Multi-hazard event typology for Western Europe
(1) Aloïs Tilloy - European Commission, Joint Research Centre, Italy & Department of Geography, King’s College London, England
(2) Bruce D. Malamud - Department of Geography, King’s College London, England
INTRODUCTION
On a global and annual scale, natural hazards can cause significant socioeconomic losses. However, they are far from independent [1,2]. The interaction or competition of different hazards may result in an impact greater than the sum of the effects of the hazards taken independently [3]. Extreme events involving multiple hazards are usually referred to as multi-hazard events or compound events. We can take two recent examples of such phenomena that impacted France:
- in 2010, Storm Xynthia hit the Atlantic coast. The storm was not particularly extreme for the season in itself, but the combination of extreme winds, high tides and heavy swells, combined with already saturated soils, caused significant damage;
- in the summer of 2022, the combination of intense drought and multiple storms one after the other contributed to the development of numerous forest fires. Recent studies have attempted to group interdependent hazards into events in order to improve understanding of the processes leading to multiple hazards [4,5].
This article is part of this effort, with the objective of understanding the multihazard landscape of a given region[6] by considering its climatic and geophysical characteristics.
This paper offers to identify and group natural hazards into multi-hazard networks relevant for Western Europe. The studied region is defined according to the concept of biogeographical region, i.e. an area with relatively homogeneous characteristics in terms of biota (fauna and flora) and climate. The European Union has defined nine biogeographic regions for Europe. Among these is the European Atlantic Region (EAR), which includes all of the UK and the western part of France (Figure 1) and which was studied in this article.
METHODOLOGY
Hazard selection
Each year the EAR is exposed to a wide range of natural hazards. For example, while coastal lowlands are more vulnerable to extratropical storms, mountainous areas are more prone to landslides. The southern parts (Spain, Portugal, France) of the EAR can experience major forest fires in summer, while the plains of England, northern France and the Rhine valley are exposed to river flooding. The identification of natural risks relevant to the EAR is carried out on the basis of three main criteria: (i) frequency of occurrence, (ii) spatial relevance (location), (iii) potential impact on energy infrastructure (production, distribution).
For each criterion, each natural hazard is assessed by a semi-quantitative score at three levels: (-) not mentioned, (*) mentioned, (**) mentioned and important. This information is derived from various sources (Emergency Events Database EM-DAT, Spatial Effects and Management of Natural and Technological Hazards in Europe and Energy Technologies Institute) and hazard maps. By combining the relevance scores on each of these three criteria, 16 natural hazards with a global relevance score of at least *** were selected to characterise the multihazard landscape of the EAR (Table 1).
Creation of multi-hazard networks
The 16 natural hazards (Table 1) are grouped into multi-hazard networks. A multi-hazard network is a set of interdependent hazards that can be triggered by the same processes and occur in a given spacetime. They can be linked to concepts developed over the past two decades to understand multihazard events: predictors (meteorological processes influencing the intensity and dynamics of compound events), triggers (factors determining the frequency and intensity of hazards); scenarios (conceptual event used for infrastructure design), physical events (physical processes that determine hazards); generic events (classification of hazard groups).
Based on the Integrated Research on Disaster Risk (IRDR) natural hazard classification report and the literature review by Tilloy et al [7], the following multihazard networks can be developed:
1. Extratropical cyclone (hydrometeorological)
2. Convective storm (hydrometeorological)
3. Ground movement (geophysics)
4. Compound dry hazards (hydrometeorological)
5. Compound cold hazards (hydrometeorological) Each hazard listed in Table 1 belongs to one (or more) of the five networks defined above.
The five multi-hazard networks used in this thesis and the natural hazards they include are summarised in Figure 2.
RESULTS
The spatial-temporal dynamics and interactions that define the five multi-hazard networks are presented in this section. The assumptions regarding the interrelationships of hazards within each multihazard network are mainly based on previous work[2-7]. Descriptions of each hazard network and interrelationship are supported by literature and a catalogue of 50 historical multi-hazard events in Western Europe.
Catalogue of historical multi-hazard events (MH)
The catalogue contains a total of 50 historical events that have occurred in EAR countries (10 events for each of the five multi-hazard networks). It is based on 32 sources dealing with the hazards presented in Table 1 (catalogues of single hazards and hydrometeorological events, natural catastrophe databases and a comprehensive literature review).
The hazards, interrelationships, spatial extent and length of each multi-hazard event are identified. The spatial scale corresponds to the total footprint of the event that caused damage according to the sources examined. The length of events is also extracted from the literature, expressed in days and associated with the duration of the reported impacts. Figure 3 highlights the variations in spatial footprint and duration between networks (localised and short convective storms vs. widespread and persistent dry hazards) but also within networks (short or long-lasting cold hazards).
# multi-hazard
# Europe
# compound hazards
# typology
# extreme events
Multi-hazard networks
The five multi-hazard networks are associated with a variable number of hazards, ranging from three for ground movements to seven for convective storms. It is important to note that some natural hazards are found in several multi-hazard networks. For example, extreme precipitation and extreme winds are both part of extratropical storms and convective storms. The relationship between these two hazards differs according to the multihazard networks to which they belong. Each hazard network is assigned a dominant hazard derived from the catalogue of historical multi-hazard events. The dominant hazard is the one most likely to occur and the most interconnected within a network (Table 2). Once the multi-hazard networks and their associated hazards have been defined, the nature of the interrelationships between the hazards and their interrelationship networks can be focused on. The classification made by Tilloy et al. [7], is used to develop interrelationships networks.
Three types of interrelationships are considered:
1) “Triggering (or cascade)”: one hazard triggers another;
2) “Change of condition”: One hazard alters the probability of a second hazard by changing environmental conditions;
3) “Compound hazards”: Different hazards arise from the same phenomenon. The interrelationships ‘trigger’ and ‘change of condition’ imply causality and are therefore represented by arrows, while ‘compound hazards’ denote non-causal dependence (Figure 4).
In Figure 4, each of the five multi-hazard networks defined in Table 2 is represented graphically with the hazards that compose it and the interrelationships between these hazards. When two hazards can have two types of interrelationships, both types are represented. For example, an existing drought may accentuate a heat wave (change in condition) or form a compound hazard with it.
Figure 1 - Physiographic map of Western Europe. The European Atlantic biogeographical region is highlighted in white (EEA, 2002).
(1) geophysical, (2) atmospheric, (3) hydrological, (4) biophysical. Relevance scores are displayed for (i) frequency of occurrence; (ii) spatial relevance; (iii) impact on energy infrastructure. The overall relevance is the sum of the 3 relevance scores. The abbreviations used for natural hazards correspond to their names in English.
Figure 2 - The five multi-hazard networks discussed, and the associated hazards divided into four natural hazard categories: (1) geophysical, (2) atmospheric, (3) hydrological, (4) biophysical.
Figure 3 - Spatial footprint and duration of 50 multi-hazard events divided into 5 networks. Ground Movement (GM), Convective Storm (CS), Extratropical Cyclone (ETC), Combined Dry Hazard (CD) and Combined Cold Hazard (CC)
Table 2 - Definition of the five multi-hazard networks, the hazards that make them up, their dominant hazard (in bold) and their number of interrelationships. The abbreviations used for natural hazards correspond to those introduced in Table 1.
THE PARTNERS
This research was carried out as part of the doctoral thesis of Aloïs Tilloy in the Department of Geography at King’s College London. The work was funded by EDF R&D and carried out in collaboration with the Natural Hazards team at the EDF UK R&D Centre. The author is now working at the European Commission’s Joint Research Centre.
CONCLUSION
Multi-hazard networks are generic events that aim to provide an initial framework to model the interrelationships between multiple hazards. These represent a step forward from interrelationship matrices by combining more than two hazards. However, the transition towards a quantitative approach requires more databases with multiple hazards and multivariate modelling methods such as Bayesian networks or copula clusters.
The concept of dominant hazard developed in this paper creates new possibilities for multivariate analysis. The study of the evolution of “satellite” hazards and their interrelationships in relation to a dominant “central” hazard would reduce modelling complexity. The typology developed synthesises interdisciplinary knowledge of hazard interrelationships, bringing together atmospheric, hydrological, geophysical and biophysical hazards in a given region. Finally, this paper provides a clear view of the interrelationships between hazards that can cause damage in the EAR. The catalogue of historical multi-hazard events and all references are available in Aloïs Tilloy’s thesis [8]./
Figure 4 - The five multi-hazard networks diagrammed with the three types of interrelationships between hazards within the networks represented. The abbreviations used for natural hazards correspond to those introduced in Table 1.
REFERENCES
1. Hewitt, K. & Burton, I. Hazardousness of a place: A regional ecology of damaging events. (University of Toronto Press, 1971).
2. Gill, J. C. & Malamud, B. D. Reviewing and visualizing the interactions of natural hazards. Rev. Geophys. 52, 680–722 (2014).
3. Terzi, S. et al. Multi-risk assessment in mountain regions: A review of modelling approaches for climate change adaptation. J. Environ. Manage. 232, 759–771 (2019).
4. Zscheischler, J. et al. A typology of compound weather and climate events. Nat. Rev. Earth Environ. 1, 333–347 (2020).
5. Schauwecker, S. et al. Anticipating cascading effects of extreme precipitation with pathway schemes - Three case studies from Europe. Environ. Int. 127, 291–304 (2019).
6. Gill, J. C., Malamud, B. D., Barillas, E. M. & Noriega, A. G. Construction of regional multihazard interaction frameworks, with an application to Guatemala. Nat. Hazards Earth Syst. Sci. 20, 149–180 (2020).
7. Tilloy, A., Malamud, B. D., Winter, H. & Joly-Laugel, A. A review of quantification methodologies for multi-hazard interrelationships. Earth-Science Rev. 196, 102881 (2019).
8. Tilloy, A., Understanding and modelling extreme multi-hazard events, Phd theses, 2021, https:// kclpure.kcl.ac.uk/portal/ en/ theses/understanding-andmodelling-extreme-multihazardevents(3fe54436-4375-4530-a4fe24c6992c09bf).html
CITATION
Tilloy et al, Multi-hazard event typology for Western Europe. In CCR 2022 Scientific Report; CCR, Paris, France, 2022, pp. 60-63