Analyzing regional disparities in dry port benefits

Dry ports play a vital role in port–hinterland settings, serving as inland intermodal terminals that connect seaports and hinterlands by road, rail, or both (Nguyen and Notteboom, 2019; Deshmukh and Song, 2023). Dry ports can also relieve congestion and capacity overflow at seaports by processing time-consuming services, including storage, consolidation, container maintenance and customs clearance, at inland locations away from seaports (Lim and Lee, 2013; Nguyen et al., 2021). Shifting from road-only to road–rail intermodal freight transportation between seaports and inland areas via dry ports can yield environmental advantages, including reduced greenhouse gas (GHG) emissions per ton-km (Roso, 2007; Lättilä et al., 2013; Bask et al., 2014). Recently, the environmental benefits of dry ports have gained increasing attention owing to international and national policies to reduce GHGs, such as achieving alternative zero or near-zero GHG fuels by 2030 in the international maritime shipping industry (IMO, 2023) and achieving carbon neutrality by 2060 in China (IEA, 2020, 2021).

According to Witte et al. (2019), dry port research has evolved from port regionalization (Notteboom and Rodrigue, 2005) to a contextualization stage that emphasizes spatial and institutional aspects (Wilmsmeier et al., 2011). However, the contribution of dry ports to shippers in the hinterland region may vary depending on the characteristics of the dry ports, shippers locations and the inland freight transportation network (Beresford et al., 2012; Nguyen and Notteboom, 2019; Nguyen et al., 2021). Shippers' choice of dry port usage can be affected by transportation costs, time sensitivity, service levels (e.g. rail frequency), environmental regulations and geographical factors (Sdoukopoulos and Boile, 2020). Consequently, dry ports may not be uniform for all shippers, leading to regional disparities in seaport accessibility (Thill and Lim, 2010).

Furthermore, not all dry ports have been effectively utilized. For example, there are more, although China has over 100 dry ports; many are not preferred by shippers due to insufficient rail connections (Beresford et al., 2012). The rail service frequency at dry ports is a key attribute of port–hinterland connectivity (Deshmukh and Song, 2023). Moreover, cargo with higher time values is typically more sensitive to the rail service frequency (Yang et al., 2020).

Dry ports have been widely examined in the literature for their role in reducing transportation costs and GHG emissions (Roso, 2007; Lättilä et al., 2013; Bask et al., 2014). However, research addressing the impacts of carbon taxation within dry port contexts remains limited. Existing studies on carbon tax policies in maritime transportation have predominantly concentrated on seaports and shipping (Wang and Zhu, 2023; Rojon et al., 2021), leaving a significant gap in understanding their implications for inland intermodal infrastructure. This gap prevents a comprehensive assessment of the potential of carbon taxation to influence dry port development and its broader implications for promoting regional spatial equity in freight distribution systems, an objective explicitly emphasized by many national and regional governments (Fried et al., 2024). To address this gap, this study aims to investigate the impact of environmental policies and dry port operations on regional disparities in dry port benefits. Accordingly, this study poses the following research questions:

This study focuses on Qingdao Port and Tianjin Port in the Bohai Sea region of Northern China, two of the country's largest container ports. Although these ports were among the earliest in China to develop inland dry port systems (Beresford et al., 2012), their hinterland regions lag behind southern China in terms of economic development and infrastructure, suggesting that dry ports may yield greater marginal improvements in this context. This makes the region particularly suitable for examining how dry ports affect regional disparities in seaport accessibility. A utility-based accessibility framework is employed to calculate destination port-specific accessibility under two scenarios: (1) road-only and (2) road–rail intermodal transportation using a dry port. The difference in generalized cost measures between the two scenarios is considered to be the benefit of dry ports for the corresponding city. This approach can be applied to identify the weaknesses in several dry port systems and improve the sustainability of the dry port sector by addressing these issues.

The remainder of this paper is organized as follows: Section 2 reviews the literature, Section 3 describes the cost-based accessibility model and data sources, Section 4 presents an empirical study of six provinces and two ports and Section 5 concludes with policy implications and directions for future research.

Accessibility is defined as “the ease of reaching any location from a specific point using a particular transport system” (Dalvi and Martin, 1976) and has been referred to as a critical factor influencing the development of regional transport networks (Taaffe et al., 1973). A widely adopted framework classifies accessibility measures into four distinct categories (Geurs and Wee, 2004). The first of these categories is infrastructure-based, focusing on travel speeds and congestion-related time losses (Linneker and Spence, 1992). The second category is location-based, analyzing the interplay between land use and transport networks, for example, by measuring distances or potential opportunities from origins to destinations (Hansen, 1959). The third category is person-based, which examines how individual travel constraints may impact feasible activities (Recker et al., 2001). The fourth category is grounded in utility theory, interpreting accessibility as the outcome of travelers' or shippers' transport choices (Koenig, 1980).

Many studies have used location-based methods to investigate freight accessibility (Thomas et al., 2003; Thill and Lim, 2010). For example, Thill and Lim (2010) used a gravity-based potential accessibility measure to examine the effect of intermodal freight networks on port accessibility and found that such networks can notably decrease regional accessibility disparities. Freight accessibility involves not only the distance to final markets but also intermediate nodes such as dry ports, which can play an important role in enhancing accessibility (Roso et al., 2009). Researchers have recently begun adopting utility-based measurement methods to capture this complexity within freight networks, incorporating factors such as service frequency or capacity constraints into their models. For instance, discrete-choice frameworks (Handy and Niemeier, 1997; Guzman et al., 2023) have been adopted to measure accessibility in scenarios involving multiple modes of transportation and route selection. For example, Guo and Yang (2018) employed a generalized cost and discrete choice-based model to reflect shippers' route selection process when measuring accessibility. Similarly, Yang et al. (2020) employed a utility theory-based model to assess the freight accessibility between Chongqing and Europe. They concluded that the use of the China–Europe Railway Express enhanced trade accessibility between Chongqing and Europe.

In practice, these accessibility metrics can be used to evaluate the effectiveness of investing in transportation infrastructure. For example, Wan et al. (2014) utilized a data envelopment analysis (DEA) model to analyze the impact of hinterland accessibility on port productivity in the US. Wei et al. (2018) employed the concept of “accessibility” as a metric to evaluate the connectivity level between nodes in inland provinces and ports within the 21st century Maritime Silk Road network, drawing on the principles of closeness centrality from complex network theory. Therefore, by adopting discrete-choice or network-based methods, policymakers or dry port investors can better evaluate how dry port services influence mode selection, transportation times and overall accessibility.

By establishing rail links with seaports, dry ports can help seaports improve service levels, capacity and storage areas (Roso, 2007; Roso and Lumsden, 2010; Lättilä et al., 2013). Consequently, dry ports are a significant component of port–hinterland settings (Sdoukopoulos and Boile, 2020; Jeevan et al., 2023), with their efficiency directly affecting the performance of seaport operations. The location of a dry port is key in defining a seaport's hinterland. Roso et al. (2009) categorized inland nodes into close, mid-range and distant dry ports based on proximity to seaports. However, despite the distance, the efficiency impact of dry ports on hinterland regions varies depending on the type of cargo, time value, operational efficiency, environmental performance and other factors. Integrating these parameters into a methodology remains a challenge (Nguyen and Notteboom, 2019; Sdoukopoulos and Boile, 2020).

From a cost-efficiency perspective, compared to road transportation, rail transportation through dry ports has cost advantages for long-distance and large-volume shipments (Bask et al., 2014). However, effectiveness depends on geographical location and operational efficiency. Inefficient rail operations may cause delays (Rodrigue and Notteboom, 2009). Wan et al. (2014) found that seaports cannot benefit fully unless the rail line is directly connected to the port yard, underscoring the importance of infrastructure planning to operational efficiency. Jensen et al. (2019) developed an intermodal freight transport chain choice model based on freight datasets from France and Sweden. Their findings indicate that direct rail terminal access at either the origin or the destination significantly increases shippers' likelihood of selecting rail-based transport. In practice, the Port of Gothenburg in Sweden has already adopted such a design, with all trains from inland ports arriving at the port yard. In contrast, at the Port of Hamburg, trains first arrive at various terminals outside the main port area and require trucks for shunting operations, an arrangement that negatively impacts the port's efficiency in handling rail cargo (Khaslavskaya et al., 2024). In order to investigate efficiency metrics further, Chang et al. (2019) revealed the technical efficiency of eight dry ports in the Chinese Ningbo Port using DEA. They found that railways enhanced dry port efficiency only when the distance between the hinterlands and seaports exceeded 500 km. These findings highlight that dry port operational efficiency is determined by multiple factors, such as location, infrastructure connectivity, scheduling and cargo characteristics. However, existing studies have yet to integrate these factors into one framework. This gap underscores the need for a comprehensive framework to assess dry port efficiency under different hinterland scenarios.

Environmental sustainability has become an emerging topic in dry port research (Miraj et al., 2021). Many studies discuss the positive impact of dry ports on GHG emission reductions by minimizing truck traffic (Roso, 2007).

In the context of sulfur emission regulations, Henttu and Hilmola (2011) examined how a dry port network could cut transportation costs and improve environmental outcomes. Subsequent research has primarily concentrated on estimating emission reductions by adopting various dry port routes (Lättilä et al., 2013; Pham and Lee, 2019) or switching to clean energy options (Aksoy and Durmusoglu, 2020). In a case study from northern Italy, Carboni and Orsini (2020) developed a simulation-based framework to quantify such environmental benefits. They analyzed the rail connection between the Port of Venice and the Padua dry port, finding that it could reduce CO₂ emissions by approximately 8,000 tons per year.

Dry port network design also affects emissions. Qi et al. (2022) found that the “last mile” link to the seaport, which refers to whether a dry port's railway is directly connected to the seaport yard, significantly influences carbon emissions within the dry port network. They also found that rail departure flexibility and frequency have a significant impact on shippers' choices, thus indirectly impacting carbon emissions. Dry ports also contribute to environmental sustainability through empty container repositioning (ECR). Castrellon et al. (2023) developed a simulation approach to assess the environmental impacts of ECR strategies. According to their study at Eskilstuna Dry Port in Sweden, with full collaboration among container owners, CO₂ emissions from inland ECR could be reduced by up to 32%.

As the environmental benefits of dry ports receive more empirical support, policymakers are looking for ways to expand these advantages across entire logistic corridors. For example, the EU's “Green Corridor” strategy sets carbon reduction targets for participants operating along specific maritime routes. An important part of this plan is to encourage inland ports to form joint performance agreements, thereby promoting shared responsibility for emission reductions (Psaraftis and Panagakos, 2012; Ismail et al., 2024). Some dry ports have already responded to these initiatives. Nuremberg Freight Village in Germany (Oláh et al., 2018), for instance, has adopted electric delivery trucks and achieved environmental certifications. Similarly, the Eskilstuna dry port in Sweden has completed rail electrification projects (Castrellon et al., 2023).

Despite these insights, many dry port-related studies rely on limited or incomplete operational data, which constrains the accuracy of carbon emission estimates. Wei et al. (2018) collated a dataset encompassing trade and distance metrics for 39 Chinese dry ports and constructed a hub-and-spoke network to quantify the impact of external policy factors on these dry ports. However, studies employing time costs and rail service data for dry ports are scarce. Even when measuring the environmental benefits of dry ports, more detailed information on dry ports should be included in the carbon measurement model, such as the authorities' environmental policies, container weight, dry port capacity and frequency of rail operations (Rodrigues et al., 2024; Lättilä et al., 2013). This study addresses this gap by using realistic transport data from dry ports to assess the impact of environmental policies on dry port operations and their benefits to hinterland regions.

Table 1 summarizes the research topics, methodologies and key issues of selected references reviewed in this chapter, illustrating the different perspectives that scholars have adopted in dry port research.

In this study, the freight network offers a road-only route and multiple intermodal transport routes, including scheduled rail lines between dry ports and seaports, as illustrated in Figure 1. In road-only transport, containers are transported directly from the city district to the seaport. When dry ports are used, containers are initially transported to one of the dry ports from the city district and subsequently transferred to the scheduled rail line to the seaport yard.

As shown in Figure 1, the shipper can choose one of n+1 routes (n road–rail via a dry port and one road transport) to transport its cargo to the destination seaport. Utilizing a discrete choice model to estimate accessibility allows one to evaluate shippers' willingness to pay for various transportation routes while considering generalized costs, such as time, environmental and transportation costs.

To facilitate understanding, the flowchart (Figure 2) shows the process from raw data collection to model output. The following sections will elaborate on each step in detail.

Discrete choice models were introduced by Ben-Akiva and Lerman (1985) and have been widely used in transportation studies to model path choices from a discrete number of alternatives that connect path utility and selection likelihood (Yang et al., 2020; Guo and Yang, 2018; Handy and Niemeier, 1997). In the context of accessibility measurement, generalized cost—covering travel time, monetary expenses, service frequency and other inconveniences—plays a pivotal role in shaping route and mode choices. According to Guo and Yang (2018), generalized cost is the central explanatory variable in discrete choice models, representing perceived disutility of travel. A lower generalized cost makes routes or destinations more attractive and thus more accessible. In this research, accessibility is defined as the inverse generalized cost: as generalized cost decreases, reaching a destination becomes easier. This inverse relationship aligns with utility-based theoretical foundations. Assuming that shippers favor routes with higher utility, this study developed a generalized transportation cost model incorporating three key attributes: time, transport (including toll fees) and environmental costs. The generalized transportation cost function for transporting a 25 ton 40 ft container from the city to the port on route k is as follows:

where

• $Tij−roadk$ is the time spent on road transport from city $i$ to port $j$

• $Tij−railk$ is the time spent on rail transport from city $i$ to port $j$ (if used)

• $k$ is the route identifier (k = 0 for road-only; k = 1, ⋅⋅⋅, n for road–rail via dry port k)

• $Dij−roadk$ is the road distance from shipper $i$ to port $j$

• $Dij−railk$ is the rail distance from shipper $i$ to port $j$ (if used)

• $δ1$ is the unit road transport cost per 40 ft container

• $δ2$ is the unit rail transport cost per 40 ft container

• $ε$ is the time value cost

• $Frailk$ is the weekly train departure frequency from dry port k

• $Ctoll$ is the toll fee

• $σroad and σrail$ are the carbon emissions per ton-km on the road and rail, respectively $(tCO2e/ton−km$⁠)

• $H$ denotes the container weight

Subsequently, as Handy and Niemeier (1997) suggested, the utility of route k between city i and seaport j can be computed as follows:

where $MAX{GCij1,GCij2,GCij3,…GCijk}$ is the path with the highest generalized cost value.

The choice probabilities of route k $(Pijk)$ can be calculated using the multinomial logit (MNL) model as follows:

The route with the highest generalized cost has the lowest accessibility from the city to the seaport. Therefore, accessibility from city $i$ to port $j$ (⁠$Aij$⁠) is expressed as follows:

In addition to port accessibility, this study also computed accessibility in the absence of dry ports, where the road-only shortest path route is considered.

The cost difference between the two scenarios can be considered as the cost benefit of using intermodal transport via a dry port.

Road freight transport data are typically acquired via surveys, questionnaires and government sources. However, the actual road freight transport is subject to traffic congestion and road weight limitations. The application programming interface (API) offered by web-based geographic information systems has recently gained popularity in the literature for obtaining more realistic and precise road network data (Guzman et al., 2023).

This study accessed road network data (distance, travel time and tolls) through the API of the Chinese mapping service AMAP location-based services (LBS). It can perform path planning based on real-time traffic conditions and weight restrictions. The routing option selected for this study was AMAP's primary recommendation, which avoids congestion, shortens travel distance and reduces travel time.

For freight route planning, AMAP LBS requires truck specifications. Table 2 lists parameters for a 40 ft container truck. According to the Ministry of Railways of the People’s Republic of China (2013), a 40-ft container's gross weight cannot exceed 25 tons. Table 3 lists the results of AMAP truck route planning.

The study area includes two seaports (Qingdao and Tianjin) and their hinterland regions in China. In 2024, Qingdao Port and Tianjin Port handled 30.87 million and 23.29 million TEUs, respectively, together accounting for approximately 16.3% of the national total (Ministry of Transport of the People's Republic of China, 2025). Based on the annual port reports, six provinces with stable cargo supply connections to these two ports were selected. Four are inland (Henan, Shanxi, Shaanxi and Anhui), and two are coastal provinces (Shandong and Hebei). Fourteen dry ports operating normally were selected within the study area.

1. Road freight rates

Based on the information provided by “China's Container Transport Network ^[1],” the unit distance cost for a 40-ft container on the road was set to 8.50 CNY (≈1.31 USD)/km. Interviews conducted with Chinese freight forwarders and truckers confirmed that most shippers pay for one-way shipments, as they tend to find truckers who do not have to return to the origin location with empty containers. Therefore, the road freight rate for one-way shipments from city districts to seaports was considered in this study.

1. Rail freight unit cost

The rail freight rate was set to 2.70 CNY (≈0.42 USD)/km for a single 40-ft container, according to the China Railway Website ^[2]. Owing to the limited public availability of railroad operation data for dry ports, some data were acquired by contacting relevant government departments and requesting information disclosure. Details of the train services between the seaports and dry ports are listed in Table 4, including the travel distance by rail, train departure frequency and average travel time.

1. Time value

A crucial factor to consider when evaluating transportation options is the time value, which affects total logistic costs (Macharis and Pekin, 2009). Woxenius (2006) defined the time value for transportation using the following terms: transportation time, order time, punctuality and frequency. According to Larranaga et al. (2017), the time value in freight transportation ranges from 0.03 to 2.88 € per ton of freight per hour. Table 5 presents the value of the time estimates for freight transport in Euro per hour/ton from different studies, reflecting significant variations influenced by factors such as freight type and distance.

The adoption of a time value of 40 CNY (≈6.15 USD)/hour/FEU followed the suggestion of Guo and Yang (2018), who investigated the logistics network connecting China and other countries, covering various transportation modes including land, sea and intermodal transport. Figure 3 shows the time values obtained from various studies in ascending order. The time values used in this study were consistent with those used in previous studies.

1. Carbon cost

This study adopted carbon tax as an environmental cost integrated into the overall transportation cost. As Figure 4 shows, carbon prices vary significantly across countries. In line with the goals of the Paris Agreement, effective carbon pricing should range between 40–80 USD/tCO₂e by 2020 and 50–100 USD/tCO₂e by 2030 (Stiglitz et al., 2017). To capture contrasting policy environments, three scenarios are examined in the present study. The first scenario—assuming no carbon tax—reflects the current condition in China, where a formal carbon tax has yet to be implemented. Additionally, scenarios incorporating carbon tax rates of 16 USD/tCO₂e (based on the California Cap-and-Trade model) and 50 USD/tCO₂e (as applied in France) are considered. The 16 USD/tCO₂e scenario simulates countries or regions where carbon pricing is relatively low. In contrast, the 50 USD/tCO₂e scenario conforms to the lower bound of international recommendations for decarbonization, representing countries or regions that implement carbon tax policies to achieve stricter emission reduction targets. These three scenarios help decision-makers understand how transportation choices may evolve under different policy intensities. This research focuses primarily on the impacts of the higher carbon pricing scenario, building on findings by Wang et al. (2020) that a higher carbon tax can effectively incentivize a modal shift from road to rail transport. The unit carbon emissions per ton of cargo transported by road and rail were set to $σroad=47.1×10−6(tCO2e/ton−km)$ and $σrail=11.4×10−6$ (⁠$tCO2e/ton−km$⁠), respectively, as suggested by Li and Zhang (2020).

Using the discrete choice-based model, the accessibility from each city district was calculated separately for each destination seaport under the three carbon tax scenarios. Table 6 presents the summary statistics of the accessibility results. The overall accessibility of Qingdao Port is higher than that of Tianjin Port. The increase in carbon tax evidently leads to higher generalized transportation costs, resulting in a slight decrease in overall accessibility.

In this study, the dry port service area is defined as cities for which shipping cargo via dry port to destination seaport yields the lowest generalized transportation cost. To determine each dry port's service area, all transport routes are evaluated, and each city is allocated to the option with the minimum generalized cost (Appendix 1). A city was not allocated to a dry port if the road-only generalized transportation cost was lower than the cost via a dry port. The service area for each influential dry port is illustrated in different colors in Figure 5. The gray color represents the service area for road-only transportation. Dry port service areas are larger for Qingdao Port than for Tianjin Port because the dry ports connected to Qingdao Port have higher rail service frequencies. Dry ports associated with Qingdao Port attract 70 cities (85.37%), whereas those associated with Tianjin Port attract 36 cities (43.90%).

In Figure 5, the black triangles represent influential dry ports, while the gray triangles indicate the non-influential dry ports based on the generalized transportation cost. Non-influential dry ports often have low rail frequency, so shippers prioritize time costs. For cargo to Qingdao Port, the Dezhou dry port occupies the largest service area, benefiting from frequent rail services (14 times per week) and relative proximity to cities in the north. The Liaocheng dry port occupies the next largest service area and extends a great distance into the inland areas in the west. Notably, four cities (Linfen, Changzhi, Lüliang and Jinzhong) in Shanxi Province choose the Liaocheng dry port (20 times per week) in Shandong Province rather than the nearby Houma dry port (a weekly rail). Xi'an dry port, despite its central location, attracts only six cities—its relatively low rail frequency to Qingdao Port limits its attractiveness.

For container cargo heading to Tianjin Port, the Taiyuan dry port occupies the largest service area because of its close proximity to northern cities and its high rail service frequency (three times per week). By contrast, the Xi'an (once a week), Zhengzhou (once a week) and Puyang (twice a week) dry ports are not selected by any city because of their low rail service frequency to Tianjin Port (once a week).

In summary, dry port service areas are more extensive when rail service frequencies are high. Although distance to a seaport remains an important factor, frequent and reliable rail connections may outweigh the advantages of closer proximity. These suggest that policymakers should invest in both rail infrastructure and service frequency at major dry ports to expand their service areas.

Figure 6 illustrates accessibility to each seaport under a 16 USD/tCO₂e carbon tax in two scenarios: (1) road-only transportation and (2) road–rail intermodal transportation via dry ports. Cities closer to seaports have higher accessibility. When comparing the accessibility maps of the two scenarios, districts with improved accessibility due to dry ports are either located far away from the seaports or close to dry ports. For container cargo to Qingdao Port, half the cities in Shaanxi Province, distant from the seaport, have improved accessibility through road–rail intermodal transportation. Hinterland areas far from the seaport achieve utility through dry ports, as longer distances result in lower variable costs (Macharis and Pekin, 2009). Simultaneously, border cities of Shandong, such as Dezhou, Liaocheng, Heze and Zaozhuang, have higher accessibility than surrounding areas because of proximity to dry ports.

Compared to Qingdao Port, the accessibility improvements to Tianjin Port are less significant. This may be due to the low operational efficiency of dry ports serving Tianjin Port. Additional analysis under 0 and 50 USD/tCO₂e carbon taxes shows that higher tax rates uniformly reduce overall accessibility but do not significantly alter the spatial distribution pattern. Therefore, they are not discussed here.

Overall, the spatial patterns demonstrate that dry ports can increase accessibility for those cities located far from seaports and those located near dry ports with frequent rail services. Although rising carbon taxes raise total transport costs, they do not significantly change regional accessibility patterns, suggesting that infrastructural factors like rail service frequency remain crucial in influencing logistics choices.

The economic and environmental benefits of dry ports can be identified by calculating the difference in the generalized cost between road-only and intermodal (via dry ports) transportation (Equation 6) under different carbon tax policies. The top 10 cities with cost reduction owing to dry ports are in Appendix 2. Figure 7 illustrates the spatial distribution of the cost reductions by city: white indicates areas where road-only transport remains more economical, while darker shades represent higher cost reductions via dry ports. Table 7 presents the number of cities by cost reduction range under carbon tax prices of 0 and 50 USD/tCO₂e.

Without a carbon tax, four cities around the Xi'an dry port farthest from Qingdao Port see the largest cost reduction (3,501–4,100 CNY/FEU) owing to dry ports. Most cities in Henan, four cities in Shanxi, and those near dry ports on Shandong's western boundary follow (2,501–3,500 CNY/FEU). For containers to Tianjin Port, the greatest cost reductions (1,001–1,286 CNY/FEU) are seen in cities near the Taiyuan dry port, which provide a relatively high rail service frequency (three times per week). Subsequently, cities near the Handan, Xingtai and Shijiazhuang dry ports in Hebei Province experienced cost reductions of 701–1,000 CNY/FEU.

When a carbon tax of 50 USD/tCO₂e is implemented and the destination is Qingdao Port, more cities achieve cost reductions of 2,501–3,500 and 3,501–4,100 (CNY/FEU). A similar pattern appears for Tianjin Port, with more cities reducing costs by 701–1,000 and 1,001–1,286 CNY/FEU.

Hence, cities located far from seaports can gain significant cost benefits when using dry ports, and the benefits grow stronger with rising carbon taxes. Therefore, policymakers should combine carbon taxes with strategic investments in dry-port infrastructure and rail service improvements, thereby encouraging intermodal transportation and reducing carbon dioxide emissions throughout inland areas.

Sensitivity analysis was conducted to examine the influence of the rail service frequency on accessibility (⁠$Aij$ in Equation 4). Three cities at different distances from Qingdao Port were selected for the analysis. The results are presented in Figure 8. The solid and dotted lines represent different values of time for container cargo: 40 CNY (≈6.15 USD)/hour/FEU and 80 CNY (≈12.30 USD)/hour/FEU, respectively. In general, accessibility is higher if the distance to the seaport is shorter and the time value is lower. With an increase in the rail service frequency, accessibility generally follows a logarithmic growth pattern. The accessibility change is relatively less sensitive to the rail service frequency change if the time value is higher or the distance between the dry port and seaport is shorter. This finding implies that intermodal transportation using dry ports is more attractive for shipping low-value cargo over long distances.

It is important to highlight that the green solid line in Figure 8 follows a different pattern compared to the other five lines. Specifically, this green solid line shows a rapid increase in accessibility when the rail frequency is below 2.5 trains per week. This implies that shippers will still choose dry ports even when the rail service frequency is low. As the frequency increases to between 2.5 and 7.5 trains per week, the growth in accessibility slows down, and beyond 7.5, it becomes relatively flat. In this case, it would be practical to provide railway services approximately five times per week. In contrast, the other five cases remain nearly unchanged at lower frequencies and only show a clear increase in accessibility after reaching a certain threshold. Beyond this threshold, the benefits from further frequency increases become smaller.

In summary, these results indicate that increasing rail service frequency improves accessibility, though the benefits diminish beyond an optimal threshold. Practically, operators and policymakers should identify and maintain this optimal frequency to balance the efficiency of resource utilization. This consideration is particularly important for transporting low-value cargo over longer distances, where the benefits of intermodal transport are clearer.

Another sensitivity analysis was conducted to investigate how the change in accessibility (⁠$Aij$ in Equation 4) changes as the proportion of carbon tax in total costs varies. Figure 9 shows that accessibility decreases as the proportion of carbon tax increases. However, the rate of decline slows after a certain threshold. Notably, the impact of the carbon tax on accessibility also depends on the city's location. Cities close to seaports experience a sharper reduction in accessibility because their baseline transportation costs are lower. As a result, when the carbon tax accounts for a larger share of total costs, it has a stronger negative impact on results that originally had higher accessibility.

In summary, policymakers should recognize that a uniform carbon tax increase could have an uneven impact, particularly creating challenges for short-distance freight operators. This insight suggests the need for more balanced carbon pricing strategies. For example, policymakers might consider incentives for using cleaner vehicles or technologies, particularly on shorter routes.

This section examines how cargo time value affects the probability of choosing a dry port. Figure 10 shows a distinct stepped declining pattern in the probability of using a dry port as the time value increases. This pattern is attributed to the longer transit time required for intermodal transport, which increases the time cost of the cargo and reduces the probability of using dry ports. Specifically, in the Qingdao Port scenario, as the cargo's time value increases from $1.54 to $15.38 per hour, the utilization rate of dry ports drops from about 92% to 62%. When the time value exceeds $24/hour, the competitiveness of dry ports drops to zero. For Tianjin Port, only about 57% of cities choose dry ports, and this proportion stays steady when the time value ranges from $2 and $6.5/hour. However, beyond $6.5/hour, the use of dry ports declines significantly, indicating a decrease in competitive advantage with higher time values.

In summary, the time value of cargo significantly affects the attractiveness of intermodal transport via dry ports. While dry ports remain competitive for low-value cargo, their advantage diminishes as cargo value increases. The main constraint in this situation is that the speed and operational reliability of rail intermodal transport are still insufficient to meet the demands of time-sensitive goods. Consequently, policies could consider railway speed upgrades and operational optimization, thereby maintaining cost advantages for low time value cargo while expanding dry port market coverage for time-sensitive goods.

The percentage of dry port usage in each city was calculated based on the approach described in Section 4.2. The generalized transportation cost to the destination seaport from each city was computed under two scenarios: road-only and road–rail intermodal transportation via dry ports. The city was classified as a dry port service area if the dry port scenario resulted in the minimum generalized transportation cost. The number of cities in the dry port service area was counted and compared with the total number of cities to obtain the percentage of dry ports in use.

Figure 11 shows how this usage percentage shifts with carbon tax for container transport to Qingdao and Tianjin Ports. A substantial carbon tax increment only slightly increases dry port adoption: for Qingdao Port, even if the carbon tax increases from 0 to 615.3 USD/tCO₂e, the percentage of cities using dry ports only increases by approximately 5%. Similarly, in the case of Tianjin Port, there is an increase of approximately 3.7% in the percentage of dry port usage under the same change in carbon tax. Therefore, a modal shift to intermodal transportation using a dry port cannot be expected under the carbon tax scenario of 50 USD/tCO₂e (Figure 6).

In summary, although carbon taxes can alter cost structures, they are insufficient on their own to drive a large-scale modal shift to intermodal solutions. Policymakers should consider other measures, such as improving the operational efficiency and service level of rail networks, to promote dry port usage (Chang et al., 2019).

This study examined the impact of environmental policies and dry port operations on regional disparities in dry port benefits by assessing the utility-based cost accessibility to Qingdao and Tianjin Ports. The cost-based freight accessibility from inland cities to ports was measured using a discrete choice model. The accessibility results of the dry port intermodal transportation scenarios were compared with those of the road freight scenarios, and then the benefits of dry ports were considered.

The main findings of this study are as follows. Compared with the road-only inland freight transportation scenario, the existing dry ports enhance the overall accessibility of container transportation from the inland region to Qingdao and Tianjin Ports, resulting in economic and environmental benefits in terms of reducing generalized transportation costs. However, such benefits are not uniform across the hinterland regions. Cities that experience relatively greater benefits are those near mid-range or distant dry ports that also have relatively frequent rail services for seaports. The implementation of a potential carbon tax policy contributes to further improvements in the benefits for dry ports in these areas.

The results of the sensitivity analysis indicate that accessibility is influenced by the cargo's time value, distance to seaports and rail service frequency. If a dry port offers only low-frequency trains, its cost advantages are limited. Cargo with a high time value tends to be transported by road to avoid delays at dry ports.

The findings of this study yield several implications for port policymakers and dry port developers aiming to enhance dry port networks and reduce regional disparities in seaport accessibility. First, efforts should be made to enhance dry port operations through a collaborative framework among maritime stakeholders such as dry port operators, port authorities, government authorities and inland transport firms. Xi'an Dry Port is a successful case supported by the local government. In 2022, it secured investments in COSCO Shipping and several high-tech companies in the medical and telecommunications sectors, leading to the establishment of a strong collaborative mechanism. Second, undeveloped dry ports may face difficulties in joining an existing port–hinterland network, possibly because maritime stakeholders do not acknowledge the prospect of cooperation or terrestrial stakeholders do not proactively pursue development prospects. To address this, maritime participants must be informed about the logistical and economic advantages offered by well-integrated dry ports, such as reducing congestion and expanding market reach. Land-based entities should actively seek partnerships, highlight their strengths and adapt their operations to align themselves with broader regional development objectives. Third, investments in infrastructure and technological innovation are crucial for dry port developers and operators. Such investments should focus on adopting ecofriendly technologies and enhancing the operational efficiency of dry ports. Finally, the importance of regional development and cooperation between dry ports and their service areas should not be overlooked. This cooperation can significantly enhance regional economic growth by improving port–hinterland connectivity and help position dry ports as key components of regional logistic networks.

In conclusion, this study reveals the benefits and regional disparities of port–hinterland connectivity, paving the way for a more detailed exploration of the complex dynamics between dry port operations and environmental regulations.

This study has some limitations that require consideration. First, the data sources primarily rely on official statistics and existing literature, which may introduce biases stemming from variations in data collection methods. Furthermore, the generalized cost calculations adopt uniform assumptions that do not consider the differences in policy, infrastructure development, dry port throughput and capacity, or the practical challenges associated with shifting transportation modes. These simplified assumptions may lead to discrepancies between the estimated outcomes and real-world conditions. Future studies should verify these findings using more detailed data and cost estimates that reflect real-world conditions. It would also be beneficial to explore the scope of the analysis to better capture regional disparities among dry ports. Future work could also explore the potential of emerging green transportation technologies—such as electric freight trucks and autonomous vehicles—in dry port operations. Additionally, it would be valuable to identify which green technology solutions align best with the varying developmental needs of dry ports across different regional contexts. Such works can examine how these innovations influence modal shifts and reduce carbon emissions, thus contributing to the sustainable development of urban logistics systems.

The authors sincerely appreciate the anonymous reviewers for their insightful and constructive comments.

The supplementary material for this article can be found online