The impact of climate change on urban transport resilience in a changing world

Abstract

The assessment of the potential impact of climate change on transport is an area of research very much in its infancy, and one that requires input from a multitude of disciplines including geography, engineering and technology, meteorology, climatology and futures studies. This paper investigates the current state of the art for assessments on urban surface transport, where rising populations and increasing dependence on efficient and reliable mobility have increased the importance placed on resilience to weather. The standard structure of climate change impact assessment (CIA) requires understanding in three important areas: how weather currently affects infrastructure and operations; how climate change may alter the frequency and magnitude of these impacts; and how concurrent technological and socio-economic development may shape the transport network of the future, either ameliorating or exacerbating the effects of climate change. The extent to which the requisite knowledge exists for a successful CIA is observed to decrease from the former to the latter. This paper traces a number of developments in the extrapolation of physical and behavioural relationships on to future climates, including a broad move away from previous deterministic methods and towards probabilistic projections which make use of a much broader range of climate change model output, giving a better representation of the uncertainty involved. Studies increasingly demand spatially and temporally downscaled climate projections that can represent realistic sub-daily fluctuations in weather that transport systems are sensitive to. It is recommended that future climate change impact assessments should focus on several key areas, including better representation of sub-daily extremes in climate tools, and recreation of realistic spatially coherent weather. Greater use of the increasing amounts of data created and captured by ‘intelligent infrastructure’ and ‘smart cities’ is also needed to develop behavioural and physical models of the response of transport to weather and to develop a better understanding of how stakeholders respond to probabilistic climate change impact projections.

Keywords

CCIA Climate Change Impact Assessment climate change climate projections socio-economic scenarios transport transport meteorology

I Background and introduction

In most cases transport is directly exposed to the atmosphere (Thornes, 1992), and as a result its activities are often subjected to various types of disruptive and hazardous weather. Transport disruption can have large-scale impacts on the economy and human life, as demonstrated by several notable events. These include the widespread and sustained flooding in Thailand in 2011, which caused loss of life as well as severe disruption to the global supply chain of many electronic components (AON Benfield, 2012), and Hurricane Sandy in October 2012 (Aerts et al., 2013), which among other impacts caused the complete shutdown of the New York subway system as a precautionary measure. The trend in the expectation of the public for efficient, safe and effective transport every day of the year and under all but the most extreme meteorological conditions was noted by Thornes (1992) in a review of the impacts of weather and climate in the UK, and this expectation has not ceased, with an observed emphasis on reducing travel time and hence increasing speeds (Banister, 2011).

Transport has become an increasingly important enabler and driver for the globalized economy, allowing the movement of goods and workers (Eddington, 2006) as well as being a social asset, particularly in the maintenance of spatially extended social networks (Cresswell, 2006). However, climate change has the potential to impact greatly on the efficiency, safety and cost of transport in a myriad of direct and indirect ways. The focus of this paper is the direct effect climate change may have on transport operations and disruption, and the progress of the research within each of the necessary components (Jaroszweski et al., 2010) to create holistic and plausible impact projections. Tol (1998) gives a useful summary of the basic conceptual framework of the Climate Change Impact Assessment (CIA) process, emphasizing that, although exact methodological details vary, to create robust and holistic visions of future impacts a CIA will ideally do three things: (1) determine the current relationships between weather and climate and the given sector; (2) project impacts using these relationships and climate change scenarios; and (3) modify these impacts with scenarios of how the sector or system might develop in the future. The final step is deemed to be critical for transport where the network, infrastructure, technology, vehicles, modal split and driver behaviour determine the exposure and sensitivity to weather (Jaroszweski et al., 2010). However, this consideration has mostly been neglected in transport-based CIA.

This review looks at the existing transport research that relates to the three components of CIA identified by Tol (1998). In particular, it attempts to identify research that is pertinent to the assessment of impacts on transport in urban areas. It is estimated that 50% of the world’s population currently live in urban areas. By 2050 this is projected to increase to 67% globally and to 86% in more developed regions (UN, 2012). This will enhance the already considerable role of cities as engines for national economies (UN, 2011), and highlights the importance of the growing research agenda around studies on urban areas, across many disciplines but especially in the climate change impacts community. The joint effects of projected climate change (such as the projected changes in the temporal and spatial distribution of rainfall and increased heatwaves; IPCC, 2013) and the rise in urban population makes an effective CIA methodology for urban weather-related transport impacts essential. This is compounded by the fact that cities and urban areas have distinct microclimatic features associated with both their material and topographical properties and the human and industrial activities that take place within them (Oke, 2006). The best known of these is the ‘urban heat island’ effect which creates higher temperatures compared to surrounding non-urbanized areas (Oke, 1973). Other effects include rainfall modification caused by the hygroscopic properties of aerosols emitted through industrial activity, which in the case of suppression allows water vapour to be spread over more cloud condensation nuclei without being precipitated out of the atmosphere (Rosenfeld, 2000).

Successful adaptation plans (Koetse and Rietveld, 2012; Lindgren et al., 2009) require realistic projections of impacts. Although several reviews have focused on either the impact of climate change on transport as a whole (Koetse and Rietveld, 2009) or for individual modes such as rail or road freight (e.g. Baker et al., 2010; Hooper and Chapman, 2012; Jaroszweski, 2012), it is important to review concisely the current knowledge about the relationships between weather and transport and the trends within CIA that influence how impact assessments are carried out. Key to this is understanding what the impact of weather on transport currently is, and this is explored in section II of this review. Where CIA has been carried out and climate impact projections produced, this is commented upon. Section III discusses the methodological requirements for improved transport-based CIAs. In investigating the impact of climate change on cities, this paper does not seek to comment explicitly on potential adaptation measures, but concurrent socio-economic change and increased adaptive capacity (the individual or institutional ability to adapt) are referred to.

II Current and future vulnerability to weather and climate

This section reviews both the current understanding of relationships between weather and transport and any existing climate change impact projections for each of the current major modes of urban surface transportation. In all modes there are broadly two types of weather impacts that can cause transport disruption: (1) those that occur as a result of behavioural changes to driver/agent caused by the additional stress brought about by weather, such as rain-induced speed reductions (Hooper et al., 2012); and (2) those that occur from a physical failure of the infrastructure, or the triggering of natural hazards, such as heat-induced rail buckling (e.g. Dobney et al., 2010). In practice, many failures occur as a result of a combination of a temporarily degraded physical operating environment and behavioural influences put on the driver, such as rain or ice-related road traffic accidents (e.g. Andersson and Chapman, 2011).

1 Road transport

a Traffic speed, activity and flow

Weather has behavioural effects which influence the propensity for travel and the speed of traffic flows in urban areas. Reductions in speed due to various adverse weather conditions have been identified in a number of studies. For example, Hooper et al. (2012) correlated traffic speed data for UK motorways with weather radar data to determine the negative effect of rainfall. Although based on the motorway route between London and Glasgow in the UK, this study included several sections which passed through urban areas. These sections form important components of inter- and intra-urban transport infrastructure. The large traffic volumes in these sections were used to explain the relatively large speed reductions during rain events compared to non-urban areas. A similar weather radar approach was used by Stern et al. (2004) in Washington DC to determine increases in journey time using publicly available online travel information. Agarwal et al.’s (2006) study on the effects of weather on urban freeway traffic operations in Minneapolis-Saint Paul uncovered strong links between reductions in speed when there is snowfall or when visibility is reduced during rain events, inducing a reduction of up to 15%. While wind is known to affect traffic, there is less understanding of the impact, although it seems that this too has a detrimental effect with studies identifying a reduction in speed of more than 10 km/h when high winds occur (Kyte et al., 2000).

Travel activity is strongly affected by a number of weather conditions, including sunshine, temperature and wind speed, and has a strong literature including observational and behavioural studies. In their wide-ranging study on the effects of weather on urban road transport in Melbourne, Keay and Simmonds (2005) identified that the greatest reductions in traffic flow due to precipitation occur at night, mostly because these journeys are largely discretionary. Travel behaviour can also be influenced by weather forecasts, with Cools and Creemers (2013) indicating that a forecast of snow elicited the highest likelihood of trip cancellation or change of destination in respondents surveyed in Belgium. Weather can also determine the level of activity of emergency services and also their ability to reach their destinations in time. Thornes et al. (2013) used ambulance call-out data for Birmingham over a five-year period along with meteorological station data to determine the relationships between weather and response times. It was found that extreme temperatures increase response times, partly as a result of increased emergencies but also due to adverse road conditions in the winter.

Precipitation causes varying degrees of capacity reduction. Capacity can be defined as the maximum traffic flow possible on a given section of road per hour. This can be reduced either through flood water directly blocking road lanes, or through a reduction in capacity caused by either the autonomous speed reductions made by individual drivers or through traffic control systems. Studies investigating the impact of rainfall on capacity in the USA suggest that the reduction can vary from 4% to 30% depending on precipitation intensity (Agarwal et al., 2006). Congestion, the reduction of speeds associated with traffic demand exceeding road capacity, is often a problem during adverse weather conditions with rain and snowfall being identified as weather variables likely to increase congestion (Chung, 2012). Indeed, Ibrahim and Hall (1994) observed that maximum flow was reduced during adverse weather conditions, in turn causing congestion to increase and journey times to be prolonged, again through reactive behavioural change on the part of the driver.

No study has used the quantitative models of the type produced by the studies above with climate scenarios to project the impact of climate change. However, the fact that many of these relationships are derived in cities, or include sections of urban and rural roads (e.g. Hooper et al., 2012), means that climate change impact assessment would be possible, providing climate projections at a suitable temporal and spatial resolution are available.

b Accidents

Several city-based studies in many countries have investigated the impact of weather on road accidents (e.g. Andreescu and Frost, 1998; Andrey and Yagar, 1993). These usually link accident counts for a given city or region to meteorological data, either from a meteorological station or a road weather information system (RWIS). The impact of weather is usually expressed as a relative accident rate (RAR), which is obtained by comparing the accident numbers during a period experiencing a hazardous weather type to an equivalent period (typically the corresponding day a week or fortnight before or after the event) in the absence of the hazard, known as matched-pairs analysis. Other studies use regression, where precipitation amount is compared to actual accident numbers. This requires a greater amount of information about confounding factors such as traffic volumes (Eisenberg, 2004).

Precipitation has been the subject of the majority of studies in the literature (Qiu and Nixon, 2008). Loss of tyre friction due to road surface ice is also associated with increased urban accident rates. This can be compounded by rain or sleet which falls on frozen road surfaces (Eriksson, 2001). Integration of information about the road network such as local topography and sky view factor (a measure of exposure) that affect the local microclimate have allowed the modelling of road surface temperatures which have in turn improved winter road maintenance (Chapman and Thornes, 2003). Potential improvements to these tools include the modelling of the warming effect that vehicles themselves have on road surface temperatures (Prusa et al., 2002).

Crosswinds are a particular hazard for high-sided vehicles (Baker, 1993; Baker and Reynolds, 1992). This has been investigated using weather station data, with gust speed and direction being key determinants of accident numbers (Young and Liesman, 2007). Although this is mainly a problem on exposed stretches of road and bridges (Chen and Cai, 2004), the built urban form can induce high street-level winds which can create a hazardous wind environment for high-sided vehicles (Gao et al., 2012). Wind-blown dust has been shown to be a contributing factor in road accidents in cities in arid regions such as Riyadh City (Nofal et al., 1996). This can be exacerbated by heat stress on drivers in conditions of extreme heat (Nofal and Saeed, 1997).

The severity of injuries associated with different weather types has been shown to vary due to the type of driving behaviour they promote. For example, snow tends to increase the number of damage-only accidents and reduces the severe injury/fatality rate as drivers reduce their speeds. A strong behavioural element is revealed in terms of the effect the frequency of exposure to weather has on the ability of drivers to cope in these conditions. Evidence of this is given by Brüde and Larson’s (1980) study in Sweden which showed that the relative accident rate on days with snow or ice cover were greater in regions which experience those conditions less often. In this example, the average percentage of winter where snow or ice is present on road surfaces is used as a metric, with relative accident rates highest in those areas which experience snow- and ice-covered roads for a low percentage of the winter. Eisenberg (2004) showed that this phenomenon can be observed over relatively short periods. Using regression rather than matched-pairs analysis and large regions rather than cities, Eisenberg showed that following an initial heavy rain event relative accident rates reduce during subsequent events in a given month. This relates to Elvik’s (2006) laws of accident causation, which suggest that relative accident rates associated with a hazard are proportional to the frequency with which it is encountered, with drivers learning to cope with conditions that they are exposed to frequently. This is encapsulated in the first snow of the season which causes more accidents than subsequent snowfall of similar amounts (Eisenberg and Warner, 2005). Additional evidence for this phenomenon comes from Edwards’ (1998) study that shows that areas which have lower numbers of wind events have a greater injury rate associated with the events.

The strong existing literature on the links between weather and road accidents has made this one of the first areas of transport to be the subject of city-based CIA. Andersson and Chapman (2011) projected large-scale reductions (50%) in ice-related accidents in the West Midlands conurbation of the UK (encompassing Birmingham, Wolverhampton and Coventry) by the 2050s (2040–2069). This study used the EARWIG (Environment Agency Rainfall and Weather Impacts Generator) weather generator to statistically downscale projected overlying climates from the UKCIP02 climate scenarios into daily time series of weather that could be used to elicit projected numbers of low temperature events. However, it was argued that there may be complacency on the part of the road authorities, with a later study using analogues suggesting that individual complacency during milder winters (hence warmer climates) may be an issue (Andersson and Chapman, 2011). It must also be mentioned that this weather generator and its successors such as the UKCP09 Weather Generator are not optimized for cities and do not take into account urban heat island or potential changes in urban form and activities. It is also true that as the weather generator is a statistical tool it does not take into consideration large-scale meteorological dynamics such as blocking, so is not well suited for projecting potential futures where colder winters become more common. Hambly et al. (2013) formed relationships based on matched-pairs analysis between different levels of rainfall (from a meteorological station) and accidents in Vancouver, Canada. Projections were made for future accident numbers using regional climate model (RCM) output. It was commented that climate change will also affect the behaviour of drivers, by either increasing or reducing the frequency with which drivers are exposed to a given weather hazard, hence altering their abilities to cope in these conditions. However, this was argued to have only a minor effect and was not explicitly modelled. Both techniques used here imagine one future climate and do not sample the full range of model uncertainty available in the latest projections such as UKCP09 (Murphy et al., 2009).

c Cycling and walking

Cycling and walking are seen as a green forms of transport, and ones that are often promoted in urban areas to reduce emissions and improve health (Tight et al., 2011). The effect of weather on cycling has been investigated by Miranda-Moreno and Nosal (2011) who determined a mixed relationship between weather and ridership, with ridership increasing with temperature but falling back with very high temperatures. Weather can also contribute to cycling accidents (Kim et al., 2007). Aultman-Hall et al. (2009) looked at the relationship between weather and walking, with precipitation having a negative influence on pedestrian numbers. It is likely that the proliferation of city-based bike-hire schemes will generate greater volumes of data (e.g. Lathia et al., 2012) for use in weather-related behavioural studies. Bocker et al (2013) present a review of the current evidence for the effect of weather conditions on daily travel behaviour.

d Infrastructure

Infrastructure is affected where weather conditions exceed design specifications. These are often based on the observed climate over the previous 30 years. The planning timescale for road infrastructure is currently around 25 years, but can be significantly longer for certain infrastructure such as bridges and underpasses (Peterson et al., 2008). For the road network to cope with a changing climate, it is important to integrate current scientific knowledge into planning, design and construction of new infrastructure, and also into the upgrade of existing infrastructure (Committee on Climate Change and US Transportation, 2008). For example, Lemonsu et al. (2013) studied the urban heat island of Paris and found that although the impact of climate change is to make the urban environment warmer, the urban heat island (the relative difference between urban and rural areas) decreases during the summer during night-time and there is occasional negative urban heat island due to countryside soil dryness.

Concern has already been expressed about how roadside slopes may be affected by climate change (He et al., 2011) with previous studies identifying that increases in the frequency of extreme weather events due to climate change are likely to have a detrimental effect on stability of slopes. This is especially so for short, heavy storm events during both summer and winter which are more likely to trigger slope failures in future (Clarke et al., 2006). Schmidt and Glade (2003) have projected landslides on the road network. Roads built on permafrost in cold regions have been observed to have degraded in recent years (Serreze et al., 2000) and are projected to increase in degradation due to climate change (Zhang et al., 2008). Bridges can also be disrupted by high winds and may be dangerous to use (Macdonald et al., 2003), again related to the increased risk of accidents (Chen and Cai, 2004). There are likely to be a number of problems for the road network as road surfaces are likely to be more susceptible to rutting. This in turn leads to water pooling on the road surface (Erlingsson, 2012) which can further damage the road surface. Bridge scour is a problem for road and railways and has been addressed by Khelifa et al. (2013) who developed a risk assessment model. This was applied to all bridges in the USA that carry vehicle traffic and projected that annual bridge failures may increase by as much as 10% due to climate change.

2 Rail

The weaknesses of rail to weather are largely determined by the type of vehicles, the traction type and whether they run above or below the surface. Weather can affect trackside signalling equipment (Medina et al., 2011). Pore water pressure can affect embankments and make them more likely to fail (Ridley et al., 2004). Line-side fires are also a problem and may increase. Although no specific projections exist for this, Liu et al. (2012) project the impact on forest fires.

Landslides are currently a problem on the rail network (Lloyd et al., 2001). Climate change is likely to impact on the stability of slopes and research has already begun to determine what the effects on slope stability may be due to climate change. Manning et al. (2008) used a novel methodology to assess the impact of climate change on slope stability along a railway line and determined that the spatial distribution of rainfall-induced landslides is likely to change in future. Ballast and sublayer can also be affected by weather (Ferreira et al., 2011).

Gaps in existing knowledge are also identified with the acknowledgement that future research needs to develop understanding of the effects of climate change on trackside vegetation (which may have implications for slope and track bed stability) and identify locations which will be most affected by any rise in sea level (Eddowes et al., 2003).

Subsurface rail travel is seen as a way of achieving efficient travel in cities, and also serves to relieve often congested roads on the surface and reduce externalities associated with road transport such as air pollution (Chen and Whalley, 2012). However, where these lines meet the surface, bricks in tunnels can be frost damaged (Thomachot et al., 2005). Heavy precipitation is likely to be a particular problem for the underground networks in future due to flash flooding making the networks increasingly vulnerable (Hunt and Watkiss, 2011). In the absence of suitable air conditioning (Tol, 1998) underground mass transit systems can be susceptible to high temperatures, decreasing thermal comfort for passengers (Abbaspour et al., 2008).

High temperatures can have significant impacts on the railway network, in particular causing rail buckling. The general increase in temperature in future, coupled with more frequent hot summers, is likely to lead to more instances of rail buckling (Arkell and Darch, 2006; Clarke et al., 2002). Further work by Dobney et al. (2009, 2010) quantified the increase in the number of rail buckles and related delays in the southeast of England, and was later extended to the entire UK indicating nationwide increases, although of a lower magnitude in Scotland (DEFRA, 2012; Palin et al., 2013).

As part of the decarbonization and general cost savings on the railways, there has been a move towards electrification. This adds a new vulnerability to the system as overhead lines are particularly vulnerable to high winds, especially during high temperatures due to line-sag. This can be remedied through closer spacing of gantries and increases in line tension during high temperatures. During high-wind events debris can be blown onto the line which may cause an obstruction or damage to the track. It has been observed that the urban environment provides a greater source of debris and detritus that may be moved during an event. Lightning can also disrupt the electrical supply systems (Theethayi et al., 2005).

III Observations on impact assessment development

1 Behavioural and physical relationships

As demonstrated, the assessment of climate change impacts on transport is very much in its infancy and requires work at all steps in the process. As with any sector, the determination of relationships between weather and transport failure is a fundamental step in forming impact projections (Tol, 1996). This literature review and other mode-specific and general reviews have identified a patchy knowledge base, with certain aspects being well-covered (e.g. accidents), while others are covered only partially (e.g. infrastructure). The effects that these component failures have on the wider transport network are often missing.

The lack of meteorological station observations at a suitable spatial and temporal resolution in the urban area and the inherent lack of representativeness in using single sites presents a barrier to creating realistic impact models. This lack of coverage in urban areas has been observed by Oke (2006), and has begun to be addressed using high-resolution arrays of temperature sensors to account for the variations in land use and topography that create the urban microclimate (Muller et al., 2013). However, the guidelines governing the locations of meteorological stations (WMO, 2008) means that, currently, the availability of representative meteorological records for cities is limited. Many of the studies in the literature such as Andrey et al. (2003) and Hambly (2013) use stations that are sited outside the city, such as at airports, which in the case of Andrey et al. (2003) can be as far as 35 km away. Additionally, where city-based stations are available, their representativeness is questionable. For example, Keay and Simmonds (2005) used a city-based station to represent the meteorological situation in Melbourne, an area of almost 10,000 km². It is likely that there will be significant variations in rainfall timing and amount across the city at any given time. A potential alternative to station data are rainfall radar images, which have previously been used by Hooper et al. (2012) to study the relationship between rainfall and road traffic speeds as well as by Jaroszweski and McNamara (2014) to determine relationships between rainfall and road traffic accidents in London and Manchester, UK. Overeem et al. (2013) demonstrate a technique using the strength of microwave signals from base stations of mobile-phone networks that can create analogous data to those produced by weather radars. This is a potentially useful technique in areas with no weather radar coverage.

The trend towards the creation and capture of more data in cities may also assist with the creation of more realistic impact models. The ‘Smart Cities’ agenda (Batty et al., 2012), the use of data created in cities to better understand their workings, is a potential source of future transport data. For example, previously mentioned bike-hire data (Lathia et al., 2012) could be an important way of understanding the effect of weather on this mode, albeit with the caveat of not capturing all journeys or all types of users. Data captured from smart card systems used for buses, trams, trains and underground networks (e.g. Kusakabe et al., 2010) can be used to determine the relationships between weather and passenger numbers, modal split and travel destination. This will add further evidence to test the theories of behavioural studies (e.g Cools and Creemers, 2013). Data can also be gathered from the developments associated with the emerging ‘Intelligent Infrastructure’ and ‘Internet of Things’ agendas (Wang et al., 2013), the move towards installing sensors in infrastructure to measure their condition.

There are also several currently under-researched ‘autonomous’ behavioural responses to climate change that may modify the projected impact on transportation. It has to be remembered that transport (at least currently) has an element of human control of varying degrees in almost all cases, and thus its response to disruptive and hazardous weather is dependent on the behaviour, decision-making and skill of the drivers, traffic controllers and maintenance engineers. All of the relationships between weather and transport behaviour that have been elicited in the literature (for instance, that between rainfall and traffic speeds) are therefore specific to the location and population from which they are derived. For example, the ability to cope in a given weather hazard is determined in part by the frequency with which a driver is exposed to that hazard (Elvik, 2006). Although there may be a projected decrease in the frequency of a given meteorological hazard, a certain part of the benefit of this will be offset by a reduced ability to cope during these events, as the driver will not have had the opportunities for learning in these conditions. This has previously been demonstrated for snow and ice-road accidents in Sweden (Elvik, 2006), where the relative accident rate during such events is greater in areas that experience this condition less frequently. Quantification of this phenomenon and its integration into climate change impact projections is an important step, as this may provide a significant offset to the types of projections made by Andersson and Chapman (2011), Hambly et al. (2013) and Jaroszweski et al. (2013).

2 Climate change

As transport has seldom been the focus of CIA, the particular requirements it has (e.g. spatially coherent, high-resolution weather generator output) have not been a priority in impact assessment toolkits. Climate change impact assessment on transportation requires a better reproduction of daily and hourly extremes in weather and improved spatial coherence for two main reasons. Firstly, behavioural responses to weather such as speed reductions or road accidents are often caused by relatively high precipitation intensities (Hambly et al., 2013; Hooper et al., 2012). Current generators (e.g. UKCP09; Jones et al., 2009) do not reproduce these extremes. For example, UKCP09 is not recommended for use when studying extremes with return periods great than 1 in 10 years at a daily resolution and 1 in 5 years at an hourly resolution (Jones et al., 2009). This absence of short-term extremes is partly an artifact of the evolution of the climate change impacts community and their changing requirements for climate input to their models. As much early CIA work was concentrated on large-scale agriculture and water resources, extremes at the hourly level were not necessary. Second, transport is vulnerable to multiple failures across the network that remove several nodes at once (Wilkinson et al., 2012), and hence requires weather input that is spatially coherent. Again, current generators produce weather for a fixed location. However, spatially coherent weather based on climate change projections is being created at a regional/city scale for the UK-based ARCADIA (Adaptation and Resilience in Cities: Analysis and Decision making using Integrated Assessment) project. Although this approach is statistical rather than dynamic and is unable to create features such as fronts, the recreation of spatially coherent weather over a city-sized region allows for the type of variations in weather that transport networks are sensitive to.

Related to this is the move towards probabilistic scenarios for climate change. Rather than the deterministic projections provided in earlier climate projection tools, there has been a trend towards including multiple model runs that take better account of the uncertainty in climate models and our understanding of the climate system (Jones et al., 2009). However, little research has been conducted on how probabilistic climate change projections will be received and used in the transport sector. Although part of the reasoning for providing probabilistic climate change projections to the stakeholder and the climate change impacts community is to give a better degree of risk and certainty (Murphy et al., 2009), little research has been conducted into how different stakeholder groups respond to the wealth of information that often results. This is especially true where, unlike the energy sector and water resources, the transport sector is generally less familiar with using probabilistic projections for long-term planning. It is not known whether these extremes will be useful as framing points, for instance by using the 90th percentiles at the highest emission scenario as the worst case scenario and basing ‘no-regret’ adaptation plans on these.

Furthermore, existing climate projections take no account of either current or future urbanization. Again, although weather generator tools are being developed which include urban microclimates, any existing CIA for urban areas should regard the current absence as a caveat. It must also be noted that the urban microclimate will change as urban planning and activity are influenced by changes in the wider socio-economic environment of the country. Saitoh et al. (1996) have looked at simulating a future urban heat island associated with the Tokyo metropolitan area using three-dimensional modelling. This looked at a simple increase in the energy release rate. However, little work has looked at the effect that changes in the urban environment might make.

Although previously treated in isolation, various studies are now looking at the interdependencies both between different modes of transport and between transport and other components of the infrastructure supply chain. The concept of ‘cascading failure’, where disruption in one location can affect other modes and sectors, has become important to governments for planning for present-day weather-related impacts, as it allows for the identification of critical infrastructure whose resilience must be ensured. The effect of climate change on the interconnected functions of a city have begun to be modelled. For instance, Jenkins et al. (2012) modelled the impact of rail buckles in London on other sectors and the productivity of the city in general. The ARCADIA Weather Generator was used with the thresholds derived by Dobney et al. (2009), with the effect on London being simulated by a city-wide economic model.

3 Transport futures

Although a fundamental part of CIA (Tol, 1998) the consideration of concurrent socio-economic change is often neglected in impact assessments. The creation of scenarios for plausible future visions of the future economy and society as well as the sector(s) in question can help to determine the potential best and worst case scenarios by combining climate and socio-economic futures (Berkhout et al., 2001). For example, Dawson et al. (2009) used climatic and socio-economic scenarios to determine the future risk of coastal erosion in East Anglia, taking into consideration different future scenarios for housing developments on the coast. Jaroszweski et al. (2013) apply these concepts to the freight and logistics sector of the UK. This study involves the determination of relationships between weather and road freight accidents, which are extrapolated onto future climate using the UKCP09 climate scenarios and weather generator (Jones et al., 2009; Murphy et al., 2009), which projects decreases in summer precipitation and winter ice-related accidents but an increase in winter precipitation-related accidents. This study also uses the UKCIP Socio-Economic toolkit to arrive at expert-led scenarios for development of the freight and logistics sector. These socio-economic scenarios are linked to the climate scenarios they promote, leading to best and worst scenarios. However, at present this is the only transport-CIA to take into consideration multiple dimensions of change.

It is also possible that changes to the urban environment brought about for other reasons, such as the planting of trees and vegetation as a substitute for traditional drainage, may have co-benefits such as reducing the strength of the urban heat island (Coutts et al., 2013). There is potential for Water Sensitive Urban Design (WSUD) to contribute to Climate Sensitive Urban Design (CSUD) through enhanced evaporation and surface cooling. Huong and Pathirana (2013) take an integrated approach and look at different aspects of future changes to flooding, including sea level rise, increases in river runoff, increased urban runoff caused by increases in imperviousness, and enhanced extreme rainfall due to microclimatic changes in the urban area. This was driven by a land-use simulation and an urban drainage model. The model included increased river flow and sea level rise, an important feature in delta regions; different development scenarios were included and worst case scenarios were identified.

In reality, transport has interdependencies with other infrastructure, particularly and increasingly energy and ICT. These dependencies already exist on electrified sections of the rail network, but could become key on the road network if battery-powered cars or directly electrified networks become widespread. If renewable energy sources are used, then the sites are the source of the dependency; if solid fuels are still the primary source, then the generation sites as well as transport links to these sites represent the dependencies (Royal Academy of Engineering, 2011). Particular modes or components of transport are also dependent on other modes and components of the transport system. Journeys often consist of several legs which may involve more than one mode of transport, hence effective transport is dependent on all modes operating in adverse conditions, as well as the resilience of the interchanges between modes and local and national infrastructure. It is therefore important to consider how all modes of transport may change in future scenarios.

It is essential to create a model framework which can incorporate all of these different failure types, and how these knock on to other modes and sectors. However, it is also important that we define what failure means; it must be remembered that there are different perspectives on resilience. Local infrastructure managers will be concerned about the resilience of their infrastructure usually in isolation, or at least in isolation in terms of the mode. Passengers, however, will be interested in whether transport can meet their needs, with each having a threshold for failure depending on their activity. Policy-makers, on the other hand, will want to ensure that transport meets the needs of the wider economy, and will be less concerned about the specific locations of failures. All of these stakeholder and user requirements will change in the future, so a consideration of the future socio-economic scenario and how this might affect attitudes is essential.

It is also important to identify which regions of the world will be most affected by the impacts of climate change on urban transportation. O’Brien (2000) discussed the consequences of climate change and globalization in the context of ‘double exposure’, the observation that some regions may face both impacts of climate change and the consequences of globalization. This links into impact of climate change on transportation in developing cities, which in some cases face the negative impact of globalization. Many are not located in flood-prone regions, and may lack the adaptive capacity to withstand the impacts.

IV Conclusion

This review has identified the current level of understanding in the requisite areas for climate change impact assessment on the transport sector. The knowledge base can generally be seen to decrease from the initial stages (determination of current relationships) to the latter (consideration of concurrent socio-economic change). The current relationships between weather and transport failure have several well-studied areas such as accidents and behaviour, although for a successful climate change impact assessment more work is needed on physical models of infrastructure performance in extreme weather. Additionally, more work is required on the knock-on effects of infrastructure failure on the performance of the transport network, as well as the effect of exogenous shocks brought about by failure in other sectors such as energy.

It is clear that many of the current climate change projection toolkits were not developed with sectors like transport in mind. The transport sector (and to some extent the energy sector) is vulnerable to short-term extreme events that are not currently captured in the existing weather generators. Spatial coherence and the reproduction of realistic fields of weather is also a very important future requirement, as transport is vulnerable to weather events that remove multiple nodes and at the same time. To create realistic impact projections much more work is needed on visioning concurrent changes in the overlying socio-economic scenario, the resulting transport network and how it is used. The creation of multiple scenarios for the transport sector will help to create best and worst case scenarios and can also be used to determine what a resilient future network may look like and how it might be achieved.

References

Abbaspour

Jafari

Mansouri

. (2008) Thermal comfort evaluation in Tehran metro using Relative Warmth Index. International Journal of Environmental Science and Technology 5 (3): 297–304.

Aerts

JCJH

Botzen

WWJ

de Moel

. (2013) Cost estimates for flood resilience and protection strategies in New York City. Annals of the New York Academy of Sciences 1294 (1): 1–104.

Agarwal

Maze

Souleyrette

(2006) The weather and its impact on urban freeway traffic operations. Paper presented at the Transportation Research Board (TRB) 2006 Annual Meeting, Center for Transportation Research and Education, Iowa.

Andersson

Chapman

(2011) The impact of climate change on winter road maintenance and traffic accidents in West Midlands, UK. Accident Analysis and Prevention 43 (1): 284–289.

Andreescu

Frost

(1998) Weather and traffic accidents in Montreal, Canada. Climate Research 9 (3): 225–230.

Andrey

Yagar

(1993) A temporal analysis of rain-related crash risk. Accident Analysis and Prevention 25 (4): 465–472.

AON Benfield (2012) 2011 Thailand floods event recap report: Impact forecasting – March 2012. Chicago, IL: Impact Forecasting LLC.

Arkell

Darch

GJC

(2006) Impact of climate change on London’s transport network. Proceedings of the Institution of Civil Engineers – Municipal Engineer 159 (4): 231–237.

Aultman-Hall

Lane

Lambert

(2009) Assessing impact of weather and season on pedestrian traffic volumes. Transport Research 2140: 35–43.

10.

Baker

(1993) The behaviour of road vehicles in unsteady cross winds. Journal of Wind Engineering and Industrial Aerodynamics 49 (1–3): 439–448.

11.

Baker

Reynolds

(1992) Wind-induced accidents of road vehicles. Accident Analysis and Prevention 24 (6): 559–575.

12.

Baker

Chapman

Quinn

. (2010) Climate change and the railway industry: A review. Proceedings of the Institution of Mechanical Engineers Part C – Journal of Mechanical Engineering Science 224 (C3): 519–528.

13.

Banister

(2011) The trilogy of distance, speed and time. Journal of Transport Geography 19 (4): 950–959.

14.

Batty

Axhausen

Giannotti

(2012) Smart cities of the future. The European Physical Journal Special Topics 214: 481–518.

15.

Berkhout

Herton

Jordan

(2001) Socio-economic futures in climate change impact assessment as ‘learning machines’. Global Environmental Change 12 (2): 83–95.

16.

Bocker

Dijst

Prillwitz

(2013) Impact of everyday weather on individual daily travel behaviours in perspective: A literature review. Transport Reviews 33 (1): 71–91.

17.

Botelle

Payne

Redhead

(2010) Squeezing the heat out of London's Tube. Proceedings of the Institution of Civil Engineers - Civil Engineering 3: 114–122.

18.

Brüde

Larsson

(1980) Samband vintertid mellan väderlek-väglag-trafikolyckor. Statistisk bearbetening och analys. VTI-rapport 210. Statens väg-och trafikninstitut (VTI), Linköping, Sweden.

19.

Chapman

Thornes

(2003) The use of geographical information systems in climatology and meteorology. Progress in Physical Geography 27 (3): 313–330.

20.

Chen

Cai

(2004) Accident assessment of vehicles on long-span bridges in windy environments. Journal of Wind Engineering and Industrial Aerodynamics 92 (12): 991–1024.

21.

Chen

Whalley

(2012) Green infrastructure: The effects of urban rail transit on air quality. American Economic Journal – Economic Policy 4 (1): 58–97.

22.

Chung

(2012) Assessment of non-recurrent congestion caused by precipitation using archived weather and traffic flow data. Transport Policy 19 (1): 167–173.

23.

Clarke

Kersey

Trevorrow

. (2002) London’s warming: The impacts of climate change on London. Technical report, London Climate Change Partnership. Available at: http://legacy.london.gov.uk/gla/publications/environment/londons_warming02.pdf.

24.

Clarke

Smethurst

(2010) Effects of climate change on cycles of wetting and drying in engineered clay slopes in England. Quarterly Journal of Engineering Geology and Hydrogeology. 43(4): 473–486.

25.

Committee on Climate Change and US Transportation (2008) Potential impacts of climate change on US transportation. Special Report 290. Washington, DC: Transportation Research Board (TRB), 296 pp.

26.

Cools

Creemers

(2013) The dual role of weather forecasts on changes in activity-travel behavior. Journal of Transport Geography 28: 167–175.

27.

Coutts

Beringer

Tapper

(2013) Changing urban climate and CO₂ emissions: Implications for the development of policies for sustainable cities. Urban Policy and Research 28 (1): 27–47.

28.

Cresswell

(2006) On the Move: Mobility in the Western World. Abingdon: Routledge.

29.

Dawson

Dickson

Nicholls

(2009) Integrated analysis of risks of coastal flooding and cliff erosion under scenarios of long-term change. Climatic Change 95: 249–288.

30.

DEFRA (2012) UK climate change risk assessment: Government report. London: DEFRA.

31.

Dobney

Baker

Chapman

. (2010) The future cost to the United Kingdom’s railway network of heat-related delays and buckles caused by the predicted increase in high summer temperatures owing to climate change. Proceedings of the Institution of Mechanical Engineers Part F – Journal of Rail and Rapid Transit 224 (F1): 25–34.

32.

Dobney

Baker

Quinn

. (2009) Quantifying the effects of high summer temperatures due to climate change on buckling and rail related delays in south-east United Kingdom. Meteorological Applications 16 (2): 245–251.

33.

Eddington

(2006) The Eddington Transport Study: Main Report, Transport’s Role in Sustaining the UK’s Productivity and Competitiveness. London: HMSO.

34.

Eddowes

Waller

Taylor

. (2003) Safety implications of weather, climate and climate change. London: Rail Safety and Standards Board. Available at: http://www.railwaysarchive.co.uk/documents/RSSB_SafetyClimateChange2003.pdf.

35.

Edwards

(1998) The relationship between road accident severity and recorded weather. Journal of Safety Research 29 (4): 249–262.

36.

Eisenberg

(2004) The mixed effects of precipitation on traffic crashes. Accident Analysis and Prevention 36 (4): 637–647.

37.

Eisenberg

Warner

(2005) Effects of snowfalls on motor vehicle collisions, injuries, and fatalities. American Journal of Public Health 95 (1): 120–124.

38.

Elvik

(2006) Laws of accident causation. Accident Analysis and Prevention 38 (4): 742–747.

39.

Eriksson

(2001) Regional influence on the occurrence of road slipperiness during winter precipitation events. Meteorological Applications 8 (4): 449–460.

40.

Erlingsson

(2012) Rutting development in a flexible pavement structure. Road Materials and Pavement Design 13 (2): 218–234.

41.

Ferreira

Teixeira

Cardoso

(2011) Impact of bituminous subballast on railroad track deformation considering atmospheric actions. Journal of Geotechnical and Geoenvironmental Engineering 137 (3): 288–292.

42.

Gao

Yao

(2012) Field studies on the effect of built forms on urban wind environments. Renewable Energy 46: 148–154.

43.

Hambly

Andrey

Mills

. (2013) Projected implications of climate change for road safety in Greater Vancouver, Canada. Climatic Change 116 (3–4): 613–629.

44.

Sun

. (2011) Environmental factors of road slope stability in mountain area using principal component analysis and hierarchy cluster. Environmental Earth Sciences 62 (1): 55–59.

45.

Hooper

Chapman

(2012) The impacts of climate change on national road and rail networks. In: Ryley

Chapman

(eds) Transport and Climate Change. Bingley: Emerald, 105–136.

46.

Hooper

Chapman

Quinn

(2012) Investigating the impact of precipitation of vehicle speed on UK motorways. Meteorological Applications 21 (2): 194–201.

47.

Huong

HTL

Pathirana

(2013) Urbanization and climate change impacts on future urban flooding in Can Tho city, Vietnam. Hydrology and Earth System Sciences 17 (1): 379–394.

48.

Ibrahim

Hall

(1994) Effect of adverse weather conditions on speed-flow-occupancy relationships. Transportation Research Record No. 1457. Washington, DC: Transportation Research Board (TRB).

49.

Intergovernmental Panel on Climate Change (IPCC) (2013) Working Group I Contribution to the IPCC Fifth Assessment Report. Climate Change 2013: The Physical Science Basis. Cambridge: Cambridge University Press.

50.

Jaroszweski

McNamara

(2014) The influence of rainfall on road accidents in urban areas: A weather radar approach. Travel Behaviour and Society 1(1): 15–21.

51.

Jaroszweski

(2012) The impact of climate change on the national freight sector. In: Ryley

Chapman

(eds) Transport and Climate Change. Bingley: Emerald, 137–173.

52.

Jaroszweski

Chapman

Petts

(2010) Assessing the potential of climate change of transportation: The need for an interdisciplinary approach. Journal of Transport Geography 18 (2): 331–335.

53.

Jaroszweski

Chapman

Petts

(2013) Climate change and road freight safety: A multidiscilinary exploration. Climatic Change 120 (4): 785–799.

54.

Jenkins

Glenis

Ford

. (2012) A probabilistic risk based approach to addressing impacts of climate change on cities: The Tyndall Centres Urban Integrated Assessment Facility. UGEC Viewpoints 8: 8–11.

55.

Jones

Harpham

Kilsby

. (2009) Projections of future daily climate for the UK from the Weather Generator. UK Climate Projections science report, University of Newcastle.

56.

Keay

Simmonds

(2005) The association of rainfall and other weather variables with road traffic volume in Melbourne, Australia. Accident Analysis and Prevention 38 (3): 445–454.

57.

Khelifa

Garrow

Higgins

. (2013) Impacts of climate change on scour-vulnerable bridges: Assessment based on HYRISK. Journal of Infrastructure Systems 2: 138–146.

58.

Koetse

Rietveld

(2009) The impact of climate change and weather on transport: An overview of empirical findings. Transportation Research Part D – Transport and Environment 14 (3): 205–221.

59.

Koetse

Rietveld

(2012) Adaptation to climate change in the transport sector. Transport Reviews 32 (3): 267–286.

60.

Kusakabe

Iryo

Asakura

(2010) Estimation method for railway passengers’ train choice behavior with smart card transaction data. Transportation 37 (5): 731–749.

61.

Kyte

Khatib

Shannon

. (2000) Effect of environmental factors on free-flow speed. In: Transportation Research Circular E-C108, presented at the Fourth National Symposium on Highway Capacity, Maui, Hawaii. Washington, DC: Transportation Research Board (TRB) and National Research Council (NRC), 108–119.

62.

Lathia

Ahmed

Capra

(2012) Measuring the impact of opening the London shared bicycle scheme to casual users. Transportation Research Part C – Emerging Technologies 22: 88–102.

63.

Lemonsu

Kounkou-Arnaud

Desplat

. (2013) Evolution of the Parisian urban climate under a global changing climate. Climatic Change 114 (3–4): 679–692.

64.

Lindgren

Jonsson

Carlsson-Kanyama

(2009) Climate adaptation of railways: Lessons from Sweden. European Journal of Transport and Infrastructure Research 9 (2): 164–181.

65.

Liu

Yang

Chang

. (2012) Spatial patterns and drivers of fire occurrence and its future trend under climate change in a boreal forest of northeast China. Global Change Biology 18 (6): 2041–2056.

66.

Lloyd

Anderson

Hussein

. (2001) Preventing landslides on roads and railways: A new risk-based approach. Proceedings of the Institution of Civil Engineers – Civil Engineering 144 (3): 129–134.

67.

Macdonald

JHG

Irwin

Fletcher

(2003) Vortex-induced vibrations of the Second Severn Crossing cable-stayed bridge: Full-scale and wind tunnel measurements. Proceedings of the Institution of Civil Engineers – Structures and Buildings 156 (3): 332–333.

68.

Manning

Hall

Kilsby

. (2008) Spatial analysis of the reliability of transport networks subject to rainfall-induced landslides. Hydrological Processes 22 (17): 3349–3360.

69.

Medina

Hajbabaie

Benekohal

(2011) Detection performance of wireless magnetometers at signalized intersection and railroad grade crossing under various weather conditions. Transportation Research Record 2259: 233–241.

70.

Miranda-Moreno

Nosal

(2011) Weather or not to cycle: Temporal trends and impact of weather on cycling in an urban environment. Transportation Research Record 2247: 42–52.

71.

Muller

Chapman

Young

. (2013) Towards a standardised metadata protocol for urban meteorological networks. Bulletin of the American Meteorological Society 94: 1161–1185.

72.

Murphy

Sexton

DMH

Jenkins

. (2009) UK climate projections science report: Climate change projections. Exeter: Met Office Hadley Centre.

73.

Nofal

Saeed

AAW

(1997) Seasonal variation and weather effects on road traffic accidents in Riyadh City. Public Health 111 (1): 51–55.

74.

Nofal

Saeed

Anokute

(1996) Aetiological factors contributing to road traffic accidents in Riyadh City, Saudi Arabia. Journal of the Royal Society of Health 116 (5): 304–311.

75.

O'Brien

Leichenko

(2000) Double exposure: assessing the impacts of climate change within the context of economic globalization. Global Environmental Change 10: 221–232.

76.

Oke

(1973) City size and urban heat island. Atmospheric Environment 7 (8): 769–779.

77.

Oke

(2006) Initial guidance to obtain representative meteorological observations at urban sites. Instruments and observing methods report no. 81. Geneva: World Meteorological Organization (WMO).

78.

Overeem

Leijnse

Uijlenhoet

(2013) Country-wide rainfall maps from cellular communication networks. Proceedings of the National Academy of Sciences of the United States of America. doi: 10.1073/pnas.1217961110.

79.

Palin

Thornton

Mathison

. (2013) Future projections of tempeature-related climate change impacts on the railway network of Great Britain. Climatic Change 120 (1–2): 71–93.

80.

Peterson

McGuirk

Houston

. (2008) Climate Variability and Change with Implications for Transportation. Washington, DC: Transportation Research Board (TRB).

81.

Prusa

Segal

Temeyer

. (2002) Conceptual and scaling evaluation of vehicle traffic thermal effects on snow/ice-covered roads. Journal of Applied Meteorology 41 (12): 1225–1240.

82.

Qiu

Nixon

(2008) Effects of adverse weather on traffic crashes systematic review and meta-analysis. Transportation Research Record 2055: 139–146.

83.

Ridley

McGinnity

Vaughan

(2004) Role of pore water pressures in embankment stability. Proceedings of the Institution of Civil Engineers – Geotechnical Engineering 157 (4): 193–198.

84.

Rosenfeld

(2000) Suppression of rain and snow by urban and industrial air pollution. Science 287 (5459): 1793–1796.

85.

Royal Academy of Engineering (2011) Infrastructure, Engineering and Climate Change Adaptation: Ensuring Services in an Uncertain Future. London: Royal Academy of Engineering.

86.

Saitoh

Schimada

Hoshi

(1996) Modeling and simulation of the Tokyo urban heat island. Atmospheric Environment 30 (20): 3431–3442.

87.

Schmidt

Glade

(2003) Linking global circulation model outputs to regional geomorphic models: A case study of landslide activity in New Zealand. Climate Research 25 (2): 135–150.

88.

Serreze

Walsh

Chapin

FS III

. (2000) Observational evidence of recent change in the northern high-latitude environment. Climatic Change 46 (1–2): 159–207.

89.

Stern

Shah

Goodwin

Pisano

(2004) Analysis of weather impacts on traffic flow in metropolitan Washington, DC. Washington, DC: US Department of Transportation Federal Highway Administration.

90.

Theethayi

Liu

Montano

. (2005) A theoretical study on the consequence of a direct lightning strike to electrified railway system in Sweden. Electric Power Systems Research 74 (2): 267–280.

91.

Thomachot

Matsuoka

Kuchitsu

. (2005) Frost damage of bricks composing a railway tunnel monument in Central Japan: Field monitoring and laboratory simulation. Natural Hazards and Earth System Sciences 5 (4): 465–476.

92.

Thornes

(1992) The impact of weather and climate on transport in the UK. Progress in Physical Geography 16 (2): 187–208.

93.

Thornes

Fisher

Rayment-Bishop

. (2014) Ambulance call-outs and response times in Birmingham and impact of extreme weather and climate change. Emergency Medicine Journal 31 (3): 220–228.

94.

Tight

Timms

Banister

. (2011) Visions for a walking and cycling focussed urban transport system. Journal of Transport Geography 19 (6): 1580–1589.

95.

Tol

RSJ

(1998) Socio-economic scenarios. In: Feenstra

Burton

Smith

Tol

RSJ

(eds) Handbook on Methods for Climate Change Impact Assessment and Adaptation Strategies. New York: United Nations Environment Programme, Chapter 2.

96.

United Nations (UN) (2011) The Economic Role of Cities. New York: UN.

97.

United Nations (UN) (2012) World Population Prospects: The 2011 Revision. New York: UN.

98.

Wang

Daneshmand

Dohler

(2013) Special issue on Internet of Things (IOT): Architecture, protocols and services. IEEE Sensors Journal 13 (10): 3505–3510.

99.

Wang

Liu

(2001) Prevention of debris flow disasters on Chengdu-Kunming Railway. Journal of Environmental Sciences – China 13(3): 333–336.

100.

World Meteorological Organization (WMO) (2008) Guide to Meteorological Instruments and Methods of Observation, seventh edition. Geneva: WMO.

101.

Young

Liesman

(2007) Estimating the relationship between measured wind speed and overturning truck crashes using a binary logit model. Accident Analysis and Prevention 39 (3): 574–580.

102.

Zhang

Chen

Riseborough

(2008) Transient projections of permafrost distribution in Canada during the 21st century under scenarios of climate change. Global and Planetary Change 60 (3–4): 443–456.