Content uploaded by Asad Asadzadeh
Author content
All content in this area was uploaded by Asad Asadzadeh on Sep 26, 2016
Content may be subject to copyright.
The 368 urban neighbourhoods (within 116 sub-regions, and 22
regions) of Tehran City (Iran) were chosen as the case study area
and validation tool of composite indicator building in this study
which consists of the following six main steps.
An Inductive Approach for Developing Disaster
Resilience Indices in the Context of Earthquake Hazard
Asad Asadzadeh1, Theo Kötter2, Dominik Weiss3
Institut of Geodesy & Geoinformation
Department of Urban Planning and Land Management
http://www.igg.uni-bonn.de/psb/
Nußallee 1
53115 Bonn
Tel.: +49 228 732610
1. asad.asadzadeh@uni-bonn.de
2. koetter@uni-bonn.de
3. weiss@igg.uni-bonn.de
Understanding and enhancing of commuity disaster resilience is
intrinsically linked to the ability to measure levels of disaster
resilience. Constructing composite indicators has often been
addressed to perform this task in literature. While the prevailing
approach in measuring disaster resilience has been the deductive
approach, using an inductive methodology, this study presents a
place-based measurement of disaster resilience that is both
conceptually and theoretically sound and easy enough to use in a
risk planning context.
Motivation & Target
1. Developing or application of a
theoretical framework as a basis
for indicator building
2. Selecting and finalizing
indicators that are sound, robust
and relevant
3. Data transformation and
overcoming incommensurability
4. Multivariate assessment (data
reduction and factor retention)
5. Weighting & Aggregation (of
indicators or group of them)
6. Visualization & Validation
(mapping and validity of results)
2. Indicators as proxies for resilience
5a. Unequal weighting of indicators using hybrid PCA-ANP
In total, 30 indicators were selected for measuring disaster resilience in the
context of earthquake hazard.
4. Extracted components and their explained variance via PCA
Components Variance (%)
Primary
Variables Abbr. Factor Loading
1. Built environment & Social
dynamics
17,67
Percent of urban deteriorated textures
Percent of the skilled employees
Percent of population with high education
Percent of population above poverty line
Percent of population without disabilities
Percent of housing with telephone
access
UDT
SE
PHE
APL
PWD
HWT
0.713
0.701
0.691
0.670
0.659
0.658
2. Urban land use &
Dependent Population 9,36
Percent of population that are not elderly
Percent of population living in hazardous areas
Percent of building density
Appropriate siting of hospitals and health
centres
NEP
PD
BD
AHH
0.916
0.916
0.799
0.540
3. Socio-cultural capacity 8,18
Number of religious and cultural land uses
Ratio of large to small business
Ratio of recreational and entertainment land uses
RCO
LSB
REI
0.692
0.690
0.540
4. Life quality 7,13
Percent Satisfaction level of neighbourhood relation
Percent population have belonging sense to the
neighbourhood
Per capita household income
Critical
infrastructure
LNR
BSN
HI
CIS
0.768
0.703
0.580
0.466
5. Open space 6,89
Number of schools
Percent of non
-built-up areas NS
NBA 0.949
0.947
6. Social capital 4,7
Percent of Satisfaction from local councils
Percent of Social
trust SLC
ST 0.867
0.776
7. Emergency Infrastructure 4,31
Access to the police stations
Access to the fire station
Number of emergency response
plan
APS
AFS
ERP
0.749
0.712
0.648
8. Economic structure
4,16
Percent of homeownership
Percent of population that are employed
HO
PE 0.677
0.539
Cumulative
variance 62.43
Methodology
Results & Discussions
Goal
Components
Indicators
Goal
Components
Indicators
5b. Linear additive aggregation
•: Disaster resilience score for neighbourhood “
•: Relative importance of indicator “ obtained from ANP
•: Normalized value of indicator “ in neighborhood “
6a. The multidimensional & multiscale patterns of disaster resilience
6b. Validation (Cross validation)
Conclusion
The inductive method of PCA has caused to analyse the interactions (causality)
between the 30 selected indicators and extracting the latent patterns of them as
well as the eight associated dimensions.
The logic of ANP predisposes to consider the interdependencies among all
indicators and provided a different relative importance for each indicator.
The study conceptualized the term disaster resilience in the context of
earthquake hazard and highlighted the hot-spots of disaster resilience level at
three urban scales.
0,1
0,2
0,3
0,4
0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
PCA-ANP DRI
BRIC DRI
Acknowledgement: the sincere gratitude goes to Prof. Dr.
Zebardast (University of Tehran, Iran) for his valuable
comments in methodological part of the study.
The results from PCA are entered
into the analytic network process
(ANP) model and pair-wise
comparisons are done between the
decision making elements to form
super matrix. But instead of the
normative scale of 1-9 in AHP or
ANP, the absolute measurements
obtained through the PCA are used
for calculations as follows:
Then, limit super matrix is calculated which displays the absolute value of
relative importance for individual variables.
1. Multifaceted concept of disaster
resilience
Disaster Resilience of a Place (DROP)
Model (Cutter et al. 2008), and its
operationalized version called Baseline
Resilience Indicators for Communities
(BRIC), (Cutter et al.2010 &2014).
3. Data Gathering & Normalization
Using Min-Max to linearize the relations among data and rescaling into a
comparable scale between 0 and 1.
BRIC model with its own
methodology was applied to the
same data set and the results were
compared with the results obtained
by PCA-ANP model. Although the
resilience scores in both models
differ, there is a strong positive
relationship between them.