Review on Energy Resilience

lable at ScienceDirect

Journal of Cleaner Production 174 (2018) 1299e1313

Contents lists avai

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Impacts and implications of climatic extremes for resilience planning of transportation energy: A case study of New York city

Arash Beheshtian a, b, *, Kieran P. Donaghy a, H. Oliver Gao b, c, Sahar Safaie d, Richard Geddes b

a Department of City and Regional Planning, Cornell University, W. Sibley Hall, 942 University Ave, Ithaca, NY 14850, USA b Cornell Program in Infrastructure Policy, Department of Policy Analysis and Management, Cornell University, 251 Martha Van Rensselaer Hall, Ithaca, NY 14853, USA c School of Civil and Environmental Engineering, Cornell University, Ithaca, USA d United Nations Office for Disaster Risk Reduction (UNISDR), 9-11 Rue de Varemb�e, CH1202, Geneva, Switzerland

a r t i c l e i n f o

Article history: Received 1 May 2017 Received in revised form 6 November 2017 Accepted 6 November 2017 Available online 10 November 2017

Keywords: Climate change Transportation energy Critical infrastructure Fuel supply chain Resilience Adaptation-mitigation nexus

* Corresponding author. Department of City and University, W. Sibley Hall, 942 University Ave, Ithaca,

E-mail address: [email protected] (A. Beheshtia

https://doi.org/10.1016/j.jclepro.2017.11.039 0959-6526/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t

An integrated alternative planning can control climate change drivers and mitigate or neutralize the adverse impacts of the changing climate on the transportation energy sector. In this article, we intro- duced an infrastructure of alternative fuel as a synergistic approach to climate-adaptation and -miti- gation, and advanced a quantitative method to simulate the dependency of travel behavior on fuel availability when the infrastructure of transportation energy is stressed or under attack.

Our approach allows both commuters traveling behavior and properties of transportation energy system to interact. Featuring a multi-stage mathematical program, the proposed model was used to study the resilience of New York City's transportation energy sector in the face of a range of climatic extremes.

We found the impact of the changing climate on energy sector is major and worsening over time. The modeling results also reveal spatial distribution of vulnerable elements in both transportation and fueling infrastructures, the system's overall resilience in time of disaster, and the daily-commuters impact of vulnerable energy infrastructure. The modeling results are crucial to successful integrated planning response to climate-adaptation and -mitigation.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

The transportation energy sector in metropolitan areas is currently challenged to address two key weaknesses it harbors. First, the primary sources of urban transportation energyegasoline and dieseleare non-renewable and considered among the main contributors of urban pollution problems (Wang, 2008). According to the U.S. Department of Transportation, twenty-nine percent of U.S. greenhouse gas emissions come from the transportation sector, and over four-fifths of polluting emissions in the transportation sector come from tailpipes.

Second, this infrastructure is vulnerable against climate-related extreme events (Beheshtian, 2016). As witnessed through the past

Regional Planning, Cornell NY 14850, USA. n).

decade's extreme-weather-episodes, the stressed or under attack infrastructure of transportation energy (i.e. motor fueling infra- structure including liquid fossil fuel supply-chain distributing fuels from terminals/refineries to gas stations) not only hamper the pre- (e.g. evacuation tasks) and post-event emergency tasks, but also interrupt the recovery process that may be in effect for several months following an extreme event. Each of these challenges, however, has been given a fair amount of attention by scientists and policy-makers investigating how to control the drivers of climate change and how to mitigate or otherwise neutralize the aftermath of climatic hazards.

Climate change mitigation has been a concern for decades and addressed by a variety of attempts including: basic technological improvements (e.g. considerable increments in fuel efficiency for conventional vehicles and improvements in electric vehicle (EV) battery range), subsidized mass transit and transit-oriented development (TOD), and a range of policy changes, such as congestion pricing and incentivizing the purchase/operation of

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e13131300

alternative fuel vehicles (AFVs). Despite these responses, the suc- cessful transition to clean mobility and the achievement of considerable market share for AFVs face a significant obstacle: demand for and use of AFVs by drivers will depend largely on the availability of charging stations which are conditioned on public- and private-sector expectations of AFV production and use.

Compared to the first major challenge, renewable energy adoption, the vulnerability of the motor fueling supply chain and its impact on travel behaviors in time of climatic hazards has been relatively understudied and under-funded. This is likely due to 1) the complexity inherent in modeling large-scale infrastructures, the stochasticity embedded in the functionality of infrastructure elements, and end-user behavior in time of disaster, and 2) the uncertainties carried by the economics of adaptation, the high politicization of planning for such transboundary urban infra- structure problems, and the capital-intensive nature of resilience- enhancing strategies.

In response to the obstacles faced by the mitigation and adap- tation approaches, integrated planning could be advanced to address both major challenges. While expansion of EV charging facilities is considered a base requirement to transition to clean inner-city mobility, a strategic spatial distribution of these facilities in hazard-prone areas would alternatively provide a planning- focused response to the climate-vulnerable infrastructure of transportation energy. Arguably, the infrastructure of alternative fuel (i.e. the network of charging stations for EVs) is more reliable in time of disasters, since the energy supplier is the power grid, which is more resilient in the face of climatic events compared with liquid-fuel, which is supplied by terminals and refineries mainly located in shoreline and flood-prone areas.

Nevertheless, the key-concern in the strategic development of charging facilities is understanding the traveling and fueling be- haviors of commuters in times of disaster. This requires modelers to investigate 1) the vulnerability of the transportation and fueling systems to extreme hazards, 2) ripple-effect failures cascading from the transportation infrastructure to the transportation energy system, and 3) travel behaviors in response to fuel shortages.

In this article, we propose a modeling approach to condition traveling behavior in time of disaster on a disrupted transportation energy infrastructure; we examine the model on the real-size network of New York City (NYC). The proposed model features a multi-stage optimization that conditions the traveling behavior based on the productivity of the fueling infrastructure, which itself is dependent on the resilience of the transportation system in time of disaster.

This work is distinguished by three key-contributions: 1) pro- posing an infrastructure of alternative fuel as a synergistic approach to adaptation and mitigation; 2) modeling fuel avail- ability in service stations with respect to end user behavior and overall systemic resilience in time of disaster; and 3) simulating the physical and functional interdependency between the trans- portation and fueling infrastructures.

The remainder of this paper is divided in sections. In Section 2 which follows, we provide a review of the relevant literature. Sec- tion 3 analyzes the impact on the built environment of climatic extremes at four levels: global, metropolitan area, critical infra- structure, and transportation energy infrastructure; it also specif- ically reviews the challenges NYC faces in adopting climate resilience planning. In Section 4 we define inoperability metrics, elaborate the proposed methodology, and discuss the modeling results. Finally, Section 5 concludes the paper by summarizing the modeling output, discussing research barriers, and proposing further improvements for future work.

2. Literature review

Here, we briefly review the literature in two subject categories: 1) behavioral change in commuting patterns in relation to a disaster; this is assessed pre- and post-disaster. It covers the evacuation process (pre-disaster), and recovery and daily commuting under a disrupted transportation network (post- disaster), and 2) the dependency of travel behavior on fuel availability.

A vast majority of available evacuation models seek minimum travel time or maximum evacuees in a defined time-period. Han et al. (2007) considered both travel time and “time- and space- based risk.” Sayyady and Eksioglu (2010) embedded the safety as- pects in evacuation model and Sbayti and Mahmassani (2006) studied phase evacuation prior to disaster time. Fu and Wilmot (2006) developed a bi-objective modeling approach seeking whether to evacuate and optimum timing to do so. Aside from the modeling approaches, many practical aspects of evacuation have also been studied by scholars such as contraflow strategies in evacuation planning (Kalafatas and Peeta, 2009), optimal evacua- tion tasks and strategies (Abdelgawad and Abdulhai, 2009), driving behavior under adverse conditions (Hoogendoorn et al., 2010), and shelter planning with respect to evacuation routing (Kongsomsaksakul et al., 2005).

The literature also includes a vast variety of modeling ap- proaches addressing the wellbeing of transportation systems following an exogenous shock and its influence on both the tasks of first responders and the traveling behavior of daily commuters. Models address resilience, robustness, availability, productivity, and recovery. He and Liu (2012) proposed a modeling approach to simulate the traffic dynamic evolution in the aftermath of unex- pected network topology change. Zhu et al. (2010) studied the impact of the link failure on travel behavior and Chang and Nojima (2001) investigated the post-disaster transportation system per- formance in the aftermath of a seismic event. Lou and Zhang (2011) modeled a game-theory between attackers, network users, and defenders simulating the travel time reliability when the trans- portation system is under attack.

Aside from the works on passenger traffic, the commodity-flow impact of disaster has also been studied extensively. Smith and Wigley (2006) studied the commodity-flow under network disruption and Jenelius and Mattsson (2012) assessed the Swedish road networks vulnerability through the covering area technique.

The association of commuting patterns with fuel availability/ shortage has been mainly through facility location/sizing problems investigating the optimum siting of fuel/charging stations with respect to the network's traffic. Conditioning the location/sizing of service station on commuting patterns, this has been investigated through a range of quantitative techniques including flow-capture (Kuby et al., 2009), fuel-travel-back model (Lin et al., 2008), set covering (Beheshtian et al., 2016), agent-based simulation (Sweda and Klabjan, 2011), maximal covering (Frade et al., 2011), p-me- dian (Nicholas et al., 2004), and activity-based models (Dong et al., 2014).

Despite a large number of articles and models, much of the literature assumes pre- and post-disaster commuting patterns that are independent of fuel availability. In addition, many models mostly adopt a business-as-usual (BAU) assumption rather than including the inoperabilities that arise in time of disaster.

3. Built environment impacts of global change

3.1. Climate change in a global context

The Earth's climate is changing at a rate that is unprecedented in

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e1313 1301

the past millennium (Crowley, 2000). Just within the past century, the global average surface temperature has risen at a rate of 0.15 �F per decade. Since the late 1970s, while the average change in sur- face temperature for the United States was similar to the global rate, the US has warmed at an average rate of 0.29e0.46 �F per decade (Laboratory for Satellite Altimetry/Sea Level Rise 2015).

Global warming is projected to continue, and climate change is likely to intensify. Worldwide, average temperature is projected to increase by 0.5e8.6 �F by the end of the century (Church et al., 2013). Some parts of the world are, however, projected to experi- ence higher temperature increases than the global average. By 2100, the average US temperature is projected to increase by about 3�e12 �F, depending on emission scenarios (aka representative concentration pathways, or RCPs) and different climate models (Melillo et al., 2014).

The rapid change in global climate is a consequence of the increasing concentration of anthropogenic greenhouse gases (GHG),1 which themselves are an inevitable derivative of bur- geoning urbanization, industrialized societies, and geometric growth.2 Since the beginning of the Industrial Revolution (1760e1820) the concentration of CO2, N2O, and CH4 have risen approximately 40, 20, and 250 days, respectively (Church et al., 2013). With respect to “time and space dependent trajectories” of emission, concentration, land use, energy sources, population growth, and other socio-economic factors (included in RCPs) (Bjørnæs, 2013), global anthropogenic GHG emissions are projected to rise by 102 percent (92 gigatons CO2-eq.) above the 2010 baseline by 2100 (Energy and Climate Outlook, 2014).

While the impact of anthropogenic activities on GHG and climate change has been investigated broadly, the bidirectional causality running between climate change and the built environ- ment3 is not fully known and remains under-studied. Despite this lack of knowledge, there is recognition of the difference between gradual and sudden impacts4 of the changing climate on the built environment and human settlements. These are mainly founded on 1) studying historic events and the possibility of future quasi- events (i.e. known knowns) and 2) forecasting future events with no track record via the past (i.e. known unknowns).5

Gradual impacts are those continuous, gradual, yet profound changes causing permanent results. Obvious examples include coastal erosion and permanent flooding due to sea-level rise (Nicholls and Cazenave, 2010), environmental migration and displacement due to drought and environmental degradation (Adamo, 2010), increasing surface temperature (Gill et al., 2007), formation of urban heat islands (UHIs) (Jones et al., 2007), and

1 GHG include CO2, N2O, CH4, and fluorinated gases, aka F-gases, including HFCs, PFCs, SF6, and NF3. The US Environmental Protection Agency (EPA) recognizes the following human activities as key-contributors to GHG: 1) CO2: fossil fuels, solid waste, deforestation, and soil degradation; 2) N2O: production and transport of oil, natural gas, and coal, livestock and agricultural practices, and the anaerobic decay of organic waste in municipal solid waste landfills; 3) CH4: agricultural and in- dustrial activities, combustion of fossil fuels, and solid waste; 4) F-gases: industrial processes and commercial and household uses.

2 According to Intergovernmental Panel on Climate Change (Church et al., 2013), “there's a better than 95 percent probability that human-produced greenhouse gases such as carbon dioxide, methane and nitrous oxide have caused much of the observed increase in Earth's temperatures over the past 50 years.”

3 As opposed to the natural environment, built environment refers to the ‘[man]- made surroundings that provide the setting for human activity, ranging from the large-scale civic surroundings to the personal places’ (Moffatt and Kohler, 2008).

4 Impacts can be direct and indirect. Indirect impacts (e.g. economic loss, casualty in the aftermath of weather extremes, or altering biotic diversity and ecosystem services which affect the ability of biological systems to support human needs (Lambin et al., 2001)), are not specifically discussed in this article.

5 Aside from the known knowns and known unknowns, there exist unknown un- knowns, which cannot be studied until they occur.

land-use and land-cover changes (Dale, 1997). Sudden impacts refer to those discrete, sudden, and temporary

pulse-disturbances in climate causing extreme variation in the built-environment's well-being. Aka ‘extreme events’6 or ‘climatic extremes,’ they have low-frequency, high amplitude, and are hard to predict. The following are some examples of climate-change- induced hazards: extreme precipitation (Alexander et al., 2006), tropical storms and hurricanes (Webster et al., 2005), heat/cold waves (Easterling et al., 2000), tropical cyclones (Knutson et al., 2010), and dust storms in desert fringe urban areas (Zhang et al., 2003).

3.2. Climate change in New York city

The International Energy Agency (IEA) estimates urban areas currently account for over 67 percent of energy-related global greenhouse gases, a figure which is expected to rise to 74 percent by 2030 (Outlook, 2013). In 2014, New York City (NYC), a “city with the world's highest total greenhouse gas emissions,” produced 49.1 million tCO2e (Inventory of New York City’s Greenhouse Gas Emissions, 2016). Within these emissions, energy use in buildings was the main contributor, with about 73 percent, followed by transportation, at approximately 21 percent.

While the drivers behind NYC's GHGs are well-observed and documented,7 the impact of changing climate on NYC is, to some extent, under-studied. This relates to the fact that former causal direction (i.e. the impact of anthropogenic activities on GHGs and climate change), being man-made, is easier for scientists to track down, while latter direction is still a black box and not fully unpacked yet.8 Despite this, few existing analyses split climate change's impact on NYC into gradual and sudden threads (see Fig. 1).

Gradual impacts influence daily metropolitan well-being, aka business as usual (BAU), in various ways. Mean annual temperature in NYC has increased by 3.4 �F since 1900 (Horton et al., 2015). The future mean annual temperature is also projected to increase by 3.2e6.6 �F by the 2050s, and 4.2 to 12.1 �F by 2100 relative to the 1971e2000 baseline of 54 �F. NYC's observed and projected tem- peratures are graphed in Fig. 2 (left).

The changing climate causes the melting of glacier polar ice caps, as well as inducing ocean thermal expansion, both contrib- uting to sea-level rise (SLR) (Mimura, 2013). By the 2080s, the sea level in NYC is projected to rise 18e39 inches, inundating

6 The US National Oceanic and Atmospheric Administration (NOAA) defines as ‘extreme’ events ‘lying in the outermost (“most unusual”) 10 percent of a place's history.’

7 Citywide and City Government GHG emission data sources are detailed (as early as 2005), accessible in the Inventory of New York City Greenhouse Gas Inventory of New York City’s Greenhouse Gas Emissions, 2016.

8 Undoubtedly, there is no single metro area across the globe subject to a larger number of research articles and technical reports to investigate climate change mitigation, adaptation, planning, and also physical investment in resilience- enhancing strategies than NYC. We believe this relates to the following excep- tional characteristics of NYC.

� The New York metropolitan area, aka the Tri-State Area, is one of the most highly-developed and populous urban agglomerations in the world, and the single largest urbanized area in North America (U.S. Census Bureau).

� Home to a leading global financial center, NYC is a hub of the largest regional economy in the U.S. Any interruption in the metro's BAU could, therefore, impose billions of dollars in losses (Hallegatte et al., 2013) which possibly could ripple throughout the national and global economy (Rinaldi et al., 2001).

� NYC is highly exposed and vulnerable to climate extremes, as evidenced through several events in the past decade (Beheshtian et al., 2018).

-2

-1

0

1

2

3

4

5

6

7

8

1900 1950 2000 2050 2100

Te m

pr ea

tu re

(° F)

RCP 2.6 RCP 4.5 RCP 6.0 RCP 8.5

250

350

450

550

650

750

850

950

800

1300

1800

2300

2800

3300

3800

1900 1950 2000 2050 2100

Em is

si on

(p pm

v/ pp

bv )

Em is

si on

(p pm

v) CH4 - RCP 2.6 CH4 - RCP 4.5 CH4 - RCP 6.0 CH4 - RCP 8.5 CO2 - RCP 2.6 N2O - RCP 2.6 CO2 - RCP 4.5 N2O - RCP 4.5 CO2 - RCP 6.0 N2O - RCP 6.0 CO2 - RCP 8.5 N2O - RCP 8.5

Fig. 1. Left. Global anthropogenic GHGs emissions; observed (1900e2016) and projected (2016e2100). Note: projections are based on four RCPs: 2.6, 4.5, 6.0, and 8.5. The CH4 curves relate to the left vertical axis and the CO2 and N2O curves relate to the right vertical axis. The unit of measurement for CO2 and CH4 concentrations is parts per million by volume (ppmv) and for N20 concentrations is parts per billion by volume (ppbv). Sources: RCP 2.6 (Van Vuuren et al., 2011); RCP 4.5 (Wise et al., 2009); RCP 6.0 (Fujino et al., 2006); RCP 8.5 (Riahi et al., 2007). Right. Global average temperature change (�F); observed (1900e2016) and projected (2016e2100). Source: Intergovernmental Panel on Climate Change, The Fifth Assessment Report, aka AR5 (Cubasch et al., 2013).

Fig. 2. Left. NYC temperature; observed (1900e2016 at Central Park) and projected (2020s, 2050s, 2080s, and 2100) for the low (10th percentile), middle range (25th percentile to 75th percentile), and high estimation (90th percentile). Source: NOAA, National Weather Service for observed trend and New York Panel on Climate Change 2015 (Horton et al., 2015) for the projected trend. Note: the projected temperature is normalized based on 1971e2000 temperature (i.e. 54 �F). Center. NYC middle range (25th to 75th percen- tiles) sea level change (inches), observed (1900e2015) and projected (2020s, 2050s, 2080s, and 2100). Source: Church and White (2011) and Laboratory for Satellite Altimetry/Sea Level Rise (2015). Note: projections are adjusted on 2000e2004 base period. Right. Low estimate (10th percentile), middle range (25th to 75th percentiles), and the high estimate (90th percentile) of annual chance of today's 100-year flood (%) and future coastal flood heights associated with 100-year flood (ft) at the Battery, NYC; projected (2020s, 2050s, and 2080s). Source: Horton et al. (2015). Note: flood heights represent the baseline value (excluding wave height) on top of the sea level rise projections. Flood height elevations are referenced to the North American Vertical Datum of 1988.

9 For additional and detailed information on this concept, the reader may consult

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e13131302

thousands of acres of low-lying areas (Horton et al., 2015). The trajectory of SLR will increase the “frequency, intensity and power dissipation” of hydro-meteorological hazards in the Atlantic basin (Lin et al., 2012, pp. 463). In the Battery section of Manhattan, for instance, by the 2080s, the annual chance of today's 100-year flood will be 4 times higher, and the corresponding surge flood height may reach up to 13 ft (Horton et al., 2015). NYC sea level change (observed and projected), the annual chance of today's 100-year flood, and flood heights associated with 100-year floods in the 2020s, 2050s, and 2080s are shown in Fig. 2.

Table 1 briefly summarizes climatic hazards NYC faces and covers their immediate consequences. While the listed threats

represent the direct consequences of the changing climate, they may trigger cascading and escalating failures9 leading to the spread of ripples with extensive, incremental, and in many cases, long- lasting impacts, some of which “may not become evident for several decades” (McMichael, 2003, pp. 12). Indirect-rippled con- sequences are, however, most uncertain, rather local in scale, and primarily studied in regional, national, or global contexts.

Indirect effects include, but are not limited to, marginal (aka social) costs of emissions and adverse impacts on welfare (Tol,

Little (2010) and Dessai et al. (2004).

Table 1 The NYC's impacts of climate change.

Phenomenon Direct threat Consequence

Gradual impacts Raising temperature

� Higher GHGs concentration � Airborne pollen production � Raising surface/air temperature

Increased peak electricity loads in summer resulting in higher frequency in utility outages (Horton et al., 2015); increased HVAC systems demand in summer; decreased water quality due to biological and chemical impacts (Murdoch et al., 2000); exacerbated UHI (Solecki et al., 2004); risen in water peak demand flow and potential water scarcity (The NYC Department of Environmental Protection Climate Change Program Assessment and Action Plan report, AAP 2008); ground-level ozone formation (Ebi and McGregor, 2008); increase particulate matter (Climate Change Adaptation Resource Center, aka ARC-X); exacerbated heat-related morbidity.

Raising total annual precipitation

� Brief-intense rainstorm Increased street, basement and sewer flooding; reduction of water quality, equipment damage from corrosion, and contaminant leaching and runoff (Horton et al., 2010); more debris in reservoir, water quality violation, and failure of drainage and wastewater treatment systems (AAP 2008)

Sea level rise � Permanent flooding � Land/soil erosion

Encroachment of salt/fresh-water sources (Cohen, 2011); inundated low-lying areas, developed lands, communities, and infrastructures such as communication, energy, solid waste facilities, roads, public transportation, motor fueling supply chain (Beheshtian et al., 2017b), waste water management systems; degraded costal ecosystem (Arkema et al., 2013); increased salt front up the Hudson and Delaware Rivers leading to reduced supply of drinking water (AAP 2008); removing natural shoreline features (e.g. wetlands, aquatic vegetation, dunes and barrier beaches) providing flood protection and storm buffering, raised water table reducing soil ability to absorb runoff, and housing degradation (New York State Sea Level Rise Task Force Report to the Legislature Task Force, 2010); population displacement (Rosenzweig et al., 2010).

Sudden impacts Extreme temperature

� Heat wavea � Cold event

Heat-related morbidity including asthma, heat stroke, heat exhaustion, respiratory difficulties, and lung failures (Rosenzweig et al., 2005) and mortality (Confalonieri et al., 2007); utility outages, and exacerbating asthma and other respiratory diseases, frostbite, and hypothermia (New York City Hazard Mitigation Plan eNYCHMP- 2014); damaged infrastructures (e.g. potential cracked pavements) (White et al., 2010); rise in water peak demand flow and potential water scarcity (AAP 2008); increased of demand on HVAC; exacerbated UHI (Rosenzweig et al., 2009); cold-related illness including cardiovascular issues (NYC Environment & Health Data Portal).

Winter storm � Ice storm, freezing rain, and blizzard Interrupted critical infrastructures; public health consequences; impassable roads (NYCHMP 2014).

Coastal stormb � Storm surge and flooding � Wind

Damaged physical infrastructures and buildings, and disrupted supply chains of critical goods including food and liquid fuels, sewage overflow in low-elevation areas and wastewater treatment plants, and interrupted emergency and first- responders tasks (NYC Special Initiative for Rebuilding and Resiliency, aka SIRR, 2013); salt water corrosion; power outages, and submerged wetlands (NYCHMP 2014); mortality and injury; health care disruption; hazardous surface releases and contaminant leaching and runoff (Horton et al., 2015); wind-induced turbidity; violation on the level of water turbidity and quality consequent to runoff into reservoirs, and failure of drainage and wastewater treatment systems (AAP 2008); unregulated release of chemicals through surface soils and ground water (ARC-X).

Downpour � Flash flood Increase of street, basement and sewer flooding; increase in nutrient loads, eutrophication, taste and odor problems and loadings of pathogenic bacteria and parasites in reservoirs; increase in combined sewer overflow events; polluting coastal waterways (Rosenzweig et al., 2011).

Drought � Water scarcity Decrease in average reservoir storage and increase in strain on upstate reservoirs; diminished source water quality (ARC-X); public health challenges (Epstein and DeFilippo, 2001).

a ‘The National Weather Service uses a heat index related to temperature and humidity to define the likelihood of harm after “prolonged exposure or strenuous activity”’ (New York State Climate Action Council, 2010).

b New York City is vulnerable to two categories of coastal storms: tropical cyclones, aka hurricanes, and nor'easters. However, the impact of climate change on frequency and intensity of the nor'easters is unknown (Horton et al., 2010).

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e1313 1303

2010), economic loss pertaining to physical property damages (Noy, 2009), business interruptions induced by climatic events (Rose and Lim, 2002), national security implications (Adger, 2010), risk of armed and violent conflicts (Barnett and Adger, 2007), environ- mental degradation (Raleigh and Urdal, 2007), serious health out- comes (McMichael et al., 2006), social-inequities (Costello et al., 2009) and injustice (Whitmarsh, 2009), and forced mass- migration and resettlement (de Sherbinin et al., 2011).

10 A few arguments exist supporting the positive impacts of climate change on the energy sector, e.g., those highlighting lower energy consumption in response to milder winter weather (Horton et al., 2014). 11 Aside from increased temperature, those physical elements of NYC's ESC sited on flood-prone coastal areas are vulnerable to SLR.

3.3. Climate change impacts on New York City's energy sector

The changing climate is expected to increase mean temperature as well as the intensity and frequency of weather extremes, both

affecting NYC's energy sector. Studies concerning the gradual im- pacts of climate change on the energy sector are largely associated with the adverse impacts10 of increased temperature11 on the vulnerability of the key-components of the energy supply chain (ESC). These components include 1) resource endowment and supply, 2) distribution and inventory, and 3) demand and end users.

On the resource endowment and supply end, increased

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e13131304

temperature is recognized as impacting five forms of power gen- eration, described below:

� Combined cycle power plants face a drop in ‘peak output’ due to the rise in ambient air temperature (Van Vliet et al., 2012);

� Thermal power plants are subject to loss in operating efficiency through “back pressure” caused by the increased average tem- perature of incoming water (Hammer and Parshall, 2009);

� Hydro power plants affected by declining stream flow and runoff, themselves triggered by insufficient precipitation and drought (Hamlet et al., 2010);

� Solar energy resources could be reduced by “increased cloud cover attributable to rising CO2 levels” which is a consequence of the changing climate (Pan et al., 2004); and

� Wind farms may lose efficiency because of lowered wind speed induced by climate change (Breslow and Sailor, 2002).

On the distribution and inventory sides, equipment (e.g. trans- formers and transmission lines in the electricity sector) are vulnerable to failure due to raised ambient air temperature and potential heat waves. On the demand and end-user end, the main concern is related to peak demands12 for electricity-based air cooling in the summer, leading to power outages, whether through brownout or blackout. For more details on the energy sector impact of the changing climate, the reader is referred to the texts by Schaeffer et al. (2012).

Aside from vulnerability from gradual impacts, NYC's energy sector (including electricity, natural gas, and petroleum in- frastructures) is vulnerable to disruptive risks triggered by climatic extremes. A rich set of literature and reports exists covering the centrality of disaster preparedness and climate-adaptation plan- ning when discussing the impacts of climatic extremes on sub- sectors of NYC's energy industry (Zamuda et al., 2013). One of NYC's most disaster-susceptible sub-sectors of the ESC is the motor fueling infrastructure (NYC Hurricane Sandy After Action Report 2013). Stretched from upstream (i.e. exploration and drilling crude oil) to downstream (i.e. refining and retailing), the robustness of the well-to-pump infrastructure serving New York's inner-city mobility has never been fully studied, especially in the metropol- itan area where liquid fuels are replenished, distributed, stored, retailed, and burned.

In the following sections, without loss of generality, we inves- tigate climate change impacts on the motor fueling infrastructure of Manhattan, one of NYC's five boroughs. For the sake of brevity, in the rest of the paper we use ‘MESC’ to indicate Manhattan's motor fueling supply chain.

13 Aerts et al. (2014) estimate $2.2 billion and $25.4 billion as direct loss to the

3.4. Climate change impacts on New York City's transportation energy sector

A home to over 8 million residents, New York City's built envi- ronment is one of the most vulnerable nationwide to the effects of the changing climate. Among a range of threads listed in Table 1, weather extremes, in particular, coastal storms, have been at the center of attention, noticeably after the strike of hurricanes Irene (2011) and Sandy (2012). The climatic extremes revealed how disaster-susceptible NYC is, and how such vulnerability could result

12 The annual energy consumption pattern does not necessarily follow the energy demand during peak temperatures. Schaeffer et al. (2012) schematize the impact of peak temperature fluctuation on indoor energy consumption as a “U-shaped fashion”: at low/high temperatures “there is a relatively high energy demand,” while “at intermediate temperature the energy demand tend to be lower.” (Schaeffer et al., 2012, pp. 7).

in mass-casualty and enormous economic loss13 (SIRR 2013). Among the elements of the built environment, however, NYC's

interdependent critical infrastructures (ICIs) face major in- terruptions when stressed or under attack. NYC's ICIs are 1) disaster-susceptible, since they are physically exposed and spatially dispersed alongside 520 miles of tidal shoreline, and 2) functionally critical, since NYC's well-being and overall resilience14 are condi- tioned on the operability of NYC's ICIs, whether directly (i.e. when their availability is vital to external entities) or indirectly (i.e. when the operability of other infrastructure(s) vital to external entities depends15 on the availability of NYC's ICIs).

Among many ICIs, the MESC is one of the most critical, yet vulnerable, infrastructures to climatic events. During hurricane Sandy, for instance, the MESC experienced extensive and broad disruptions; refineries and terminals lost power and were physi- cally damaged, roads were inundated, all of which led to wide- spread gas station closures. According to an emergency survey conducted by the U.S. Energy Information Administration, from 4th through 11th following Sandy, 67-28% of gas stations across the New York metropolitan area (larger in the borough of Manhattan) did not have gasoline available for sale.

Despite the early conclusion many reached that these closures were due primarily to power outages that prevented stations from pumping gas, the larger problem turned out to be that stations simply had no gas to pump. The disruptions to the MESC yet began well upstream from the gas stations, and at nearly every compo- nent of the supply chain. Petroleum imports, which account for 27% of the total 2.2 million barrels of daily consumption on the East Coast, were shut down by Sandy (DOE 2013). Shipments from Gulf Coast refineries and terminals in the Buckeye and Colonial pipelines were down due to safety precautions that took place. Sandy also affected 94% of area terminals and refinery capacities (NACS 2013).

Additionally, thousands of roads were closed due to downed power lines and tree limbs that hampered the ability of trucks to get to open terminals or reservoirs for fuel deliveries. At the downstream MESC, retail was also interrupted by dysfunctional endpoints. Some service stations were inundated, had canopies toppled, roof panels shredded, or faced power cuts, resulting in inoperable pumps and hampered credit card transactions. The failure within the incapacitated MESC rippled across the interde- pendent network of transportation, and it was immediately recognized by first responders and other agents, whether directly or indirectly, who rely on a functional transportation system. Full restoration of the depressed MESC, which took more than a month, slowed down the recovery in other sectors.

3.5. Resilience-enhancing strategies and adaptive measures for MESC

Following the hurricane Sandy and to protect the operability of the MESC in time of extreme weather episodes, two physical resilience-enhancing strategies have been pursued.

1) transfer switch for back-up power generator in every down- state gas station within a half-mile of a highway exit or a

NYC economy caused by 100- and 1000-year potential flooding events, respectively. They also estimate $174 million expected economic loss, annually, in absence of the flood management strategies in NYC. 14 We recognize term resilience as “a system's property to better withstand and absorb, efficiently adapt to, and quickly/cheaply recover from the inoperability imposed by extreme events” (Beheshtian et al., 2017a). 15 This dependency includes ‘principal classes’ of physical, cyber, geographic and logical (Rinaldi et al., 2001).

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e1313 1305

hurricane evacuation route and on-site backup generators in gas stations in “strategic locations” to provide power when the utility is not available, as partially funded by the New York State Energy Research and Development Authority (NYSERDA PON 2924); 2) fuel reservoir tank, a pilot program on Long Island with 3 million gallons of fuel, to supply gasoline during a fuel shortage or prolonged disruption to MESC, as proposed through a $10 million project by the New York State governor's office (NYS Governor 2013).

16 The MFSC in Manhattan, NY, has a complex-unique network topology. The distribution system's wellbeing relies on availability of only twelve “entry points” (Omer et al., 2011) which are either tunnels or bridges. Therefore, while 99.9% and 99.4% of the roads across the case study area remain operable following the least- and most-severe extremes, respectively, the location (i.e. siting) of the flood- vulnerable transportation arcs and their criticality in distribution of the liquid- fuel demonstrate the bottlenecks' determining role in the system's overall functioning.

3.6. Technical, regulatory, and economics barriers to adaptation planning in New York city

Forward-thinking planning for an adaptive MESC in NYCe similarly modeling many problem sets featuring system of systems (SOSs) aspects e is quite a complex task for multiple reasons. First, critical-goods supply chains, such as those associated with the MESC, are inherently complex to design and operate due to their large size and their dynamic, time-variant behavior, heterogeneity in end-users, and extensive interdependencies with other critical facilities. Second, the MESC, as well as other NYC ICIs, are nearing the end of their service lives (New York Infrastructure Report Card, 2015). Aging and often over-utilized facilities introduce new types of operational difficulties, where the infrastructure nominal service rate is either unachievable or not adequate. Third, projecting the MESC's long-term demand is challenging when the en route future composition of vehicle fuel sources is unknown and subject to many uncertainties.

Lastly, existing regulatory and economics barriers to infra- structure investment in NYC are enormous and private investment in the New York region is challenging mainly due to two reasons. The first is the issue funding. For any infrastructure project to attract private financing, it must be bankable. This means that it must have a funding source that is sufficiently predictable to attract both debt and equity investors to the project. Funds to compensate investors for risk borne can come from one of two broad sources. The first is some type of project-specific user fee, such as a toll, price, or in some cases a mileage-based user fee. The second is from tax revenues that are dedicated to the project but are unrelated to the intensity of infrastructure use. Those may include dedicated sales tax revenues, property taxes, or income taxes. The critical issue for investors is that government guarantees of payment to investors, usually based on performance, be reliable. Given the strong property and contractual protections enjoyed by investors in U.S. infrastructure, adequate project funding will attract global capital. The challenge is to structure projects so that adequate funding is assured.

Another major challenge relates to the multi-jurisdictional na- ture of the New York metro area. The area stretches across three states: New York, New Jersey, and Connecticut. Some facilities cross state lines, such as many of NYC's bridges and tunnels. Others fall within the city's limits only. Depending on project details, several jurisdictions may need to cooperate. Moreover, New York State makes extensive use of the authority structure, including the Port Authority of New York and New Jersey and the New York State Thruway Authority. Many of the bridges, tunnels, and airports in the NYC region fall under the control of the Port Authority, which streamlines project delivery. However, the Port Authority is controlled by the governors of both New York and New Jersey, and mayor of New York City may also come into play in many projects. The uniquely complex multi-jurisdictional nature of many projects in the NYC area is likely to be a major challenge for many infra- structure investors.

4. Vulnerability simulation and modeling results

Fig. 3 shows fueling elements of the MESC, including all of Manhattan's gas stations, 29 in total, in addition to 28 terminals and 8 refineries supplying Manhattan's service stations. These elements are linked through a transportation network of 1268 bidirectional arcs, 2440.95 miles in total.

We first investigate the vulnerability of the MESC0 elements, individually, and disregard their functioning roles as system com- ponents. Furthermore, we define 8 discrete scenarios representing the random extremes that potentially threaten the case study area. These scenarios simulate 100- and 500-year flooding (YF) events in 2017, 2020s, 2050s, and 2080s. Fig. 4 shows the projection of flooding events on a part of the MESC (lower Manhattan) and for the scenarios described above. Among the infrastructure's ele- ments, refineries and terminals are the most vulnerable, since many of these developments are sited in low-lying coastal areas and are thus vulnerable to flooding from storm surge and SLR.

Manhattan's gas stations are relatively less vulnerable. Among 29, 4 are vulnerable to today's (2017) 100-YF. However, under the worst-case scenario that we simulate (i.e. 2080s 500-YF), just 13, or around 21%, of service stations remain operable. By contrast, the transportation infrastructure is more sprawling and well- developed inland where inundation is less of a concern. Accord- ingly, 99.9% and 99.4% of the roads remain functional during the less and most severe scenarios, respectively. Yet the small portion of roads situated in flood-prone areas will turn into distribution system bottlenecks16 in times of flooding.

Investigating the MESC's productivity (output) for every sce- nario, we model fuel distribution assignment through a truck and trailer vehicle routing problem (TTVRP). Based on a generic shortest path problem, the TTVRP model simulates a series of scenario- specific fuel distribution assignment to include 1) locating the supply nodes and sizing the shipments subject to distribution, 2) routing tankers on the transportation network with respect to the shortest-open-path between operating supply nodes and func- tioning service stations, and 3) locating open delivery stations, and sizing shipments subject to unload in every station.

With respect to the TTVRP model, five metrics are analyzed representing the operability of the MESC's critical elements (i.e. terminals, refineries, gas stations, and roads), individually, as well as the system's overall output (i.e. the ratio of fuel delivered to fuel demand rate) which is conditioned on the first four metrics. The MESC's overall output is modeled in the absence of consumer behavior. This assumption will be relaxed in the following section. As well, the estimated outputs are based on a commonly-used assumption in risk analysis: elements located in flood-prone areas become fully-dysfunctional in time of flooding. The scenario-based MESC outputs are shown in Fig. 5.

The supply nodes (i.e. terminals and refineries) are the most vulnerable during the severe flooding scenarios. During a 2080s 500-YF event, nearly four-fifths of the terminals become inoper- able. Yet, more than eighty-five percent of the refineries experience dysfunctionality during the same scenario. The reason, clearly, is the fact that a majority of the terminals and refineries are supplied by delivery barges and tanker vessels, and hence are required to be

Fig. 3. The MESC.

Fig. 4. Floodplain projections for a part of the case study area (lower Manhattan); (a): 2020s 100-YF; (b): 2050s 100-YF; (c): 2080s 100-YF; (d): 2020s 500-YF; (e): 2050s 500-YF; (f): 2080s 500-YF.

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e13131306

in close proximity to the flood-prone shoreline. Compared to sup- ply nodes, service stations are less vulnerable.

Through the most severe event (i.e. a 2080s 500-YF), nearly half of the service stations remain functional, while, for less severe events (i.e. 2017 flooding scenarios), this would be something in the region of 80e86 percent. Drawing a comparison between the MESC output (black solid line in Fig. 5) and the functionality of the MESC's individual components (vertical bars in Fig. 5), the MESC as an in- tegrated system entails higher vulnerability since several

bottlenecks arise across critical elements that exacerbate the in- frastructure's overall throughput.

Next, we simulate the MESC functionality interfaced with end- user behavior in three time frames: pre-, during, and post- extremes. To this end, we assume 1) consumer anxiety, as driven by hazard-triggered psychological stress, increases fuel demand by 20% as early as 20 and 36 h (hr) prior to a 100- and 500-YF, respectively. Consumer anxiety is assumed to end when hazards strike; 2) any activity in the supply chain (e.g. refining, blending,

Fig. 5. The MESC overall output and the elements' vulnerability against flooding scenarios. Note: values corresponding to the transportation arcs are associated to the secondary (right) vertical axis.

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e1313 1307

distributing, retailing, etc.) ceases for a period of 4 h during a storm strike; 3) refineries and terminal operators shutter their facility 12 and 24 h prior to a 100- and 500-YF, respectively, to minimize damage to equipment (as experienced during hurricane Sandy (SIRR 2013)).

The simulation is sited at the intersection of the TTVRP and a consecutive-fuel-replenishments model. The TTVRP is simplified in Equations (1)e(3) and its detailed formulation is provided in Appendix.

Min: WP* unmet fuel demand þ distribution cost (1)

s:t: unmet fuel demand ¼ f ðMESC functionalityÞ (2)

distribution cost ¼ ð̂shotrest; open pathÞ (3) Expression (1) represents the objective function. To be mini-

mized is the fuel shipment cost in addition to penalty associating to every gallon of unmet fuel demand. Scalar WP, a sufficiently large number, is embedded to prioritizing the maximum met demand over the distribution cost.

The optimization model is subject to two constraints. Constrain 2 conditions the overall output on the MESC's availability during a given scenario. The MESC's output, defined as unmet fuel demand, hence would be a function of system's overall wellbeing ði:e: MESC functionalityÞ. Constraint 3 describes the fuel distribu- tion cost which is a distance-related variable based on the shortest, unflooded (open) path between operable elements of the MESC.

The consecutive-fuel-replenishments model is shown in Equa- tion (4) for two consecutive time periods of t and tþ1.

TIt�1 þ CSIt�1 � CDOt�1 ¼ TIt (4) Denoting TI for total inventory (i.e. Manhattan-wide fuel avail-

ability), CSI for cumulative supply inflow (i.e. terminals/refineries outflow of shipment), and CDO for the cumulative demand outflow (i.e. fuel retailed at end-user points), Equation (4) schedules fuel shipment with respect to the availability of suppliers and demand points. This constraint holds the flow conservation in every service station, meaning in time t, the available inventory in a given service station equals to inventory in addition to the balance of inflow and

outflow in the previous time frame t-1. The model simulates the MESC functionality for a period of

152 h starting 36 h prior to the onset of flooding. Fig. 6 displays fuel availability (FA), with the MESC throughput assuming end-user behaviors, during the 8 scenarios defined earlier. Aside from the FA, we measure two more indicators best representing the impact of the disturbed MESC on the interdependent infrastructure of transportation: transportation network input (TNI) and vehicle miles traveled (VMT).

To bridge the gap between FA and these two metrics, we assume service stations control the size of the ration by meeting just a fraction of demands, no more than the ration of fuel they have available on-site. For example, a sale rate of a given service station with 65% fuel available in the underground tank does not exceed 65% of demand in the BAU state. As experienced following Sandy, this could be implemented through an odd-even license plate fuel rationing program resulting in shorter lines at fueling stations, prolonged service time, and keeping price below the equilibrium price.

The TNI, in general, reflects the cumulative number of trips being generated across the network and in origins or traffic analysis zones (TAZs). Under a fuel scarcity and rationing program, however, travelers are assumed to be more cautious in consuming fuel, hence, forgoing unnecessary trips. Therefore, we maximize the number of work-related trips (i.e. necessary trips, in contrary to non-work related trips such trips for shopping or leisure) backed by available fuel, FA, and assigned to operable, unflooded, trans- portation arcs. The second metric, VMT, is a function of TNI and defined as the cumulative miles traveled by trips assigned through TNI. The TNI and VMT metrics are illustrated and compared with the FA graphs in Fig. 6.

In all scenarios, the FAs gradually decline as early as 24 and 40 h prior to 100- and 500-YFs, respectively, the time end-users reveal concerns. The FAs, however, experience steeper drops during the time window where end-user anxiety escalates demand, and when supply is interrupted by shut down facilities. The steeper drops discussed above, however, do not immediately respond to the ter- minal closures, since the inventory left in service stations maintains some level of availability across the stations for a short period of time. During the 2017 scenarios, the FA is expected to decline 46e50 percent by the time extreme events arrive, while these rates

Fig. 6. MESC's behavior post flooding events; projected for 2017, 2020s, 2050s, and 2080s. Note: The values are normalized as a percent of the value in BAU.

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e13131308

are expected to be 22e45 percent during 2080s scenarios. The post-hazard decline of FAs is relatively steady when the

lifted anxiety alleviates demand and uninterrupted terminals/re- fineries start operating. Making a comparison between scenarios, less aggressive drops are observed on FAs associated with 100-YFs as well as the extremes projected for earlier times. Despite the less- aggressive drops that the metric of FA experiences following extreme events, dysfunctional supply nodes, flooded arcs, and insufficient inventory levels in service stations turn the MESC into full-inoperability. The MESC total shutdown is expected to happen as early as 80, 64, 48, and 20 h following a 500-YF in 2017, and the 2020s, 2050s, and 2080s, respectively.

The ratios of the trips generated across the network, TNI metric, are higher than the FAs in the same time frame. This is due to the assumption that non-work-related trips (accounts for nearly 32 percent of the overall trips) forgo due to fuel scarcity. Accordingly, fuel shortages adversely impact just a portion of the work-related travel demand. The gaps between TNIs and FAs merge in time, meaning that, when the FA is not considerable, a lower ratio of demand associated with work-related trips is met.

The ratio of TNI to FA shows the dependence of travel demand upon fuel availability. This ratio (TNI over FA) during 100-year flood scenarios, on average, is slightly higher than the same measures

during 500-YF scenarios, despite the fact that the TNI is linearly conditioned on FA. Interestingly, this shows that following 500-YF scenarios, the level of forgone trips is only partially related to fuel availability, and in part may be triggered by the lack of connect- edness between origin and destination pairs.

The VMT metrics are also associated with TNIs. When the TNI is conditioned on FA and larger, reasonably, the corresponding VMT is lower, assuring a larger ratio of travel demand met by a smaller ratio of available fuel. This is due to traveler precautions in fuel consumption: making more, yet shorter-distance trips, avoiding non-work-related trips, or work-related, but long-distance trips.

We earlier assumed binary-elimination of the MESC's elements, meaning infrastructure components sited in flood-prone zones become fully-dysfunctional in time of flooding. To relax this assumption, we simulate random failures incapacitating vulnerable elements through different levels (with the exception of roads). since even if a portion of the road sited in flood-prone areas, the whole road is considered closed). To model such random failures across the MESC's critical elements, we pre-define a series of pos- sibilities for each of the elements. Each possibility is assigned with an input value representing the element's dysfunctionality/in- capacity level as well as an associated occurrence likelihood. Pos- sibilities and their properties are listed in Table 2.

Table 2 Possibilities.

Element subject to partial failure Incapacity properties

Variation from BAU's state (%) Occurrence chance (%)

100-year flood 500-year flood

Gas stations �10 50 10 �30 25 20 �60 15 30 �100 10 40

Terminals and refineries �30 70 70 �100 30 30

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e1313 1309

Having randomized the (in)capacity of critical elements, we develop a set of eight discrete sub-scenarios under every 100- and 500-year flooding event. Each sub-scenario covers a unique combination of two possibilities (i.e. separate incapacity condition for gas stations and terminals/refineries) and carries one occur- rence chance, which is the product of the possibilities’ occurrence chance.

The expression (5), new objective function, reflects the weighted average across the occurrence probabilities associated with sub-scenarios. Fig. 7 shows the MESC's outputs against a range of random failures in each scenario.

Min: X8

sub�scenario¼1 ðWP�unmet fuel demand þ distribution costÞ

(5)

As discussed, two types of randomness were simulated to show the system's overall functionality in the face of disrupting sce- narios: whether the elements fully-absorb the shock, and if not, to what extent the shock hampers the capacity of elements. Drops in the MESC's output are mainly caused by road closures. The vulnerability of terminals, refineries, and gas stations have rela- tively less impact on the system's overall performance. The reason relates to the topology of Manhattan's bridges connecting this is- land to other boroughs and New Jersey, which are all situated in flood-prone areas.

Because of this, the drops are steeper even after the failure of small ratio of critical elements. Through intensive flooding sce- narios, i.e., the 2080s’ 100- and 500-YFs, the MESC outputs drop very sharply in response to 2e3% inoperable elements. However, the MESCs drop smoother under more recent or less intensive scenarios. The MESCs throughputs in all scenarios reach points of

Fig. 7. MESC output under random elimin

saturation when 30e85% of flood-vulnerable elements become fully- or partially-inoperable.

5. Conclusion

Decentralized infrastructure incorporating renewable energy and large-scale adoption of alternative fuel vehicles could be an example of strategic planning by fulfilling mitigation and adapta- tion requirements simultaneously, in particular in the case of NYC where 1) expanded access to transportation alternatives and electricity-based mobility is a recognized and well-supported mitigation strategy, and 2) the motor fueling supply chain is asset intensive and highly vulnerable to climatic extremes.

Through the proposed research, we modeled the vulnerability of the NYC's transportation fueling infrastructure to climatic hazards and studied the travel behavior in time of disaster, which are necessary steps toward synergetic approaches to climate adapta- tion and mitigation. The modeling approach advanced in three steps: 1) vulnerability assessment of critical elements and the transportation energy system's overall resilience; 2) fuel availabil- ity across the service stations and its impact on travel-met demand and vehicle mile traveled by commuters between 36 h prior to and 5 days following the events; 3) the transportation energy sector's response to a range of randomness in both the arrival and intensity of climatic events.

Modeling results detailed the vulnerability of NYC's existing energy system to climate, and the adaptation challenges the city faces in tackling this challenge. Even through current-day potential flooding events (which are less severe in comparison with the in- tensity of floods projected for the future), the borough of Man- hattan will run out of fuel in less than four days, impacting millions of daily commuters, hampering the recovery process, slowing

ation of the infrastructure elements.

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e13131310

down the economy, and, more importantly, blocking first- responders emergency tasks. Models also revealed how and to what extent the transportation energy system in Manhattan de- pends on functioning bridges connecting the borough to other lands.

The unique geographical dispersion and spatial distribution of the motor fuel supply chain in NYC, which makes resilience- enhancing strategies capital intensive, along with a range of regu- latory barriers NYC faces, keep NYC's transportation energy sector far from being considered climate-adaptable. This condition, however, is not NYC-specific. On a different scale, this could be the challenge for other metro areas, such as Houston and Miami, which recently experienced widespread dysfunctionality across their transportation energy sectors in the aftermath of hurricane Harvey and Irma, respectively.

We suggest several directions for future research and opportu- nities for theory development using a synergistic approach to adaptation and mitigation. Above, we mainly discussed adaptation challenges for transportation energy infrastructure. However, our modeling results should be used as input parameters for investi- gating the optimal location and sizing of charging stations across the case study area.

Table A1 Nomenclature

Sets {n} ¼ {1, 2, …, N} Scenarios {a} ¼ {1, 2, …, A} Capacity-enhancing (investment) strategies {s} ¼ {1, 2, …, S} Supply nodes (terminals/refineries) {t} ¼ {1,2, …, T} Transshipment nodes {d} ¼ {1,2, …, D} Demand nodes (gas stations) {k} ¼ {1,2, …, K} Fuel types Parameters ℙn Chance of occurrence of scenario n L A large number

Fkd;n Demand of commodity k in demand node d

Fks;n Supply of commodity k in supply node s dur

NOðdkÞ Number of demand nodes (gas stations) in t Cks;t;n Distance between supply node s and trans-s

Ckt;tt;n Distance between trans-shipment nodes dur

Ckt;d;n Distance between trans-shipment node t an

Oks;n Operability of node s during scenario n. 1 if

Okd;n Operability of node d during scenario n. 1 if

Os;t;n Operability of arc (s,t) during scenario n. 1 if Ot;tt;n Operability of trans-shipment arc during sce Ot;d;n Operability of arc (t,d) during scenario n. 1 i Ws;t;n Nominal capacity of arc (s,t) during scenario Wt;tt;n Nominal capacity of the transshipment arc d Wt;d;n Nominal capacity of arc (t,d) during scenario Decision variables Ukd;n

UDR (unfilled demand rate) corresponding t

Xks;t;n Flow type k on arc (s,t), during scenario n

Xkt;tt;n Flow type k on transshipment arc, during sc

Xkt;d;n Flow type k on arc (s,d), during scenario n

V Number of investment strategies

Ik;as Binary variable; 1 if investment strategy a is

Ik;ad Binary variable; 1 if investment strategy a is

Ias;t Binary variable; 1 if investment strategy a is Iat;tt Binary variable; 1 if investment strategy a is Iat;d Binary variable; 1 if investment strategy a is

Second, mitigation policies are heavily incentivized by the fed- eral government, states, and nearly all metropolitan areas across the nation. On the mitigation side, regulations are also in place and well-supported. However, this is not the case for neither adaptation capacities, nor synergistic models. Therefore, we suggest advancing regulatory and finance models needed to deliver a successful fully integrated planning approach to the critical needs of the trans- portation energy sector in metropolitan areas. Therefore, we sug- gest further studies on regulatory and finance models which are needed to deliver a successful fully integrated planning approach to the critical needs of the transportation energy sector in metropol- itan areas.

Acknowledgements

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Appendix

during scenario n

ing scenario n

he borough of Manhattan

hipment node t during scenario n

ing scenario n

d demand node d during scenario n

operable, 0 otherwise

operable, 0 otherwise

operable, 0 otherwise nario n. 1 if operable, 0 otherwise f operable, 0 otherwise n uring scenario n n

o fuel type k in demand node d during scenario n

enario n

chosen for facilities corresponding to fuel type k at supply node s, 0 otherwise

chosen for facilities corresponding to fuel type k at demand node d, 0 otherwise

chosen for arc (s,t), 0 otherwise chosen for transshipment arc, 0 otherwise chosen for arc (t,d), 0 otherwise

Employing the notation synopsized in Table A.1, Manhattan's MESC vulnerability analysis is formulated as the following bi-stage stochastic program.

Min

XN n¼1

ℙn

" L

PD d¼1 PK

k¼1U k d;nPK

k¼1NO � dk � þ XS

s¼1

XT t¼1

XK k¼1

Xks;t;nC k s;t;n þ

XT t¼1

XTT tt¼1

XK k¼1

Xkt;tt;nC k t;tt;n þ

XT t¼1

XD d¼1

XK k¼1

Xkt;d;nC k t;d;n

# (1)

s.t.

V ¼ XS s¼1

XK k¼1

Ik;as þ XD d¼1

XK k¼1

Ik;ad þ XS s¼1

XT t¼1

Ias;t þ XT t¼1

XTT tt¼1

Iat;tt þ XT t¼1

XD d¼1

Iat;d (2)

V � 1 (3)

Ukd;n ¼ Fkd;n �

XT t¼1

Xkt;d;n

!, Fkd;n (4)

Fks;n h Oks;n þ ð1Þ Ik;as

i � XT t¼1

Xks;t;ncs; n; k (5)

Fkd;n h Okd;n þ

� 1 � Okd;n

� Ik;as i � XT t¼1

Xkt;d;n c d; n; k (6)

XK k¼1

Xks;t;n � Ws;t;n h Os;t;n þ

� 1 � Os;t;n

� Ias;t i c s; t; n (7)

XK k¼1

Xkt;tt;n � Wt;tt;n h Ot;tt;n þ

� 1 � Ot;tt;n

� Iat;tt i c t; tt; n (8)

XK k¼1

Xkt;d;n � Wt;d;n h Ot;d;n þ

� 1 � Ot;d;n

� Iat;d i c s; t; n (9)

Xks;t;n � 0 ck; s; t; n (10)

Xkt;tt;n � 0 ck; t; tt; n (11)

Xkt;d;n � 0 c k; t; d; n (12)

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e1313 1311

References

Abdelgawad, H., Abdulhai, B., 2009. Emergency evacuation planning as a network design problem: a critical review. Transp. Lett. 1 (1), 41e58.

Adamo, S.B., 2010. Environmental migration and cities in the context of global environmental change. Curr. Opin. Environ. Sustain. 2 (3), 161e165.

Adger, W.N., 2010. Climate change, human well-being and insecurity. New Polit. Econ. 15 (2), 275e292.

Aerts, J.C., Botzen, W.W., Emanuel, K., Lin, N., de Moel, H., Michel-Kerjan, E.O., 2014. Evaluating flood resilience strategies for coastal megacities. Science 344 (6183), 473e475.

Alexander, L.V., Zhang, X., Peterson, T.C., Caesar, J., Gleason, B., Klein Tank, A.M.G., Haylock, M., Collins, D., Trewin, B., Rahimzadeh, F., Tagipour, A., 2006. Global observed changes in daily climate extremes of temperature and precipitation. J. Geophys. Res. Atmos. 111 (D5).

Arkema, K.K., Guannel, G., Verutes, G., Wood, S.A., Guerry, A., Ruckelshaus, M., Kareiva, P., Lacayo, M., Silver, J.M., 2013. Coastal habitats shield people and property from sea-level rise and storms. Nat. Clim. Change 3 (10), 913.

Barnett, J., Adger, W.N., 2007. Climate change, human security and violent conflict. Polit. Geogr. 26 (6), 639e655.

Beheshtian, A., 2016. Planning Resilient Infrastructure. Beheshtian, A., Donaghy, K.P., Rouhani, O.M., 2016. Flood-resilient deployment of

fueling stations: extension of facility location problem. Transp. Res. Rec. J. Transp. Res. Board (2599), 81e90.

Beheshtian, A., Donaghy, K.P., Geddes, R.R., Rouhani, O.M., 2017a. Planning resilient motor-fuel supply chain. Int. J. Disaster Risk Reduct. 24, 312e325.

Beheshtian, A., Donaghy, K.P., Geddes, R.R., Rouhani, O.M., 2017b. Adaptation Planning for Climate-resilient Urban Infrastructure (No. 17-06809).

Beheshtian, A., Donaghy, K.P., Geddes, R.R., 2018. Motor Fuel Supply Chain

Vulnerability to Climatic Extremes (No18-07636). Bjørnæs, C., 2013. A guide to representative concentration pathways. Cicero. Cent.

Int. Clim. Environ. Res. http://cicero.uio.no/images/A%20guide%20to%20RCPs. pdf. (Accessed 11 November 2017).

Breslow, P.B., Sailor, D.J., 2002. Vulnerability of wind power resources to climate change in the continental United States. Renew. Energy 27 (4), 585e598.

Chang, S.E., Nojima, N., 2001. Measuring post-disaster transportation system per- formance: the 1995 Kobe earthquake in comparative perspective. Transp. Res. Part A Policy Pract. 35 (6), 475e494.

Church, J.A., White, N.J., 2011. Sea-level rise from the late 19th to the early 21st century. Surv. Geophys. 32 (4e5), 585e602.

Church, J.A., Clark, P.U., Cazenave, A., Gregory, J.M., Jevrejeva, S., Levermann, A., Merrifield, M.A., Milne, G.A., Nerem, R.S., Nunn, P.D., Payne, A.J., 2013. Climate Change 2013: the Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Sea Level Change, pp. 1137e1216.

Climate Change Adaptation Resource Center (ARC-X), last access 4/29/2017, https:// www.epa.gov/arc-x.

Climate Change Program Assessment and Action Plan (AAP), 2008. The NYC Department of Environmental Protection (DEP). Accessed 11.04.17. http:// home2.nyc.gov/html/dep/html/news/climate_change_report_05-08.shtml.

Cohen, N. (Ed.), 2011. Green cities: an A-to-Z Guide, vol. 4. Sage. Comparing the Impacts of Northeast Hurricanes on Energy Infrastructure Office of

Electricity Delivery and Energy Reliability U.S. Department of Energy, last accessed 7/9/2017, https://energy.gov/sites/prod/files/2013/04/f0/Northeast% 20Storm%20Comparison_FINAL_041513b.pdf.

Confalonieri, U., Menne, B., Akhtar, R., Ebi, K.L., Hauengue, M., Kovats, R.S., Revich, B., Woodward, A., 2007. Human Health. Climate Change 2007: Impacts, Adaptation and Vulnerability: Contribution of Working Group II to the Fourth Assessment

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e13131312

Report of the Intergovernmental Panel on Climate Change. Costello, A., Abbas, M., Allen, A., Ball, S., Bell, S., Bellamy, R., Friel, S., Groce, N.,

Johnson, A., Kett, M., Lee, M., 2009. Managing the health effects of climate change. Lancet 373 (9676), 1693e1733.

Crowley, T.J., 2000. Causes of climate change over the past 1000 years. Science 289 (5477), 270e277.

Cubasch, U., Wuebbles, D., Chen, D., Facchini, M.C., Frame, D., Mahowald, N., Winther, J.G., 2013. Climate Change 2013: the Physical Science Basis. Contri- bution of Working Group I to the Fifth Assessment Report of the Intergovern- mental Panel on Climate Change Cambridge. United Kingdom and New York, NY, USA, pp. 119e158.

Dale, V.H., 1997. The relationship between land-use change and climate change. Ecol. Appl. 7 (3), 753e769.

de Sherbinin, A., Castro, M., Gemenne, F., Cernea, M.M., Adamo, S., Fearnside, P.M., Krieger, G., Lahmani, S., Oliver-Smith, A., Pankhurst, A., Scudder, T., 2011. Pre- paring for resettlement associated with climate change. Science 334 (6055), 456e457.

Dessai, S., Adger, W.N., Hulme, M., Turnpenny, J., K€ohler, J., Warren, R., 2004. Defining and experiencing dangerous climate change. Clim. Change 64 (1), 11e25.

Dong, J., Liu, C., Lin, Z., 2014. Charging infrastructure planning for promoting battery electric vehicles: an activity-based approach using multiday travel data. Transp. Res. Part C Emerg. Technol. 38, 44e55.

Easterling, D.R., Evans, J.L., Groisman, P.Y., Karl, T.R., Kunkel, K.E., Ambenje, P., 2000. Observed variability and trends in extreme climate events: a brief review. Bull. Am. Meteorol. Soc. 81 (3), 417e425.

Ebi, K.L., McGregor, G., 2008. Climate change, tropospheric ozone and particulate matter, and health impacts. Environ. health Perspect. 116 (11), 1449.

Energy and Climate Outlook, 2014. https://globalchange.mit.edu/publications/ signature/2014-energy-and-climate-outlook. (Accessed 29 April 2017).

Epstein, P.R., DeFilippo, C., 2001. West Nile virus and drought. Glob. change Hum. health 2 (2), 105e107.

Frade, I., Ribeiro, A., Gonçalves, G., Antunes, A., 2011. Optimal location of charging stations for electric vehicles in a neighborhood in Lisbon, Portugal. Transp. Res. Rec. J. Transp. Res. board (2252), 91e98.

Fu, H., Wilmot, C., 2006. Survival analysis-based dynamic travel demand models for hurricane evacuation. Transp. Res. Rec. J. Transp. Res. Board (1964), 211e218.

Fujino, J., Nair, R., Kainuma, M., Masui, T., Matsuoka, Y., 2006. Multi-gas mitigation analysis on stabilization scenarios using AIM global model. Energy J. 343e353.

Gill, S.E., Handley, J.F., Ennos, A.R., Pauleit, S., 2007. Adapting cities for climate change: the role of the green infrastructure. Built Environ. 33 (1), 115e133.

Governor Cuomo Launches First-ever Strategic Gasoline Reserve to Prevent Supply Gaps during Emergencies, 2013. https://www.governor.ny.gov/news/governor- cuomo-launches-first-ever-strategic-gasoline-reserve-prevent-supply-gaps- during. (Accessed 3 November 2017).

Hallegatte, S., Green, C., Nicholls, R.J., Corfee-Morlot, J., 2013. Future flood losses in major coastal cities. Nat. Clim. change 3 (9), 802e806.

Hamlet, A.F., Lee, S.Y., Mickelson, K.E., Elsner, M.M., 2010. Effects of projected climate change on energy supply and demand in the Pacific Northwest and Washington state. Clim. Change 102 (1), 103e128.

Hammer, S.A., Parshall, L., 2009. Climate Change and New York City's Energy Sector: Vulnerabilities, Impacts and Potential Adaptation Strategies. Tech. rept.. Columbia University Center for Energy, Marine Transportation and Public Pol- icy. vulnerabilities, impacts and adaptation strategies. Contract, 16, p.32.

Han, L.D., Yuan, F., Urbanik, T., 2007. What is an effective evacuation operation? J. urban Plan. Dev. 133 (1), 3e8.

He, X., Liu, H.X., 2012. Modeling the day-to-day traffic evolution process after an unexpected network disruption. Transp. Res. Part B Methodol. 46 (1), 50e71.

Hoogendoorn, R.G., Tamminga, G., Hoogendoorn, S.P., Daamen, W., 2010, September. Longitudinal driving behavior under adverse weather conditions: adaptation effects, model performance and freeway capacity in case of fog. In: Intelligent Transportation Systems (Itsc), 2010 13th International Ieee Confer- ence on. IEEE, pp. 450e455.

Horton, R., Little, C., Gornitz, V., Bader, D., Oppenheimer, M., 2015. New York City Panel on Climate Change 2015 report chapter 2: sea level rise and coastal storms. Ann. N. Y. Acad. Sci. 1336 (1), 36e44.

Horton, R., Rosenzweig, C., Gornitz, V., Bader, D., O'Grady, M., 2010. Climate risk information. Ann. N. Y. Acad. Sci. 1196 (1), 147e228.

Horton, R.M., Bader, D.A., Rosenzweig, C., DeGaetano, A.T., Solecki, W., 2014. Climate Change in New York State Updating the 2011 ClimAID Climate Risk Information Supplement to NYSERDA Report 11-18 (Responding to Climate Change in New York State).

Inventory of New York City’s Greenhouse Gas Emissions, 2016. April 2016, by Cventure LLC, Cathy Pasion, Mikael Amar, and Yun Zhou, Mayor's Office of Sustainability, New York. https://www1.nyc.gov/assets/sustainability/ downloads/pdf/publications/NYC_GHG_Inventory_2014.pdf. (Accessed 29 April 2017).

Jenelius, E., Mattsson, L.G., 2012. Road network vulnerability analysis of area- covering disruptions: a grid-based approach with case study. Transp. Res. part A policy Pract. 46 (5), 746e760.

Jones, P.D., Trenberth, K.E., Ambenje, P., Bojariu, R., Easterling, D., Klein, T., Parker, D., Renwick, J., Rusticucci, M., Soden, B., 2007. Observations: surface and atmo- spheric climate change. Clim. change 235e336.

Kalafatas, G., Peeta, S., 2009. Planning for evacuation: insights from an efficient network design model. J. Infrast. Syst. 15 (1), 21e30.

Knutson, T.R., McBride, J.L., Chan, J., Emanuel, K., Holland, G., Landsea, C., Held, I., Kossin, J.P., Srivastava, A.K., Sugi, M., 2010. Tropical cyclones and climate change. Nat. Geosci. 3 (3), 157e163.

Kongsomsaksakul, S., Yang, C., Chen, A., 2005. Shelter location-allocation model for flood evacuation planning. J. East. Asia Soc. Transp. Stud. 6, 4237e4252.

Kuby, M., Lines, L., Schultz, R., Xie, Z., Kim, J.G., Lim, S., 2009. Optimization of hydrogen stations in Florida using the flow-refueling location model. Int. J. hydrogen energy 34 (15), 6045e6064.

Laboratory for Satellite Altimetry: Sea level rise, National Oceanic and Atmospheric Administration, 2015 (Accessed 8 November 2017). www.star.nesdis.noaa.gov/ sod/lsa/SeaLevelRise/LSA_SLR_timeseries_global.php.

Lambin, E.F., Turner, B.L., Geist, H.J., Agbola, S.B., Angelsen, A., Bruce, J.W., Coomes, O.T., Dirzo, R., Fischer, G., Folke, C., George, P., 2001. The causes of land- use and land-cover change: moving beyond the myths. Glob. Environ. change 11 (4), 261e269.

Lin, N., Emanuel, K., Oppenheimer, M., Vanmarcke, E., 2012. Physically based assessment of hurricane surge threat under climate change. Nat. Clim. Change 2 (6), 462e467.

Lin, Z., Ogden, J., Fan, Y., Chen, C.W., 2008. The fuel-travel-back approach to hydrogen station siting. Int. J. hydrogen energy 33 (12), 3096e3101.

Little, R.G., 2010. Managing the risk of cascading failure in complex urban in- frastructures. Disrupted cities When Infrastruct. fails 27e39.

Lou, Y., Zhang, L., 2011. Defending transportation networks against random and targeted attacks. Transp. Res. Rec. J. Transp. Res. Board (2234), 31e40.

McMichael, A.J., 2003. Global climate change and health: an old story writ large. In: Climate Change and Human Health: Risks and Responses. World Health Orga- nization, Geneva, Switzerland.

McMichael, A.J., Woodruff, R.E., Hales, S., 2006. Climate change and human health: present and future risks. Lancet 367 (9513), 859e869.

Melillo, J.M., Richmond, T.T., Yohe, G., 2014. Climate Change Impacts in the United States. Third National Climate Assessment.

Mimura, N., 2013. Sea-level rise caused by climate change and its implications for society. Proc. Jpn. Acad. Ser. B 89 (7), 281e301.

Moffatt, S., Kohler, N., 2008. Conceptualizing the built environment as a social- eecological system. Build. Res. Inf. 36 (3), 248e268.

Murdoch, P.S., Baron, J.S., Miller, T.L., 2000. Potential effects of climate change on surface-water quality in North America. JAWRA J. Am. Water Resour. Assoc. 36 (2), 347e366.

National Association of Convenience Stores How Hurricane Sandy Affected the Fuels Industry, 2013. http://www.nacsonline.com/YourBusiness/FuelsReports/ GasPrices_2013/Pages/How-Hurricane-Sandy-Affected-the-Fuels-Industry. aspx. (Accessed 24 September 2017).

New York City Hazard Mitigation Plan, 2014. http://www1.nyc.gov/site/em/ready/ hazard-mitigation.page. (Accessed 4 November 2017).

New York Infrastructure Report Card, 2015. http://www.infrastructurereportcard. org/state-item/new-york/. (Accessed 21 March 2017).

New York State Climate Action Council, 2010. Chapter 2, Climate Projections and Vulnerabilitieshttp. Interim Report. Accessed 11.04.17. www.dec.ny.gov/docs/ administration_pdf/irchap2.pdf.

New York State Sea Level Rise Task Force. Report to the Legislature, December 2010. http://www.dec.ny.gov/docs/administration_pdf/slrtffinalrep.pdf. (Accessed 29 April 2017).

Nicholas, M., Handy, S., Sperling, D., 2004. Using geographic information systems to evaluate siting and networks of hydrogen stations. Transp. Res. Rec. J. Transp. Res. Board (1880), 126e134.

Nicholls, R.J., Cazenave, A., 2010. Sea-level rise and its impact on coastal zones. Science 328 (5985), 1517e1520.

Noy, I., 2009. The macroeconomic consequences of disasters. J. Dev. Econ. 88 (2), 221e231.

NYC Hurricane Sandy after Action Report, May 2013. http://www.nyc.gov/html/ recovery/downloads/pdf/sandy_aar_5.2.13.pdf. (Accessed 4 November 2017).

Omer, M., Mostashari, A., Nilchiani, R., 2011. Measuring the resiliency of the Man- hattan points of entry in the face of severe disruption. Am. J. Eng. Appl. Sci. 4 (1), 153.

Outlook, S.A.E., 2013. World Energy Outlook Special Report. France International Energy Agency (IEA).

Pan, Z., Segal, M., Arritt, R.W., Takle, E.S., 2004. On the potential change in solar radiation over the US due to increases of atmospheric greenhouse gases. Renew. Energy 29 (11), 1923e1928.

Raleigh, C., Urdal, H., 2007. Climate change, environmental degradation and armed conflict. Polit. Geogr. 26 (6), 674e694.

Riahi, K., Grübler, A., Nakicenovic, N., 2007. Scenarios of long-term socio-economic and environmental development under climate stabilization. Technol. Forecast. Soc. Change 74 (7), 887e935.

Rinaldi, S.M., Peerenboom, J.P., Kelly, T.K., 2001. Identifying, understanding, and analyzing critical infrastructure interdependencies. IEEE Control Syst. 21 (6), 11e25.

Rose, A., Lim, D., 2002. Business interruption losses from natural hazards: con- ceptual and methodological issues in the case of the Northridge earthquake. Glob. Environ. Change Part B Environ. Hazards 4 (1), 1e14.

Rosenzweig, C., Solecki, W., Hammer, S.A., Mehrotra, S., 2010. Cities lead the way in climate-change action. Nature 467 (7318), 909e911.

Rosenzweig, C., Solecki, W.D., Blake, R., Bowman, M., Faris, C., Gornitz, V., Horton, R., Jacob, K., LeBlanc, A., Leichenko, R., Linkin, M., 2011. Developing coastal adap- tation to climate change in the New York City infrastructure-shed: process,

A. Beheshtian et al. / Journal of Cleaner Production 174 (2018) 1299e1313 1313

approach, tools, and strategies. Clim. Change 106 (1), 93e127. Rosenzweig, C., Solecki, W.D., Cox, J., Hodges, S., Parshall, L., Lynn, B., Goldberg, R.,

Gaffin, S., Slosberg, R.B., Savio, P., Watson, M., 2009. Mitigating New York City's heat island: integrating stakeholder perspectives and scientific evaluation. Bull. Am. Meteorol. Soc. 90 (9), 1297e1312.

Rosenzweig, C., Solecki, W.D., Parshall, L., Chopping, M., Pope, G., Goldberg, R., 2005. Characterizing the urban heat island in current and future climates in New Jersey. Glob. Environ. Change Part B Environ. Hazards 6 (1), 51e62.

Sayyady, F., Eksioglu, S.D., 2010. Optimizing the use of public transit system during no-notice evacuation of urban areas. Comput. Ind. Eng. 59 (4), 488e495.

Sbayti, H., Mahmassani, H., 2006. Optimal scheduling of evacuation operations. Transp. Res. Rec. J. Transp. Res. Board (1964), 238e246.

Schaeffer, R., Szklo, A.S., de Lucena, A.F.P., Borba, B.S.M.C., Nogueira, L.P.P., Fleming, F.P., Troccoli, A., Harrison, M., Boulahya, M.S., 2012. Energy sector vulnerability to climate change: a review. Energy 38 (1), 1e12.

SIRR, N., 2013. A Stronger, More Resilient New York. Tech. Rep. (The City of New York, 2013).

Smith, S.J., Wigley, T.M.L., 2006. Multi-gas forcing stabilization with Minicam. En- ergy J. 373e391.

Solecki, W.D., Rosenzweig, C.Y.N.T.H.I.A., Pope, G., Chopping, M.A.R.K., Goldberg, R.I.C.H.A.R.D., Polissar, A.L.E.X., 2004. Urban heat island and climate change: an assessment of interacting and possible adaptations in the Camden, New Jersey Region. Accessible le, 13.

Sweda, T., Klabjan, D., 2011. September. An agent-based decision support system for electric vehicle charging infrastructure deployment. In: Vehicle Power and Propulsion Conference (VPPC), 2011 IEEE. IEEE, pp. 1e5.

Tol, R.S., 2010. The economic impact of climate change. Perspekt. Wirtsch. 11 (s1), 13e37.

Van Vliet, M.T., Yearsley, J.R., Ludwig, F., V€ogele, S., Lettenmaier, D.P., Kabat, P., 2012.

Vulnerability of US and European electricity supply to climate change. Nat. Clim. Change 2 (9), 676e681.

Van Vuuren, D.P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt, G.C., Kram, T., Krey, V., Lamarque, J.F., Masui, T., 2011. The representative concentration pathways: an overview. Clim. change 109 (1e2), 5.

Wang, M., 2008. The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) Model: Version 1.5. Center for Transportation Research, Argonne National Laboratory.

Webster, P.J., Holland, G.J., Curry, J.A., Chang, H.R., 2005. Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309 (5742), 1844e1846.

White, P., Golden, J.S., Biligiri, K.P., Kaloush, K., 2010. Modeling climate change impacts of pavement production and construction. Resour. Conserv. Recycl. 54 (11), 776e782.

Whitmarsh, L., 2009. What's in a name? Commonalities and differences in public understanding of “climate change” and “global warming”. Public Underst. Sci. 18 (4), 401e420.

Wise, M., Calvin, K., Thomson, A., Clarke, L., Bond-Lamberty, B., Sands, R., Smith, S.J., Janetos, A., Edmonds, J., 2009. Implications of limiting CO2 concentrations for land use and energy. Science 324 (5931), 1183e1186.

Zamuda, C., Mignone, B., Bilello, D., Hallett, K.C., Lee, C., Macknick, J., Newmark, R., Steinberg, D., 2013. US Energy Sector Vulnerabilities to Climate Change and Extreme Weather (No. DOE/PI-0013). Department of Energy Washington DC.

Zhang, X.Y., Gong, S.L., Zhao, T.L., Arimoto, R., Wang, Y.Q., Zhou, Z.J., 2003. Sources of Asian dust and role of climate change versus desertification in Asian dust emission. Geophys. Res. Lett. 30 (24).

Zhu, S., Levinson, D., Liu, H.X., Harder, K., 2010. The traffic and behavioral effects of the I-35W Mississippi River bridge collapse. Transp. Res. part A policy Pract. 44 (10), 771e784.