ARTI CLEVuWiki:AnOntology-BasedSemanticWikiforVulnerabilityAssessmentsBijanKhazaiTinaKunz-PlappChristianBu scherAntjeWegnerPublishedonline:29March2014TheAuthor(s)2014.ThisarticleispublishedwithopenaccessatSpringerlink.comAbstract Three decades of vulnerability research havegenerated a complex and growing body of knowledge. Theconcept of vulnerability, aswell asitsimplementationinvulnerability assessments, is used in various disciplines andcontexts. Correspondingly, a multitude of ideas andframeworksabouthowtoconceptualizeandmeasurevul-nerabilityexists. Toprovideastructuredrepresentationofvulnerability, wehaveset upanontology-basedsemanticwiki for reviewingandrepresentingvulnerabilityassess-ments (www.vuwiki.org). Basedona surveyof 55vul-nerability assessment studies, we rst developed anontology as an explicit reference systemfor describingvulnerabilityassessments. Theontologywas thenimple-mentedinasemanticwiki whichallows for theclassi-cation and annotation of vulnerability assessment. Theresultingsemantic wiki, VuWiki, does not aimat syn-thesizing a holistic and overarchingmodel of vulnera-bility, but at (1) providingboth scientists andpractitionerswith a uniformontology as a referencesystem; (2) providing easy and structured access to theknowledge eld of vulnerability assessments with thepossibilityforanyusertoretrieveassessmentsusingspe-cic researchcriteria; and(3) servingas acollaborativeknowledge platform that allows for the active participationof those generating and using the knowledge represented inthevulnerabilitywiki.Keywords Knowledge representation Ontology Semantic wiki Vulnerability assessment1 IntroductionThe notion of vulnerability draws on the distinctionbetweena forceful event andsomethingaffectedbytheevent, and, subsequently, it emphasizes the object exposed,for example, its characteristics, properties, or quality.Vulnerabilityasatermandasaconcepthasbeenusedforthelast 30yearsandoriginatesfromdifferent conceptuallineages, suchaspoliticalecological, politicaleconomic,and risk hazard approaches (for example, Hewitt andBurton 1971; OKeefe et al. 1976; Kates 1985; Blaikieetal. 1994;Hewitt1997;Cutteretal. 2003;Wisneretal.2004; Adger 2006; Eakin and Luers 2006, and many more).Vulnerabilityassessmentshavebecomeakeyresourcetodevelop measures and pathways for reducing risk andvulnerability and a key instrument to monitor vulnerabilityovertime. Theyhavebeenintegratedasakeyconceptincentral documents of global efforts and action plans toreduce disaster risk, such as the Hyogo Framework forAction(UNISDR2005),andclimatechangeimpacts,suchas theIPCCs Special Report onManagingtheRisks ofExtreme Events andDisasters to Advance ClimateChangeAdaptation(SREX) (IPCC2012). Moreover, the assess-ment of vulnerability has increasingly become theB. Khazai T. Kunz-Plapp(&)Geophysical Institute, Center for Disaster Management and RiskReductionTechnology(CEDIM), KarlsruheInstituteofTechnology(KIT),76187Karlsruhe, Germanye-mail:tina.kunz-plapp@kit.eduC. BuscherInstituteforTechnologyAssessmentandSystemsAnalysis(ITAS), KarlsruheInstituteofTechnology(KIT),76133Karlsruhe, GermanyA. WegnerInstituteofRegionalScienceandInstituteofUrbanandSpatialPlanning, KarlsruheInstituteofTechnology(KIT),76131Karlsruhe, Germany1 3IntJDisasterRiskSci(2014)5:5573 www.ijdrs.comDOI10.1007/s13753-014-0010-9 www.springer.com/13753touchpointinthedebatebetweenresearchcommunitiesinclimatechange anddisaster riskreduction(for example,Bohle et al. 1994; Adger 1999; KellyandAdger 2000;Thomallaet al. 2006; Kienberger et al. 2009; Birkmannetal. 2013).In addition, various communities have adopted theconcept of vulnerability and adjusted it to the needs of theirrespective discipline and elds of work. Attempts todescribe andtomeasure vulnerabilitywere made inthecontext ofassessingthevulnerabilityofsystems, thetrig-geringevents,ordriversspecictocertainresearchelds,suchasthevulnerabilityof transportation, industrial pro-duction, andenergysupplytohydrometeorologicalevents(Fekete 2009; Lebel et al. 2006) or the vulnerabilityofbuildings to earthquakes (Grunthal 1998). Vulnerableentitiesof thesamekindwereassessed, forexample, thevulnerability of communities in different regions to climatechange(Wuetal. 2002;Hahnetal. 2009)orframeworkswere developed to better understand the complex andmulti-facetedcharacteristicofsocial vulnerability(HewittandBurton1971;Hewitt1983;Blaikieetal.1994;Cutteretal. 2003;Wisneretal. 2004).Whenexaminingvulnerabilityinthecontextofnaturalhazardsanddisastersfromdisciplinary, multidisciplinary,andinterdisciplinaryperspectives, researchersandpracti-tioners haveused a multitudeofframeworks(forexample,Birkmann(2006); Fussel (2007); Hufschmidt (2011), andBirkmannet al. (2013) providedifferent exampleframe-works) and a variety of methods and technologies to gatherknowledge on the different dimensions. This has resulted inincompatibilities andinconsistencies amongvulnerabilitystudies and has made it difcult to discover, access, and usedata and information on vulnerability (NRC 2006; Giulianiet al. 2011). Consequently, developinguniversal metricsfor vulnerability assessments across disciplines is chal-lenging, whichispartlyduetothemulti-facetednatureofvulnerabilityitself, thediverseanddynamicnatureofthecomponents,andthe changing scalesofanalysis (temporaland spatial). In addition, various aspects of data availabilityand knowledge integration may potentially impede theeffectiveandefcientuseofvulnerabilityassessmentsfordisasterriskreduction(Giulianietal. 2011).Despite all the differences in theoretical frameworks,metrics, scales,and levels ofanalysis,a highlyfragmentedandwidespreadbodyof knowledgepertainingtothedif-ferentdimensionsofnatural hazardriskhasbeencreated,which may serve as a basis of science-based decision-making by individuals and households, policy makers,emergency managers, and various stakeholders in theprivate sector. Before such knowledge can be used orapplied by potential users, some conditions have to befullled: theknowledge shouldbe organized, structured,anddisseminatedeffectively; collaborationofthedifferentcommunities generating and using this knowledge has to besignicantlystrengthenedtofacilitatelearningandinfor-mationexchangebetweenthem;theinformationshouldberelevant to stakeholders; and stakeholders have to bemotivatedtouseit. Theabsenceof theseconditions cancontribute to the underutilization of knowledge, the so-calledimplementationgap(NRC2006).These considerations highlight the need to shareknowledge and data sources in an interoperable way and toensurethattheyareeasilyaccessibleanddiscoverableforusebydifferent stakeholder communitiesasoftenandaswidely as possible. Despite the myriad of vulnerabilitystudies, thereiscurrentlynoknowledgebasethat focusesexplicitlyondata, methods, current andpast researchini-tiatives, theory, and ancillary information that may behelpful for researchers andpractitioners tobetter under-standthe variedandcontextuallyspecic approaches tovulnerabilityassessmentfoundintheliterature.In the context of structuring and sharing knowledge,semantic web technologies, such as ontologies andsemantic wikis, emerged as a newtype of knowledgemanagement (forexample, SWEETontologyfororganiz-ingthevast knowledgebaseinearthandenvironmentalsciences or the ontology developed in Ontoverse formanaging knowledge and networks in life sciences).Ontologies, inshort, areaformal, hierarchicalrepresenta-tion of concepts and their interrelations in a specicknowledgedomain(Gruber 1993; RaskinandPan2005;Mainz et al. 2008). Semantic wikis use ontologies asunderlyingmodelstoembedformalizedknowledge, con-tent, structures, and links in wiki pagesvia a special mark-up language (Krotzsch et al. 2007). Currently, they areamong the most popular practical application of ontologies(Buffaetal. 2008).Tohelpaddressthecaveatsoutlinedabove,andtopro-videstructuredandguidedaccess tothefragmentedandscattered knowledge on vulnerability and vulnerabilityassessmentsfornewcomersintheeld, aswell aspracti-tioners and researchers fromother elds, a web-based,interactive knowledge platform, VuWiki (www.vuwiki.org), has been developed as a framework for the descriptionof vulnerabilityassessments. It allows for the structuredstorageandretrieval of informationbyannotationof keycategoriesandpropertiesofvulnerability.Hence,vulnera-bility assessments are comparable and easily accessible at aglance. When developing the ontology and VuWiki, the aimthereforeexplicitlywasnot tosynthesize aholisticoroverarching model for vulnerability assessments or to derivean integrated vulnerability framework.In this article, we present the attempt to develop anontology for vulnerability assessments ina theoreticallycontrolled manner. Important inuences in gaining anabstractunderstandingofthenotionofvulnerabilitycome56 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3from outside of vulnerability research, from the direction ofsystems theory (von Bertalanffy 1950; Luhmann 1995).Wetakeideasfromthesetheoriesinorder toexposetheconceptual problemsbehindthebasicquestionsof Vul-nerabilityof What? andVulnerabilitytoWhat? Bothquestions refer to problems of describing and analyzing, ontheonehand, theexposedentity(asystem), ontheotherhand, system/environment relations. Systems theory ishelpfulbecauseitallowsonetoderiveaformalrepresen-tation of concepts and their interrelations in a specicknowledge domain, which is what we mean by anontology.In addition, we present the implementation of theontologyinthesemanticwiki platform. First wedescribethe general method of developing the ontology and thewiki. Thenwepresent theresultingontologyforvulnera-bility assessments and its implementation in a semanticwiki. Thediscussionfocusesonthemethodological chal-lenges of developing the ontology and on the potentialapplication for civil protection and disaster (risk) man-agement.Thearticleconcludeswithanoutlookonfurtherdevelopments of VuWiki, inparticular byinvolvingtheresearchandpracticecommunity.2 MethodologyTheaimofdevelopingtheontologyandthesemanticwikiwas to conceive a solution-oriented framework and,through the denition of semantic rules, to allowfor aprocess of comparability across different vulnerabilityassessments and applied concepts and methods. Thedevelopment of the ontology and the implementation in thesemantic wiki platform was realized in a four step iterativeprocess that built progressively upon each step. First, asurveyofexistingliterature,data,models,andmethodstoconductassessmentsofvulnerability(bothqualitativeandquantitativeassessments)wasperformed. Inthisway, weobtained an overview of the use of vulnerability in differentresearch elds and extracted relevant classes and categoriesto structure the semantic elds. The second step wasdevelopingtheontologyitself, which(inthesenseusedhere) allowsfor theexplicit descriptionofmethods, con-cepts, andmodelsthat areuseful for theclassicationofvulnerability assessments. Third, the semantic wikiknowledgeplatformwasdesignedandimplementedusingSemantic MediaWiki (SMW). This also included thedevelopmentoftoolsfordatasearchandretrievalaswellasinformationsharinginamanner that provideslinkstoexistingtheory, data, research, andassessment initiatives.Finally, thewiki was initiallypopulatedwith55assess-ments. TheapproachtodevelopingVuWiki isoutlinedinFig. 1andbrieydescribedinthefollowingsubsections.2.1 SurveyofVulnerabilityAssessmentsWorkstartedwithathoroughsurveyofexistingvulnera-bility studies to determine the principles for organizing andpresentingkeycomponentsofquantitativeandqualitativevulnerabilityassessmentsinliterature.Thesurveycoveredover 70 articles and books from the mid-1990s to 2012 andwas not limited to standard, widely known assessments, butalso identied new, integrated multidisciplinary approa-ches. Ingeneral, theinitial reviewfocusedonbothcon-ceptual andoperational studiesandconsideredstudiesonthevulnerabilityof natural, technical, andsocial systemstakingintoaccount abroadrangeofdeterminantsofvul-nerabilityaswellastheinteractionsbetweenthedifferentdeterminants. To develop the ontology, 55 vulnerabilityassessments were selectedfrom a broadrange ofacademicdisciplines (geography, economics, social sciences, andearthquake engineering). The assessments were made indifferent elds (development studies, disaster riskreduc-tion, climatechangeadaptation, andenvironmental man-agement). The selection process for the basic stock ofliteraturewasguidedbythefollowingideasandcriteria:(1) Toincludeprimarilyvulnerabilitystudies that havebeen operationalized and empirically implementedratherthanstudiesthat focusondevelopingtheoret-icalframeworksofvulnerability.(2) Tocoverabroadrangeofvulnerabilityassessmentsinordertoobtainanoverviewoftheorganizationofvulnerability in different research elds and to extractrelevant categories to structure the knowledgedomain.(3) Toincludestudiesthataccordingtohowoftentheyare referred to by other studiesrepresent keyreferencesfortheknowledgedomainofvulnerabilityassessment.Due to the authors main elds of expertise, the 55assessments selected for developing the ontology wereslightly focused on vulnerability of the social system in thedisaster risk reduction context, but it was ensured thatenoughstudiesofvulnerabilityofecologicalandtechnicalsystemswereincluded.2.2 CollaborativeOntologyDevelopmentBasedona sub-selectionof 45vulnerabilitystudies, aninitial version of the ontology was developed and then lateradjusted using another 10 studies from our selection as testcases. Developing the ontology for vulnerability assess-mentswasguidedmainlybyvedevelopment principles:applicationindependence, natural languageindependence,orthogonality, scalability, and community involvement(Raskin and Pan 2005). The principle application-IntJDisasterRiskSci 571 3independence meansthat thestructureandcontentsofanontologyshouldbebasedupontheinherentknowledgeof the discipline, rather than on how the domain knowledgeis used (Raskin and Pan 2005, p. 1121). We implementedthis principle in our theoretically controlled approach basedon general systems theory (von Bertalanffy 1950; Luhmann1995) todevelop the structure of the ontology fromanabstract meta-perspective. Natural language indepen-dence emphasizes representing concepts rather thanterms, slang, and technical jargon and requires the deningaxiomstobelogicallyconsistent. Applyingthiskeyprin-ciple inour ontology development meant the strict dis-tinction between thinking formally in structures thatrepresent the knowledge domain vulnerability andthinkinginschoolsortheoretical conceptswhenreferringto the content of the knowledge body. Orthogonalityaddresses compound concepts, which should be decom-posable into their component parts and enable users toreusethemindifferentcontexts. Theaimofapplyingthisprinciplewasthatthetermdenitionsfortheontologyarecoherent andclear enoughinorder tobereusedwithoutrequiringotherstocreatetheirowndenitions. Scalabil-ityreferstothefactthatanyknowledgebodygrowsandontologies should therefore be easily extendable to enablespecialized domains to build upon more general ontologiesalready generated (Raskin and Pan 2005, p. 1121).Communityinvolvement,nally, referstotheideathatontologies as structural, hierarchical representations of aknowledge domain should be developed by involving thosewho contribute to the knowledge domain and are part of theusercommunityofthatknowledge.As mentioned above, the applicability of the initialontologywastestedagainstanothersetof10vulnerabilitystudies and subsequently modied in an iterative manner toaccount for gaps andissues raised ingroup discussionsamongthe authors through a seriesof hands-on workshopsin 2011. Finally, the ensuing semantic structure andontologywereevaluatedduringaworkshopwithagroupofresearcherswithabackgroundindisasterriskresearchandinformatics at theKarlsruheInstituteof Technology(KIT). Additionally, the authors used the ontology as alearningtool for analyzingvulnerabilityassessments andsubsequently interacted with an extended group ofresearchers and students through the Systemic Vulnera-bilitySeminarofferedat theKITinthefall of2012. Theuse and evaluation of the ontology and semantic wikiplatforminthe seminar allowedfor a collaborative andparticipatoryapproachtofurtherimprovethevulnerabilityontology and related tools to reach the formcurrentlypresented in thisarticle. It should benotedthat developingacomprehensiveontologyof vulnerabilityassessmentsisan adaptive process, which will continue to growandincreaseinscalewithmoreinput fromtheresearchandpracticecommunity.2.3 SemanticInterfaceandToolDevelopmentThe next stepwas to denemechanisms for translating theontology to represent and visualize the knowledge in aknowledgebaseplatform, suchas awiki. As aunifyinginterfacetoanalyzing, describing, andcompilingvariousmethods for vulnerability assessment, a platform wasSurvey of vulnerability assessmentsReview of nomenclature and basic conceptsCollaborative ontology development Iterative revision based on test-cases and discussing their fit to the ontologyDevelopment and implementation of semantic interface and tools based on Semantic MediaWikiFeedback loop including user surveyPopulation of wiki Initial database of about 55 vulnerability assessments added to wikiDraws on reviewed literatureSelection of test casesAdaption based on test-casesFig.1 TheworkowindevelopingVuWiki58 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3established according to the state-of-the-art SMW, which iseasytouse.Whiletraditionalwikiscontainonlytextsthatcomputers canneither understandnor evaluate, semanticwikisaddsemanticannotationsthat organizeinformationbasedonstructurethat canbeunderstoodbycomputers.Thisstructureisprovidedbytheontology.2.4 PopulatingtheWikiSofar, 55assessmentshavebeenaddedtoVuWiki asaninitial database. As aneffect of the selection of the 55assessmentsfor developingtheontology, most studiesinVuWiki focus on vulnerability of the social system.However, somestudies of thevulnerabilityof ecologicalandtechnical systems are alsoincluded. The process ofpopulating the wiki with more content is expected to comefromresearchersandpractitioners, asVuWiki isintendedtobefurther utilizedthroughtheinvolvement of awiderresearchandpracticecommunity.3 OntologyforVulnerabilityAssessmentsInthecontextofknowledgesharing, thetermontologyisused to mean a formal specication of shared knowledge sothat vast amounts of information, data, and concepts can bestructuredandorganizedforstorage, querying, andretrie-val(Gruber1993).Thestructureprovidedbytheontologycanbeunderstoodasaformal, hierarchical representationof concepts and their interrelations in a specic knowledgedomain(RaskinandPan2005; Mainzet al. 2008). Com-mon components of ontologies are individuals, instances orobjects(thebasisorgroundlevelobjects),classes(sets,collections, concepts, classes in programming, types ofobjects, orkindsofthings), properties(aspects, attributes,features, characteristics or parameters of theobjects andclasses), and relations (ways inwhich classes andindi-vidualscanberelatedtoeachother).Oncedeveloped,thisabstract structure enables the user to depict the structure ofitsknowledgedomainbycollectingsynonyms, capturinghierarchies likeintaxonomies, andestablishingrelationsbetweenclassesandindividuals(Mainzetal. 2008).TheontologyillustratedinFig. 2showsthat eachvul-nerabilityassessmentisthebasicobjectorInstanceoftheontology, which belongs to a Category X. In the laterimplementation in the semantic wiki, each wiki article alsoisaninstanceofaClassCandhasaValueVforPropertyP. Accordingly, the Instance: Vulnerability Study ofElectrical Systems for Category: Vulnerable System can beclassiedtohaveClass: Technical Systemthat hasProp-erty:InfrastructurewithValue:ElectricalSystem.Four keyquestions formtherst level branches orcategories of the ontology and correspond to the basic,abstract structure of the knowledge domain of vulnerabilityassessmentsand, hence, totheentrypointofthesemanticwiki later on. The four questions are simple, yet conse-quential questions, andhavebeendeducedfromvarioustheoriesandconceptsfromamultitudeofdisciplines: (1)Vulnerability of what? (2) Vulnerability to what? (3) Whatreferenceframeworkwasusedinthevulnerabilityassess-ment?and(4)Whatmethodologicalapproachwasusedinthevulnerabilityassessment?Inthefollowingsubsections,we introduce andexplainthe ontologyalongthese fourbasic questions. While reading them, it is important to keepin mind that when describing the ontology we use languagein a formal way independent of technical jargon in the eldand regardless of the fact that in some cases the same word(for example, driver) might alsobe usedina certainvulnerabilityconcept witha specic meaning. Toavoidmisunderstandings, it isthereforeimportant todistinguishbetween the formal level of the ontology and the content oftheknowledgedomaindescribedintheontology.3.1 VulnerableSystemsVulnerabilityofWhat?ToanswerthequestionVulnerabilityofwhat,weuseasystemsapproach(systemasacollectionofpartsorsub-systems) and begin with the classic concept of risk as wenditinnaturalscienceorengineeringdomains:riskisafunctionof hazardandvulnerability. Whilethehazardiscommonly referred to as the occurrence potential of atriggeringevent, thenotionofvulnerabilitydesignatesthepredisposition of people, processes, infrastructure, services,organizations, or systems to be affected, damaged, ordestroyed by the event. In this concept, hazard is theexogenousandvulnerabilityistheendogenousvariableofrisk. Somethingis at risk, exposedtoor affectedbyanoccurrence(perturbation, stress) andsomethingpossessesthe potential to change its state, a degree of sensitivity, andFig.2 Relationships among category, class, property, and instance intheontologyIntJDisasterRiskSci 591 3the capacityof response. This qualityexists a priori. Ingeneral,theobjectofobservationisthoughtofinabstractterms as a system. In developing the ontology, it is thereforeassumedthateveryresearchintovulnerabilitymustimplythedistinctionof systemandenvironment andmust, fur-thermore, distinguish types of systems and subsystemsinvestigated in the study, as this is the most basic premise ingeneral systems theory (von Bertalanffy 1950). Conse-quently, the ontology on vulnerable systems shown in Fig. 3explicitlyreferstofourclassesofvulnerablesystems: (1)natural systems for vulnerability studies referring to a setof subclasses that include physical systems (Calvalieri et al.2012), biological systems(DeLangeet al. 2010), and/orbiophysicalsystems(OBrienetal.2004);(2)socialsys-tems for vulnerability studies referring to the subclasses ofpopulationingeneral (Adger 1999; Carrenoet al. 2007),social groups, for example, communities (Cutter et al. 2003;BollinandHidajat 2006), functional systems, suchastheeconomy (Patt et al. 2010), the public nancial sector(Mechler et al. 2006) or the health sector (Hahn et al. 2009;FewandTran2010);and(3)technicalsystems,suchasvulnerability studies referring to critical infrastructure(Hellstrom2007; Kroger andZio2011). Inaddition, theontologyalsoaccountsforaseparateclassofhybridcon-ceptsreferringtointeractionsbetweenandwithinsystems,such as in societal and ecological (biophysical) subsystems(Turner et al. 2003; Gallop n 2006) or societal and technicalsubsystems(Khazaietal.2013).Overall, theontologyonvulnerablesystems showninFig. 3 mirrors some classic approaches of hazard andvulnerability research, but also includes sociological theoryintheformofastrict distinctionbetweenmodesofoper-ation of natural, social, and technical systems as well as thethesis of functional differentiation of modern society(Luhmann1997). The strictness of this argument (func-tional differentiation) could not be maintained in someFig.3 Ontologyforvulnerablesystems60 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3cases, sinceit wouldhaveleft out orpushedasideestab-lishednomenclatureinmainstreamvulnerabilityresearch.Forexample, thebranchesofIndustry,Agriculture&Forestry,orFinancialSystemarecertainlypartoftheoverall economic system, but in most studies they arereferred to as complementary systems on their own. Inrecentyears,attemptsweremadetointroducesociologicalterminologies in vulnerability research (Zehetmair 2012). Itremainstobeseentowhatextenttheseattemptswillmeetwithacceptance.3.2 VulnerabilityDriversVulnerabilitytoWhat?One of the basic traits of the concept of vulnerability is theneed to analyze the relationship between system andenvironment regardingcontingent occurrences (shock) orrather slowlydevelopingchanges leadingtounsafecon-ditions (continuous stressors). However, there are manynuances in the nature of the correlation between hazard andvulnerability. Byassertingthat hazardandvulnerabilityaremutuallyconditioningsituationsandneithercanexiston its own, Cardona (2003) raises awareness towardsconceptualissueswiththeaprioriexistenceofhazardandvulnerability separate fromeach other. Therefore, it isimportant to highlight the theoretical model behind thevulnerabilityanalysis.The dominant concept in vulnerability research is that offactor-theoretical models of anexplanationof causeandeffectrelationships,whichrefertotheideaofcausality.In our ontology, the termdriver was chosen as anabstract termtoanswer the secondbasic question: Vul-nerability to what? In the ontology, driver refers toinstantaneouseventsand/orlong-termprocessesaswellasto external and/or internal causes. Among many otherfeatures, general systems theory claims that systemsmaintaincontacttotheirrespectiveenvironmentinaveryselective fashion, despite sustainingaboundarybetweenthesystemanditsenvironment. Intermsofcausality, theargumentscontendthat insustainingaboundary, systemscut off manycausalities, whilesimultaneouslytheymustcontrol some, but not all, causalitiesvital fortheir repro-duction (Luhmann 1995). Those productive causes must beemployed to some extent within the system (as endogenousfactors), while others remain environmental causes (asexogenous factors). In this sense, potentially hazardouseffects on the systemmust be dened as unproductivecausesthatcanoccureitheroutsideof(external)orinsidethesystem(internal). Driver inour ontologythereforeindicateshowatriggeringevent orprocesscaninuence,affect,ordeviatethe stability/equilibrium ofa system, thatis, establish the conditions for maintenance of physicalstructuresorthereproductionoflivingsystems.Insteadofdiscussing negative and positive effects, which is, as ajudgment,alwaysobserver-related,wecandistinguishinamore abstract way a driver as a productive or unproductivecauserelatedtothesysteminfocus.Forfurtherstructuringthedriverintheontology,wechosetheclassesnatural andsocial drivers. Typicaldriversinnatural hazardresearch, whichact outsideof asystem, are called natural drivers. In the ontology shownin Fig. 4, they are furthersubdivided intothreesubclasses:geophysical drivers(earthquakes, volcaniceruption, land-slides, tsunamis), hydrometeorological drivers (tropicalcyclones, tornados, oods, coastal stormsurges, droughts,andsoon), or biological/ecological drivers(for example,infestationor loss of biodiversity). Whether a systemisvulnerable to processes of endogenous risk production(self-endangering) of a systemitself is of importance.While, for example, fromtheperspectiveof theeldofengineeringthedominatingcanonofvulnerabilityassess-mentsisconcernedwithexogenousnaturalhazards(forexample,thevulnerabilityofabuildingtoanearthquake),studiesthat analyzevulnerabilityfromasocietal perspec-tivefocusonendogenousprocessesofsocialdriversofvulnerability. Those assessments typically cover socialinequalities, political systems, andpoliciesasdrivers(forexample, Pelling2003; Brookset al. 2005; Wisner 2006;Hahnet al. 2009) or theyconcentrate onhowdecision-makingprocessescontributetocreatingvulnerability, likeineconomics(forexample,Smithson1993).Ourontologytriestointegrateclassicfeaturesofvulnerabilityresearch,whileremainingopentorecent theoretical developmentsthat maybeimplementedinvulnerabilityassessmentsinthenearfuture. Next tothenatural driver,weattributeconsiderable importance to the social driver and identifysocial inequality, governance, war and conict, andanthropogenic impact as different subclasses of driverswithinthesocialdriverclass(Fig.4).Furthermore, theconceptual decisionsontheproper-tiesof thevulnerabilitydriversshouldbemadeclear inevery study by using a temporal scope of the drivers(observed as a continuous stressor or discrete shock),spatial scope of the driver (local, regional or globalimpacts), and, in case of hybrid events where there is morethanonedriver, theinteractionbetweendifferent drivers(forexample,cascadingandlinkedhazards)(Fig.4).3.3 ReferenceFrameworkThe framework of reference of all vulnerability studiescorrelates with the answers to the core questions ofVulnerabilityofwhat?andVulnerabilitytowhat?Ingeneral, wedistinguishthreedimensionsofassessmentfactual(andmorespecic,spatial),temporal,andsocialwhendescribingthereferenceframeworkofvulnerabilitystudies (Fig.5). In this way, the assessments differ inIntJDisasterRiskSci 611 3regardtothescopeofassessment inthesocial dimension(individuals: Adger1999; households: Turneret al. 2003;EriksenandSilva2009; communities: BollinandHidajat2006; Wisner2006), thespatial dimension(region: RanciandMigliavacca2010; country: Brookset al. 2005; sub-city: Armas2008), andthetemporal dimension(point oftime: Kienberger et al. 2009; mediumterm: Hahnet al.2009;longterm:Lietal. 2010).An additional class, the target users (for example, sci-entists, policy makers, local authorities, emergency manag-ers, insurance companies) for whom the vulnerabilityassessment is madeis alsodescribed ineachstudy. It is anadditional class of the reference framework in our ontology.Eachclassvariesregardingthescopeof assessment withwhichresearchersoperate.Tobetterillustratesomeofthedistinctions used in the ontology, three examples arepresentedbelowforeachofthethreedimensions(spatial,temporal, and social) used in the reference framework.3.3.1 Example1:Fact/SpatialDimensioninVulnerabilityofCriticalInfrastructuresThefactdimensionofthevulnerablesysteminwhichthespatial aspect isthemost important specicationreferstothe distinction of elements within the systemand to spatialdistinctions, for example, thespatial realizationof inter-relatedelements. Whilethespatial dimensionof thevul-nerability of geographical or political units or entitiesmight be rather simple and the focus of vulnerabilityanalysesmightbecities(Pelling2003;Prasadet al.2009),regions (Birkmann et al. 2012) or whole nations (Birkmannet al. 2011; GAO2011; Welleet al. 2012), thesituationFig.4 Ontologyforvulnerabilitydrivers62 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3becomesmorecomplexwhenthevulnerabilityofafunc-tionalsystemisassessed.Oneexampleisthevulnerabilityof critical infrastructures, whereobserversareconfrontedwiththefact that physical installationsorcommunicationnetworksarespreadout inadistinctivemanner. Wendsystems, networksof systems, or internetworks(Edwardset al. 2007). Critical infrastructures encompass locally sitedwater supply systems (Moderl and Rauch 2011), regionallyimplemented power grids (Hines 2010), or globallyexpanding information and communication grids (Hell-strom2007). Fromamethodological point of view, it isvery difcult to distinguish sharp boundaries of infra-structuresystems, inwhichtechnical andsocial elementsare included and interact in a complex manner. Conse-quently,theanalyticalframeworkissomewhatdifferentineachandeverystudy.3.3.2 Example2:TemporalDimensioninVulnerabilitytoClimateChangeThe temporal dimension of assessments correlates with thesysteminfocus,butespeciallywiththedriverasystemisexposed to. Research into vulnerability and climate changeexemplies the needfor a distinctive temporal scope ofobservation. Research in this domain is driven by (at least)twotheses: (1) It iswidelyassumedthat climatechangeand the occurrence of extreme weather events correlate(for example, KunreutherandMichel-Kerjan2009). Asaconsequence, the scope of hazard and vulnerabilityassessment must includeshort-term, instantaneouseventsaswell aslong-termdevelopments. Scienticestimationsof signicant changes in the dynamics of the climatesystemareintherangeofdecadesandcenturies(Lentonetal.2008);(2)Anydesignandimplementationofactionplansmust alsoconsider distinctivetemporal horizonsinpreparingfor immediate threats or for the adaptationtolong-termclimate change as well as in responding tosudden weather events and using long-termmitigationstrategies(Fussel2007). Forexample, researcherscall formultiple perspectives when analyzing large urbanagglomerations: A resilient community is one thatmaintains a current information base to understandpotential hazards, andiswell informedinthepreparationandimplementationofitsfuturegrowthandimprovementplans(Prasadet al. 2009, p. 4).Fig.5 OntologyinthereferenceframeworkofthevulnerabilityassessmentIntJDisasterRiskSci 631 33.3.3 Example3:SocialDimensioninVulnerabilitySince the mid-1970s, research into vulnerability hasincluded the analysis of situations of vulnerable people andvulnerablegroups(OKeefeet al. 1976) andincreasinglyimplemented means of assessing social realities (Hewitt1983; Blaikieet al. 1994; Bohleet al. 1994; Adger1999,2006; Pelling 2003; Wisner et al. 2004). Blaikie et al.(1994) andWisneret al.(2004),for example, useda set ofvariables to distinguish root causes, dynamic pressures, andunsafeconditionsandgeneratedageneralizeddescriptionof being affected of individuals or social groups (asfamilies, households, neighborhoods, or groups as thepoorormigrants). Linkingvulnerabilityonthemicrolevel(individuals,households,groups)toprocessesanddistant root causes on the macro level has been an immenseimprovementinexplanatorypowerconcerningtheoverallcomplexity of hazardous situations, yet it is associated withmethodological challengesregardingthesocial dimensionof assessing vulnerability. When analyzing the vulnera-bilityof small, concrete social units as the level wherevulnerability is revealed, the analysis at the same timereferstothelevel ofthelarger, moreabstractsocialunitsand levels that help shape and propagate dynamic pressuresand root causes. These forces, in turn, determine the unsafeconditionsonthesmall social scale, suchasthegloballyoperating economy, the development of large urbanagglomerations, or the transformationof modernsocietydrivenbyfunctionaldifferentiation.Intheend, thescopeforempiricalresearchinvulnera-bilityassessmentsinmostcasesisrelatedtosmallerunits,likeindividuals, households, neighborhoods, andcommu-nities. Consequently, we used these levels also as sub-classesforthesocialdimensionintheontology.3.4 MethodologicalFrameworkThemethodologicalframeworkdomainoftheontologyissubdividedintotheoperationalapproachandtheunderly-ing theoretical concept implemented in vulnerabilityassessments.3.4.1 OperationalApproachTheontologyfor operational approachesusedinthevul-nerability assessment is characterized by the researchdesignclass: wedistinguishbetweenlongitudinal, cross-sectional case studies and assessments, which have a strongfocus on dening indicators that measure vulnerability.Since indicators are a key element in operationalizingvulnerabilityassessmentsandhaveastrongimpactonthevalidity of the assessment, a special class in the operationalapproach ontologyis dedicated toindicatorsand is usedtoprovideanoverviewof theactual indicatorsusedinaparticular vulnerability study. Sometimes, the choice ofindicatorsisrestrictedtosecondarydataprovidedbyof-cial statistics, whereas in other contexts researchersdevelopadhoc indicators. This domainof the ontologyshown in Fig. 6 provides an overviewof all capturedaspects of the operational approaches of vulnerabilityassessments. In addition to the research design andindicator classesalreadydescribed, thisincludesdatacollection anddataanalysis methods. Thedatacol-lection classdescribesthemethodsandsourcesusedtogather information about the vulnerability of a certain placeor system. Theassessments differ intechniques for datacollection, such as remote sensing (Eckert et al. 2011),mapping (Boruff et al. 2005; Collins et al. 2009), availablesocioeconomic data as input for multivariate statistics(Cutter et al. 2000, 2003), focus groups (Brooks et al.2005)orcontentanalysis(Turneretal.2003).Methodsofinquiry that focus on in-depth understanding of humanbehavior and its reasons are labeled as qualitative.Often,thesemethodsusenonstandardizedinstrumentsandrather ask why or how something happened instead ofwhere, when or what. Insome cases, concernedpeopleorstakeholdersparticipateinstepsoftheresearchprocess and the relationship between researchers andinterested parties is less or even non-hierarchical. Thedata analysis class describes various methods usedtoanalyzedatainthevariousvulnerabilityassessments.Thisincludes attributes, such as multivariate statistics (forexample, regressionanalysis, principal component orfac-tor analysis); content analysis; historical or policy analysis;uncertainty treatment; modeling and simulation; spatialanalysis; spatial or temporal mapping; andindexing. Forthelatter,differentapproachestoaggregatingindicatorstoanindexaredistinguished: (1)methodofweightingindi-cators (for example, statistical, expert opinion, multi-cri-teria decision analysis (MCDA)); (2) method ofaggregationofindicators(forexample, additive, multipli-cative, geometric); (3) selection of indicators that areincluded inindex; and(4) accuracy andvalidityof theapproach.3.4.2 TheoreticalApproachConcepts of vulnerability and the corresponding denitionsof vulnerabilityvaryacross researchdomains anddeter-mine the choice and design of research instruments. Hence,a discussion of an assessment should always take intoaccount the theoretical framework and the underlyingdenition of vulnerability. Each conceptual framework cancompriseamultitudeoffactorswhichdeterminevulnera-bility. Unfortunately, these frameworks are incompatiblewitheachother andnooverall frameworkexists. Fussel64 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3(2007)arguesthatterminologicalconfusionmainlyresultsfrom an unclear distinction between the dimensionssphereandknowledgedomainandproposesamini-mal structure to classify the multitude of approaches.Whereastherstdimensionspheredescribeswhetheravulnerability factor is considered as internal or external, thesecond dimension knowledge domain distinguishesbetween socioeconomic and biophysical factors, which cananddooverlap.Socioeconomicfactorsencompassaspectslike access to power and resources, social networks as wellas policies, internationalaid, or economic globalization. Incomparison, biophysical aspects of vulnerabilityrefer totopography, environmental conditions, landcover or haz-ards like earthquakes, stormor sea level rise. BasedonFusselsideas(Fussel2007),thevulnerabilityassessmentswereclassiedaccordingtotheirmainconceptuallineage(Fig.7): (1) Risk hazard approach (Burton et al. 1978;Kates 1985; Hewitt 1997; Fussel 2007); (2) Politicaleconomyapproach(AdgerandKelly1999;Pelling2003);(3) Pressure and release model (Blaikie et al. 1994;Wisner et al. 2004; Rauken and Kelman 2010); (4)Resilience approaches, such as the MCEERFrameworkfor quantifyingresilience (Bruneauet al. 2003) andtheBricModel of communityresilience(Cutter et al. 2010);and(5)Integratedapproaches, suchasCuttersHazardofPlace model (Cutter 1996), Turners VulnerabilityFramework(Turneretal. 2003),andtheBBCConceptualFramework (Birkmann 2006, based on Bogardi andBirkmann (2004) and Cardona (2001)). Integratedapproachesarenot ahomogeneousclass, but differ fromeach other in complexity and abstractness of the theo-retical concept, hazardconceptualization, andthedegreeto which they can be made operational. Regarding thedenition of vulnerability, the ontology distinguisheswhethervulnerabilityisdenedexplicitlyorimplicitlyinavulnerabilityassessment.Fig.6 OntologyfortheoperationalapproachIntJDisasterRiskSci 651 3It would be too space-consuming to list in this article allthevulnerabilityassessmentsencodedinVuWiki. Abriefgeneral overviewreferringtoalimitednumberofvulner-abilitystudiesusingthekeycomponentsof theontologydescribedinthissectionispresentedinTable1.4 ImplementationoftheOntology:VuWikiPlatformThe ontology outlined in the previous section allows for thestructuring of information on vulnerability assessments in awaythat theseconceptscanbequeriedor assessedviaastructuredandrationalmanner.Inthissectionwedescribehowthe ontology was applied and implemented in theplatformVuWiki.4.1 SemanticWikisCurrently, semantic wikis are among the most popularpractical application of ontologies (Buffa et al. 2008).Semanticwikisuseontologiesastheunderlyingmodeltoembed formalized knowledge, content, structures, and linksin the wiki pages via a special mark-up language (Krotzschetal. 2007). Incontrasttothis, conventionalwikisenabletheiruserstocollect andshareknowledgebystoringandretrievingindividual information, but arelessappropriateforobtainingaggregatedorqueriedinformationandtheircontent is often only weakly structured and not easilymachine-interpretable.Therisinginterest ofscienticcommunitiesandwork-inggroupsinthesemanticwebasanewlyemergingtypeofknowledgemanagementmaterializesinagrowingbodyof ontologies which are often publicly accessible (forexample,theSWEETontologybyRaskinandPan(2005)fororganizingthevastknowledgebaseinearthandenvi-ronmental sciences). For implementationof the vulnera-bilityontologyintoasemanticwiki, weusedtheSMWplatform,afree, open-sourceextensionofthewell-knownMediaWiki (whichisalsousedtorunWikipedia). SMWuses the stability and established usage patterns of theexistingMediaWikisystemfortheseamlessintegrationofsemantic technologies into a wiki (Volkel et al. 2006;Krotzsch et al. 2007). While many semantic wikis areunder development, SMW currently is the only one that hasbeen deployed in large-scale semantic wiki applicationsandisusedwidelyonpublicwebsites(Buffaetal. 2008).4.2 OntologyinVuWikiBeyond its functions as a content management system,VuWiki isttedwithaterminologicalbackboneinwhichthe tangible representation of a vulnerability study (a singlearticle/publication) is a unique wiki page specied bynumerous propertiesbasedonthevulnerabilityontology.Inthepractical implementationof theontologyintotheobject-oriented language of SMW, we used Categories as asimpleformof annotationfor classicationof eachvul-nerability assessment article into four main categoriesaccordingtothefourentryquestions: Category: Vulnera-bility Driver, Category: Vulnerable System, Category:ReferenceFramework, and Category:Assessment Method.Inturn, eachCategoryhasanumber of Properties, Sub-Fig.7 Ontologyforthetheoreticalapproach66 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3Table1Examplesofvulnerabilityassessmentstudiesorganizedbykeyelementsofthevulnerabilityassessmentsontology;simpliedpresentationonthelevelofallindividuals,selectedclasses,andselectedattributesVulnerabilityassessmentstudyVulnerabilityofwhat?Vulnerabilitytowhat?ReferenceframeworkAssessmentmethodsSpatialunitSocialscopeDatacollectionmethodDataanalysisTheoreticalapproachAdger(1999)PopulationingeneralTropicalcycloneCity/village/municipalityIndividuals,householdsSurvey,expertinterviews,secondarydataMultivariatestatisticsPoliticaleconomyapproachBirkmannetal.(2011)IndividualsEarthquake,storms,ood,drought,climatechangeCountry,globalIndividualsAuxiliarydata/ofcialstatisticsMultivariatestatistics,indexingIntegratedapproachBollinandHidajat(2006)Communities,economy,politicsEarthquake,volcaniceruption,landslide,tsunami,droughtCommunitiesIndexingRiskhazardapproachCarren oetal.(2007)Populationingeneral,communitiesEarthquakeNeighborhoodsAuxiliarydata/ofcialstatistics,GISdataMapping,indexingIntegratedapproachCutteretal.(2003)CommunitiesCounty,districtAuxiliarydata/ofcialstatisticsMultivariatestatistics,indexingIntegratedapproachFekete(2009)Populationingeneral(River)oodCountryCommunitiesSecondarydata/ofcialstatisticsMultivariatestatistics,indexingHahnetal.(2009)Communities,healthTropicalcyclone,drought,climatechangeHouseholdsSurvey,auxiliarydataIndexingKienbergeretal.(2009)Populationingeneral(River)oodGridSecondarydata,mappingModeling/simulation,spatialanalysis,indexingIntegratedapproachSchneiderbaueretal.(2013)Agriculture&forestryClimatechangeRegionInterviews,secondarydata,contentanalysisIndexingIntegratedWisner(2006)Population,ethnicgroupsEarthquake,ood,droughtGlobalCommunitiesContentanalysisContentanalysisIntJDisasterRiskSci 671 3Properties, andValues. For example, thecategoryvul-nerabilitydrivercanhavenaturaldriverasaproperty,geological driverasasubproperty, andearthquakes,volcaniceruptions, landslides or tsunamis asval-ues. In this way, the ontology in VuWiki is translated into 4keycategories, 77different properties andsubproperties,and6,089uniquevalues.4.3 ProvidingStructuredAccesstoandOrganizingKnowledgeinVuWikiTheSMWplatformprovides aversatileset of tools thatallows users to search, organize, tag, browse, evaluate, andsharethewikis vulnerabilityassessment content. Strongemphasis was put onembeddinguser-friendlyoptions toenrichthe accompanyingdatabaseof VuWiki, sincethisisconsideredanecessarypreconditiontocreateaviableandlivelyplatform. Besidestheopportunitytoadd, edit, andlink conventional texts, tables, illustrations or external datasources,severalfeaturesinVuWikiarehighlightedbelow,whichsupport easy, structuredaccess toexistingknowl-edge and help organize the knowledge according to specicqueriesofusersandaddnewknowledge.(1) Semantic Forms The semantic form (Fig.8) servesas the backbone of VuWiki. It guides the usersthrougha step-by-stepqueryformwithdrop-downmenusfordescribingaselectedvulnerabilityassess-ment. Thesemanticformthusembedstheontologyand automatically generates the semantic markup thatallows users to add new studies or edit and query anyof the existing vulnerability assessment methodsdescribed in the wiki articles. The semantic formallows for new knowledge to be added to the VuWikiby other researchers or practitioners, for example,newassessment studiesoradditional informationonassessments already stored in VuWiki. These newdatacanbeintegratedintoVuWiki inaformat thatcanreadilybeincorporatedinSMWandsupportedby tools such as Dynamic Tables or WikiDrilldown.(2) Dynamic Tables Users can sort and compare allvulnerability assessments in VuWiki throughdynamic tables, which can easily be customizedwithout any actual programming. For example, atablecanbegeneratedtoprovidedetails about thereference framework of all vulnerability studies in thewiki, showing the geographical areas, spatial unit,temporal scope, andsocial dimensionof assessmentof these studies as columns in a dynamic and sortabletable. Thesetables inparticular facilitateaccess byproviding a structured overviewof the knowledgedomainas representedbytheassessment studies inVuWiki.Fig.8 Screenshot of the semantic form based on the vulnerability ontology and implemented in VuWiki to guide user while adding a new study68 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3(3) Wiki DrilldownThewiki drilldownenablesuserstoformulateindividualqueriesbymenuoptionsthatprovide selections from main categories (for example,data, theoretical frameworks, or assessment stan-dards) throughahierarchyof properties, subproper-ties, andvalues.Forexample,a user canuse the wikidrilldownfeaturetoquicklylocateall vulnerabilityassessmentsthatemployanindicator-basedapproachandlter theinformationtoidentifywhichof theseuse expert judgment versus statistical methods forweighting indicators or how many use the Hazards ofPlacemodelastheirtheoreticalframework.5 DiscussionInthetwoprevioussectionswehavepresentedtheontol-ogyasaresult of aniterativediscussionprocessandthenal application in the VuWiki platform. Correspondingly,therstpartofthediscussionfocusesonthedevelopmentof theontologyandthesecondpart onpossibleapplica-tionsincivilprotectionanddisaster(risk)management.5.1 MethodologicalChallengesinDevelopingtheOntologyDuring the process of developing the ontology for vul-nerabilityassessmentsandimplementingit inasemanticwiki, anumber of challengeswerefacedintermsof theprinciples of ontology creation (Raskin and Pan 2005) usedfor guidance (see Sect. 2.2): natural language indepen-dence, application independence, orthogonality, scalability,and community involvement. The principle of naturallanguage independence emphasizes the representationofconcepts rather thanterms, slang,andtechnicaljargonandrequires the dening axioms to be logically consistent.However, whenconceptsareinherentlyfuzzyandambig-uousas in the case of vulnerabilitycoming to anagreementonhowtopresenttheminaformalstructureischallengingandnotmerelyaprocessoftranslationthatneeds keepingapart thinkinginformal structures andindisciplines, theories or concepts. It is also a generic part oftheprocessofknowledgegenerationinthedomainitself,which in turn challenges the principle of applicationindependence. Creatingavulnerabilityontologywill rep-resent thebiasesandinuencesoftheknowledgedomainrepresented by this community. For example, Gallop n(2006) demands a decision on including or excludingexposure into/from the concept of vulnerability, becauseit isconsequential for thecourseof theresearchandtheinterpretation of novel insights. If researchers includeexposure,thefocusofvulnerabilityanalysisshiftstowardsthe relationship between system and environment. Ifexposure is excluded, vulnerability becomes a propertysolely of the systemand exposure as a componentcontributingtothevulnerabilitybecomespart oftheana-lysisonly,if,andwhenthepotentiallyhazardouseventorprocessoccurs.Another principle in ontology design that addressescompound concepts is orthogonality, meaning that thecompound concepts should be decomposable into theircomponent parts and enable users to reuse them in differentcontexts. Theaimofthisprinciplewasthatthetermde-nitions developed in VuWiki are coherent and clear enoughin order to be reused without requiring others to create theirowndenitions. Yet, while the thrust of the discussionsamongtheauthorsdevelopingtheontologywastosatisfythis requirement, it was not always straightforward to breakdown compound concepts into their subcomponentsmeasured against the demands of a formal structuringprocessduetothelackofconceptual clarityinherent invulnerability studies. Thinking of vulnerability studiesreferring to the human systemor individuals as beingvulnerable,forexample,itisnotalwaysclearwhataspectof the individuals existence exactly is susceptible toexogenous inuence: thephysical (facingdamagetotheorganism) or psychological health (facing damage tomental integrity), integrationintosocial groups, families,neighborhoods (facingisolation), or inclusionintofunc-tional and/ororganizational spheresofsociety(facingthedanger of exclusionfromlegal rights, political participa-tionor economic transaction). Obviously, there is adif-ferenceinanalyzingindividualsasasolitaryentityor associal beings with characteristic roles and functions insociety. But decomposing a complex system, such as socialsystems,intodistinctivesubclassesisaprocessinherentlyfraught with conceptual difculties in the eld and can be amatter of debatethat cannot besolvedbydevelopinganontology.Since the body of knowledge in the eld of vulnerabilitygrows steadily and more and more research elds areintersectingindealingwithvulnerabilityandrelatedcon-cepts, suchas resilience, theontologyshouldbeabletokeeppacewiththisdevelopment. Thisiscapturedbytheprinciple of scalability, which refers to the fact thatontologiesshouldbeeasilyextendabletoenablespe-cializeddomains tobuilduponmore general ontologiesalreadygenerated(RaskinandPan2005, p. 1121). Thisprinciplewasincludedinkeepingalevelofopennessandawarenessof relatedconcepts (such asresilience), butalsointhedifferenteldsofvulnerabilityassessment(disasterrisk reduction, climate change, development studies) in thegroup discussions. Despite selecting key vulnerabilityassessments from as wide a range of disciplines as possible,the ontology proposed in VuWiki is shaped (1) by theIntJDisasterRiskSci 691 3selectionoftheinitial 55assessmentstodevelopandtestthe ontology, which has a focus on social vulnerabilityassessmentsinthecontext of disaster riskreduction; and(2) by our interpretation of the literature and how the initial45assessmentsandthe10testcaseswereusedfordevel-opingtheterminologicalstructureoftheontology.Tothisextent,likeinmosttaxonomicapproachesinvulnerabilityresearch or social science in general, we will alwaysencounter some difculties inclassifying all potentialobjects bya predenedset of categories andproperties.Duetotheprincipleofscalabilityandduetoourselectionand interpretation of the initial assessments, and also due tothe principles of natural language and application inde-pendence,the ontologydevelopedheremay not satisfytherepresentationofconceptsandtheirinterrelationsasusedinthestrict perspectiveor technical jargonof onedisci-pline. Nevertheless, weexpectthat thefourkeyquestionsthat formthebasicstructureof theontology(ontherstlevel)capturethecoredimensionsandapproachesinvul-nerabilityassessment regardless of particular disciplinarybackgrounds. At thesametime, weexpect that theontol-ogyprovidesaconceptualfoundationthatcanincorporatearangeofadditionaldimensionsandconceptsinvulnera-bility, which currently are not considered, and that theontologycanbeextendedratherexiblyinawaythatwillnotrequireatotalrevisionoftheexistingstructure.Inviewofthesechallenges,theprincipleofcommunityinvolvement (RaskinandPan2005) gains importance, iftheontologyisexpectedtorepresentthecommonstateofknowledgeintheeld.Additionalinputisneededthroughinvolvementandparticipationofthevulnerability researchand practice community in order to extend and populate theontologywithregardtoperspectives that were not con-sideredinthecurrentwork. InVuWiki, wepresentarstscalableversionof anontologytodescribevulnerabilityassessments, keepinginmindthat developingthevulner-ability ontology is an adaptive process that will continue togrowwith more input fromthe research and practicecommunity. Toencouragetheexchangeof ideas, severaldiscussionpageshavebeendedicatedinVuWiki, alsoforeditingandfurtherdeveloping ourontology.Butreectingthestate-of-the-artofthetoolsavailableforimplementingthe principle of community involvement to its full extent, itisalsorecognizedthat thereisaneedforbettertoolsforcollaborativeontologydevelopment andfor manipulationofontologiesingeneral(Buffaetal. 2008).5.2 PossibleApplicationforCivilProtectionandDisaster(Risk)ManagementAsuccessful implementation of disaster risk reductionoptionsandstrategiesdemandsappropriatemechanismstocommunicateandtransfer theoverall knowledgeonriskand its underlying drivers to the various stakeholdersinvolved in the decision-making process. Vulnerabilityassessments are the product of the state-of-the-art in scienceandintegratelargevolumesofdataandsophisticatedana-lysis. However, astheknowledgeandthevolumeof sci-entic works on vulnerability assessments multiplysteadily, it is becoming increasingly difcult for the practiceandsciencecommunitytokeeptrackofallthesedevelop-mentseffectivelyandtouseittowardsdisasterriskreduc-tion. Consequently, the main intention of makingvulnerabilityassessmentscomparablebyusingapracticaland structured access to the existing, complex, and growingknowledge eld is to foster exchange of knowledge, as wellas learning among disciplines. VuWiki can serve as aknowledge management tool for a broad research andpractice community and, in this capacity, contribute tointeractions between science and practice in terms ofknowledge transfer. VuWiki also comprises the potential tobridgetheimplementationgapbyservingasaninterac-tive platform that helps sort through and convey the relevantknowledgeforaspeciccontext sothat theknowledgeisused and put into practice. For example, a national authoritythat maywant todevelopnewguidelinesfor communityood risk management based on risk and vulnerabilityassessmentscanusethefeaturesinVuWiki suchaswikidrilldown (see Sect. 4.3) to get an overview of the relevantparameters for ood vulnerability at the community scale ofassessment and be able to discern which of the many studiesare the most applicable to their particular needs. It should benoted that the usefulness of VuWiki in this regard isdependent on community involvement criteria, and thedepthandextenttowhichthewikiispopulatedwithaddi-tional vulnerability assessments beyond those currentlyrepresented.Second,VuWikiisjustatoolthatprovidesaccesstoknowledgeinastructuredway.Duetocopyrightreasons, data licensing, and other legal limitations, it cannotprovidetheactualjournalarticlesitself.6 ConclusionInthisarticle, wepresentedthedevelopment ofanontol-ogyforvulnerabilityassessmentsanditsimplicationinacollaborative semantic wiki platform vuwiki.org. Theontology proposedinVuWiki was developed iterativelyandrevisedinnumeroussessionsbytheauthorsbasedonitst inclassifyingthevulnerabilitystudiesinthelit-erature, whichwere selectedtobe organizedbyit. Theontology was subsequently validated with a group ofexperts at KIT who were not involved in the creation of theontology and later evaluated by students who used VuWikiasalearningtool inaseminar offeredonSystemicVul-nerabilityAnalysisinthefallof2012atKIT.70 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3The aim of VuWiki development was not to synthesize anewoverarchingmodelofvulnerability,buttoaddresstherelations between the existing vulnerability studies throughdevelopinganontologythat canultimatelybeusedas acomparative reference system for vulnerability assess-ments. In the end, the ontology itself, as provided andimplemented in the semantic media wiki platform VuWiki,isnot just adatabasefor collectingvulnerabilitystudies,but also a heuristic for qualifying vulnerability studies. ThevulnerabilityontologythatformsthebackboneofVuWikihelpsidentifythedecisionsandassumptionsmadebytheresearchers concerning theories, concepts, and methodsusedintheir assessments. Inthissense, VuWiki supportsscholarsindeterminingtheexplicit andimplicit assump-tions made in a particular assessment with respect toacomparative reference system. VuWiki does this by askingthe authors (or the person annotating the study) to trace theconceptual lineage(s) of the study and distinguish keycomponents, suchas thevulnerabilitydrivers, vulnerablesystems, spatial scope, temporal scope, and target usersaddressed(ornotaddressed)inthestudyathand.VuWikiis onlineandavailable atwww.vuwiki.org, andwe inviteinterestedpractitioners, researchers, and studentstovisit,use,andenrichthesite.Witheasyaccessibilitytostructured knowledge and the chances associated withbeingasemanticandcollaborativewikiplatform,VuWikicancontributetothedisseminationanduseofexistingknowledgeaswell astothepromotionofdataidentica-tion, access, interoperability, and the sharing of key sourcesof information on vulnerability assessment methodsthroughout theworld. Duetothecollaborativewiki plat-form, VuWiki has thepotential tolinktogether experts,institutions, and programs that focus on vulnerabilityassessments, and it is intended that the platformanddatabase will be utilized by the research and practicecommunity. The sustainability of the VuWiki itselfdepends on its ability to create benet for its users, but alsoonits abilitytoincorporatenewknowledge. This meansVuWiki must be exible enough to adapt vulnerabilitystudies fromother contexts, for exampleclimatechangeadaptation, andexpandtheontologytomakeitusableforother vulnerability-relatedconcepts, suchas resiliencethereby testing the scalability of the ontology. Furthermore,VuWikiisbuiltonthepremiseofcollaborativeauthoring,whichreliesonsocial incentivesandcommunitybuildingto grow and mature as a useful tool for researchers. Thus, acentraltaskinthe next stageis todevelopanddisseminateapromotionstrategyintargetingpractitioners, academics,andother stakeholders interestedinvulnerabilityassess-ment to actively contribute and enrich the knowledge base.This will requirecommunity-building work, recognition ofactivecontributors, andintegrationoftoolsthat allowforcommunity-drivenontologydevelopment.Acknowledgments Wethankthetworeviewersfortheirthoroughassessment of thisarticleandhelpingusimproveit, aswell astheeditorsofthisspecialissue, AlexanderFeketeandGabrieleHufsch-midt. The major work in developing this wiki was funded by KIT as aStartupBudget programinthecompetenceeldEarthandEnviron-ment and supported by the Center for Disaster Management and RiskReductionTechnology(CEDIM). WethankMarkusGartnerfortheprogrammingof thesemanticwiki andGautamMondal aswell asNadia Anoumoufor populatingthe wiki withthe rst 55studies.Furthermore, we would like to thank James Daniell, Sandra Las-kowski, FriedemannWenzel, andall the students inthe SystemicVulnerability Analysis seminar for their feedback and contributions tothe development of the ontology and the wiki. Finally, we would liketo thank Christopher Burton for his valuable comments and insights inthedevelopmentofthisarticle.Open Access This article is distributed under the terms of theCreativeCommonsAttributionLicensewhichpermitsanyuse, dis-tribution, and reproduction in any medium, provided the originalauthor(s)andthesourcearecredited.ReferencesAdger, W.N. 1999. Social vulnerability to climate change andextremes in coastal Vietnam. World Development 27(2):249269.Adger, W.N. 2006. Vulnerability. Global Environmental Change16(3):5763.Adger, W.N., andP.M. Kelly. 1999. Social vulnerabilitytoclimatechange and the architecture of entitlements. Mitigation andAdaptationStrategiesforGlobalChange4(34):253266.Armas, I. 2008. Social vulnerability and seismic risk perception. Casestudy: The historic center of the Bucharest Municipality/Romania. NaturalHazards47(3):397410.Birkmann, J. (ed.). 2006. Measuring vulnerability to natural hazards.Tokyo:UnitedNationsUniversityPress.Birkmann, J., T. Welle, D. Krause, J. Wolfertz, D-C. Suarez, N.J. Setiadi. 2011. World risk index: Concepts and results. InWorldrisk report 2011, 1342. Berlin: Bundnis EntwicklungHilft(AllianceDevelopmentWorks) in cooperationwith UnitedNations University, Institute for Environment and HumanSecurity. http://www.ehs.unu.edu/le/get/9018.Birkmann, J., C. Bach, andM. Vollmer. 2012. Toolsfor resiliencebuildingandadaptivespatial governance. RaumforschungundRaumordnung70(4):293308.Birkmann,J.,O.D.Cardona,M.L.Carreno,A.H.Barbat,M.Pelling,S. Schneiderbauer, S. Kienberger, M. Keiler, D. Alexander, P.Zeil, and T. Welle. 2013. Framing vulnerability, risk and societalresponses: The MOVE framework. Natural Hazards 67(2):193211.Blaikie, P., T. Cannon, I. Davis, and B. Wisner. 1994. At risk: Naturalhazards, peoples vulnerability, and disasters. London:Routledge.Bogardi, J., andJ. Birkmann. 2004. Vulnerabilityassessment: Therst steptowards sustainable riskreduction. InDisasters andsociety:Fromhazardassessmenttoriskreduction, ed. D. Mal-zahn, andT. Plapp, 7582. Berlin:Logos.Bohle, H.G., T.E. Downing, andM.J. Watts. 1994. Climatechangeandsocial vulnerability: Towardasociologyandgeographyoffoodinsecurity. GlobalEnvironmentalChange4(1):3748.Bollin, C., and R. Hidajat. 2006. Community-based disaster riskindex:PilotimplementationinIndonesia. InMeasuringvulner-ability tonatural hazards, ed. J. Birkmann, 271289. Tokyo:UnitedNationsUniversityPress.IntJDisasterRiskSci 711 3Boruff, B.J., C.T. Emrich, and S.L. Cutter. 2005. Erosion hazardvulnerability of US coastal counties. Journal of CoastalResearch21(5):932942.Brooks,N.,W.N.Adger,andP.M.Kelly.2005.Thedeterminantsofvulnerability and adaptative capacity at the national level and theimplications for adaptation. Global Environmental Change15(2):151163.Bruneau, M., S.E. Chang, R.T. Eguchi, G.C. Lee, T.D. ORourke,A.M. Reinhorn, M. Shinozuka, K. Tierney, W.A. Wallace, andD.vonWinterfeldt. 2003. Aframeworktoquantitativelyassessandenhancetheseismicresilienceofcommunities. EarthquakeSpectra19(4):733752.Buffa, M., F. Gandon, G. Ereteo, P. Sander, andC. Faron. 2008.Sweetwiki:Asemanticwiki. WebSemantics6(1):8497.Burton, I., R.W. Kates, andG.F. While. 1978. Theenvironment ashazard.NewYork:OxfordUniversityPress.Calvalieri, F., P. Franchin, B. Khazai, and P. Gehl. 2012. Quantitativeassessment of buildingshabitabilitybasedonphysical damageandfunctional interactionwithinfrastructural systems. Earth-quake Engineering and Structural Dynamics 41(11): 15691589.Cardona, O.D. 2001. Holistic estimation of seismic risk using complexdynamic systems (Estimation Holistica del Riesgo SismicoUtilizandoSistemasDinamicosComplejos). Barcelona:Techni-calUniversityofCatalonia.(inSpanish).Cardona, O.D. 2003. The need for rethinking the concepts ofvulnerabilityandriskfromaholisticperspective: Anecessaryreviewandcriticismforeffectiveriskmanagement.InMappingvulnerability: Disasters, development and people, ed. G. Bankoff,G. Frerks, andD. Hilhorst, 3751. London:Earthscan.Carreno, M.-L., O.D. Cardona, and A.H. Barbat. 2007. Urban seismicrisk evaluation: Aholistic approach. Natural Hazards 40(1):137172.Collins, T.W., E.S. Grineski, andM.L.R. deAguilar. 2009. Vulner-abilitytoenvironmentalhazardsintheciudadJuarez(Mexico)ElPaso(USA)Metropolis:Amodelforspatialriskassessmentintransnationalcontext. AppliedGeography29(3):448461.Cutter, S.L. 1996. Vulnerabilitytoenvironmental hazards. ProgressinHumanGeography20(4):529539.Cutter, S.L., J.T. Mitchell, and M.S. Scott. 2000. Revealing thevulnerabilityofpeopleandplaces:AcasestudyofGeorgetownCounty, SouthCarolina. AnnalsoftheAssociationofAmericanGeographers90(4):713737.Cutter, S.L., B.B. Boruff, and W.L. Shirley. 2003. Social vulnerabilityto environmental hazards. Social Science Quarterly 84(2):242261.Cutter,S.L.,C.G.Burton,andC.T.Emrich.2010.Disasterresilienceindicators for benchmarking baseline conditions. Journal ofHomelandSecurityandEmergencyManagement 7(1): Article51.DeLange, H.J., S. Sala, M. Vighi, andJ.H. Faber. 2010. Ecologicalvulnerability in risk assessment: Areviewand perspectives.ScienceoftheTotalEnvironment408(18):38713879.Eakin, H., and A.M. Luers. 2006. Assessing the vulnerability ofsocial-environmental systems. Annual Reviewof EnvironmentandResources31:365394.Eckert, S., R. Jelinek, G. Zeug, andE. Krausmann. 2011. Remotesensing-basedassessment of tsunami vulnerabilityandriskinAlexandria, Egypt. AppliedGeography32(2):714723.Edwards, P.N., S.J. Jackson, G.C. Bowker, andC.P. Knobel. 2007.Understandinginfrastructure: Dynamics, tensions, anddesign.AnnArbor:DeepBlue.Eriksen, S., and J.A. Silva. 2009. The vulnerability context of asavannaareainmozambique: Householddrought copingstrat-egies and responses to economic change. Environmental Science&Policy12(1):3352.Fekete,A.2009.Validationofasocialvulnerabilityindexincontexttoriver-oodsinGermany.NaturalHazardsandEarthSystemsSciences9(2):393403.Few, R., and P.G. Tran. 2010. Climatic hazards, health risk andresponse in Vietnam: Case studies on social dimensions ofvulnerability. GlobalEnvironmentalChange20(3):529538.Fussel,H.-M.2007.Vulnerability:Agenerallyapplicableconceptualframeworkfor climatechangeresearch. Global EnvironmentalChange17(2):155167.Gallop n,G.C.2006.Linkagesbetweenvulnerability,resilience,andadaptive capacity. Global Environmental Change 16(3):293303.GAO(Government Accountability Ofce). 2011. Electricity gridmodernization: Progress being made on cybersecurity guide-lines, but keychallengesremaintobeaddressed. Washington:US Government Accountability Ofce. http://www.gao.gov/new.items/d11117.pdf.Giuliani,G.,N.Ray,S.Schwarzer,A.DeBono,P.Peduzzi,H.Dao,J.VanWoerden, R.Witt,M.Beniston,andA.Lehmann.2011.Sharing environmental data through GEOSS. InternationalJournalofGeospatialResearch2(1):117.Gruber, T.R. 1993. Atranslation approach to portable ontologies.KnowledgeAcquisition5(2):199220.Grunthal, G. 1998. EuropeanMacroseismic Scale1998(EMS-98).CahiersduCentreEuropeendeGeodynamiqueet deSeismol-ogie15,CentreEuropeendeGeodynamiqueetdeSeismologie,Luxembourg.Hahn, M.B., A.M. Riederer, andS.O. Foster. 2009. Thelivelihoodvulnerability index: Apragmatic approach to assessing risksfromclimatevariabilityandchange: AcasestudyinMozam-bique. GlobalEnvironmentalChange19(1):7488.Hellstrom, T. 2007. Critical infrastructure and systemic vulnerability:Towardsaplanningframework.SafetyScience45(3):415430.Hewitt, K. (ed.). 1983. Interpretations of calamity from the viewpointofhumanecology.Boston:AllenandUnwin.Hewitt, K. 1997. Regions of risk: Ageographical introduction todisasters.London:Longman.Hewitt, K., andI. Burton. 1971. Thehazardousness of aplace: Aregional ecologyof damagingevents. ResearchPublication6,Toronto:UniversityofTorontoPress.Hines,P.2010.Survivabilityandreciprocalaltruism:Twostrategiesforintelligentinfrastructurewithapplicationstopowergrids.InIntelligent infrastructures, ed. R.R. Negenborn, Z. Lukszo, andH. Hellendoorn. Dordrecht:Springer.Hufschmidt, G. 2011. A comparative analysis of several vulnerabilityconcepts. NaturalHazards58(2):621643.IPCC(Intergovernmental Panel onClimateChange). 2012. Specialreport: Managingtherisks of extremeevents anddisasters toadvanceclimatechangeadaptation(SREX). IntergovernmentalPanelonClimateChange. http://ipcc-wg2.gov/SREX/.Kates,R.W.1985.Theinteractionofclimateandsociety.InClimateimpact assessment 27, ed. R.W. Kates, J. Ausubel, and M.Berberian, 337. Chichester:Wiley.Kelly, P.M., and W.N. Adger. 2000. Theory and practice in assessingvulnerability to climate change and facilitating adaptation.ClimaticChange47(4):325352.Khazai, B., M. Merz, C. Schulz, andD. Borst. 2013. Anintegratedindicator framework for spatial assessment of industrial andsocial vulnerabilitytoindirect disasterlosses. Natural Hazards67(2):145167.Kienberger, S., S. Lang, and P. Zeil. 2009. Spatial vulnerability units:Expert-basedspatial modellingofsocio-economicvulnerabilityintheSalzachCatchment, Austria. Natural HazardsandEarthSystemSciences9(3):767778.Kroger, W., and E. Zio. 2011. Vulnerable systems. London: Springer.72 Khazaietal. AnOntology-BasedSemanticWikiforVulnerabilityAssessments1 3Krotzsch, M., D. Vrandecic, M. Volkel, H. Haller, andR. Studer.2007. Semantic wikipedia. Web Semantics Science, Services andAgentsontheWorldWideWeb5(4):251261.Kunreuther, H.C., E.O. Michel-Kerjan. 2009. At war with theweather. Managinglarge-scalerisksinaneweraof catastro-phes.Cambridge:MITPress.Lebel, L., E. Nikitina, V. Kotov, andJ. Manuta. 2006. Assessinginstitutionalisedcapacities andpractices toreducetherisksofood disaster. In Measuring vulnerability to natural hazards, ed.J.Birkmann,359379.Tokyo:UnitedNationsUniversityPress.Lenton, T.M., H. Held, E. Kriegler, J.W. Hall, W. Lucht, S.Rahmstorf, andH.J. Schellnhuber. 2008. Tippingelements inthe earths climate system. Proceedings of the National AcademyofSciencesoftheUnitedStatesofAmerica105(6):17861793.Li, Z., F. Nadim, H. Huang, M. Uzielli, and S. Lacasse. 2010.Quantitative vulnerability estimation for scenario-based land-slidehazards. Landslides7(2):125134.Luhmann, N. 1995. Social systems. Stanford: Stanford UniversityPress.Luhmann,N.1997.Globalizationorworldsociety:Howtoconceiveof modern society? International Reviewof Sociology 7(1):6779.Mainz, I., K. Weller, I. Paulsen, D. Mainz, J. Kohl, and A. vonHaeseler. 2008. Ontoverse: collaborativeontologyengineeringfor the life sciences. Information Wissenschaft und Praxis 59(2):9199.Mechler,R.,S.Hochrainer,J.Linnerooth-Bayer,andG.Pug.2006.Public sector nancial vulnerability to disasters: The IIASACATSIMmodel.InMeasuringvulnerabilitytonaturalhazards,ed. J. Birkmann, 380398. Tokyo: UnitedNations UniversityPress.Moderl,M.,andW.Rauch.2011.Spatialriskassessmentforcriticalnetwork infrastructure using sensitivity analysis. Frontiers ofEarthScience5(4):414420.NRC (National Research Council). 2006. Facing hazards anddisasters: Understanding human dimensions. Washington:NationalAcademiesPress.OBrien, K., R. Leichenko, U. Kelkar, H. Venema, G. Aandahl, H.Tompkins, A. Javed, S. Bhadwal, S. Barg, L. Nygaard, andJ.West. 2004. Mapping vulnerability to multiple stressors: Climatechange and globalization in India. Global EnvironmentalChange14(4):303313.OKeefe, P., K. Westgate, and B. Wisner. 1976. Taking thenaturalnessout of natural disasters. Nature 260(5552):566567.Patt, A.G., M. Tadross, P. Nussbaumer, K. Asante, M. Metzger, J.Rafael, A. Goujon, and G. Brundrit. 2010. Estimating least-developed countries vulnerability to climate-related extremeevents over the next 50years. Proceedings of the NationalAcademyof Sciences of theUnitedStates of America107(4):13331337.Pelling, M. 2003. Thevulnerabilityof cities: Natural disastersandsocialresilience.London:Earthscan.Prasad, N., F. Ranghieri, F. Shah, Z. Trohanis, E. Kessler, andR.Sinha. 2009. Climate resilient cities: A primer on reducingvulnerabilitiestodisasters.Washington:WorldBank.Ranci, C., M. Migliavacca. 2010. Social vulnerability: Amultidi-mensional analysis. InSocial vulnerabilityinEurope. Thenewconguration of social risks, ed. C. Ranci, 219249. Basing-stoke:PalgraveMacmillan.Raskin, R.G., andM.J. Pan. 2005. Knowledgerepresentationinthesemantic web for earth and environmental terminology(SWEET). Computers&Geosciences39(9):11191125.Rauken, T., and I. Kelman. 2010. River ood vulnerability in Norwaythroughthepressureandreleasemodel. Journal of FloodRiskManagement3(4):314322.Schneiderbauer, S., L. Pedoth, D. Zhang, and M. Zebisch. 2013.Assessing adaptive capacity within regional climate changevulnerability studiesan Alpine example. Natural Hazards67(3):10591073.Smithson, M. 1993. Ignoranceandsciencedilemmas, perspectives,andprospects. ScienceCommunication15(2):133156.Thomalla, F., T. Downing, E. Spanger-Siegfried, G. Han, and J.Rockstrom. 2006. Reducing hazard vulnerability: Towards acommonapproachbetweendisaster riskreductionandclimateadaptation. Disasters30(1):3948.Turner, B.L., P.A. Matson, J.J. McCarthy, R.W. Corell, L. Christen-sen, N. Eckley, G.K. Hovelsrud-Broda, J.X. Kasperson, R.E.Kasperson, etal. 2003. Illustratingthecoupledhuman-environ-ment systemfor vulnerability analysis: Three case studies.ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica100(14):80808085.UNISDR(UnitedNationsInternationalStrategyforDisasterReduc-tion). 2005. Hyogoframeworkforaction20052015: Buildingtheresilienceof nations andcommunities todisasters. Unitednations. http://www.unisdr.org/we/inform/publications/1037.Volkel, M., M. Krotzsch, D. Vrandecic, H. Haller, andR. Studer.2006. SemanticWikipedia. InProceedingsof the15thinterna-tional conferenceonWorldWideWeb, 585594, ACM(Asso-ciationforComputingMachinery), Edinburgh.vonBertalanffy, L. 1950. Anoutlineofgeneral systemtheory. TheBritishJournalforthePhilosophyofScience1(2):134165.Welle, T., J. Birkmann, J. Rhyner, M. Witting, andJ. Wolfer. 2012.Worldriskindex2012:Concept,updatingandresults.InWorldrisk report 2012, 1126. Berlin: Bundnis Entwicklung Hilft(Alliance Development Works). http://www.worldriskreport.com/.Wisner, B. 2006. Self-assessment of copingcapacity: Participatory,proactive, andqualitativeengagement of communities intheirown risk management. In Measuring vulnerability to naturalhazards, ed. J. Birkmann, 316328. Tokyo: United NationsUniversityPress.Wisner, B., P. Blaikie, T. Cannon, and I. Davis. 2004. At risk: Naturalhazards, peoplesvulnerabilityanddisasters, 2nded. London:Routledge.Wu, S.Y., B. Yarnal, andA. Fisher. 2002. Vulnerabilityof coastalcommunities to sea-level rise: A case study of Cape MayCounty, NewJersey, USA.ClimateResearch22(3):255270.Zehetmair, S. 2012. Societal aspects of vulnerability to naturalhazards. RaumforschungundRaumordnung70(4):273284.IntJDisasterRiskSci 731 3