events: The Impact of the beAWARE Solution 181Anastasios Karakostas, Stefanos Vrochidis, and Ioannis Kompatsiaris9.1 Introduction 1819.2 Overall Objectives of the Project 1829.3 The Impact of beAWARE 1889.3.1 Scientific and Innovation Impact 1889.3.2 Economic Impact 1919.3.3 Safety Impact 1919.3.4 Training Impact 1919.3.5 Policymakers 1939.3.6 First Responders 1949.3.7 General Public (Citizens) 1959.4 Conclusion 196Acknowledgement 197References 19710 Innovative Visual Analysis Solutions to Support Disaster Management 199Emmanouil Michail, Panagiotis Giannakeris, Ilias Koulalis, Stefanos Vrochidis, and Ioannis Kompatsiaris10.1 Introduction 19910.2 Related Work 20010.3 Methodology 20310.3.1 Disaster Detection 20410.3.2 Object Detection 20510.3.3 River Level Monitoring 20610.3.4 Drone Analysis 20610.3.5 Traffic Analysis and Management 20910.4 System Evaluation 21110.4.1 Disaster Detection 21210.4.2 Object Detection and Tracking 21310.4.3 River Level Monitoring 21510.4.4 Drone Analysis 21710.4.5 Traffic Analysis and Management 21910.5 Conclusions 221References 22111 Social Media Monitoring for Disaster Management 224Stelios Andreadis, Ilias Gialampoukidis, Stefanos Vrochidis, and Ioannis Kompatsiaris11.1 Introduction 22411.2 Social Media Analysis 22511.2.1 Framework Overview 22511.2.2 Data Collection from Twitter 22611.2.3 Analysis of Social Media Data 22711.2.4 Data Representation 23211.3 Social Media Clustering 23411.3.1 Evaluation of Spatial Clustering Techniques 23411.3.2 The Proposed Spatiotemporal Clustering 23611.4 Visualizations 23711.4.1 Annotation Tool 23711.4.2 Demonstration Tool 23911.5 Conclusion 240Notes 241References 24112 Human- Centred Public Warnings 243Claudio Rossi and Antonella Frisiello12.1 Introduction 24312.2 Risk Communication 24512.2.1 Risk Communication Key Aspects 24612.2.2 United Nation Guidelines 24912.3 Technical Standards and Recommendations 25012.3.1 Standards and Requirements for Public Warning Systems Implementation 25012.3.2 The Common Alerting Protocol 25112.3.3 Recommended System Architecture 25212.3.4 Use of Technical Standards 25712.3.5 Media Adaptation and Usability of Alerts 26012.4 Future Outlooks in Public Warning and Risk Communication 26712.4.1 Crowdsourcing Approaches 26712.4.2 Organizational Best Practices 269Note 271References 27213 A DRM Solution for Professionals and Citizens 275Claudio Rossi, Antonella Frisiello, Gianluca Marucco, and Marco Pini13.1 A Novel Mobile Application for DRR 27513.2 The I- REACT Co- Design Approach 27613.2.1 The Co- Design Process in the I- REACT Project 27713.2.2 From Data to Specifications: The Results of I- REACT Co- Design Activities 28013.3 The Development and Implementation of the I- REACT Mobile Solution 28513.4 Gamified Crowdsourcing for Disaster Risk Management 29013.5 The I- REACT Wearable Solution for First Responders 29313.5.1 Ad- hoc Positioning Wearable Device for Enhanced Localization 29413.5.2 Operational Scenario 29513.5.3 Device Operating Modes 29713.5.4 Communication Flow 29913.5.5 Wearable Device Implementation and Prototyping Cycles 29913.5.6 Wearable Device Performance Validation 30113.6 Improved Positioning of First Responders Using EGNSS Technologies 30213.6.1 A Service- Oriented Cloud- Based Architecture for Mobile Geolocated Emergency Services (EGNOS in the Cloud) 30413.6.2 EDAS Service Selector, Decoder and Storage 30613.6.3 Augmented PVT and Integrity Computation 30713.6.4 Implementation of the Architecture of the Cloud Software Module 30813.6.5 Performance Evaluation of the Implementation 30913.6.6 Positioning Integrity Computation for Consumer- Grade GNSS Receivers 313References 32314 Transforming Data Coming from Social Media Streams into Disaster- Related Information 326Claudio Rossi, Edoardo Arnaudo, Dario Salza, Giacomo Blanco, and Lorenzo Bongiovanni14.1 Introduction 32614.2 Natural Language Processing Methods for Emergency- Related Text Processing 33114.2.1 Document Representation 33214.2.2 Document Classification 33314.2.3 Named Entity Recognition 33414.3 Model Architecture 33514.4 Classification Results 33614.4.1 Bag of Words with SVM 33614.4.2 CNN with Multilingual Word Embeddings 33714.4.3 CNN with XML- T Contextual Word Embeddings 33814.5 Image Filtering and Classification for Contextual Awareness 33914.5.1 Filtering Unwanted Images 33914.5.2 Methodology for NSFW Classification 34014.5.3 Classifying Relevant Images 34114.5.4 Methodology for Image Classification 34314.6 Event Detection 34514.6.1 Related Work 34614.6.2 Methodology 34914.6.3 Evaluation of the Event Detection Pipeline 35114.7 Impact Extraction 35414.7.1 Related Work 35414.7.2 Methodology 35614.7.3 Aggregating the Information 35714.7.4 Evaluation Results 35814.8 Annex 1: Definition of Yara Rules for Impact Estimation 360Funding 362Notes 362References 36215 Conclusions and Perspectives 368Philippe Quevauviller15.1 Introduction 36815.2 Policy Background 36915.2.1 Civil Protection Policies 37015.2.2 EU Strategy on Adaptation to Climate Change 37215.2.3 Water Framework and Marine Policies 37315.2.4 Links with Projects Subject to this Book 37415.3 Actor's Interactions and Community Building 37515.3.1 Who are the Actors? 37515.3.2 Community Building 37715.4 Research Trends Related to Disaster Risks (Including Climate Extremes) in the Security Research Area 37915.4.1 Societal Resilience 37915.4.2 Tools for Integrated Risk Reduction for Extreme Climate Events 38115.5 Conclusions, Gaps and Recommendations 383Notes 384References 384Index 386

Daniel Sempere-Torres is Professor at Universitat Politècnica de Catalunya, Barcelona, Spain. Anastasios Karakostas is Director at DRAXIS Environmental S.A. and former Senior Researcher at the Centre for Research and Technology Hellas, Thessaloniki, Greece. Claudio Rossi is Program Manager and Senior Researcher at LINKS Foundation, Turin, Italy. Philippe Quevauviller is Former Professor at Vrije Universiteit Brussel, Brussels, Belgium and Research Programming and Policy Officer at the European Commission.

• List of Contributors xiiSeries Preface xvi1 The ANYWHERE Paradigm Shift in Responding to Weather and Climate Emergencies: Impact Forecasting, Dynamic Vulnerability and the Need for Citizen's Involvement 1Daniel Sempere- Torres and Marc Berenguer1.1 Disaster Risk Management in Times of Climate Change: The Need of a Proactive Approach 11.2 Adapting Risk Management to the 'New Normality': The Case of Flood Risk Management 21.3 Changing the Paradigm: Impact- Based Multi- Hazard Early Warning Systems to Move from Reactive to Pro- Active Emergency Response Strategies 41.3.1 From Reactive to Proactive Emergency Response Strategies 51.3.2 The ANYWHERE MH- IEWS 91.4 The New Paradigm: Dynamic Vulnerability 131.5 Future Work: From Multi- Hazards to Multi- Risk IEWS 16Notes 17References 182 Hydrometeorological Drought Forecasts: Lessons Learned from ANYWHERE and Next Steps to Improve Drought Management 23Samuel J. Sutanto and Henny A.J. Van Lanen2.1 Introduction 232.2 Method for Forecasting Hydrometeorological Droughts 252.2.1 The Climate (ECMWF SEAS5) and Hydrological (LISFLOOD) Models 252.2.2 The Drought Indices 262.2.3 The Drought Forecast Algorithms 282.3 Hydrometeorological Drought Forecasts 302.3.1 Meteorological Drought Forecasts 302.3.2 Hydrological Drought Forecasts 312.4 Drought Forecast Performance 332.4.1 The Origin of Seasonal Drought Forecast Skill 332.4.2 Examples of Assessment of Seasonal Drought Forecast Performance 342.5 Importance of Catchment Memory 382.6 Outlook and Future Improvements 402.6.1 Drought Impact Forecasts 412.6.2 Compound and Cascading (CC) Dry Hazards 43References 443 Experiences and Lessons Learnt in Wildfire Management with PROPAGATOR, an Operational Cellular- Automata- Based Wildfire Simulator 49Andrea Trucchia, Mirko D'Andrea, Francesco Baghino, Nicolò Perello, Nicola Rebora, and Paolo Fiorucci3.1 Introduction 493.1.1 Mathematical Models for Wildfire Management 503.2 Synopsis of Propagator Development: More than a Decade of Wildfire Simulations 523.3 Propagator Model 553.4 Case Studies 623.4.1 Data Retrieval 623.5 Results and Discussion 653.5.1 Performance Indicators 653.5.2 Performances of Test Cases 703.5.3 An Example of Continuous Improvement and Operational Deployment: Implementation in Ireland 713.6 Conclusions 71References 734 Building an Operational Decision Support System for Multiple Weather- Induced Health Hazards: ANYWHERE Developments and Future Applications 77Claudia Di Napoli4.1 Introduction 774.2 Heatwave Prediction in ANYWHERE 794.2.1 The Universal Thermal Climate Index 804.2.2 Forecasting Algorithms 804.2.3 Heatwave Products 814.2.4 Integration in the MH- EWS 814.2.5 Temperature Products 814.3 Air Pollution Prediction in ANYWHERE 834.3.1 Air Quality 834.3.2 Forecasting Algorithms 854.3.3 Air Quality Products 854.3.4 Integration in the MH- EWS 854.4 ANYWHERE MH- EWS in Action: The European 2017 Heatwave 864.5 Implementation at Pilot Sites 874.5.1 Integration of Local Heatwave and Air Pollution Products 904.5.2 Evaluation at Pilot Sites 924.6 Future Applications 934.6.1 Impact- Based Warnings 934.6.2 Multi- Hazard Forecasting 954.6.3 Cold Spells as a Health Hazard 974.6.4 Social Sensing 974.6.5 Protecting the Vulnerable 984.7 Conclusions 98Funding 99Acknowledgements 99Notes 99References 995 The EUMETNET OPERA Radar Network – European- Wide Precipitation Composites Supporting Rainfall- Induced Flash Flood Emergency Management 105Shinju Park, Marc Berenguer, Daniel Sempere- Torres, and Annakaisa Von Lerber5.1 Introduction 1055.2 The EUMETNET OPERA Radar Precipitation Composites 1065.3 Monitoring the Quality of the Opera Precipitation Composites 1085.4 Application of Opera Precipitation Composites for Flash Flood Hazard Nowcasting 1105.5 Conclusions and Outlook 113References 1166 Towards Impact- Based Communication During Climate Emergencies: A Community- Based Approach to Improve Flood Early Warning Systems 119Erika Meléndez- Landaverde, Daniel Sempere- Torres, and Shinju Park6.1 Introduction 1196.2 Impact- Based Early Warning Systems (IB- EWS) for Actionable Decisions: Key Aspects 1216.2.1 Partnerships for an Effective Co- Design IB- EWS 1226.2.2 End Users: Identifying Needs for Emergency Response 1236.2.3 Risk Identification and Impact Data Collection 1246.2.4 Evaluation of IB- EWSs 1256.3 The Next Step for Community- Based EWS: The Site- Specific EWS Framework (SS- EWS) 1256.3.1 The Site- Specific Early Warning System Framework (SS- EWS) 1266.4 The SS- EWS in Catalonia, NE Spain: Experiences and Lessons Learned 1286.4.1 Community- Based Sessions in Terrassa: The Co- Design Process and Experiences 1296.4.2 Community- Based Emergency Response: SS- EWS Real- Time Application in Terrassa 1326.4.3 The Site- Specific Warnings (SSWs): Their Influence on the Risk Perception and Understanding of Users in Blanes 1326.4.4 A4alerts: Mobile Application for Emergency Communication 1346.5 An Outlook on Future Community and Impact- Based Communication Tools for Floods 135Notes 137References 1377 Challenges for a Better Use of Crowdsourcing Information in Climate Emergency Situational Awareness and Early Warning Systems 141Milan Kalas, Joy Ommer, Amin Shakya, Saša Vraníc, Denys Kolokol, and Tommaso Sabattini7.1 Introduction 1417.2 Crowd- Generated Content to Support Emergency Management and Early Warning 1437.2.1 Examples of the Citizen Science in Disaster Risk Management 1437.2.2 Tools 1447.2.3 Challenges in the Integration and Application of Citizen- Generated Content in DRM 1457.3 ANYWHERE Applications and Their Lessons Learnt 1467.3.1 Crowd Mapping to Support Real- Time Risk Assessment 1477.3.2 Social Media Streaming to Increase Emergency Situational Awareness 1477.3.3 A Crowdsourcing Solution for Collecting Information on the Magnitude and Impact of Disasters 1537.3.4 Towards a Holistic System 1557.3.5 Facilitating Communication Between Actors in Emergency Management 1577.4 Conclusion 158Note 159References 1598 Co- Evaluation: How to Measure Achievements in Complex Co- Production Projects? ANYWHERE's Contribution to Enhance Emergency Management of Weather and Climate Events 163Oliver Gebhardt and Christian Kuhlicke8.1 Introduction 1638.2 Application of the ANYWHERE Co- Evaluation Framework 1658.2.1 Step 1: Context Analysis 1668.2.2 Step 2: Description of Baseline Scenario and ANYWHERE Scenario 1668.2.3 Step 3: Selection of Suitable and Feasible Criteria 1668.2.4 Step 4: Selection of Appropriate Co- Evaluation Method 1678.2.5 Step 5: Data Collection 1678.2.6 Step 6: Data Aggregation and Analysis 1688.3 Discussion of Co- Evaluation Results 1688.4 Discussion 1768.5 Conclusion 177Notes 177References 1789 Using Artificial Intelligence to Manage Extreme Weather Events: The Impact of the beAWARE Solution 181Anastasios Karakostas, Stefanos Vrochidis, and Ioannis Kompatsiaris9.1 Introduction 1819.2 Overall Objectives of the Project 1829.3 The Impact of beAWARE 1889.3.1 Scientific and Innovation Impact 1889.3.2 Economic Impact 1919.3.3 Safety Impact 1919.3.4 Training Impact 1919.3.5 Policymakers 1939.3.6 First Responders 1949.3.7 General Public (Citizens) 1959.4 Conclusion 196Acknowledgement 197References 19710 Innovative Visual Analysis Solutions to Support Disaster Management 199Emmanouil Michail, Panagiotis Giannakeris, Ilias Koulalis, Stefanos Vrochidis, and Ioannis Kompatsiaris10.1 Introduction 19910.2 Related Work 20010.3 Methodology 20310.3.1 Disaster Detection 20410.3.2 Object Detection 20510.3.3 River Level Monitoring 20610.3.4 Drone Analysis 20610.3.5 Traffic Analysis and Management 20910.4 System Evaluation 21110.4.1 Disaster Detection 21210.4.2 Object Detection and Tracking 21310.4.3 River Level Monitoring 21510.4.4 Drone Analysis 21710.4.5 Traffic Analysis and Management 21910.5 Conclusions 221References 22111 Social Media Monitoring for Disaster Management 224Stelios Andreadis, Ilias Gialampoukidis, Stefanos Vrochidis, and Ioannis Kompatsiaris11.1 Introduction 22411.2 Social Media Analysis 22511.2.1 Framework Overview 22511.2.2 Data Collection from Twitter 22611.2.3 Analysis of Social Media Data 22711.2.4 Data Representation 23211.3 Social Media Clustering 23411.3.1 Evaluation of Spatial Clustering Techniques 23411.3.2 The Proposed Spatiotemporal Clustering 23611.4 Visualizations 23711.4.1 Annotation Tool 23711.4.2 Demonstration Tool 23911.5 Conclusion 240Notes 241References 24112 Human- Centred Public Warnings 243Claudio Rossi and Antonella Frisiello12.1 Introduction 24312.2 Risk Communication 24512.2.1 Risk Communication Key Aspects 24612.2.2 United Nation Guidelines 24912.3 Technical Standards and Recommendations 25012.3.1 Standards and Requirements for Public Warning Systems Implementation 25012.3.2 The Common Alerting Protocol 25112.3.3 Recommended System Architecture 25212.3.4 Use of Technical Standards 25712.3.5 Media Adaptation and Usability of Alerts 26012.4 Future Outlooks in Public Warning and Risk Communication 26712.4.1 Crowdsourcing Approaches 26712.4.2 Organizational Best Practices 269Note 271References 27213 A DRM Solution for Professionals and Citizens 275Claudio Rossi, Antonella Frisiello, Gianluca Marucco, and Marco Pini13.1 A Novel Mobile Application for DRR 27513.2 The I- REACT Co- Design Approach 27613.2.1 The Co- Design Process in the I- REACT Project 27713.2.2 From Data to Specifications: The Results of I- REACT Co- Design Activities 28013.3 The Development and Implementation of the I- REACT Mobile Solution 28513.4 Gamified Crowdsourcing for Disaster Risk Management 29013.5 The I- REACT Wearable Solution for First Responders 29313.5.1 Ad- hoc Positioning Wearable Device for Enhanced Localization 29413.5.2 Operational Scenario 29513.5.3 Device Operating Modes 29713.5.4 Communication Flow 29913.5.5 Wearable Device Implementation and Prototyping Cycles 29913.5.6 Wearable Device Performance Validation 30113.6 Improved Positioning of First Responders Using EGNSS Technologies 30213.6.1 A Service- Oriented Cloud- Based Architecture for Mobile Geolocated Emergency Services (EGNOS in the Cloud) 30413.6.2 EDAS Service Selector, Decoder and Storage 30613.6.3 Augmented PVT and Integrity Computation 30713.6.4 Implementation of the Architecture of the Cloud Software Module 30813.6.5 Performance Evaluation of the Implementation 30913.6.6 Positioning Integrity Computation for Consumer- Grade GNSS Receivers 313References 32314 Transforming Data Coming from Social Media Streams into Disaster- Related Information 326Claudio Rossi, Edoardo Arnaudo, Dario Salza, Giacomo Blanco, and Lorenzo Bongiovanni14.1 Introduction 32614.2 Natural Language Processing Methods for Emergency- Related Text Processing 33114.2.1 Document Representation 33214.2.2 Document Classification 33314.2.3 Named Entity Recognition 33414.3 Model Architecture 33514.4 Classification Results 33614.4.1 Bag of Words with SVM 33614.4.2 CNN with Multilingual Word Embeddings 33714.4.3 CNN with XML- T Contextual Word Embeddings 33814.5 Image Filtering and Classification for Contextual Awareness 33914.5.1 Filtering Unwanted Images 33914.5.2 Methodology for NSFW Classification 34014.5.3 Classifying Relevant Images 34114.5.4 Methodology for Image Classification 34314.6 Event Detection 34514.6.1 Related Work 34614.6.2 Methodology 34914.6.3 Evaluation of the Event Detection Pipeline 35114.7 Impact Extraction 35414.7.1 Related Work 35414.7.2 Methodology 35614.7.3 Aggregating the Information 35714.7.4 Evaluation Results 35814.8 Annex 1: Definition of Yara Rules for Impact Estimation 360Funding 362Notes 362References 36215 Conclusions and Perspectives 368Philippe Quevauviller15.1 Introduction 36815.2 Policy Background 36915.2.1 Civil Protection Policies 37015.2.2 EU Strategy on Adaptation to Climate Change 37215.2.3 Water Framework and Marine Policies 37315.2.4 Links with Projects Subject to this Book 37415.3 Actor's Interactions and Community Building 37515.3.1 Who are the Actors? 37515.3.2 Community Building 37715.4 Research Trends Related to Disaster Risks (Including Climate Extremes) in the Security Research Area 37915.4.1 Societal Resilience 37915.4.2 Tools for Integrated Risk Reduction for Extreme Climate Events 38115.5 Conclusions, Gaps and Recommendations 383Notes 384References 384Index 386