Nat. Hazards Earth Syst. Sci.

, 15, 789–803, 2015

www.nat-hazards-earth-syst-sci.net/15/789/2015/
doi:10.5194/nhess-15-789-2015
© Author(s) 2015. CC Attribution 3.0 License.

Open space suitability analysis for emergency shelter

after an earthquake
J. Anhorn1,3 and B. Khazai2,3
1 South Asia Institute (SAI), Heidelberg University, INF 330, 69120 Heidelberg, Germany
2 Center for Disaster Management and Risk Reduction Technology (CEDIM), Karlsruhe Institute of Technology,
Hertzstrasse 16, 76187, Karlsruhe, Germany
3 HEiKA – Heidelberg Karlsruhe Research Partnership, Heidelberg University and Karlsruhe

Institute of Technology, Heidelberg and Karlsruhe, Germany

Correspondence to: J. Anhorn (anhorn@sai.uni-heidelberg.de)

Received: 2 April 2014 – Published in Nat. Hazards Earth Syst. Sci. Discuss.: 11 June 2014
Revised: 14 October 2014 – Accepted: 8 March 2015 – Published: 10 April 2015

Abstract. In an emergency situation shelter space is crucial 1 Introduction

for people affected by natural hazards. Emergency planners
in disaster relief and mass care can greatly benefit from a
As the impacts of natural disasters continue to increase
sound methodology that identifies suitable shelter areas and
around the world, experts agree that post-event response has
sites where shelter services need to be improved. A method-
to become more efficient and draw on science (Balcik et al.,
ology to rank suitability of open spaces for contingency plan-
2010; Bharosa et al., 2010; McEntire, 2007; Rawls and Turn-
ning and placement of shelter in the immediate aftermath of a
quist, 2010). This becomes visible, for example, through the
disaster is introduced. The Open Space Suitability Index uses
formulation of the United Nations International Strategy for
the combination of two different measures: a qualitative eval-
Disaster Reduction (UN ISDR) in 2000. Pre-disaster plan-
uation criterion for the suitability and manageability of open
ning as well as risk mitigation measures have gained inter-
spaces to be used as shelter sites and another quantitative cri-
est in both scientific and practitioner communities. As the
terion using a capacitated accessibility analysis based on net-
primary international agreement for disaster reduction, the
work analysis. For the qualitative assessment implementation
Hyogo Framework of Action prioritizes “strengthening pre-
issues, environmental considerations and basic utility sup-
paredness for response” as one of its five priorities of action
ply are the main categories to rank candidate shelter sites. A
identified for 2005 to 2015 (UN ISDR and UN OCHA, 2008,
geographic information system is used to reveal spatial pat-
p. 1).
terns of shelter demand. Advantages and limitations of this
One important concern of strategies to improve prepared-
method are discussed on the basis of an earthquake hazard
ness for response is the identification and provision of suit-
case study in the Kathmandu Metropolitan City. According
able areas for emergency shelter before disasters unfold
to the results, out of 410 open spaces under investigation,
(Chandler, 2007; Chien et al., 2002; Donohou, 2012; Perry,
12.2 % have to be considered not suitable (Category D and E)
1979; Perry and Green, 1982; Tai et al., 2010). Especially
while 10.7 % are Category A and 17.6 % are Category B. Al-
in urban contexts the availability of such areas is often lim-
most two-thirds (59.55 %) are fairly suitable (Category C).
ited and there is increasing demand for risk-sensitive land use
planning which are often lacking (e.g., Global Communities,
2012).
Shelter needs can be divided, according to the time elapsed
from the onset of the disaster event, into emergency shelter,
temporary shelter, temporary housing, and permanent hous-
ing (Chou et al., 2013; Donohou, 2012; Félix et al., 2013;
Johnson, 2007, 2009; Lizarralde et al., 2009; Quarantelli,

Published by Copernicus Publications on behalf of the European Geosciences Union.

790 J. Anhorn and B. Khazai: Emergency shelter after an earthquake

1995). The timeline for transitioning from these different earthquake scenario” in which shelter placement is exclu-
phases of shelter needs – for example from emergency shel- sively based on open spaces, as very few buildings, such as
ter to temporary shelter – is often variable; however, the un- schools and shopping malls, can be considered stable enough
derlying sequential process seldom becomes reality (John- to be used for shelter purposes.
son, 2007; Ritchie and Tierney, 2011). Earthquakes confront The paper is structured as follows: first, the rationale of
emergency managers with special challenges due to their a combined method to investigate capacity-based suitabil-
rapid onset and relatively short duration. Furthermore, as ity of shelter sites is given. As such, existing methods to
earthquakes are inherently unpredictable, there is usually no calculate displaced and shelter-seeking populations resulting
lead time for preemptive evacuation, which results in emer- from earthquakes as the fundamental prerequisite of such
gency shelter placement becoming mostly a post-event is- a methodology are reviewed. Second, a set of categories to
sue (e.g., Wright and Johnson, 2010). Pre-event planning and characterize site suitability based on qualitative indicators is
preparedness for emergency shelter placement is thus critical proposed. Furthermore, a methodology to derive capacitated
for ensuring a coordinated response during the complex and accessibility using spatial network analysis as a key measure
changing risk contexts after a large earthquake. to evaluate further shelter needs in a spatial context is in-
Planning for emergency shelter placement draws on stan- troduced. Third, the combined OSSI is outlined. Fourth, the
dards, criteria, and guidelines developed for emergency man- proposed methodology for open space suitability analysis for
agers and humanitarian organizations which have been based emergency shelters is applied to our case study in KMC. The
mostly on post-disaster assessments (e.g., Da Silva, 2007; final sections reflect on the results from the case study and
SPHERE Project, 2011; UNDRO, 1982; UN OCHA et al., discuss limitations and the transferability of the method to
2010). For example, the SPHERE Project provides minimum other hazards.
standards and general guidance for use in any of several re-
sponse scenarios and includes provisions for strategic plan-
ning, settlement planning, covering living space, construc- 2 Shelter suitability
tion, and environmental impact for shelter and settlements
(SPHERE Project, 2011). While the minimum standards pro- 2.1 Shelter need
vide the basis for developing an emergency shelter placement
plan, optimal siting and accessibility of shelter sites based The initial estimation of the potential number of displaced
on shelter needs from comprehensive risk assessments are population after a disaster is a major step in emergency man-
also required (Indriasari et al., 2010). There is still a lack of agement and a prerequisite for calculating temporary shel-
combined approaches to investigate demand for public emer- ter demand. While many casualty estimation methodologies
gency shelter sites with their suitability and accessibility in- exist in earthquake engineering that provide estimates of
corporating capacity constraints of (candidate) shelter sites. both injuries and fatalities by relating the intensity of the
In this paper we propose a methodology that examines the earthquake and/or damaged buildings to casualty potential
capacity of open spaces to be used as public emergency shel- (Coburn and Spence, 2006; FEMA, 1999, 2011; Samardjieva
ter sites, which takes into account both how well a site meets and Badal, 2002), methods for estimating displaced popula-
demand for public shelter as well as the level of accessibility tion and population in need of shelter are far fewer. Examin-
of the site using a deterministic earthquake risk assessment. ing data from 457 historic earthquakes from 1900 to 2012 in
Alongside the quantitative capacity analysis, a set of qualita- the CATDAT damaging earthquake database (Daniell et al.,
tive suitability criteria (SI) are proposed for open spaces to be 2011; Khazai et al., 2014) shows that while a general linear
used as temporary shelter sites during an earthquake emer- trend on logarithmic scale is observable between damaged
gency. The combined Open Space Suitability Index (OSSI) buildings after an earthquake and the number of homeless
will rank candidate sites according to their accessibility, tak- people, for many events there are scalar differences from this
ing into consideration the available capacity and also their trend that not only depend on external factors like building
suitability for earthquake shelter purposes based on expert damage, loss of utilities, and weather conditions but also on
knowledge. internal socioeconomic and individual factors such as safety
We showcase this methodology on officially identified concerns or fear of aftershocks (Khazai et al., 2014).
open spaces by the National Society for Earthquake Tech- Most earthquake loss estimation (ELE) software for cal-
nology (NSET), Nepal, and the International Organization culating shelter needs is based on the HAZUS methodology
for Migration (IOM) within Kathmandu Metropolitan City (ABAG, 1996; Harrald and al Hajj, 1992) and accounts for
(KMC) using the combination of two different measures: a several variables on the census track level influencing the ten-
qualitative evaluation criterion for the suitability and man- dency to seek short-term shelter, including income, ethnicity,
ageability and a second quantitative criterion using a capaci- age, and ownership (FEMA, 2011). Chou (2013) proposes
tated accessibility analysis based on both an earthquake risk the use of three variables determining higher tendency to
analysis and a network analysis based on a geographic in- seek shelter out of all displaced people affected by an earth-
formation system (GIS). We thereby assume a “worst-case quake, namely low household income, rented housing tenure,

Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015 www.nat-hazards-earth-syst-sci.net/15/789/2015/

J. Anhorn and B. Khazai: Emergency shelter after an earthquake 791

and belonging to either the youngest ( < 16 years) or the old- tion, and health, community infrastructure, security, and sec-
est (> 65 years) age group. Chien et al. (2002) use contextu- ondary risks (SPHERE Project, 2011, p. 247 ff.). All in all,
alized weights explored in a shelter survey after the Chi-Chi the standards aim to use strategic settlement planning to en-
Earthquake in Taiwan to revise the HAZUS default values. able “safe and secure use of accommodation and essential
Shelter needs are mostly calculated directly as a function services by the affected population” (SPHERE Project, 2011,
of structural damage to buildings not taking into account p. 254). Especially in densely populated urban areas, the
household decision making or social and demographic fac- shelter demand can exceed the supply in close vicinity of
tors, which is considered a deficit by some authors (Khazai the affected population if these standards are enforced rigor-
et al., 2011, 2014; Tierney et al., 2001). Besides building ously. Limited available space may urge adjustments on the
damage, social factors have emerged as crucial in forming applied average shelter space per person.
the decision to seek shelter or not on a household and indi- Suitability of open spaces for shelter purposes depends on
vidual level (Chang et al., 2009; Chou et al., 2013; Khazai the perspective from which it is evaluated. Da Silva stresses
et al., 2014). Riad et al. (1999) state that besides risk char- the need to consider the “shelter occupant’s perspective” (Da
acteristics, territorial tendencies (house ownership) and per- Silva, 2007, p. 25) when shelter site quality is evaluated.
sonal characteristics – like social support, education, finan- Emergency planners and affected population may have a dif-
cial wellbeing – are influencing people’s decision to seek fering perception of relevant considerations in the immediate
shelter. Additionally, they state that “social influences on aftermath of an earthquake. For example, emergency plan-
evacuation behavior may vary according to the resident’s net- ners evaluate shelter sites with respect to a longer time frame.
work size and ethnicity” (Riad et al., 1999, p. 921). Another Limitations for implementation, existing secondary hazards
important determinant of the number of people seeking shel- as well as future construction plans play a predominant role.
ter was found inter alia by Wright and Johnston (2010) and A shelter-seeking person, however, focuses on accessibility
Chang and Chamberlin (2003) to be the loss of lifelines. In- and space availability when looking for an immediate emer-
teractions between the physical damage state of buildings gency shelter site. Hence sheltering in the close vicinity of
and the combined residual service level in the utility net- one’s own plot or house may be of greater importance than
works have been considered in a system approach to as- mid-term perspectives.
sess the habitability of buildings from which the number of The qualitative suitability indicators used for the OSSI are
displaced persons can be computed (Cavalieri et al., 2012; described in the following section and are mostly inferred
Khazai et al., 2013). The rationale for this is that people are from the SPHERE standards. As a matter of course other rel-
likely to seek refuge in a public shelter if they are cut off evant indicators should be added if applicable (e.g., differ-
from basic necessities such as water supply or electricity, ent local context, data constraints, or expanding to other haz-
even if buildings are otherwise intact. For example, during ards). The selection and weighting of indicators remains the
the L’Aquila earthquake of 2009, shelter seekers originated greatest difficulty and needs to incorporate expert judgment.
not only from non-usable (collapsed or cut off from lifelines) The weighting of categories and indicators was done in a par-
buildings but also from partly damaged and non-damaged ticipatory way in four consecutive expert group discussions
buildings (up to 54 %) (Khazai et al., 2012). Furthermore, involving emergency management researchers from NSET
risk perception and access to resources are identified as in- and from the Center for Disaster Management and Risk Re-
fluential factors by Chang et al. (2009) and Chakraborty et duction Technology (CEDIM) in Karlsruhe, Germany. The
al. (2005). Upreti (2009, p. 52) shows that 87 % of Kath- rationale behind the choice of individual factors often lies in
mandu’s citizens do not believe an earthquake is going to the stakeholders’ or experts’ experiences and available data.
happen during their lifetime even though they are aware of Consequently, it is important to state concisely the scope and
the possibility. This shows that most people will be hit un- objectives of such an index. The methodology we propose
prepared should a major disaster occur in the near future. focuses on suitability for immediate emergency shelter, with
weighting of indicators and categories applicable to this con-
2.2 Suitability text. The following three core categories have therefore been
identified to explore suitability of open spaces in an urban
The SPHERE shelter and settlement standard for covered liv- context for immediate shelter after an earthquake: implemen-
ing space provides a guideline for emergency managers to tation issues, environmental considerations, and basic utili-
evaluate or plan for immediate, short-, and long-term shel- ties supply (Table 1).
ter (SPHERE Project, 2011). It recommends an area in ex- The category implementation issues consists of owner-
cess of 3.5 m2 per person to meet requirements of typi- ship, existing use, and future plans. Generally, publicly
cal household activities. The overall surface area per per- (governmental) owned spaces should be preferred, as these
son, including communal space for cooking, roads and foot- can be managed easier than privately owned open spaces
paths, educational facilities, administration etc., within tem- (cf. FEMA, 2007). Another indicator is the current type of
porary communal settlements should be 45 m2 . Besides sur- use. A playground or a park for example is best suited for
face area, special consideration is given to water, sanita- shelter, since their existing type of use does not hamper camp

www.nat-hazards-earth-syst-sci.net/15/789/2015/ Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015

792 J. Anhorn and B. Khazai: Emergency shelter after an earthquake

Table 1. Overview of suitability categories and indicator criteria for immediate shelter sites.

Category Weight Indicator Score Explanation

Implementation 0.1 Ownership 1 public (governmental, community, religious, institutional, educational)
issues 0.7 private
0.06 Future plan 1 no plan, planned park, planned garden, planned
playground, or long-term structure plan
0.5 short-term structure plan
0 under partial or full construction
0.2 Existing use 1 non-used, park, garden, or playground
0.7 religious
0.5 agricultural or institutional
0.4 educational
0.1 dumping site
Environmental 0.18 Secondary 1 no secondary hazard
considerations hazards 0.7 fire or flood hazard
0.5 fire and landslide hazard
0.4 fire and flood hazard
0.2 fire and landslide and flood hazard
0.1 Pollution 1 Category 0: no pollution
issues 0.9 Category 1: noise pollution or air pollution
0.8 Category 2: river pollution
0.5 Category 3: urban waste pollution
0.4 Category 1 and Category 3
0.3 Category 2 and Category 3
0.2 Category 1, Category 2, and Category 3
Basic 0.1 Electricity 1 distribution line and generator(s) or alternative source
utilities 0.9 generator(s) or alternative source
supply 0.7 distribution line
0.1 no electricity available
0.11 Water 1 some type of source and tank and piped water
supply 0.8 some type of source and tank
0.7 some type of source and piped water
0.6 some type of source (natural source, ground water, or deep boring)
0.5 tank and piped water
0.4 tank
0.2 piped water
0 no water supply available
0.15 Nearness to 0.9 hospital(s) within less than 1 km distance
critical 0.8 hospital(s) within more than 1 km distance but less than 2 km
facilities 0.6 hospital(s) within more than 2 km distance but less than 3 km
0.4 hospital(s) within more than 3 km distance
0 unknown distance to next hospital

erection. If the space has an institutional or educational func- areas were considered the least suitable ones. The rationale
tion it should not be prioritized for immediate shelter in or- behind this category indicates possible restrictions in access
der to not delay the resumption of daily activities and not to or continuing use.
endanger people due to potentially unstable building condi- As an environmental consideration we include pollution
tions (c.f. SPHERE Project, 2011). The future planning indi- and secondary hazard criteria. Air, water, or waste pollution
cator gives a last indication regarding usage complications. needs to be taken into account. Only shelter sites with a clean
Some sites have existing long- or even short-term plans in air supply (i.e., not in areas with high air pollution discharge
place, are already under partial or complete construction, and due to debris or road traffic) and cleared grounds (i.e., no
should thus not be preferred. During the site visits, some of dumping areas) should be chosen as shelter sites. Possible
the places turned out to be used as dumping sites and such shelter sites should not include areas prone to other haz-

Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015 www.nat-hazards-earth-syst-sci.net/15/789/2015/

J. Anhorn and B. Khazai: Emergency shelter after an earthquake 793

ards (floods, landslides, etc.). Additionally, secondary haz- Indriasari et al. (2010) use a similar approach to identify
ards like potential fire eruption from nearby hazardous mate- the optimal siting of emergency facilities like fire brigades
rials (e.g., gas and petrol sellers) need to be taken into ac- or hospitals. They argue that maximum coverage is more ap-
count and avoided in the choice of a suitable shelter area plicable for identifying suitable emergency facilities among
(FEMA, 2007; Tai et al., 2010). a larger set of candidate sites than methods minimizing the
Access to basic utility supply systems and critical infras- distance between demand and supply. In general, the main
tructure such as hospitals need to be considered as part of site difference between the approaches is the spatial domain:
suitability. Sufficient and continuous supply with basic utili- Gall (2004) uses a raster-based model with continuous fric-
ties such as water and electricity are crucial immediately af- tion data, while Indriasari et al. (2010, p. 2014) apply the
ter a disastrous event (Chang et al., 2009; Chu and Su, 2011; facility location problem on a street network “taking into ac-
Daley et al., 2001; FEMA, 2011). The proximity to medi- count the road access, barriers, and road network attributes”.
cal services has also found wide acceptance as an important All these methods focus on emergency facility location prob-
factor (FEMA, 2007; SPHERE Project, 2011). Hospitals are lems for preemptive evacuation, which differ from the chal-
particularly important due to high numbers of injuries incur- lenges the shelter-seeking population faces in the aftermath
ring during an earthquake and to prevent high numbers of of an earthquake.
post-event “fade-away” people (Coburn et al., 1992). Network analysis has been proven to be a valuable tool for
analyzing the strengths and weaknesses of manifold types of
spatial and non-spatial networks (cf. Crucitti et al., 2006).
2.3 Accessibility
With its theoretical foundation in graph theory, road net-
works are defined as elements of nodes and edges, either
People seeking shelter rely on some sort of existing and suit- using street segments as edges (primal representation) or as
able network (roads) to access available shelter areas within a nodes (dual representation) (Porta et al., 2006a, b). The most
certain time (Tai et al., 2010). Kongsomsaksakul et al. (2005) important feature and analytic strength of network analysis
use a two-level mathematical representation to show author- is the inherent importance of relational topological informa-
ities’ selections of best-suited evacuation sites on the one tion. Results often comprise of the summed costs (e.g., time,
hand and evacuees’ decisions on the escape route to that de- length) or turns of nodes between predefined sets of origins
termined site on the other hand. The number of people seek- (demand) and destinations (supply). Network analysis, for
ing shelter and the decision to access a particular shelter site example, allows calculating least-cost distances in terms of
relies on several factors ranging from socio-cultural and eco- travel time or distance using impedance values for different
nomic factors to physical constraints like road network ac- node types from/to destinations. Other measures are service
cessibility after the event and availability of motorized or areas to determine the extent of business relations or run cal-
non-motorized vehicles. culations for logistic fleet management or manifold facility
Many studies focus on transportation issues in terms of location problems (e.g., Toregas et al., 1971). The usability
time constraints to reach evacuation sites during preemp- of network analysis in the emergency context has been shown
tive evacuations (Cova and Church, 1997; Cova and John- on different examples, like optimal siting of emergency facil-
son, 2002; Kar and Hodgson, 2008). Others focus on differ- ities (Indriasari et al., 2010) and emergency routing services
ent variables determining the “evacuation assistance needs” on near-real-time basis (Neis et al., 2010; Weiser and Zipf,
(Chakraborty et al., 2005, p. 23) based on social vulnerability 2007). Differences in accessibility constraints during or af-
and earthquake risk patterns. ter extreme events affecting road networks can be investi-
Kar and Hodgson (2008) use a GIS-based suitability gated using, for example, volunteered geographic informa-
model to investigate the number and location of predefined tion (VGI) (Neis and Zielstra, 2014).
shelter areas for preemptive hurricane evacuation. They iden- With their Urban Network Analysis toolbox, Sevtsuk and
tify a set of factors from official and unofficial guidelines Mekonnen (2012) introduce an additional level of analysis to
and determine the suitability of shelter sites using weighted the traditional calculation of network centrality: the building
linear combination and a pass/fail screening on raster basis. level. Previous studies focused solely on the capabilities and
The shelter sites used in their study are mostly public multi- centrality measures of the network itself (nodes and edges),
purpose assembly facilities, like cultural or civic centers, ignoring individual elements along the edges. They promote
and healthcare facilities. Factors included are proximity mea- adding buildings as supplementary nodes and establishing
sures and vulnerability profiles of the population (percentage links between single buildings and the adjacent (closest) road
of children, elders, minorities, and low-income households). network.
Gall (2004) highlights the importance of shelter sites for hu- We use the Maximize Capacitated Coverage analysis (im-
manitarian assistance in terms of relief good distribution. The plemented in ESRIs ArcGIS™ 10.1 Network Analyst) to de-
model follows some basic assumptions that are only applica- termine the maximum coverage of selected sites, taking into
ble in rural areas where transportation friction can be mod- consideration network impedance, building weight, and shel-
eled as a result of land cover and distance only. ter capacity. The method uses Dijkstra’s algorithm for finding

www.nat-hazards-earth-syst-sci.net/15/789/2015/ Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015

794 J. Anhorn and B. Khazai: Emergency shelter after an earthquake

(POPservedOS ). The CAMOS determines the “pressure” on

each candidate site to be overcrowded due to the surround-
ing undersupply. It shows a spatial representation of shelter-
demanding population that can be served with appropriate
shelter space. The 1 km cutoff value is used as a standard-
ization factor and determined by the overall size of the used
network. It helps to identify hot spots of unserved populated
areas within the urban environment. Therefore, the optimal
facility location problem is modified to address existing suit-
ability constraints and limited capacity of shelter areas.

4 Case study Kathmandu Metropolitan City

Figure 1. Open Space Suitability Index (OSSI) evaluation scheme.
The territory of Nepal spans about one-third of the length of
the Himalayan arc, leading to a long history of devastating
the shortest paths and solves the location–allocation problem
earthquakes in Nepal. Over the last century, the Himalayan
by choosing a subset of facilities (candidate shelter sites)
arc has experienced four earthquakes with magnitude around
such that the sum of the weighted distances from each de-
8.5 in 1897, 1905, 1934, and 1950. As one of the most dev-
mand point (with a certain weight) to the closest shelter site is
astating earthquakes in the recent past, the 1934 Bihar earth-
minimized (ESRI, 2013). Thus it assigns each demand point
quake of magnitude Mw 8.3 caused the collapse of 20 % of
(building) to the closest candidate shelter facility (supply) ac-
all buildings in the Kathmandu Valley; another 40 % were
cording to the number of people seeking shelter (weight),
severely damaged (EMI, 2010; JICA and MoHA, 2002). To-
taking into consideration the overall capacity and the total
day, the total population in the Kathmandu Valley is 8-fold
length network distance of all buildings. Capacity of candi-
what it was in 1934; its density has quadrupled. Expansion
date shelter sites is deduced using existing standards for cov-
took place without political supervision, despite various ef-
ered living space as described earlier. The number of people
forts to enforce spatial planning (Gutschow and Kreutzmann,
seeking shelter is used as the weighting factor for each build-
2012; Thapa et al., 2008; Thapa and Murayama, 2009). As
ing.
the political and cultural capital of Nepal, KMC within the
Kathmandu Valley is particularly at risk. With its fertile
3 Open Space Suitability Index land the valley has attracted many people living off farm-
ing (Gutschow and Kreutzmann, 2012). However, with rapid
The objective of this study is to model shelter site suitabil- growth of urban centers much of the open land has vanished
ity considering road network accessibility, capacity, and suit- in favor of built-up living space in the recent decades (Haack
ability of shelter. We focus on immediate shelter placement and Rafter, 2006). Today, with an average annual growth rate
with a time frame up to several days following an earthquake. of 4.59 % between 2000 and 2005, Kathmandu is one of the
The final suitability index OSSI consists of two factors: first fastest-growing city in the world, facing high earthquake risk
an expert-based weighting procedure of SI and second a GIS- (UN DESA, 2012). Owing to this unimpeded urban growth,
based accessibility and capacity measure (CAMOS ). Figure 1 an earthquake of similar magnitude as the Bihar earthquake
shows the evaluation scheme applied. It is calculated using would result in significantly higher losses in the form of casu-
the following equations: alties and physical destruction (Dixit et al., 2000; Guragain
n
X et al., 2008). Experts estimate that at least 1 million home-
OSSIOS = (Wi × Ii + Wi+1 × Ii+1 + . . . + Wn × In ) less people in need of immediate assistance can be expected
n=1 and all routes into and out of the Kathmandu Valley will be
× CAMOS , (1) blocked for weeks if not months (NRRC, 2013). Assuming
this holds true, all emergency services need to be supplied
POPservedOS
CAMOS = , (2) from within the valley – without external help.
POPOS The above-mentioned 2-fold suitability analysis of open
where Ii is the suitability indicator scores and Wi is the space shelter sites is implemented in a case study for the
respective weight for each indicator. CAMOS is calculated KMC. The open spaces used in the analysis (Fig. 2) are based
as the ratio between the total shelter-seeking population on 887 open spaces identified by NSET as potential sites for
within a 1 km service area of each candidate shelter site emergency purposes, out of which 410 are located within
(POPOS ) derived from an earthquake risk assessment and the KMC (NSET, 2010, 2012). In the assessment, most pub-
people accommodated within the same spatial unit accord- licly owned cleared areas and smaller open spaces or court-
ing to the Maximize Capacitated Coverage analysis result yards were included. The qualitative suitability information

Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015 www.nat-hazards-earth-syst-sci.net/15/789/2015/

J. Anhorn and B. Khazai: Emergency shelter after an earthquake 795

models used, the mid-Nepal earthquake with Mw 8.0 would

lead to MMI VIII within the valley (JICA and MoHA, 2002).
This is seen as the “worst-case scenario” in terms of the mod-
eled building damages and casualties. The SEDM building
damage estimates were carried out in 2000–2002 and re-
flected the population in 2001 and the building stock from
1998. The first step in computing shelter demand for KMC
was to update the 2001 population with the latest population
data of the 2011 census. Due to the lack of detailed recent
building data including building types, the population ratio
(r) serves as scaling factor to estimate building numbers for
2011 using the ward building inventory of 1998 according to
Eq. (3):

Pop2011
i NB2001
r= ∼ . (3)
Pop1998
i NB1991

This simplification can be made since a comparison of the

1991 Housing Survey and the 2001 National Census re-
vealed that the ratio of building stock to population has not
changed significantly, and population growth between 2001
and 2011 was similar to the previous decade (CBS, 1995,
2002, 2012; NSET, 2012). Unfortunately there are no recent
data on building stock composition available for KMC. This
poses a serious limitation which we accounted for by assum-
Figure 2. Distribution of open spaces in Kathmandu Metropolitan ing a linear increase without specifying details about replace-
City. ments, upgrade, or deterioration of building structures during
the last years. The Nepalese building code developed in 1994
was only in 2003 approved by the government and has never
was obtained using structured data entry forms. The criteria been implemented across the country. Most new buildings
for the identification can be found in the “Shelter Response (private and public) do not comply with earthquake safety
Strategy and Plan for Earthquake Disasters for Kathmandu standards due to the lack of resources (enforcing governmen-
Valley, Nepal” (NSET, 2010, 2012). Other places (like pri- tal chapters, trained masons, financial resources, etc.) despite
vate areas, agricultural land inside the city boundaries) are various efforts (Dixit, 2009). The actual composition of the
not considered. Additionally, IOM and the Ministry of Home building stock, therefore, had to be derived using the simplis-
Affairs jointly identified 83 open spaces for medium-term tic linear upscaling.
post-disaster needs including larger facilities for camp es- The need for public emergency shelter was computed
tablishing (IOM and GoN, 2012). In their assessment, only based on a modified HAZUS methodology in a two-step ap-
publicly owned sites and areas controlled by commercial en- proach. First, the number of displaced persons in each ward
tities with which the government could enter a formal contin- from the scenario earthquake are computed by assuming all
gency agreement were considered. The qualitative data avail- occupants of heavily damaged buildings will be displaced.
able from both data sets were combined and converted using Additionally, even for building damages that may be moder-
the weighting scheme formulated in four consecutive expert ate, some buildings may not be habitable, as lifeline breaks
round-table discussions (Table 1). They form the basis for the (e.g., water and electricity utilities) for an extended time of-
qualitative part of the OSSI. The available area of 2285 km2 ten leads to people seeking shelter outside of their other-
supplies a maximum of 253 900 persons as shelter, applying wise usable homes (e.g., Khazai et al., 2013). As of today,
a standard of 9 m2 per person. many people, especially in the core area of KMC, rely on
The Japan International Cooperation Agency (JICA) water tankers servicing the area once a week or less (UN-
“Study on Earthquake Disaster Mitigation for Kathmandu HABITAT, 2008). A high proportion of displaced persons
Valley, Nepal”(SEDM) has been used to deduce building can be assumed from partially damaged buildings since it is
damages for a potential earthquake scenario (JICA and expected that secondary damages to water pipelines will af-
MoHA, 2002). The respective ground motion, seismicity, and fect 80 % of water users (JICA and MoHA, 2002; cf. NRRC,
fault model used can be found there. Unfortunately this study 2013). Finally, partially damaged buildings of low-strength
is the most recent published earthquake assessment for the masonry made of fired bricks in mud mortar are treated
Kathmandu Valley at this time. Out of the three different fault as a special category. Even where partially damaged build-

www.nat-hazards-earth-syst-sci.net/15/789/2015/ Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015

796 J. Anhorn and B. Khazai: Emergency shelter after an earthquake

ings of this type could provide some shelter, past earthquake demand is 342 300 persons. Especially the core wards with
events show that aftershocks threaten to collapse these types their weak building structure and very high population den-
of buildings and most survivors remain outside (Khazai and sities are expected to have large numbers of casualties and a
Hausler, 2005). Thus, the total number of displaced persons very high shelter demand.
in 2011 (DP2011 ) in KMC is given by the sum of displaced The spatial representation used to calculate the CAMOS
persons in each ward i minus the casualties (C) in ward i as consists of building blocks and roads. A detailed road net-
given by Eq. (4): work provided by the Kathmandu Valley Development Au-
thority (KVDA) was utilized. It consists of 1250 km roads
X X classified into nine different types. The established topolog-
DP2011 = r × HD_All1998
i + 0.9 PD_BM1998
i ical network has 27 724 nodes and 67 118 edges. Addition-
i i
! ally, 72 783 building footprints based on Quickbird® satellite
X
1998
X
1998 imagery from 2006 were included as demand points for all
+0.8 PD_nonBMi − Ci . (4)
i i
network-based measures. Within the core area, many build-
ings are not directly connected to the nearest road segment
According to Eq. (4), 100 % of people from highly damaged but through a sequential arrangement of courtyards and nar-
buildings of all types (HD_All), 90 % of people from par- row passages. In extreme cases, several high-rise dwellings
tially damaged brick in mud mortar buildings (PD_BM), and share a single courtyard with only one exit point towards
80 % of people from partially damaged buildings of all other other courtyards before even reaching a road or trail. Map-
types (PD_nonBM) will be displaced. While some displaced ping of such narrow trails from satellite imagery is almost
people will seek to use public shelter, experience in Nepal impossible. In these cases, courtyards were used as build-
has shown that a fraction of the population will access other ing block centroids with a higher weight and manually con-
forms of shelter such as staying with friends and family or nected to the main road network. Shelter demand calculated
migrate to their original cities and villages. Likewise, a por- on ward level in the first step had to be spatially disaggre-
tion of the population will use their property or nearby ar- gated onto the building blocks. This is done by neglecting
eas as makeshift shelter sites (NSET, 2012). In a 2012 study day- and nighttime population and occupancy rates for dif-
on shelter response strategies by NSET it was determined ferent building use. Additional knowledge on population dy-
that approximately 5 % of the population will take shelter namics, as in Freire and Aubrecht (2012) might, be benefi-
with their families and friends; approximately 5 % will take cial for a more detailed study. Optionally, remote sensing has
shelter in damaged houses or self-managed temporary shel- shown advantages in assessing the urban fabric and popula-
ters nearby original houses; and approximately 2 % will mi- tion distribution in larger agglomerations (e.g., Aubrecht et
grate to outside cities and villages (NSET, 2012). Two fac- al., 2013; Kubanek et al., 2010; Taubenböck, 2008). Some
tors of residential urban fabric and migration to rural areas key numbers and characteristics of the data used in this ex-
are thus considered here in determining a ward level distribu- emplary case study can be found in Table 3.
tion of populations seeking shelter in planned, public emer- Using this spatial representation of the urban environment,
gency shelter sites from the computed displaced population. each building with its allocated weight corresponding to the
First, the shelter-seeking population is obtained by reducing number of persons seeking shelter is assigned to the near-
the total displaced population by 2, 10, or 15 % depending est open space, taking into consideration network impedance
on the corresponding levels of residential urban fabric (Ta- and the sheltering capacity of that particular space. The loca-
ble 2). The assumption is that in sparsely built urban areas tion problem is solved so that (a) the nearest site is selected,
where there is more outdoor space, a greater portion of the (b) the overall weighted distances along the network for all
displaced population (up to 15 %) is likely to take up shel- buildings is minimized across the study area, and (c) no site
ter on their own property or nearby areas rather than seek- remains unselected as long as there are buildings which are
ing shelter in the designated emergency shelter sites. In more not served or the capacity is not reached.
dense urban areas, however, there is little or no space for One main obstacle to most network analysis methods are
self-managed shelter, thus only 2 % of the displaced popu- spatial boundary problems. The complete network and build-
lation may seek temporary shelter on non-designated open ing database was available for inside KMC only. Thus peo-
spaces. Next, the displaced population seeking shelter is fur- ple from outside KMC seeking shelter in any open space in-
ther reduced by the internal migration rate from each ward side the municipal boundary or persons inside KMC seeking
based on the 2001 population census (Subedi, 2010). Here shelter outside the city boundaries were not considered. In
the assumption used is that 5 % of the internal migrants in special cases along the ring road, the identified open spaces
each ward will migrate to outside cities and villages instead consist mostly of two parts on both sides of the lane. To ac-
of seeking public shelter. count for intrusion of people towards KMC, we only used the
The total displaced population within KMC derived from ones towards KMC for the analysis. To the south, KMC bor-
the modified ELE considering social factors and urban fabric ders the Bagmati river forming a physical barrier, which can
settings is thus estimated as 406 500, while the total shelter only be traversed at a few bridges all considered not earth-

Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015 www.nat-hazards-earth-syst-sci.net/15/789/2015/

J. Anhorn and B. Khazai: Emergency shelter after an earthquake 797

Table 2. Shelter-seeking class definition.

Residential urban fabric Shelter-seeking class

Sparse density residential urban fabric approx. 15 % of displaced population will not seek public shelter
Medium density residential urban fabric approx. 10 % of displaced population will not seek public shelter
Dense to very dense residential urban fabric approx. 2 % of displaced population will not seek public shelter

Table 3. Key characteristics of the used database.

Data Value
Number of open spaces inside KMC 410
Available open space 2 284 731 m2
Overall capacity 253 859 pers.
Shelter demand (ELE) 342 299 pers.
Served population (GIS) 253 806 pers.
Unserved population (GIS) 88 493 pers.
Number of buildings (GIS) 72 783
Served buildings (GIS) 54 742
Unserved buildings (GIS) 18 031
Road network length (GIS) 1250 km
Road network nodes (GIS) 27 294
Road network edges (GIS) 66 576

quake safe (JICA and MoHA, 2002; NSET and GeoHazards

International, 1998). Hence for the chosen scenario, it can be
assumed that from or to this side, no movement of population
seeking shelter can be expected.

5 Results

As can be drawn from the raw numbers used for the analysis
(Table 3), there is a lack of shelter space in terms of capacity. Figure 3. The first and last 15 open spaces ranked according to the
342 300 persons were estimated seeking public shelter within suitability indicators.
KMC, using 9 m2 covered living space per person as a stan-
dard. Out of these, 253 900 persons (74 %) can be accommo-
dated using the above set restrictions in terms of distance and and 19.2 % for nearness to critical facilities. This is similar
capacity. within all categories except Category C, where existent use
Figure 3 shows the ranking results of the qualitative suit- gains importance (28.1 %) and nearness to critical facilities
ability criteria for the upper and lower 15 ranks, only dis- drops (6.4 %). Existing future plans for the sites and near-
playing the cumulative value of SI. The OSSI ranking re- ness to critical facilities form an exception for Category A
sults are grouped in 0.2 ranges from Category A (> 0.8 to compared to the average of all categories (7.2 and 12.8 %,
1.0, green) to Category E (below 0.2, red). The most suitable respectively). Water (5.7 %) and electricity supply (5.4 %) as
open spaces in categories A and B add up to a total of 116 well as nearness to critical facilities (12.8 %) all contribute
open spaces, which accounts for almost one-third of all open on average across all categories (A to E) less than the applied
spaces (28.3 %). Categories D and E (not suitable) account weights (11, 10, and 18 %, respectively).
for 50 open spaces (12.2 %). The distribution of OSSI values The map representation of OSSI reveals some hot spots
for all 410 open spaces is shown in Fig. 4. of shelter needs within KMC (Fig. 5). It shows the distribu-
Using the expert-based weighting scheme, the average tion of building blocks that can be served by one of the open
contribution from each of the qualitative indicators for Cat- spaces (light blue in the background) compared to the ones
egory A is 21.1 % for existent use, 12.0 % for ownership, that remain unserved (light orange).
2.1 % for future plan, 17.9 % for secondary hazard, 14.2 % As a result, some wards are very well prepared in terms of
for pollution, 6.3 % for water supply, 7.2 % for electricity, suitable open space for shelter purposes, while others have a

www.nat-hazards-earth-syst-sci.net/15/789/2015/ Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015

798 J. Anhorn and B. Khazai: Emergency shelter after an earthquake

Figure 4. Distribution of OSSI values for all open spaces.

lack in terms of either the capacity of the sites or their suit-

ability. Especially to the west of the core wards, where high-
rise dwellings and extremely dense areas are located, shel-
ter deficits can be observed. Clusters of well-connected and
high-capacity sites, e.g., around Pashupati Temple area in the
east, are important in reducing people’s shelter vulnerability.

6 Discussion and conclusion

Figure 5. Spatial representation of the Open Space Suitability Index
In this paper we analyze 410 open spaces identified as emer- for Kathmandu Metropolitan City.
gency shelter sites within KMC in terms of their suitabil-
ity for shelter. Four aspects are evaluated: shelter imple-
mentation issues, environmental considerations, availability
of basic utilities, and the capacity-based coverage analysis. with improved semi-automatic (object-oriented) feature ex-
The methodology offers a straightforward way to identify traction tools offer wide applications (e.g., Wieland et al.,
hotspots in urban settings in terms of areas under-served 2012). Likewise, local governments often do see a benefit
by open spaces that can be used for emergency immedi- in establishing and maintaining spatial databases which then
ately after an earthquake. It combines an approach to clas- can be used firsthand. In this Kathmandu test case, the data
sify and rank depth-qualitative information on the suitability sets used (road network and building footprints) were readily
of open spaces for emergency shelter available through site available through official governmental units (e.g., KVDA,
visits with knowledge from local experts of quantitative in- Dept. of Survey) and only minor corrections had to be con-
formation on shelter capacity, derived from shelter need cal- ducted. Hence not all earthquake-prone urban areas have to
culations using earthquake risk analysis and site accessibility be considered data sparse.
from a GIS-based network accessibility model. The selection of criteria for qualitative evaluation of open
On the demand side, a comprehensive database of avail- space suitability is based on thorough literature review and
able candidate sites is needed, spatially covering the study the latest design standards (e.g., FEMA, 2007; SPHERE
area. Such data might be available through local agencies as Project, 2011). However, the criteria, sub-criteria, and indi-
in the Kathmandu case but for some areas need to be com- cators used can be taken as a model and customized to fit
piled from other sources or researched. The CAMOS at the the particular needs of a different context. Shelter suitabil-
same time relies on fully functional and topologically cor- ity is calculated for this case study as a function of immedi-
rect road network. As such the proposed methodology de- ate shelter needs derived from structural earthquake damage,
pends on detailed spatial data which might not be available availability of critical services, and accessibility. The con-
in some places. However, advancing tools for deriving data cept behind OSSI could be used for many other hazards,
from remote sensing and/or VGI data (e.g., OpenStreetMap) if shelter needs and suitability criteria (including the pro-
offer huge opportunities for acquiring data. Taubenböck and posed scoring and weighting) as well as the time horizon
Strunz (2013) provide a conceptual framework for some of are contextualized accordingly. As one limitation, this paper
the pertinent questions of earthquake risk reduction using re- focuses on immediate shelter suitability and not reconstruc-
mote sensing. Thus high-resolution satellite imagery together tion of settlements, which would most likely need different

Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015 www.nat-hazards-earth-syst-sci.net/15/789/2015/

J. Anhorn and B. Khazai: Emergency shelter after an earthquake 799

qualitative parameters. While the open space suitability indi- Chang et al. (2012), or Franchin et al. (2006), or incorpo-
cators were developed to be transferable to other urban set- rating the robustness and redundancy of street networks into
tings, they were arrived at and influenced to some extent by the overall suitability might advance the proposed method.
the Kathmandu context. We used an expert-based approach (e) The selected qualitative evaluation criteria their scoring
to identify the most important criteria and evaluated the im- and weighting should always be based on local experts, tak-
portance with the help of group discussions. Besides multi- ing into account contextualized conditions. This also applies
faceted objectives of different stakeholders and experts, data to the potentially necessary incorporation of additional cri-
availability also shapes the selection of certain criteria. Addi- teria. (f) People’s needs and preferences for selecting shelter
tionally, Chien et al. (2002), for example, found that climate places change over time. We only considered a limited num-
and weather conditions in different seasons influence peoples ber of factors influencing suitability for immediate shelter,
shelter-seeking behavior. This and other temporally set fac- taking a mixed planner’s and inhabitant’s position in evalu-
tors were not taken into consideration so far. We have consid- ating them. Medium- and long-term shelter may need differ-
ered flood, landslide, and fire hazard as part of the suitabil- ent factors. The adjustment to such dynamic circumstances
ity indicators under the environmental considerations cate- is what we understand as contextualization of models and is
gory (see Table 1). By considering at least the distance to not yet part of the KMC case study.
critical sources of fire like gas and petrol stations, we aim The hotspot map that was derived according to the OSSI
to avoid exposing people in earthquake shelter to secondary rankings of open spaces can guide decision-makers to de-
threats. Nevertheless, emergency response services (e.g., fire velop strategies and earthquake contingency plans for shel-
brigade) in Kathmandu are known to be very limited in per- ter placement. The analysis specifically addresses the emer-
sonnel and equipment. We would like to highlight the neces- gency shelter logistics and resource allocation problem:
sity to use the most recent available hazard information also where do we expect shelter deficits and where do we need
considering cascading effects to avoid putting people at risk to improve site suitability or identify alternative sites.
in designated shelter areas. In general the indicator-based It has been argued that optimal site selection for emer-
methodology allows for any incorporation of more detailed gency planning needs to consider two main aspects: first,
data (e.g., from flood hazard models) and is transferable to a sufficient quantity of accessible shelter area and second,
other hazards with respect to shelter placement problems. site quality in terms of people-centered shelter needs. An
The proposed methodology to investigate the suitability of indicator-based methodology for combining both the qualita-
open spaces poses some limitations due to data constraints tive suitability criteria and the quantitative shelter needs and
and therefore provides a methodological framework with po- site accessibility measured has been presented through the
tential for further enhancement. Some suggestions are given Open Space Suitability Index. The potential of such meth-
here with respect to the case study in Kathmandu and fu- ods lies in its applicability to further areas, variable types
ture transfer of the method to other cases: (a) population dis- of candidate sites, and/or changing time frames of shelter-
tribution usually varies across time and space within an ur- ing. Therefore, the initial weights and scores of the suitabil-
ban area, so we recommend adjusting this parameter to the ity index need to be contextualized according to the specific
best available model (e.g., Aubrecht et al., 2013). (b) Earth- purpose and possibly different hazard(s). One main recom-
quake risk scenarios highly depend on detailed seismological mendation is to engage local experts and decision-makers in
studies (i.e., microzonation) as well as information about the a participatory approach in the selection and weighting pro-
fragility of elements at risk (critical infrastructure, buildings, cess to achieve consensus about the structure and perceived
etc.). In this case study we only considered one so-called importance of the different indicators. To this end the ap-
worst-case scenario. More ideally, cascading secondary ef- proach outlined here and the assumptions made are based on
fects as well as multiple scenarios should be integrated. consultations with local experts at NSET and developed as a
However, the number of these scenarios must then be re- preparedness tool for emergency shelter allocation in KMC.
duced to become manageable for shelter planning processes. Further studies are needed to test these assumptions both for
(c) The proposed methodology relies on detailed geospatial shelter suitability and shelter demand and to understand bet-
data which are prone to be outdated, fragmented, and lim- ter patterns and behavior of displaced populations in seeking
ited in detail. Users have to identify the most comprehensive public shelter.
data set or make use of promising data capturing tools avail-
able (e.g., Pittore and Wieland, 2013; Wieland et al., 2012).
(d) The road network is considered a fully functional rela-
tional network, the potential failure/disruption of accessibil-
ity due to debris cover or damages. No actual damage of the
road network is accounted for in the capacitated accessibil-
ity measure so far. Modeling road blockage due to debris and
damages, as well as accessibility of building blocks in post-
disaster situations proposed by Caiado et al. (2011, 2012),

www.nat-hazards-earth-syst-sci.net/15/789/2015/ Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015

800 J. Anhorn and B. Khazai: Emergency shelter after an earthquake

Acknowledgements. This study is part of the “Integrated Earth- CBS (Ed.): National Population and Housing Census 2011: Na-
quake Risk Assessment for the Himalayan Region” project (IERA- tional Report, Government of Nepal, National Planning Commis-
Himal), funded by the Heidelberg–Karlsruhe Research Partner- sion Secretariat, Central Bureau of Statistics, Kathmandu, 2012.
ship (HEiKA). We acknowledge financial support by Deutsche Chakraborty, J., Tobin, G., and Montz, B.: Population Evacuation:
Forschungsgemeinschaft and Ruprecht-Karls-Universität Heidel- Assessing Spatial Variability in Geophysical Risk and Social
berg within the funding programme Open Access Publishing. Vulnerability to Natural Hazards, Nat. Hazards Rev., 6, 23–33,
The authors would like to acknowledge Julia Schaper for her doi:10.1061/(ASCE)1527-6988(2005)6:1(23), 2005.
input in the development of this paper. The authors also thank the Chandler, P. J.: Environmental Factors Influencing the Siting
National Society for Earthquake Technology, Nepal, for their great of Temporary Housing in Orleans Parish, Master of Science,
support, provision of data, and insight. We would especially like to Louisiana State University, Department of Environmental Stud-
thank Ganesh K. Jimee, Amod M. Dixit, Surya N. Shresta, Ramesh ies, Louisiana, 2007.
Guragain, and Gopi K. Basyal for their insights and expert advice. Chang, L., Elnashai, A. S., and Spencer, B. F.: Post-earthquake
modelling of transportation networks, Struct. Infrastruct. Eng.,
Edited by: F. Guzzetti 8, 893–911, doi:10.1080/15732479.2011.574810, 2012.
Reviewed by: N. Kerle and two anonymous referees Chang, S. E. and Chamberlin, C.: Assessing the role of lifeline
systems in community disaster resilience, Res. Prog. Accompl.,
2003–2004, 87–94, 2003.
Chang, S., Pasion, C., Yavari, S., and Elwood, K.: Social Im-
References pacts of Lifeline Losses: Modeling Displaced Populations and
Health Care Functionality, in: Lifeline Earthquake Engineering
ABAG: Shaken awake! Estimates of uninhabitable dwelling units in a Multihazard Environment, edited by: Tang, A. K. K. and
and peak shelter populations in future earthquakes affecting the Werner, S., 563–572, American Society of Civil Engineers, Oak-
San Francisco Bay region, Association of Bay Area Govern- land, 2009.
ments, Oakland, 1996. Chien, S., Chen, L., Chang, S., Chiu, G., and Chu, C.: Development
Aubrecht, C., Özceylan, D., Steinnocher, K., and Freire, S.: of an after earthquake disaster shelter evaluation model, J. Chin.
Multi-level geospatial modeling of human exposure pat- Inst. Eng., 25, 591–596, doi:10.1080/02533839.2002.9670733,
terns and vulnerability indicators, Nat. Hazards, 68, 147–163, 2002.
doi:10.1007/s11069-012-0389-9, 2013. Chou, J.-S., Ou, Y.-C., Cheng, M.-Y., Cheng, M.-Y., and Lee, C.-
Balcik, B., Beamon, B. M., Krejci, C. C., Muramatsu, K. M., and M.: Emergency shelter capacity estimation by earthquake dam-
Ramirez, M.: Coordination in humanitarian relief chains: Prac- age analysis, Nat. Hazards, 65, 2031–2061, doi:10.1007/s11069-
tices, challenges and opportunities, Int. J. Prod. Econ., 126, 22– 012-0461-5, 2013.
34, doi:10.1016/j.ijpe.2009.09.008, 2010. Chu, J. Y. and Su, Y. P.: Comprehensive Evaluation In-
Bharosa, N., Lee, J., and Janssen, M.: Challenges and obstacles in dex System in the Application for Earthquake Emer-
sharing and coordinating information during multi-agency disas- gency Shelter Site, Adv. Mater. Res., 156–157, 79–83,
ter response: Propositions from field exercises, Inf. Syst. Front., doi:10.4028/www.scientific.net/AMR.156-157.79, 2011.
12, 49–65, doi:10.1007/s10796-009-9174-z, 2010. Coburn, A. and Spence, R.: Earthquake Risk Modelling, in Earth-
Caiado, G., Macário, R., and Oliveira, C. S.: A New Paradigm quake Protection, 311–352, John Wiley & Sons Ltd, Chichester,
in Urban Road Network Seismic Vulnerability: From a Link- doi:10.1002/0470855185.ch9, 2006.
by-link Structural Approach to an Integrated Functional As- Coburn, A. W., Spence, R. J. S., and Pomonis, A.: Factors deter-
sessment, edited by: Santos, M. A., Sousa, L., and Portela, mining human casualty levels in earthquakes: Mortality predic-
E., National Civil Engineering Laboratory, Lisbon, avail- tion in building collapse, in: Proceedings of the Tenth World
able at: http://www.iscramlive.org/ISCRAM2011/proceedings/ Conference Earthquake Engineering, Vol. 10, 5989–5994, A.A.
papers/221.pdf (last access: 9 July 2014), 2011. Balkema, Rotterdam, 1992.
Caiado, G., Oliveira, C. S., Ferreira, M. A., and Sá, F.: As- Cova, T. J. and Church, R. L.: Modelling community evacuation
sessing Urban Road Network Seismic Vulnerability: An Inte- vulnerability using GIS, Int. J. Geogr. Inf. Sci., 11, 763–784,
grated Approach, Indian Institute of Technology, Lisbon, avail- doi:10.1080/136588197242077, 1997.
able at: http://www.iitk.ac.in/nicee/wcee/article/WCEE2012_ Cova, T. J. and Johnson, J. P.: Microsimulation of neighborhood
1105.pdf (last access: 9 July 2014), 2012. evacuations in the urban – wildland interface, Environ. Plan. A,
Cavalieri, F., Franchin, P., Gehl, P., and Khazai, B.: Quantitative as- 34, 2211–2229, doi:10.1068/a34251, 2002.
sessment of social losses based on physical damage and interac- Crucitti, P., Latora, V., and Porta, S.: Centrality measures in
tion with infrastructural systems, Earthq. Eng. Struct. Dyn., 41, spatial networks of urban streets, Phys. Rev. E, 73, 036125,
1569–1589, doi:10.1002/eqe.2220, 2012. doi:10.1103/PhysRevE.73.036125, 2006.
CBS (Ed.): National Population Census 1991 (National Report), Daley, W. R., Karpati, A,. and Sheik, M.: Needs Assessment of
2nd Edn., Government of Nepal, National Planning Commission the Displaced Population Following the August 1999 Earthquake
Secretariat, Central Bureau of Statistics, Kathmandu, 1995. in Turkey, Disasters, 25, 67–75, doi:10.1111/1467-7717.00162,
CBS (Ed.): Population Census 2001 (National Report), His 2001.
Majesty’s Government of Nepal, National Planning Commis- Daniell, J. E., Khazai, B., Wenzel, F., and Vervaeck, A.: The CAT-
sion, Central Bureau of Statistics, Kathmandu, 2002. DAT damaging earthquakes database, Nat. Hazards Earth Syst.
Sci., 11, 2235–2251, doi:10.5194/nhess-11-2235-2011, 2011.

Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015 www.nat-hazards-earth-syst-sci.net/15/789/2015/

J. Anhorn and B. Khazai: Emergency shelter after an earthquake 801

Da Silva, J.: Quality and standards in post-disaster shelter, Struct. Haack, B. N. and Rafter, A.: Urban growth analysis and modeling
Eng., 85, 25–32, 2007. in the Kathmandu Valley, Nepal, Habitat Int., 30, 1056–1065,
Dixit, A. M.: Challenges of Building Code Implementation in doi:10.1016/j.habitatint.2005.12.001, 2006.
Nepal, in: From Code to Practice: Challenges for Building Code Harrald, B. F. and al Hajj, S. F.: Estimates for Demand for Mass
Implementation And the Further Direction of Housing Earth- Care Services in Future Earthquakes Affecting the San Francisco
quake Safety, edited by: Ando, S., Subedi, J. K., and Nakamura, Bay Region, George Washington University for The American
H., 61–66, UNCRD, available at: http://www.preventionweb.net/ Red Cross, 1992.
files/10591_HESITokyoPapers.pdf (last access: 13 May 2014), Indriasari, V., Mahmud, A. R., Ahmad, N., and Shariff, A.
2009. R. M.: Maximal service area problem for optimal siting of
Dixit, A., Dwelley-Samant, L., Nakarmi, M., Pradhanang, S. B., emergency facilities, Int. J. Geogr. Inf. Sci., 24, 213–230,
and Tucker, B. E.: The Kathmandu Valley Earthquake Risk doi:10.1080/13658810802549162, 2010.
Management Project: An Evaluation, in: Proceedings, 12th IOM and GoN: Report on Identification of Open Spaces for Hu-
World Conference on Earthquake Engineering, Auckland, New manitarian Purposes in Kathmandu Valley, International Organi-
Zealand, available at: http://www.iitk.ac.in/nicee/wcee/article/ zation for Migration, Kathmandu, 2012.
0788.pdf (last access: 28 August 2013), 2000. JICA and MoHA: The Study on Earthquake Disaster Mitigation in
Donohou, C.: Strategic Planning for Post-Earthquake Temporary the Kathmandu Valley, Kingdom of Nepal, Final Report, Japan
Housing: Best Practices, edited by: Hall, S. and Hapeman, K., International Cooperation Agency, Kathmandu, 2002.
Humanit, Briefs Spanning Field Relief Aid Dev., 16, 2012. Johnson, C.: Strategic planning for post-disaster temporary housing,
EMI: Risk-Sensitive Land Use Plan: Kathmandu Metropolitan City, Disasters, 31, 435–458, doi:10.1111/j.1467-7717.2007.01018.x,
Nepal, Earthquakes and Megacity Initiative (EMI), Manila, avail- 2007.
able at: http://www.preventionweb.net (last access: 3 June 2014), Johnson, C.: Planning for temporary housing, in: Rebuilding Af-
2010. ter Disasters?: From Emergency to Sustainability, edited by:
ESRI: ArcGIS Help 10.1, ArcGIS Help 101, available at: http:// Lizarralde, G., Johnson, C., and Davidson, C. H., 70–87, Taylor
resources.arcgis.com/en/help/main/10.1/index.html, last access: & Francis, Independence, 2009.
21 August 2013. Kar, B. and Hodgson, M. E.: A GIS-Based Model to Determine
Félix, D., Branco, J. M., and Feio, A.: Temporary housing after Site Suitability of Emergency Evacuation Shelters, Trans. GIS,
disasters: A state of the art survey, Habitat Int., 40, 136–141, 12, 227–248, doi:10.1111/j.1467-9671.2008.01097.x, 2008.
doi:10.1016/j.habitatint.2013.03.006, 2013. Khazai, B. and Hausler, E.: Intermediate Shelters in Bam and
FEMA: HAZUS99 Technical Manual: Earthquake Model, FEMA, Permanent Shelter Reconstruction in Villages Following the
Washington, D.C., 1999. 2003 Bam, Iran, Earthquake, Earthq. Spectra, 21, 487–511,
FEMA: Emergency temporary group housing site selection guide- doi:10.1193/1.2098907, 2005.
lines – Minimizing environmental/historic/safety problems, Khazai, B., Daniell, J. E., and Wenzel, F.: The March 2011 Japan
available at: http://www.fema.gov (last access: 13 August 2013), Earthquake: Analysis of Losses, Impacts, and Implications for
2007. the Understanding of Risks Posed by Extreme Events, Tech.-
FEMA: HAZUS-MH 2.1 Technical Manual: Earthquake Model, Teor. Prax., 20, 22–33, 2011.
available at: http://www.fema.gov (last access: 14 August 2013), Khazai, B., Daniell, J. E., Franchin, P., Cavalieri, F., Vangelsten, B.
2011. V., Lervolino, L., and Esposito, S.: A New Approach to Modeling
Franchin, P., Lupoi, A., and Pinto, P. E.: On the Role of Road Net- Post-Earthquake Shelter Demand: Integrating Social Vulnerabil-
works in Reducing Human Losses After Earthquakes, J. Earthq. ity in Systemic Seismic Vulnerability Analysis, in: Proceedings
Eng., 10, 195–206, doi:10.1080/13632460609350593, 2006. of the 15th World Conference on Earthquake Engineering, 2105–
Freire, S. and Aubrecht, C.: Integrating population dynamics into 2115, National Information Centre of Earthquake Engineering,
mapping human exposure to seismic hazard, Nat. Hazards Earth Lisbon, 2012.
Syst. Sci., 12, 3533–3543, doi:10.5194/nhess-12-3533-2012, Khazai, B., Argyroudis, S., Hancilar, U., Taucer, F., and Kakderi,
2012. K.: Guidelines for the consideration of socio-economic impacts
Gall, M.: Where to Go? Strategic Modelling of Access to in seismic risk analysis, SYNER-G Reference Report 5, Euro-
Emergency Shelters in Mozambique, Disasters, 28, 82–97, pean Commission Joint Research Centre, Institute for the Pro-
doi:10.1111/j.0361-3666.2004.00244.x, 2004. tection and Security of the Citizen, Ispra, available at: http:
Global Communities: Shelter and Settlements: Post-Disaster Re- //www.syner-g.eu (last access: 3 June 2014), 2013.
sponse in Urban Environments, available at: http://www. Khazai, B., Daniell, J. E., Düzgün, S., Kunz-Platt, T., and Wenzel,
globalcommunities.org/node/37242 (last access: 8 August 2013), F.: Framework for Systemic Socio-Economic Vulnerability and
2012. Loss Assessment, in: SYNER-G: Systemic Seismic Vulnerabil-
Guragain, R., Jimee, G. K., and Dixit, A. M.: Earthquake Aware- ity and Risk Assessment of Complex Urban, Utility, Liefeline
ness and Effective Planning Through Participatory Risk Assess- Systems and Critical Facilities, edited by: Pitilakis, K., Franchin,
ment: An Experience from Nepal, in: Proceedings of 14th World P., Khazai, B., and Wenzel, H., Springer, Heidelberg, 2014.
Conference on Earthquake Engineering, 12–17, NICEE, Bei- Kongsomsaksakul, S., Chen, A., and Yang, C.: Shelter Location-
jing, available at: http://www.iitk.ac.in/nicee/wcee/article/14_ Allocation Model for Flood Evacuation Planning, J. East. Asia
07-0086.pdf (last access: 28 August 2013), 2008. Soc. Transp. Stud., 6, 4237–4252, 2005.
Gutschow, N. and Kreutzmann, H.: Handlung schlägt Plan, Geogr. Kubanek, J., Nolte, E.-M., Taubenböck, H., Kappas, M., and Wen-
Rundsch., 4, 42–49, 2012. zel, F.: Modelling of population dynamics: GIS versus Remote

www.nat-hazards-earth-syst-sci.net/15/789/2015/ Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015

802 J. Anhorn and B. Khazai: Emergency shelter after an earthquake

sensing – a case study for Istanbul, in: Proceedings of TIEMS, Ritchie, L. A. and Tierney, K.: Temporary Housing Planning and
1–10, Beijing, China, 2010. Early Implementation in the 12 January 2010 Haiti Earthquake,
Lizarralde, G., Johnson, C., and Davidson, C. H. (Eds.): Rebuild- Earthq. Spectra, 27, 487–507, doi:10.1193/1.3637637, 2011.
ing After Disasters: From Emergency to Sustainability, Taylor & Samardjieva, E. and Badal, J.: Estimation of the Expected Number
Francis, Independence, 2009. of Casualties Caused by Strong Earthquakes, Bull. Seismol. Soc.
McEntire, D. A.: Disaster response and recovery: strategies and tac- Am., 92, 2310–2322, doi:10.1785/0120010112, 2002.
tics for resilience, Wiley, Hoboken, NJ, 2007. Sevtsuk, A. and Mekonnen, M.: Urban network analysis, Rev. Int.
Neis, P. and Zielstra, D.: Recent Developments and Future Géomatique, 22, 287–305, doi:10.3166/RIG.22.287-305, 2012.
Trends in Volunteered Geographic Information Research: SPHERE Project (Ed.): SPHERE Handbook: Humanitarian Char-
The Case of OpenStreetMap, Future Internet, 6, 76–106, ter and Minimum Standards in Disaster Response, 3rd Edn.,
doi:10.3390/fi6010076, 2014. SPHERE Project, Bourton on Dunsmore, available at: www.
Neis, P., Singler, P., and Zipf, A.: Collaborative mapping sphereproject.org (last access: 17 July 2013), 2011.
and Emergency Routing for Disaster Logistics-Case stud- Subedi, B. P.: Ethnic/caste diversification in Kathmandu metropoli-
ies from the Haiti earthquake and the UN portal for tan: Changing social landscape of a capital city, J. Geogr. Reg.
Afrika, in: Proceedings of the Geoinformatics Forum Plan., 3, 185–199, 2010.
Salzburg, Vol. 10, edited by: Car, A., Griesebner, G., Tai, C.-A., Lee, Y.-L., and Lin, C.-Y.: Urban Disaster Prevention
and Strobl, J., 239–248, Wichmann, Salzburg, available at: Shelter Location and Evacuation Behavior Analysis, J. Asian Ar-
http://koenigstuhl.geog.uni-heidelberg.de/publications/2010/ chit. Build. Eng., 9, 215–220, doi:10.3130/jaabe.9.215, 2010.
Neis/un-osm-emergency-routing.gi-forum2010.full.pdf (last Taubenböck, H.: Vulnerabilitätsabschätzung der erdbebengefährde-
access: 21 August 2013), 2010. ten Megacity Istanbul mit Methoden der Fernerkundung, Dis-
NRRC: Earthquake Scenario of Kathmandu Valley, Interview sertation, Universität Würzburg, Philosophische Fakultät I, In-
with Moira Reddick (Video), available at: http://youtu.be/ stitut für Geographie und Geologie, Würzburg, 4 July, available
HN3T7fYVJYk, last access: 12 August 2013. at: http://opus.bibliothek.uni-wuerzburg.de/volltexte/2008/2804/
NSET: Shelter Response Strategy and Plan for Earthquake Disasters (last access: 15 April 2012), 2008.
For Kathmandu Valley, Nepal, Draft Report, National Society for Taubenböck, H. and Strunz, G.: Widening a narrow road: remote
Earthquake Technology – Nepal, Kathmandu, 2010. sensing contributing to the multifaceted problem of earthquake
NSET: Shelter Response Strategy and Plan for Earthquake Disasters risk reduction, Nat. Hazards, 68, 1–5, doi:10.1007/s11069-013-
For Kathmandu Valley, Nepal, Final Report, National Society for 0604-3, 2013.
Earthquake Technology – Nepal, Kathmandu, 2012. Thapa, R. B. and Murayama, Y.: Examining Spatiotemporal Ur-
NSET and GeoHazards International (Eds.): The Kathmandu Valley banization Patterns in Kathmandu Valley, Nepal: Remote Sens-
Earthquake Risk Management Action Plan, National Society for ing and Spatial Metrics Approaches, Remote Sens., 1, 534–556,
Earthquake Technology – Nepal, Kathmandu, 1998. doi:10.3390/rs1030534, 2009.
Perry, R. W.: Evacuation decision-making in natural disasters, Mass Thapa, R. B., Murayama, Y., and Ale, S.: City Profile Kathmandu,
Emerg., 4, 25–38, 1979. Cities, 25, 45–57, doi:10.1016/j.cities.2007.10.001, 2008.
Perry, R. W. and Green, M. R.: The Role of Ethnicity in the Tierney, K. J., Lindell, M. K., and Perry, R. W.: Facing the Unex-
Emergency Decision-Making Process, Sociol. Inq., 52, 306–334, pected: Disaster Preparedness and Response in the United States,
doi:10.1111/j.1475-682X.1982.tb01257.x, 1982. The National Academies Press, Washington, D.C., 2001.
Pittore, M. and Wieland, M.: Toward a rapid probabilistic seismic Toregas, C., Swain, R., ReVelle, C., and Bergman, L.: The Loca-
vulnerability assessment using satellite and ground-based remote tion of Emergency Service Facilities, Oper. Res., 19, 1363–1373,
sensing, Nat. Hazards, 68, 115–145, doi:10.1007/s11069-012- doi:10.1287/opre.19.6.1363, 1971.
0475-z, 2013. UN DESA: World Urbanization Prospects The 2011 Revision,
Porta, S., Crucitti, P., and Latora, V.: The network analysis of urban available at: http://esa.un.org/unup/index.html (last access: 12
streets: A dual approach, Phys. Stat. Mech. Its Appl., 369, 853– August 2013), 2012.
866, doi:10.1016/j.physa.2005.12.063, 2006a. UNDRO (Ed.): Shelter after disaster: guidelines for assistance,
Porta, S., Crucitti, P., and Latora, V.: The network analysis of ur- United Nations, available at: http://sheltercentre.org (last access:
ban streets: A primal approach, Environ. Plan. B, 33, 705–725, 5 August 2013), 1982.
doi:10.1068/b32045, 2006b. UN-HABITAT: Water Movements in Patan with Reference to Tra-
Quarantelli, E. L.: Patterns of sheltering and housing ditional Stone Spouts, UN-HABITAT, Kathmandu, 2008.
in US disasters, Disaster Prev. Manag., 4, 43–53, UN ISDR and UN OCHA: Disaster preparedness for effective re-
doi:10.1108/09653569510088069, 1995. sponse: guidance and indicator package for implementing pri-
Rawls, C. G. and Turnquist, M. A.: Pre-positioning of emergency ority five of the Hyogo Framework, United Nations, New York,
supplies for disaster response, Transp. Res. Part B Methodol., Geneva, 2008.
44, 521–534, doi:10.1016/j.trb.2009.08.003, 2010. UN OCHA, DFID and Shelter Centre (Eds.): Shelter after disaster:
Riad, J. K., Norris, F. H., and Ruback, R. B.: Predicting Evacua- strategies for transitional settlement and reconstruction, United
tion in Two Major Disasters: Risk Perception, Social Influence, Nations, available at: http://sheltercentre.org (last access: 5 Au-
and Access to Resources, J. Appl. Soc. Psychol., 29, 918–934, gust 2013), 2010.
doi:10.1111/j.1559-1816.1999.tb00132.x, 1999. Upreti, N.: Comparative Analysis of Levels of Earthquake Risk Per-
ception in Kathmandu Valley in 1998 and 2009, Master’s Thesis,
Sikkim Manipal University, New Delhi, 2009.

Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015 www.nat-hazards-earth-syst-sci.net/15/789/2015/

J. Anhorn and B. Khazai: Emergency shelter after an earthquake 803

Weiser, A. and Zipf, A.: Web Service Orchestration of OGC Web Wright, K. and Johnston, D.: Post-earthquake Sheltering Needs:
Services for Disaster Management, in: Geomatics Solutions for How Loss of Structures and Services Affects Decision Making
Disaster Management, edited by: Li, J., Zlatanova, S., and Fab- for Evacuation, in: Proceedings NZSEE Conference, Vol. Ses-
bri, A. G., 239–254, Springer Berlin Heidelberg, 2007. sion 7B, Paper 31, New Zealand Society for Earthquake Engi-
Wieland, M., Pittore, M., Parolai, S., Zschau, J., Moldobekov, B., neering Inc., Wellington, available at: http://www.nzsee.org.nz/
and Begaliev, U.: Estimating building inventory for rapid seis- db/2010/ (last access: 3 June 2014), 2010.
mic vulnerability assessment: Towards an integrated approach
based on multi-source imaging, Soil Dyn. Earthq. Eng., 36, 70–
83, doi:10.1016/j.soildyn.2012.01.003, 2012.

www.nat-hazards-earth-syst-sci.net/15/789/2015/ Nat. Hazards Earth Syst. Sci., 15, 789–803, 2015