2.1. Urban Underground Space Utilization

As China’s urbanization progresses, urban planning has gradually shifted from “incremental expansion” to “optimization of existing resources,” especially in areas with high-intensity development, where the scarcity of space has become a key factor restricting sustainable development [20]. The growing demand for urban transportation has made it difficult for surface transportation infrastructure to meet these needs, leading to problems such as overcrowding, insufficient functional space, and traffic system chaos [21, 22]. While measures such as building highways and elevated roads can ease some traffic pressure, the efficiency of urban commuting continues to decline. Surface space has reached saturation, and the expansion of roads is limited by restrictions and negative landscape impacts. Moreover, in areas with high traffic flow, motor vehicle emissions exacerbate environmental degradation and further reduce air quality.

In this context, the use of underground space (UUS) offers an effective solution to address these challenges. Underground space can alleviate land scarcity, optimize transportation supply, and reduce environmental pollution, making it an important choice for sustainable urban development [23, 24]. Classic cases, such as the underground road networks in Shanghai’s CBD, the underground expressways in Hong Kong, and the underground parking lots in Chongqing’s Jiefangbei CBD, have become important references for underground space development due to their large-scale construction and effectiveness in easing traffic congestion. Chen et al. proposed a multicriteria framework combining economic, social, and underground space factors, offering an in-depth analysis of the current status and development trends of underground space in Chinese cities and advancing research from traditional engineering cases to a more theoretical and systematic approach [25]. Later, Peng et al. and Dong et al. further quantified the impact of economic and social factors on the distribution of underground space, revealing the close relationship between the concentration of underground space and economically developed areas or areas with high population density, providing theoretical support for underground space planning [4, 26].

In recent years, research on underground space has gradually shifted from traditional engineering and technical approaches to interdisciplinary, comprehensive methods. With the advancement of big data and information technology, new data sources like POI data have become important tools for building accurate assessment frameworks. For example, Ma and Peng proposed an evaluation framework based on POI data that analyzed the supply-demand ratio of underground space in 217 streets in Shanghai, revealing the differences in underground space demand between central urban areas and suburbs, providing a quantitative basis for underground space planning [27].

In addition, research on underground space has progressed from economic evaluations to demand quantification, planning optimization, and risk management. Ma and Peng first proposed a method for evaluating the economic benefits of complex underground road networks, laying the groundwork for understanding the value of underground infrastructure [28]. Building on this, Ge et al. quantified the demand for underground space in undeveloped areas, emphasizing the importance of aligning spatial planning with demand distribution [29]. Extending this perspective, Dong et al. analyzed subway-dominated underground spaces using multisource data, stressing the need for planning optimization tailored to different temporal and spatial contexts [30]. Further advancing the field, Deng et al. evaluated underground space suitability by integrating 3D geological modeling with multicriteria decision-making methods, offering technical support for sustainable development [31]. Addressing safety challenges, Zhong, Li, and Jiang explored strategies to enhance resilience against extreme events, providing insights into risk mitigation for underground spaces [32].

In summary, underground urban space has been widely recognized as a critical solution to challenges such as land scarcity, transportation inefficiencies, and environmental degradation. Existing studies have provided valuable insights into the general utilization of UUS, offering interdisciplinary frameworks for spatial planning, demand assessment, and resilience strategies. While much of this research has focused on broader contexts, this study narrows its focus to underground road networks in high-intensity urban areas. By addressing the unique traffic demands and resilience challenges of such environments, this study builds on previous findings to deliver reliable and targeted analyses, offering practical strategies and theoretical support for optimizing the planning and utilization of underground road networks in densely developed cities.

2.2. Transportation Network Resilience

Resilience analysis mainly focuses on the resilience attributes of systems and methods to enhance them, including metrics, evaluation methods, and improvement strategies [33]. The profound impacts of external shocks highlight the urgent need to assess the resilience of various systems. System resilience is usually defined as a multidimensional and dynamic structure, often evaluated by the system’s ability to resist external shocks and restore its function [34]. Resilience is typically quantified by determining whether a system can recover to its predisruption performance or effectiveness after an interference event [35]. In the context of sudden disasters, scholars often divide resilience into different stages—preparation, absorption, adaptation, and recovery—each reflecting the system’s performance during different stages of disruption [36].

In the study of resilience in road transportation systems, many scholars focus on the impact of disaster events (such as floods [37], hurricanes, and earthquakes) on transportation networks. These studies generally explore the ability of transportation networks to recover during extreme disruptions [38, 39]. For example, Lu et al. analyzed the impact of traffic disruption on urban road network resilience through simulations of idealized networks [40]. Although the study provided valuable insights, its simulation setup failed to fully consider the complexity and variability of real-world situations. To address this, Wei et al. developed a transportation system resilience curve using postdisaster actual data, revealing the dynamic changes and key turning points of the system during disasters [41].

This data-driven approach greatly improved the practical significance and applicability of resilience assessments. Zhang et al. proposed a new evaluation method based on a network cascading failure model, capable of dynamically evaluating the resilience of the entire road network across different time and space scales [42]. This method breaks through the limitations of traditional evaluation models, not only improving the accuracy of resilience assessments but also providing a more comprehensive framework for analyzing transportation network resilience under various scenarios.

In summary, the evolution of resilience analysis in road transportation systems reflects a shift from simplified theoretical models to more practical, data-driven approaches. By incorporating real-world data and dynamic modeling techniques, recent studies have significantly improved the accuracy and applicability of resilience evaluations, offering robust tools for enhancing system performance under various disruption scenarios. Building on these advancements, the present study distinguishes itself by introducing a threshold model tailored to the initiation of underground road network construction in high-intensity urban development areas. By integrating a spatiotemporal consumption model with traffic demand analysis and resilience assessment, this study provides a holistic framework for decision-making in urban transportation planning, addressing the unique challenges of densely developed urban environments.

3.1. Supply and Demand Model

3.2. The Road Network Resilience Model

In the face of traffic peaks, large-scale public events, or other situations leading to a surge in traffic demand, a resilient transportation system can more effectively absorb and adapt to these pressures [43]. Its core concept includes two aspects: first, the ability to absorb disturbances, reflected by the degree of degradation of system performance under disturbances; and second, the ability of the system to recover from disturbances, reflected by the speed and extent of system recovery. This helps avoid traffic paralysis. The changes in the state of the road traffic system under disturbances are illustrated in Figure 1. At the moment of disturbance occurrence, the system performance degrades to its lowest level, and the moment when the system performance returns to the expected state, dividing the road traffic system under disturbances into three stages and four states: the predisturbance stage (before time t_e), the disturbance impact stage (from time t_e to t_r), and the disturbance impact elimination stage (after time t_r), the four system states are reliable state, degraded state, recovery state, and postrecovery state. This study mainly focuses on the recovery performance curve after failure.

To effectively reflect the resilience of the road network in high-intensity development areas, this study adopts structural entropy, intermediary degree, and closeness centrality as key indicators to characterize the resilience of the road network.

4.1. Study Area

The Jiangbei new area is located north of the Yangtze River in Nanjing, Jiangsu Province, China. It is a national-level new area comprising the Baqiao Street of Pukou District, Liuhe District, and Qixia District. Serving as a strategic pivot in the hinterland of eastern China, it boasts convenient transportation connections via highways, railways, waterways, and air hubs. Situated at the intersection of the Yangtze River Economic Belt and the eastern coastal economic belt, it serves as a vital hub radiating to the central and western regions of the Yangtze River Delta. It plays a crucial role in connecting the northern part of Nanjing to the central and western regions, contributing significantly to the integration of the Nanjing metropolitan area and the development of the Ninggao and Ningzhenyang urban agglomerations. Covering an area of approximately 2451 square kilometers, it accounts for 8% of the total area of southern Jiangsu Province and 37% of Nanjing’s total area.

As one of the most rapidly urbanizing cities in China, Nanjing has undergone extensive high-intensity development, particularly in the Jiangbei New Area. This district is not only a key area within Nanjing but also a representative example of China’s broader urbanization and industrialization trends. The core area and surrounding regions of the Jiangbei New Area are positioned as the urban center of Nanjing’s Jiangbei New Area, serving as a leading area for independent innovation, a vibrant agglomeration zone, and a multifunctional demonstration area. With high construction density and a concentration of various facilities, it exemplifies the city’s broader strategy of concentrated urban development. Researching this area can provide theoretical basis and guidance for similar high-intensity development areas, as it reflects Nanjing’s strategic emphasis on urban growth and its role as a critical node in regional development. The study area is illustrated in Figure 2.

4.2. Results and Discussion

In the study area, the length of expressways is 3.88 km, main roads is 6.1 km, arterials is 11.1 km, and local roads is 29.11 km. Local roads are the predominant road type in the area, constituting the highest proportion of total road length. Although local roads typically handle lower traffic volumes, they play a crucial role in distributing and dispersing traffic throughout the network. Parameters are set, including the reduction coefficients for different road categories and intersection utilization coefficients as shown in Tables 2 and 3. These coefficients directly affect the final calculation of traffic carrying capacity. Higher reduction coefficients or lower intersection utilization coefficients may imply greater restrictions on traffic flow.

Using the time-space consumption method model, the calculated traffic carrying capacity of the road network in the study area is 48,821 (pcu/h). While theoretical traffic carrying capacity provides an important reference, actual traffic capacity may be lower than the calculated value due to the fluctuation and unpredictability of traffic flow in real-world road networks. If actual traffic volumes approach or exceed the calculated capacity, it may be necessary to consider improvement measures, such as increasing road capacity or optimizing intersection design to enhance the overall traffic carrying capacity of the network.

A current traffic survey conducted on road segments indicates that the traffic volume within the area is 24,987 vehicles per hour (pcu/h). This traffic volume is below the road network’s capacity, suggesting that the road network’s capacity can effectively accommodate the traffic volume and maintain normal road traffic operations. However, an analysis and calculation of various land parcels (mainly residential, commercial, and educational land) within the planned area, as shown in Figure 3, using a traffic demand calculation model, yield a traffic volume of 51,620 vehicles per hour (pcu/h). The projected traffic demand for the planned land use significantly exceeds the road network’s capacity, indicating that it cannot meet the normal operational requirements of the road network.

Figure 4 illustrates the overall trend of road network resilience performance under simulated failures involving the most critical nodes and segments. In both charts, the performance function F (t) is observed over time t. The graphs reveal a pattern where the curves initially decline, then rise, and eventually plateau. This trend reflects a typical sequence in which an initial failure or damage reduces network performance, followed by a recovery or reconstruction phase that gradually restores and stabilizes performance. During the early phase of the curves (approximately between t = 0 and t = 4), a slight decline in performance is evident. This phase likely represents the initial impact of the failure on the road network’s resilience, as the system responds to the disturbance. For failures affecting critical nodes or segments, certain components of the network may immediately lose functionality, resulting in a noticeable performance drop.

The curves exhibit a steep upward trend in the subsequent stage (approximately between t = 4 and t = 8), indicating that the network has entered a recovery phase following the initial impact of the failure. After t = 8, the curves level off, suggesting that the road network’s resilience performance has reached a new stable state. This implies that the network has adapted to a “new normal,” maintaining a certain level of service or functionality despite the failure. While the performance in this stable state is lower than in a fully functional state, it remains sufficient to support the network’s basic operations. The shapes of the two graphs (Figures 4(a) and 4(b)) are highly similar, but closer examination reveals slight differences in the curves. These differences provide insights into the varying impacts of node failure and segment failure on network performance. For instance, the recovery speed and the stabilization point in Figures 4(a) and 4(b) may differ slightly, suggesting that node failures and segment failures influence the network’s recovery dynamics and ultimate performance in distinct ways.

Integrating equation (7) into the resilience performance curve yields the following results: (a) for a node failure, the resilience value of the road network is 1.557, with a maximum performance of 1.8. Thus, the relative maximum performance is calculated as R = 1.557/1.8 = 0.865. (b) For a segment failure, the resilience value of the road network is 1.598, with a maximum performance of 1.835. Consequently, the relative maximum performance is determined as R = 1.598/1.835 = 0.870. When the performance falls within this range, it indicates that the road network can accommodate relative operational demands. Conversely, if the performance exceeds the relative maximum performance, it suggests that the road network is unable to meet the operational demands of road traffic. Therefore, the relative maximum performance of road network resilience for the area is established as 0.865–0.870. This range represents the peak resilience relative to the road network’s maximum capacity in high-intensity development areas. Within this range, the network can recover from disturbance events and maintain operational functionality without experiencing traffic paralysis.

The current ratio of land use demand to road traffic carrying capacity in the area is 1.057, which exceeds the resilience relative maximum performance range of the road network in this region (0.865–0.870). Traditional transportation planning and management often rely on the road saturation index, which mainly assesses road service levels by measuring the ratio of traffic flow to road capacity. According to Table 1, when the ratio exceeds 0.9, the traffic service level is considered poor; when it exceeds 1, the service level is regarded as extremely poor, indicating that the road network is overloaded. However, this standard primarily focuses on the immediacy of traffic flow and static load capacity, without adequately considering the system’s recovery capability and overall resilience in the event of a disruption or partial network failure.

In contrast, the resilience relative maximum performance index proposed in this paper evaluates not only the load-bearing capacity of the road network under normal operating conditions but also takes into account its elastic recovery capability and redundancy when subjected to disruptions. The advantage of this new index lies in its ability to more comprehensively assess the performance of the road network in the face of potential risks. Specifically, the resilience relative maximum performance index integrates both structural and functional characteristics of the road network to evaluate its potential to maintain or restore to a certain functional level after partial failure.

When the demand-to-supply ratio reaches 1.057, it not only indicates that the road network in this area has exceeded its optimal service level under designed load conditions but also suggests that, in the event of disruptions or damage, the network may not be able to quickly recover to a normal functional state. This contrasts sharply with traditional saturation analysis, which only indicates the current traffic load situation but cannot predict the recovery capacity and long-term service capability of the network. Therefore, the resilience relative maximum performance index provides more forward-looking guidance for urban traffic planning, particularly suitable for assessing and enhancing traffic safety and stability in urban core areas under high-density development and complex environments.

Introducing private vehicles to underground roadways while reserving surface road space for public transportation not only reduces surface traffic congestion but also provides pedestrians and cyclists with safer and more spacious pathways. This approach enhances the overall mobility efficiency of the city, supports a public-transit-oriented shared transportation model, and fosters low-carbon, sustainable travel modes. It aligns with global carbon reduction trends, decreases cities’ reliance on fossil fuels, and alleviates the environmental impact of transportation systems. With diminished surface traffic, air quality improves, and noise pollution is significantly reduced. Moreover, optimizing surface space enables the expansion of urban greenery, public parks, and leisure areas, creating a quieter, more comfortable environment for residents. This contributes to a healthier and more active urban lifestyle.

If the ratio of planned land demand to the road traffic carrying capacity in the analyzed area falls within the range of optimal resilience performance, it can be concluded that the road network is adequate to meet its transportation demand. Even when subjected to external disturbances, the network can sustain smooth operations due to its inherent recovery capabilities. At this stage, the focus can shift to strategies for enhancing road network capacity or mitigating surface traffic flow. To increase road network capacity, measures such as expanding peripheral surface road networks around high-intensity development areas, increasing lane numbers, and introducing one-way traffic systems to reduce intersection congestion can be incorporated into road network planning. Alternatively, expanding the coverage of medium-capacity or conventional public transportation services can be explored to ensure efficient traffic operations.

To optimize urban traffic systems and improve land use dynamics, several key urban design interventions can be implemented. One effective approach is to promote public-transit-oriented development by designating surface roads exclusively for public transportation modes such as buses and trams. This will enhance transit efficiency, reduce reliance on private vehicles, and foster sustainable shared mobility. In densely developed areas, peripheral road networks can be expanded, and interventions such as adding more lanes or introducing one-way traffic systems can help alleviate congestion at intersections and improve traffic flow.

In addition, expanding medium-capacity public transportation options can reduce the demand for private vehicles, ensuring smoother traffic operations while aligning with global low-carbon and sustainable development goals. With a reduction in surface road traffic, urban spaces can be repurposed for green areas, public parks, and pedestrian-friendly zones. These changes will not only reduce air and noise pollution but also promote healthier urban lifestyles, contributing to the overall improvement of urban livability.

To ensure the long-term sustainability of urban traffic systems, it is essential to incorporate resilience-based planning metrics, such as the resilience relative maximum performance index. This tool allows planners to assess both the load-bearing capacity and recovery potential of road networks, ensuring that transportation systems remain robust and capable of recovering from disruptions. Furthermore, integrating land use and transportation strategies by balancing the ratio of land use demand to road traffic capacity is vital. When this ratio exceeds the optimal range, measures such as reducing surface traffic flow, optimizing land use, or enhancing public transit systems should be prioritized. These interventions help maintain a balance between development intensity and transportation system capacity, ensuring a resilient, efficient, and sustainable urban environment.