Journal Pre-proof A multi-hazard risk model with cascading failure pathways for the Dawlish (UK) railway using historical and contemporary data Keith Adams, Mohammad Heidarzadeh PII: S2212-4209(21)00048-0 DOI: https://doi.org/10.1016/j.ijdrr.2021.102082 Reference: IJDRR 102082 To appear in: International Journal of Disaster Risk Reduction Received Date: 10 July 2020 Revised Date: 19 January 2021 Accepted Date: 23 January 2021 Please cite this article as: K. Adams, M. Heidarzadeh, A multi-hazard risk model with cascading failure pathways for the Dawlish (UK) railway using historical and contemporary data, International Journal of Disaster Risk Reduction, https://doi.org/10.1016/j.ijdrr.2021.102082. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2021 The Author(s). Published by Elsevier Ltd. 1 A multi-hazard risk model with cascading failure pathways for 2 the Dawlish (UK) railway using historical and contemporary data 3 4 Keith Adams* a, Mohammad Heidarzadeh a 5 6 a Department of Civil and Environmental Engineering, 7 Brunel University London, Uxbridge UB8 3PH, UK 8 9 Confidential manuscript submitted to: 10 “International Journal of Disaster Risk Reduction” 11 Original submission: 10 th July 2020 12 13 14 15 16 * CorrespondenceJournal to: Pre-proof 17 Mr Keith Adams, 18 Doctoral Researcher, 19 Department of Civil & Environmental Engineering, 20 Brunel University London, 21 Uxbridge, UB8 3PH 22 UK 23 Tel: +44-1895-265756 24 Email: [email protected] 25 Webpage: https://www.brunel.ac.uk/people/keith-adams1 1 26 Abstract 27 The failure of the vital economic railway link between London and the southwest of the United 28 Kingdom in the 2014 storm chain incurred up to £1.2bn of economic losses. This incident highlighted 29 the urgent need to understand the cascading nature of multi hazards involved in storm damage. This 30 study focuses on the Dawlish railway where a seawall breach caused two months of railway closure in 31 2014. We used historical and contemporary data of severe weather damage and used failure analysis 32 to develop a multi-hazard risk model for the railway. Twenty-nine damage events caused significant 33 line closure in the period 1846-2014. For each event, hazards were identified, the sequence of failures 34 were deconstructed, and a flowchart for each event was formulated showing the interrelationship of 35 multiple hazards and their potential to cascade. The most frequent damage mechanisms were 36 identified: (I) landslide, (II) direct ballast washout, and (III) masonry damage. We developed a risk 37 model for the railway which has five layers in the top-down order of: (a) root cause (storm); (b) force 38 generation (debris impact, wave impact, overtopping, excess pore pressure, wind impacts); (c) 39 common cause failure (slope instability, rail flooding, coping and parapet damage, foundation failure 40 and masonry damage); (d) cascading failure (landslide, ballast washout, upper masonry seawall 41 failure, loss of infill material), and (e) network failure forcing service suspension. We identified five 42 separate failure pathways and damage mechanisms by analysing these 29 major events. 43 Journal Pre-proof 44 Keywords: Storm Surge; Cascading Failure Pathway; Dawlish Railway; Multi-hazard; Structural 45 Vulnerability; Damage Mechanism. 46 2 47 1. Introduction 48 The United Kingdom has nearly 16,000 km of open rail routes [1] with a significant proportion of 49 coastal alignment. Most of these are strategically important as they often are the only regional rail 50 connection, or they provide logistical support to critical national infrastructure. For instance, the 51 Cumbrian line in the lake district in north-west England (Figure 1c, d), is a vital link providing public 52 transport as well as freight capacity to the nuclear reprocessing plant at Sellafield. The area is a 53 UNESCO world heritage site and a major tourist destination [2]. In Wales, the south and west coastal 54 railways (Figure 1d) were instrumental in providing infrastructure to support the coal mining and 55 shipping businesses of the nineteenth century [3]. Where coastal railways are subject to direct wave 56 action, they are particularly vulnerable to climate change effects including sea level rise (SLR) and 57 increases in storminess and rainfall [4]. Recent studies have indicated that there is a 20% increase in 58 the number and intensity of storms [5, 6] while Castelle et al. [7] showed that the winter ‐mean wave 59 height has increased in recent years. The latest marine climate projections for the UK [8] predict mean 60 sea level rise between 0.39 m and 0.70 m by 2100 dependant on emission scenarios. Network Rail, the 61 infrastructure owner in the UK, has acknowledged weather resilience and climate change as a major 62 risk to future operations, and in response has produced a series of adaptation plans. The latest is the 63 “Second Climate Change Adaptation Report ” [9], with a third due in 2021 [10]; the organisation has 64 also contributed to the Journal“Tomorrow's Railway and Pre-proof Climate Change Adaptation” research programme 65 [11]. 66 The section of railway between Exeter and Newton Abbott, UK (Figure 1a) is particularly 67 vulnerable to climate change effects. In 1845, a vertical seawall was built along the coastal margin to 68 support the railway alignment in Dawlish [12]. Soft red sandstone cliffs were blasted along the 69 coastline, with the unconsolidated material used to backfill the area between the cliff face and the 70 frontage of the masonry seawall. Following the second report by Beeching [13], this section of 71 railway became part of the Great Western mainline strategically connecting London to the south west 72 of England. This sole vital economic link was broken during the February 2014 storm when 73 successive deep extra-tropical cyclonic storms caused strong winds and violent waves [14] to cause 3 74 structural failure on the seawall (Figure 1b, 4, 7d) and precipitated a route closure that lasted two 75 months [15]. The cost of reinstatement was £50m with associated economic losses estimated at up to 76 £1.2bn for the South West region [16]. 77 In this research, we study historical and contemporary data on the failure of the Dawlish railway 78 by storm induced forces to establish a multi-hazard risk model with cascading failure pathways 79 (FPW) which could be used with an exposure database to evaluate risk to the structural assets. We use 80 primary historical accounts and identify damage mechanisms (DM) and FPW that cascade between 81 separate DM. The innovation in this study is the identification of storm initiated multi-hazards and the 82 development of cascading structural vulnerabilities of rail network infrastructure in the UK. To our 83 knowledge, this is one of the pioneering multi-hazard risk models with cascading FPW for rail 84 networks in coastal settings. The combination of historical damage data with contemporary 85 engineering understanding of cascading risk is a particular strength of this research. A multi-hazard 86 risk model such as this would be beneficial for improving the resilience of the railway network to 87 severe weather events by providing a tool that predicts possible FPW to inform future engineering 88 interventions. Journal Pre-proof 4 89 90 Journal Pre-proof 91 Figure 1. a) The location of the railway at Dawlish. b) Major damage after the February 2014 storm 92 (Matt Clark, Met Office UK) . c) The coastal railway near Sellafield in Cumbria. d) Map of Great 93 Britain showing the extent of coastal railways (red lines). 94 2. Literature Review 95 Differences of language used in Disaster Risk Management studies by diverse disciplines has led 96 to an effort for harmonisation in terminology as identified by Kappes et al. [17] and more recently 97 Monte et al. [18]. For clarity we define the terms used in this paper in Table 1. 5 98 Risk () is defined by the Intergovernmental Panel for Climate Change [19] as the product of 99 hazard () and consequence (): 100 = (1) 101 Civil and Structural Engineers often define risk by expanding the definition of consequence as the 102 product of structural vulnerability () and exposure () [20]: 103 = (2) 104 In this study we have developed a model which combines the hazard and vulnerability of the 105 engineering assets, respecting the intrinsic link between hazardous force generation and propensity for 106 damage as represented by structural vulnerability. The benefit of this approach is that the exposure 107 has been decoupled from the structural vulnerability. This allows multiple stakeholders to interrogate 108 the model using their own exposure metrics to provide tailored evaluations of overall risk (e.g. 109 insurance providers for loss quantification, Network Rail for maintenance and capital expenditure and 110 local authorities for disaster risk planning). This infers the risk model proposed is specific to similar 111 hazard events (i.e. extratropical cyclonic storms temporally coincident with high spring tides and 112 strong onshore winds) for infrastructure with common characteristics (coastal railways with vertical 113 masonry seawalls subject to direct wave action). 114 The challenges of Journalanalysing multi-hazards havePre-proof been comprehensively reviewed and discussed 115 by Kappes et al. [17] who make the relation between multi-hazard analysis and the objective of risk 116 reduction.