Socio-Economic Vulnerability and Losses

Transcription

1 Systemic Seismic Vulnerability and Risk Analysis for Buildings, Lifeline Networks and Infrastructures Safety Gain Socio-Economic Vulnerability and Losses B. Khazai, J.E. Daniell, F. Wenzel Karlsruhe Institute of Technology, Germany Sebnem Duzgun, Kivanc Ertugay Middle East Technical University, Turkey B.V. Vangelsten Norwegian Geotechnical Institute, Norway I. Iervolino & S. Esposito University of Naples Federico II, Italy SYNER-G Final Workshop, Milano, March

2 SYNER-G Socio-economic Losses (WP4) Shelter and Health Short-term Emergency Relief and Recovery Urban Scale SYNER-G Final Workshop, Milano, March

3 Physical Risk Parameters / Inputs Inputs and Outputs of WP4 Socio-Economic Loss Parameters / Outputs Casualty Model Building Habitability Model Accessibility Model Shelter Needs Model Health Impact Model Building Occupancy Building Damage Utility Loss Road Block Possibility Healthcare Functionality Data Harmonization Calibration/Validation SYNER-G Final Workshop, Milano, March

4 Post-earthquake Shelter Placement Disaster displacement is a complex, dynamic process Timeline of displaced popl. after Tohoku Coupled interactions between environmental, physical and socio-economic processes State-of-the-Art Earthquake Loss Estimation (ELE): linear relationship of building damage Khazai et al., 2011 Effects of a super-aging Society ; 3 under 10; ; ; ; ; ; or older; ; or older under 10 SYNER-G Final Workshop, Milano, March 2013

5 Relationship between Building Damage and Displacement after Earthquakes Relationship between severely damaged and destroyed buildings and displaced persons after Relationship earthquakes between (n = severely 457 earthquakes damaged from and destroyed , buildings Source and CATDAT displaced 2012) persons after earthquakes (n = 457 earthquakes from , Source CATDAT 2012) SYNER-G Final Workshop, Milano, March 2013

6 Shelter Needs Decision Model Logic Model to simulate Decision-making process for Dislocation and Seeking Shelter SYNER-G Final Workshop, Milano, March

7 Selection of Indicators Decision Model Literature Review Multivariate Statistical Analysis EU Urban Audit Database SYNER-G Final Workshop, Milano, March

8 Selection of Indicators Example from Urban Audit data at sub-city level. PCA gives information on - inter-correlation structure between variables - which variables represent the most variation in the data set Results: - Close to 75% of variation in data represented by 8 dimensions - Identify key indicators for creating socioeconomic profiles of European urban areas No Subjective descriptor for principal component 1 Mortality/Age 2 Education 3 Lone Parent with Children 4 Population Density 5 Migration/Ethnicity 6 Gender Strongest correlated indicator Strongest correlation value SA2016I: Mortality rate for <65 per year TE2028I: Prop. of working age population qualified at level 3 or 4 ISCED DE3005I: Prop. of households that are lone-parent households EN5101I: Population density: total resident pop. per square km DE2006I: Proportion of Residents who are not EU Nationals and citizens of a country with a medium or low HDI DE1003I: Proportion of females to males in total population Unemployment EC1020I: Unemployment rate Sub-standard Housing SA1018I: Proportion of dwellings lacking basic amenities SYNER-G Final Workshop, Milano, March 2013

9 Multi-criteria Decision Analysis (MCDA) Decision Factors Household Tenure (Owner vs. Renter) Housing Type (Single, Multi-family) Household Type (Large Families with Children, Single Parents) Age (Children and Elderly) Perceived Security Urban Audit Indicators for Desirability to Evacuate -Prop. of households living in private rented housing -Proportion of households living in owned dwellings -Number of houses per 100 apartments -Proportion of households living in social housing -Proportion of dwellings lacking basic amenities -Proportion of non-conventional dwellings -Avg. Size of households -Lone-parent households with children aged 18 or under -Proportion of households living in social housing -Proportion of total population aged 0-4 -Proportion of total population aged 75 and over -Total Number of Recorded Crimes per 1000 population Decision Factors Income Unemployment Migration/ Ethnicity Education Urban Audit Indicators for Shelter Seeking Index -Percent of households with less than 60% of national median annual disposable income -Proportion of households reliant upon social security -Unemployment rate -Number of residents born abroad (not only nationals) -Residents who are not EU Nationals and citizens of a country with a medium or low HDI -Prop. of working age population qualified at level 1, 2, 3 4, 5 and 6 ISCED SYNER-G Final Workshop, Milano, March

10 Uninhabitable Buildings Index (UBI) n = 305, Pettino Usability Damage Level Ratio D0 D1 D2 D3 D4 D5 FU Fully Usable 0,87 0,69 0,17 0,02 0,00 0,01 PU Partially Usable 0,13 0,31 0,45 0,21 0,04 0,01 NU Non Usable 0,00 0,00 0,38 0,77 0,96 0,98 Utility Loss in each system j defined as one minus the ratio of satisfied to required demand Utility Loss Tolerance Thresholds UL T Weather Conditions Good Bad FU Fully Usable PU Partially Usable Displaced Persons (UB): SYNER-G Final Workshop, Milano, March

11 Displaced Persons Index (DPI) DPI = UBI ( 1 + LRE) Lack of Resistance to Evacuation (LRE) UBI = Ratio of persons in Uninhabitable Buildings Full evacuation of urban core or city center (Urban Audit Boundaries) Full evacuation of neighborhoods with damage levels above user-defined thresholds Full evacuation of geographic areas based on a user-defined evacuation area SYNER-G Final Workshop, Milano, March 2013

12 Shelter Needs Index (SNI) push factors - factors that account for reasons to seek public shelter pull factors - factors that account for the preference of seeking public shelter SYNER-G Final Workshop, Milano, March 2013

13 MCDA Indicator Toolkit The toolkit is based on an open source multi-criteria decision analysis tool developed at IKET (KIT) by Susan Vaziri and Tim Mueller IndicatorToolkit allows users to combine their local knowledge of shelter demand with socio-economic databases, to develop their own Shelter Needs Index SYNER-G Final Workshop, Milano, March 2013

14 MCDA Indicator Toolkit Key Features SYNER-G Final Workshop, Milano, March 2013

15 Implementation and Benchmarking L Aquila Mixed Operations Centers (COM): Centers with overall coordinating role in its own territory for all rescue operations. AEDES Survey (Building Damage, Usability and Evacuation Data) Italian Civil Defense Shelter Site Data (156 Shelter sites, April Nov. 2009) Socio-Economic Data for Municipalities and Fractions (104 Fractions; 131 Variables reduced to 27 indicators) Shelter population evolution in each COM SYNER-G Final Workshop, Milano, March

16 Displaced Persons Index (DPI) SYNER-G Final Workshop, Milano, March

17 Shelter Seeking Index (SSI) SYNER-G Final Workshop, Milano, March

18 Health Impact Model Casualty Model No. Dead / Injured SYNER-G Final Workshop, Milano, March

19 Semi-Empirical Casualty Model Methodology for Semi-empirical Casualty Ratios: 1. Building stock of any region is grouped in terms of its distribution in Building- Casualty super classes at the event year 2. A database for each historic earthquake event is constructed, where: Building Damage State is allocated to each Building Class Casualty Rate is allocated to each Building Class Seismic Intensity is allocated to each of the Building Class Building Occupancy is allocated to each Building Class 3. Using above parameters, #dead is simulated using published C.R. values initally 4. These values are optimized itteratively to achieve best fit between simulated and recorded casualties Intensity 6 D0 D1 D2 D3 D4 D5 1-BC BC BC Intensity 7 D0 D1 D2 D3 D4 D5 1-BC BC BC Intensity 8 D0 D1 D2 D3 D4 D5 1-BC BC BC Intensity 9 D0 D1 D2 D3 D4 D5 1-BC BC BC SYNER-G Final Workshop, Milano, March 2013

20 Deaths Casualty Estimation where: t = building-casualty type (t = 1-BC. 2-BC. 3-BC) d = damage level (d = D0, D1, D2, D3, D4, D5) i = seismic intensity level (i = VI, VII, VIII, IX, X) N t,d,i = number of buildings of type t having damage level d at seismic intensity level i CR t,i = proportion of deaths by building type, damage level and seismic intensity Casualty and building damage data has been collected for over 70 municipalities (commune) for three large Italian earthquakes: - Irpinia 1980, 214 municipalities - Friuli 1976, 26 municipalities - L Aquila 2009, 8 COMS ) NO t = number of occupants (at the time of the event) by building type t SYNER-G Casualty vs. Coburn & Spence Real Simulated Coburn and Spence Average Error (Opt.) Average Error (C&S) Intensity SYNER-G Final Workshop, Milano, March 2013

21 Conclusions Framework provides a holistic evaluation of Shelter needs and Health Impacts by integrating physical systems model with socio-economic vulnerability based on a stratified decision criteria and vulnerability parameters in each model. Multi-criteria frameworks for shelter and health impacts can be used as a decision support and planning tool: - provide a rational but easily understood benchmark - can be linked directly to the damage loss parameters of the city - can communicate gaps in shelter demand for administrative districts Urban Audit: data completeness problem, however provides a common indicator database at the European Level to harmonize selection of indicators - Validation of Model with L Aquila - Implementation in Thessaloniki SYNER-G Final Workshop, Milano, March 2013