risk city_de multi hazard risk assessment

8/12/2019 Risk City_DE Multi Hazard Risk Assessment

1/4

RiskCity: a training package on the use of GIS for urbanmulti-hazard risk assessment.Cees van Westen (UNU-ITC DGIM, The Netherlands)

Abstract. As part of the capacity building activities of theUnited Nations University ITC School on DisasterGeo-Information Management (UNU-ITC DGIM) theInternational Institute for Geoinformation Science and EarthObservation (ITC) has developed a training package on theapplication of GIS for multi-hazard risk assessment. The

package, called RiskCity comprises a complete suite ofexercise descriptions, together with GIS data and presentationmaterials on the various steps required to collect and analyzerelevant spatial data for hazard, vulnerability and riskassessment in an urban environment. The exercises deal withfour different types of hazards: earthquakes, flooding,technological hazards, and landslides. The exercises deal witha hypothetical case study, hence the name RiskCity. Most ofthe data was derived from the city of Tegucigalpa inHonduras, but has been adapted to make the exercises moredidactical. They aim to give an understanding of the basicconcepts involved in risk assessment, and allow the studentsto have an idea on how GIS can be used for analyzing thedifferent types of hazards, creating an elements at riskdatabase, assessing vulnerability, making loss estimationsusing qualitative and quantitative methods, carry out a cost

benefit analysis, and use the data for urban planning anddisaster risk management. The package has been developed incollaboration with several partner organizations in differentcontinents, and is used as the basis for a series of courses.Currently it is developed into a distance education course.

Keywords. Risk assessment, capacity building, landslide risk,

training package, Open source software.

1. Introduction

One of the important components of disaster riskmanagement is capacity building and training of disastermanagement experts and professionals working in manydifferent disciplines that have an important disaster reductioncomponent, such as planners, engineers, architects etc. TheHyogo framework of action 2005-2015 of the UN-ISDRindicates risk assessment and education as two of the keyareas for the development of action in the coming years,

Worldwide a number of organizations are specialized inproviding short training courses on disaster risk managementrelated issues (ADPC, 2005). A number of organizationshave also prepared training materials that are accessible

through the internet, for example the Disaster ManagementTraining Programme (DMTP), or the International Federationof Red Cross and Red Crescent Societies (IFRC). Most ofthese however are concentrating on community-basedmethods. Disaster risk management courses at BSc or MSclevel are now available in many Universities in all continents.

Relatively few training materials are available onmulti-hazard risk assessment. Good textbooks on the subjectare still not available. Online training materials can beobtained for example from the websites of FEMA (2008) andEMA (2008). The development of innovative forms of

learning and teaching oriented towards building new curriculain the field of natural risk has attracted attention in Europeaninitiatives such as DEBRIS (2006) and NAHRIS (2006).

As far as GIS-related material related to multi-hazard riskassessment is concerned, the HAZUS methodology developedin the US can be considered the standard. This comprehensiveloss estimation software which runs under ARCGIS is a verygood tool for carrying out loss estimations for earthquakes,flooding and windstorms (FEMA, 2008b), but is restricted touse in the USA because of data constraints. Courses in the useof HAZUS can be followed online from the ESRI VirtualCampus (ESRI, 2008). However, complete GIS basedtraining packages on spatial hazard and risk assessment usinglow-cost or free GIS software are still very scarce, to theknowledge of the author. One example is a training packagein English and Spanish developed for Central America in theframework of the UNESCO RAPCA project (ITC, 2004)

This paper describes the main aspects of a GIS-basedtraining package on multi-hazard risk assessment, which has

been developed by the UNU-ITC School for DisasterGeoinformation Management.

2. United Nations University ITC collaboration

The International Institute for Geo-Information Scienceand Earth Observation (ITC) is an institute for postgraduatetraining and research in the field of geoinformation directedto capacity building and institutional development of

professional and academic organizations from developingcountries. In 2005, ITC and the United Nations University

have established a collaborative programme on the use ofspatial information for disaster management, which resultedin the formation of the UNU-ITC School for DisasterGeo-Information Management. The main activities of theDGIM School focus on training, education, curriculumdevelopment, knowledge development and researchcollaboration. This is done through the establishment ofUniversity networks in Asia, Africa and Latin America, ofwhich the member Universities exchange spatial information,course materials and jointly carry out training and research

projects.The DGIM School develops training packages and

courses that are given jointly with the partners of thenetworks in various countries. The materials are uniform, andhave been developed in different languages, and the support is

given by local University staff who have followed earliertraining, and by staff from the UNU-ITC DGIM School.

One of these courses is on Multi-hazard risk assessment,which is centered around a case study on the use ofGeographic Information Systems, and Remote Sensing for theassessment of hazard, vulnerability and risk in a typical urbanarea representative of situations in many developing countries.Rapid urbanization combined with a lack of planning oftenleads to the spreading of squatter areas located in hazardousareas, such as steep slopes, flood prone areas etc.


2/4

3. RiskCity training package

The package, called RiskCity comprises a complete suiteof exercise descriptions, together with GIS data and

presentation materials on the various steps required to collectand analyze relevant spatial data for hazard, vulnerability andrisk assessment in an urban environment.

The exercises deal with a hypothetical case study, hence

the name RiskCity. The exercises are based on a case studyfrom Tegucigalpa in Honduras. Tegucigalpa suffered severedamage from landslides and flooding during Hurricane Mitchin October 1998 when the city received 281 mm of rain in 3days (Mastin and Olsen, 2002). Due to river flooding, an oldlandslide was reactivated and an entire neighborhood on topof it was destroyed. The landslide caused the damming of theriver and resulted in severe flooding in large parts of the citycenter for several weeks (Harp et al., 2002). These events arewell identifiable on the high resolution image which serves asthe basis for the exercises (See figure 1). After HurricaneMitch USGS and JICA carried out extensive work inHonduras and produced extensive datasets.

Fig. 1 High-resolution image of the center of RiskCitywith some of the hazard and vulnerability featuresindicated.

Only part of the exercises is based on the actual situationin Tegucigalpa. In order to be able to reach the learningobjectives, modifications and additions were made to theoriginal data. It is very difficult to have a dataset for a

particular area where all aspects of multi-hazard riskassessment can be properly demonstrated. Either because

particular hazard types do not happen in the city or becauseparticular data sets are incomplete, restricted or erroneous.

One important consideration in designing the exercises isthat people from developing countries should not be restricted

in using it due to financial burdens for software acquisition.Therefore the aim was to use Open Source software as a basis.The exercises are written for the ILWIS software. ILWIS isan acronym for the Integrated Land and Water InformationSystem. It is a Geographic Information System (GIS) withintegrated image processing capabilities. It also has its ownattribute data analysis, spatial data entry and conversion

modules. New modules have been recently added for SpatialMulti Criteria Evaluation, analysis of Digital ElevationModels and for digital stereo image interpretation. Thestrongest point of the software is the map calculation modulethat allows extensive modeling with raster maps, also usingscripts. The software has extensive help functions anddocumentation, and can be downloaded from the followingweb-site: http://52north.org/ilwis. ILWIS is very user-friendly,and allows the participants to concentrate on the riskassessment application rather than on the specifics of thesoftware.

Each of the exercises in RiskCity has its own dataset. Thedata are all provided in separate directories, including theresults of the previous exercises that are needed to make asubsequent one. So it is possible to carry out each of the

exercises separately. Also result files and PowerPointpresentations with instructions are included.

4. Structure of RiskCity

The overall structure of the RiskCity training package isgiven in figure 2, and an overview of the various componentsis presented in table 1. Four different types of hazards areevaluated: landslides, floods, earthquakes and technologicalhazards. The training package starts with introductoryexercises dealing with the software and with the study area,where students learn the various hazard problems byevaluating high resolution images. An important componentof RiskCity is the generation of a database of elements at riskin order to evaluate the vulnerability of buildings and

population. Here two options are considered. The first is that

there is no spatial data available, except for a high resolutionimage, and the students have to generate mapping units withhomogeneous types of buildings by stereo interpretation anddigitizing on the image. Sampling is then carried out to definethe number of buildings and population per mapping unit.The second option for generating the elements at riskdatabase assumes there are digital data available in the formof building footprint maps, census information and detailedelevation data from a Lidar survey. These are used tocalculate the number of buildings per mapping unit andlanduse type, and to characterize the buildings, for instance

by calculating their height and floorspace using Lidar data.The floorspace is used then to distribute the census populationover the mapping units, and population estimates are made forday- and nighttime scenarios. The elements at risk database

contains information on the buildings, with importantattributes such as urban land use type, construction type,floorspace and height, as well as on population for a daytimeand nighttime scenario.

There are a wide range of hazard assessment exercises,not only the ones dealing with the four types of hazardmentioned before, which are based on the data of RiskCity,

but also other ones which use data from other areas fortsunami, cyclone, volcanic, forest fires and land degradationhazard assessment. They use a variety of approaches, such as


3/4

inventory based, heuristic, statistical and deterministic ones.The vulnerability assessment includes exercises on the

use of vulnerability curves for assessing physicalvulnerability, as well as the use of expert based Spatial MultiCriteria Evaluation for the evaluation of social vulnerabilityand capacity.

Table 1 Overview of the exercise structure of RiskCity.Component ExerciseIntroduction Introducing the ILWIS software

Introducing the study area and the mainproblems based on a high resolution imageand main spatial data .

Generating anElements at r iskdatabase

Focusing on buildings and population Twodifferent approaches:- when no data is available: use of high

resolution images & screen digitizing- when data is available: use of census

data, building footprint maps and LiDARdata.

Hazardassessment

Selection of a range of exercises dealing withlandslides, earthquakes, floods, technologicalhazard, volcanic hazard, coastal hazards,tsunami etc.

Vulnerabilityassessment

Application of vulnerability curves andmatrices for physical damage assessment

Loss estimation:qualitativemethods

Application of risk matrices, combiningsusceptibility and vulnerability; use of SpatialMulti Criteria Evaluation for vulnerability andcapacity assessment including a range ofindicators.

Loss estimation:quantitativemethods

Annual loss estimation using risk curves, forearthquakes, landslides, floods andtechnological hazards

Cost-benefitanalysis fordisaster reductionmeasures

Converting of building losses in monetaryvalues for different hazard types; selection ofpossible risk reduction measures; cost-benefitanalysis to select the optimal measures.

Using riskinformation inurban planning

Spatial Multi Criteria Evaluation in which riskinformation is combined with other indicatorfor the planning of new neighbourhoods andinfrastructure

Using riskinformation inemergencypreparedness

Modelling of potential sites for evacuationcentres, medical support, and emergencycentres. Planning of damage assessmentcampaign.

Fig. 2 General structure of the RiskCity training package.See text for explanation

The loss estimation is done using the formula:Risk = Hazard * Vulnerability * Amount, in which the variouscomponents analyzed in the previous exercises are combined,and risk curves are generated, plotting annual probabilityagainst expected losses.

The risk curves form the basis of subsequent cost-benefitanalysis, in which for each hazard type, a number of risk

reduction measures is evaluated. The investments toimplement certain measures (e.g. relocation of houses, floodcontrol) are estimated and compared to the reduction inannual losses that would result if they are implemented.Based on this the most appropriate methods for risk reductionare selected.

The last part of the RiskCity exercises deal with a finalproject in which the participants are given a particularproblem they have to solve with the risk information obtainedearlier.

5. Example: landslide component of RiskCity

Part of the RiskCity training package deals with landslidesusceptibility, hazard and risk assessment. These componentsare also used for giving a separate course on the use of GIS

for landslide risk assessment. They are based on methodspresented by Soeters and Van Westen (1996) and Castellanos(2008). The landslide related exercises deal with thefollowing tasks: Downloading of a high resolution image from Google

Earth, georeferencing it, and converting it into a digitalstereo image through the combination with a detailedDEM derived from a Lidar survey;

Use of stereo-image interpretation, using the anaglyphimage generated and a number of other ones for older

periods to digitally map landslides from different times aswell as elements at risk;

Generating a landslide inventory map, with attributeinformation related to type, age, volume and activity,which is used in the susceptibility assessment;

Generating a basic landslide susceptibility map, using asimple bivariate statistical method, with the landslideinventory, slope, lithology, landuse and other relevantfactor maps. Use of success rate curves to validate thesusceptibility map;

Application of a basic deterministic method, using theinfinite slope method, to calculate the factor of safetyunder different groundwater scenarios;

The conversion from susceptibility to hazard bymultiplication of event probability, based on the successrates, spatial probability, based on the landslide density inthe susceptibility classes, and temporal probability, basedon estimated return periods.

Combination of the elements at risk database with thelandslide susceptibility map to estimate the number of

buildings falling in each of the susceptibility classes; Generation of a qualitative landslide risk map using a

matrix that combines vulnerability classes withsusceptibility classes;

Generation of a quantitative landslide risk map, bymultiplying for each susceptibility class, the spatial andtemporal probability of landslide occurrence, thevulnerability and the number of buildings.

Plotting a risk curve, and analyzing through cost-benefitanalysis the most suitable options for risk reduction;


4/4

Conclusions

The RiskCity training package is intended as a tool todemonstrate the utility and requirements of spatial data inurban multi-hazard risk assessment. Since the preparation ofsuch a training package takes quite some time, it is normallynot possible to adapt the dataset easily to local conditionseach time a course is given in another place. This is also one

of the reasons why the exercises have been made as genericas possible, by excluding most of the references to the actualcity where the dataset is obtained.

The RiskCity training package is constantly beingupdated and further improved. The plan is to incorporatemore Participatory GIS approaches in the training package, aswell as to include more hazard modeling, using Open Sourcesoftware. There is also a plan to make a separate versionwhich is focusing on risk occurring in rural areas taking intoaccount flooding, forest fires, drought and land degration asthe main types of hazards.

Discussions are ongoing to make the text of the RiskCitytraining package available in time on the internet page of aninternational UN organization. The training package is usedregularly in courses and is available in English, with major

parts also in Spanish and Chinese (see figure 5). Currently itis developed into a distance education course.

Fig. 3 Cover pages of the landslide risk part of the guidefor the RiskCity training package, available in English,Spanish and Chinese.

Acknowledgments

We would like to thank Gonzalo Funes from Hondurasfor providing the initial data sets. The Digital Surface Modeland flood information was obtained from a study by theUnited States Geological Survey. The high resolution imagewas obtained from a project funded by JICA. Ruben VargasFranco, Dinand Alkema, Lorena Montoya, Michiel Damen,

Nanette Kingma, Antonio Naverette, Jean Pascal Iannacone,Manzul Hazarika and Norman Kerle are thanked for theircontributions on various aspects of this case study.Colleagues from ADPC, AIT, CDUT, ICIMOD, UNAM,CLAS, IIRS and UGM are thanked for the friendlycollaboration and the testing of the training package invarious training courses. The following persons are thanked

for the translation in Spanish: Carlos Saavedra, Jose AntonioNavarrete, Ruben Vargas, Edward Gonzalez, Estuardo Lira,and Manolo Barillas. Fan Xuanmei is thanked for translating

part of it in Chinese. This work is part of the United NationsUniversity ITC School for Disaster GeoInformationManagement (www.itc.nl/unu/dgim)

References

ADPC 2005. Knowledge Development, education, publicawareness training and information sharing. A Primer ofDisaster Risk Management in Asia. Asian DisasterPreparedness Center. URL: http://www.adpc.net

Castellanos Abella, E.A., Multiscale landslide risk assessmentin Cuba, Utrecht, Utrecht University, 2008. ITCDissertation 154, 273 p. ISBN: 978-90-6164-268-8

DEBRIS 2006. Innovative education for risk management.Development of innovative forms of learning andteaching oriented towards building a family of newcurricula in the field of natural risk EU Leonardo da Vinci

programme. URL: http://www.e-debris.netEMA, 2008. Disaster loss assessment guidelines. Emergency

Management Australia. URL:http://www.ema.gov.au/agd/EMA/rwpattach.nsf/VAP/(383B7EDC29CDE21FBA276BBBCE12CDC0)~Manual+27A.pdf/$file/Manual+27A.pdf

ESRI 2007. ESRI Virtual Campus Courses. HAZUS-MH(Multi-Hazards) for Decision Makers. URL:http://training.esri.com/Courses/

FEMA, 2008a. Mitigation Planning :How To Guides.Federal emergency Management Agency. URL:http://www.fema.gov/plan/mitplanning/planning_resources.shtm

FEMA, 2008b. Hazus, FEMAs software for estimatingpotential losses from disasters. Federal EmergencyManagement Agency URL:http://www.fema.gov/plan/prevent/hazus/

Harp, E.L., Castaneda, M., Held, M.D. 2002. Landslidestriggered by Hurricane Mitch in Tegucigalpa, Honduras.USGS Open-File Report 02-0033.http://pubs.usgs.gov/of/2002/ofr-02-0033/

ITC 2004. Mdulo de capacitacin. Aplicacin de Sistemasde Informacin Geogrfica y Sensores Remotos para elAnlisis de Amenazas, Vulnerabilidad y Riesgo. Spanishand English version. International Institute forGeoinformation Science and Earth Observation (ITC).URL: http://www.itc.nl/external/unesco-rapca/start.html

Mastin, M.C., and Olsen, T.D., 2002, Fifty-yearflood-inundation maps for Tegucigalpa, Honduras: U.S.Geological Survey Open-File Report 02-261. URL:http://pubs.usgs.gov/of/2002/ofr02261/

NAHRIS 2006. Dealing with NAtural Hazards and RISks.

Swiss Virtual Campus Project. URL:http://www.nahris.ch/

Soeters, R. and van Westen, C.J., 1996. Slope Instability.Recognition, analysis and zonation. In: A.K. Turner andR.L. Schuster (Editors), Landslide: Investigations andMitigation. Special Report 247. Transportation ResearchBoard. National Research Council. National AcademyPress., Washington, D.C, pp. 129-177.

risk city_de multi hazard risk assessment

Documents