Content uploaded by C.J. Van Westen
Author content
All content in this area was uploaded by C.J. Van Westen on Apr 16, 2019
Content may be subject to copyright.
Cees J. van Westen and Stefan Greiving
The earth is shaped by endogenic processes, caused by forces from within the earth,
resulting in hazardous events like earthquakes or volcanic eruptions, and exogenic
processes, caused by forces related to the earth’s atmosphere, hydrosphere, geosphere,
biosphere and cryosphere and their interactions. Anthropogenic activities have had
a very important inuence on a number of these processes, especially in the last
two hundred years, for instance through the increase of greenhouse gasses, leading
to global warming, but also through dramatic changes in the land cover and land
use, and overexploitation of scarce resources. The above mentioned processes from
endogenic, exogenic and anthropogenic nature may lead to potentially catastrophic
events, even in locations that may be far away. For instance earthquakes might trigger
landslides which may lead to landslide-dammed lakes that may break out and cause
ooding downstream. Or the dams of large reservoirs in mountains, constructed for
hydropower, irrigation or drinking water, may fail under an earthquake or extreme
rainfall event and cause a similar ood wave.
These potentially harmful events are called hazards. They pose a level of
threat to life, health, property, or environment. They may be classied in different
ways, for instance according to the main origin of the hazard in geophysical,
meteorological, hydrological, climatological, biological, extra-terrestrial and
technological (See Table 2.1, from Guha-Sapir et al. 2016). Such classications
are always a bit arbitrary, and several hazard types could be grouped in different
categories, e.g. landslides could be caused by earthquakes, extreme precipitation
and human interventions.
Hazards have a number of characteristics that should be understood in order to be
able to assess and subsequently reduce their potential damage. Hazards with certain
magnitudes may occur with certain frequencies, as small events may occur often,
and large events seldom. In order to be able to establish a magnitude-frequency
Chapter 2
Multi-hazard risk assessment
and decision making
32 Environmental Hazards Methodologies for Risk Assessment
Table 2.1 Classification of hazard types as used by the International Disaster Database EM-DAT (Guha-Sapir et al. 2016),
which is based on and adapted from the he IRDR Peril Classification and hazard Glossary (IRDR, 2014).
Main Group Main Sub-group Main Type Sub-Type
Natural Geophysical: A hazard originating from solid earth.
This term is used interchangeably with the term
geological hazard.
Earthquake Ground shaking, tsunami
Mass movement
Volcanic Ash fall, lahar, pyroclastic flow,
lava flow
Meteorological: A hazard caused by short-lived,
micro- to meso-scale extreme weather and
atmospheric conditions that last from minutes to days.
Storm Extra-tropical storm, tropical
storm, convective storm
Extreme
temperature
Cold wave, heat wave, severe
winter conditions
Fog
Hydrological: A hazard caused by the occurrence,
movement, and distribution of surface and subsurface
freshwater and saltwater.
Flood Coastal flood, riverine flood,
flash flood, ice jam flood.
Landslide Avalanche (snow, debris),
mudflow, rockfall
Wave action Rogue wave, seiche
Climatological: A hazard caused by long-lived,
meso- to macro-scale atmospheric processes
ranging from intra-seasonal to multi-decadal climate
variability.
Drought
Glacial Lake
outburst
Wildfire Forest Fire, land fire (bush,
pasture)
Multi-hazard risk assessment and decision making 33
Biological: A hazard caused by the exposure to living
organisms and their toxic substances or vector-
borne diseases that they may carry. Examples are
venomous wildlife and insects, poisonous plants, and
mosquitoes carrying disease-causing agents such as
parasites, bacteria, or viruses (e.g. malaria).
Epidemic Viral , bacterial, parasitic,
fungal, prion disease
Insect infestation Grasshopper, locust
Animal accident
Extraterrestrial: A hazard caused by asteroids,
meteoroids, and comets as they pass near-earth,
enter the Earth’s atmosphere, and/or strike the Earth,
and by changes in interplanetary conditions that
effect the Earth’s magnetosphere, ionosphere, and
thermosphere.
Impact
Space weather Energetic particles,
geomagnetic storm
Technologic al Industrial accident Chemical spills, collapse,
explosion, fire, gas-leak,
poisoning, radiation, other
Transport accident Air, road, rail, water
Miscellaneous accident Collapse, explosion, fire, other.
34 Environmental Hazards Methodologies for Risk Assessment
relationship for hazard events, it is generally necessary to collect historical data
(e.g. from seismographs, meteo-stations, stream gauges, historical archives, remote
sensing, eld investigations etc.) and carry out statistical analysis (e.g. using extreme
event analysis such as Gumbel analysis) (Van Westen etal. 2008). The magnitude
of the hazard gives an indication of the size of the event, or the energy released,
whereas the intensity of a hazard refers to the spatially varying effects. For example
earthquake magnitude refers to the energy released by the ruptured fault (e.g.
measured on the Richter scale) whereas the intensity refers to the amount of ground
shaking which varies with the distance to the epicentre (e.g. measured on Modied
Mercalli scale). The magnitude of oods may be measured as the discharge in the
main channel at the outlet of a watershed before leaving the mountainous area,
whereas the intensity may be measured as the water height or velocity which is
spatially distributed, and depends on the local terrain. For some types of hazards
there is no unique intensity scale dened, e.g. for landslides (Corominas etal. 2015).
These events may be potentially harmful to people, property, infrastructure,
economy and activities, but also to the environment, which are all grouped
together under the term Elements-at-risk or assets. Also the term exposure is used
to indicate those elements-at-risk that are subject to potential losses. Important
elements-at-risk that should be considered in analysing potential damage of hazards
are population, building stock, essential facilities and critical infrastructure. Critical
infrastructure consists of the primary physical structures, technical facilities
and systems, which are socially, economically or operationally essential to the
functioning of a society or community, both in routine circumstances and in the
extreme circumstances of an emergency (UN-ISDR, 2009). Elements-at-risk have a
certain level of vulnerability, which can be dened in a number of different ways. The
general denition is that vulnerability describes the characteristics and circumstances
of a community, system or asset that make it susceptible to the damaging effects
of a hazard (UN-ISDR, 2009). There are many aspects of vulnerability, related to
physical, social, economic, and environmental conditions (see for example Birkmann,
2006). When considering physical vulnerability only, it can be dened as the degree
of damage to an object (e.g. building) exposed to a given level of hazard intensity (e.g.
water height, ground shaking, and impact pressure).
2.1 RISK
Risk is dened as the probability of harmful consequences, or expected losses
(deaths, injuries, property, livelihoods, economic activity disrupted or environment
damaged) resulting from interactions between natural or human-induced hazards
and vulnerable conditions (UN-ISDR, 2009; EC, 2011). Risk can presented
conceptually with the following basic equation indicated in Figure 2.1. Risk can
presented conceptually with the following basic equation:
Risk Hazard Vulnerability Amount of elements at risk=× × -- (2.1)
Multi-hazard risk assessment and decision making 35
Figure 2.1 Schematic representation of risk as the multiplication of hazard,
vulnerability and quantification of the exposed elements-at-risk. The various
aspects of hazards, vulnerability and elements-at-risk and their interactions are
also indicated. This framework focuses on the analysis of physical losses, using
physical vulnerability data.
The equation given above is not only a conceptual one, but can also be actually
calculated with spatial data in a GIS to quantify risk from geomorphological
hazards. The way in which the amount of elements-at-risk are characterized (e.g.
as number of buildings, number of people, economic value) also denes the way
in which the risk is presented. Table 2.2 gives a more in-depth explanation of the
various components involved. In order to calculate the specic risk equation 2.1
can be modied in the following way:
RP P V A
ST L
=×××
( :Hs) ( :Hs) (Es|Hs) ES (2.2)
in which:
P(T:Hs) is the temporal (e.g. annual) probability of occurrence of a specic hazard
scenario (Hs) with a given return period in an area;
P(L:Hs) is the locational or spatial probability of occurrence of a specic hazard
scenario with a given return period in an area impacting the elements-at-risk;
V(Es|Hs) is the physical vulnerability, specied as the degree of damage to a specic
element-at-risk Es given the local intensity caused due to the occurrence of
hazard scenario HS;
36 Environmental Hazards Methodologies for Risk Assessment
AEs is the quantication of the specic type of element at risk evaluated (e.g.
number of buildings).
Table 2.2 Components of risk with definitions, equations and explanations.
Term Definition Equations & Explanation
Natural
hazard (H)
A potentially damaging physical
event, phenomenon or human
activity that may cause the
loss of life or injury, property
damage, social and economic
disruption or environmental
degradation. This event has a
probability of occurrence within
a specified period of time and
within a given area, and has a
given intensity.
P(T:HS) is the temporal
(e.g. annual) probability of
occurrence of a specific
hazard scenario (Hs) with
a given return period in an
area;
P(L:HS) is the locational
or spatial probability of
occurrence of a specific
hazard scenario with a
given return period in
an area impacting the
elements-at-risk
Elements- at-
risk (E)
Population, properties,
economic activities, including
public services, or any other
defined values exposed to
hazards in a given area”. Also
referred to as “assets”.
Es is a specific type of
elements-at-risk (e.g.
masonry buildings of 2
floors)
Vulnerability
(V)
The conditions determined
by physical, social, economic
and environmental factors or
processes, which increase the
susceptibility of a community to
the impact of hazards. Can be
subdivided in physical, social,
economical, and environmental
vulnerability.
V(Es|Hs) is the physical
vulnerability, specified as
the degrees of damage
to ES given the local
intensity caused due to
the occurrence of hazard
scenario HS
It is expressed on a scale
from 0 (no damage) to 1
(total loss)
Amount of
elements-at-
risk (AE)
Quantification of the elements-
at-risk either in numbers
(of buildings, people etc), in
monetary value (replacement
costs etc), area or perception
(importance of elements-at-risk).
AES is the quantification of
the specific type of element
at risk evaluated (e.g.
number of buildings)
Consequence
(C)
The expected losses (of which
the quantification type is
determined by AE) in a given
area as a result of a given
hazard scenario.
CS is the “specific
consequence”, or expected
losses of the specific
hazard scenario which is
the multiplication of VS × AES
Multi-hazard risk assessment and decision making 37
The term risk mapping is often used as being synonymous with risk analysis
in the overall framework of risk management. Risk assessments (and associated
risk mapping) include: a review of the technical characteristics of hazards
such as their location, intensity, frequency and probability; the analysis of
exposure and vulnerability including the physical social, health, economic and
environmental dimensions; and the evaluation of the effectiveness of prevailing
and alternative coping capacities in respect to likely risk scenarios (UN-ISDR,
2009; EC, 2011; ISO 31000). In the framework of natural hazards risk
assessment, the term risk mapping also indicates the importance of the spatial
aspects of risk assessment. All components of the risk equation (Figure 2.1)
are spatially varying and the risk assessment is carried out in order to express
the risk within certain areas. To be able to evaluate these components there
is a need to have spatially distributed information. Computerized systems for
the collection, management, analysis and dissemination of spatial information,
so-called Geographic Information Systems (GIS) are used to generate the data
on the various risk components, and to analyse the risk (OAS, 1991; Coppock,
Table 2.2 Components of risk with definitions, equations and explanations
(Continued).
Term Definition Equations & Explanation
Specific risk
(RS)
The expected losses in a
given area and period of time
(e.g. annual) for a specific
set of elements-at-risk as a
consequence of a specific
hazard scenario with a specific
return period.
RS = HS × VS × AES
RS = HS × CS
RS = P(T:Hs) × P(L:Hs)
× V(Es|Hs) × AES
Total r i sk (RT)The probability of harmful
consequences, or expected
losses (deaths, injuries,
property, livelihoods,
economic activity disrupted
or environment damaged)
resulting from interactions
between natural or human-
induced hazards and vulnerable
conditions in a given area and
time period.
It is calculated by first
analyzing all specific risks.
It is the integration of all
specific consequences over all
probabilities.
RT ≈ ∑(RT) = ∑(HS × VS × AES)
Or better:
RT = ∫(VS × AES)
– For all hazard types
– For all return periods
– For all types of
elements-at-risk.
It is normally obtained by
plotting consequences
against probabilities, and
constructing a risk curve.
The area below the curve is
the total risk.
38 Environmental Hazards Methodologies for Risk Assessment
1995; Cova, 1999; van Westen, 2013). Hazard data is generally the most difcult
to generate. For each hazard type (e.g. ooding, debris ow, rock fall) so-called
hazard scenarios should be dened, which are hazard events with a certain
magnitude/intensity/frequency relationship (e.g. ood depth maps for 10, 50 and
100 year return periods). Different types of modelling approaches are required
for the hazard scenario analysis, depending on the hazard type, scale of analysis,
availability of input data, and availability of models. Generally speaking a
separate analysis is required to determine the probability of occurrence for
a given magnitude of events, followed by an analysis of the initiation of the
hazard (e.g. hydrological modelling or landslide initiation modelling), and of the
run-out or spreading of the hazard (e.g. hydrodynamic modelling or landslide
run-out modelling). Overviews of hazard and risk assessment methods for
landslides for example can be found in Corominas etal. (2014), and for oods in
Prinos (2008). Elements-at-risk data are very often based on building footprint
maps, which represent the location of buildings, with attributes related to their
use, size, type and number of people during different periods of the year (e.g.
daytime, night-time). Remote Sensing is often used to extract these building
maps if existing cadastral maps are not available. For other elements-at-risk like
transportation infrastructure and land cover maps also remote sensing data are
used as important inputs. Vulnerability data is often collected in the form of
vulnerability curves, fragility curves or vulnerability matrices, which indicate
the relationship between the levels of damage to a particular type of element-at-
risk (e.g. single storey masonry building) given intensity levels of a particular
hazard type (e.g. debris ow impact pressure). Generation of vulnerability
curves is a complicated issue, as they can be generated empirically from past
damage event for which intensity and damage is available for many elements-at-
risk, or through numerical modelling (Roberts etal. 2009). Table 2.2 gives an
overview of the various components of risk.
Risk can be presented in a number of different ways, depending on the
objectives of the risk assessment (Birkmann, 2007). Risk can be expressed in
absolute or relative terms. Absolute population risk can be expressed as individual
risk (the annual probability of a single exposed person to be killed) or as societal
risk (the relation between the annual probability and the number of people that
could be killed). Absolute economic risk can be expressed in terms of Average
Annual Loss, Maximum Probable Loss, or other indices that are calculated from
a series of loss scenarios, each with a relation between frequency and expected
monetary losses (Jonkman etal. 2002). It is also possible to differentiate between
direct risk (which is the risk directly resulting from the impact of the hazard) and
indirect risk (which may occur later as a consequence of the direct impact). Some
examples of direct risk are the destruction of physical objects (e.g. buildings,
transportation infrastructure), and examples of indirect losses are loss of revenues
and economic production, disruption of transportation networks leading to longer
travel time etc. A signicant component of the losses are intangible (difcult
Multi-hazard risk assessment and decision making 39
or impossible to quantify), for example the societal or psychological impact of
disaster events.
2.2 MULTI-HAZARD RISK
One of the difcult issues in natural hazards risk assessment is how to
analyse the risk for more than one hazard in the same area, and the way they
interact. Figure 2.2 shows an illustration of how different sets- of triggering
factors can cause a number of different hazards. There are many factors
that contribute to the occurrence of hazardous phenomena, which are either
related to the environmental setting (topography, geomorphology, geology,
soils etc.) or to anthropogenic activities (e.g. deforestation, road construction,
tourism). Although these factors contribute to the occurrence of the hazardous
phenomena and therefore should be taken into account in the hazard and risk
assessment, they are not directly triggering the events. For these, there is a need
for triggering phenomena, which can be of meteorological or geophysical origin
(earthquakes, or volcanic eruptions). A generally accepted denition of multi-
hazard still does not exist. In practice, this term is often used to indicate all
relevant hazards that are present in a specic area, while in the scientic context
it frequently refers to “more than one hazard”. Likewise, the terminology that is
used to indicate the relations between hazards is unclear. Many authors speak of
interactions (Tarvainen etal. 2006; de Pippo etal. 2008; Marzocchi etal. 2009;
Zuccaro & Leone, 2011; European Comission, 2011), while others call them
chains (Shi, 2002), cascades (Delmonaco etal. 2006a; Carpignano etal. 2009;
Zuccaro & Leone, 2011; European Comission, 2011), domino effects (Luino,
2005; Delmonaco etal. 2006a; Perles & Cantarero, 2010; van Westen, 2010;
European Comission, 2011), compound hazards (Alexander, 2001) or coupled
events (Marzocchi etal. 2009).
Compared to single processes, standard approaches and methodological
frameworks for multi-hazard risk assessment are less common in the literature
(Kappes et al. 2012), which is related to the complex nature of the interaction
between the hazards, and the difculty to quantify these.
2.2.1 Independent events
The simplest approach is to consider that the hazards are independent and caused
by different triggers. This means that the expected losses from one hazard type
are independent from the losses expected from the other hazard type. If that is
the case, the risk can be calculated by adding the average annual losses for the
different types of hazard. This would be the case for example for earthquake
hazard and ood hazard. They have different triggering mechanisms, which
do not directly interact. Therefore, earthquake hazard is independent of ood
hazard and may be analysed separately. Also the risk may be analysed separately
and the resulting losses could be added. Other examples of independent hazard
40 Environmental Hazards Methodologies for Risk Assessment
are for instance technological hazards and ood hazards. Many of the existing
software tools for multi-hazard risk assessment (See section 2.9) deal with these
hazard independently, and sum up the losses. However, when these apparent
independences are examined in detail, the relation may be more complicated.
For instance, an earthquake may trigger landslides that may block a river
leading to ooding, which makes that the earthquake and ood risk cannot be
considered entirely independent. Even ooding and technological hazards cannot
be considered completely independent: during ood events there may be a higher
risk of technological accidents.
Figure 2.2 Schematic representation of multi-hazards interactions between the
main triggering events (volcanic eruptions, Earthquakes, Meteorological extremes,
and anthropogenic activities) and secondary hazards.
2.2.2 Coupled events
The second multi-hazard relationship is between different hazard types that are
triggered by the same triggering event. These are what we would call coupled
Multi-hazard risk assessment and decision making 41
events (Marzocchi et al. 2009). The temporal probability of occurrence of such
coupled events is the same as it is linked to the probability of occurrence of the
triggering mechanism. For analysing the spatial extent of the hazard, one should
take into account that when such coupled events occur in the same area and the
hazard footprints overlap, the processes will interact, and therefore the hazard
modelling for these events should be done simultaneously, which is still very
complicated. In order to assess the risk for these multi-hazards, the consequence
modelling should therefore be done using the combined hazard footprint areas,
but differentiating between the intensities of the various types of hazards and
using different vulnerability-intensity relationships. When the hazard analyses
are carried out separately, the consequences of the modelled scenarios cannot be
simply added up, as the intensity of combined hazards may be higher than the
sum of both or the same areas might be affected by both hazard types, leading
to overrepresentation of the losses, and double counting. Examples of such types
of coupled events is the effect of an earthquake on a snow-covered building (Lee
& Rosowsky, 2006) and the triggering of landslides by earthquakes occurring
simultaneously with ground shaking and liquefaction (Delmonaco et al. 2006b;
Marzocchi etal. 2009). Within multi-hazard risk assessment the best way to treat
coupled risk is to take the maximum of the risks that are coupled. For example,
during the same tropical storm a village may be hit by ash oods or debris ows.
Once it is hit by one type there is damage, and buildings cannot be destroyed twice
during the same event.
2.2.3 One hazard changes conditions for the next
A third type of interrelations is the inuence one hazard exerts on the disposition
of a second hazard, though without triggering it (Kappes et al. 2010). An
example is the “re-ood cycle” (Cannon & De Graff, 2009): forest res alter the
susceptibility to debris ows and ash oods due to their effect on the vegetation
and soil properties. This problem highlights the fact that the conditions that make
certain areas more susceptible to hazards may change constantly. For instance,
land cover and land use have a large effect on hydro-meteorological hazards, such
as ooding and landslides. When these change as a result of other hazards (like
forest res), also the susceptibility to landslides, debris ows or oods increases.
Many of the hazard relations are of this type. For instance, volcanic eruptions
may lead to the deposition of volcanic ash, which will increase the susceptibility
to landslides and ooding. Earthquakes may trigger landslides, and the landslide
scars that are unvegetated may lead to increased erosion and debris ows. It is very
difcult to take this type of relationship into account before one particular hazard
has changed the conditions that make the terrain more susceptible to the second
hazard. The practice is to update a multi-hazard risk assessment each time after
the occurrence of a major hazard event (like a volcanic eruption, major earthquake
or hurricane).
42 Environmental Hazards Methodologies for Risk Assessment
2.2.4 Domino or cascading hazards
The fourth type of hazard relationships consists of those that occur in chains: one
hazard causes the next. These are also called domino effects, concatenated, or
cascading hazards. These are the most problematic types to analyse in a multi-
hazard risk assessment. Hazard may occur in sequence, where one hazard may
trigger the next. These hazard chains or domino effects are extremely difcult to
quantify over certain areas, although good results have been obtained at a local
level (e.g. Peila & Guardini, 2008). The best approach for analysing such hazard
chains is to use so-called event-trees (See section 2.3.2). However, it is often very
difcult to apply such event-trees in a spatial manner, where in fact different parts
of an area may require different event-trees. This is true for instance for the chain:
earthquake-landslide-damming-dam break ood. Each part of the terrain has a
different susceptibility to landslides. But also each earthquake, which a given
depth and magnitude, may trigger different landslide patterns. If a landslide may
be generated, the next step is to evaluate whether the size is large enough to dam a
river. This also depends on the location of the landslide with respect to the river, the
width of the river and the river discharge. Once the river is dammed it depends on
the type of material in the dam and the strength of the river, whether the landslide
dam is broken fast or whether there is a possibility for a lake to develop, which may
cause more severe ooding when the dam breaks later. This sequence is described
by Fan etal. (2012).
2.2.5 Example of multi-hazard chain: Layou Valley
landslides in Dominica, Caribbean
The Layou-Carholm landslide, located on the Caribbean island of Dominica,
represents an example of a multi-hazard situation that achieved climactic
proportions in 1997 and 2011. The Layou River, with a length of 17 km is one
of the largest watersheds in Dominica (70 km2) and drains about 9% of the land
(ACOE, 2004). The Layou Tuff forms vertical walls along the lower Matthieu and
Layou Rivers through these reaches. This welded tuff resulted from ignimbrite
eruptions in the Late Pleistocene (Roobol & Smith, 2004). Landslides were
common in the area, with specic reports occurring between 1987 and 1997. There
is an eyewitness account of a slide following Hurricane Hugo in 1989 and also
following Hurricanes Iris, Luis and Marilyn in 1995. There was a major change to
the pattern of small landslides. Dramatic slumping occurred between November
18 and 25, 1997. Two major slides blocked the river and created a natural dam. The
dam was breached on November 21 with mudows reaching the sea accompanied
by extensive ooding of the lower river valley. The larger of the Layou ood
events which happened on November 28, 1997, measured 1,325,000 m3. A wall
of material estimated at 50 feet high was washed downstream. The riverbed rose
dramatically in its lower reaches. This elevation was estimated at 10 meters at the
Multi-hazard risk assessment and decision making 43
location of the bridge. The river had dried up between 18 and 20 November 1997
and then ooded on November 21. Further landslides occurred on November 25,
1997 and October 8 and 11, 1998 with subsequent dam breaks being signicant
events. End-of-the-year measurements show that the lake depth increased from
22 m in 1998 to nearly 40 m in 2008 (DeGraff etal. 2010). The maximum volume
estimate is 3,611,985 m3, assuming failure by overtopping and complete draining
of the lake (Breheny, 2007). A major dam break event occurred on 28/06/2011. The
road along the Layou River to Pont Casse was closed, due to ood hazard. R Also
in later years the Layou valley was heavily affected by oods and landslides. In
August 2015, during tropical storm Erika, severe ooding damaged the road in a
number of places (Figure 2.3).
Figure 2.3 Mathieu landslide dam development. (a) Carholm landslide blocking
the Mathieu River and forming a lake in 1997 (Satellite image from 3-8-2005), (b)
Google Earth image from 21-12-2012 after the breaching of the landslide dam in
2011 (c) View of the river valley just below the breaching point. (d) Downstream
part of the valley where the road was washed out by heavy flooding on 27 August
2015 during Tropical Storm Erika.
Table 2.3 shows the main multi-hazard relationships for a number of hazards
occurring in the Caribbean countries.
44 Environmental Hazards Methodologies for Risk Assessment
Table 2.3 Main hazard types and their interactions. The relationship should be read starting from the left and reading
horizontally (Source: www.charim.net).
Earthquake Volcanic
Eruption
Tsunami Storm
Surge
River
Flooding
Landslides Forest Fires
Earthquake –Independent Chain Independent Independent Chain Independent
Volcanic eruption Independent – Chain Independent Disposition Disposition Chain
Tsunami Caused by Caused by – Independent Independent Chain along
coast
Independent
Storm surge Independent Independent Independent – Chain Chain Independent
River flooding Independent Independent Independent Coupled – Coupled Independent
Landslides Caused by Independent Independent Coupled Coupled –
Forest Fires Independent Coupled Independent Independent Disposition Disposition –
Multi-hazard risk assessment and decision making 45
2.3 RISK ANALYSIS APPROACHES
Risk assessment is a process to determine the probability of losses by analysing
potential hazards and evaluating existing conditions of vulnerability that could pose
a threat or harm to property, people, livelihoods and the environment on which they
depend (UN-ISDR, 2009). ISO 31000 (2009) denes risk assessment as a process
made up of three processes: risk identication, risk analysis, and risk evaluation.
Risk identication is the process that is used to nd, recognize, and describe the
risks that could affect the achievement of objectives. Risk analysis is the process
that is used to understand the nature, sources, and causes of the risks that have
been identied and to estimate the level of risk. It is also used to study impacts and
consequences and to examine the controls that currently exist. Risk evaluation is
the process that is used to compare risk analysis results with risk criteria in order to
determine whether or not a specied level of risk is acceptable or tolerable.
Risk mapping for natural hazard risk can be carried out at a number of scales
and for different purposes. Table 2.4 and Figure 2.4 give a summary. In the
following sections four methods of risk mapping will be discussed: Quantitative
risk assessment (QRA), Event-Tree Analysis (ETA), Risk matrix approach (RMA)
and Indicator-based approach (IBA).
Table 2.4 Indication of scales of analysis with associated objectives and data
characteristics.
Scale of
Analysis
Scale Possible Objectives Possible
Approaches
International,
Global
<1:1 million Prioritization of countries/
regions; Early warning
Simplified
RMA & IBA
Small: provincial
to national scale
<1:100,000 Prioritization of regions;
Analysis of triggering events;
Implementation of national
programs; Strategic environmental
assessment; Insurance
Simplified
EVA, RMA &
IBA
Medium:
municipality to
provincial level
1:100000 to
1:25000
Analyzing the effect of changes;
Analysis of triggering events;
Regional development plans
RMA/IBA
Local: community
to municipality
1:25000 to
1:5000
Land use zoning; Analyzing the
effect of changes; Environmental
Impact Assessments; Design of
risk reduction measures
QRA/EVA/
RMA IBA
Site-specific 1:5000 or
larger
Design of risk reduction
measures; Early warning systems;
detailed land use zoning
QRA/EVA/
RMA
Approaches: QRA = Quantitative risk assessment, EVA = Event-Tree Analysis, RMA = Risk
matrix approach, IBA = Indicator-based approach.
46 Environmental Hazards Methodologies for Risk Assessment
Figure 2.4 Components relevant for risk assessment, and the four major types of risk mapping that are presented in this section.
Multi-hazard risk assessment and decision making 47
2.3.1 Quantitative risk assessment
If the various components of the risk equation can be spatially quantied for a
given set of hazard scenarios and elements-at-risk, the risk can be analysed using
the following equation:
Risk (HS)
= 0
= 1
(HS)
All E
=×××
()
()
∫PPAV
T
P
P
S
T
T
||||(ER HS) (ER HS)
aaRAll hazards ∑∑ ⎛
⎝
⎜⎞
⎠
⎟
⎛
⎝
⎜
⎜
⎞
⎠
⎟
⎟
(2.3)
In which:
P(THS) = the temporal probability of a certain hazard scenario (HS). A hazard
scenario is a hazard event of a certain type (e.g. ooding) with a certain
magnitude and frequency;
P(SHS) = the spatial probability that a particular location is affected given a
certain hazard scenario;
A(ERHS) = the quantication of the amount of exposed elements-at-risk, given
a certain hazard scenario (e.g. number of people, number of buildings,
monetary values, hectares of land) and
V(ERHS) = the vulnerability of elements at risk given the hazard intensity under
the specic hazard scenario (as a value between 0 and 1).
The method is schematically indicated in Figure 2.5. GIS operations are used to
analyse the exposure as the intersection between the elements-at-risk and the hazard
footprint area for each hazard scenario. For each element-at-risk also the level
of intensity is recorded through a GIS-overlay operation. These intensity values
are used in combination with the element-at-risk type to nd the corresponding
vulnerability curve, which is then used as a lookup table to nd the vulnerability
value. The way in which the amount of elements-at-risk are characterized (e.g.
as number of buildings, number of people, economic value) also denes the
way in which the risk is calculated. The multiplication of exposed amounts and
vulnerability should be done for all elements-at-risk for the same hazard scenario.
The results are multiplied with the spatial probability that the hazard footprint
actually intersects with the element-at-risk for the given hazard scenario P(SHS) to
account for uncertainties in the hazard modelling. The resulting value represents
the losses, which are plotted against the temporal probability of occurrence for the
same hazard scenario in a so-called risk curve. This is repeated for all available
hazard scenarios. At least three individual scenarios should be used, although it
is preferred to use at least 6 events with different return periods (FEMA, 2004)
to better represent the risk curve. The area under the curve is then calculated by
integrating all losses with their respective annual probabilities. It is possible to
create risk curves for the entire study area, or for different spatial units, such as
administrative units, census tracks, road or railway sections etc.
48 Environmental Hazards Methodologies for Risk Assessment
Figure 2.5 Schematic representation of Quantitative Risk Assessment.
The components that are involved in risk assessment have a high degree of
uncertainty. Aleatory uncertainty is associated with the variation of the input data
used in the risk assessment. For example the variations in soil characteristics used
to model landslide probability, surface characteristics, building characteristics
etc. These are normally incorporated in probabilistic risk analysis (Bedford &
Cook, 2001), which calculates thousands of hazard and risk scenarios taking
the variations of the input factors and calculating exceedance probabilities
using techniques, such as Monte Carlo simulation. Epistemic uncertainty refers
to uncertainty associated with incomplete or imperfect knowledge about the
processes involved, and lack of sufcient data. This is often a serious problem as
there may not be enough data available to determine individual hazard scenarios,
or there are no vulnerability curves for the types of elements-at-risk within the
study area. Probabilistic risk assessment takes into account all possible hazard
scenarios and the uncertainty of the input factors, by running thousands of loss
scenarios, and calculate eventually the loss exceedance curve. For a number
Multi-hazard risk assessment and decision making 49
of hazards, such as landslides or ooding, it is very complicated to develop a
large number of hazard scenarios due to the large epistemic uncertainty caused
by lack of data. In such cases uncertainty can be taken into account using the
method illustrated in Figure 2.6. In this method data are used showing the range
of possible values for the temporal probability, spatial probability, intensity of
the hazard, value of the elements-at-risk and vulnerability. The uncertainty
range in the temporal probability of the hazard scenario is reected by a range
of possible values on the Y-axis of the risk curve. The uncertainty in the hazard
intensity (e.g. water height for ooding, impact pressure for landslides) combined
with the uncertainty in the vulnerability curve will results in larger uncertainty
ranges in vulnerability, which are then multiplied with the uncertainty range
of the quantication of elements-at-risk (e.g. building costs). This then gives a
range of values for the expected losses. Thus, instead of a single point in the risk
curve, each hazard scenario will result in a rectangle, dened by the variation in
probability and losses. The upper right corners of the rectangle are connected to
provide the most pessimistic risk curve, and the lower right corners are connected
to provide the most optimistic risk curve. When calculating the area under the
curves it is then possible to show the range in annual expected losses.
Figure 2.6 Method for including uncertainty in Quantitative Risk Analysis in cases
where it is not possible to define many hazard scenarios.
50 Environmental Hazards Methodologies for Risk Assessment
Figure 2.7 gives an example of a quantitative risk assessment. In this simple
example we are taking a ood situation. The gure shows a cross section through
a ood plain. There are three hazard scenarios, which have been modelled using a
ood model. They have different return periods (10 years, 20 years and 50 years).
In this simple example there are 3 elements at risk only (buildings) that are of
two types. Building A and building B are wooden and relatively weak buildings.
They have also lower replacement values. They are located in different elevations.
Building C is a concrete building, which is stronger. It is also located at a higher
elevation than building B. It is also larger and more expensive. In the exposure
analysis, there is overlaying the ood heights with the building heights and the water
height is calculated for each hazard scenario and for each building. For the 10 years
return period, building A is not ooded, and building C only 0.1 meter. For the 20
year return period, all buildings are ooded, but with different degrees. For the 50
year return period, all buildings are ooded, building B and C very much. For the
vulnerability analysis, there is a need for vulnerability curves, which are related
for each type of building the degree of loss to a building with a given water height.
These curves are generated from past event damage assessment, by correlating the
water height with damage. For example: building B has an exposed intensity of 5.6
for the scenario of 20 years return period. And it is a wooden building so we take the
value of 5.6 on the X-axis of the vulnerability curve, representing the ood depth.
Because it is a wooden building, the curve for the wooden buildings is looked up, and
then the damage value on the y-axis is read. This is done for all buildings and for all
return periods. The replacement values (amount) are lled in, and the replacement
values (amount) are multiplied with the vulnerability to calculate losses. The losses
for the buildings are summed up for the same hazard scenario (return period). The
annual probability is calculated: 1 divided by the return period. The probability
is plotted for each scenario against the losses, and t a curve through the points,
which links all probabilities with all losses. The area below the curve represents
the Average Annual Losses. It is the integration of all losses over all probabilities.
2.3.2 Event-tree approaches
As mentioned in section 2.2.4 a number of hazards may occur in chains: these are
also called domino effects, or concatenated hazards. These are the most problematic
types to analyse in a multi-hazard risk assessment. The best approach for analysing
such hazard chains is to use so-called event-trees. An event tree is a system,
which is applied to analyse all the combinations (and the associated probability of
occurrence) of the parameters that affect the system under analysis. All the analysed
events are linked to each other by means of nodes (See Figure 2.4) all possible states
of the system are considered at each node and each state (branch of the event tree)
is characterized by a dened value of probability of occurrence. Figure 2.8 gives an
example of an event tree for a situation where a rockfall in a lake may trigger a ood
wave that would impact a village (from Lacasse etal. 2008).
Multi-hazard risk assessment and decision making 51
Figure 2.7 Schematic presentation of the steps involved in quantitative risk
analysis. See text for explanation.
AQ13
52 Environmental Hazards Methodologies for Risk Assessment
Figure 2.8 Bayesian Event tree for tsunami propagation, given that rock slide in Åknes has occurred (V = rockslide volume,
R = run-up height). From Lacasse et al. (2008).
Multi-hazard risk assessment and decision making 53
2.3.3 Risk matrix approach
Risk assessments are often complex and do not allow to develop a full numerical
approach, since many aspects are not fully quantiable or have a very large degree
of uncertainty. This may be related to the difculty to dene hazard scenarios, map
and characterize the elements-at-risk, or dene the vulnerability using vulnerability
curves. In order to overcome these problems, the risk is often assessed using the
so-called risk matrices or consequences-frequency matrices (CFM), which are
diagrams with consequence and frequency classes on the axes (See Figure 2.4).
They permit to classify risks based on expert knowledge with limited quantitative
data (Haimes, 2008; Jaboyedoff etal. 2014). The risk matrix is made of classes of
frequency of the hazardous events on one axis, and the consequences (or expected
losses) on the other axis. Instead of using xed values, the use of classes allows
for more exibility and incorporation of expert opinion. Such methods have
been applied extensively in natural hazard risk assessment, e.g. in Switzerland
(Figure2.9 from Jaboyedoff etal. 2014). This approach also permits to visualize
the effects and consequences of risk reduction measures and to give a framework
to understand risk assessment. The system depends on the quality of the group
of experts that are formed to identify the hazard scenarios, and that carry out the
hazard ltering and ranking in several sub-stages characterized by frequency
(probability) and impact classes and their corresponding limits (Haimes, 2008).
Figure 2.9 Example of potential building area in a high hazard area and illustration
of the proposed solutions. The risk matrix is used to represent the degree of
risk. The scope of tolerable risk (yellow) is between the limits of tolerance and
of acceptability. The initial situation 0 is a combination of very high frequency of
debris flows with a high impact. After construction of a deflection dike or wall the
frequency doesn’t change but the impact decreases considerably. The areas Z1
and Z2 on the other hand will get a higher frequency of occurrence and higher
consequence as a result of the mitigation works (Jaboyedoff et al. 2014).
54 Environmental Hazards Methodologies for Risk Assessment
2.3.4 Indicator-based approach
There are many situations where the (semi)-quantitative methods for risk mapping
are not appropriate. This could be because the data are lacking to be able to quantify
the components, such as hazard frequency, intensity, and physical vulnerability.
For instance, when the risk assessment is carried out over large areas, or in areas
with limited data. Another reason is that one would like to take into account a
number of different components of vulnerability that are not incorporated
in (semi-) quantitative methods, such as social vulnerability, environmental
vulnerability and capacity. In those cases, it is common to follow an indicator-
based approach to measure risk and vulnerability through selected comparative
indicators in a quantitative way in order to be able to compare different areas or
communities. The process of disaster risk assessment is divided into a number
of components, such as hazard, exposure, vulnerability and capacity (See Figure
2.4), through the so-called criteria tree, which list the subdivision into objectives,
sub-objectives and indicators (Figure 2.11). Data for each of these indicators are
collected at a particular spatial level, for instance by administrative units. These
indicators are then standardized (e.g. by reclassifying them between 0 and 1),
weighted internally within a sub-objective and then the various sub-objectives are
also weighted amongst themselves. Although the individual indicators normally
consist of quantitative data (e.g. population statistics), the resulting vulnerability,
hazard and risk results are scaled between 0 and 1. These relative data allows to
compare the indicators for the various administrative units. These methods can
be carried out at different levels, ranging from local communities (e.g. Bollin &
Hidajat, 2006) cities (Greiving etal. 2006) to countries (Van Westen etal. 2012).
The approaches are mostly based on the development of the so-called risk
indices, and on the use of spatial multi-criteria evaluation. One of the rst attempts
to develop global risk indicators was done through the Hotspots project (Dilley
et al. 2005). In a report for the Inter-American Development Bank, Cardona
(2005) proposed different sets of complex indicators for benchmarking countries
in different periods and to make cross-national comparisons. Peduzzi et al.
(2005, 2009) have developed global indicators, not on the basis of administrative
units, but based on gridded maps. The Disaster Risk Index (UN-ISDR, 2005b)
combines both the total number and the percentage of killed people per country in
large- and medium-scale disasters associated with droughts, oods, cyclones and
earthquakes. In the DRI, countries are indexed for each hazard type according to
their degree of physical exposure, their degree of relative vulnerability, and their
degree of risk. Also at local scale, risk indices are used, often in combination
with spatial multi criteria evaluation (SMCE). Castellanos and Van Westen (2007)
present an example of the use of SMCE for the generation of a landslide risk index
for the country of Cuba, generated by combining a hazard index and a vulnerability
index. Van Westen etal. (2012) developed a similar approach for national scale
vulnerability and risk assessment for the country of Georgia (Figure 2.10 and
Multi-hazard risk assessment and decision making 55
Figure 2.11). The hazard index is made using indicator maps related to triggering
factors (earthquakes and rainfall) and environmental factors. The vulnerability
index was made using ve key indicators: housing condition and transportation
(physical vulnerability indicators), population (social vulnerability indicator),
production (economic vulnerability indicator) and protected areas (environmental
vulnerability indicator). The indicators were based on polygons related to political-
administrative areas, which are mostly at municipal level. Each indicator was
processed, analysed and standardized according to its contribution to hazard and
vulnerability. The indicators were weighted using direct, pair wise comparison
and rank ordering weighting methods and weights were combined to obtain the
nal landslide risk index map. The results were analysed per physiographic region
and administrative units at provincial and municipal levels. Another example at
the local level is presented by Villagrán de León (2006), which incorporates 3
dimensions of vulnerability, the scale or geographical level (from human being to
national level), the various sectors of society, and 6 components of vulnerability.
The method uses matrices to calculate a vulnerability index, which was grouped in
qualitative classes (high, medium and low).
Figure 2.10 National scale multi-hazard risk assessment using an indicator-based
approach for the country of Georgia (Caucasus). The method is developed in a
web-based platform (http://drm.cenn.org/index.php/en/) and a risk atlas (https://
issuu.com/grammallc/docs/atlas_of_risk?pageNumber=1&e=5243266/2932778).
56 Environmental Hazards Methodologies for Risk Assessment
Figure 2.11 National scale multi-hazard risk assessment using an indicator-
based approach for the country of Georgia (Caucasus). A multi-hazard map was
generated using a hazard criteria tree, which was combined with a vulnerability
criteria tree. The vulnerability criteria tree is composed of a number of main groups:
physical vulnerability, social vulnerability, environmental vulnerability and economic
vulnerability. These are subdivided into subgroups and eventually in a number of
spatial indicators, which are standardized and weighted. (Van Westen et al. 2012).
Multi-hazard risk assessment and decision making 57
2.4 RISK ANALYSIS AND DECISION MAKING:
A CASE STUDY
After the introduction parts, the second half of this chapter deals with the
application of risk analysis in decision making. This is done through illustrating
the procedures with a hypothetical case study.
The overall aim of the second part of this chapter is to illustrate how to analyse
possible changes in risk to multi-hazards. These changes may be related to
possible risk reduction measures, but also to possible future scenarios related to
land use change, population change, and climate change, and the effect of possible
int er vention a lternat ives on t op of t hese p ossible future scenario s. This i s illustrated
in Figure 2.12, which has four possible workows:
(A) Analysing the current level of risk. In this workow the stakeholders
(e.g. local authorities) are interested to know the current level of risk in
their municipality. They request expert organizations to provide them
with hazard maps, asset maps, and vulnerability information, and use this
information in risk modelling. They use the results in order to carry out a
risk evaluation.
(B) Analysing the best alternatives for risk reduction. In this workflow
the stakeholders want to analyse the best risk reduction alternative, or
combination of alternatives. They define the alternatives, and request the
expert organizations to provide them with updated hazard maps, assets
information and vulnerability information reflecting the consequences
of these alternatives. Once these hazard and asset maps are available
for the alternatives, the new risk level is analysed, and compared with
the existing risk level to estimate the level of risk reduction. This is
then evaluated against the costs (both in terms of finances as well as
in terms of other constraints) and the best risk reduction scenario is
selected.
(C) The evaluation of the consequences of possible future scenarios.
Possible future scenarios can be formulated that project possible changes
related to climate, land use change or population change due to global and
regional changes, and which are only partially under the control of the local
planning organizations. Stakeholders would like to evaluate the effect of
these changes on the hazard and assets (again here the updated maps should
be provided by expert organizations) and how these would translate into
different risk levels.
(D) The evaluation how different risk reduction alternatives will perform
under different future scenarios. (trends of climate change, land use
change and population change). This is the most complicated workow,
as it requires to calculate the present risk level, the effect of different risk
reduction alternatives, and the overprinting of these on the scenarios. For
58 Environmental Hazards Methodologies for Risk Assessment
each of these combinations of alternatives & scenarios new hazard, assets
and risk maps need to be made.
These four different workows will be presented more in detail in the coming
sections. First, the study area and case study data set will be presented.
Figure 2.12 Four situations where risk assessment is used in decision making. (a)
Analysing the current level of risk. (b) Analysing the best risk reduction alternative;
(c) Analysing changing risk due to possible future scenarios. (d) Analysing the best
“change-proof” alternative: the alternative that behaves best under possible future
scenarios. See text for more explanation.
2.4.1 The case study data set
The case study data set was based on an original dataset from Nocera Inferiore,
located between Naples and Salerno in Italy. This data set was prepared for the EU
FP7 project SAFELAND (2011a, 2011b). The procedure and results for the hazard
Multi-hazard risk assessment and decision making 59
and risk assessment of the existing situation were reported by Ferlisi etal. (2016).
The various risk reduction alternatives and the involvement of various stakeholders
in the selection of the optimal measures was presented by Linnerooth-Bayer etal.
(2016). Narasimhan et al. (2016) presented results on Cost-Benet Analysis for
some of the risk reduction measures. The Nocera dataset is used as a start, but it
has modied the data and methods so that it allows to show the various procedures
for multi-hazard risk assessment dealing with a hypothetical case study of a
mountainous slope along the coast of a Caribbean island.
Therefore, the original hazard maps are modied to reect the situation for
the various alternatives, and scenarios, and additional tsunami hazard maps are
added. Building attributes were modied by us, and all land parcel maps have been
generated by ourselves based on available high resolution images. It was decided
to choose a hypothetical case because of the difculty in getting the right data for
achieving all objectives in the same study area. This analysis requires local scale
hazard intensity maps, detailed element-at-risk maps, vulnerability curves, risk
reduction alternatives and future scenarios. Many of these data are still not available
for a single area. Nevertheless, it is hoped that by following the examples in this
chapter, readers get a better idea of the procedure and can eventually also apply it
in their own situation, once data is available. There is also a series of GIS exercises,
with descriptions, open source GIS software and a spatial dataset, available that
follow the procedures step-by-step and downloaded from http://www.charim.net/
use/41. The available maps are illustrated in Figure 2.13 and 2.14, respectively.
Figure 2.13 Hazard intensity maps for 4 hazard maps (debrisflows, floods,
landslides and tsunami) and 3 return periods.
60 Environmental Hazards Methodologies for Risk Assessment
2.4.2 Hazard input data
The case study area is a mountainous area along the coast, with steep forested
slopes in the south part, and a fault-related mountain front with triangular facets,
from which landslides may be triggered. The steep slopes have a number of steep
gulleys, and landslides may form in the upper parts that could lead to debris ows.
During heavy rainfall also ash oods may occur in these gulleys, which may affect
the atter areas in the north. The atter northern part of the area has agricultural,
and residential areas. The area is affected by landslides, debris ows, ash oods
and possibly also by tsunami. Frequency estimation of hazardous events was carried
out based on the analysis of historical hazard data and rainfall data, and hazard
assessment for landslides, oods and debris ows were carried out for different return
periods using various models, such as OpenLISEM, SEEP/W, Slope/W, TRIGRS
and FLO-2d (Ferlisi et al. 2016). The debris ow, ood and tsunami maps have
intensity data (impact pressure for debris ows, and depth for ooding and tsunami).
The landslide hazard maps do not have intensity maps, but only spatial probability
maps indicating the chance that a particular area will be affected by a landslide.
The hazard maps have a specic code, which consist of the following
components:
• Hazard type: the name of the hazard map start with two characters
referring to the hazard type: LS = landslides, MF = mudow, FL = Flood,
DF = Debris ows, TS = Tsunami
• Return Period: this is the average frequency with which the events are
expected to occur. This is based on the analysis of the magnitude and
frequency of the triggering rainfall, or of the events themselves (e.g. ood
discharge, or the number of landslides occurring in a particular period). The
following return periods were used: 20, 50 and 100 years. For some specic
situation also longer return periods (200 years) were used.
• Intensity: the intensity indicates the spatially distributed effect of the hazard
event. This can be water depth (DE) for ooding or tsunami, or impact
pressure (IP) for debris ows. These have been modelled using specic
hazard modelling software. These models require quite a lot of input data
and assumptions. In this chapter we will not deal with the methods how the
hazard maps were created. For some types of hazards (e.g. landslides) it may
also not be possible to generate intensity maps, as data or models are lacking.
• Spatial probability: the spatial probability indicates the chance that a
particular location would actually be affected by the hazard. This could be
the result of uncertainty in the ood modelling or runout modelling. Or it
could also represent (in the absence of an intensity map) the likelihood that a
particular area will be affected by landslides based on the area of the units,
divided by the area of landslides that have occurred in the past. In this way,
it can be used to reclassify the so-called landslide susceptibility maps into
spatial probability maps.
Multi-hazard risk assessment and decision making 61
2.4.3 Input data: elements-at-risk
It is possible to use four types of elements-at-risk maps: building footprints, land
parcels (land use related units), line elements (e.g. transportation lines) and point
elements (e.g. individual objects). In the case study, the work involves only building
footprints and land parcels as elements-at-risk types (Figure 2.14). Each of them
have information on:
• Land use: land use is one of the most important characteristics and it is used
to model a number of other attributes (e.g. population).
• Element-at-risk types: for land parcels this would be the same as the land
use type (e.g. forest, residential, commercial etc.) and for building footprints
this would be the construction types. Different types of elements-at-risk
can be affected differently by hazard events. For the risk analysis this is
important as this provides the link to vulnerability curves, which will be
explained later.
• Value: this is the replacement value of the elements-at-risk in monetary units
(Euros, US dollars etc.). These are mostly estimated by multiplying unit costs
(e.g. per m2) with the area of land parcels or oor space for building footprints.
• People: the number of people that might be present in the element-at-risk.
Here it is relevant to decide whether to take the maximum number of people
or the people present at a given time, or to use specic population scenarios
(in case of dealing with rapid events, the time of day/year is also important
for the population loss estimation).
Figure 2.14 Elements-at-risk maps: building footprints and land parcels with
related attributes, and administrative units.
The building footprint map and the land parcel maps have the same number
of people for the parcels in which buildings are located. For the other parcels,
population values per m2 are used and multiplied these with the area of the land
parcel, so that an estimate can be obtained of the total number of people per
parcel. For the economic assessment, the values of the buildings are taken from
the building footprint map, and used these for the value of the land parcels. For the
parcels without buildings, an estimation is conducted based on the value per m2
and multiplied this with the area.
62 Environmental Hazards Methodologies for Risk Assessment
2.4.4 Input data: vulnerability curves
Vulnerability curves are also very important components in the risk analysis. A
vulnerability curve expresses the relation between the hazard intensity (e.g. water
depth) and the degree of damage which is expressed between 0 and 1 for a specic
type of element-at-risk. Vulnerability curves are derived from past disaster events
by correlating observed intensities with observed damage and deriving average
regression lines from these. Vulnerability curves may also be derived through
computer modelling (e.g. nite element models where a particular building is exposed
to a particular intensity and the effect is calculated) or through expert opinion. For
this case study, a number of vulnerability curves have been developed for all the
combinations of the hazard intensity types and the elements-at-risk types. Existing
curves have been used for the literature, but there is a need for a lot of changes,
as the curves were not for all of the units. For the analysis these curves should be
implemented in GIS, as tables (See Figure 2.15). Curves have been developed for
buildings, and land parcels, and separate curves for the physical losses (required for
the economic risk analysis) and for the population losses (people killed). The codes
in the table indicate the various aspects of the vulnerability curves. For example:
VUL_FL_DE_LP_PH: Vulnerability curves for ooding, expressed in water depth,
for land parcels, and showing the physical vulnerability.
2.4.5 Input data: administrative units
For the calculation of risk, there is also need for an administrative unit map,
as there is going to be an aggregation of the losses for particular units, and the
decision making is based on the risk within these units. The administrative unit
map contains 19 administrative units (Figure 2.14).
2.5 ANALYSING THE CURRENT LEVEL OF RISK
This section will further present and discuss the workow for multi-hazard risk
analysis of the current situation, illustrated in Figure 2.12A.
2.5.1 Stakeholders
Central in the whole process are the stakeholders. The envisaged users of the
system are organizations involved in spatial planning, planning of risk reduction
measures, or emergency preparedness and response. They work in a country
with a specic legislation and planning process organizations that have different
mandates. These could be subdivided into:
• Government departments responsible for the construction, monitoring,
maintenance and protection of critical infrastructure (e.g. the Ministry of
Public Works). Their mandate is to:
Multi-hazard risk assessment and decision making 63
Figure 2.15 Vulnerability curves and related vulnerability tables. Here only some examples are shown.
AQ14
64 Environmental Hazards Methodologies for Risk Assessment
{ Plan the (re)location of critical infrastructure (roads, buildings and other
critical infrastructure);
{ Design guidelines for construction of roads and buildings in potentially
dangerous areas;
{ Design of structural and non-structural mitigation measures against
ooding and landslides and other hazards;
{ Design of non-structural mitigation measures (e.g. watershed
management);
• Physical planning departments responsible with the mandate to make land
development plans at different scales. Their mandate is to:
{ Develop national or regional physical development plans;
{ Develop local development plans;
{ Develop guidelines for building construction in hazard areas;
{ Evaluate relocation options for settlements in endangered areas;
{ Develop zoning maps with relevant hazard information for building
control;
{ Provide relevant hazard and risk information for land subdivision;
• National Emergency Management Organizations with the mandate to:
{ Design disaster response plans;
{ Organize volunteers for disaster response planning;
{ Develop and management Early Warning systems;
{ Shelter planning and management.
2.5.2 Hazard modelling and elements-at-risk/
vulnerability assessment
Most government organizations normally have a few persons capable of visualizing
spatial data using GIS, but are not sufciently capable of carrying out the actual
spatial hazard and risk analysis required as the basis for their work. Therefore,
they will work with external consultants that will carry out this type of analysis
for them, and they have to specify the exact Terms of Reference of the work of the
consultants. These consultants may work on hazard modelling and the generation
of elements-at-risk maps. This should be done for a specic scale of analysis:
• National scale (for the entire country) with output at a scale of 1:25.000–
1:1.000.000 depending on the size of a country;
• Local scale (for specic areas) with output at a scale of 1:5.000 to 1:10.000
for specic planning areas (such as settlement areas). Our case study is at
local scale;
• Site-investigation scale (1:1000) for specic problem sites.
Part of this work should be done by the government organizations themselves
as they have the mandate to collect spatial information on the elements-at-risk.
For instance the Public works Department should develop a spatial database
Multi-hazard risk assessment and decision making 65
of the roads including all relevant characteristics related to road type, culverts,
bridges, drainage, road cuts and embankment lls, and slope stabilization works.
The Department of Physical Planning, together with other relevant Departments
is responsible for collecting and maintaining a national building database with
the relevant characteristics of buildings. In the establishment of these databases
external consultants may play an important role, however on the long run for
the respective government organizations would have the mandate to maintain
and update these databases, e.g. using web-based GIS solutions (e.g. http://www.
charim-geonode.net/).
2.5.3 Risk analysis
The crucial stage in the analysis of risk, which use the available information
toestimate the risk to people, property, or the environment, from hazards. Also
this type of work generally is not carried out by the government organizations
themselves, but rather by consultants, that have the right expertise to carry out
this type of analysis for one or more types of hazards, in combination with one or
more types of elements-at risk. This work is done at the appropriate scale related to
the objectives of the stakeholders. The risk assessment can be subdivided into the
following components:
• Exposure analysis. In this analysis hazard scenarios worked out by the
hazard assessment consultants for different return periods (e.g. once in 20,
50 and 100 years) are combined spatially with the elements-at-risk and the
number of elements-at-risk exposed to a certain hazard intensity is calculated.
Also for each element-at-risk the (maximum) intensity is calculated given a
certain return period.
• Vulnerability analysis. The results from the exposure analysis in terms of
the maximum intensity per return period are then used in combination with
vulnerability curves or matrices for the respective elements-at-risk types.
Through the vulnerability curves a translation is made from the intensities of
the hazard to the expected degree of loss for the elements at risk.
• Loss analysis. The results from the vulnerability assessment are then used
in combination with the quantication of the elements-at-risk to calculate the
expected losses. In the case of economic losses, the replacements costs for the
elements-at-risk are used, resulting in specic losses per return period. In the
case of population losses the number of people are used in combination with
the population vulnerability resulting from the vulnerability assessment.
• Risk assessment. The resulting losses for different return periods are
summed up for given administrative units if the hazards are independent
and integrated for the different return periods to provide the average
annual losses. These are used as the basis for the risk evaluation and for the
formulation of possible risk mitigation measures.
66 Environmental Hazards Methodologies for Risk Assessment
Figure 2.16 gives a detailed overview of the steps involved in loss analysis and
risk analysis. The Loss analysis has to be done for each combination of a hazard
map (for a given return period) and an elements-at-risk map, and for population
and/or economic losses. Each loss estimation requires a number of steps, and
doing this repetitive analysis manually is very time consuming. Therefore, an
automated script is developed in GIS, which combines a number of calculations
and operations, and uses parameters. Figure 2.16 shows the input screen of the
script. Some of these (alternative, scenario and future year) will be discussed later.
This makes that the same script can be used in many different situations. It follows
a number of steps:
• Overlay the element-at-risk map with the hazard intensity map;
• The resulting cross table (joint frequency table) contains all combinations of
the land parcel/building code and the intensity values (e.g. water depths). The
results are classied, according to the classication of the hazard intensity,
so that the result is in the form of classes, which can be used to join with the
vulnerability tables;
• As land parcels are sometimes large and only part of them might be actually
exposed to hazard intensity the script calculates the losses rst for the parts
of the land parcels with the same intensity;
• In order to know which fraction of each land parcel has a certain intensity,
the script reads in in the area of the whole land parcel from the attribute table
of the land parcel map, and then calculates the fraction of the land parcel
(Area of the unit in the joint frequency table divided by the area of the entire
land parcel);
• The joint frequency table is joined with the attribute table of the land
parcels/buildings and the amount column (either value or people) is entered,
depending on the input provided by the user. Also the land use is joined;
• The joint frequency table is then joined with the vulnerability table (of the
hazard type indicated) and the vulnerability values for all lands use types
are read;
• The vulnerability for each record is calculated by taking the vulnerability
value of the column that has the same land use code as in the record;
• A column is calculated that has an indication whether is dealing with a
spatial probability map. This is done by creating a column SPCheck and
then checking if the entered value is SP (Spatial Probability) or not. If this is
the case, the spatial probability is used, otherwise a value of 1;
• The script then calculates the loss by multiplying the amount × vulnerability
× spatial probability;
• In order to bring back the information at the level of the land parcels, the
script aggregates the loss for the land parcels;
• The script also aggregates the loss for the administrative units , and for the
entire study area;
Multi-hazard risk assessment and decision making 67
Figure 2.16 Procedure for loss analysis of all combinations of hazard scenarios
for specific return periods with elements-at-risk, and the combination in the
subsequent risk analysis. The following abbreviations are used: FL = Flooding,
DF = Debris flow, LS = Landslides, TS = Tsunami, DE = Water Depth, IP = Impact
Pressure, SP = Spatial probability, BU = Building footprints, LP = Land parcels. The
two input screen are those used in the loss analysis and risk analysis.
The loss analysis results in a loss database, with the calculated losses for each
combination of hazard type, return period, and elements-at-risk type. The user can
then select which combinations of losses are used in the subsequent risk analysis.
The risk analysis can be carried out if the loss estimation was done for at least three
return periods of the hazard type(s) that were selected.
The risk analysis can also be done using a script, that uses a number of
parameters, and which has the following steps:
• The risk type is determined: either economic risk or population risk is
calculated;
68 Environmental Hazards Methodologies for Risk Assessment
• The types of hazard and their dependencies are selected. Here the user can
interactively select which hazard types are taken into account. This allows to
carry out single risk or multi-hazard risk. In the case of multi-hazard risk, the
user has to evaluate whether the hazard types belong to the same triggering
event group (See section on Multi-hazard risk in this chapter). For example,
tsunami hazard is independent of the other three hazards, whereas ooding
and debris ows can occur in the same area, and are triggering by the same
event (extreme rainfall).
• Based on the dependencies that are dened, the script will either sum up
the losses for different independent hazards, or take the maximum for
dependent hazards in the same group, to avoid double counting of losses.
This is done at the lowest aggregation level: the individual land parcels or
building footprints;
• After that the multi-hazard losses are aggregated for the administrative units;
• Annual probability is calculated from the return periods, and the losses are
plotted against the probability in loss curves (for economic risk) and FN
curves (for population).
• Average annual loss is calculated for economic risk using the equation
presented in Figure 2.7.
2.5.4 Risk evaluation
The results of the risk analysis for the case study area are presented in Figure
2.17, both for land parcels, as well as for building footprints. As can be seen from
this gure the results differ when we use only building footprints, or land parcels.
Economic losses are lower for buildings, because the risk analysis for land parcels
also takes into account many other assets (e.g. roads, agriculture, and forest). Also
the population risk for land parcels is larger, as it takes into account the persons
not present in buildings. After estimating the risk, it is important to determine
whether the risk is too high, and where the risk is too high. This is called the risk
evaluation stage, and is the stage at which values and judgements enter the decision
process, explicitly or implicitly, by including consideration of the importance
of the estimated risks and the associated social, environmental, and economic
consequences, in order to identify a range of alternatives for managing the risks.
Important considerations in this respect are:
• Risk perception among stakeholders. Risk perception is about how
individuals, communities, or authorities perceive/judge/evaluate/rank the
level of risk, in relation to many factors;
• Risk communication. An important component in determining the risk
perception is the communication between the stakeholders of the levels of
risk. Do government organizations actively involve other stakeholders in the
consultation process of the actual level of risk and the possible risk reduction
measures?
Multi-hazard risk assessment and decision making 69
• Risk acceptability. An acceptable risk is a risk, which the society or
impacted individuals are prepared to accept. Actions to further reduce such
risk are usually not required unless reasonably practicable measures are
available at low cost in terms of money, time and effort. The denition of
acceptability levels is a responsibility of the national or local government
in a country. Risk acceptability depends on many factors, and differs from
country to country. Therefore, it is also not possible to simply export them
to other countries. Risk acceptability levels are generally done on this basis
of individual risk levels or societal risk levels (using so-called F-N curves),
some of which are shown in Figure 2.14, based on Ho (2009).
Figure 2.17 Results of the multi-hazard risk analysis for the existing situation in
the case study area. Left: economic risk. Right: societal risk with risk acceptability
threshold of the United Kingdom (UK) and the Netherlands (NL). Results are shown
for two different elements-at-risk data sets: land parcels and building footprints.
Below the risk acceptability thresholds for a number of different countries are
presented (Source: Ho, 2009).
70 Environmental Hazards Methodologies for Risk Assessment
2.6 ANALYSING THE BEST PLANNING ALTERNATIVE
Once the multi-hazard risk for the current situation is considered unacceptable,
which is the case in the case study, a new workow is introduced for the evaluation
of optimal risk reduction measures (Figure 2.12B).
2.6.1 Defining possible planning alternatives
In this workow the stakeholders want to analyse the best planning alternative,
or combination of alternatives. They dene the alternatives, and request the
expert organizations to provide them with updated hazard maps, elements-at-
risk information and vulnerability information reecting the consequences of
these alternatives. Once these hazard and asset maps are available, the new
risk level is analysed, and compared with the existing risk level to estimate the
level of risk reduction. This is then evaluated against the costs (both in terms
of nances, as well as in terms of other constraints) and the best risk reduction
scenario is selected. The planning of risk reduction measures (alternatives)
involves:
• Disaster response planning: focusing on analysing the effect of certain
hazard scenarios in terms of number of people, buildings and infrastructure
affected. It can also be used as a basis for the design of early warning
systems;
• Planning of risk reduction measures: which can be engineering measures
(such as dikes, check-dams, sediment catchment basins), but also non-
structural measures, such as relocation planning, strengthening/protection
of existing buildings etc.;
• Spatial planning: focusing on where and what types of activities are
planned and preventing that future development areas are exposed to natural
hazards.
The methods for risk reduction planning can be subdivided into:
• Structural measures refer to any physical construction to reduce or avoid
possible impacts of hazards, which include engineering measures and
construction of hazard-resistant and protective structures and infrastructure.
The strategy is to modify or reduce the hazard;
• Non-structural measures refer to policies, awareness, knowledge
development, public commitment, and methods and operating practices,
including participatory mechanisms and the provision of information,
which can reduce risk and related impacts. With the aim of modifying the
susceptibility of hazard damage and disruption and/or modifying the impact
of hazards on individuals and the community.
Multi-hazard risk assessment and decision making 71
The planning alternatives that are evaluated may be designed without considering
the possible impact of hazard and risk, and in these situation the analysis is carried
out to evaluate the impact of the different alternatives on the hazard and risk (will
it increase or decrease).
There are mostly different planning alternatives that can be formulated, and
each of them may have advantages and disadvantages. The aim of this analysis to
quantify their advantages and disadvantages in terms of hazard and risk reduction,
and to evaluate these against the costs for implementation through a cost-benet
analysis. Also other criteria that cannot be quantied can be used in deciding the
best alternative, using a multi-criteria evaluation.
For example, for the case study are the stakeholders dened three possible sets
of risk reduction measures, which are presented in Table 2.5 and Figure 2.18.
Table 2.5 Characterization of risk reduction alternatives in the case study area.
The alternatives are also shown in Figure 2.18.
Alternative 1 Alternative 2 Alternative 3
Name Engineering
Alternative
Eco-DRR
Alternative
Relocation
Alternative
Characteristics • Constructing
six retention
basins,
designed
to retain
maximum 100
year events;
• Soil removal
work in
selected areas
for slope
stabilization;
• Expropriation
of land and
existing
buildings
where
construction
will take place;
• Early Warning
System;
• Planting of
protection forest
at the foot of the
steep slope and
on fans of the
side streams;
• Construction
of drainage
channels and 10
water tanks;
• Expropriation of
land and existing
buildings where
construction will
take place
• Soil removal
work in selected
areas for slope
stabilization;
• Relocation of
buildings with
people in the
most dangerous
sectors;
• Compensation
of owners of
buildings or
construction of
new buildings
elsewhere;
• Expropriation
of existing
buildings
• Lawsuits of
residents that
do not want to
leave the area;
• Early Warning
System for
other areas;
Time of
construction
3 years 2 years 4–10 years,
depending on the
resistance
(Continued)
72 Environmental Hazards Methodologies for Risk Assessment
Table 2.5 Characterization of risk reduction alternatives in the case study area.
The alternatives are also shown in Figure 2.18 (Continued).
Alternative 1 Alternative 2 Alternative 3
Name Engineering
Alternative
Eco-DRR
Alternative
Relocation
Alternative
Year w h en
benefit starts
4th year From third year
increasing to
maximum after 15
years
From fourth year
increasing until the
tenth year
Annual
maintenance
Cleaning retention
basins, Slope
monitoring. Starts
at fourth year
Forest management,
cleaning water tanks
and channels, slope
monitoring. Starts in
third year.
Monitoring of
areas, prevention
of illegal
resettlements.
Starts at fourth
year.
Does the
hazard
change?
Yes, new
hazard maps
are needed for
landslides,
debris flows and
floods. Events
with return
periods of 100
years or less
are retained.
Higher return
periods may
overtop the
structures. No
effect on tsunami
Yes, new hazard
maps are needed
for landslides, debris
flows and floods.
Smaller events are
completely retained
by protection forest,
and larger events
are retained, but
there is still a certain
hazard. No effect on
tsunami
No.
Do the
elements-at-
risk change?
Only those
units where the
retention basins
are constructed
Only those
land parcels
where the protection
forest will be
constructed.
Yes, the
relocation areas
will change. A new
land parcel map is
needed.
Items to
estimate costs
Construction
costs retention
basin: unit
costs × number;
Soil removal: unit
costs × area;
Expropriation
costs: unit
cost × area;
Tree planting: unit
cost × area;
Soil removal: unit
costs × area;
Water tanks: unit
cost × number;
Expropriation costs:
unit cost × area;
Number of buildings
to relocate;
Compensation per
building;
Number of lawsuits
and value of legal
costs;
Early warning
system;
Multi-hazard risk assessment and decision making 73
2.6.2 Re-analysing hazards and elements-at-risk
The implementation of certain structural or non-structural risk mitigation measures
might lead to a modication of the hazard, exposure and vulnerability. Risk is a
function of Hazard × vulnerability of exposed elements-at-risk × the quantication
of the elements-at-risk (Figure 2.1). Therefore, there are several possibilities that
risk mitigation measures will inuence:
• The hazard. In terms of the probability (or return period) of specic hazard
events, the spatial distribution of the hazard and the intensity of the hazards.
For instance, the construction of a ood wall along a river may reduce the
area that will be ooded. For certain lower return periods the ood wall may
retain all ood water, and therefore the intensity (ood height) outside of the
ood wall will become zero for these return periods. For more extreme events
the ood intensity may become lower as a result of the ood wall. Therefore,
it is required to re-analyse the hazard given the implementation of the risk
reduction measure.
• The exposure of elements-at-risk. The number of elements-at-risk might
change as a result of the risk mitigation measure, or planning alternative. For
instance, if one of the alternatives involves relocation, the number of exposed
elements-at-risk will decrease, whereas the hazard might stay the same. In
other planning alternatives the effect of future development on the number of
exposed elements-at-risk might also be evaluated.
• The vulnerability of the elements-at-risk. The type of elements-at-risk
might change as a result of implementing the planning alternative. For
instance, when retrotting is considered, the number of elements-at-risk
might be the same, as well as the hazard, but the vulnerability of these
elements-at-risk might decrease, leading to a lower risk level. The same can
be said for the implementation of an Early Warning System. It will decrease
the number of exposed people, but also their vulnerability, if they would
move to shelters where they are better protected.
Figure 2.18 Three sets of risk reduction measures presented as three alternatives.
The characteristics of these alternatives are given in Table 2.5.
74 Environmental Hazards Methodologies for Risk Assessment
• The quantication of the elements-at-risk. might change. This might refer
to planning alternatives where the value of the exposed elements-at-risk
changes, e.g. they could increase when more expensive housing is considered
in a certain planning alternative.
Therefore, experts should evaluate together with the stakeholders what would be
the effect of the proposed alternative on the hazard, elements-at-risk location and
characteristics and the vulnerability. If needed new hazard modelling should be carried
out, or new elements-at-risk maps should be made representing the new situation.
2.6.3 Analyse risk reduction
After re-analysing the hazard, elements-at-risk and vulnerability for the situation
after the implementation of the planning alternative, the next step is to analyse the
resulting level of risk, and compare this with the current risk level. The difference
between the average annual losses before and after the implementation of the
planning alternative, provides information on the risk reduction. This should be
done for all the possible planning alternatives. The risk reduction should be done
preferable both in terms of economic risk reduction (reduction in the average
annual losses in monetary values) as well as in population risk reduction (reduction
in the expected casualties or exposed people).
The analysis of risk requires a repetitive procedure which has to be carried
out for each hazard scenarios (different hazard types and return periods)
in combination with elements-at-risk types, and then also for each possible
alternative. This requires the use of the same automated procedures using scripts
in GIS described in section 2.5. Figure 2.19 shows the results of the analysis of the
three risk reduction alternative for the case study. It is clear from the gure that
alternative 1 (engineering solutions) and 2 (Eco-DRR) result in more risk reduction
than alternative 3 (relocation).
Figure 2.19 Left: Results of the multi-hazard risk analysis for the current situation
(A0) and the three risk reduction alternatives: A1 (Engineering alternative), A2
(Eco-DRR alternative) and A3 (Relocation alternative). Right: Results of the Cost-
Benefit Analysis showing the Net Present Value (NPV) for the three alternatives for
different interest rates.
Multi-hazard risk assessment and decision making 75
2.6.4 Compare alternatives
Once the effects of the various planning alternatives are analysed, in terms of their
risk reduction, the next step is to compare them and decide which of the alternatives
would be the best to implement. This could be done through different methods:
• Cost-benet analysis. Here both the benets and the costs can be quantied.
The benet of a risk reduction alternative is represented by its annual risk
reduction in monetary values, which was calculated in the previous step (risk
after implementation minus current risk) (Figure 2.19). The costs for the
planning alternative can be quantied, as well in terms of their investment
costs, maintenance costs, project life time etc. (See Table 2.5). Cost-benet
analysis can be carried out by calculating relevant indicators, such as the Net
Present Value, Internal Rate of Return or Cost-Benet ratio.
• Cost-effectiveness analysis. This is carried out when the costs can be
quantied and compared, but the benets in terms of risk reduction cannot be
quantied in monetary values. This is the case for instance when population
risk is calculated, as it is generally considered not ethical to represent human
lives in monetary values.
• Multi-Criteria Evaluation. When both the costs and benets cannot
be quantied in monetary values, or when additional to cost-benet
or cost-effectiveness also other non-quantiable indicators are used, a
(spatial) multi-criteria evaluation is generally considered the best option.
In this analysis also social, ecological, cultural and other criteria can be
incorporated in the decision making process.
The comparison of alternatives is generally carried in a process, where the other
stakeholders are also involved before a decision is taken on the optimal alternative
that will be implemented.
2.6.5 Final decision and implementation
The last step of this workow related to the selection of the optimal planning
alternative in relation to the reduction of risk to hydro-meteorological hazards is the
consultation with the various stakeholders involved (Figure 2.12B). This includes
public hearings with the population, private sector, non-governmental organizations,
and various social network groups (e.g. communities, churches). The stakeholders
have the opportunity to request adjustment to the proposed plan of action, and if these
adjustments are considered valid, and substantial, a new round of evaluation might be
needed if the change of expected hazard and risk impact is substantial. Once the plan
is approved, the procedures will start for the implementation of the plan.
2.7 ANALYSING POSSIBLE FUTURE SCENARIOS
Risk is changing continuous, as both hazard and elements-at-risk change. Over a
longer period of time this may result in considerable changes in multi-hazard risk.
76 Environmental Hazards Methodologies for Risk Assessment
Figure 2.12C shows the workow to analyse changing risk due to possible future
scenarios.
2.7.1 Identification of possible future scenarios
The scenarios are related to possible changes related to climate, land use change or
population change due to global and regional changes, and which are only partially
under the control of the local planning organizations. The stakeholders might like
to evaluate how these trends have an effect on the hazard and elements-at-risk and
how these would translate into different risk levels. The possible future could be of
the following types:
• Climate change scenarios. In that case, the stakeholders require the
involvement of experts that would indicate which climate change scenarios
would be evaluated and what would be the expected effects in terms of
changes in frequency and magnitude of hydro-meteorological triggers would
be expected, such as changes in sea-level and extreme precipitation;
• Land use change scenarios. In this case, the stakeholders require the
involvement of experts that would indicate possible land use changes based on
macro-economic and political developments which would be translated into
local changes. For instance, scenarios could be envisaged where an increase in
tourism would be translated into possible future expansion of tourist facilities
would be evaluated. The future land use scenarios would also involve possible
changes in population which should also be taken into account.
• Future planning scenarios. In the national physical development plans also
possible future developments will be outlined and priorities for development
indicated which have implications for the spatial distribution of land use and
population.
Also combinations of these drivers might be considered. The possible future
changes should be expressed for certain years in the future, for instance for 2020
and 2030 and are considered as a basis for long term planning. Table 2.6 shows
four possible future scenarios for the case study area, which are combining land
use changes and climate changes.
2.7.2 Re-analysing hazards and elements-at-risk for
possible future scenarios
The possible future scenarios might lead to a modication of the hazard, exposure and
vulnerability in certain future years from now. Therefore it is required to re-analyse:
• The hazard. Possible future scenarios of climate change might lead to a change
in the frequency and magnitude of triggering events for oods and landslides.
Therefore, a new magnitude-frequency analysis might be required, that take into
account changing trends in frequencies of extreme events. The same hazard event
Multi-hazard risk assessment and decision making 77
Table 2.6 Definition of four possible future scenarios for the case study.
Scenario 1 2 3 4
Name Business as
Usual
Risk-Informed
Planning
Worst Case Most Realistic
Land use
change
Rapid growth
without taking
into account the
risk information
Rapid growth that takes
into account the risk
information and extends
the alternatives in the
planning
Rapid growth
without taking
into account the
risk information
Rapid growth that
takes into account the
risk information and
extends the alternatives
in the planning
Climate
change
No major
change
in climate
expected
No major change in
climate expected
Climate change
expected, leading
to more frequent
extreme events
Climate change
expected, leading to
more frequent extreme
events
2020 Return period: 20 20 (±5)20 (±5)17 (±6)17 (±6)
Return period: 50 50 (±10)50 (±10)45 (±12)45 (±12)
Return period: 100 100 (±20)100 (±20)90 (±23)90 (±23)
Return period: 200 200 (±40)200 (±40)180 (±44)180 (±44)
2030 Return period: 20 20 (±5)20 (±5)14 (±7)14 (±7)
Return period: 50 50 (±10)50 (±10)35 (±14)35 (±14 )
Return period: 100 100 (±20)100 (±20)75 (±26)75 (±26)
Return period: 200 200 (±40)200 (±40)150 (±49)150 (±49)
2040 Return period: 20 20 (±5)20 (±5)11 (±8)11 (±8)
Return period: 50 50 (±10)50 (± 10)25 (±16)25 (±16)
Return period: 100 100 (±20)100 (±20)55 (±30)55 (±30)
Return period: 200 200 (±40)200 (±40)110 (±53)110 (±53)
78 Environmental Hazards Methodologies for Risk Assessment
that now has an average return period of 50 years, might have an average return
period which is much smaller in a number of years from now. Also the intensities
of the hazard might change for instance due to changes in land use, which might
affect the hazardous processes (e.g. deforestation scenarios). In the case study
(Table 2.6), this is represented by a decrease in return periods in future years, and
the increasing standard deviations illustrate the increasing level of uncertainty.
• The exposure of elements-at-risk. The possible land use scenarios might lead
to substantial changes in land use/land cover, which also has an important
effect on the number of elements-at-risk within the various land use classes.
The analysis of future changes in land use/land cover is generally carried out
based on land parcel maps, rather than on the basis of building footprints
maps, as it is generally very difcult to translate the land use changes directly
into possible locations of buildings. Figure 2.20 shows the possible land use
maps for 2020, 2030 and 2040 for the case study area, following the possible
scenarios described in Table 2.6.
Figure 2.20 Possible land use changes for 2020, 2030 and 2040 for the possible
future scenarios described in Table 2.6.
Multi-hazard risk assessment and decision making 79
Therefore, experts should evaluate together with the stakeholders what would be the
effect of the possible future scenarios on the hazard, elements-at-risk location and
characteristics and the vulnerability. If needed new hazard modelling should be carried
out, or new elements-at-risk maps should be made representing the new situation.
2.7.3 Analyse possible changes risk for possible future
scenarios
After analysing the hazard, elements-at-risk and vulnerability for (a) future year(s)
given certain possible future scenarios, the next step is to analyse the resulting
change in risk, and compare this with the current risk level. The difference
between the current average annual losses and those in a future year under a given
change scenario provides information for decision makers on the possible negative
consequences of climate change and land use change scenarios. They can be used as
a basis for designing appropriate strategies for adaptation. The risk reduction should
be done preferable both in terms of changes in economic risk (average annual losses
in monetary values), as well as in population risk reduction (expected casualties
or exposed people). It is also important to incorporate the uncertainty levels in
this type of analysis, thus providing a range of change rather than concrete values.
Figure 2.21 shows the differences in risk for the four scenarios for the study area.
2.7.4 Changing risk evaluation
After assessing the possible changes in risk that might result from a number
of possible future scenarios related to climate change and land use change,
stakeholders should analyse these changes carefully in terms of:
• Spatial location of changes in risk. Some areas might be much more
impacted by these possible future changes than others. Based on the
outcomes of the analysis stakeholders could then prioritize certain areas for
critical interventions.
• Critical sectors. Changes in risk could be analysed for different sectors of
society, such as economy, agriculture, tourism, education, transportation etc.
• Development of adaptation strategies. The analysis of the expected level of
changes in risk and areas where an increase in risk is expected according to
the possible scenarios, should lead to the formulation of adaptation strategies
that aim to reduce these possible impacts through planning alternatives that
could be implemented now.
When the change of risk is compared for the four scenarios, the risk can be
observed including the climate change effects (by reducing the return periods as
indicated in Table 2.6) and has a very signicant increasing trend. When the effect
of land use changes (as shown in Figure 2.17) are examined, it can be seen that the
worst case scenario indeed has the largest increase in risk, and the risk informed
80 Environmental Hazards Methodologies for Risk Assessment
planning scenario has a lower risk. However, this last scenario focuses on the
avoidance of the most hazardous areas for ooding, debris ows and landslides, and
the development takes place mostly along the coast (See Figure 2.17). Therefore,
tsunami hazard is also included in the analysis, as shown in Figure 2.18, Scenario
2 and 4 have the highest risk levels. This stresses the importance of including all
hazard types in multi-hazard risk assessment.
Figure 2.21 Top: changes in population and economic values from 2016 to 2040
according to two land use change scenarios. Middle: changes in some of the land
use types from 2016 to 2040 according to two land use change scenarios. Below
left: Changes in multi-hazard risk from 2016 to 2040 for the four possible future
scenarios, without including tsunami risk. Lower right: Changes in multi-hazard risk
from 2016 to 2040 for the four possible future scenarios, including tsunami risk.
2.8 ANALYSING PLANNING ALTERNATIVES UNDER
POSSIBLE FUTURE SCENARIOS
The analysis as shown in Figure 2.21 shows that it is relevant to analyse how multi-
hazard risk might change in future according to different possible trends. However,
it also demonstrates the need to implement risk reduction measures now. In the last
workow, illustrated in Figure 2.12D, it is analysed which of the risk reduction
alternatives performs best under different possible future scenarios.
Multi-hazard risk assessment and decision making 81
2.8.1 Selection of alternatives, scenarios
and future years
The evaluation how different risk reduction alternatives will lead to risk reduction
under different future scenarios (trends of climate change, land use change and
population change) is the most complicated workow, as it requires to calculate
the present risk level, the effect of different risk reduction alternatives, and the
overprinting of these on the scenarios. For each of these combinations of alternatives
and scenarios new hazard, assets and risk maps need to be made. This type of
analysis allows stakeholders to make the most optimal “change proof” selection
of planning alternatives. This type of analysis is entirely based on experts and
consultants, which should evaluate both the effects of the planning alternatives, as
well as the associated effects of possible future scenarios on hazard, vulnerability
and risk. Such type of analysis could be applied to specic critical areas, such as the
capitals or important critical infrastructure. Table 2.7 combines all combinations
in a matrix. This table indicates the combination of the four scenarios (S1, S2, S3,
S4 as described in Table 2.6) and the three risk reduction alternatives (A1, A2, A3
described in Table 2.5) in three future years (2020, 2030, 2040). Due to the large
number of input maps it is important to use the coding of the les in a similar
way, so that it is possible to use the calculation script to analyse the loss and multi-
hazard risk all combinations. Therefore, for example: LP_2020_A1_S2 refers to
the land parcels for future year 2020 under alternative A1 (Engineering solutions)
and for scenario S2 (risk informed planning).
2.8.2 Re-analysing hazards and elements-at-risk for
alternatives/scenarios
The combination of the implementation of certain planning alternatives (structural
or non-structural risk mitigation measures) in combination with certain possible
future scenarios will certainly lead to a modication of the hazard, exposure and
vulnerability. This is why both the hazard maps and the elements-at-risk maps
should be updated for each combination.
• The hazard. In terms of the probability (or return period) of specic hazard
events, the spatial distribution of the hazard and the intensity of the hazards.
For instance, the construction of the retention basins reduce the area that will be
impacted by ooding and or debris ows. The retention basins are designed for
a specic maximum discharge (e.g. 100 years). With climate change scenarios
the same discharge may occur more frequently, and events with higher return
periods may overtop the basins. Therefore, it is required to re-analyse the
hazard given the implementation of the risk reduction measure and the possible
future climate change/land use change scenario. In the case study area, different
hazard intensity maps are considered for the various risk reduction measures,
and change the return periods for future years following Table 2.6.
82 Environmental Hazards Methodologies for Risk Assessment
Table 2.7 Combination of four scenarios, for future years 2020, 2030 and 2040, with three risk reduction alternatives. Each
combination is a situation where multi-hazard risk is analysed.
Scenario: Possible
Future Trends
Alternative: Risk
Reduction Options
Now 2014 Future Years
2020 2030 2040
S0 (Without including any
future trends)
A0 (no risk reduction) 2014_A0_S0 No future trends are taking into account, and all
hazards, elements at risk and vulnerabilities are
considered constant in future.
A1 Engin eering 2014_A0_S1
A2 Ecological 2014_A0_S2
A3 Relocation 2014_A0_S3
S1 Business as usual A0 (no risk reduction) Does not exist: use
existing situation
2020_A0_S1 2030_A0_S1 2040_A0_S1
A1 Engineering 2020_A1_S1 2030_A1_S1 2040_A1_S1
A2 Ecological 2020_A2_S1 2030_A2_S1 2040_A2_S1
A3 Relocation 2020_A3_S1 2030_A3_S1 2040_A3_S1
S2 Risk informed planning A0 (no risk reduction) Does not exist: use
existing situation
2020_A0_S2 2030_A0_S2 2040_A0_S2
A1 Engineering 2020_A1_S2 2030_A1_S2 2040_A1_S2
A2 Ecological 2020_A2_S2 2030_A2_S2 2040_A2_S2
A3 Relocation 2020_A3_S2 2030_A3_S2 2040_A3_S2
S3 Worst case (Rapid
growth + climate change)
A0 (no risk reduction) Does not exist: use
existing situation
2020_A0_S3 2030_A0_S2 2040_A0_S3
A1 Engineering 2020_A1_S3 2030_A1_S3 2040_A1_S3
A2 Ecological 2020_A2_S3 2030_A2_S3 2040_A2_S3
A3 Relocation 2020_A3_S3 2030_A3_S3 2040_A3_S3
S4 Climate resilience
(informed planning under
climate change)
A0 (no risk reduction) Does not exist: use
existing situation
2020_A0_S4 2030_A0_S3 2040_A0_S4
A1 Engineering 2020_A1_S4 2030_A1_S3 2040_A1_S4
A2 Ecological 2020_A2_S4 2030_A2_S3 2040_A2_S4
A3 Relocation 2020_A3_S4 2030_A3_S3 2040_A3_S4
Multi-hazard risk assessment and decision making 83
• The exposure of elements-at-risk. The number of elements-at-risk might
change as a result of the risk mitigation measure, or planning alternative,
and also as a result of the possible future scenario. For instance, if one of the
alternatives involves relocation, the number of exposed elements-at-risk will
decrease, whereas the hazard might stay the same. However, under certain
land use scenarios the pressure on the land might be so high that previously
related areas might become occupied again. In other planning alternatives,
the effect of future development on the number of exposed elements-at-
risk might also be evaluated. Figure 2.22 shows the land use maps for one
scenario (business as usual). Each combination of possible future scenario,
future year and risk reduction alternative requires a separate land use map.
From this gure it can be observed that under this scenario, the relocation
alternative will not be very effective, as the areas that are relocated in 2016
will become occupied again in later years. The different land use situations
is combined with increases in economic values and population numbers in
the various scenarios (As was shown in Figure 2.22). These will also have
an important inuence on the estimated multi-hazard risk.
Figure 2.22 Land parcel maps for different years in a possible future scenario (Business
as usual) without risk reduction measures (left) and three risk reduction alternatives.
Therefore, experts should evaluate together with the stakeholders what would be
the effect of the proposed alternatives and scenarios on the hazard, elements-at-risk
84 Environmental Hazards Methodologies for Risk Assessment
location and characteristics and the vulnerability for a given future year. If needed
new hazard modelling should be carried out, or new elements-at-risk maps should
be made representing the new situation.
2.8.3 Analyse risk reduction for alternatives/scenarios
After re-analysing the hazard, elements-at-risk and vulnerability for the specic
combinations of planning alternative, possible future scenario and future year, the
next step is to analyse the resulting level of risk, and compare this with the current
risk level. The difference between the average annual losses before and after the
implementation of the planning alternative, provides information on the risk
reduction. This should be done for all the possible planning alternatives/scenario
combinations. The risk reduction should be done preferable both in terms of economic
risk reduction (reduction in the average annual losses in monetary values), as well as
in population risk reduction (reduction in the expected casualties or exposed people).
2.8.4 Compare alternatives under different scenarios
Once the effect of the various planning alternatives has been analysed, under different
future years and future scenarios, in terms of their risk reduction, the next step is to
compare them and decide which of the alternatives would be the best to implement. In
the cost-benet analysis both the benets and the costs can be quantied. The benet
of a risk reduction alternative is represented by its annual risk reduction in monetary
values, which was calculated in the previous step (risk after implementation minus
current risk). However, whereas the benet would remain constant in the analysis
which was presented earlier under “Analysing planning alternative”, when the risk
reduction of planning alternatives are analysed for different future years under
possible change scenarios, the risk reduction might also change considerably over
time. The costs for the planning alternative can be quantied, as well in terms of
their investment costs, maintenance costs, project life time etc. Cost-benet analysis
can be carried out by calculating relevant indicators, such as the Net Present Value,
Internal Rate of Return or Cost-Benet ratio. When possible future changes are
taken into account, the cost-benet ratios of the various alternatives might be quite
different than if no future changes are considered, which might lead to the selection
of another planning alternative that may be the most “change proof”.
2.8.5 Final decision and implementation
The last step of this workow related to the selection of the optimal planning
alternative in relation to the reduction of risk to hydro-meteorological hazards
is the consultation with the various stakeholders involved. This includes public
hearings with the population, private sector, non-governmental organizations, and
various social network groups (e.g. communities, churches). The stakeholders have
the opportunity to request adjustment to the proposed plan of action, and if these
Multi-hazard risk assessment and decision making 85
adjustments are considered valid, and substantial, a new round of evaluation might
be needed if the change of expected hazard and risk impact is substantial. Once
the plan is approval the procedures will start for the implementation of the plan.
Figure 2.23 Changes in Average Annual Loss for four scenarios and three risk
reduction alternatives for future years.
AQ15
Here, the environmental impact assessment comes into play (Greiving, 2004).
The recently amended Environmental Impact Assessment Directive (2014/52/EU)
stats with Art 3 § 2: “Precautionary actions need to be taken for certain projects
which, because of their vulnerability to major accidents, and/or natural disasters
(such as ooding, sea level rise, or earthquakes) are likely to have signicant
adverse effects on the environment. For such projects, it is important to consider
their vulnerability (exposure and resilience) to major accidents and/or disasters,
the risk of those accidents and/or disasters occurring and the implications for the
likelihood of signicant adverse effects on the environment.”
In this context, a broad involvement of the public is requested (see Art 6 §
2 Environmental Impact Assessment Directive and Art. 6 §§ 1–6 Strategic
Environmental Assessment Directive) that needs to be adjusted to the different
steps of analysis of a risk assessment as shown by Figure 2.24.
However, in many European countries, a legal basis for hazard (and partly
risk) zoning exists which means that both, the methods chosen for delineating the
zones, but also the legal consequences for subsequent planning activities are laid
down by law. Hence, there is no discussion possible about suitable risk assessment
alternatives. Table 2.9 gives an overview about existing zoning models and
discusses their advantages and disadvantages.
86 Environmental Hazards Methodologies for Risk Assessment
Figure 2.24 Integration of risk assessment into environmental assessments.
2.9 SUMMARY AND CONCLUSIONS
2.9.1 Which method to choose?
The four methods for risk assessment that were treated in the previous sections all
have certain advantages and disadvantages, which are summarized in Table 2.8. The
Quantitative Risk Assessment method is the best for evaluating several alternatives
for risk reduction, through a comparative analysis of the risk before and after the
implementation followed by a cost-benet analysis. The event-tree analysis is the best
approach for analysing complex chains of events and the associated probabilities.
Qualitative methods for risk assessment are useful as an initial screening process to
identify hazards and risks. They are also used when the assumed level of risk does not
justify the time and effort of collecting the vast amount of data needed for a quantitative
risk assessment, and where the possibility of obtaining numerical data is limited. The
risk matrix approach is often the most practical approach as basis for spatial planning,
where the effect of risk reduction methods can be seen as changes in the classes within
the risk matrix. The indicator-based approach, nally, is the best when there is not
enough data to carry out a quantitative analysis, but also as a follow-up of a quantitative
analysis as it allows to take into account other aspects than just physical damage.
The decision depends among other factors in particular from the spatial scale of
the project, plan or program, the risk assessment was done for (see Table 2.8). In this
Multi-hazard risk assessment and decision making 87
Table 2.8 Advantages and disadvantages of the four risk assessment methods discussed.
Method Advantages Disadvantages Suitability for Specific
Spatial Scales
Quantitative risk
assessment
(QRA)
Provides quantitative risk
information that can be used
in Cost-benefit analysis of risk
reduction measures.
Very data demanding. Difficult
to quantify temporal probability,
hazard intensity and vulnerability.
Normally used as basis for
investments in structural
mitigation measures on
project level
Event-tree
analysis
Allow modelling of a sequence of
events, and works well for domino
effects
The probabilities for the different
nodes are difficult to assess, and
spatial implementation is very
difficult due to lack of data.
Normally used as basis for
plan approval procedures
of dangerous facilities
(e.g. nuclear power plants,
chemical establishments) on
project level
Risk matrix
approach
Allows to express risk using classes
instead of exact values, and is
a good basis for discussing risk
reduction measures.
The method doesn’t give
quantitative values that can be
used in cost-benefit analysis of
risk reduction measures. The
assessment of impacts and
frequencies is difficult, and one area
might have different combinations of
impacts and frequencies.
Basis for hazard zoning
in many countries like
Austria France, Italy and
Switzerland. Good fit for
regional and local spatial
planning as basis for keeping
hazard prone areas free of
further development
Indicator-based
approach
Only method that allows to carry out
a holistic risk assessment, including
social, economic and environmental
vulnerability and capacity.
The resulting risk is relative and
doesn’t provide information on
actual expected losses.
Suitable for comparing the
level of risk on national level
(see e.g. World Risk Index)
88 Environmental Hazards Methodologies for Risk Assessment
Table 2.9 Hazard zoning models in Europe.
Model Coordinated Zoning in General
Land-use Plan (Applied e.g. in
Finland, Poland, Germany Except
Floods)
Specific Hazard Map in General
Land-use Plan with Binding
Effects (Applied e.g. in Austria,
France, Italy, Switzerland,
Germany – Floods Only)
Independent Map without
Binding Effects (Applied e.g.
in Greece, Spain, and U.K.)
Description Consideration of the hazard prone
areas during the compiling or review
of the local land-use plan (informed
i.e by Strategic Environmental
Assessment)
The hazard zones are displayed
as a separate map which has a
direct effect on land ownership
rights
Definition of hazard zones
within the scope of expert
planning – objections may be
raised to decisions that are
made on this basis
Advantages At the local level, no additional
instruments are needed; hazards are
weighted- up against other concerns
and interests
The hazard can be considered in
a uniform manner for the whole
municipality.
Definition of hazard zones can
be applied directly in building
approval procedures
A simple alteration of a hazard
zone plan is possible. Suitable
for a cooperative strategy that
aims at influencing existing
building structures by individual
building protection
Disadvantages Land-use plans only contain
information about hazard prone
areas when a specific reference is
made. An alternation of the danger
situation means the plan must be
adapted accordingly
An alteration of the risk means
that the complete zoning plan
has to be adapted accordingly.
For legally binding effects, a very
exact evidence basis is needed
Not effective if private
stakeholders do not want to
follow the advise
Multi-hazard risk assessment and decision making 89
context, the subsidiarity principle plays a considerable role. Art. 5 § 2 of the Strategic
Environmental Assessment Directive (2001/42/EC) lays down: “The environmental
report […] shall include […] the level of detail in the plan or programme, its stage
in the decision-making process and the extent to which certain matters are more
appropriately assessed at different levels in that process in order to avoid duplication
of the assessment.” This means that a quantitative risk assessment would not t to
the scope of a strategic environmental assessment at the level of preparatory plans
or programs, but should be used as evidence basis for an environmental impact
assessment at the project level, as long as the project may be threatened by any kind
of natural or technological hazard.
2.9.2 Tools for multi-hazard assessment
The analysis of risk requires a repetitive procedure which has to be carried out for
each hazard scenarios (different hazard types and return periods) in combination
with elements-at-risk types, and then also for each possible alternative. This
requires the use of automated procedures using Geographic Information Systems.
Risk assessment is computationally intensive. It can be carried out using
conventional GIS systems, although it is advisable to use specic software tools.
Loss estimation has been carried out in the insurance sector since the late 1980’s
using geographic information systems. Since the end of the 1980’s risk modelling
has been developed by private companies resulting in a range of proprietary software
models for catastrophe modelling for different types of hazards. Unfortunately,
these are not publicly available, which is a major obstacle to the development of
risk assessment for many parts of the world by government organizations. The
best initiative for publicly available loss estimation thus far has been HAZUS
developed by the Federal Emergency Management Agency (FEMA) together with
the National Institute of Building Sciences (Schneider & Schauer, 2006). The rst
version of HAZUS was released in 1997 with a seismic loss estimation focus, and
was extended to multi-hazard losses in 2004, incorporating also losses from oods
and windstorms. HAZUS was developed as a software tool under ArcGIS. Several
other countries have adapted the HAZUS methodology to their own situation.
The HAZUS methodology has also been the basis for the development of several
other software tools for loss estimation. One of these is called SELENA. Also an
interesting example is RiskScape developed in New Zealand (Schmidt etal. 2011).
Another interesting development has been going on in the development
of standalone software modules for multi-hazard risk assessment, which are
not running as a component of an existing GIS. A good example of this is the
CAPRA Probabilistic Risk Assessment Program supported by the World Bank
(CAPRA, 2013). The methodology focuses on the development of probabilistic
hazard assessment modules, for earthquakes, hurricanes, extreme rainfall, and
volcanic hazards, and the hazards triggered by them, such as ooding, windstorms,
landslides and tsunamis.
90 Environmental Hazards Methodologies for Risk Assessment
Another recent development is towards Open Source web-based modules for
multi-hazard risk assessment.
A tool which is currently under development as part of the Global Earthquake
Initiative (GEM), called OpenQuake (https://www.globalquakemodel.org/openquake/
support/documentation/), is most probably going to be the standard for earthquake loss
estimation, as there are also plans to expand it into a multi-hazard risk assessment tool.
2.9.3 Development of a spatial decision support system
Within the framework of the EU FP7 Marie Curie Project CHANGES (www.
changes-itn.eu) and the EU FP7 Copernicus project INCREO (http://www.increo-fp7.
eu) a spatial decision support system was developed with the aim to analyse the
effect of risk reduction planning alternatives on reducing the risk now and in the
future, and support decision makers in selecting the best alternatives. The Spatial
Decision Support System is composed of a number of integrated components. The
Risk Assessment component allows to carry out spatial risk analysis, with different
degrees of complexity, ranging from simple exposure (overlay of hazard and assets
maps) to quantitative analysis (using different hazard types, temporal scenarios and
vulnerability curves) resulting into risk curves. The platform does not include a
component to calculate hazard maps, and existing hazard maps are used as input
data for the risk component. The second component of the SDSS is a risk reduction
planning component, which forms the core of the platform. This component includes
the denition of risk reduction alternatives (related to disaster response planning,
risk reduction measures and spatial planning) and links back to the risk assessment
module to calculate the new level of risk if the measure is implemented, and a cost-
benet (or cost-effectiveness/Spatial Multi Criteria Evaluation) component to compare
the alternatives and make decision on the optimal one. The third component of the
SDSS is a temporal scenario component, which allows to dene future scenarios
in terms of climate change, land use change and population change, and the time
periods for which these scenarios will be made. The component doesn’t generate
these scenarios but uses input maps for the effect of the scenarios on the hazard and
assets maps. The last component is a communication and visualization component,
which can compare scenarios and alternatives, not only in the form of maps, but also
in other forms (risk curves, tables, graphs). The envisaged users of the platform are
organizations involved in planning of risk reduction measures, and that have staff
capable of visualizing and analysing spatial data at a municipal scale. The Decision
Supper System RiskChanges is accessible at: http://www.charim.net/use_case/46.
ACKNOWLEDGEMENTS
The following persons have developed the original hazard maps: Leonardo
Cascini, Settimio Ferlisi, Sabatino Cuomo, and Giovanna De Chiara. We also
would like to thank Anna Scolobig from IIASA for her work on the risk reduction
alternatives (which we have taken as they were) the stakeholder involvement and
Multi-hazard risk assessment and decision making 91
the stakeholder roleplay exercise. Hari Narasimhan (ETH) and Emile Dopheide
are thanked for their input in the cost-benet analysis. Also we would like to thank
Andrea Tripodi for his work in the development of the case study. Luc Boerboom
and Ziga Malek are thanked for their input in the thinking about possible future
scenarios. Kaixi Zhang is thanked for her feedback on the risk calculation method.
Vera Andrejchenko, Julian Berlin, Irina Cristal, Kaixi Zhang and Roya Olyazadeh
are thanked for developed the RiskChanges Spatial Decision Support System.
Funding was provided through the EU FP7 Marie Curie Project CHANGES (www.
changes-itn.eu) the EU FP7 Copernicus project INCREO (http://www.increo-fp7.eu),
and the World Bank GFDRR project CHARIM (http://www.charim.net)
REFERENCES
ACOE (ARMY CORPS OF ENGINEERS) (2004). Water Resources Assessment of
Dominica, Antigua, Barbuda, St. Kitts and Nevis: Mobile District and Topographic
Engineering Center Report, U.S. Army Corps of Engineers, Mobile, AL, p. 94.
Alexander D. (2001). Encyclopedia of Environmental Science, Chapter Natural Hazards.
Kluwer Academic Publishers.
Bedford T. and Cooke R. M. (2001). Probabilistic Risk Analysis: Foundations and Methods.
Cambridge University Press.
Birkmann J. Ed. (2006). Measuring Vulnerability to Natural Hazards: Towards Disaster
Resilient Societies. UNU Press, Tokyo, New York.
Birkmann J. (2007). Risk and vulnerability indicators at different scales: applicability,
usefulness and policy implications. Environmental Hazards, 7, 20–31.
Bollin C. and Hidajat R. (2006). Community-based disaster risk index: pilot implementation
in Indonesia. In: Measuring Vulnerability to Natural Hazards—Towards Disaster
Resilient Societies, J. Birkmann (ed.), UNU-Press, Tokyo, New York, Paris.
Breheny P. (2007). Hydrogeologic/Hydrological Investigation of Landslide Dam and
Impounded Lake in the Matthieu River Valley, Commonwealth of Dominica, West
Indies. Unpublished M.S. Dissertation, University of Leeds, United Kingdom.
Cannon S. and DeGraff J. (2009). The increasing wildre and post-re debris-ow threat
in Western USA, and implications for consequences of climate change. In:Landslides
– Disaster Risk Reduction, K. Sassa and P. Canuti (eds), Springer Verlag, pp. 177–190.
CAPRA (2013). Probabilistic Risk Assessment Program. http://www.ecapra.org/.
Carpignano A., Golia E., Di Mauro C., Bouchon S. and Nordvik J.-P. (2009). A
methodological approach for the denition of multi-risk maps at regional level: rst
application. Journal of Risk Research, 12, 513–534.
Coppock J. T. (1995). GIS and natural hazards: an overview from a GIS perspective. In:
Geographical Information Systems in Assessing Natural Hazards, A. Carrara and
F. Guzzetti (eds), Volume 5 of the series Advances in Natural and Technological
Hazards Research, pp. 21–34.
Corominas J., van Westen C. J., Frattini P., Cascini L., Malet J.-P., Fotopoulou S., Catani
F., van den Eeckhaut M., Mavrouli O., Agliardi F., Pitilakis K., Winter M. G., Pastor
M., Ferlisi S., Tofani V., Hervas J. and Smith J. T. (2014). Recommendations for the
quantitative analysis of landslide risk: open access. Bulletin of Engineering Geology
and the Environment IAEG, 73(2), 209–263.
AQ1
AQ2
AQ3
AQ3
AQ4
AQ5
92 Environmental Hazards Methodologies for Risk Assessment
Cova T. J. (1999). GIS in emergency management. In: Geographical Information Systems:
Principles, Techniques, Applications, P. A. Longley (ed.), 2nd edn, Management Issues
and Applications.
DeGraff J. V. (2012). Solving the dilemma of transforming landslide hazard maps into
effective policy and regulations. Natural Hazards Earth System Sciences, 12, 53–60.
DeGraff J. V. and Rogers C. T. (2003). An unusual landslide dam event in Dominica, West
Indies. Landslide News, 14(15), 8–11.
DeGraff J. V., James A. and Breheney P. (2010). The formation and persistence of the
Matthieu Landslide-Dam Lake, Dominica, W.I. Environmental & Engineering
Geoscience, 16(2), 73–89.
Delmonaco G., Margottini C. and Spizzichino D. (2006a). ARMONIA Methodology
for Multi-Risk Assessment and the Harmonisation of Different Natural Risk Maps.
Deliverable 3.1.1, ARMONIA.
Delmonaco G., Margottini C. and Spizzichino D. (2006b). Report on New Methodology
for Multi-Risk Assessment and the Harmonisation of Different Natural Risk Maps.
Deliverable 3.1, ARMONIA.
De Pippo T., Donadio C., Pennetta M., Petrosino C., Terlizzi F. and Valente A. (2008).
Coastal hazard assessment and mapping in Northern Campania, Italy. Geomorphology,
97, 451–466.
EC (2011). Risk Assessment and Mapping Guidelines for Disaster Management. European
Commission Commission staff working paper, European Union.
EC (2011). Risk Assessment and Mapping Guidelines for Disaster Management. European
Commission Commission staff working paper, European Union.
Fan X., van Westen C. J., Xu Q., Gorum T. and Dai F. (2012). Analysis of landslide dams
induced by the 2008 Wenchuan earthquake. Journal of Asian Earth Sciences, 57, 25–37.
FEMA (2004). HAZUS-MH. FEMA’s Methodology for Estimating Potential Losses from
Disasters. US Federal Emergency Management Agency. http://www.fema.gov/plan/
prevent/hazus/index.shtm.
Ferlisi S., De Chiara G. a nd Cascini L. (2016). Quantitative risk an alysis for hyperconcentrat ed
ows in Nocera Inferiore (Southern Italy). Submitted to Natural Hazards—special
issue: landslide risk Nocera Inferiore. Nat Hazards, 81, S89–S115.
Garcia-Aristizabal A. and Marzocchi W. (2013). Software for Multi-Hazard Assessment.
Deliverable 3.5 of the EU Matrix Project: New Methodologies for Multi-Hazard and
Multi-Risk Assessment Methods for Europe. http://matrix.gpi.kit.edu/downloads/
MATRIX-D3.05.pdf.
Granger K., Jones T. G., Leiba M. and Scott G. (1999).Community Risk in Cairns: A
Multi-Hazards Risk Assessment. Technical Report, Australian Geological Survey
Organisation (AGSO). http://www.ga.gov.au/corporate_data/33548/33548.pdf.
Greiving S. (2004). Risk assessment and management as an important tool for the EU
Strategic Environmental Assessment. DISP, 157, 11–17.
Greiving S., Fleischhauer M. and Lückenkötter J. (2006). A Methodology for an integrated
risk assessment of spatially relevant hazards. Journal of Environmental Planning and
Management, 49(1), 1–19.
Guha-Sapir D., Below R. and Hoyois P. H. (2016). EM-DAT: The CRED/OFDA International
Disaster Database. www.emdat.be. Université Catholique de Louvain, Brussels, Belgium.
Haimes Y. Y. (2009) Risk Modeling, Assessment, and Management. 3rd edn. John Wiley
& Sons, p. 1009.
AQ6
AQ7
AQ8
Multi-hazard risk assessment and decision making 93
Ho K. (2009). Risk Acceptability Criteria. Lecture Held at the International LARAM
School. Government of Hong Kong SAR, CEDD, Geotechnical Engineering Ofce.
IRDR (2014) Integrated Research on Disaster Risk. Peril Classication and Hazard
Glossary (IRDR DATA Publication No. 1). Integrated Research on Disaster Risk,
Beijing. http://www.irdrinternational.org/wp-content/uploads/2014/04/IRDR_DATA-
Project-Report-No.-1.pdf.
ISO 31000. Risk Management – Principles and Guidelines. http://www.iso.org/iso/home/
standards/iso31000.htm.
Jaboyedoff M., Aye Z. C., Derron M.-H., Nicolet P. and Olyazadeh R. (2014). Using
the consequence – frequency matrix to reduce the risk: examples and teaching. In
International Conference Analysis and Management of Changing Risks for Natural
Hazards, 18–19 November 2014, Padua, Italy.
Jonkman S. N., Van Gelder P. and Vrijling H. (2002). An overview of quantitative risk
measures and their application for calculation of ood risk. In: λμ13 – ESREL 2002
European Conference.
Kappes M. S., Keiler M., Von Elverfeldt K. and Glade T. (2012). Challenges of analyzing
multi-hazard eisk: a review. Natural Hazards, 64(2), 1925–1958.
Lacasse S., Eidsvik U., Nadim F., Hoeg K. and Blikra L. H. (2008). Event tree analysis
of Aknes rock slide hazard. In: IV Geohazards Quebec, 4th Canadian Conf. on
Geohazards, pp. 551–557.
Lee K. and Rosowsky D. (2006). Fragility analysis of woodframe buildings considering
combined snow and earthquake loading. Structural Safety, 28, 289–303.
Linnerooth-Bayer J. and Pratt A. (2016). Introduction to the special issue on rethinking
participatory processes: the case of landslide risk in Nocera Inferiore. Nature Hazards,
81, 1–6.
Linnerooth-Bayer J., Scolobig A., Ferlisi S., Cascini L. and Thompson M. (2015). Expert
engagement in participatory processes: translating stakeholder discourses into policy
options. Submitted to Natural Hazards—special issue: landslide risk Nocera Inferiore.
Nature Hazards, 81, 69–88.
Luino F. (2005). Sequence of instability processes triggered by heavy rainfall in the
Northern Italy. Geomorphology, 66, 13–39.
Marzocchi W., Mastellone M. and Di Ruocco A. (2009). Principles of Multi-Risk
Assessment: Interactions Amongst Natural and Man-Induced Risks. European
Commission. http://cordis.europa.eu/documents/documentlibrary/106097581EN6.pdf.
Narasimhan H., Ferlisi S., Cascini L., De Chiara G. and Faber M. H. (2016). A cost–benet
analysis of mitigation options for optimal management of risks posed by ow-like
phenomena. Nat Hazards, 81, S117–S144.
OAS (1991) Geographic Information Systems in Natural Hazard Management. Organization
of American States. Primer on Natural Hazard Management in Integrated Regional
Development Planning. https://www.oas.org/dsd/publications/Unit/oea66e/ch05.htm.
Peila D. and Guardini C. (2008). Use of the event tree to assess the risk reduction obtained
from rockfall protection devices. Natural Hazards Earth System Sciences, 8, 1441–1450.
Perles Roselló M. and Cantarero Prados F. (2010). Problems and challenges in analyzing
multiple territorial risks: methodological proposals for multi-hazard mapping. Boletín
de la Asociación de Geógrafos Espanoles, 52, 399–404.
Prinos P. (2008). Review of Flood Hazard Mapping. European Community Sixth Framework
Programme for European Research and Technological Development. FLOODsite.
AQ9
AQ10
94 Environmental Hazards Methodologies for Risk Assessment
Roberts N. J., Nadim F. and Kalsnes B. (2009). Quantication of vulnerability to natural
hazards. Georisk: Assessment and Management of Risk for Engineered Systems and
Geohazards, 3(3), 164–173.
Roobol M. J. and Smith A. L. (2004). Geologic Map of Dominica, West Indies: Geology
Department, University of Puerto Rico at Mayaguez. Electronic document, Available
at http://www.caribbeanvolcanoes.com/dominica/content/dominicamap.pdf.
SafeLand (2011a) Deliverable 2.11 QRA Case Studies at Selected “hotspots”. Synthesis
of critical issues. Edited for the EU-funded SafeLand Project by Corominas J and
Mavrouli O. http://www.safeland-fp7. Eu.
SafeLand (2011b) Deliverable 5.7 Design and Testing: A Risk Communication Strategy and a
Deliberative Process for Choosing a Set of Mitigation and Prevention Measures. Edite