Winter Storms in Germany

RESEARCH REPORT 2015 - 2016

KIT – The Research University in the Helmholtz Association www.kit.edu Content 3

Content

Imprint...... 5

Preface...... 6

Vorwort...... 8

I. FDA Activities...... 10

FDA Reports...... 10

CEDIM Investigations on the 2015 Nepal Earthquake...... 10

Severe Thunderstorm Episode May/June 2016 Germany...... 13

Hurricane Matthew, September 2016...... 15

Short Reports...... 17

Earthquake Amatrice 2016...... 17

Excessive Rain in Southern Germany, 19. - 21. November 2015...... 18

Lefkada Earthquake Greece 2015...... 20

Illapel Earthquake Chile 2015...... 21

II. Research...... 22

CEDIM Projects...... 22

Crowdsourcing - Using Social Media for Rapid Damage Assessment...... 22

Improving Communication with the Public during Disaster Situations...... 23

Towards the Development of a Global Tsunami Risk Model...... 25

The Web Service “Wettergefahren-Frühwarnung”/ Weather Hazards - Early Warning...... 27

ATMO Forensic Prediction and Analysis - Winter Storms in Germany...... 29

New CEDIM Projects...... 31

Resilience of Cities in the Course of Time...... 31

Critical Social Interactions in Case of Emergency - Vulnerabilitiy of the Chilean Street Network to Social Interaction in Case of Emergency of Natural Disasters...... 32

Effects of Extreme Events on EMI-Systems...... 33

Energy Risks toward 2025...... 34

Earthquake Aftershock Modeling...... 36

Loss Estimation and Analysis...... 37 4 Content

III. Cooperations and Outreach...... 39

Research Institutions...... 39

University of Adelaide...... 39

Institute for Environmental Studies, VU University Amsterdam...... 39

Deltares...... 39

Insurance Industry...... 39

Willis Research Network (WRN)...... 39

Sparkassenversicherung SV...... 40

International Programs...... 41

IRDR International Center of Excellence...... 41

WMO’s High Impact Weather Project HIWeather...... 41

CEDIM and the World Bank & GFDRR (Global Facility for Disaster Reduction and Recovery)...... 41

GADRI...... 42

Spin-Off Company Risklayer...... 42

Outreach...... 42

European Geosciences Union (EGU)...... 42

Earth System Knowledge Platform ESKP...... 43

Media Presence (Television, Radio, Print)...... 43

FDA and Short Reports...... 44

CEDIM at the “The Change Agenda” Symposium in Oxford, England...... 44

Networking Initiative at KIT...... 44

Strategic Workshops on „Risks, Catastrophes and Security”...... 44

IV. Publications 2015 and 2016...... 46

Articles in Journals and Books 2015...... 46

Articles in Journals and Books 2016...... 47

Conference Abstracts 2015...... 48

Conference Abstracts 2016...... 49

CEDIM Reports 2015...... 50

CEDIM Reports 2016...... 50 Content 5

Imprint

CEDIM Research Report 2015 - 2016

Editors: Prof. Dr. Michael Kunz Dr. André Dittrich tech. Assistant: Johannes Hilpert

DOI: 10.5445/IR/1000069933 State: June 2017

CEDIM Cover picture: Devastated street in Braunsbach, Germany after flood events end of Center for Disaster Management and May and beginning of June 2016. Risk Reduction Technology Image by Dr. Andreas Kron Hermann-von-Helmholtz-Platz 1 Captured 7 June 2016 76344 Eggenstein-Leopoldshafen, Germany [email protected] Printed by: Systemedia GmbH, Wurmberg 6 Preface

Preface

The Center for Disaster Management and Risk and flow velocity to draw conclusions about the Reduction Technology (CEDIM) is an interdiscipli- temporal sequence of the event by combining nary research institution in the field of disaster direct observations, statements of individuals management. After 10 years of successful coop- affected, and video analyses. After Hurricane eration with the Deutsches GeoForschungsZen- Matthew on 16 September 2016, which caused trum Potsdam (GFZ), the Karlsruhe Institute of more than 1000 fatalities in Haiti alone, its me- Technology (KIT) became solely responsible for teorological background and socioeconomic ef- CEDIM in January 2016. At present, 16 KIT insti- fects were analysed. tutes participate in CEDIM. This report provides an overview of the research work and activities In addition to these broad forensic analyses, CE- of CEDIM during 2015 and 2016. DIM researchers have examined further extreme events and disasters primarily disciplinarily, which For several years, CEDIM’s research has focused are also described briefly in Chapter I. This in- on Forensic Disaster Analyses (FDA)1 in near-real- cludes the earthquake in Amatrice (Italy) in Au- time. The objectives of this new research are to gust 2016, with an estimated damage of 2.6 assess a disaster directly after its occurrence, an- billion USD, the heavy rain episode of Novem- alyse its effects, track its temporal development, ber 2016 in Germany, the weaker earthquake and identify the factors most relevant to its impli- on Lefkada (Greece) in November 2015 and in cations. In the course of a CEDIM FDA-Task Force the same month, the heavy Mw=8.3 earthquake Activity, reports with different foci are written in Illapel (Chile), where the new Tsunami-model contemporaneously, i.e., a few days to weeks TsuKIT was applied for the first time. In addition after the disaster occurs. In some cases, specific to these CEDIM short reports, 131 reports about field studies on site complement these analyses. forthcoming extreme or unusual weather events worldwide were published on the web portal, Various CEDIM FDA activities have been per- “Wettergefahren-Frühwarnung.” formed in the last two years, and are described in detail in Chapter I. A team of researchers ana- Different CEDIM projects constitute the scien- lysed the effects on various regions of the severe tific basis of the FDA activities, where methods earthquake in Nepal on 25 August 2015, as well have been developed or improved during the as the intensity and frequency of the aftershocks. last years to help analyse disasters in near-real- During three field studies (June and November time, and which are applied in the case of an 2015; April 2016), different surveys were con- FDA activity (Chapter II). Within the scope of a ducted on site to shed light on the situation in project on crowdsourcing, natural disasters are the emergency shelters, establish criteria for an detected, located, and classified in real-time via evacuation, and evaluate existing information a self-developed software and taxonomy using gaps for the public. On this basis, a conceptual social media posts. Based on past FDA activities, model was constructed to assess the vulnerability a further CEDIM project investigated factors that of critical infrastructures and the emergency shel- hamper effective communication during disasters, and for emergency planning purposes. With ters and lead to significant information gaps, respect to the thunderstorm episode in Germany and identified the primary reasons for these during May/June 2016, which lasted almost two situations. Two further projects are designed to weeks and caused heavy flooding in many local- estimate the risk of winter storms and tsunamis ities (e.g., Braunsbach and Simbach), CEDIM’s using probabilistic methods. While in the case of research focused on analyses of the probability winter storms, a simplified approach is used that of comparable events in the long-term. Through estimates damage by means of the intensity and surveys in several municipalities affected, CEDIM spatial expansion of an event, combined with researchers estimated runoff, flood wave height, data about residential buildings, the estimation of the risk of a tsunami is based on the com- putation of wave propagation using up-to-date 1 The term “Forensic Investigation of Disasters” (FORIN) has already been coined in 2010 in connection with natural dis- parallel processing hardware. asters by the international research programme IRDR/ICSU (Integrated Research on Disaster Risk / International Council Forensic Disaster Analyses also are a focus of for Science). IRDR, with its forensic approach, follows up on the current funding phase 2016-2018. There- the question of how natural hazards turn into disasters – or not. fore, the research approach has been developed Preface 7 further by integrating social science studies and tional and international research landscape and sharpening the profile to include society’s areas is visible widely to various user groups, such as of demand of energy, mobility, and information. the insurance industry, relief organizations, and The objectives of the new CEDIM projects, which governmental institutions. With its new focus, began at the end of 2016 and also are introduced CEDIM contributes to all three of society’s areas briefly in Chapter II, are to scrutinize changes in of demand—energy, mobility, and information— risk and resilience attributable to social change, identified by the KIT umbrella strategy 2025, and especially concerning energy, mobility, and sup- is thereby in an excellent position to continue to ply systems, and critical infrastructures in urban pool risk research at KIT. For this reason, CEDIM, areas. Work on the effects of natural hazards and together with the KIT-Center Climate and Envi- social interactions on the street network of Chile, ronment, has begun a process to improve the in- post-disaster rapid loss estimation, and the mod- tegration and networking of the various research elling of aftershocks complement these projects. activities at KIT in the fields of Risks, Catastro- phes, and Security, and to develop a common Several new collaborations were established in strategy. the last two years with international research centers, such as the University of Adelaide, the Michael Kunz VU university of Amsterdam, and Deltares, an Stefan Hinz independent institute that focuses on water re- Franz Nestmann search. Close connections continue with the insurance industry, especially with the Willis Re- search Network (WRN), where CEDIM is involved in a Flagship Project to develop a hail risk model for Europe and Australia. Further, CEDIM has developed risk models for both types of hazards, hail and flood, for the Sparkassenversicherung SV. CEDIM also has joined various national and international programs: the IRDR International Center of Excellence on Critical Infrastructures and Strategic Planning established in 2015, the new 10-years-program High Impact Weather of the World Meteorological Organization (WMO), and different programs of the World Bank and the Global Facility for Disaster Reduction and Recovery (GFDRR). Since 2016, CEDIM has been a member of the newly founded network, the Global Alliance of Disaster Research Institute (GADRI).

Outreach is another important pillar of CEDIM. This is being promoted actively, for example, in its own session at the European Geoscience Union (EGU), by several articles on the website Earth System Knowledge Platform (ESKP: www. eskp.de), as well as by numerous contributions in high-ranking media (e.g., New York Times, Die Zeit, FAZ, ARD, ZDF, ARTE, Deutschlandradio, and others). James Daniell, a CEDIM researcher, contributed an article, “Natural Disasters since 1900: Over 8 Million Deaths and 7 Trillion US Dollars damage,” that garnered the highest number of quotations for the KIT in 2016. A summary of these activities is provided in Chapter III.

With its near-real-time analyses of major disasters, a research program that addresses new topics repeatedly, and its active public relations, CEDIM is now established successfully in the na- 8 Vorwort

Vorwort

Das Center for Disaster Management and Risk zung der Resilienz kritischer Infrastrukturen, für Reduction Technology (CEDIM) ist eine interdiszi- die Notfallvorsorge und die Vulnerabilität inner- plinäre Forschungseinrichtung im Bereich des Ka- halb der Notunterkünfte. Bezüglich der fast zwei tastrophenmanagements. Nach über 10 Jahren Wochen andauernden Gewitterepisode im Mai/ erfolgreicher Kooperation mit dem Deutschen Juni 2016 in Deutschland mit schweren Über- GeoForschungsZentrum Potsdam (GFZ) ist CE- schwemmungen vieler Ortschaften (beispielswei- DIM im Januar 2016 in die alleinige Verantwor- se Braunsbach und Simbach) konzentrierten sich tung des Karlsruher Instituts für Technologie (KIT) die Forschungsarbeiten von CEDIM vor allem auf übergegangen. Derzeit sind daran 16 Institute Analysen der Wahrscheinlichkeit vergleichbarer des KIT beteiligt. Der vorliegende Bericht gibt ei- Ereignisse in einem langzeitlichen Kontext. Bei Er- nen Überblick über die Forschungsarbeiten und hebungen vor Ort in mehreren betroffenen Kom- Aktivitäten von CEDIM in den Jahren 2015 und munen konnten Wissenschaftler des CEDIM aus 2016. der Kombination von direkten Beobachtungen, Aussagen von Betroffenen und Videoanalysen Seit einigen Jahren liegt der Schwerpunkt der Abfluss, Scheitelhöhe und Strömungsgeschwin- Forschungsarbeiten von CEDIM auf forensischen digkeit schätzen und Rückschlüsse auf die Er- Katastrophenanalysen in Nahe-Echtzeit (Foren- eignisgenese ziehen. Nach Hurrikan Matthew sic Disaster Analysis, FDA)1. Die Ziele dieses For- im September 2016 mit über 1000 Todesopfern schungsansatzes sind es, unmittelbar nach dem allein in Haiti wurden in einer weiteren FDA Akti- Eintreten einer Katastrophe diese zu bewerten, vität sowohl die meteorologischen Hintergründe die Folgen abzuschätzen, die zeitliche Entwick- als auch die sozioökonomischen Auswirkungen lung nachzuverfolgen und die wichtigsten Fak- dieser Katastrophe analysiert. toren zu identifizieren, die für die Auswirkungen maßgeblich sind. Im Rahmen einer CEDIM FDA- Neben den breit angelegten forensischen Ana- Task Force Aktivität werden zeitnah, also wenige lysen untersuchten CEDIM Wissenschaftler Stunden bis Tage/Wochen nach dem Eintreten weitere Extremereignisse und Katastrophen einer Katastrophe, Berichte mit unterschied- vor allem disziplinär, die ebenfalls in Kapitel I lichen Schwerpunkten verfasst. In einigen Fällen kurz dargestellt sind. Hierzu zählen das Erdbe- werden diese Analysen ergänzt durch spezifische ben von Amatrice (Italien) im August 2016 mit Feldeinsätze vor Ort. einem geschätzten Schaden von 2,6 Mrd. $, die Starkniederschlagsepisode im November 2016 in Auch in den vergangenen beiden Jahren kam es Deutschland, das schwächere Erdbeben auf Lef- zu mehreren CEDIM FDA Aktivitäten, die aus- kada (Griechenland) im November 2015 sowie führlich in Kapitel I beschrieben sind. So analy- das schwere Mw=8,3 Erdbeben in Illapel (Chile) sierte ein Team von Wissenschaftlern nach dem im gleichen Monat, bei dem erstmalig auch das schweren Erdbeben in Nepal am 25. April 2015 eigene Tsunami-Modell TsuKIT zum Einsatz kam. detailliert die damit verbundenen Auswirkungen Neben diesen CEDIM Kurzberichten wurden wei- in verschiedenen Regionen und die Intensität und tere 131 Berichte zu bevorstehenden extremen Häufigkeit von Nachbeben. Bei insgesamt drei oder ungewöhnlichen Wetterereignissen welt- Feldeinsätzen (Juni und November 2015; April weit verfasst und über das Webportal „Wetter- 2016) wurden verschiedene Befragungen vor Ort gefahren-Frühwarnung“ veröffentlicht. durchgeführt, um die Situation in den Notunter- künften näher zu beleuchten, Entscheidungskri- Die wissenschaftliche Basis der FDA-Aktivitäten terien für eine Evakuierung nachzubilden und bilden verschiedene CEDIM-Projekte, deren Ziel bestehende Informationsdefizite der Bevölke- es ist, Methoden zu entwickeln oder zu verbes- rung zu eruieren. Darauf aufbauend wurde ein sern, mit denen Katastrophen im Rahmen einer konzeptionelles Modell erstellt für die Abschät- FDA Aktivität zeitnah analysiert werden können (Kapitel II). Im Rahmen eines Projekts zu Crowd- sourcing werden Naturkatastrophen in Echtzeit 1 Der Begriff “Forensic Investigation of Disasters” (FORIN) im Zusammenhang mit Naturkatastrophen wurde im Jahr mit Hilfe einer eigens entwickelten Software und 2010 durch das internationale Forschungsprogramm IRDR/ Taxonomie aus Beiträgen sozialer Medien detek- ICSU (Integrated Research on Disaster Risk / International tiert, verortet und klassifiziert. In einem weiteren Council for Science) geprägt. IRDR untersucht mit Hilfe des CEDIM-Projekt wird aus vergangenen FDA Aktivi- forensischen Ansatzes, wie aus Extremereignissen Katastro- phen werden oder nicht. täten untersucht, welche Faktoren eine effektive Vorwort 9

Kommunikation während einer Katastrophensi- Impact Weather der World Meteorological Orga- tuation erschweren und welche maßgeblichen nization (WMO) und an verschiedenen Program- Informationslücken aus welchen Gründen vor- men der World Bank und der Global Facility for herrschen. Zwei weitere Projekte haben zum Ziel, Disaster Reduction and Recovery (GFDRR). Seit das Risiko durch Winterstürme und Tsunamis mit 2016 ist CEDIM außerdem Mitglied des neu ge- probabilistischen Methoden zu schätzen. Wäh- gründeten Netzwerks Global Alliance of Disaster rend im Fall von Winterstürmen ein vereinfach- Research Institute (GADRI). ter Ansatz zum Tragen kommt, mit dem aus der Intensität und räumlichen Ausdehnung eines Weiterhin ist die Außendarstellung eine wichtige Sturms zusammen mit Daten über Wohngebäu- Säule von CEDIM. Dies wird aktiv befördert etwa de auf den Schaden geschlossen wird, erfolgt bei durch eine eigene Session bei der European Geo- der Risikoschätzung von Tsunamis eine genaue science Union (EGU), durch vielfältige Artikel bei Berechnung der Wellenausbreitung mit Hilfe der Internetpräsenz Earth System Knowledge neuester Hardware zur Parallelprozessierung. Platform (ESKP; www.eskp.de) sowie durch zahlreiche Beiträge in hochrangigen Medien (z.B. Forensische Katastrophenanalysen stehen auch New York Times, Die Zeit, FAZ, ARD, ZDF, ARTE, im Mittelpunkt der aktuellen Förderphase 2016- Deutschlandradio u.a.). Auch die Presseinforma- 2018. Dabei wurde der Forschungsansatz wei- tion, die im vergangenen Jahr die meisten Zitie- terentwickelt - durch verstärkte Integration rungen für das KIT insgesamt einbrachte, war gesellschaftlicher und sozialwissenschaftlicher von einem Wissenschaftler von CEDIM („Bilanz Fragestellungen und durch eine Profilschärfung von Naturkatastrophen seit 1900: acht Millionen auf die gesellschaftlichen Bedarfsfelder Energie, Tote, sieben Billionen Dollar Schaden“ von James Mobilität und Information. Die erst Ende 2016 Daniell). All diese Aktivitäten sind in Kapitel III zu- begonnenen neuen CEDIM Projekte, die eben- sammengefasst. falls in Kapitel II kurz vorgestellt werden, haben zum Ziel, die mit dem gesellschaftlichen Wan- Mit seinen zeitnahen Analysen zu Katastrophen, del verbundenen Änderungen des Risikos und einem Forschungsprogramm, das immer wieder der Resilienz näher zu untersuchen, insbeson- neue Themenfelder federführend erschließt, und dere hinsichtlich von Energie-, Mobilitäts- und der aktiven Öffentlichkeitsarbeit hat sich CEDIM Versorgungssystemen bzw. urbanen kritischen erfolgreich in der nationalen und internationalen Infrastrukturen. Ergänzt werden diese Projekte Forschungslandschaft etabliert und ist weithin durch Arbeiten über die Auswirkungen von Na- sichtbar bei verschiedenen Nutzergruppen wie turgefahren und sozialen Interaktionen auf die Versicherungen, Hilfsorganisationen oder staatli- Funktionalität des Straßennetzwerkes von Chile, chen Einrichtungen. Durch die neue Ausrichtung durch schnelle Schadenschätzung nach Natur- trägt CEDIM außerdem zu allen drei gesellschaft- katastrophen und durch die Modellierung von lich relevanten Bedarfsfeldern der KIT Dachstra- Nachbeben. tegie 2025, Energie, Mobilität und Information, bei und ist damit bestens aufgestellt, die Risiko- In den vergangenen beiden Jahren wurden meh- forschung am KIT weiter zu bündeln. Gemein- rere neue Kooperationen mit verschiedenen in- sam mit dem KIT-Zentrum Klima und Umwelt hat ternationalen Forschungseinrichtungen etabliert, CEDIM daher auch im vergangenen Jahr einen etwa mit der Universität Adelaide, der VU Univer- Prozess angestoßen, die am KIT bestehenden sität Amsterdam oder Deltares, einem unabhän- und breit gestreuten Forschungsaktivitäten zu gigen Institut mit Schwerpunkt in der Wasserfor- den Themenfeldern Risiken, Katastrophen und schung. Enge Verbindungen bestehen nach wie Sicherheit besser zu vernetzen und eine gemein- vor mit der Versicherungsindustrie, insbesondere same Strategie zu entwickeln. mit dem Willis Research Network (WRN), bei dem CEDIM in einem sog. „Flagship Project“ an der Michael Kunz Entwicklung eines Hagelrisikomodells für Euro- Stefan Hinz pa und Australien beteiligt ist. Weiterhin erstellt Franz Nestmann CEDIM für die Sparkassenversicherung SV eigene Risikomodelle für die beiden Gefährdungsarten Hagel und Hochwasser. Außerdem ist CEDIM verschiedenen nationalen und internationalen Programmen beigetreten: dem in 2015 einge- richteten IRDR International Center of Excellence zu kritischen Infrastrukturen und strategischer Planung, dem neuen 10-Jahres Programm High 10 FDA Reports I. FDA Activities FDA Reports

CEDIM Investigations on the 2015 Nepal Earthquake

Trevor Girard, Bijan Khazai, Geophysical Institute (GPI)

Sebnem Duzgun Middle East Technical University, Department of Mining Engineering, Ankara

Introduction combined research on communication issues and road blockages due to landslides Following the devastating 7.8 magnitude Gorkha Earthquake on 25 April 2015 and its aftershocks, Aims / Objective approximately 2.3 million people were displaced. A team of CEDIM researchers responded to the Through multiple field investigations, the CE- April 2015 Nepal Earthquakes by producing DIM team pursued a better understanding of valuable reports within days and weeks of the the post-disaster situation in Nepal and aimed to disaster and executing two separate field inves- develop and test conceptual models on the as- tigations within months. Four consecutive FDA sessment of critical infrastructure vulnerabilities reports were issued on April 27th, May 5th, May following earthquake triggered landslides, emer- 12th and July 15th. gent issues and vulnerability factors in temporary shelters, and the communication behaviour of A CEDIM research team carried out field inves- the earthquake affected communities and local tigations in Nepal to examine shelter protec- government, respectively. tion options and information needs following the devastating Nepal Earthquake. In an initial field survey from June 6 – 20th 2015, CEDIM researchers carried out a household level survey of displaced populations to analyse the shelter response situation and decision factors influenc- ing displaced households in seeking shelter and temporary housing. The team conducted 284 household surveys in 177 locations spanning 27 Municipalities/VDCs and 7 districts. In a second three-week mission starting on November 11th 2015, a survey of 420 individuals investigated the information seeking and communication behaviour of earthquake affected communities. In a third mission from April 20-30th 2016, one year after the devastation, a team of CEDIM re- Fig. 1: People from Ree and Lapa VDC Dhading searchers, including Dr. Bijan Khazai, Mr. Trevor whose villages were destroyed by landslides are Girard and Prof. Dr. H. Sebnem Duzgun con- sheltering on the hills near the district capital ducted another 45 interviews with government Dhading Besi (Photo: J. Anhorn). officials, police, and news media broadcasters to examine the key issues in the recovery process, Project status including breakdowns in the disaster information communication chain. In September 2016, the Shelter field investigations CEDIM team supported by Utsav Upreti and Shy- am Thapa from AAROH, an experienced Nepali Individuals in urban and rural areas of Nepal NGO, launched a new round of surveys which sought different shelter options from spontane- ous tent camps close to their homes, to scattered FDA Reports 11 unofficial shelter areas to designated official contact local government offices than anyone shelter sites (Figure 1). The Gorkha earthquake else. Hence, the local government plays a critical created an unprecedented need for emergency role in addressing the issues of all households. shelter as well as temporary and transitional It is important however, that the government housing. The purpose of the study was to better reaches out to those vulnerable groups men- understand the factors that increase vulnerabil- tioned above to ensure their particular needs are ity to being displaced from a disaster event. The met, as well as those isolated in rural areas who research team investigated the emergent issues may have limited capacity to communicate their with respect to decision processes of displaced needs. households seeking shelter and temporary housing. The household shelter survey (HSS) conduct- Disaster communication field investigations ed about 6 weeks after the earthquake across 7 of the most affected districts has revealed several A key focus of the interviews, particularly in the factors that aggravate population displacement. 2016 field investigations, were on disaster re- These include insufficient protection from heat sponse information and communication related and inclement weather (flooding from heavy questions. In total, 401 individuals and 20 local rains); emotional difficulties; issues with privacy officials participated in the quantitative survey including access to safe sanitation; and social and 25 key informants in the qualitative inter- difficulties including experiencing discrimination views. The results provided evidence of patterns and crime. in the information seeking and communication behaviour of the different types of respondents. While confirming the importance of social media to those with internet access, the survey revealed that individuals from rural areas and women in general have disproportionately lower access to the internet than those from urban areas and men, respectively. The findings provide evidence to support the use of local government officials and radio stations to communicate with individuals who lack internet access.

It was also observed that some forms of ICT, such as television, were more vulnerable to impacts than others like smartphones and mobile Fig. 2: Destroyed homes after the earthquake in phones. The information sources which proved Bhaktapur, Nepal. (Foto: S. Duzgun/KIT). to be most useful after the earthquake were radio, friends/family, and government officials To understand the factors that aggravate disas- (Figure 3). Challenges to obtaining information ter displacement, it is necessary to consider the in the first week after the earthquake were as- process of displacement holistically. The physi- sociated with a lack of access to communication cal impact of a hazard often will lead directly to channels, issues with disaster message content population displacement. However, factors such and not knowing what information was avail- as proximity of shelter site to damaged homes, able or where to find it. In contrast, the big- fear of aftershocks and landslides and the up- gest challenge to collecting information about coming monsoon influenced people’s decision to impacts and relief efforts for key informants- seek different shelter options. In addition, socio- was blocked roads. Since officials rely on face- economic vulnerability can exacerbate the sever- to-face communication in Nepal, blocked roads ity of disaster displacement in different ways. We prevented those in the disaster communication found that among the displaced population, the system from travelling to meet one another. elderly, female-headed households, people with disabilities, and some ethnic groups were more An official flow of disaster response information adversely affected through increased vulnerabil- was also observed in Nepal for the dissemination ity to violence, discrimination, inadequate shelter of information about aid. The official flow used or other hazards. In particular, displaced house- two-way channels of communication over one- holds in settlements whose homes and liveli- way channels, which is critical to seeking feedback hoods were wiped out by landslides are at risk from communities. Overall, individuals trusted of long-term displacement. We found that when government officials more than other agencies, households have issues, they are more likely to and expected officials to warn them of future risks. 12 FDA Reports

Finally, the surveys and interviews identified lo- power and maintenance equipment were major cal and district officials, police, radio stations obstacles to road clearance. Some remote vil- and emergent groups to be key actors in disaster lages were inaccessible for up to 10 months fol- communication. lowing the earthquake, but for the urban areas that were interviewed the road blockages were cleared within the first month following the earthquake. However, during this time prices of commodities doubled due to the double impact of road disruptions by the earthquake, and the fuel blockade. It was also found that the relief efforts and recovery activities were considerably affected by road blockages due to landslides. One of the major roads connecting Nepal to China has still not come back to its original service due to intense and large landslides triggered by the earthquake. Fig. 3: Best information sources before and after the 2015 Gorkha Earthquake. Outlook

Critical Infrastructure Vulnerability The CEDIM team presented the results of their research at the International Conference on CEDIM researchers also investigated the role of Earthquake Engineering and Post Disaster Recon- accessibility and disruptions in transportation in struction Planning in Bhaktapur, Nepal. The team the recovery process. Adverse conditions created plans to continue its research collaborations with by India’s September blockade of transportation local Nepali organizations such as the National networks across the India-Nepal border, created Society for Earthquake Technology (NSET) and a second wave of disruptions for delivery of fuel, Kathmandu Living Labs (KLL). By transferring the food and goods following the widespread road knowledge gained by the in-depth research to disruptions caused by the earthquake and sub- these local organizations who are already work- sequent landslides. CEDIM researchers found ing with government authorities, the research the average response time to clear roads due to can help influence local disaster risk reduction medium-size and large landslides was 1-2 and initiatives. 5-7 days, respectively, but lack of skilled human

Die Untersuchung des Erdbebens in Gorkha tel genutzt wurden. Interviews mit Regierungs- (Nepal) 2015 durch CEDIM beamten, Polizisten und Nachrichtendiensten wurden im April 2016 durchgeführt, um die Durch das verheerende Erdbeben der Stärke 7,8 Bewältigung der Katastrophe einschließlich der in Gorkha, Nepal am 25. April 2015 und den Schwierigkeiten in der Katastrophenkommuni- zahlreichen Nachbeben in den Folgewochen kation zu beleuchten. Die Interviews halfen un- wurden etwa 2,3 Millionen Menschen aus ihren ter anderem den CEDIM Wissenschaftlern, die Häusern vertrieben. Bei verschiedenen Feldein- verschiedenen Katastropheninformationen und sätzen hat ein Forscherteam des CEDIM in Nepal Kommunikationssysteme, die in Nepal nach dem die Vor- und Nachteile innerhalb der unterschied- Erdbeben etabliert wurden, mosaikartig zusam- lichen Notunterkünfte sowie den Informations- menzusetzen. Außerdem hat das interdisziplinä- bedarf der Menschen in Notsituationen unter- re CEDIM Wissenschaftlerteam untersucht, wel- sucht. Bereits im Juni 2015 wurde eine Umfrage che Rolle beim Wiederaufbau die Zugänglichkeit vor Ort durchgeführt, um Information über die bzw. Zerstörung der Transportwege gespielt hat. Situation in den Notunter-künften zu erhalten So haben nicht nur die Zerstörungen durch die und um mehr über Entscheidungskriterien bei Erdbeben und Hangrutschungen zu erheblichen der Auswahl der Notunterkünfte bzw. kurzzei- Schwierigkeiten bei der Versorgung der Bevöl- tigen Unterkünften zu erfahren. In einer weite- kerung geführt, sondern auch die indische Blo- ren Umfrage im November 2015 ging es insbe- ckade der indisch-nepalesischen Grenze brachte sondere darum, zu erfahren, wie die Menschen die Lieferung von Treibstoff, Lebensmitteln und nach Informationen über die Katastrophenlage weiterer Güter erheblich ins Stocken. gesucht haben und welche Kommunikationsmit- FDA Reports 13 Severe Thunderstorm Episode May/June 2016 Germany

Michael Kunz, David Piper, Bernhard Mühr, Florian Ehmele, Susanna Mohr Institute of Meteorology and Climate Research (IMK)

James Daniell Geophysical Institute (GPI)

Andreas Kron Institute for Water and River Basin Management (IWG)

During an FDA activity, CEDIM researchers focused on the meteorological aspects of the event, especially on the analysis of the persistency of specific weather patterns, and estimated inundation and runoff in Braunsbach from two surveys.

Atmospheric conditions

Prior to the thunderstorm episode, moist and warm subtropical air masses were advected towards Germany, where insolation additionally led to instability. The remarkable length of the storm episode can be attributed to a strong high-pressure ridge flanked by two low-pressure troughs persisting over several days. Moreover, the wind speed was unusually low allowing for thunderstorms to become almost stationary, yielding lo- cally extreme rain accumulations. Fig. 1: Weather phenomena associated with severe convective storms between 26 May and Hydrological and hydraulic aspects 9 June 2016 collected from various sources of information (European Severe Weather Da- During two field trips of CEDIM researchers, over tabase, newspaper articles, weather services; 700 recordings were made of water heights with Piper et al., 2016). 450 points used in the end to derive the inundation at each building in Braunsbach affected Introduction by flooding. In the catchment of the Schlossbach like in several other creeks struck by the thun- During a two-week period in May and June derstorms, driftwood and floating debris was 2016, large parts of Germany were repeatedly washed into the v-shaped valley and caused a affected by severe thunderstorms (Fig. 1). jam in a culvert upstream of Braunsbach. The Extreme rain exceeding 100 mm within 1-2 following runoff was dammed up and released hours led to dramatic rise of water levels of spontaneously comparable to a dike breach. Ve- various creeks. On 29 May, a flash flood with a locities of 10-15 ms-1 were estimated upstream wave height of up to 3.5 m damaged around 80 between Orlach and Braunsbach from a combi- buildings in Braunsbach (Baden-Württemberg; nation of video, local estimates from vegetation Fig. 2). Only three days later, an extreme flood and from local people. struck the town of Simbach (Bavaria). The Ahr valley (Rhineland-Palatinate) experienced heavy In Braunsbach, the Orlacher and Schlossbach rain events even on multiple days leading to the creeks provided the most severe flooding of the worst flood ever reported. According to extreme town, with velocities of 7-10 ms-1. The observed value analyses for 24-hour and 7-day rain totals, flood wave in the Schlossbach reached heights return periods exceeded 200 years or even more of 3.5 m scouring a new pathway in the built-up in several regions. Around 7,000+ structures area of Braunsbach as well as overtopping the were affected in some way, leading to total dam- street causing high velocity flooding down the age in the order of € 1.2 bn, mostly due to a few very steep slope, affecting many buildings on major business losses (GDV, 2016; www.gdv.de). the ground flood and flooding cellars. While the 14 FDA Reports

Orlacher Bach is designed upstream to accom- modate flows of 15-20 m³s-1, more than 80 m³s-1 were estimated. Additional runoff of 20 m³s-1 at Schlossbach contributed to the flash flood. Both estimates are consistent with a flat area check. Assuming a height of 110 mm on average across the 6.4 km² catchment, the total water volume affecting Braunsbach is estimated to 700,000 m³. The mass of debris from upstream landslides, hydraulic effects, road washouts and sedi- ment transport contributed greatly to the impact downstream in Braunsbach.

Persistence of weather and atmospheric patters

To put the 2016 thunderstorm episode in the historical context, the probability of different atmospheric patterns were statistically assessed (Piper et al., 2016) with respect to a 55-year ref- Fig. 2: Structural damage in Braunsbach after erence period. It was found that clusters of 10 the 29 May 2016 flash flood (Image credit: consecutive days exhibiting extreme precipitation James. E. Daniell). anywhere in Germany have occurred only three times before (exceedance of the 99.9% percen- 55-year reference period. Furthermore, the ap- tile; REGNIE data provided by German Weather plication of a new weather pattern method that Service DWD). Similar results were obtained re- allows us to determine the potential for thunder- garding compound events with low atmospheric storms to occur yielded a cluster of 11 consecu- stability and weak mid-troposphere flow. A 13- tive convective days. Evaluating the entire refer- day cluster for that combination, as observed ence period statistically results in a probability of in 2016, has been found only twice during the less than 1% for such a cluster as observed in 2016.

Die schwere Gewitterepisode im Mai/Juni schwindigkeit schätzen. Ein geschätzter Ab- 2016 in Deutschland flusswert von fast 100 m³s-1 in Braunsbach erklärt die schweren Auswirkungen mit 80 Während einer fast 15-tägigen Episode vom zum Teil erheblich beschädigten Gebäuden. 26. Mai bis 9. Juni 2016 kam es in Deutsch- Die Wahrscheinlichkeit für das Auftreten ei- land zu einer außergewöhnlich großen Zahl von ner Gewitterepisode vergleichbar zu der in schweren Gewittern, die mit schweren Nie- 2016 ist sehr gering. Cluster mit 10 aufein- derschlägen, teilweise auch mit Hagel und ander folgenden Tagen, an denen in einer be- Tornados, verbunden waren. Die schwersten liebigen Region in Deutschland Extremnieder- Hochwasser und Sturzfluten ereigneten sich in schlägen fielen, wurden in einem 55-jährigen Braunsbach (29. Mai) und in Simbach/Inn (1. Juni). Kontrollzeitraum nur dreimal registriert. Ähn- Bei Erhebungen vor Ort in und um Braunsbach liches gilt für die Kombination von geringer ließen sich aus der Kombination von direkten atmosphärischer Stabilität und geringer Wind- Beobachtungen, Aussagen von Betroffenen geschwindigkeit bzw. gewitterförderlichen und Videoanalysen, der Abfluss, die Schei- Großwetterlagen. telhöhe des Abflusses und die Strömungsge- FDA Reports 15 Hurricane Matthew, September 2016

Bernhard Mühr, Jan Wandel, Michael Kunz Institute of Meteorology and Climate Research (IMK)

Introduction

On 23 September 2016, hurricane Matthew arose from a tropical wave west of West Africa. Travelling westward, the area of deep convection organized into a tropical storm on 28 Septem- ber 2016 and was named “Matthew”. While in- tensifying into a category 1 hurricane, Matthew continued its westerly track and moved over the easterly Caribbean. The hurricane showed a rap- Fig. 1: Devastation in Haiti after Matthew has id intensification and deepened from a category made landfall. Image source: Engineers Without 2 into a category 5 hurricane (highest category Borders. on the Saffir-Simpson hurricane wind scale) within just 15 hours. 1-minute sustained wind speeds ricane Matthew struck South Carolina, USA, increased from 85 kt (157 kph) to 140 kt (259 near McCleanville on 8 October 2016, 15 UTC. kph). The hurricane kept category-5-intensity for Measurements with satellite and satellite based a few hours only and then became upper cat- precipitation radar showed values of more than egory 4. Matthew changed its track following a 200 mm all along Matthew’s path through the northerly direction, and on 4 October 2016 the Caribbean to the Bahamas. When Matthew was storm’s centre crossed Haiti’s coastline. After rag- in its rapid intensification stage, precipitation ing through the Bahamas Matthew followed the amounts exceeding 625 mm have been derived. coastline of Florida into a north westerly direc- Over land and along the south coast of Haiti and tion. Being a category 3 hurricane, its centre was the Dominican Republic the widespread rainfall very close to the coast but stayed offshore, shift- amounts were 400-500 mm. Hourly rain rates ing along Georgia’s Atlantic coast before making were as high as 229 mm in the inner rain bands. landfall in South Carolina for a short while as a category 1 storm on 8 October 2016. The next Matthew also brought extremely large volumes day, Matthew followed a track into an easterly of rain to the Southeast of the USA and set nu- direction away from the US-coast and lost hur- merous new records. Torrential rain fell from ricane status. Florida all the way up to Virginia and even across parts of eastern Canada. These regions not only With Matthew being a hurricane of the highest saw coastal floods, but many of the rivers had category, affecting many countries and causing new all-time high water levels causing severe billions of US$ of damage and many casualties, flooding. CEDIM started an FDA activity to investigate the hazard in terms of both meteorological and so- Loss analysis cioeconomic aspects in near real-time. Matthew claimed many fatalities reported from Track of Matthew, wind and precipitation the United States (46), the Dominican Republic (4), Colombia (1) and St. Vincent and the Gren- Matthew set many new records for intensity, adines (1). But most of all, Hurricane Matthew longevity and landfall. During its lifetime, Mat- has left behind widespread destruction across thew had multiple (4) landfalls, in Haiti, Cuba, Haiti and the death toll rose to more than 1000. the Bahamas and in South Carolina (USA). The first and most fatal landfall was near Les An- For the US, Matthew is the most expensive Atlantic glais in Haiti as a category 4 storm on 4 October hurricane since Sandy in 2012, causing an overall 2016 around 11 UTC with maximum sustained loss of more than $ 15 bn. In total, 26 Million peo- winds of 125 kt (232 kph).The next two landfalls ple were affected by the hurricane in the US. For took place in Cuba (near Juaco, 5 October 2016, the Bahamas Matthew likely was responsible for 00 UTC as category 4) and on the Bahamas the largest insured losses from a hurricane ever. (close to Freeport, 07 October 2016, 00 UTC, as In Haiti, these losses were in the order of $ 2 bn. category 4). Finally, the centre of category-1-hur- 16 FDA Reports

Further reading

http://www.wettergefahren-fruehwarnung.de/ http://www.cedim.de/download/FDA_mat- Ereignis/20161004_e.html thew_2016_report1_update2.pdf

Fig. 2: Satellite image, 06 October 2016, 08:37 UTC. (Image credit: NASA GSFC GOES Project).

Hurrikan "Matthew", September 2016 ne in Haiti. Sturm und sintflutartige Regenfälle führten zu großen Überschwemmungen. In den Mit Hurrikan “Matthew” zog Ende September betroffenen Staaten kam es zu umfangreichen / Anfang Oktober 2016 ein tropischer Wirbel- Evakuierungsmaßnahmen. Hurrikan Matthew sturm über den Atlantik und durch die Karibik, war eines der teuersten Naturereignisse für die der sich in vielerlei Hinsicht als ungewöhnlich USA mit Schäden von mehr als 15 Mrd. $. Auch erwies. Er hinterließ nicht nur ein Schneise der auf den Bahamas sorgte Matthew für größere Verwüstung, sondern stellte in Bezug auf Inten- versicherte Schäden als je ein Hurrikan zuvor. sität, Langlebigkeit und der Übertritte auf Land Angesichts der Intensität des Wirbelsturms, sei- zahlreiche neue Rekorde auf. Matthew erreich- ner Zugbahn, und seines enormen Schadenspo- te die höchste Wirbelsturm-Intensität, soweit tentials startete CEDIM eine FDA - Aktivität. In südlich gelang das keinem Hurrikan zuvor (100 Nahe-Echtzeit wurden dabei sowohl die meteo- km nördlich von Kolumbien). Matthew trat ins- rologischen Hintergründe als auch die sozioöko- gesamt 4 Mal auf Land über (Haiti, Kuba, Baha- nomischen Auswirkungen analysiert. mas, USA). Entlang seiner Zugbahn forderte der Hurrikan viele Todesopfer, mehr als 1000 allei- Short Reports 17

Short Reports

Earthquake Amatrice 2016

James Daniell, Andreas Schäfer Geophysical Institute (GPI)

Bernhard Mühr Institute of Meteorology and Climate Research (IMK)

Antonios Pomonis World Bank, Social, Urban, Rural & Resilience (GSURR)

A reduced FDA activity was undertaken post- earthquake for the 2016 Amatrice earthquake. Significant damage was seen in the epicentral area caused by one of the largest earthquakes of the last 50 years in Italy.

The earthquake caused about 295 fatalities and an estimated loss of about $ 2.58 bn, however with the recent aftershocks, the loss bill has been put at much higher with damage now in Nor- cia and other more densely populated regions in October 2016 and January 2017 earthquake sequences. In the August quakes, most affected were the towns of Amatrice and Accumoli. In contrast to the L’Aquila earthquake in 2009, the 2016 event was not preceded by a significant swarm activity, which lasted in 2009 for Fig.1: Buildings per town vs. intensity bounds as several months and caused wide-spread discus- collected from census data for this earthquake sions about earthquake predictability, awareness in August showing the small town nature. and public warning. The earthquake sequence of 2009 not only shares common characteristics Being August, the population of the mountain with the 2016 events but also those in 1997, top towns would have been higher than at the the Umbria Marche earthquake sequence, just census, with unknown numbers of people stay- 35 km north of Amatrice. This was significantly ing in hotels, hostels, with family and other lo- similar to the 2016 but caused only 11 fatalities. cations throughout the region. The region is a bit poorer than most in Italy with around likely $ Two key historical events occurred around this lo- 2.5-3 bn stock exposed in the intensities over VII. cation, including the 1639 earthquakes in Octo- It is important to note that this is a replacement ber where great damage occurred and buildings cost as many older masonry buildings may have and livestock were greatly affected and over 500 become dilapidated over the years. Many build- deaths and 800,000 crowns of building damage ings in the mountains stay vacant during other were seen. Similarly in January 1703, much dam- times. It can be seen that the population is likely age in Amatrice and Accumoli was seen with underestimated by 3+ times perhaps using build- 2000+ deaths and over 1.4 m crowns in dam- ing ratios from the V-VI region. The GDP in the ages. region over intensity VI is ca. $ 3.3 bn. The Capi- tal stock over VI is in the order of $ 18 bn. Intensities reached VIII-IX on the MMI scale – very well built structures received slight damage, The affected area was sparsely populated and whereas older buildings suffered great damage. mountainous, and is around 40 km to the North The damage seen corresponds to VIII and per- of the city of L’Aquila that was devastated in the haps very isolated VIII-IX locations on the MMI April 2009 earthquake of similar magnitude. scale. From north to south, within a distance of about 18 Short Reports

15 km, the worst affected were the villages of The October and January earthquake events in- Pescara del Tronto (part of the Arquata del Tronto cluding an avalanche caused damage and deaths commune) – population 135, Accumoli – popu- that are still being evaluated. lation 670 and Amatrice – population 2650 (incl. the surrounding villages). All three villages were situated on steep mountain ridges where slope instabilities and ground motion amplification can cause excess damage as has been seen in past events in Italy (e.g. 1976 Friuli and 1980 Irpinia). Inspection of damage photos suggests that Pescara del Tronto and Amatrice were devastated, losing a large proportion of their mostly residential building stock. In Amatrice there is a distinct new part of the village, with large public buildings and other facilities that is appar- ently less affected. The overwhelming majority of the buildings that collapsed were 2 to 4 story unreinforced stone masonry constructions with wooden floors. Some severe damage to rein- forced concrete or hybrid construction buildings has also been seen, but it is limited in number, e.g. the Hotel Roma in Amatrice. Most of the masonry buildings are very old (built prior to 1920) and in need of substantial strengthening. Fig.2: GDP per capita per province as a percent- This is a common problem in most of Italy’s old age of national, showing the disparity in income rural settlements. In the affected villages most of across the region. Amatrice in Rieti is a poorer these buildings collapsed either partly or entirely region than many in Italy. and very few will be salvaged.

Das Amatrice Erdbeben 2016 erreichten VIII – IX auf der Modified Mercalli Sca- le (MMI), bei welchen sogar sehr stabil gebaute Das Erdbeben in Amatrice, Italien, am 24. Au- Gebäude leichtere Schäden davontragen. Ältere gust 2016 um 1:36 UTC mit einer Magnitu- Gebäude erlitten hingegen sehr starke Schäden. de von 6,0 verursachte ca. 295 Todesfälle und Am stärksten wurden die Ortschaften Amat- einen geschätzten ökonomischen Schaden rice und Accumoli getroffen. Dieses gebirgige von 2,58 Mrd. $. In Anbetracht der folgenden Gebiet ist nur dünn besiedelt und liegt knapp Sequenzen an Nachbeben bis in den Janu- 40 km nördlich der Stadt L‘Aquila, die im April ar 2017 hinein, liegt der tatsächliche Schaden 2009 von einem Erdbeben ähnlicher Magnitude aber noch höher. Das Beben trat in einer Tie- zerstört wurde. fe von etwa 4,2 km auf und die Intensitäten

Excessive Rain in Southern Germany, 19th to 21st of November 2015

Bernhard Mühr, Florian Becker, Michael Kunz Institute of Meteorology and Climate Research (IMK)

Introduction 48-hour precipitation all-time records at some locations, no major flooding emerged. This fact An extreme rain event occurred in November is even more surprising since the rain event was 2015 and mainly affected the southern parts of widespread and ranked in second place for any Germany (Baden-Württemberg and Bavaria). De- day since 1951 in respect of the entire area of the spite rain amounts that exceeded previous 24- or Baden-Württemberg and for a 48 hours period. Short Reports 19

Weather pattern and previous conditions

Before the rain event, the month of November 2015 showed extreme high temperatures across central Europe with temperature deviations between 6 and 10 K above normal. During this time there was only little or even no rain in Germany, especially in Baden-Württemberg and Bavaria. Moreover, a drought that had already began in February 2015, continued in southern Germany. Some locations got as little as 60% of the usual rain amount in 2015 until mid-November (e.g. Stuttgart).

From 19 to 21 November 2016 a frontal system established over central Europe and stretched along central parts of Germany and France all the way to the central North Atlantic Ocean. The frontal system became quasi-stationary and sep- arated cold and dry air masses from subtropical moist and warm air masses to the south. With strong surface winds from south-westerly direc- Fig. 1:72-h accumulated and interpolated rain tions an effective moisture transport was estab- amount (mm), based on station data, 19.- lished. Upper level short wave troughs provided 22.11.2015, 06 UTC. some extra lifting and gave rise to small surface lows along the frontal system which moved rap- The extraordinary high rainfall volumes had only idly from west to east. On 21 November the minor impacts. Before the rain event nearly all frontal system finally crossed the Alps. Cold and rivers were running with low water levels During drier air could then penetrate from the north and the event some rivers were in flood, but the flood made its way across entire Germany. It replaced was not dangerous, as the recurrence interval of the subtropical air masses and started a wintry nearly all gauges remained below 2 years. episode. Further reading Rain amounts and river gauges Baden-Württemberg got an interpolated rain http://www.wettergefahren-fruehwarnung.de/ amount (48 hours) of 61.8 mm, which is the sec- Ereignis/20151121_e.html ond highest following the record rainfall of May 1978 (85.9 mm). At Simonswald-Obersimons- http://www.cedim.de/download/CEDIM_FDA_ wald (Black-Forest) 115 mm within 24 hours was RainEventGermany_20151120.pdf a new daily record.

Starkes Regenereignis in Süddeutschland, Baden-Württemberg registrierte den nassesten 19. - 21. November 2015 48-Stunden Zeitraum seit 1951, bezogen auf die gesamte Landesfläche gingen 61,8 mm Regen Ab dem 19. November 2015 ging eine mona- nieder. telange Trockenzeit im Süden Deutschlands zu Trotz der enormen Wassermassen blieb ein grö- Ende. Seit Februar war in Baden-Württemberg ßeres Hochwasser aus. Vor dem Regen-ereignis und Bayern gebietsweise nur gut die Hälfte der führten die Flüsse Niedrigwasser und auch mit in diesem Zeitraum sonst üblichen Regen-menge dem Starkregen stiegen die Pegel lediglich in gefallen. den Bereich eines 2-jährigen Hochwassers. Es Entlang einer quasi-stationären Luftmassen- ist vor allem der außerordentlich trockenen Vor- grenze, die sich quer über Deutschland erstreck- witterung und der geringen Bodenfeuchten zu- te, kam es zu anhaltenden, großflächigen und zuschreiben, dass sich kein großes Hochwasser ergiebigen Regenfällen bis in höchste Lagen. bilden konnte. 20 Short Reports Lefkada Earthquake Greece 2015

Andreas Schäfer Geophysical Institute (GPI)

On 17th November 2015, a Mw=6.4 earthquake other one by falling debris; 8 more were injured. was reported at the island of Lefkada in Western The impact was smaller compared to the earth- Greece. It was the strongest earthquake in that quake in 2003 and other events, e.g. in 1948 or region since 1997, although in 2003 a similarly 1953. The economic impact was estimated to be large earthquake (Mw=6.3) had hit the same re- about $ 20 m on average (range $ 5-65 m), but gion. Based on the observed aftershock pattern the indirect long-term damage may exceed the and moment centroid results, a SW-NE strike- direct damage. For Lefkada, tourism is one of the slip rupture was determined. Most quakes were most important business sectors and after any hereby located to the South-West of Lefkada. The larger earthquake, a significant decrease in the peak ground motion was modelled to be about number of visitors can be expected. In total, only 0.3 g in the epicentral region and about 0.14 g about 120 houses have been damaged, leaving for the surrounding areas. This caused in total an 20 of them uninhabitable. The whole event was earthquake intensity of MMI=VII. Two fatalities associated by a number of landslides. have been reported, one of them by rockfall, an-

Fig.1: Earthquakes and aftershocks from the last 24 hours indicating the possible fault mechanism direction (fault length not indicated). The fault projection extends through the centre of Lefkada according to aftershock activity.

Lefkada Erdbeben 2015 Gebäude). 8 weitere Menschen wurden verletzt. Insgesamt wurden ca. 120 Gebäude beschädigt, Am 17. November 2015 ereignete sich ein 20 davon sind unbewohnbar. Das Erdbeben war Mw=6,4 Erdbeben im Südwesten von Lefkada das stärkste seit 1997, die Folgen blieben jedoch im Westen von Griechenland. Dabei entstand ein deutlich geringer im Vergleich zu anderen Beben Schaden von ca. 20 Mio. $, sowie 2 Todesfälle (1 in der Region, bspw. 2003, 1948 oder 1953. durch einen Erdrutsch, 1 durch ein einstürzendes Short Reports 21 Illapel Earthquake Chile 2015

Andreas Schäfer Geophysical Institute (GPI)

During the night of the 16th to the 17th of No- indirectly during an evacuation procedure. Al- vember in 2015, a massive Mw=8.3 earthquake most 5000 buildings were damaged, with a ma- hit the coast of Central Chile. The region has jority in Chile, but also some in Argentina and been frequently hit by very strong earthquakes Brazil. Several landslides were reported, blocking in recent decades, including the 2010 Mw=8.8, streets. A tsunami was recorded in the Pacific, Maule earthquake and the 1960 Mw=9.5 Val- with maximum wave heights of at least 4 - 6 m divia earthquake. Compared to these disasters, in the epicentral region, 0.5 - 1.5 m for most of the 2015 Illapel earthquake had only a minor- the Chilean coast and <0.5 m for the remaining moderate impact. Nonetheless, it caused moder- Pacific. The TsuKIT model used proved its appli- ate damage, with an average expected economic cability for rapid impact assessments and will be loss of about $ 58 m. (range $ 12 m - $ 176 m). used for future events. In total, 16 fatalities were recorded, one them

Fig.1: Up to 4.5 m waves have been reported in Coquimbo close to the epicentre. Modelled using TsuKIT by Schaefer et al. (2015).

Chile Erdbeben 2015 zwar überschaubar, aber dennoch beachtlich mit einer Höhe von ca. 58 Mio. $. Außerdem wurden In der Nacht zum 17. November 2015 ereigne- 16 Todesfälle vermeldet. Der Tsunami erreichte te sich ein Mw=8,3 Erdbeben vor der Küste von eine maximale Höhe von ca. 4,5 m in der Nähe Chile, welches darüber hinaus einen Tsunami zur des Epizentrums und überflutete mehrere Küs- Folge hatte. Im Vergleich zu anderen Starkbeben tenabschnitte. in der Region (1960, 2010) blieben die Schäden 22 CEDIM Projects II. Research CEDIM Projects

Crowdsourcing - Using Social Media for Rapid Damage Assessment

André Dittrich, Stefan Hinz Institute of Photogrammetry and Remote Sensing (IPF)

Introduction and reliable information for further analysis and decisions. In the case of natural and man-made disasters the temporal aspect of collecting information on the event as a whole as well as local on-site information is often crucial for decision support in disaster response organizations. Thus, for example, they are able to better coordinate their actions and provide help to the affected people.

The exploitation of social media streams has been proven to yield valuable information for this purpose in a near real-time manner, such as eye- witness accounts of first responders, estimated impact areas, and identification of smaller scale hot-spots within large-scale disaster areas. This information, e.g. from micro-blogs such as Twit- ter, which is difficult or impossible to be detected by conventional sensors, can be used to quickly detect disastrous events, to complement tradi- tional sensors or to validate damage scenarios.

Fig.1:Detection of an earthquake event in Still- Aims / Objective water, Oklahoma on April 19th at 5:28:10 UTC. The map depicts the cells of the grid (dashed The objective of the project is to acquire real- gray lines), the tweets (white dots), urban areas time observations from eyewitnesses from a di- (medium gray), affected cells (green shaded rec- verse range of Social Media sites. These obser- tangles) and the earthquake epicentre (red dot). vations are used to provide rapid damage and impact estimation. Additionally, other related, Project status digitally accessible, platforms can be exploited to gather relevant (background) information on an In order to extract valuable event information event to semantically enrich Social Media data. from the mass of available social media data and Within our approach, we are aiming at different provide a real-time analysis in a fully automated natural disaster events such as storms, floods, way, several processing modules were designed volcanic eruptions, and earthquakes. The specific and implemented. challenge in using social media as an information source is to automatically extract the relevant in- Our research currently focuses on the exten- formation from the huge amount of data. There- sion and optimization of the developed TENAS fore, our goal is to develop algorithms that are software (Twitter Event Notification and Analy- able to a) handle the massive amount of data sis Service). It is a framework that monitors in real-time, b) filter the incoming data stream geo-referenced Twitter messages based on a for disaster-relevant content and c) provide quick world-wide grid (20x20 km2) in real-time and CEDIM Projects 23 operates day and night. The software is able to subscribers of the service including the most im- robustly detect significant increases in tweet vol- portant context information, such as affected ume for any area around the world compared area, the time of the detection and the automati- to a weekly updated three month moving aver- cally classified disaster type. age baseline. The test for significance is based on different statistical models such as the Pois- Outlook son distribution, the negative binomial distribution and the zero-inflated Poisson distribution, In a next step, the team will investigate the usage which are suitable for modelling count data. of the information provided by TENAS as context The subsequent analysis of the tweets’ textual knowledge for a more detailed analysis of single content identifies disaster related terms in more messages. The goal is a machine learning ap- than 60 languages and classifies the event ac- proach that is able to distinguish between mes- cordingly, e.g. as earthquake, thunderstorm, sages that are relevant for disaster management tornado, etc. The textual analysis relies on meth- and those that are not. Depending on the ex- ods from information extraction and natural periences with this approach, a finer categoriza- language processing. The classification of the tion of the relevant messages in groups - such as event is then based on similarity measures of the damaged infrastructure, food/water supply, help retrieved messages with a domain taxonomy of arrived - might be beneficial. However, in order natural disasters. to retrieve more qualitative social media data, other platforms than Twitter need to be incorpo- Within minutes after the start of a detected rated in the future (e.g. Instagram or Facebook). event, an automatic e-mail alert is sent to the

Crowdsourcing - Verwendung sozialer Me- der Domäne Naturkatastrophen analysiert und dien zur schnellen Schadensbewertung einer spezifischen Katas-trophenart zugeordnet.

Ziel des Projektes ist es, aus (Augenzeugen) Bei- Bei Erkennen eines möglichen Katastrophener- trägen in sozialen Medien, Naturkatastrophen in eignisses wird eine Benachrichtigung über E-Mail Echtzeit zu detektieren, verorten und klassifizie- an alle Abonnenten des Service gesendet. Diese ren zu können. beinhaltet die wichtigsten Informationen wie die ermittelte Ereignisart, die betroffene Region, den Hierzu wurde die Software TENAS entwickelt, Zeitpunkt der Detektion sowie eine Abschätzung welche zur Zeit erweitert und optimiert wird. der Detektions- und Klassifikationszuverlässig- TENAS extrahiert, speichert und wertet geo- keit. referenzierte Nachrichten aus dem Microblog- ging-Dienst Twitter aus. Mittels statistischer Ver- Zukünftig soll der Output von TENAS verstärkt fahren werden zunächst Gebiete identifiziert, als Vorwissen zur Einzeltextanalyse und deren die aktuell eine signifikant erhöhte Nachrichten- automatisierter Kategorisierung bezüglich ihrer dichte aufweisen. Anschließend werden diese Relevanz für das Katastrophenmanagement die- Nachrichten basierend auf Schlüsselbegriffen in nen. 60 verschiedenen Sprachen und einer Taxonomie

Improving Communication with the Public during Disaster Situations

Trevor Girard Geophysical Institute (GPI)

Introduction als access the aid they need to recover. Though the actors in the disaster communication system Effective communication of disaster response attempt to inform the public, a variety of com- information before, during and after a disaster munication problems can prevent their disaster event can save lives and help affected individu- messages from eliciting the desired response by 24 CEDIM Projects

the public (e.g., evacuating or seeking aid). The result is that many individuals are impacted by disasters and lives are lost, despite warnings being issued. Likewise, many individuals suffer after disasters, despite information about aid being available.

Three case studies were implemented to gain an Fig.1: Typhoon Haiyan surveys: top selected in-depth understanding of the communication answers by individuals to explain why warnings behaviour of disaster affected individuals and lo- failed to convince them to evacuate prior to cal government during actual disaster situations. landfall (N = 55). The disasters investigated were typhoons Haiyan (2013) and Hagupit (2014) in the Philippines and a variety of information seeking and communica- the Gorkha Earthquake (2015) in Nepal. During tion behaviour were dependent on gender, loca- 2015 and 2016, more than 800 quantitative sur- tion (urban vs. rural) and age. Information gap veys were conducted with individuals and local analyses were also completed for Cyclone Pam officials from disaster affected communities, and (2015) in Vanuatu and the Gorkha Earthquake 75 qualitative interviews were conducted with (2015) in Nepal. Almost 100 disaster updates for key informants from various government and these two disasters were analysed for their con- non-government agencies. tent and further analysis was conducted to identify potential gaps where more information could Aims / Objective improve the situational awareness, and hence response, of the public. These are in addition to A key purpose and result of earlier research was 19 disasters previously investigated. One of the the development of a methodology to analyse assumptions that was verified by comparing the disaster response information in near-real-time in information gap analyses with the survey results order to identify potential information gaps. The was that there is a wealth of valuable informa- purpose of this latest research was to identify key tion available online that fails to help disaster barriers to effectively communicate disaster re- affected communities because it never reaches sponse information (i.e., disaster warnings and them. Furthermore, information that does reach relief information) to the public during disaster the public may not be fully understood. In the situations Typhoon Haiyan case study, warnings were not understood because the hazards were nei- Project status ther explained nor emphasized (see Figure 1).

The three case studies from the Philippines and Outlook Nepal demonstrated that although the system for distributing disaster response information is Integrating the results of the case studies, infor- complex, with multiple actors and communica- mation gap analyses and disaster literature has tion channels, clear patterns for obtaining disas- culminated in the development of a conceptual ter information were observed. For instance, in model of the typical disaster communication sys- Nepal, significant variances were observed in the tem (see Figure 2). The intent of the model is to relation between location (urban vs rural) and enhance discussion of disaster communication access to information and communication chan- plans and problem solving activities. The results nels. The findings support the argument that in- of this research have practical implications for formation sent through online channels are high- any individual or agency attempting to commu- ly unlikely to reach rural areas. For all three cases, nicate with the public during a disaster situation. a clear shift was also observed in the information The goal for further research is to expand near- sources used before the disaster event and those real-time information gap analyses to include an used afterwards. Many of those who previously investigation of the variety of information sourc- tuned into television and radio began relying on es and strategies for disseminating messages to local officials, family and friends for disaster relief the public during disaster situations. information. In general, the results revealed that CEDIM Projects 25

Fig.2: Typical pathways for a national government to disseminate disaster response information to the public.

Verbesserung der Kommunikation mit der wurden qualitative Befragungen durchgeführt. Öffentlichkeit bei Katastrophensituationen Die Ergebnisse der Fallstudien zeigen wie Betrof- fene und Regierungsvertreter in beiden Ländern Das Ziel dieses Projekts ist es, festzustellen, wel- im Verlauf einer Katastrophensituation einerseits che Faktoren effektive Kommunikation innerhalb Informationen bereitstellen und verarbeiten und einer Katastrophensituation erschweren. Mit andererseits miteinander kommunizieren. Eben- Hilfe von drei Fallstudien wurde ein tiefgreifen- so zeigen die Ergebnisse, dass Verhaltensweisen des Verständnis erlangt, wie Gefahrenabwehr- hinsichtlich Beschaffung und Kommunikation informationen die Allgemeinheit während einer von Informationen von Geschlecht, Standort und Katastrophensituation tatsächlich erreichen. Alter abhängen. Zugleich haben die Fallstudien In diesem Zusammenhang wurden die Taifune verdeutlicht, wie und in welcher Form die ver- Haiyan und Hagupit auf den Philippinen (2013 schiedenen Akteure miteinander kommunizie- und 2014), sowie das Gorkha Erdbeben in Ne- ren. Schließlich zeigen die Ergebnisse der Befra- pal (2015), untersucht. Quantitative Umfragen gungen, dass nicht immer die Kernbotschaften mit Betroffenen und lokalen Amtsträgern aus der Warnungen von der Öffentlichkeit tatsäch- Katastrophengebieten wurden durchgeführt. lich verstanden werden. Mit Schlüsselpersonen aus verschiedenen Re- gierungs- und Nichtregierungsorganisationen

Towards the Development of a Global Tsunami Risk Model

Andreas Schäfer, James Daniell, Friedemann Wenzel, Robert Eschbach Geophysical Institute (GPI)

Introduction institutes with supercomputing access, or the modellers are forced to reduce model resolu- The assessment of tsunami risk is on many lev- tion either quantitatively or qualitatively. Further- els still ambiguous and under discussion. Over more, data on the vulnerability of infrastructure the last two decades, various methodologies and buildings is empirically limited to a few disas- and models have been developed to quantify ters that have occured in the recent years. tsunami risk, most of the time on a local or regional level, with either deterministic or proba- Aims / Objective bilistic background. Probabilistic modelling has significant difficulties, as the underlying tsunami Thus, a reliable quantification of socioeconomic hazard modelling demands an immense amount vulnerability is still questionable. Nonetheless, of computational time and thus limits the assess- significant improvements have been made re- ment substantially, being often limited to either cently on both the methodological aspect as 26 CEDIM Projects

well as computationally. Recently, we introduced with respect to their sensitivity on inunda- a methodological framework for a globally uni- tion results, showing that the source is inform probabilistic tsunami risk assessment. Here, deed the most important aspect of tsunami the power of recently developed hardware for impact assessment. desktop-based parallel computing (GPUs) plays a crucial role in the calculation of numerical tsuna- • In the Caribbean, potential inundation im- mi wave propagation, accelerating the computa- pacts on several Lesser Antilles states have tion speed in the order of magnitudes. A large- been quantified. scale parametric study on tectonic, geologic and paleo-seismological data was taken out for the • The potential for future big earthquakes quantification of megathrust earthquake poten- along the Puysegur Trench was analysed to- tials on more than 70 subduction zones world- gether with its respective possible impact on wide. the coast line of South-East Tasmania.

• A sensitivity study was undertaken for the city of Lima, analysing the variability of inundation impact with respect to varying source slip distributions and epicentre location of a potential Magnitude 9 event. Adaptation of empirical tsunami vulnerability functions in conjunction with methodologies from flood modelling support a more reliable vulnerability quantification. In addition, methodologies for exposure modelling in coastal areas are introduced focusing on the diversity of coastal exposure landscapes and data avail- Fig.1: Potential impact pattern of a mega-thrust ability. earthquake scenario along the Northern Lesser Antilles. Overall, the framework introduces a first overview of how a global tsunami risk modelling Project Status framework may be accomplished, while covering methodological, computational and data-driven The framework has been successfully applied for aspects. various studies: Outlook • The framework was tested by considering wave propagation benchmarks (e.g. run- Further development on the global tsunami risk up on conical island) as well as comparison model focuses on a global compilation of tsuna- of recent buoy data observations from the mi-genic earthquakes sources and their respec- 2013 Santa Cruz Island tsunami as well as tive risk potential on major coastlines around the the 2011 Tohoku tsunami. world. Future tsunami events can then be assessed quickly as part of a CEDIM's FDA activity. • The 3 main parameters in tsunami wave In addition, the numerical assessment of under- propagation and inundation modelling, the water mass movements is about to be added to source model, spatial resolution and avail- the computational framework. able roughness/friction data were compared

Fig.2: Peak Coastal Amplitude Maps with a 2% (right) and 10% (left) occurrence probability in 50 years. CEDIM Projects 27

Globales Tsunami Risiko Modell von Supercomputern in adäquater Auflösung realisiert werden. Aus diesem Grund wurde im Die Abschätzung des Tsunami Risikos ist in wei- Rahmen des Projekts eine Methode für eine glo- ten Teilen noch nicht gut umgesetzt. Während bal einheitliche, probabilistische Tsunami Risiko- der letzten zwanzig Jahre wurden verschiedene abschätzung entworfen. Dabei spielt die Kapazi- Methoden und Modelle entwickelt, um das Tsu- tät neuester Hardware zur Parallelprozessierung nami Risiko zu quantifizieren. In den meisten Fäl- (auf GPUs) eine entscheidende Rolle zur schnellen erfolgte dies auf lokaler oder regionaler Ebe- len Berechnung der Tsunamiwellenausbreitung. ne, entweder deterministisch oder eingeschränkt mit einem probabilistischen Ansatz. Die Methode wurde bereits erfolgreich für mehrere Studien eingesetzt und gegenüber beste- Die probabilistischen Ansätze können aufgrund hender Benchmarks für Wellenausbreitungsbe- ihrer Rechenintensität allerdings nur mit Hilfe rechnungen getestet.

The Web Service „Wettergefahren-Frühwarnung“ / Weather Hazards - Early Warning

Bernhard Mühr, Michael Kunz Institute of Meteorology and Climate Research (IMK)

Overview events that are enriched by images and meas- ured values, are the hallmarks of the internet The internet service “Wettergefahren-Früh- project. warnung” provides information on imminent or just occurring unusual or extreme weather Routine operation started on 1 February 2004 events worldwide; of particular interest are those and has since been continuously maintained. All weather events that are ruinous and associated warnings, special notes and detailed reports can with heavy losses. Permanent availability, daily be found in an ever-growing archive and more updated (warning) information, editorially en- than 1000 extreme events have been investigat- hanced reports of extreme or unusual weather ed so far.

Fig. 1: Interconnections between "Wettergefahren-Frühwarnung" and different Users and Stakeholders. 28 CEDIM Projects

The weather warnings of “Wettergefahren- affected were Zambia, Malawi, Zimbabwe and Frühwarnung” South Africa. According to the South African Weather Service, the year 2015 was the driest It is neither the task of “Wettergefahren-Früh- year since records began and many new all-time warnung” to publish detailed and minute-by- temperature records were set. Food prices rose minute updated warnings nor to pronounce and a famine became imminent for many people codes of conduct. It is the German (DWD) or in southern Africa. other National Weather Services that are responsible for those warnings. The key aspect of As a response to the strong 2015-2016 El Niño, “Wettergefahren-Frühwarnung” is the broader at the beginning of May 2016 devastating for- view of extreme weather events, especially when est fires developed around Fort McMurray in Al- they are associated with potentially extensive berta, Canada. 90.000 people had to flee and damage. Textual notes about forthcoming ex- were evacuated in a dangerous rescue operation traordinary events are made usually a few days – the biggest evacuation from wildfires in the before their arrival and include general informa- history of the province of Alberta. For the first tion on the nature of the extreme event, their time in the history of North America, a city with intensity and their route. Typically, a short warn- at least 60.000 inhabitants was surrounded by ing text supplemented by informative forecast severe wildfires. The resulting damage was es- maps, indicate the affected areas and what has timated to be above C$ 9 bn (€ 6.1 bn). At the to be expected. These alerts are usually updated end of November 2016 there was an extreme daily. One to three days after the event a detailed rain event in the western Mediterranean area. editorial article finalizes the activities; these arti- Within 5 days rain accumulation was more than cles contain the main findings of the event and 700 mm at some locations in northern Italy. In are enriched with data, maps, illustrations and particular Piedmont and Liguria experienced ex- figures. treme floods causing losses of more than € 1 bn.

Notable extreme weather events in 2016 Wettergefahren-Frühwarnung - part of an extensive network 68 extreme or unusual weather events have been investigated in the course of 2016 by Wet- Wettergefahren-Frühwarnung is not only a cru- tergefahren-Frühwarnung. Apart from hurricane cial part of CEDIM’s Forensic Disaster Analysis, Matthew in October 2016 and a series of flash but is also connected with several other users, floods in May/June 2016 in Germany which both stakeholders and institutions. Data, information required CEDIM FDA activities (see this issue), and advice are transferred in both directions. some of the extreme weather events are particularly worth mentioning: Further reading:

Southern Africa experienced a summer http://www.wettergefahren-fruehwarnung.de/ 2015/2016 that was extremely hot and dry. Most

Fig. 2: Screenshot taken from the archive of “Wettergefahren-Frühwarnung”. CEDIM Projects 29

Der Webservice "Wettergefahren-Frühwar- Deutet sich in den Modellvorhersagen ein ext- nung" remes Wetterereignis an, wird wenige Tage vor dem wahrscheinlichen Eintritt des Ereignisses Seit dem 1. Februar 2004 informiert das Internet- eine Vorwarnung oder Warnung formuliert. Projekt „Wettergefahren-Frühwarnung“ über In ihr wird darauf hingewiesen, welche Gebie- bevorstehende extreme oder ungewöhnliche te betroffen sein können und was auf sie zu- Wetterereignisse. Zwar richtet sich der Blick auf kommt. Etwa ein bis drei Tage nach dem Ereignis alle Vorgänge weltweit, das Haupt-augenmerk steht in Form eines redaktionellen Artikels eine liegt allerdings auf extreme Wetterereignisse in ausführliche Analyse mit umfangreichen Infor- Mitteleuropa, vor allem, wenn sie ein großes mationen zur Verfügung. Das ständig wachsen- Schadenspotential in sich bergen. „Wetterge- de Archiv umfasst bereits Analysen von mehr als fahren-Frühwarnung“ wird unabhängig von den 1000 extremen Wetterereignissen; im Jahr 2016 offiziellen Warnseiten der nationalen Wetter- kamen 68 hinzu. Wettergefahren-Frühwarnung dienste und ohne Gewähr betrieben. Detaillierte, ist Teil des FDA Konzepts von CEDIM und steht landkreisbezogene Warnungen, Verhaltensre- darüber hinaus in regem Austausch mit vielen geln oder Handlungshinweise an Einzelpersonen, weiteren Akteuren (z.B. Medien, Forschungsein- Betriebe oder Behörden sind nicht Gegenstand richtungen, externe Firmen). der Warnungen.

ATMO Forensic Prediction and Analysis - Winter Storms in Germany

Bernhard Mühr, Michael Kunz Institute of Meteorology and Climate Research (IMK)

Overview tics may be combined into an easy storm index. The letters A, B, C represent the storm’s intensity In Germany winter storms are responsible for (percentage of the area of entire Germany with a significant amount of loss and damage in re- gusts exceeding 135 kph). The extension of the spect to all natural hazards. For example, in 2007 storm is given by the numbers 1, 2, 3, depending the winter storm Kyrill caused insured losses on the fraction of the countries area with storm of buildings of more than € 2 bn. Other well- gusts of at least 75 kph. known representatives of winter storms are Ana- tol or Lothar in 1999. In addition to the routine forecasting and analysing activities of worldwide extreme weather events within the project “Wettergefahren-Frühwarnung” (see section) a method will be developed to record, classify and evaluate winter storms in Germany with the pos- Fig. 1: Storm index matrix based on wind gust sibility to extend the system to central or even measurements (extension and intensity). the whole of Europe. For winter storms practical and routine methodologies will be implemented Weather stations record the wind usually with which allow an a priori and a real-time assess- different sensor heights, and locations are often ment of possible damage (e.g. total or insured not representative of a larger area. Furthermore, losses) before and during a winter storm event the same wind speeds do not cause the same affecting Germany. losses at various locations. Coastal or mountainous areas are prone to higher wind speeds Storm intensity and potential impact of win- and minor damages have to be expected. In a ter storms further step, wind gusts will be normalized by using a threshold value, e.g. the 98% quantile The intensity of a winter storm can be described of maximum wind speeds during the recording by its maximum wind gusts which are recorded period of the specific station (Klawa, 2003). By at stations within the network of national weath- taking into account only wind gusts exceeding a er services. Another indicator is the extension of given threshold value, the local wind climate at the storm’s wind field. Both storm characteris- a specific site will gain particular importance. 30 CEDIM Projects

Fig 2: Characteristics of some recent winter storms affecting Germany. Data source: DWD (wind data), GDV (insurance data).

Additionally, population density and (insured) values are crucial for the evaluation of the loss that might occur during a severe winter storm. Both will also be considered in future steps.

Implementing an operational forecast sys- Fig 3: Storm risk matrix based on probability of tem for winter storms occurrence and storm index. quality of peak wind gusts for upcoming events With the availability of ensemble forecast data may be enhanced. (e.g. Global Forecast System, GFS) both deter- All data about past and future winter storm ministic and probabilistic forecasts on incoming events will become part of a database. The da- winter storms will be given. A storm scenario tabase entries give information about individual may be described well in advance and with de- storm events and their characteristics such as creasing uncertainty the closer the storm gets. name, date of occurrence, affected regions, Once completed and operational, the forecast intensity, damage and loss. With the use of this system predicts the probabilities of storms, their database we can compare and search for storm intensities and resulting loss amount classes. events with similar or same attributes Moreover, with the use of high-resolution forecast models further improvement of damage References assessment may be expected at least for specific smaller regions of interest and should be verified Klawa, M., Ulbrich, U.: A model for the estima- by case studies of upcoming storm events. By tion of storm losses and the identification of se- validating and adjusting gust forecast data with vere winter storms in Germany, Natural Hazards respect to past winter storm events the forecast and Earth System Sciences, 3, 725–732, 2003.

Atmo Forensic and Prediction Systems - präsentativität der Messungen an den jeweili- Winterstürme in Deutschland gen Standorten erfordern eine Normierung der Böendaten. Mit Hilfe aktueller und historischer Immer wieder ziehen winterliche Sturm- oder Or- Böen-Messdaten, Daten der Bevölkerungsdich- kantiefs über Mitteleuropa hinweg und richten te und versicherter Werte sowie Schadendaten auch in Deutschland erhebliche Schäden an; mit vergangener Sturmereignisse kann ein Vorher- dem besonders heftigen Orkan Kyrill im Januar sagesystem implementiert werden, das aktuelle 2007 erreichten die Schäden alleine an versi- Ensemble-Wind-Vorhersagedaten von Wetter- cherten Gebäuden mehr als 2 Mrd. €. Das Scha- modellen in eine Sturm-Risikomatrix überführt. denspotential eines Wintersturms lässt sich mit Diese gibt Auskunft darüber, mit welcher Wahr- einem einfachen Index beschreiben, der die vom scheinlichkeit, Intensität und daraus resultieren- Sturm insgesamt betroffene Fläche und die Flä- dem Schaden bei einem sich nähernden Sturm- che mit extremen Orkanböen miteinander kom- tief in Deutschland gerechnet werden muss. Das biniert. Dazu werden die gemessenen Spitzenbö- Vorhersagesystem soll operationell betrieben en an Stationen des Messnetzes des Deutschen werden und eine schnelle und verlässliche Scha- Wetterdienstes herangezogen. Unterschiedliche densprognose bei Winterstürmen möglich ma- Messhöhen sowie die große Variabilität der Re- chen. New CEDIM Projects 31

New CEDIM Projects

Resilience of Cities in the Course of Time

Sadeeb Ottenburger, Wolfgang Raskob Institute for Nuclear and Energy Technologies (IKET)

Introduction erality that allows dynamic model parametri- zation, dynamic specification of disruption It is desirable to gain a profound understanding scenarios, dynamic implementation of coun- on how disruptions of Critical Infrastructures (CIs), ter measures and a wide range of graphical for example, those caused by weather events or tools for analysis has to be established. terrorist attacks, affect the welfare of an urban area nowadays and in the future. Interdependen- cies of CIs may have unexpected impacts on the vulnerability of certain CIs and thus on the city as a whole. For this purpose an Agent Based Mod- elling (ABM) approach seems to be most promising i.e. CIs such as hospitals, pharmacies or power grid components may be modelled as agents.

Aims / Objective

The main goal is to establish a software tool based on ABM which supports decision making during a crisis as well as facilitates city planners or CI owners to design or redesign future or ex- Fig.1: Two versions of smart grid installations. isting CIs, respectively, in an optimal way in the Red and yellow coloured buildings indicate dif- sense of robustness and mitigation. ferent CI types in a segment of a city where varying sizes imply varying levels of importance Therefore this software tool should comprise the of the CI type. The union of green framed and following key features: the union of blue framed areas subdividing this urban segment represent the two different kinds • Flexible specification of disruption scenarios. of smart grid decompositions. • Flexible specification of model parameters in order to define different instances of certain Project status future trends. • Visualization of urban resilience related Currently we focus on smart grid technologies quantities on the level of CIs, CI-complexes - a conception of modelling topologically dif- or the city as a whole. ferent smart grid instances is work in progress. • Dynamic options for interaction: counter A generic framework for configuring disruption measures may be implemented and evalu- scenarios and adjusting model parameters is also ated during run time. work in progress but partially implemented in a testing environment. In order to achieve the main goal the following intermediate steps have to be done: Outlook

• CIs as they exist nowadays have to be mod- A workshop is scheduled for January 2017 whose elled as agents – this also includes network outcome shall be a preliminarily complete list of and environment modelling. all basic model parameters such as: time in the • The most promising future trends related to year, where a simulated disruption takes place, CIs have to be identified, evaluated and if methods for decision making, pattern of behav- applicable modelled by, e.g., smart grid tech- iour or negotiation, smart grid topology, etc. nology or car sharing • A holistic simulation framework with a gen- 32 New CEDIM Projects

Building and adjusting new and existing models Environmental aspects including the simulation of CIs is ongoing work: among others, hospitals of urban traffic and demand for services of CIs and smart grids will be central topics. will be included as well as models for negotia- tions and decision making will be evaluated and implemented.

Resilienz von Städten im Laufe der Zeit roranschläge verursacht werden. Mithilfe agen- tenbasierter Simulationen, in denen kritische Das Ziel des Projektes ist es, ein tiefgehendes Infrastrukturen oder Komponenten kritischer In- Verständnis über die Auswirkungen von Funkti- frastrukturen als Agenten modelliert werden, ist onsstörungen von kritischer Infrastrukturen auf es möglich, sowohl Krisenmanager Hilfe bei der das Wohlergehen einer städtischen Umgebung Entscheidungsfindung zu geben als auch Stadt- zu erlangen – aus der Gegenwarts- aber auch planer und Betreiber kritischer Infrastrukturen der Zukunftsperspektive. Funktionsstörungen bei der Gestaltung von kritischen Infrastrukturen können z.B. durch Wetterereignisse oder Ter- zu unterstützen.

Critical Social Interactions in Case of Emergency - Vulnerability of the Chilean Street Network to Social Interaction in Case of Emergency of Natural Disasters

Susanne Kubisch, Andreas Ch. Braun Institute of Regional Science (IfR)

Introduction Aims / Objective Infrastructure systems form the basis of the functioning of modern societies. If their destruction Due to the multiple impacts of natural disasters or degradation as well as their outage for a cer- and their interaction between the natural and tain period of time have a significant impact on human system, the analysis of the impact of nat- social and economic well-being as well as on na- ural disasters needs an integrated approach as it tional security and protection, they are termed to forms part of a project between the Institute of as critical infrastructures (CI; Attorney-General’s Photogrammetry and Remote Sensing (IPF) and Department, 2003). The street network of a the Institute of Regional Science (IfR). This project country captures a key role in the realm of critical aims to investigate the impacts of natural disas- infrastructures, as it is important for the acces- ters and social interactions on the functionality sibility and therefore for the restoration, as well of the street network in Chile. as the evacuation of the affected society in case of emergency (Anbazhagan et al., 2012). Due to Project status the social relevance and in order to understand how their functionality can be affected in emer- The project is currently collecting basic informa- gencies, the analysis of risk and vulnerability of tion and theoretical knowledge in order to pre- critical infrastructures should gain more impor- pare the field work in Chile. tance. It is not only the visibly increasing frequency of natural disasters in recent decades, but also Outlook the population movements and evacuation thus triggered, which may affect the functionality of The results of the overall project may give an the street network. Also, if social interactions in important insight on the interdependencies be- highly stressed extreme situations are character- tween the natural, technological and social sys- ized by large variability, we assume that they are tems. The results can be used in order to improve down to structural patterns due to geographical risk management and reduce the vulnerability and social constraints. Therefore, the prediction and strengthen the resilience of the street net- of social interactions and mobility should become work in Chile, with a synergetic effect on the re- crucial for effective risk management of natural silience of the social system. disasters, as the social interactions may influence the magnitude of the impact of a natural hazard. New CEDIM Projects 33

References Attorney-General’s Department. (2003). Trusted Anbazhagan, P., Srinivas, S., & Chandran, D. information sharing network for critical infra- (2012). Classification of road damage due to structure protection. Canberra ACT: Attorney- earthquakes. Natural Hazards(60), S. 425–460. General’s department, Australian Government.

Die Auswirkungen von Naturgefahren und gefahren die Funktionalität des Straßennetzwer- sozialen Interaktionen auf die Funktionali- kes beeinträchtigen können. Dem Straßennetz- tät des Straßennetzwerkes in Chile werk kommt als Teil kritischer Infrastrukturen eine Sonderrolle zu, da es für die Erreichbarkeit Das integrative Projekt zwischen dem Institut für und die Restaurierung weiterer kritischer Infra- Photogrammetrie und Fernerkundung (IPF) und strukturen sowie für die Evakuierung im Ereig- dem Institut für Regionalwissenschaft (IfR) un- nisfall von großer Bedeutung ist (Anbazhagan et tersucht die Auswirkung von Naturgefahren und al., 2012). Der Vorhersage sozialer Interaktionen sozialen Interaktionen auf die Funktionalität des und Mobilität im Ereignisfall von Naturgefahren Straßennetzwerkes in Chile. In dem Teilprojekt sollte neben der Auswirkung der Naturgefahren des IfR geht es speziell darum, wie und welche an sich daher eine immer größere Bedeutung zu- sozialen Interaktionen im Ereignisfall von Natur- kommen.

Effects of Extreme Events on EMI-Systems

Mariana Bartsch, Frank Schultmann Institute for Industrial Production (IIP)

Christina Wisotzky, Kay Mitusch Institute of Economics (ECON)

Introduction and strong inter-connectedness around the world. Facing the increasing probability of extreme events and their potentially tremendous impacts Aims / Objective on societies, it is imperative to investigate their impacts on current and future energy, mobility This project has two subprojects: (i) to investigate and information systems. This is also more than the impacts of natural disasters on supply chain valid since through the interwoven character of performance and (ii) the changed mobility be- these systems, extreme events cascade through- haviour of households in the aftermath of a dis- out. That is why natural disasters can also have aster. Furthermore recommendations on supply severe impacts far away from their place of ori- chain design, sourcing strategies and on mobility gin. This is aggravated by current globalization regulations will be given.

Fig.1: Possible scenario: flooding in urban regions. 34 New CEDIM Projects

Project status To understand mobility behaviour after extreme events it is crucial to examine every day mobil- Subproject 1: Effects of extreme events on ity behaviour and mobility patterns, i.e. to first supply chains look at behaviour under normal conditions. For this, survey data from the German Mobility Panel For a detailed investigation, different vulnerabil- (MOP) will be investigated with a special focus ity drivers for a supply chain have been identi- on trip purposes and modal split for specified fied and classified according to different types of groups of persons. In a second step, people who risk. The aforementioned extreme events are part suffered from extreme events in the past will be of the external supply chain risks, which cannot interviewed or an online questionnaire will be be influenced directly by, e.g., a corporation or used to gain information about changing mobil- country. Furthermore, the vulnerability of a place ity preferences. or location is determined by different pre-event characteristics. Those factors can be social, eco- Outlook nomic, physical and environmental characteristics. Through an intensive literature review, these For both models a common sample region will factors have been identified and a selection made be defined, where the effects on supply chains regarding the topic. The next step is the formula- as well as on changed mobility behaviour can tion and implementation of the model to assess be demonstrated. The chosen region shall have the impacts of natural disasters on supply chains a significant mobility demand as well as signifi- and the application to defined case studies. cant industry, so that impacts are substantial. In addition, a common extreme scenario has to be Subproject 2: Effects of extreme events on defined and implemented. After a current state consumer mobility requests scenario, the defined scenario shall be applied to a future city with their different needs as well From the literature review, information about as their different conditions, e.g. new forms of general mobility behaviour and mobility behav- mobility systems. Meanwhile for each project iour during extreme events could be found, but further investigations and literature reviews will studies dealing with mobility patterns after ex- be done, leading to the formulation and imple- treme events have not been found so far. mentation of models for both projects.

Die Auswirkungen extremer Naturereignis- Diese Kaskadeneffekte werden durch die starke se auf EMI-Systeme Globalisierung und Vernetzung der Welt noch weiter verstärkt. Um diesen Aspekten Rechnung Die zu erwartende Zunahme extremer Naturer- zu tragen, werden innerhalb von zwei Teilmodel- eignisse und die erheblichen potenziellen Effekte len die indirekten Auswirkungen von Naturka- für die Gesellschaft erfordern es, die Auswirkun- tastrophen untersucht. Unter Berücksichtigung gen extremer Naturereignisse für gegenwärtige globale Abhängigkeiten werden zum einen die und zukünftige Energie-, Mobilitäts- und In- Auswirkungen von Naturkatastrophen auf Sup- formationssysteme zu untersuchen. Durch den ply Chains und zum anderen die veränderte Mo- Netzwerkcharakter dieser Systeme können sich bilitätsnachfrage privater Haushalte nach einer Störungen und Unterbrechungen weit über den Katastrophe untersucht. Ort einer Naturkatastrophe hinweg ausbreiten.

Energy Risks toward 2025

James Daniell Geophysical Institute (GPI)

Introduction in the systems have often been reduced and Identifying risks from natural hazards towards shown to be vulnerable to natural perils. The in- energy sources are a key issue moving into the creasing influence of renewable energy systems future. As networks become further intercon- demands increasing focus on these renewable nected and optimised, the degrees of redundancy power sources. New CEDIM Projects 35

Aims / Objective Outlook

1. Development and extension of natural dis- In the near future, the development of histori- aster databases focusing on energy systems. cal loss databases for energy and infrastructure systems (earthquakes and all) will be undertaken, 2. Modelling of energy systems and infrastruc- including a focus on geophysical risks first in or- ture damage post-disaster (PDNA style loss der to derive preliminary loss functions for en- estimation). ergy systems. Blackouts in 2003/2004/2006 Eu- rope; 2003 USA; 1999/2009/2011 Brazil; 2005 3. Examination of small scale energy systems to Indonesia; 2001 India; 1998 New Zealand, 2011 examine the risk from natural hazards. Japan and other locations provide learning expe- riences for disasters in the energy sector. Project status For the second half of the year, the continuation Currently, the project is just beginning, and re- of the development of loss functions will be un- sources and a literature review are being undertaken in addition to the extension of histori- dertaken. In addition, collaborations with other cal energy loss database work to meteorological groups around the world are being set up includ- risks. As disasters occur, forensic disaster analy- ing VU Amsterdam, the University of Adelaide, ses will be undertaken into the energy systems Griffith University, Oxford University and the Aus- specifically. tralian National University to better understand either various perils or energy systems such as the Tesla powerwalls.

Fig.1: The rising price of power since the Fukushima disaster in Japan; and the nuclear reactors in Europe.

Energierisiken bis zum Jahr 2025 Der Fokus des Projekts liegt auf erneuerbaren Energiequellen, da diese einen wachsenden Ein- Durch die zunehmenden Abhängigkeiten und fluss am künftigen Energiemix haben werden. In steigenden Verbindungen in optimierten Netz- der Projektentwicklung wurden drei Ziele defi- werken sinkt oft die Redundanz der Energiesys- niert: teme und macht diese so anfälliger gegenüber natürliche Gefahren. Folglich steigt in Zukunft • Entwicklung und Erweiterung einer Daten- das Risiko der Energieversorgung durch Naturge- bank für Naturkatastrophen mit einem Fo- fahren. kus auf Energiesysteme. 36 New CEDIM Projects

• Modellierung von Schäden an Energiesyste- In den nächsten Schritten wird die Entwicklung men und -infrastrukturen nach einer Katas- von historischen Datenbanken zu Schäden an trophe. Energie- und Infrastruktursystemen – zunächst • Betrachtung kleinskaliger/kleiner Energie- bzgl. geophysikalischer Risiken – vorangetrieben. systeme zur Untersuchung des Risikos bzgl. Anschließend soll der Fokus auf meteorologische Naturgefahren. Risiken erweitert werden.

Earthquake Aftershock Modelling

Andreas Schäfer, James Daniell, Friedemann Wenzel Geophysical Institute (GPI)

Introduction Project Status

The occurrence of earthquakes is usually associ- Building on a previously developed earthquake ated with spatial and temporal clustering, espe- cluster classification procedure, temporal occur- cially in the case of aftershocks, when a strong rence parameters of past earthquake sequences earthquake triggers several smaller ones in its have been collected and analysed. With the help vicinity. This activity is often associated with a of various nonlinear optimization techniques it relaxation process of the earth’s crust or the ac- was possible to quantify the expected number tivation of nearby fault systems. In many cases, of aftershocks, e.g. for the next 24 hours. This aftershocks cause additional impact both in estimate is used to compute stochastic permuta- terms of socioeconomic losses but also in terms tions of possible aftershock occurrences. Based of a psychological burden arising from the fear of on various likelihood-based tests, the method- more earthquakes and thus hindering the ongo- ology proved its applicability in many cases and ing recovery process. computes an easily communicable metric for the average expected aftershock impact. Aims / Objective Outlook With the quantification of potential future aftershock impacts, the psychological burden could For future cases of strong-seismic sequences, the be lowered with adequate estimates of time-de- current version of the ARI can help to understand pendent possible aftershock risk. Thus, a quan- the socioeconomic impact of aftershocks on the tification scheme has been developed on the local society. In addition, the methodology can basis of aftershock occurrence statistics from the be further tested and improved to become a fre- present and the past, the Aftershock Risk Index quent part of upcoming earthquake FDA activi- (ARI). ties.

Fig.1: Example compilation of aftershock activity metrics for the 2010 Baja California earthquake, predicting the aftershock activity within 24h and 42h after the mainshock occurrence. New CEDIM Projects 37

Nachbeben Modellierung Aus diesem Grund wurde unter Einbezug zeitli- cher Parameter vergangener Erdbebenserien der Im Allgemeinen treten Erdbeben räumlich und Aftershock Risk Index (ARI) entwickelt. So kann zeitlich gruppiert auf. Dies ist insbesondere dann bspw. für die folgenden 24 Stunden nach einem der Fall, wenn ein starkes Hauptbeben mehrere Erdbeben die Anzahl von Nachbeben und deren Nachbeben in der näheren Umgebung auslöst. durchschnittlich zu erwarteten Auswirkungen Nachbeben ziehen oft zusätzliche sozioökono- abgeschätzt werden. Zukünftig soll diese Metho- misch Verluste nach sich und behindern den de weiterentwickelt werden und regelmäßig bei Erholungsprozess durch die verursachte psycho- Erdbeben-bezogenen FDA Aktivitäten zum Ein- logische Belastung. Letzteres könnte durch eine satz kommen. adäquate Abschätzung des Nachbebenrisikos sowie der Quantifizierung der potentiellen Aus- wirkungen abgemildert werden.

Loss Estimation and Analysis

James Daniell, Andreas Schäfer Geophysical Institute (GPI)

Bernhard Mühr Institute of Meteorology and Climate Research (IMK)

Introduction

2016 was another quiet year for earthquakes but 2 or 3 main events were seen. The Kumamo- to earthquakes in April caused damage likely in excess of $ 30 bn. In addition, the Muisne earthquake in Ecuador in the same month caused in excess of 600 deaths and exacted at least $ 1.5 bn in damage. The Amatrice earthquake (see separate article) caused 295 deaths in Au- gust, and then additional losses have occurred through large sequential earthquakes in October near the larger town of Norcia.

Aims / Objective

1. Rapid Loss Estimation post-disaster of socioeconomic losses

2. Summary information as to hazard, exposure etc. post-disaster as part of the Foren- sic Disaster Analysis effort Fig.1: The MDR of the residential buildings as part of the Ecuador risk model pro- 3. Provide a link to the PDNA methodology of duced as part of the rapid response effort. the World Bank and ECLAC The Kumamoto earthquakes were analysed in Project status real time, with the initial Kumamoto earthquake The project is ongoing and is carried out on an ad being expected to have losses only in the order of hoc basis as part of the Forensic Disaster Analysis $ 2-5 bn. However, the second shock was much project. Potentially there will be links to a new stronger and closer to Kumamoto and causing Information Portal project if possible. more damage, giving an expected total of $ 22 38 New CEDIM Projects

bn - $ 41 bn with already weakened structures many buildings were very quickly demolished. being damaged. In addition, 110 deaths oc- The Amatrice earthquake is written about in curred and the destruction of many historical chapter I. landmarks. The old adage “Earthquakes don’t kill people, The Taiwan earthquake in Kaohsiung in early buildings do” proved true this year. February caused the destruction of one building which was poorly constructed through corrupt Outlook processes such as weak concrete and filling voids with old paint cans. In the future, this work will continue as part of the “Energy Risks towards 2025” project and po- The Muisne earthquake in Ecuador was tentially as part of an Information Portal. In 2016, worked on in conjunction with a team from the number of fatalities was significantly lower the World Bank, and loss analyses were un- than the average from 1900-2016 (ca. 20,000 dertaken. This earthquake had an interest- per year). In terms of damage, the earthquake ing medium period effect where buildings economic losses were less in absolute numbers around 3 to 4 storeys in Pedernales were most than the normalized average since 1900 of ca. $ affected. The reconstruction was swift and 90 bn per year.

Schadensabschätzung und -analyse glichen mit dem Durchschnitt von 1900 – 2016 (ca. 20.000 pro Jahr). Auch die absoluten ökono- Im Projekt erfolgt im Nachgang zu einer Kata- mischen Schäden waren geringer als der norma- strophe eine schnelle Abschätzung sozioökono- lisierte Jahresdurchschnitt seit 1900 (ca. 90 Mrd. mischer Schäden für das betroffene Gebiet. Im $). Zuge einer FDA Aktivität wird darüber hinaus eine umfassende Übersichtsinformation bzgl. der Die Arbeiten basieren auf einem ad-hoc Ansatz Gefährdungslage erarbeitet. als Teil des Forensic Disaster Analysis und werden im neuen Projekt “Energy Risks towards 2025” Im Jahr 2016 wurden unter anderem verschie- fortgeführt. Des Weiteren sollen die Ergebnisse dene Erdbeben analysiert. Hier zeigte sich eine in ein Informationsportal integriert werden. signifikant niedrigere Anzahl an Todesopfern ver- Cooperations and Outreach 39 III. Cooperations and Outreach

In recent years, CEDIM researchers have col- lowing, we briefly introduce the most important laborated with more than 50 national and in- partnerships and cooperations that have contin- ternational partners, such as other universities, ued or have been newly established over the last research institutions, or companies. In the fol- two years on various risk-related topics.

Research Institutions

University of Adelaide

Various overlaps between CEDIM-GPI and Ade- through the Australian Earthquake Engineering laide have been set up with collaboration pro- Society (AEES); and research papers written to- jects within BNHCRC (Bushfire and Natural Haz- gether. There have also been collaborations be- ards CRC). In addition, various meetings have tween IIP and the University of Adelaide. occurred in Europe and Australia; collaborations

Institute for Environmental Studies, VU University Amsterdam

The collaboration between CEDIM and IVM which have considerable co-supervision overlap has increased greatly in the last few years with as well as initial workshops at both Amsterdam a combined EGU session on „Global and conti- and Karlsruhe. The complementary flood, cli- nental scale risk assessment for natural hazards: mate change and drought studies at Amsterdam methods and practice“ and another joint session. overlap well with the earthquake and meteoro- In addition, a set of cross-cutting PhD theses and logical foci in Karlsruhe. research projects are in hand at both KIT and IVM

Deltares

and subsoil resources, environment). In several different projects, such as the Afghan- istan multi-hazard risk assessment, open source risk software packages and ECA risk profiles, Deltares is an independent institute for applied there is strong cooperation with Deltares. Among research in the field of water and subsurface the various outputs are joint research papers and processes with five areas of expertise (flood risk, abstracts, as well as EGU sessions that are part of adaptive delta planning, infrastructure, water the extended research.

Insurance Industry

Willis Research Network (WRN)

CEDIM has been a member of the Willis Research lis (renamed in 2016 as Willis Towers Watson). Network (WRN) since 2009, a network of excel- WRN cooperates with CEDIM by funding a Willis lence funded by the global insurance broker Wil- fellow (full position) in the working group “At- 40 Cooperations and Outreach

mospheric Risks” (IMK-TRO) on the topic of hail success of WEHM, the model is now being ad- risk modeling. justed and applied to other hail-prone regions of WRN provides an open forum for the advance- the world such as Australia. ment of the science of extreme events through close collaboration between highly-ranked uni- References versities, insurers, reinsurers, catastrophe modelling companies, government research institu- Punge, H.J. and M. Kunz, 2016: Hail observations tions and non-governmental organizations. As and hailstorm characteristics in Europe: A review. the only German partner of the network, CEDIM Atmos Res., 176–177, 159–184, doi: 10.1016/j. researchers gather experience in the cooperation atmosres.2016.02.012 with the insurance industry and learn about their needs and research priorities. Punge, H.J., K. Bedka, M. Kunz, and A. Werner, Within a flagship project, CEDIM has developed 2014: A new physically based stochastic event a unique method to identify and characterize hail catalog for hail in Europe. Nat. Hazards, 73, events using a satellite processing algorithm in 1625-1645. DOI 10.1007/s11069-014-1161-0. cooperation with NASA. The resultant Willis Re European Hail Model (WEHM), covering 34 Eu- WRN, 2016: The Research Network: The Science ropean countries, is applied by several insurance of Managing Extremes. companies to estimate hail risk for property, mo- http://www.willis.com/willisresearchnetwork/ tor and agricultural portfolios. Building on the flagship_project/FlagBook/index.html

Sparkassenversicherung SV

In the project HARIS-SV (Hail Risk SV), CEDIM / leading to extreme loss situations. Precipitation IMK-TRO have developed a novel hail risk mod- and related return periods are quantified based el for the SV Sparkassenversicherung insurance on stochastically generated precipitation fields. company. The model estimates the probable These fields are obtained from a linear combina- maximum loss from a 200-year event (PML200) tion of orographic, background, convective and that insurers need to provide in the context of frontal rain. Hydrological modelling for different the regulatory directive Solvency II. The hail risk catchments allows relationships to be established model can also be applied to quantify damage between rainfall and runoff. Probable maximum from single events. It is based on several thou- loss (PML) for different return periods is estimat- sands of footprints from past hailstorms in Ger- ed from SV portfolio data in combination with many estimated from radar reflectivity from the modelled inundation areas. Involved in the pro- German Weather Service (DWD) radar network. ject are IMK-TRO (Working group Atmospheric The stochastic event set is created from poly- Risks), IWG (Department of Water Resources gons that are randomly constructed from the Management and Rural Engineering), and James radar-derived event set. In the last year, the hail E. Daniell. model was further improved and calibrated by considering adjusted hail-size distributions and References different damage functions based on loss expe- riences from past hailstorms. The HARIS-SV pro- Puskeiler, M., M. Kunz and M. Schmidberg- ject ended in January 2017, but will continue on er, 2016: Hail statistics over Germany derived a reduced basis in the future. from single-polarization Radar Data. Atmos. Res., 178-179, 459-470, doi:10.1016/j.atmos- Within the cooperative project FLORIS-SV (Flood res.2016.04.014 Risk SV), CEDIM aims at estimating damage related to widespread extreme flood scenarios Kunz, M., and P.S.I. Kugel, 2015: Detection of based on stochastic modelling of precipitation hail signatures from single-polarization C-band processes and resulting floods. A special focus Radar Reflectivity. Atmos. Res., 153, 565-577. is put on events that occur more or less simul- doi:10.1016/j.atmosres.2014.09.010 taneously in different catchment areas, thus Cooperations and Outreach 41

International Programs

IRDR International Center of Excellence

Since 2012, CEDIM has been cooperating with Strategic Planning was established in Germany the program Integrated Research on Disaster Risk as a network hub between the Institute of Spatial (IRDR, www.irdrinternational.org), an initiative of and Regional Planning (IREUS; Prof. J. Birkmann, the International Council for Science (ICSU), the leader), University of Stuttgart, CEDIM, the Unit- International Social Science Council (ISSC) and ed Nations University in Bonn, and the University the United Nations International Strategy for Dis- of Potsdam. This ICoE will support integrated re- aster Reduction (UNISDR). The IRDR Forensic In- search on the resilience of critical infrastructures, vestigations of Disasters (FORIN) project proposes their management and the increasing dependen- an approach that aims to uncover the root causes cy of societies on the functioning of critical infra- of disasters through in-depth investigations that structures. In October 2015, CEDIM researchers go beyond the typical reports and case studies were invited to a DFG roundtable at IREUS to conducted post-disaster events. With its focus on discuss and identify open research questions on near real-time disaster analysis to identify major the topic of integrated research for enhancing risk drivers, CEDIM’s FDA research activity is com- the resilience of critical infrastructures through plementary to IRDR’s more long-term analysis. strategic assessments and innovative planning In June 2015, a new IRDR International Centre of approaches and to develop a framework for fu- Excellence (ICoE) on Critical Infrastructures and ture research and cooperation activities.

WMO’s High Impact Weather Project HIWeather

In 2016, the World Meteorological Organization and evaluation. CEDIM (M. Kunz) is member of (WMO), part of the United Nations, launched a the Human Impacts, Vulnerability and Risk Task new 10-year World Weather Research Program Team. During the kick-off meeting at UK Met on High Impact Weather events (HIWeather). The Office in April 2016, where CEDIM presented its overall objective of HIWeather is to build resil- FDA approach, specific activities for the future, ience to a wide range of selected weather-relat- possible cooperations, and specific stakeholder ed hazards through research and its application needs and how they can be satisfied were dis- in key aspects of meteorology, hazard predic- cussed. tion, human impacts, warnings communication

CEDIM and the World Bank & GFDRR (Global Facility for Disaster Reduction and Recovery)

• A risk assessment study for Eastern Europe and Central Asia (33 countries) for earthquake and floods (https://openknowledge. worldbank.org/handle/10986/25858); • Disaster risk profiles for over 12 countries in Over the past few years as part of the Forensic Latin America and the Caribbean for earth- Disaster Analysis work and their research, a natu- quakes and tropical cyclones; ral overlap has occurred with the interests of the • Rapid post-disaster assessment of losses for World Bank and CEDIM Researchers (J. Daniell inidividual earthquakes, such as those in Ne- and A. Schaefer). Various projects including: pal 2015, Ecuador and Haiti 2016; • Review of open source global natural disas- • The Afghanistan Multi-Hazard Risk Assess- ter risk modelling software (https://www.gf- ment (F. Wenzel, B. Khazai & volunteers) in drr.org/sites/gfdrr/files/publication/UR-Soft- cooperation with Deltares, ENEA and GRF ware_Review-Web_Version-rev-1.1.pdf); Davos. 42 Cooperations and Outreach GADRI

In 2016, CEDIM became a member of the Global Framework for Disaster Risk Reduction 2015- Alliance of Disaster Research Institutes (GADRI). 2030 of the United Nations Office for Disaster Inaugurated in March 2015, GADRI is a collab- Risk Reduction (UNISDR). CEDIM participated orative platform for discussion, sharing knowl- with an active role at both the 2nd Global Sum- edge and promoting networks on topics related mit of Research Institutes for Disaster Risk Reduc- to disaster risk reduction and resilience to disas- tion in March 2015 at Kyoto University, and the ters. This program is embedded into The Sendai 1st GADRI General Assembly in March 2017.

Spin-Off Company Risklayer

Risklayer’s staff are also involved in intense aca- demic and governmental research, teaching and training activities with various universities including KIT, which allow Risklayer to remain at the The new spin-off Risklayer (www.risklayer.com) cutting edge of scientific and technical solutions. is an independent think tank established in 2016 Risklayer aims to provide tailor-made services and and based in Karlsruhe. The Risklayer team orig- practical solutions for their clients to help them inates from within CEDIM as well as the Gen- uncover short- and long-term risks and build dis- eral Sir John Monash Foundation in Australia. aster resilience.

Outreach

European Geosciences Union (EGU)

Since the foundation of CEDIM, its members more holistic perspective, James Daniell has over have been and are still actively contributing on several years compiled a database that includes the annual General Assembly of the European all natural perils, for the first time globally. In ad- Geosciences Union (EGU). dition to his regular talk at the EGU General As- sembly in April 2016, he was an invited speaker “CEDIM” session and special issue to the EGU press conference on the topic “Im- pacts and costs of natural hazards”. In recent years the assembly featured a dedicated Over 50,000 events make up the database with session to the core research fields of CEDIM. In at least 34,000 damaging events since 1900 2017 the session was renamed “Natural hazard sourced from over 90 languages. Up to $7 trillion event analyses for risk reduction and adaptation” worth of economic losses have occurred globally and is co-convened by CEDIM members. In ad- since 1900. This value was revealed by compar- dition, the submissions to this year’s session will ing economic costs for various natural disasters be published in a special issue of the open-ac- including floods, earthquakes, volcanoes, storms cess journal “Natural Hazards and Earth System and droughts using a collection of socio-eco- Sciences (NHESS)” featuring the CEDIM spokes- nomic indicators called the CATDAT Damaging man Michael Kunz as one of the editors. Natural Disaster database. Of this $7 trillion, most of financial costs have been from flooding CATDAT goes global at the EGU 2016 disasters, which accounted for around 40% of losses. Since the 1960’s, however, this trend has In the news, disaster statistics are commonly re- started to shift, with storms and storm surges ported as isolated figures, placed into context by accounting for 30% of the losses. In terms of the tragic human costs of floods, earthquakes fatalities (8 million+ deaths globally since 1900), and droughts. In order to be able to provide a earthquakes have caused the highest percentage Cooperations and Outreach 43 since 1950. By combining the data for disasters (http://www.bbc.com/news/science-environ- reported between 1900 and 2015, interesting ment-36078527) and “Der Spiegel”. trends in vulnerability across the globe are re- (http://www.spiegel.de/wissenschaft/natur/acht- vealed. Although absolute losses are increasing, millionen-tote-durch-naturkatastrophen-seit- relative losses proportionate to the global econ- 1900-a-1087842.html). omy seem to be quite constant or even decreas- ing, depending on the peril looked at. In terms Outstanding ECS Award of fatalities these are reasonably constant versus the global death rate, with a decrease versus the At the 2017 EGU General Assembly, James Dan- global population. iell will be awarded the “Outstanding Early Ca- In the fortnight following the EGU, over 800 reer Scientist Award” from the Natural Hazards newspaper articles, TV spots, science reports in Division for his distinguished scientific achieve- over 40 languages in more than 70 countries ments in the field of natural hazards. reported upon the article including the BBC

Earth System Knowledge Platform ESKP

The Earth System Knowledge Platform (ESKP) is of the FDA activities to the public and several part of the Helmholtz-Association’s Portfolio pro- stakeholders. With its expertise on natural haz- cess in the Research Field Earth and Environment ards and risks, and in particular with the current coordinated by the Helmholtz Centre Potsdam focus on Forensic Disaster Analysis in near re- GFZ. Via the ESKP webpage (www.eskp.de), CE- al-time, CEDIM is an important partner for ESKP. DIM disseminates the main results and outcomes

Media Presence (Television, Radio, Print)

CEDIM researchers are esteemed experts for re- peared on 3 March 2016 in the weekly news- gional, national and international media such as paper „Die Zeit“ with an estimated 2 million TV, radio and print media. On several different readers. For the 5-year anniversary of the Fuku- events in the past two years their evaluation and shima disaster, statistics on the changes in Fuk- assessment comments were in high demand. ushima Province with respect to power prices, Numerous appearances on TV shows, interviews radiation, employment, indirect deaths, pop- on radio shows and citations in newspapers and ulation and housing changes as well as other news portals, exhibit the wide range of visibility socioeconomic issues in various sectors such as of CEDIM’s work. Among the TV appearances tourism were analyzed. The non-exhaustive list were well-known and popular national and re- of broadcasting stations, news agencies and gional formats such as: newspapers that quoted CEDIM or reported on work or members of CEDIM shows that also on • “[W] wie Wissen” on ARD with the topic an international level CEDIM is extremely visible. “Wasserflut und Hagelschlag”; • “planet e.” on ZDF with the topic “Hagel, Fluten, Wirbelwinde - Wetterchaos in • ABC • MDR Deutschland”; • Al Jazeera • ORF • “SWR extra” on SWR with the topic “Un- • ARD • Pro Sieben wetterkatastrophe – Der Tag danach”; • Arte • Russia Today • “odysso - Wissen im SWR”on SWR with the • BBC • Scientific Ameri- topic “Wetterkatastrophen”; • BR can • “Im Gespräch” on Baden TV featuring the • DER SPIEGEL • Sky news CEDIM members James Daniell and Andreas • DIE ZEIT • Stern Schäfer in an interview with focus on earth- • Discovery Channel • SWR quake and tsunami risk modelling; • Frankfurter Allge- • The Japan Times meine • The New York Another highlight that shows CEDIM’s high pro- • HR Times file, is the full page article/infographic that ap- • L’OBS • ZDF 44 Cooperations and Outreach FDA and Short Reports

During the last two years the Forensic Disaster well as different foci for follow-up reports and Analysis Task Force has published near real-time scientific papers. Forty-one different authors reports on seven different disasters; among them from five different institutes within the KIT and were four earthquakes two thunderstorm events six external organizations worked jointly on and one hurricane (see Section II.). To some ex- these reports. tent these activities included several updates as

CEDIM at “The Change Agenda” Symposium in Oxford, England

As part of his John Monash Award work, James shifts and unique research. The attendees were Daniell represented CEDIM at the inaugural John from Uni. Cambridge, Uni. Oxford, UCL, KIT (CE- Monash Scholars’ Symposium held at Oxford DIM), University of Paris, University of Copenha- University, England, from 31 March to 1st April gen, ETH Zurich, London School of Economics, 2016, discussing topics such as natural disas- University of Amsterdam, the Australian and UK ters, refugee crises, economics, small scale wall Governments, various industry partners such as batteries and quantum computing. The attend- Commonwealth Bank, and other companies. A ees collaborate on projects, white papers and Collaboration on energy systems losses via the reports throughout the year that provide advice natural hazards project was negotiated with to the Australian government as well as other some attendees as well as with institutions back global entities on climate change, technology in Australia.

Networking Initiative at KIT

Strategic Workshops on „Risks, Catastrophes and Security“

Various institutes at KIT conduct high-level re- KIT-Center Climate and Environment, has there- search on Risks, Catastrophes, and Security, fore started a process to improve the cross-link- enabling high visibility at national and interna- ing of the various research activities at KIT and to tional levels. So far, however, there is no aligned develop a common strategy. In 2016, two work- cross-linking and networking strategy among shops were set up, which serve as kinds of nuclei the different disciplines, KIT-Centers, and KIT-Di- for the foreseen networking strategy. Based on visions on those topics. An improved networking the presentations and discussions during these is considered necessary by a large number of re- workshops, a topic paper will be compiled pre- searchers to better address the challenges that senting and highlighting the common research arise from societal changes such as the energy at KIT in the field of Risks, Catastrophes, and Se- turnaround or the decisions of the UN World Cli- curity. mate Conference COP21 in Paris 2015 for security and risk research. CEDIM, together with the Cooperations and Outreach 45

James Daniell at the press conference at the EGU General Assembly on the topic „Impacts Strategic workshop at KIT on the topics and costs of natural hazards”. of „Risk, Catastrophe and Security“ on 8 December 2016.

CEDIM at the “Wissenschaftsfestival” EFFEKTE on 27 June 2015.

CEDIM expert team in Nepal to investigate the effects of the Gorkha earthquake on 25 April 2015.

James Daniell during a field trip in Brauns- bach, Germany, after the severe flashflood on 29 May 2016.

CEDIM spokesman Michael Kunz in the ZDF science show „planet e.“ as expert for severe convective storms. 46 Publications IV. Publications 2015 and 2016

Articles in Journals and Books 2015

Anhorn J. & Khazai B. (2015): Open Space Suit- Temporary and Intermediate Shelters Following ability Analysis for Emergency Shelter after an the 2015 Nepal Earthquake”, International Jour- Earthquake, Nat. Hazards Earth Syst. Sci., 15, nal for Mass Emergencies and Disasters, submit- 789-803, doi:10.5194/nhess-15-789-2015 ted January 2016 (Manuscript in review).

Brauner F., Münzberg T., Wiens M., Fiedrich F., Kunz M., P.S.I. Kugel (2015): Detection of Hail Lechleuthner A., Schultmann F. (2015): Critical In- Signatures from Single-Polarization C-Band Ra- frastructure Resilience: A Framework for Consid- dar Reflectivity, Atmos. Res., 153, 565-577. ering Micro and Macro Observation Levels, Pro- doi:10.1016/j.atmosres.2014.09.010 ceedings of the 12th International Conference on Information Systems for Crisis Response and Mohr S., Kunz M., Geyer B. (2015): Hail Poten- Management, 24-27 tial in Europe Based on a Regional Climate Mod- el Hindcast, Geophys. Geophys. Res. Lett., 42, Daniell, J.E. (2015): “Global View of Seismic doi:10.1002/2015GL067118 Code and Building Practice Factors”, Michael Beer, Edoardo Patelli, Ioannis Kougioumtzoglou Mohr S., Kunz M., Keuler K. (2015): Develop- and Ivan Siu-Kui Au (Eds.) in “Encyclopaedia of ment and Application of a Logistic Model to Earthquake Engineering”. Springer. Estimate the Past and Future Hail Potential in Germany, J. Geophys. Res., 120, 3939-3956. Daniell, J.E. (2015): “The Evolving Risk of Earth- 10.1002/2014JD022959 quakes: Past, Present, and Future”, eds. Stuart Fraser, upcoming book: Evolving Risk, World Müller T., Meng S., Raskob W., Wiens M., Schult- Bank Press, in publication. mann, F. (2015): A Transparent Approach for Pri- oritising Security Measures, Beyerer, J. (Eds.) Se- Dittrich A., Richter D., Lucas C. (2015): “Ana- curity Research Conference: 10th Future Security lysing the Usage of Spatial Prepositions in Short Conference, Berlin, September 15-17, 2015 Pro- Messages”, eds. Gartner G. & Huang H., Spring- ceedings, 217-224 er International Publishing, Cham, 153-169. Münzberg T., Wiens M., Raskob W., Schultmann Fohringer J., Dransch D., Kreibich H., Schröter K. F. (2015): Measuring Resilience - The Benefits of (2015): Social Media as an Information Source for an Empirical Study of Power Outages by Media Rapid Flood Inundation Mapping, Nat. Hazards Data, 10th Future Security Conference 2015, Earth Syst. Sci., 15, 2725–2738, doi:10.5194/ Berlin, September 15-17, 2015, Proceedings, nhess-15-2725-2015 393 – 400

Goda K., Wenzel F., De Risi R. (2015): Empirical Münzberg T., Wiens M., Schultmann F. (2015): Assessment of Nonlinear Seismic Demand of The Effect of Coping Capacity Depletion on Crit- Mainshock-Aftershock Ground Motion Sequenc- ical Infrastructure Resilience, Proceedings of the es for Japanese Earthquakes, Front. Built Environ. 12th International Conference on Information 1:6. doi: 10.3389/fbuil.2015.00006 Systems for Crisis Response and Management, L. Palen, M. Buscher, T. Comes, and A. Hughes, Gunasekera, R., Ishizawa, O., Aubrecht, C., Blank- eds. espoor, B., Murray, S., Pomonis, A., & Daniell, J.E. (2015): “Developing an adaptive global ex- Raskob W., Bertsch V., Ruppert M., Strittmatter posure model to support the generation of coun- M., Happe L., Broadnax B., Wandler S., Deines try disaster risk profiles”, Earth-Science Reviews, E. (2015): Security of Electricity Supply in 2030, 150, 594-608. Critical Infrastructure Protection and Resilience Europe (CIPRE) Conference & Expo, The Hague, Khazai B., Anhorn J., Girard T., Brink S., Jimee the Netherlands. G.K., Parajuli B., Wagle S., Khanal O. (2015): “Emergent Issues and Vulnerability Factors in Publications 47

Raskob W., Wandler S., Deines E. (2015): Agent- Germany an exceptional event? A hydro-mete- based Modelling to Identify Possible Measures orological evaluation, Hydrol. Earth Syst. Sci., 19, in Case of Critical Infrastructure Disruption, Pro- 309-327, doi: 10.5194/hess-19-309-2015 ceedings of the 12th International Conference on Information Systems for Crisis Response and Sokolov V., Wenzel F. (2015): On the Relation Management, L. Palen, M. Buscher, T. Comes, between Point-Wise and Multiple-Location Prob- and A. Hughes, eds. abilistic Seismic Hazard Assessments, Bulletin of Earthquake Engineering, 13, 5, 1281-1301 Schröter K., Kunz M., Elmer F., Mühr B., Merz B. (2015): What made the June 2013 flood in

Articles in Journals and Books 2016

Daniell J.E., Kron A., Brand J., Muehr B., Vecere Puskeiler, M., Kunz M., Schmidberger M. A., Piper D., Kunz M., Ehmele F., Mohr S.: Ex- (2016): Hail statistics over Germany derived ceptional sequence of severe thunderstorms and from single-polarization Radar Data, Atmos. related flash floods in May and June 2016 in Res., 178-179, 459-470, doi:10.1016/j.atmos- Germany – Part 2: Hydrology and Damage As- res.2016.04.014 sessment, Nat. Hazards Earth Syst. Sci., 17 (in preparation) Richter A., Ruck B., Mohr S. & Kunz M. (2016): Interaction of Severe Convective Gusts and Typ- Girard T., Kämpf C. (submitted September 2016): A ical Urban Structures, Urban Climate, in print, Semiotic-Process Model to Enhance the Commu- http://dx.doi.org/10.1016/j.uclim.2016.11.003 nication of Disaster Response Information. Int. J. of Mass Emergencies and Disasters. Thieken A.H., Bessel T., Kienzler S., Kreibich H., Müller M., Pisi S., Schröter K. (2016): The flood Kunz-Plapp T., Hackenbruch J., Schipper J.W. of June 2013 in Germany: how much do we know (2016): Factors of Subjective Heat Stress of Ur- about its impacts?, Nat. Hazards Earth Syst. Sci. 16, ban Citizens in Contexts of Everyday Life, Natural 1519-1540; doi:10.5194/nhess-16-1519-2016 Hazards and Earth Systems Sciences, 16, 977- 994, doi:10.5194/nhess-16-977-2016 Thieken A.H., Kienzler S., Kreibich H., Kuhlicke C., Münzberg T., Wiens M., Schultman F. (2016): Un- Kunz M., Mühr B., Müller M., Otto A., Petrow derstanding Resilience: A Spatio-temporal Vul- T., Pisi S., Schröter K. (2016): Review of the flood nerability Assessment of a Population affected risk management system in Germany after the by Sudden Lack of Food, Springer International major flood in 2013, Ecology & Society 21(2): Series in Operations Research & Management 51, DOI: http://dx.doi.org/10.5751/ES-08547- Science: Advances in Managing Humanitarian 210251 Operations, 235, 257-280 Tsang Hing-Ho, Wenzel F. (2016): Setting Struc- Piper D., Kunz M., Ehmele F., Mohr S., Mühr tural Safety Requirement for Controlling Earth- B., Kron A. & Daniell J. (2016): Exceptional se- quake Mortaility Risk, Safety Science, 86, 174- quence of severe thunderstorms and related flash 183 floods in May and June 2016 in Germany – Part 1: Meteorological background, Natural Hazards Wagner Th., Goda K., Erdik M., Daniell J., Wen- and Earth System Sciences 16(12), 2835-2850 zel F. (2016): A Spatial Correlation Model of Peak Ground Acceleration and Response Spectra Punge H.J., Kunz M. (2016): Hail Observations Based on Data of the Istanbul Earthquake Rap- and Hailstorm Characteristics in Europe: A review, id Response and Early Warning System, Soil Dy- Atmos Res., 176-177, 159-184, doi: 10.1016/j. namics and Earthquake Engineering, in print atmosres.2016.02.012 48 Publications Conference Abstracts 2015

Daniell, J.E., Schäfer, A., Wenzel F. (2015a): “A York, NY, USA, 78-86. tail of eight cities: earthquake scenario risk assessment for major Australian cities”, Australi- Schäfer A.M., Daniell, J.E., Wenzel F. (2015a): an Earthquake Engineering Society Conference, “The seismic hazard of Australia - a venture into Sydney, Australia. an uncertain future”, Proceedings of the Tenth Pacific Conference on Earthquake Engineering, Daniell, J.E., Schäfer, A., Wenzel, F. (2015b): Building an Earthquake-Resilient Pacific, Sydney, “The value of life in earthquakes and other nat- Australia. ural disasters: historical costs and the benefits of investing in life safety”, Australian Earthquake Schäfer A.M., Daniell, J.E., Wenzel F. (2015b): Engineering Society Conference, Sydney, Aus- “M9 returns – towards a probabilistic Pan-Pacific tralia. Tsunami Risk Model”, Proceedings of the Tenth Pacific Conference on Earthquake Engineering, Daniell, J.E., Schaefer, A.M., Wenzel, F. (2015): Building an Earthquake-Resilient Pacific, Sydney, “What is the best use of 100 Euros to reduce the Australia. earthquake risk of a residential masonry building in a developed nation? Optimisation and Quan- Schaefer, A.M., Daniell, J.E., Wenzel, F. (2015): tification of the benefits of risk reduction”, EGU “Earthquake Forecasting Methodology Cata- General Assembly Conference Abstracts, 17, logue-A collection and comparison of the state- 1881. of-the-art in earthquake forecasting and prediction methodologies”, EGU General Assembly Daniell, J.E., Kunz-Plapp, Tina, Schaefer, A.M., Conference Abstracts, 17, 5346 Wenzel, F., Khazai, B. (2015): “What is the cost of a life in a disaster?-Examples, Practice and Schaefer, A.M., Daniell, J.E., Wenzel, F. (2015): Global Analysis”, EGU General Assembly Con- “Present, Past, Future-What earthquake clusters ference Abstracts, 17, 1882. can tell us about an upcoming Marmara Sea earthquake”, EGU General Assembly Confer- Daniell, J.E., Simpson, A., Gunasekara, R., Baca, ence Abstracts, 17, 5314. A.l, Schaefer, A.M., Ishizawa, O., Murnane, R., Tijssen, A., Deparday, V., Forni, M. (2015): Schaefer, A.M., Daniell, J.E., Wenzel, F. (2015): “Global review of open access risk assessment “The Asia-Pacific effects of a Megatsunami along software packages valid for global or continental the Tonga Trench”, EGU General Assembly Con- scale analysis”, EGU General Assembly Confer- ference Abstracts, 17, 5284. ence Abstracts, 17, 1889. Schaefer, A.M., Daniell, J.E., Wenzel, F. (2015): Daniell, J.E., Schaefer, A.M., Toro, J., Murnane, “Earthquake Cluster Analysis for Turkey and its R., Tijssen, A., Simpson, A., Saito, K., Winsemius, Application for Seismic Hazard Assessment”, H., Ward, P. (2015): “A stochastic risk assess- EGU Gen. Assem. Conf. Abstracts, 17, 5301. ment for Eastern Europe and Central Asian countries for earthquakes”, EGU General Assembly Schaefer, A.M., Daniell, J.E., Wenzel, F. (2015): Conference Abstracts, 17, 1886. “State-of-the-Art in Tsunami Risk Modelling for a global perspective”, EGU General Assembly Daniell, J.E., Skapski, J.-U., Vervaeck, A., Wenzel, Conference Abstracts, 17, 5239. F., Schaefer, A.M. (2015): “Global Earthquake and Volcanic Eruption Economic losses and costs Schmidt H.-M., Wiens M., Schultmann F. (2015): from 1900-2014: 115 years of the CATDAT da- Industrial Vulnerability against Climate Change tabase-Trends, Normalisation and Visualisation”, Impacts – An indicator-Based Analysis for Ger- EGU General Assembly Conference Abstracts, many and the Stuttgart Metropolitan Area. The 17, 8119. 5thWorld Sustainability Forum, Basel, 2015.

Dittrich A., Vasardani M., Winter S., Baldwin T. Schmidt H-M., Wiens M., Schultmann F. (2015): & Liu F. (2015): A Classification Schema for Fast Models for Regionalizing Economic Data and Disambiguation of Spatial Prepositions, in ‘Pro- their Applications within the Scope of Forensic ceedings of the 6th ACM SIGSPATIAL Interna- Disaster Analyses. Poster session at the European tional Workshop on GeoStreaming’, ACM, New Union General Assembly 2015, Vienna. Publications 49 Conference Abstracts 2016

Arnst, I., Wenzel, F., Daniell, J.E. (2016): Trends tion options in Nepal using the CEDIM FDA mod- in global earthquake loss, EGU General Assembly elled and empirical analysis following the 2015 Conference Abstracts, 18, 1812. earthquake, EGU General Assembly Conference Abstracts, 18, 1900. Girard T. (2016): “Information Seeking and Communication Behavior of Communities Af- Düzgün H.S., Wenzel F., Aydin N.Y. (2016): “As- fected by the 25 April 2015 Gorkha Earthquake sessment of Critical Infrastructure Vulnerabilities in Nepal”, Proc. of International Conference on due to Cascading Impacts of Multiple Natural Earthquake Engineering and Post Disaster Re- Hazards: A Conceptual Framework for Earth- construction Planning in Bhaktapur, Nepal. quake Triggered Landslides”, Proc. of Interna- tional Conference on Earthquake Engineering Daniell, J.E., Daniell, T.M. (2016): Updated Risk and Post Disaster Reconstruction Planning in Indices in the Asia-Pacific Region for Earthquakes Bhaktapur, Nepal. and Floods, CECAR7 Conference, Honolulu, HI. Schaefer, A.M., Daniell, J.E., Wenzel, F. (2016): Daniell, J.E., Wenzel, F., Werner, A., Schäfer, Development of a global tsunami source data- A.M., Tsang, H-H. (2016): An introduction of base – initial results, EGU General Assembly Con- tools for evaluating the influence of building ference Abstracts, 18, 7847. codes on the level of earthquake fatalities globally, 24th Australasian Conference on the Mechan- Schaefer, A.M., Daniell, J.E., Wenzel, F. (2016): ics of Structures and Materials, Perth, Australia. The SMART CLUSTER METHOD – adaptive earthquake cluster analysis and declustering, EGU Daniell, J.E., Wenzel, F., McLennan, A., Daniell, General Assembly Conference Abstracts, 18, K., Kunz-Plapp, T., Khazai, B., Schaefer, A., Kunz, 7830. M., Girard, T. (2016): The global role of natural disaster fatalities in decision-making: statistics, Schaefer, A.M., Daniell, J.E., Khazai, B., Wenzel, trends and analysis from 116 years of disaster F. (2016): The Aftershock Risk Index – quantifi- data compared to fatality rates from other caus- cation of aftershock impacts during ongoing es, EGU General Assembly Conference Abstracts, strong-seismic sequences, EGU General Assem- 18, 2021. bly Conference Abstracts, 18, 7719.

Daniell, J.E., Wenzel, F., Schaefer, A.M. (2016): Schaefer, A.M., Daniell, J.E., Mazengarb, C., The economic costs of natural disasters global- Rigby, E., Wenzel, F. (2016): The Tsunami Risk ly from 1900-2015: historical and normalised of South-East Tasmania, Australia – the impact floods, storms, earthquakes, volcanoes, bush- of the Puysegur Trench on greater Hobart, EGU fires, drought and other disasters, EGU General General Assembly Conference Abstracts, 18, Assembly Conference Abstracts, 18, 1899. 7807.

Daniell, J.E., Schäfer, A.M., Wenzel, F. (2016): Tsang, H-H., Wenzel, F., Daniell, J.E. (2016): Are The historical socioeconomic cost of earthquakes Our Buildings Safe Enough for Controlling Earth- vs. other natural disasters types globally – an quake Mortality Risk?, EGU General Assembly argument for greater funding for research and Conference Abstracts, 18, 1809. prevention, Australian Earthquake Engineering Conference 2016, Melbourne, Australia. Wagener, T., Goda, K., Erdik, M., Daniell, J.E., Wenzel, F. (2016): A Spatial Correlation Model of Daniell, J.E., Daniell, T.M., Daniell, K.A., Khaz- Peak Ground Acceleration and Response Spectra ai, B., Schaefer, A., Wenzel, F. (2016): A contin- Based on Data of the Istanbul Earthquake Rapid uation of the Asia-Pacific Risk Index for natural Response and Early Warning System, EGU Gen- disasters: extending the record for an updated eral Assembly Conference Abstracts, 18, 1810. analysis from 1900-2015, EGU General Assembly Conference Abstracts, 18, 1901. Wenzel, F., Leder, J., Daniell, J.E., Gottschäm- mer, E. (2016): Loss to residential buildings by Daniell, J.E., Schaefer, A.M., Wenzel, F., Khazai, a re-eruption of the Laacher See Volcano, EGU B., Girard, T., Kunz-Plapp, T., Kunz, M., Muehr, General Assembly Conference Abstracts, 18, B. (2016): The costs and benefits of reconstruc- 2016. 50 Publications CEDIM Reports 2015

Daniell J.E., Khazai B., Power C. (2015): Earth- Mühr B., Becker F., Kunz-Plapp T., Kunz M. quake Vulnerability Indicators for the Determina- (2015): Central European extreme rain events tion of Indirect Loss Potential. CEDIM, 73p. 19th to 21st November 2015 - Report No 1, 23rd November 2015 Daniell J.E. et al. (2015): “Nepal Earthquake – Report #1”, 27.04.2015, Nepal Report No. 1, Schäfer M., Daniell J., Wenzel F., Pomonis A., 20p. Mühr B. (2015): Greece Earthquake 17 Novem- ber 2015 – Report No 1, 17.11.2015 – Situation Daniell J.E. et al. (2015): “Nepal Earthquake – Report No. 1 Report #3”, 13.05.2015, Nepal Report No. 3, 26p. Schäfer M., Daniell J., Wenzel F., Vervaeck A. (2015): Chile Earthquake 16 Sep 2015 – Report Khazai B., Anhorn J., Girard T., Brink S., Dan- No 1, 17.09.2015 – Situation Report No. 1 – iell J.E., Bessel T., Mühr B., Flörchinger V., Kunz- 05:30 GMT Plapp T. (2015): “Shelter response and vulnerability of displaced populations in the April 25, 2015 Nepal Earthquake.”, Nepal Report No. 2.

CEDIM Reports 2016

Daniell J.E., Pomonis A., Schaefer A.M., Vervae- ck A., Gunasekera R., Ishizawa O., Skapski J-U., Mühr B., Wenzel F., Dittrich A. (2016): Amatrice Earthquake, 26.08.2016 – Situation Report No. 1 – 16:00 GMT, CEDIM FDA short report.

Mühr B., Daniell J.E., Wisotzky C., Wandel J., Becker F., Buchholz M., Baumstark S., Schäfer A.M., Dittrich A. (2016): Hurricane Matthew - 24 October 2016 – Report No. 1 – Update, CE- DIM FDA report.

Mühr B., Daniell J.E., Ehmele F., Kron A., Dittrich A., Kunz M. (2016): High water/flooding South- ern Germany May/June 2016 20 June 2016, CE- DIM FDA Report. Contact CEDIM Head Office Hermann-von-Helmholtz-Platz 1

76344 Eggenstein-Leopoldshafen paper with the quality label „Der Blaue Engel“ % recycled [email protected] 100

Issued by www.cedim.de