Towards a unified seismic-flood-hazard model for risk assessment of roadway networks in Greece

Abstract

Roadway networks, playing a vital role in the economic prosperity of modern societies. Recent hazardous events in Greece, for instance, the 2021 Thessaly earthquake and floods and the heavy 2019 rainfall in Crete, have demonstrated the vulnerability of roadway networks to natu-ral hazards, resulting in severe physical damage and important economic and societal losses. Severe damage on bridges and tunnels of roadway networks is commonly related to the effects of multiple hazards that may act independently during their life. However, the literature on risk assessment of the above elements is commonly focused on the effects of one hazard, disregard-ing the potential interaction effects of diverse hazards in a multi-hazard environment. In this context, there is an increasing need for reasonable and effective evaluation of the multi-hazard risk of transportation infrastructure. Research project INFRARES aspires to bridge this gap by gaining further insight into the risk assessment of bridges and tunnels of transportation net-works in Greece, when subjected to separated and subsequent hazards, with particular empha-sis being placed on seismic and flood hazards. The present paper briefly presents a unified methodology to homogenize the single seismic and flood hazard scenarios and develop appro-priate single- and multi-hazard maps for Greece to be used in risk assessment of roadway net-works. Seismic hazard data, referring to rock site conditions, developed within the SHARE re-search project (www.share-eu.org), is initially selected and is properly amplified to account for site effects, by employing a simplified Vs,30 model originating from morphology and topography data of each region in Greece. The seismic hazard is estimated for a return period of 475 years. Flood hazard zones are derived for whole Greece using newly developed data from the Joint Research Center of the European Commission (https://data.jrc.ec.europa.eu/dataset) for the 100-years return period scenario. Using the above input, both single hazard and multiple haz-ard models are developed and provided in terms of maps in GIS format. The model developed within this study is expected to be a valuable contribution towards the generation of a uniform multiple hazard model for the risk assessment of critical elements of transportation infrastruc-ture in a multi-hazard environment.

Full text

COMPDYN 2021
8th ECCOMAS Thematic Conference on
Computational Methods in Structural Dynamics and Earthquake Engineering
M. Papadrakakis, M. Fragiadakis (eds.)
Streamed from Athens, Greece, 27–30 June 2021
TOWARDS A UNIFIED SEISMIC- FLOOD- HAZARD MODEL FOR
RISK ASSESSMENT OF ROADWAY NETWORKS IN GREECE
Anna C. Karatzetzou1, Sotiria P. Stefanidou2, Stefanos P. Stefanidis3, Grigorios K.
Tsinidis4, Dimitrios K. Pitilakis 5
1, 2, 4, 5 School of Civil Engineering, Aristotle University of Thessaloniki, 54124, Greece
e-mail: {akaratze, ssotiria, gtsinidi, dpitilak}@civil.auth.gr
3 Faculty of Forestry and Natural Environment, Aristotle University of Thessaloniki, 54124, Greece
e-mail: ststefanid@gmail.com
Abstract
Roadway networks, playing a vital role in the economic prosperity of modern societies. Re-
cent hazardous events in Greece, for instance, the 2021 Thessaly earthquake and floods and
the heavy 2019 rainfall in Crete, have demonstrated the vulnerability of roadway networks to
natural hazards, resulting in severe physical damage and important economic and societal
losses. Severe damage on bridges and tunnels of roadway networks is commonly related to
the effects of multiple hazards that may act independently during their life. However, the lit-
erature on risk assessment of the above elements is commonly focused on the effects of one
hazard, disregarding the potential interaction effects of diverse hazards in a multi-hazard en-
vironment. In this context, there is an increasing need for reasonable and effective evaluation
of the multi-hazard risk of transportation infrastructure. Research project INFRARES aspires
to bridge this gap by gaining further insight into the risk assessment of bridges and tunnels of
transportation networks in Greece, when subjected to separated and subsequent hazards, with
particular emphasis being placed on seismic and flood hazards. The present paper briefly
presents a unified methodology to homogenize the single seismic and flood hazard scenarios
and develop appropriate single- and multi-hazard maps for Greece to be used in risk assess-
ment of roadway networks. Seismic hazard data, referring to rock site conditions, developed
within the SHARE research project (www.share-eu.org), is initially selected and is properly
amplified to account for site effects, by employing a simplified Vs,30 model originating from
morphology and topography data of each region in Greece. The seismic hazard is estimated
for a return period of 475 years. Flood hazard zones are derived for whole Greece using new-
ly developed data from the Joint Research Center of the European Commission
(https://data.jrc.ec.europa.eu/dataset) for the 100-years return period scenario. Using the
above input, both single hazard and multiple hazard models are developed and provided in
terms of maps in GIS format. The model developed within this study is expected to be a valua-
ble contribution towards the generation of a uniform multiple hazard model for the risk as-
sessment of critical elements of transportation infrastructure in a multi-hazard environment.
Keywords: Natural hazards, earthquakes, floods, multi-hazard risk assessment.

Anna C. Karatzetzou, Sotiria P. Stefanidou, Stefanos P. Stefanidis, Grigorios K. Tsinidis, Dimitrios K. Pitilakis
1 INTRODUCTION
The reliability of roadway networks and their components, exposed to multiple natural
hazards, is on the frontline of engineering research during the last three decades since poten-
tial damage on their critical components (e.g., bridges and tunnels) is strongly related to im-
portant direct and indirect economic losses. In this context, enhancing the resilience of
roadway networks is key for a safety and an economic viewpoint. Disaster resilience is de-
fined by the National Academies as “the ability to prepare and plan for, absorb, recover from,
and more successfully adapt to adverse events,” while “enhanced resilience allows better an-
ticipation of disasters and better planning to reduce disaster losses – rather than waiting for
an event to occur and paying for it afterward” [1]. To achieve such enhanced resilience, civil
infrastructure systems must not only survive natural disasters, but also recover to functional
levels within acceptable time and cost limits.
The spatial extent of most civil infrastructure systems, including roadway networks, and
the disparity of their elements make them susceptible to a wide range of natural hazards.
Bridges and tunnels are considered to be the most critical components of urban and interurban
transportation systems, and as such should ensure mobility and intercity connection after ex-
treme hazard events. Bridge damage may cause significant disruption to a transportation sys-
tem, resulting in severe substantial direct and indirect losses; for instance, the Loma Prieta
1989 earthquake resulted in more than 40 deaths due to bridge damage and $1.8 billion mone-
tary direct losses due to damage to the transportation infrastructure [2]. Flood due to heavy
rainfall may result in substantial losses, as well; for example, in 2007, heave rainfall in the
UK affected the road network with estimated cost £60 million (The Parliamentary Office of
Science and Technology, Post Note Number 362, October 2010). Extreme weather conditions,
associated with recorded climate changes, e.g., floods and extreme temperatures, are expected
to worsen the performance of many bridges in the near future [3]. Damage on bridges due to
extreme weather conditions in Greece (e.g., reported damage due to heavy rainfall in Trikala
in 2016 and in Crete in2019) is more frequently recorded during the last years. Βridge damage
related to flood, scouring and ground failures may result in collapse and traffic disruption.
Although to a lesser extent, natural hazards may result in damage on tunnels as well. For in-
stance, a large number of mountain tunnels suffered significant damage during the 1999 Chi-
Chi earthquake in Taiwan, as well as during the 2008 Wenchuan earthquake in China [4].
During the last 30 years, several methods have been developed for the assessment of per-
formance and vulnerability of bridges [5] and tunnels [4, 6] against seismic and flood hazard
[7]. Recognizing the significant effects of multiple hazards, as well as of climate change, on
the vulnerability of civil infrastructure, the research interest has been recently shifted upon the
derivation of multi-hazard fragility curves [8]. However, the lack of knowledge in this field
remains significant, when referring to transportation infrastructure, including roadway net-
works.
Regardless of the examined system or element at risk, one of the most critical steps of any
multi-risk assessment methodology is the appropriate definition of multiple hazard scenarios
under which the examined system or element may be subjected throughout its life.
Based on the above considerations, the main objective of the present paper is to present
briefly a framework for the development of combined seismic- flood-hazard scenarios to be
used for the risk assessment of critical elements of roadway networks, i.e., bridges and tunnels,
referring to whole Greece. The proposed framework helps towards a unified seismic- flood
hazard model, which will be used in within the research project INFRARES
(https://www.infrares.gr/) that aims at assessing the risk and resilience of bridges and tunnels
of roadway networks in Greece against the aforementioned hazards.

Anna C. Karatzetzou, Sotiria P. Stefanidou, Stefanos P. Stefanidis, Grigorios K. Tsinidis, Dimitrios K. Pitilakis
2 INFRARES PROJECT
A comprehensive methodology for the risk and resilience assessment of roadway networks
in a multi-hazard environment, will be developed in the framework of INFRARES, focusing
on bridges and tunnels. To meet the objectives of the project, various methodological frame-
works will be used, associated with the following steps enclosed in the definition and assess-
ment of risk: (i) exposure: an inventory of crucial elements of transportation systems, i.e.,
bridges and tunnels, which may be affected by diverse natural hazards, will be developed ac-
counting for typologies found commonly in Greece. (ii) Multi-Hazard assessment: various
scenarios of distinct and multiple natural hazards will be defined, focusing on earthquakes and
floods since these hazards are considered more relevant for the risk assessment of the trans-
portation infrastructure in Greece. This step will include also the definition of appropriate
measures to describe the intensity of examined hazards. (iii) Vulnerability assessment: the de-
gree of loss on the given element or set of elements at risk, when subjected to a specific natu-
ral hazard or to a combination of diverse hazards will be calculated, by employing
comprehensive numerical analyses of the selected elements, while accounting thoroughly for
the effects of ageing-related degradation phenomena of the elements, as well as of Soil-
Structure Interaction (SSI) effects.
Α fully parametrized software will accompany the methodology, allowing for its easier ap-
plication by providing the provided time-dependent, multi-hazard fragility curves for roadway
bridges and tunnels. Figure 1 presents a first draft of the general flowchart of the methodology
that is being developed in the framework of the INRARES project.
Figure 1: Flowchart of INFRARES project

Anna C. Karatzetzou, Sotiria P. Stefanidou, Stefanos P. Stefanidis, Grigorios K. Tsinidis, Dimitrios K. Pitilakis
In particular, INFRARES is expected to contribute on the State of the Art by providing an
innovative methodology for the generation of multi-hazard maps, to be used for the definition
of relevant scenarios in the framework of multi-hazard risk assessment of transportation or
other civil infrastructure. In addition, new, analytical fragility curves and functions will be
developed for various typologies of bridges and tunnels, and for distinct and/or combined
hazards, considering in the latter case combinations of hazards that are relevant for Greece.
Moreover, the damage state definitions within fragility analysis of bridges and tunnels will be
case- and hazard-specific, considering different failure modes and damage mechanisms, to fill
relevant knowledge gap. Finally, a resilience index will be proposed for bridges and tunnels
typologies, as well as for roadway networks, referring to distinct hazards and various hazard
combinations. In this context, the index will be appropriately modified so that to be applicable
for multi-hazard assessment purposes. The methodology and the software could be used for
rapid and rigorous pre- or post-event assessment of infrastructure or for post-event risk man-
agement, constituting very useful tools for stakeholders, operators, consultancies and public
authorities.
In the following sections, we focus on one of the main steps of the methodology, namely the
multi-hazard assessment. More specifically, we present an innovative methodology for the
generation of multiple seismic- flood- hazard models in GIS format.
3 MULTIPLE HAZARD SCENARIOS FOR TRANSPORTATION
INFRASTRUCTURE IN GREECE
In the framework of the multi-hazard assessment approach adopted herein, diverse natural
hazards and combined hazards scenarios are carefully selected, prioritized on the basis of the
potential to cause damage on transportation infrastructure and further examined. After a thor-
ough literature review and lessons learned from past hazards events, earthquakes and floods
are selected as the most critical natural hazards for the assessment of the transportation infra-
structure in Greece. Initially, each natural hazard is studied separately, while in a second
phase we examine the combination of the two hazards, proposing a unified framework for
multi-hazard scenarios. Both single or multiple hazards scenarios are visualized in the form of
maps in GIS format, referring to whole Greece.
The main steps of the herein proposed framework are:
Step 1: Establish a ranking of the different types of natural hazards, considering potential in-
teractions between them.
Step 2: Identify single-hazard models.
Step 3: Identify multi-hazard models, covering all potential intensities and relevant hazard
interactions.
Step 4: Generate single- and multiple- hazard scenarios and relevant maps in GIS format for
the assessment of transportation infrastructure in Greece.
For the sake of presentation of the framework, in the present paper, we consider for the
seismic hazard the standard design/assessment seismic scenario with a return period equal to
Tms=475 years. For the flood hazard, we examine a design/assessment flood hazard scenario
with a return period equal to Tmf=100 years.

Anna C. Karatzetzou, Sotiria P. Stefanidou, Stefanos P. Stefanidis, Grigorios K. Tsinidis, Dimitrios K. Pitilakis
3.1 Single hazard maps
In the present work, a unified methodology is employed to homogenize the single seismic
and flood hazard scenarios. These single-hazard scenarios are displayed in the form of GIS
maps and can be used for the risk assessment of any examined element. Moreover, they may
be used to and develop appropriate maps for multi-hazard scenarios as discussed in section
3.2.
Seismic hazard data for rock site conditions, developed within the EU-funded research pro-
ject SHARE (https://www.share-eu.org), is initially selected. This data is subsequently
properly amplified to account for site effects, by employing a simplified Vs,30 model, originat-
ing from morphology and topography data of the examined region. In the present study, the
seismic hazard is estimated for a return period of 475 years.
More specifically, the seismic hazard estimates were extracted from the global seismic
hazard map produced by Pagani et al. [89]. The hazard is expressed in terms of the Peak
Ground Acceleration (PGA, as a fraction of g) for a probability of exceedance of 10% in 50
years (equivalent to a 475-year return period) on rock (average shear-wave velocity down to
30 m - Vs30 = 760 m/s). A detailed description of how the global seismic hazard model was
developed may be found in Pagani et al. [9]. The seismic hazard analysis is performed using
the OpenQuake engine [10], an open-source seismic hazard and risk calculation software de-
veloped, maintained and distributed by the Global Earthquake Model (GEM) Foundation. Re-
cent studies on the use of the European Seismic Hazard ESHM13 [11] may be found in Riga
et al. [12] and Karatzetzou et al. [13].
With reference to the flood hazard, flood hazard zones were derived for one scenario based
on 100 years return period. In particular, a newly developed dataset for Europe referring to
river flood hazard, was derived from the Join Research Center of the European Commission
[14]. The flood hazard zones were based on the combination of river flow data, estimated
through the hydrological model LISFLOOD, while the inundation simulations were per-
formed with the 2D hydrodynamic modelling LISFLOOD-FP. Subsequently, the hazard was
expressed in terms flood extent zones in different flood frequencies (e.g., T=100 means fre-
quency 1-in-100-years).
Figure 2 portray the resulted single hazard maps for Greece for the examined seismic and
flood hazards scenarios, respectively.
3.2 Multi-hazard maps
For the homogenization of the single hazard maps of Greece (i.e., seismic hazard map and
flood hazard map, presented in section 3.1), we used a bivariate scaling system. This qualita-
tive approach is often used to depict pairs of variables, whose mathematical combination
might not be straightforward or possible, for instance, social vulnerability and natural hazards
[15] or seismic and biological hazards [16]. In the herein proposed framework, we define four
thresholds to classify each variable into low, moderate, high and very high hazard. These
thresholds are created automatically in Arc-Gis using the quantile method. Then, we create a
color matrix comprising all the combinations between the two variables. In this study, we
combine the seismic hazard in terms of the PGA for the selected return period (475 years)
with the river flood hazard in the terms of the percentage distribution of flood hazard zones,
respectively for return period equal to 100 years. Figure 3, presents the resulted multiple-
hazard map for Greece, for the examined scenarios.

Anna C. Karatzetzou, Sotiria P. Stefanidou, Stefanos P. Stefanidis, Grigorios K. Tsinidis, Dimitrios K. Pitilakis
Figure 2: (a) Seismic hazard map for a return period equal to Tms=475 years and (b) flood hazard map for a re-
turn period equal to Tmf=100 years for Greece
Figure 3: Bivariate map depicting the combination of seismic (Tms=475 years) and flood (Tmf=100 years) haz-
ard for Greece.

Anna C. Karatzetzou, Sotiria P. Stefanidou, Stefanos P. Stefanidis, Grigorios K. Tsinidis, Dimitrios K. Pitilakis
4 CONCLUSIONS
The present paper presented a unified framework to homogenize the single seismic and
flood hazard scenarios and develop appropriate single- and multi-hazard maps for Greece to
be used in an under-development methodology for the risk assessment of critical elements of
transportation infrastructure in a multi-hazard environment. The work was developed within
the INFRARES research project, which is also briefly presented.
With reference to seismic hazard; seismic hazard data for rock site conditions, developed
within the SHARE research project (www.share-eu.org), was initially selected and is properly
amplified to account for site effects of various regions in Greece, by employing a simplified
Vs,30 model originating from morphology and topography data. The seismic hazard was esti-
mated for whole Greece and for a 475-years return period scenario and plotted in a relevant
map in GIS format.
Regarding flood hazard; flood hazard zones were derived for Greece from a newly devel-
oped database of the Joint Research Center of the European Commission [14] for one scenario
based on 100 years return period. The resulted flood hazard was estimated for whole Greece
and plotted in a relevant map in GIS format.
A bivariate scaling system was used for the homogenization of single-hazard maps refer-
ring to the above hazards. For the development of the multi-hazard map, four thresholds were
automatically created in Arc-Gis to classify each hazard into low, moderate, high and very
high level. A color matrix was then created comprising all the combinations between the two
variables-hazards, leading to the creation of the multi-hazard map for whole Greece.
The framework presented within this study is expected to be a valuable contribution to-
wards the generation of a uniform multiple hazard model for the risk and resilience assess-
ment of roadway networks.
ACKNOWLEDGEMENTS
The work reported in this paper was carried out in the framework of the research project
INFRARES ‘‘Towards resilient transportation infrastructure in a multi‐hazard environment’’
(https://www.infrares.gr/), funded by the Hellenic Foundation for Research and Innovation
(HFRI) and General Secretariat for Research and Innovation (GSRI) under the “2nd Call for
H.F.R.I. Research Projects to support Post-Doctoral Researchers”. (Project Number: 927).
REFERENCES
[1] S. L. Cutter & S. Carolina, Vulnerability to environmental hazards, Progress in Human
Geography, 20(4), 529-539, 1999.
[2] Y. Zhang, W. F. Cofer, D. I. Mclean, Analytical Evaluation of Retrofit Strategies for
Multicolumn Bridges. Journal of Bridge Engineering, ASCE, 4(2), 143-150, 1999.
[3] F. Nemry & H. Demirel, Impacts of Climate Change on Transport: A focus on road and
rail transport infrastructures Impacts of Climate Change: A focus on road and rail
transport infrastructure. JRC Report, Luxembourg, European Union, 2012.

Anna C. Karatzetzou, Sotiria P. Stefanidou, Stefanos P. Stefanidis, Grigorios K. Tsinidis, Dimitrios K. Pitilakis
[4] G. Tsinidis, F. De Silva, I. Anastasopoulos, E. Bilotta, A. Bobet, Y.M.A. Hashash, C.
He, G. Kampas, J. Knappett, G. Madabhushi, N. Nikitas, K. Pitilakis, F. Silvestri, G.
Viggiani, R. Fuentes. Seismic behavior of tunnels: From experiments to analysis. Tun-
nelling and Underground Space Technology, 99, 103334, 2020.
[5] S. P. Stefanidou & A. J. Kappos, A., Methodology for the development of bridge-
specific fragility curves. Earthquake Engineering & Structural Dynamics, 46, 73-93,
2017.
[6] S. Argyroudis, G. Tsinidis, F. Gatti, & K. Pitilakis, Effects of SSI and lining corrosion
on the seismic vulnerability of shallow circular tunnels. Soil Dynamics and Earthquake
Engineering, 98, 244-256, 2017.
[7] Lee, Jaebeom, Lee, Young-Joo, Kim, Hyunjun, Sim, Sung-Han, and Kim, Jin-Man. “A
New Methodology Development for Flood Fragility Curve Derivation Considering
Structural Deterioration for Bridges.” Smart Structures and Systems 17, no. 1 (January
25, 2016): 149–65. doi:10.12989/SSS.2016.17.1.149.
[8] P. Gehl, & D. D’Ayala, System loss assessment of bridge networks accounting for mul-
ti-hazard interactions. Structure and Infrastructure Engineering, 2479, 1–17, 2018.
[9] Μ. Pagani, J. Garcia-Peláez, R. Gee, KL Johnson, V. Poggi, V. Silva, M. Simionato,
RH Styron, D. Viganò, L. Danciu, D. Monelli, G. Weatherill G, The 2018 version of the
global earthquake model: hazard component. Earthquake Spectra, doi:
10.1177/8755293020931866., 2020.
[10] M. Pagani, D. Monelli, G. Weatherill, L. Danciu, H. Crowley, V. Silva, P. Henshaw, L.
Butler, M. Nastasi, L. Panzeri, M. Simionato, D. Vigano, OpenQuake-engine: an open
hazard (and risk) software for the global earthquake model. Seismological Research Let-
ters, 85, 692-702, 2014.
[11] J. Woessner, L. Danciu, D. Giardini, H. Crowley, F. Cotton, G. Grünthal, G. Valensise,
R. Arvidsson, R. Basili, M.B. Demircioglu, S. Hiemer, C. Meletti, RMW Musson, AN
Rovida, K. Sesetyan, M. Stucchi, The SHARE Consortium (2015). The 2013 European
Seismic Hazard Model: Key Components and Results, Bulletin of Earthquake Engineer-
ing, 13(12), 3553-3596, 2015.
[12] E. Riga, A. Karatzetzou, S. Apostolaki, H. Crowley, K. Pitilakis. Verification of seismic
risk models using observed damages from past earthquake events. Bulletin of Earth-
quake Engineering, doi: 10.1007/s10518-020-01017-5, 2021.
[13] A. Karatzetzou, E. Riga, S. Apostolaki, S. Karafagka, K. Pitilakis, E. Lekkas. A Priori-
tization Scheme for Seismic Intervention in School Buildings. The case of Thessaloniki,
Greece. Proceedings of the 7th International Conference on Civil Protection & New
Technologies (SAFE GREECE 2020), 14-16 October, on-line, 2020.
[14] F. Dottori, L. Alfieri, A. Bianchi, J. Skoien, & P. Salamon, A new dataset of river flood
hazard maps for Europe and the Mediterranean Basin region. Earth System Science Da-
ta Discussions, 1-35, 2021.
[15] C. Emrich and S. Cutter, Social vulnerability to climate-sensitive hazards in the south-
ern United States. Weather, Climate, and Society, 3(3),193-208, 2011.
[16] V. Silva and N. Paul, Potential impact of earthquakes during the 2020 COVID-19 pan-
demic. Earthquake Spectra, 37(1) 73-94, doi: 10.1177/8755293020950328, 2020