Review of quantitative methods to assess impacts of changing climate and socioeconomic conditions on Arctic transportation systems

Abstract

Rapid climate and socioeconomic changes are transforming Arctic human–earth systems. An integral part of these systems is mobility, which encompasses the transport of humans and goods into, out of, and between Arctic regions. Impacts of climate and socioeconomic drivers on Arctic mobility are heterogenous. Methodologies are needed to quantify these impacts in measures that can be linked with broader socioeconomic systems. This article reviews existing methods and organizes them into a conceptual framework to understand trends and gaps in the literature. We found methods quantifying impacts of a range of climate drivers on most transportation modes present in the Arctic, but few methods focused on socioeconomic drivers. In addition, underrepresented were methods explicitly considering adaptive capacity of transportation systems. We provide insight into the data and relationships relevant to understanding impacts of Arctic change on transportation systems, laying a foundation for future work that investigates how these impacts fit into broader human–arth systems.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13280-023-01853-0.

Introduction

Arctic human-Earth systems have entered a phase of rapid change driven by climate change and globalization (Stephen 2018). The impacts of compounding climate and socioeconomic drivers on Arctic societies and the need to respond and adapt have been recognized on an international level (AMAP 2017a, b, 2018). Impacts on mobility, including the transport of both people and goods, are a pressing concern given the unique challenges of transportation in the Arctic. Such challenges include the remoteness of many Arctic communities necessitating long travel distances to essential services (Berman 2013; Larsen and Fondahl 2015), dependence in certain cases on one or few transportation modes with limited alternatives (Olsen et al. 2021; Tretheway et al. 2021), dependence of many regions on imported fuel and/or food (Szymoniak et al. 2010; Rodnina 2022), geophysical constraints such as permafrost and sea ice (Vincent et al. 2017; Khan et al. 2018), and harsh weather conditions (Barjouei et al. 2020). Understanding climate and socioeconomic impacts on Arctic mobility, and linking them to broader human-Earth systems, is a multi- disciplinary problem requiring data and approaches spanning the natural sciences, social sciences, and economics.

The Arctic is particularly vulnerable to climate change, experiencing warming at a rate of two to four times the global average rate (IPCC 2021; Rantanen et al. 2022). Sea ice thickness and extent is in decline (Kwok 2018) along with glacier extent and terrestrial snow cover (Mudryk et al. 2018; Box et al. 2019). Hydrology is intensifying with increasing precipitation and river discharge (Rawlins et al. 2010), and ground temperatures have risen substantially over the past decade, indicating widespread permafrost degradation (Biskaborn et al. 2019). These climate trends have and will continue to impact travel and the delivery of goods over Arctic land, ice, and water.

Socioeconomic change constitutes another major transformation for the Arctic that has implications for transportation and mobility. Population and economic conditions can impact transportation systems through public and private sector investment as well as labor market dynamics (Brooks and Frost 2012; Khoreva et al. 2018). Many regions have experienced persistent outmigration due to high costs of living and limited employment and education opportunities; other regions are seeing in-migration as resource extraction and other economic activities emerge (Schmidt et al. 2015; AMAP 2018; Hamilton et al. 2018). Historical patterns have varied regionally, with consistent out-migration from much of the Russian Arctic since the beginning of the century and more mixed patterns in North America and other Arctic regions; all Arctic regions, however, have seen recent increases in migration to urban centers (Larsen and Fondahl 2015). Economic activity in much of the Arctic is tied to resource extraction, and thus is vulnerable to changes in global energy demand and fluctuating market prices of oil and gas (AMAP 2017b). Global market dynamics including fuel prices are also driving changes for motorized passenger travel in Arctic communities (van Lanen 2018).

Rapid climate and socioeconomic changes in the Arctic have impacts across multiple scales and sectors from local community resilience to regional and global economic dynamics. Particularly vulnerable to these impacts are transportation and mobility, which already face challenges due to the Arctic’s unique and harsh physical conditions. The transport of both goods and people to, from, and between Arctic regions relies on temporary and seasonal navigation modes as well as permanent transportation infrastructure. Temporary and seasonal mobility includes marine navigation and shipping, winter roads, and informal trail systems for travel by foot, snowmobile, off-road vehicle, and dogsled. Permanent infrastructure includes both paved and unpaved roads, railroads, airports, ports, and pipelines. The relative importance and abundance of these modes varies greatly among Arctic regions (Table S1).

Many Arctic communities are geographically isolated and thus rely on complex transportation systems for inter-community connectivity, access to essential services including healthcare, and delivery of food and supplies (Brooks and Frost 2012). Local travel also allows access to subsistence hunting and gathering areas and can be an important component of cultural identity for indigenous communities (Gearheard et al. 2011). On the regional scale, many Arctic economies rely on extractive industries that require robust transportation systems for supplying raw materials and exporting outputs (McGregor et al. 2008). Other regions rely on tourism, another transportation-heavy industry (Lasserre and Têtu 2015). Arctic transportation systems also play an important role in regional and global trade dynamics and meeting global energy demand through shipping routes and oil and gas exports (PAME 2020).

Planning agencies and national governments have recognized the importance of assessing and addressing climate change impacts on Arctic transportation systems. However, many such efforts have either focused on qualitative relationships between climate stressors and transportation systems (National Research Council 2008) or assessed specific vulnerabilities at individual locations to directly inform policy and engineering solutions (Stockton et al. 2021).

Another way to assess climate impacts on Arctic transportation systems is vulnerability indices, which provide a more integrated approach (Debortoli et al. 2019; Cold et al. 2020). In particular, Debortoli et al. (2019)’s Arctic Climate Change Vulnerability Index considers both the sensitivity of different transportation modes to various climate exposures and the socioeconomic determinants of adaptive capacity to these exposures. While indices highlight important drivers of Arctic transportation performance and can be used to identify vulnerable communities, their relative nature limits how directly they can be linked with other human–earth systems.

To understand the changing dynamics of Arctic transportation from an integrated systems perspective, methods are needed to quantify impacts of both climate and socioeconomic drivers in terms of “real-world” metrics. Such metrics include changes to the volume of goods and/or people transported and costs of damages to infrastructure, among others. These metrics can be used to connect impacts across transportation modes and with other systems and to understand feedbacks and tipping points. To our knowledge, there have been no comprehensive reviews of existing quantitative methods that provide these capabilities. The present paper aims to fill this gap by collecting such methods and organizing them into a high-level framework to identify trends, gaps, and priorities for future work.

We will begin by describing our literature review methodology and the framework used to organize the resulting quantitative methods. We will then summarize the findings of the review and how they fit into the framework. We will end by highlighting gaps in the literature and discussing future opportunities to apply our findings, including linking to broader systems and developing typologies of transportation vulnerability.

Methods

Our methodology consists of two main steps: (1) conducting a systematic literature review of quantitative methods that capture relationships between climate and socioeconomic drivers and Arctic transportation system performance and (2) organizing these methods in a conceptual framework to identify trends, connections, and gaps.

Systematic literature review

We conducted a systematic literature review in January 2022 on Scopus. We used the terms below to search the titles, abstracts, and keywords of peer-reviewed articles written in English and published no earlier than 2000:

(arctic OR “far north” OR [(permafrost OR “sea ice”) AND (Alaska OR Russia OR Greenland OR Canada OR Norway OR Sweden OR Finland)]) AND (transportation OR road OR shipping OR airport OR aviation OR port OR barge OR railroad OR pipeline OR bridge OR mobility OR trail OR snowmobile OR boat) AND (climat* OR resilience OR socioeconomic).

From the 1564 initial results, we selected 36 papers that described methods to quantify impacts of climate and socioeconomic factors on Arctic transportation systems. The selection criteria were that the paper must include at least one quantifiable climate or socioeconomic factor and report at least one quantitative metric describing the impact of the factor(s) on one or more mode of transportation in the Arctic. We selected papers both that described past impacts and that projected future ones. We included one additional paper that did not appear in our search (Streletskiy et al. 2019) because it quantified impacts on roads, railroads, airports, and pipelines but only included general terms such as “infrastructure” in the title, abstract, and keywords.

Some search results included only qualitative descriptions of impacts, but we did not include these in the final selection; we selected only quantitative methods as they can provide more opportunities to link impacts between systems. We did not include papers that focused on unitless vulnerability indices; these indices are appropriate for comparing relative impacts on different communities, but do not provide opportunities to link these impacts with systems outside of transportation. For example, Debortoli et al. (2019)’s Arctic Climate Change Vulnerability Index assigns each community an index value between 0 and 1 representing relative climate change vulnerability of aviation and marine transportation. This method is generalizable and can be used to identify priorities but does not provide more concrete information such as infrastructure damage costs or changes to start and end dates of navigation seasons.

Organizing methods and identifying gaps

Following the literature review, we organized methods to identify trends and gaps in the literature and to determine further research needs. We adopted a framework based on the widely used concepts of exposure, sensitivity, and adaptive capacity (Debortoli et al. 2018) (Fig. 1). In this framework, a transportation system’s exposure consists of the external factors that impact the system’s performance. These factors include temperature, precipitation, permafrost thaw, snow cover, sea ice concentration, and other climate variables as well as socioeconomic drivers including population changes and economic pressures. Sensitivity includes the system’s characteristics that mediate the effect of the exposure on the system. Exposure and sensitivity together determine the potential impact on the system, indicated by quantitative metrics such as season length and cost of infrastructure damages. Lastly, adaptive capacity is the system’s ability to prepare for and/or respond to the exposure factors, resulting in a realized impact that may differ from the potential impact (Brooks 2003).

Fig. 1.

Arctic transportation system performance and resilience framework

Results

We identified 37 papers that use quantitative methods to describe the impacts of climate and socioeconomic factors on transportation modes in the Arctic. Figure 2 shows the distribution of methods across modes and exposure variables, and Fig. 3 summarizes the framework components addressed by methods for each mode. Some methods included multiple modes and/or exposure variables. Marine transportation was the most well represented mode in the search, followed by ice roads, while railroads were the least commonly included. Modes for which we did not find quantitative methods in our search include bridges and ports.

Fig. 2.

Summary of quantitative methods, categorized by mode and exposure. Ice conditions includes sea ice concentration and thickness as well as the thickness and quality of inland ice. Snow conditions include snow cover and depth. Income includes both household income and wage income opportunity measures. Note that some quantitative methods include multiple transportation modes and/or exposure variables and thus are counted multiple times

Fig. 3.

Summary of quantitative methods, categorized by mode and the transportation resilience framework components included in impact calculation. Note that some quantitative methods include multiple transportation modes and thus are counted multiple times

A range of exposure variables were addressed, although most methods only considered climate exposures with just two methods including socioeconomic exposure variables (Fig. 2). Most methods only quantified exposure and potential impact or exposure, sensitivity, and potential impact, while few methods considered adaptive capacity (Fig. 3). The sections below describe each mode in the Arctic context and synthesize the framework components (Fig. 1) addressed by the corresponding methods. These include exposure and impact for each mode, as well as sensitivity and adaptive capacity of modes for which we identified methods addressing these components (Fig. 3). Refer to Table S2 for a summary of these components for each method.

Marine transportation

Marine navigation is a prevalent seasonal mode of both freight and passenger Arctic transportation. Marine travel is central to local mobility, subsistence activities, and tourism in coastal areas as well as local, regional, and global trade (AMAP 2017b). The opening and closing dates of marine routes each year depends primarily on sea ice concentration conditions, with vessels’ ice breaking abilities determining conditions needed for safe navigation (Transport Canada 2018). For small passenger boats, weather conditions such as wind and visibility also impact day-to-day accessibility (Rolph et al. 2018). Marine transportation was the best represented mode in our search, accounting for 19 quantitative methods. Of these methods, 8 considered only the system’s exposure and potential impact while 11 also incorporated sensitivity; none explicitly included adaptive capacity in their quantitative analyses.

Most methods focused on trans-Arctic shipping routes including the Northern Sea Route, the Northwest Passage, and the possible future Transpolar Sea Route. Some methods addressed shipping within local regions including parts of the Canadian arctic (Andrews et al. 2017, 2018) and Prudhoe Bay, Alaska (Drobot et al. 2009). Two methods addressed marine transportation in the community context, one focusing on resupply operations and tourism (Mudryk et al. 2021) and the other addressing marine travel for subsistence hunting (Rolph et al. 2018).

Exposure and sensitivity

For all methods assessing impacts on commercial shipping, exposure was quantified by measures of sea ice conditions; Drobot et al. (2009) additionally considered freezing degree days derived from air temperature. Sea ice conditions include average ice thickness within a given spatial extent, as well as ice concentration, which is the area of ice within a given spatial extent relative to the total sea surface area. Thresholds determined whether a certain area was navigable at a given time depending on sea ice conditions. For example, Khon et al. (2017) divided the Northern Sea Route region into grid cells and considered a cell to be “ice-free,” or navigable by ship, as long as the sea ice concentration in the cell was under 15%. Characteristics of shipping vessels are factors in sensitivity for marine shipping since vessels’ ice breaking abilities determine the sea ice conditions through which they can navigate. Many methods considered these differences, adopting ice thickness and concentration thresholds specific to different vessel classes. Mudryk et al. (2021) used vessel characteristics to explicitly address community impacts by including an analysis of a vessel class representing ships used for resupply and tourism.

A common standard adopted by marine shipping methods is the Arctic Ice Regime Shipping System (AIRSS), developed by Transport Canada (Transport Canada 2018), which considers both exposure and sensitivity. AIRSS uses a region’s ice conditions, along with a vessel classification system that accounts for a vessel’s strength, power, and displacement, to calculate the “Ice Numeral” index for a given vessel class in the region. A positive Ice Numeral represents safe navigation while a negative Ice Numeral indicates that the region is not navigable for the ship.

One method addressed marine travel for local subsistence activities rather than commercial shipping (Rolph et al. 2018). Exposure variables included wind speed and sea ice concentration. The authors adopted both a sea ice concentration threshold to determine safe open water conditions and a wind speed threshold to define conditions for safe hunting by boat.

Impact

Most methods focusing on major trade routes and trans-Arctic commercial shipping used AIRSS and other sea ice thresholds to project the length and timing of navigable seasons for shipping routes under future emissions scenarios. These methods consistently projected sea ice decline and corresponding lengthening of shipping seasons as well as opening of new routes. While exact opening and closing dates varied, methods agreed that sea ice conditions under RCP4.5 and RCP8.5 will allow passage along the Northern Sea Route for both open water and ice breaking ships for multiple months each year by mid-century. While the Northwest Passage will be less accessible to open water ships, the probability of passage for these ships will increase and the season for ice breaking ships will extend substantially, including possible year-round accessibility by the end of the century under RCP8.5 (Oh et al. 2017). Some methods also indicated an increasing feasibility of passage along the Transpolar Sea Route by mid- and late- century (Smith and Stephenson 2013; Aksenov et al. 2017; Wei et al. 2020). Projected increases in total accessible Arctic marine area are in agreement with these route accessibility trends (Stephenson et al. 2011).

While most methods addressed major trade routes, some quantified community impacts. One method applied the shipping season length approach to local resupply operations, finding that the marine resupply seasons will extend for coastal Canadian Arctic communities under future warming scenarios. Another method focused on subsistence travel, quantifying historical climate impacts with indicators of the length and timing of the open water season and the number of days with safe wind conditions for hunting by boat (Rolph et al. 2018). They found that while the open water season has increased due to earlier ice break-up and later freeze-up, this increase in marine accessibility has been limited by an increase in the number of days with unsafe wind conditions.

Inland waterways

Arctic inland waterway shipping is accessible seasonally depending on ice conditions and river water levels and provides an essential link between marine routes and inland areas (Scheepers et al. 2018; Arkhipov et al. 2020). We identified two quantitative methods assessing climate change impacts on inland waterway transportation systems in the Arctic (Zheng and Kim 2017; Scheepers et al. 2018). Both quantified impacts of changing water level and streamflow on transportation along Canada’s Mackenzie River.

Exposure and sensitivity

Both methods used projected hydrological variables to represent future exposure, and inferred impacts on inland waterway shipping using operability thresholds. The main variable for Scheepers et al. (2018) was water level; the authors assumed that water levels of 3, 4, and 5 m would act as thresholds for ships carrying various loads. Zheng and Kim (2017) focused on stream flow, assuming that a minimum of 6,000 cubic meters per second was required for barge operation and that various stream flow thresholds would also trigger reduced travel speeds. They considered economic sensitivity factors including the cost of delivery delays and failures associated with stream flows below these thresholds.

Impact

The methods found that future changes in the Mackenzie River’s hydrology will impact both the length and timing of the river’s shipping season. Both methods projected that the most favorable conditions for shipping (high water levels and high stream flow) will shift earlier in the season. Scheepers et al. (2018) additionally projected a decrease in the total number of days per season above operability thresholds, including a 16% and 43% decrease in days with a water level over 5 m under RCP4.5 and RCP8.5, respectively. Zheng and Kim (2017) combined stream flow thresholds with assumptions about logistical costs of shipping to infer economic impacts of the shifting hydrology, which include added costs associated with delays, slower travel speeds, rescheduling, and failed deliveries during low-flow periods.

Adaptive capacity

Zheng and Kim (2017)’s economic model of shipping costs on the Mackenzie River was used to infer the most economically efficient shipping schedule given shifting future streamflow trends. This “optimal” schedule involves rescheduling deliveries planned for late in the season to earlier periods, resulting in an increased peak freight volume in July. While the authors state that they do not intend their results to directly inform specific future schedules, they demonstrate that a general trend toward increasing freight deliveries early in the season and decreasing deliveries late in the season could mitigate increased costs associated with climate change impacts. Whether shipping companies can adapt to climate impacts in this way could depend on the nature of goods being shipped and logistical challenges such as availability of equipment and personnel earlier in the season.

Winter (snow and ice) roads

Winter road systems, available only during the winter and rebuilt each year, are another important seasonal transportation mode in the Arctic. They consist of ice roads constructed over a range of surfaces including frozen ground, compacted snow, and frozen rivers, lakes, and coastal sea ice. Ice roads stretch over 10,000 km in Canada alone and account for two-thirds of the road network in Russia’s Chukotka region (AMAP 2017b; Barrette 2018). They transport large freight loads of materials and equipment to industrial operations and deliver food and supplies to remote communities (Dimayuga et al. 2021). Some ice roads also play an important role in local passenger travel needs during the winter season (Hori et al. 2018b). Ice road opening and closing dates are largely dependent on surface temperature and, for ice roads built over frozen water bodies, sufficient ice thickness (Mullan et al. 2017; Hori et al. 2018a); thus, major climate change concerns for winter roads include shortening season lengths and declining ice thickness (Prowse et al. 2009). We identified six methods quantifying impacts of climate change on ice road performance and vulnerability in the Arctic.

Exposure and sensitivity

Climate exposure metrics depended on the type of ice road assessed by each method. Methods addressing only ice road segments over frozen water bodies focused on ice thickness (Mullan et al. 2017, 2021; Kiani et al. 2018), while methods considering winter road systems built over both frozen water bodies and land considered additional metrics such as surface temperature, soil temperature, and snow depth (Stephenson et al. 2011; Hori et al. 2018a; Gädeke et al. 2021). Most methods used thresholds to determine the viability of ice road construction or travel given these indicators. For example, Mullan et al. (2021) considered a day to be navigable for heavy-haul vehicles on Canada’s Tibbitt to Contwoyto Winter Road if the ice thickness was above 107 cm. Winter roads were considered suitable by Stephenson et al. (2011) given an air temperature of below 30 °F and either a snow depth of at least 20 cm or an ice thickness of at least 22.4 cm. A requirement for ice road construction in the method developed by Gädeke et al. (2021) was an air temperature below -28 °C, and Hori et al. (2018a) compared freezing degree days calculated from air temperature projections to a 380 freezing degree day threshold for a range of winter roads. Kiani et al. (2018) did not adopt a threshold approach, but rather used historical data to examine the relationships between air temperature and ice thickness and ice road performance.

Two methods incorporated sensitivity in addition to exposure. One described impacts of ice thickness on different vehicle types, adopting a threshold of at least 107 cm for heavy vehicles but only 70 cm for light vehicles (Mullan et al. 2017). Another method considered a range of soil types on which ice roads are built, using different soil temperature thresholds to determine viable construction conditions given the soil type in a region (Gädeke et al. 2021).

Impact

Most methods quantified climate impacts on the timing and length of winter road construction and operational seasons. Results indicated declines in historical and future projected seasons, with season opening dates particularly impacted. Kiani et al. (2018) observed a historical negative trend in operating season length for Finland’s Oulu-Hailuoto ice road driven mainly by the season opening date shifting later. Under future warming scenarios, Mullan et al. (2017) and Mullan et al. (2021) projected declining season lengths and particularly rapid opening date delays for Canada’s Tibbitt to Contwoyto Winter Road. While these methods focused on ice thickness, temperature conditions during winter road preconditioning and construction periods also impact opening dates and season lengths; warming during these periods will also reduce future winter road viability (Hori et al. 2018a). Similarly, warming-driven declines in the number of days per season suitable for ice road construction are projected to threaten winter road viability across the entire Arctic (Gädeke et al. 2021).

One method measured impacts on ice roads spatially rather than temporally, calculating the total pan-Arctic land area suitable for winter road construction and extrapolating overall land accessibility proxied by travel time to the nearest settlement (Stephenson et al. 2011).

Adaptive capacity

One winter road methodology explicitly addressed adaptive capacity (Mullan et al. 2021). The potential impact of season length was used to determine the realized impact of the economic viability of operating the ice road each year. Given the season length, viability depended on the whether a flexible scheduling system could be implemented. Flexible scheduling allows for increasing daily loads on the road to meet the total freight demand within a shorter season. A season was determined to be viable given more than 45 navigable days in the presence of flexible scheduling, but more than 50 days were required otherwise.

Flexible scheduling is an important adaptive capacity consideration for ice roads, particularly those whole economic viability depends on required volumes of freight transported. In the absence of flexible scheduling, costlier alternate modes such as air freight may be needed to supplement shortening ice road seasons (McCartan and Kent 2011). For winter roads meeting community mobility and resupply needs that are less temporally flexible, adapting to winter road declines may require investing in permanent roads and railroads or switching to inland waterway transport when possible (Stephenson et al. 2011).

Trails

Arctic trails encompass the informal and often seasonal transportation routes used for both motorized and non-motorized local travel over land and sea ice. These include snowmobile, off-road vehicle, and dogsled routes as well as footpaths. Trails can be an essential component of Arctic inter-community connectivity (Steiro et al. 2021); they also provide access to subsistence hunting areas and cultural sites for indigenous communities (Ford et al. 2019). Trail accessibility depends on weather conditions and may also require specific snow and/or ice conditions (Ford et al. 2019; Cold et al. 2020). Five of the quantitative methods identified in our search assessed climate impacts on Arctic trail systems.

Exposure and sensitivity

Most trail methods focused on off-road vehicle and snowmobile travel, and climate exposure metrics differed for overland and sea ice trails. For overland trails, snow depth and soil temperature thresholds determined navigability in conjunction with soil type as a sensitivity metric (Gädeke et al. 2021). Exposure metrics for sea ice trails included ice concentration and thickness, as well as over-ice snow cover (Ford et al. 2019; Cooley et al. 2020; Steiro et al. 2021). One method also considered weather conditions relevant to both overland and sea ice travel, including temperature, precipitation, wind, and visibility (Ford et al. 2019). This method incorporated user risk tolerance as a sensitivity metric that determined the thresholds of each exposure variable past which travel was not possible. Finally, since the accessibility of off-road vehicle use depends on individuals’ ability to purchase fuel and equipment, one method considered a socioeconomic metric of wage income opportunity (Ehrich et al. 2019).

Impact

Historical and future climate impacts on sea ice trail accessibility yielded consistently negative trends, while impacts on overland trails were more mixed. Steiro et al. (2021) found that sea ice travel by snowmobile between communities in a Greenland fjord has become slower and more unpredictable in recent years due to changes in snow and ice conditions, and particularly a trend towards earlier on-ice snowmelt. A trend of declining sea ice trail access was also detected in Inuit Canadian communities, where the number of annual trail access days decreased due to later freeze-ups, earlier break-ups, and decreasing ice concentration (Ford et al. 2019). Cooley et al. (2020) projected that coastal sea ice trail accessibility in Arctic Canada will continue to decline under future emissions scenarios due to shortening shorefast ice seasons and earlier ice break-ups.

Access to overland trails, however, has changed minimally over the past two decades for Canadian Inuit communities, with slight increases for some communities and decreases for others due to changing local weather conditions such as wind speeds and visibility (Ford et al. 2019). However, wider pan-Arctic trends in snow and soil temperature conditions could pose challenges for overland trails under future emissions scenarios; across the Arctic, the average number of days with conditions allowing overland travel by offroad vehicle is projected to decline by up to 26% under RCP6.0 and 40% under RCP8.5 by the end of the century (Gädeke et al. 2021).

Wage income opportunity was found to be positively related with the extent of offroad vehicle use in several settlements throughout Arctic Alaska, Canada, and Russia (Ehrich et al. 2019), suggesting that access to wages in a mix subsistence-wage economy allows for greater access to the transportation modes central to subsistence hunting. However, none of the methods in our search projected future socioeconomic impacts on Arctic trail use.

Permanent infrastructure

Year-round Arctic transportation modes include railroads, roads, aviation, and pipelines. Road networks are underdeveloped in much of the Arctic (AMAP 2017b, 2018), necessitating reliance on the seasonal transportation modes described above and on aviation. In many cases, aviation is the only mode available year-round for inter-community travel and resupply and thus plays an important role in community resilience (Widener et al. 2017; Dimayuga et al. 2021). Air freight can also serve as an alternative to ice roads for supplying industrial operations (McCartan and Kent 2011). Pipelines are an important freight transportation mode, exporting oil and gas from Arctic drilling operations. While roads, railroads, airports, and pipelines are generally operational year-round, weather conditions can impact day-to-day operations and climate change pressures including permafrost thaw and sea level rise can affect long-term costs and viability (Bardal 2017; Melvin et al. 2017; Streletskiy et al. 2019). Six quantitative methods examined climate change impacts on permanent Arctic infrastructure including roads, railroads, airports, and pipelines.

Exposure and sensitivity

The most common exposure variable in these methods was permafrost thaw (Melvin et al. 2017; Shojae Ghias et al. 2017; Porfiriev et al. 2019; Streletskiy et al. 2019; Bartsch et al. 2021). Measures of permafrost thaw included active layer thickness (ALT), ground ice content (GIC), and ground temperature (Melvin et al. 2017) as well as ground subsidence, which is a function of ALT and GIC (Porfiriev et al. 2019; Streletskiy et al. 2019). Streletskiy et al. (2019) also included change in foundation bearing capacity as a permafrost thaw metric for buildings including railway stations and airports; this was inferred from ALT, ground temperature, and assumptions about foundation characteristics. Methods used thresholds of permafrost thaw metrics, such as ground subsidence of 0.1 m, to determine when transport infrastructure would need to be replaced (Melvin et al. 2017; Streletskiy et al. 2019) or repaired (Porfiriev et al. 2019).

Beyond permafrost thaw, other climate variables that were assumed to trigger infrastructure replacement and repair included flooding, precipitation, and freeze–thaw dynamics (Melvin et al. 2017). Exposure variables affecting day-to-day road operations, rather than long-term replacement and repair needs, included temperature, precipitation, wind, and visibility (Bardal 2017). Finally, one method addressed the impact of socioeconomic exposure variables, including population and median income, on communities’ airport infrastructure (Widener et al. 2017).

One method incorporated sensitivity by assessing permafrost thaw impacts under a range of future infrastructure development scenarios, ranging from the maintenance of existing roads to both moderate and high levels of road network expansion (Porfiriev et al. 2019). Another method focusing on day-to-day road operations used vehicle type (passenger versus freight) as a sensitivity factor (Bardal 2017).

Impact

Most methods quantified long-term future climate change impacts by projecting the amount of damage caused to infrastructure up to mid- or late- century. Measures of damage included total costs of damages and maintenance including replacement costs (Melvin et al. 2017; Porfiriev et al. 2019; Streletskiy et al. 2019), the extent of infrastructure affected (Bartsch et al. 2021), and physical indicators of damage to a particular site (Shojae Ghias et al. 2017).

Estimates of future damage costs varied between regions, types of infrastructure, and climate change scenarios. In the Russian Arctic, estimates of mid-century permafrost thaw- related damages to transportation infrastructure under RCP8.5 range from $1.5 billion per year for all infrastructure (Streletskiy et al. 2019) to $223 million per year for only roads (Porfiriev et al. 2019). In Alaska, undiscounted infrastructure damages under RCP8.5 through 2100 range from a minimum of $2 million per year for pipelines to a maximum of $8.3 million per year for roads (Melvin et al. 2017). However, under RCP4.5, Melvin et al. (2017) find that airport damages will outweigh road damages in Alaska. Finally, while Bartsch et al. (2021) did not explicitly estimate damage costs, they did find that over half of pan-Arctic human impact areas, including roads and railways, will be on ground with above freezing mean annual ground temperature by mid-century, which will induce major maintenance and repair costs.

The expected level of future infrastructure development could be a crucial sensitivity variable when estimating permafrost thaw- related costs; Porfiriev et al. (2019) found that annual costs increased by 50% to 100% for moderate to high road network expansion scenarios compared to a scenario requiring only maintenance of existing Russian Arctic roads.

Impacts on day-to-day operations for permanent infrastructure were estimated by one method using traffic volume along a mountain pass in Norway (Bardal 2017). While they found that high wind speeds and low temperatures slightly decreased passenger traffic, overall impacts of weather conditions on both passenger and freight traffic were moderate. This may have been a result of the lack of alternative transport options and therefore the necessity to take on increased risks during adverse weather conditions.

Finally, Widener et al. (2017) used impact metrics of runway length, flight frequency, and proportion of scheduled flights that arrived (flight reliability) to assess socioeconomic impacts on air travel in Arctic communities. The focal communities in this study were “fly-in” communities, or those for which air travel is the only available mode of transportation and resupply for much of the year. They found positive correlations between median household income and both runway length and flight reliability, suggesting that communities with more economic activity and wage income opportunity may also have increased access to essential flight services.

Adaptive capacity

In their assessment of infrastructure damage costs, Melvin et al. (2017) considered the benefits of proactive adaptation, a type of adaptive capacity consisting of engineering approaches to protect infrastructure from damages. They calculated total costs of damages both with and without these adaptive capacity measures. This framework could be integrated into other methods assessing economic damages to permanent infrastructure provided that the extent of damages avoided by proactive adaptation measures can be estimated.

Discussion

Our literature review identified existing methods that measure the impacts of several climate drivers, and some socioeconomic drivers, on Arctic transportation. Many methods incorporated factors determining sensitivity to these drivers; however, few considered adaptive capacity. All methods calculated “real-world” measures of transportation performance and vulnerability. These measures could be used to incorporate transportation system challenges into broader frameworks of pan-Arctic vulnerability and resilience. Below we highlight gaps in the literature including socioeconomic exposures and adaptive capacity. We also discuss future opportunities for linking transportation impacts to broader Arctic human–earth systems and for applying our findings to detailed typologies of transportation vulnerability.

Gaps in the literature

Adaptive capacity

More research is needed to incorporate adaptive capacity into quantitative methods, particularly for trails and marine travel. Adaptive capacity can include the ability to switch modes, routes, and schedules in response to changing transport accessibility as well as opportunities to prevent and mitigate negative impacts. Examples of adaptive capacity measures already taken include developing technologies to manage risks associated with changing climate conditions (Bell et al. 2015), behavioral changes such as route switching and increased use of safety equipment (Ford and Goldhar 2012), and large-scale investments such as rebuilding or improving existing infrastructure (McGregor et al. 2008) and building new permanent infrastructure, partly to replace seasonal modes impacted by climate change (Bennett 2018).

Only three of the 37 methods in our literature review explicitly incorporated adaptive capacity into their methods addressing ice roads (Mullan et al. 2017), inland waterways (Zheng and Kim 2017), and permanent infrastructure (Melvin et al. 2017). While transportation systems facing similar pressures may share potential impacts, realized impacts are likely to be more heterogeneous and location-specific due to varying adaptive capacity constraints. Thus, developing more quantitative methods that integrate adaptive capacity is an important next step towards a holistic understanding of Arctic transportation vulnerability. Examples of important adaptive capacity factors that were not addressed by any of the quantitative methods are provided below.

Regarding trails, income and employment opportunity are important drivers of adaptive capacity to climate change impacts on travel for subsistence activities (Pearce et al. 2010; van Lanen 2018). The ability to switch routes in response to changing trail accessibility depends on access to wages for purchasing equipment and fuel. Additionally, factors influencing risk tolerance, such as investment in training and safety equipment, could increase adaptive capacity for trail users as weather and climate conditions change (Ford et al. 2019).

As to marine travel, many methods quantified future increases in shipping season length, but none considered the adaptive capacity factors influencing whether shipping operations can adjust routes and schedules to take advantage of this increasing accessibility. For example, the economic feasibility of using the Northern Sea Route in addition to or in place of existing global shipping routes depends on competition with existing routes (Zhang et al. 2016; Zeng et al. 2020; Liu et al. 2021) and policy factors such as carbon taxes (Ding et al. 2020). It is important to consider these and other factors when projecting changes to Arctic shipping traffic since declining sea ice by itself is not sufficient to drive large scale Arctic shipping development, particularly for non-destinational traffic (Lasserre 2019).

Socioeconomic exposures

Our literature search also highlights that there are relatively few existing methods to quantify impacts of socioeconomic exposure variables on Arctic transportation systems. Globalization and its related socioeconomic impacts, including migration, resource extraction, political changes, and cultural trends, are a major force of change for Arctic societies (AMAP 2017a, b, 2018; Stephen 2018). For example, as mixed subsistence-wage Arctic communities have adopted motorized transport modes for local travel and subsistence activities, they have become increasingly reliant on imported fuel; their travel is therefore impacted by global economic markets as well as local socioeconomic conditions (van Lanen 2018). Additionally, levels of government investment in port infrastructure can impact the quality and frequency of freight shipping services to communities relying on regular marine resupply (Brooks and Frost 2012). Declining population and economic activity in some Arctic regions can reduce availability of local public transportation services (Olsen et al. 2021); this can in turn hinder economic activity and trigger outmigration, creating feedbacks and highlighting the bidirectional nature of relationships between socioeconomics and mobility. The economic and well-being consequences of Arctic transportation infrastructure development projects further demonstrate these bidirectional relationships (Povoroznyuk et al. 2022).

More methods are needed to quantify impacts of these and other socioeconomic factors as well as compounding socioeconomic and climate exposures. Ultimately, an understanding of the complex interactions between climate and socioeconomic drivers is needed to thoroughly assess Arctic transportation vulnerability. Both types of drivers can contribute to a transportation system’s exposure, sensitivity, and adaptive capacity and can form complex, bidirectional interactions (Debortoli et al. 2018). Thus, a holistic assessment of Arctic transportation vulnerability and resilience must integrate both socioeconomic and climate factors and consider how these factors interact. Such an assessment could provide valuable insight into the most effective ways to mitigate impacts and build capacity for Arctic mobility and broader human-Earth systems.

Future opportunities

Linking to broader systems

The framework presented here (Fig. 1) can be used to identify the most relevant external drivers (“Exposure”) and internal characteristics (“Sensitivity”) and to calculate impacts on transportation system performance, considering the effects of adaptive capacity when applicable. The resulting “real-world” quantitative impact metrics lay the foundation for integrating transportation system impacts within broader Arctic systems including local food and socioeconomic systems, regional industry viability, and global trade, among others. Below we discuss a non-comprehensive selection of examples illustrating how the metrics could be used in future work to connect transportation impacts to these systems.

First, shortening winter road seasons could impact the amount and cost of freight transport. These impacts have been projected in the subarctic, where logging companies in boreal forests could see 5.2% to 11.4% increases in logging costs by the 2080s under various warming scenarios (Kuloglu et al. 2019). Similar assessments could reveal impacts of climate change on the costs of Arctic industrial operations such as mining, as well as on food and fuel costs for Arctic communities relying on winter resupply via ice roads.

In the case of shipping, lengthening open water seasons and emerging accessibility of new transpolar shipping routes is relevant to regional and global trade dynamics and may also impact local employment and costs of goods shipped to Arctic communities (Bennett et al. 2020). Season changes could also affect tourism industries, with implications for local income and employment (Lasserre and Têtu 2015). For Arctic communities relying on trail-accessed subsistence food sources, changing trail accessibility has implications for the proportion of income spent on imported food. Changing travel distances on trails for motorized vehicles could also be linked to income spent on fuel and equipment.

Lastly, costs of damages to permanent transportation infrastructure could be reflected in costs of public transportation as well as costs of food, fuel, and supplies delivered to Arctic communities via roads, railroads, and air. Infrastructure damage may also have implications for regional and global supply chains; for example, costly damage to pipelines could destabilize oil and gas exports (Streletskiy et al. 2019).

Typologies of transportation vulnerability

While the quantitative methods presented here cover a range of climate and socioeconomic impacts on transportation modes, they cannot capture the full extent of challenges. The Arctic has heterogeneous geography, physical conditions, economics, and other factors affecting the ways in which socioeconomic and climate drivers impact transportation systems. While some of these factors can be captured by sensitivity and adaptive capacity, many are unique to specific localities. This poses challenges to gaining a holistic understanding of pan-Arctic mobility trajectories. However, the high-level framework presented here provides insight into the data needed and the relationships relevant to understanding impacts of Arctic change on a range of transportation modes.

Future research could further address variability and uncertainty by applying our framework and insights from our literature review to detailed typologies of Arctic transportation. A set of dimensions could classify Arctic communities and industries according to their transportation needs, options, and challenges; our framework could then identify the most important exposure and sensitivity measures (Fig. 1) to monitor and the associated impacts. Table 1 provides illustrative examples of such typologies. This typology approach could be used to identify needs for additional data in targeted areas, to expand our understanding of Arctic transportation system vulnerability, and to highlight opportunities for adaptive capacity. It could also be used as a roadmap to identify the relevant data and relationships needed for incorporating Arctic transportation dynamics into integrated assessment modeling frameworks. This is an important next step towards improving representations of human-Earth system interactions in the Arctic context.

Table 1.

Non-exhaustive examples of typologies of Arctic transportation, their associated transportation options and needs, and relevant exposures and impacts from our literature review

Transportation typology dimensions			Transportation options and needs	Example exposures and impacts
Geophysical characteristics	Economic activity	Remoteness
- Coastal - Inland with river access	Mixed subsistence-wage economy	Small community without road connection	- Resupply via barge - Subsistence travel on trails	- Changing sea ice and river flow regimes could impact cargo barge seasons (Scheepers et al. 2018; Mudryk et al. 2021) - Changing sea ice regimes and wind conditions could impact marine subsistence hunting accessibility (Rolph et al. 2018)
Inland without river access	Mixed subsistence-wage economy	Small community without road connection	- Resupply via air - Subsistence travel on trails	- Local socioeconomic conditions could impact airport reliability (Widener et al. 2017) - Wage income opportunity could impact access to motorized overland subsistence travel (Ehrich et al. 2019)
Inland without river access	Resource extraction	No road connection	Import and export via ice road	Increasing temperatures and declining snow and ice could shorten ice road seasons (Mullan et al. 2021)
Permafrost		Connected to larger road network	Year-round road access	Permafrost thaw could increase road maintenance costs (Porfiriev et al. 2019)
	Trans-Arctic shipping		Marine vessels with varying icebreaking abilities	Declining sea ice could increase temporal and spatial marine accessibility along major trans-Arctic trade routes (Stephenson et al. 2013; Oh et al. 2017)

Supplementary Information

Below is the link to the electronic supplementary material.

Biographies

Taryn Waite

is a research associate at the Joint Global Change Research Institute. Her research interests include assessing multisector climate change impacts and analyzing net-zero pathways through integrated assessment modeling on both city and national scales.

Meredydd Evans

is a senior staff scientist at the Joint Global Change Research Institute, where she is managing a program on international sustainable energy, including analysis to reduce short-lived forcers, efforts on building energy efficiency, and energy data for policy and clean energy investments, among others. Her work covers both global assessments as well as regional and national research in the Arctic, Russia, China, India, and elsewhere. Her work often bridges multiple scales, such as the impact of local actions on national and global trends.

Nazar Kholod

is an Earth Scientist at the Joint Global Change Research Institute. His current research focuses on energy policy, energy efficiency, energy security, emissions of short-lived climate forcers, and integrated assessment of climate change.

Nina Blahut

is a research associate at the Joint Global Change Research Institute, where her research interests include assessing climate change impacts on infrastructure and modeling decarbonization pathways in the industrial sector.

Joel Rowland

is a scientist at Los Alamos National Laboratory where he researches the role of landsurface dynamics and hydrology in the transport, storage and cycling of sediment, water, and biogeochemical constituents with a focus on river and floodplains, permafrost landscapes, and coastal regions.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest with this work.