REVIEW ARTICLE
Challenges of analyzing multi-hazard risk: a review
Melanie S. Kappes
•
Margreth Keiler
•
Kirsten von Elverfeldt
•
Thomas Glade
Received: 14 September 2010 / Accepted: 9 July 2012 / Published online: 31 July 2012
Springer Science+Business Media B.V. 2012
Abstract Many areas of the world are prone to several natural hazards, and effective risk
reduction is only possible if all relevant threats are considered and analyzed. However, in
contrast to single-hazard analyses, the examination of multiple hazards poses a range of
additional challenges due to the differing characteristics of processes. This refers to the
assessment of the hazard level, as well as to the vulnerability toward distinct processes, and
to the arising risk level. As comparability of the single-hazard results is strongly needed, an
equivalent approach has to be chosen that allows to estimate the overall hazard and
consequent risk level as well as to rank threats. In addition, the visualization of a range of
natural hazards or risks is a challenging task since the high quantity of information has to
be depicted in a way that allows for easy and clear interpretation. The aim of this con-
tribution is to give an outline of the challenges each step of a multi-hazard (risk) analysis
poses and to present current studies and approaches that face these difficulties.
Keywords Multi-hazard risk Hazard Vulnerability Risk Hazard cascades
Hazard chains
1 Introduction
The use of the term multi-hazard is in most cases closely related to the objective of risk
reduction. For example, within international politics, one of the first references to this term
M. S. Kappes M. Keiler K. von Elverfeldt T. Glade
Department of Geography and Regional Research, University of Vienna, Vienna, Austria
M. S. Kappes (&)
The World Bank, 1818 H St. NW, Washington, DC 20433, USA
e-mail: kappes.melanie@googlemail.com
M. Keiler
Institute of Geography, University of Bern, Bern, Switzerland
K. von Elverfeldt
Department of Geography and Regional Science, Klagenfurt University, Klagenfurt, Austria
123
Nat Hazards (2012) 64:1925–1958
DOI 10.1007/s11069-012-0294-2
has been made in the Agenda 21 for sustainable development (UNEP 1992). This docu-
ment calls for ‘‘complete multi-hazard research’’ as a part of human settlement planning
and management in disaster-prone areas (UNEP 1992, paragraph 7.61). The term reappears
in the Johannesburg Plan in the context of ‘‘protecting and managing the natural resource
base of economic and social development’’ (UN 2002, p. 14). It refers to ‘‘[a]n integrated,
multi-hazard, inclusive approach to address vulnerability, risk assessment and disaster
management, including prevention, mitigation, preparedness, response and recovery’’ as
‘‘an essential element of a safer world in the twenty-first century’’ (UN 2002, p. 20). In the
following, the Hyogo Framework of Action (UN-ISDR 2005, p. 4) adopted this aspect and
suggests an ‘‘integrated, multi-hazard approach for disaster risk reduction […] into poli-
cies, planning and programming related to sustainable development, relief, rehabilitation,
and recovery activities in post-disaster and post-conflict situations in disaster-prone
countries.’’ Furthermore, also FEMA (1995) uses this term in the U.S. national mitigation
strategy with the goal to lower risks and reduce the effects of disasters due to natural
hazards by focusing on the application of multi-hazard building approaches in the design
and construction of buildings.
The awareness of the necessity to investigate and manage the whole range of natural
hazards that pose a risk to humans, assets, and societies continued to grow in the last. Thereby,
the identification of the risks to be taken into account is mostly based on a spatial approach,
that is, a certain area is considered and all threats within this zone are taken into account
(c.f. Greiving et al. 2006, p. 1). Hewitt and Burton (1971, p. 5) refer to this concept as the ‘‘all-
hazards-at-a-place’’ approach. According to DHS (2011, 1–7), all-hazards ‘‘encompasses all
conditions, environmental or manmade, that have the potential to cause injury, illness, or
death; damage to or loss of equipment, infrastructure services, or property; or social, eco-
nomic, or environmental functional degradation’’. Consequently, ‘‘a first definition of the
term multi-hazard in a risk reduction context could read as follows: the totality of relevant
hazards in a defined area’’ (Kappes 2011, pp. 6 & 7). However, whether a hazardous process
is relevant has to be defined according to the specific setting of the respective area and to the
objective of the study. For instance, Hewitt and Burton (1971) propose a cut-off point for the
hazard-related damages: Depending on the respective scale, a process is considered irrelevant
if it causes damages below a certain point. The larger the observed area, the higher is this cut-
off point. Another example is given by the European Commission (2011, p. 24) in their
guidelines for risk assessment and mapping. These guidelines propose a set of criteria for the
determination of all significant hazards at a national level. For example, those threats with an
annual probability of at least 1 % ‘‘and for which the consequences represent significant
potential impacts, i.e.: number of affected people greater than 50, economic and environ-
mental costs about € 100 million, and political/social impact considered significant or very
serious [need to be taken into account]. Where the likely impacts exceed a threshold of 0.6 %
of gross national income (GNI) also less likely hazards or risk scenarios should be considered
(e.g., volcanic eruptions, tsunamis)’’. In the context of spatial planning, Greiving et al. (2006)
and Greiving et al. (2006, p. 4) define relevant according to differing criteria and restrict the
set of considered processes to ‘‘hazards that are closely tied to certain areas that are especially
prone to a particular hazard,’’ whereby ubiquitous threats such as meteorite impacts are
excluded.
However, not all studies on multiple hazards share the aim of involving all relevant
processes of a defined area, but can rather be described as more-than-one-hazard approaches.
This is especially true for scientific studies and supposedly is, among multiple reasons, due to
the strict separation of disciplines (with all the consequences for differing terminology, partly
conflicting definitions, and approaches, etc.) that hamper multi-hazard studies (c.f. WMO 1999).
1926 Nat Hazards (2012) 64:1925–1958
123
Nevertheless, in certain contexts, the joint investigation of two or more processes is indis-
pensable, that is, whenever one hazard triggers a second process, for example, earthquakes
leading to landslides (e.g., Bommer and Rodrı
´
guez 2002; Keefer 2002; Lin et al. 2006; Chang
et al. 2007; Chang et al. 2007; Lee et al. 2008; Miles and Keefer 2009). Moreover, an event
may cause multiple threats such as a volcanic eruption resulting in lava flows, lahars, and ash
and lapilli fallout (e.g., Zuccaro et al. 2008; Thierry et al. 2008). Another reason for con-
sidering several hazards jointly are common characteristics, as, for example, within the
system RAMMS (RApid Mass MovementS) that spans snow avalanches, debris flows, and
rock fall (Christen et al. 2007). In the SEDAG project (SEDiment cascades in Alpine Geo-
systems), Wichmann and Becht (2003) focus on the sediment cascade, thereby considering
soil erosion, rock falls, full-depth avalanches, shallow landslides, and debris flows.
In summary, two approaches to multi-hazard can be distinguished. The first one is
primarily spatially oriented and aims at including all relevant hazards. The second, in
contrast, is primarily thematically defined.
One challenge related to multi-hazard risk
1
analyses is related to the fact that while for
many, if not most, single processes a multitude of well-established approaches is available
(please refer to review articles provided by Ancey et al. (2004) and Bru
¨
ndl et al. (2010) for
snow avalanches; Hunter et al. (2007) for river floods; Dai et al. (2002), Glade and Crozier
(2004), and Fell et al. (2005) for landslides; WMO (1999) for meteorological, volcanic,
and seismic hazards), much fewer studies analyze multiple hazards. In consequence,
experience with associated problems is rare, and also, standard approaches are not avail-
able. This is problematic, because multi-hazard risk analyses are not just the sum of single-
hazard risk examinations:
1. hazard characteristics differ, and thus also the methods to analyze them (c.f.
Carpignano et al. 2009),
2. hazards are related and influence each other. This results in phenomena often
described as hazard chains, cascades, etc. (c.f. Tarvainen et al. 2006; Marzocchi et al.
2009; Kappes et al. 2010),
3. natural processes exert diverging impacts on elements at risk, and methods to describe
vulnerability vary between hazards (c.f. Hufschmidt and Glade 2010; Papathoma-
Ko
¨
hle et al. 2011; Kappes et al. 2011), and
4. a variety of risk description and quantification measures exists and has to be adapted to
enable the comparison of multiple risks (c.f. Marzocchi et al. 2009; Marzocchi et al.
2012).
These issues are major challenges for the analysis of multi-hazard risks. Therefore, the
aim of this contribution is threefold: Firstly, it aims at detailing the difficulties and challenges
associated with multi-hazard risk analysis. Secondly, the objective is to give an overview of
existing approaches that meet these challenges. Thereby, not only multi-hazard strategies
with a focus on risk reduction are considered, but also those studies that deal with more-than-
one-hazard. These studies were incorporated since they provide profound insight into spe-
cific aspects that are mostly neglected by studies focusing on a very large number of pro-
cesses. And thirdly, the paper seeks to give a coherent overview of all steps in multi-hazard
analysis. Thus, this paper is structured accordingly: (1) the joint hazard analysis of multiple
natural threats, (2) the assessment of the physical vulnerability of elements at risk toward
1
The term multi-hazard risk refers to the risk arising from multiple hazards. By contrast, the term multi-risk
would relate to multiple risks such as economic, ecological, social, etc.
Nat Hazards (2012) 64:1925–1958 1927
123
multiple hazards,
2
(3) the analysis of risk arising from multiple natural hazards, combining
the aspects hazard and vulnerability of exposed elements at risk, and (4) the joint visuali-
zation of multiple hazards. For each step, the multi-hazard-specific challenges are described,
discussed, and current studies and approaches are presented. Thereby, this paper rather
provides a comprehensive introduction into the field of multi-hazard risk analyses and its
specificities, and the authors do not claim completeness. Furthermore, the main focus is on
the multi-hazard aspect rather than on the vulnerability and risk issue. Finally, exclusively
physical vulnerability is considered, while social vulnerability and other types of vulnera-
bility are not included in this article. The analysis of social or community vulnerability is a
topic in its own right and thus cannot be covered in this review. Hereby, it is by no means
intended to foster the rather outdated perception that nature is the ‘problem’ while engi-
neering measures are the solution. Accordingly, linking respective social and nature-sci-
entific approaches is yet another challenge that needs to be covered elsewhere.
The terms hazard, vulnerability, and risk exhibit multiple definitions and are described
by various authors and institutions that often refer to Varnes (1984) and UNDHA (1992).
To avoid confusion, definitions of the main terms are highlighted at the beginning of each
section. Furthermore, the described concepts are classified in qualitative, semiquantitative,
and quantitative approaches (c.f. Altenbach 1995; Borter 1999; DIN 2009):
Qualitative: Description in words (e.g., high, medium, and low) which relate to, or
involve quality or kind. Qualitative judgments rank in higher and lower without the
information on how much higher or lower and are commonly based on expert appraisals.
Semiquantitative: Description by means of a scale that consists of words or numbers.
This scale allows a relative ranking and provides a measure for how much more one
scenario contributes over the next.
Quantitative: Relates to or can be expressed in terms of quantities or totals. It allows the
determination of absolute values on a determined scale.
2 Challenges and current approaches in the field of multi-hazard risk analyses
According to Varnes (1984, p. 10), hazard is defined as the ‘‘probability of occurrence within
a specified period of time and within a given area of a potentially damaging phenomenon.’’ In
addition to the hazard aspect, risk involves the vulnerability of the elements at risk and is
established as the ‘‘expected degree of loss due to a particular natural phenomenon,’’
thus suggesting that the product of hazard and vulnerability of exposed elements at risks
(Varnes 1984, p. 10). This section focuses on the three steps of a multi-hazard risk analysis,
namely the analysis of multi-hazard (2.1), vulnerability of elements at risk for multiple
processes (2.2) and multi-hazard risk (2.3). Furthermore, examples of existing methods for
the joint investigation of multiple hazards are presented (2.4).
2.1 Multi-hazard analyses
Delmonaco et al. (2006b, p. 15) define multi-hazard analyses as the ‘‘[i]mplementation of
methodologies and approaches aimed at assessing and mapping the potential occurrence of
different types of natural hazards in a given area.’’ The employed methods ‘‘have to take
2
The exposure of elements at risk is not considered separately since this article focuses on the issues and
challenges that arise in multi-hazard context in contrast to single hazard analysis, and exposure does not
change.
1928 Nat Hazards (2012) 64:1925–1958
123
into account the characteristics of the single hazardous events […] as well as their mutual
interactions and interrelations (e.g., landslide induced earthquake, floods and landslides
triggered by extreme rainfall, natural disasters as secondary effect from main disaster
types)’’ (Delmonaco et al. 2006b, p. 15). This description indicates two major challenges:
(1) differing process characteristics so that it becomes difficult to compare multiple haz-
ards, and (2) the existence of relations and interactions between hazards. In the following,
both aspects are presented and current approaches are exemplified.
2.1.1 Comparability of hazards due to differing process characteristics
Within multi-hazard analyses, it is essential to assess the respective level of each threat of
the investigated multiple hazards. However, hazards ‘‘differ by their nature, intensity,
return period and by the effects they may have on exposed elements. […] Their magnitudes
are also measured in different ways, using different units of reference, for example, dis-
charge or inundation depth for floods, ground motion or macro-seismic intensity for seism’’
(Carpignano et al. 2009, p. 515). Thus, the principal difficulty in the comparison of
multiple hazards is the distinct reference units. An approach to overcome this problem is
the standardization to a common measure. Reviewing numerous studies (e.g., Heinimann
et al. 1998; Odeh Engineers, Inc 2001; Delmonaco et al. 2006b; El Morjani et al. 2007;
Bartel and Muller 2007; Thierry et al. 2008), two major standardization approaches can be
distinguished: (1) the classification of hazards (qualitative approach) and (2) the devel-
opment of indices (continuous, semiquantitative approach).
(1) The standardization by means of classification is the most frequently used approach to
enable the comparison of different hazards. Intensity and frequency thresholds are defined in
order to classify the respective hazards into a predefined number of hazard classes. In order to
determine thresholds, a framework of shared objectives or criteria describing the classes has
to be established. This assures the equivalence and comparability of, for example, high
earthquake and high flood hazard (Delmonaco et al. 2006a). At the same time, however, it
becomes very difficult or even impossible to compare information from different sources
since, most probably, different criteria were applied (Marzocchi et al. 2009). In the following,
an overview of a number of studies employing diverse classification schemes is presented.
With their study, Moran et al. (2004, p. 185; based on Heinimann et al. 1998; Fuchs
et al. 2001) aim at a ‘‘conceptual approach to natural hazard investigations in regions
lacking hazard zoning or where only rudimentary hazard assessments exist’’ to ‘‘identify
the risk potential at a regional scale’’ as ‘‘foundation for further detailed studies.’’ In order
to reach this aim, Moran et al. (2004) used a worst-case scenario for the regional scale
modeling of avalanches and rock falls. Due to the common basis (worst-case scenario), the
resulting areas of potential impact can be compared and jointly visualized. Additionally, by
overlay with elements at risk, the number of endangered buildings or affected road kilo-
meters by each process can be determined and compared.
One objective of the ARMONIA project (Applied Multi Risk Mapping of Natural Hazards
for Impact Assessment) was to define a ‘‘new harmonised methodology for integrated
management of data from different risk analysis approaches and set-up basic principles for a
EU directive on harmonized risk maps aimed at spatial planning’’ (Delmonaco et al. 2006b,
p. 5). In this context, a classification scheme was proposed for hazard intensities at a regional
scale. The hazard is classified in low, medium, and high intensity with regard to spatial
planning purposes (Table 1). Subsequently, the importance of hazards can be compared and
consequences for the spatial planning process can be defined (Delmonaco et al. 2006b).
Nat Hazards (2012) 64:1925–1958 1929
123
