EN中文

Journal of Geographical Sciences >

2022 , Vol. 32 >Issue 12: 2475 - 2502

DOI: https://doi.org/10.1007/s11442-022-2057-7

Research Articles

Spatial differentiation and influence mechanism of the connection-distribution performance of urban high-speed railway hub in the Yangtze River Economic Belt

  • WANG Degen , 1 ,
  • XU Yinfeng 1 ,
  • ZHAO Meifeng , 2, *
Expand
  • 1. School of Architecture, Soochow University, Suzhou 215123, Jiangsu, China
  • 2. School of Geographic and Environmental Sciences, Tianjin Normal University, Tianjin 300387, China
* Zhao Meifeng (1986-), Associate Professor. E-mail:

Wang Degen (1973-), Professor, specialized in urban and rural development and regional planning. E-mail:

Received date: 2021-09-28

  Accepted date: 2022-03-28

  Online published: 2022-12-25

Supported by

National Natural Science Foundation of China(41930644)

Fold

Abstract

A transportation hub is the key link in the construction of the comprehensive three-dimensional transportation corridor of the Yangtze River Economic Belt, and is the basic factor responsible for the promotion of this belt. A high-speed railway hub has the “last kilometer of time-space compression” effect and is the key to building an efficient, convenient, modern, and comprehensive transportation system. This study constructed a model for measuring the connection-distribution performance of the high-speed railway hub, determined the connection-distribution performance of the urban high-speed railway hub in the Yangtze River Economic Belt, and analyzed its spatial differentiation characteristics, further revealed the influencing mechanism of the connection-distribution performance of the high-speed railway hub. The main results are as follows: (1) The connection-distribution performance of the high-speed railway hub in the Yangtze River Economic Belt presented an “olive-shaped pattern” grade structure with two small ends and a large middle section, that is, the number of high-speed railway stations with high performance and average performance was small, and the number of high-speed railway stations with good performance and medium performance was large. (2) The connection-distribution performance of the high-speed railway hub in the Yangtze River Economic Belt showed a regional differentiation pattern of “high in the east and low in the west” and “high in the north and low in the south”, and showed an urban agglomeration differentiation pattern of “high in the core areas but low in the marginal areas”; moreover, spatial differences were prominent in the distribution of nine evaluation indexes of the connection-distribution performance of the high-speed railway hub. (3) GDP, urbanization rate, city level, station passenger flow and frequency of shuttle bus were key driving factors affecting the connection-distribution performance of the high-speed railway hub. At the same time, there were significant differences in the key driving factors for the connection-distribution performance grades of high-quality, good, medium and average.

Key words： high-speed railway hub; performance evaluation of the connection-distribution; spatial differentiation; mechanism; Yangtze River Economic Belt

Cite this article

WANG Degen , XU Yinfeng , ZHAO Meifeng . Spatial differentiation and influence mechanism of the connection-distribution performance of urban high-speed railway hub in the Yangtze River Economic Belt[J]. Journal of Geographical Sciences, 2022 , 32(12) : 2475 -2502 . DOI: 10.1007/s11442-022-2057-7

1 Introduction

As the first-level development axis of China’s “T” spatial structure, the Yangtze River Economic Belt has become one of the regions with the strongest comprehensive strength and the largest strategic support, and one of the national strategies implemented by the central government. In September 2014, the State Council promulgated “the guiding opinions on promoting the development of the Yangtze River Economic Belt by relying on the golden waterways”, and “the comprehensive three-dimensional traffic corridor planning of the Yangtze River Economic Belt” (2014-2020) (hereinafter referred to as “the planning”); it is the first time to emphasize the first lines of traffic construction in the development of the Yangtze River Economic Belt (Yu et al., 2015). Traffic construction was emphasized because the traffic is the base of space connection, which decides the strength and extent of interaction between social and economic spaces (Jin et al., 1993), and the development of transportation systems directly affects the socio-economic development of the agglomeration effect, diffusion effect, and their synergistic effect which are three dominant factors; these three factors have close correlation with resource development, urban systems and industrial layout (Wu, 2010). Therefore, the construction of an interconnected transportation system has become an important prerequisite for coordinated development of the middle and lower reaches of the Yangtze River Economic Belt and the interaction and cooperation between the east and west. By the end of 2016, the “eight vertical and eight horizontal” high-speed railway lines of Shanghai-Wuhan-Chengdu and Shanghai-Kunming have been completed, the regional inter-city high-speed railway network has been continuously improved, and the comprehensive three-dimensional transport corridor with fast and convenient high-speed railway as the core has been formed, laying a foundation for accelerating the coordinated development of the Yangtze River Economic Belt region.
Although the Yangtze River Economic Belt has initially formed a fast, efficient and interconnected high-speed railway network across different parts in the east, central region, and west, “the planning” clearly states that “the Yangtze River Economic Belt is not smooth in connection with various transport modes, and the construction of comprehensive transport hub needs to be strengthened urgently”. On the one hand, the connection between railway, highway and aviation is not smooth enough; on the other hand, inter-city traffic and urban traffic connection is not perfect. The lack of connectivity of transportation facilities in these two aspects reduces the comprehensive utilization efficiency of transportation resources. Therefore, the connection system is the key link of the comprehensive three-dimensional traffic corridor construction. Therefore, “the planning” emphasized that, in accordance with the requirement of the “zero distance transfer, seamless connection”, to improve the efficiency of the operation of the comprehensive transportation system by strengthening the construction of comprehensive transport hub, and effectively solve the “last kilometer” problem, especially the “last kilometer of time-space compression” effect associated with the high-speed connection system, plays an important supporting role in the guidance of industrial layout and urban development. The high-speed connection system is a key link of the comprehensive three-dimensional transportation corridor in the Yangtze River Economic Belt. In-depth study of high-speed connection system can fully respond to and serve the national policies, and provide reference for advancing the construction of comprehensive urban transport hub along the Yangtze River Economic Belt scientifically and improving the operation efficiency of the comprehensive transportation system.
As a portal of moving in and out of the city, high-speed railway station is an important traffic hub area with the functions of node and place. The former shows the transport hub itself as an important traffic facility reflecting transportation and functional properties, while the latter shows the influence of the transport hub on the development of the city in terms of its function and value attributable to the catalytic action of the two functions (Zheng et al., 2007), which exist in close synergy (Bertolini, 1996). The Bertolini “node-places” model is widely used in the evaluation of the coordinated development of transport hub areas (Bertolini, 1999), and its core idea is to promote the collaborative coupling and interactive promotion of node and place functions in the hub area, so as to realize the structural optimization and sustainable development of the regional space in the hub area (Lu et al., 2019; Song et al., 2016).The ontology of a transportation hub includes an aggregation node of people flow, logistics, capital flow, etc. With increasing scale of its ontology and maturity of its functions, the identification function of the transportation hub can be strengthened, more factor flows can be gathered, and several functional aggregates can be formed (Deng et al., 2018). According to Bertolini’s “node-place” concept (Bertolini, 1996; 1999), the core idea of the node function of a high-speed railway hub is to realize the rapid distribution of passenger flow (He and Guo, 2014b), which is related to the layout pattern of station hubs, diversified transfer transportation distribution system and humanized detailed service. The detailed are reflected in the following aspects.
Firstly, the layout pattern of high-speed railway station hubs is a key factor affecting the spatial transfer convenience (Duan et al., 2014). Generally speaking, the layout patterns of high-speed railway station hubs mainly include centralized, three-dimensional, station front, side and tandem, etc. Among them, the centralized layout pattern has a high degree of station space transfer convenience, which is the basis for the realization of “integrated connection” (Wang et al., 2010). The core idea of transfer layout is to take the transfer space as the core of transfer layout according to the location and type of high-speed railway hub, and arrange the transfer facilities around the transfer space (He and Guo, 2014a), such as the designing of the layout of traffic facilities in the front area of the station according to the supporting traffic circulation system of “core traffic circulation”, and the motor vehicle traffic streamline should move around the station building in a “clockwise” cycle (Xu et al., 2013).
Secondly, the transfer convenience in the space of high-speed railway stations is directly proportional to the type of transfer (Duan et al., 2013). Therefore, the construction of a core transfer system of diversified public transport including urban rail, bus rapid transit and conventional bus (Cao et al., 2019) can not only enable full play to the functions of various modes of transportation connection, but also meet the personalized needs of different groups (Zhang et al., 2015). High-speed railway public transport consolidation systems should consider urban economic conditions, urban spatial form, total population, and distribution, etc. (Teng et al., 2013). Large and medium-sized cities are mainly connected by public transport, supplemented by taxis and rapid public transport tends to replace ordinary public transport (Niu et al., 2016).
Thirdly, the transfer of high-speed railway station is a complex system, and the transportation organization, station room management, schedule matching, signage and other humanized detail service settings all affect spatial transfer convenience (Duan et al., 2013). The seamless connection between “people-oriented” high-speed railway stations and urban public transport needs to consider the personalized services required by passengers’ transfer (Loo et al., 2012; Carreira et al., 2013; Lu et al., 2018). A multi-dimensional streamline organization model, such as diversion station, can improve transfer efficiency (Xu et al., 2013). In the market for public traffic transfer point/line connection point (Lu et al., 2018), changes to large open spaces should include full connection of the line of sight to all sorts of functions such as distribution, waiting time, and business services in the same space, optimization of the environment of the transfer, and improvement of the transfer efficiency. At the same time, the high-speed transfer space should also be equipped with perfect passenger service facilities, such as information technology, information identification guide, barrier-free design and adequate lifting facilities, which should all fully embody the people-oriented concept of the green interchange design (Xu et al., 2013).
Fourthly, there is a certain correlation between high-speed railway connection and location selection of high-speed railway stations. In the city center, the connection distance between high-speed railway stations is relatively short, the public transportation system is developed, and the transportation facilities are perfect, especially the subway connection; however, transport facilities and their functions are relatively inferior in the urban expansion type and urban periphery type (Li et al., 2019; Xu et al., 2019). Moreover, some scholars have carried out researches on the public transport distribution network of high-speed railway hub. The core function of high-speed railway hub public transport distribution network is the fast distribution of passenger flow (He et al., 2014). The distribution mode with rail transit as the main body and the distribution mode with the equal emphasis on the public transportation and the individual motorized transportation should be the main development pattern of distribution mode of high-speed railway hub (He and Guo, 2014b). The coverage of public transportation network is an important factor to determine the overall competitiveness of the distribution mode of high-speed railway hub. Therefore, it is of great significance to improve the radiation scope and influence of high-speed railway hub at the regional level (He et al., 2014).
In conclusion, existing studies have mainly planned and designed the transfer function of high-speed railway hub from the perspective of architecture and engineering planning. The design of high-speed railway hub transfer functions, focusing on the microscopic aspects, and high-speed railway hubs are an important part of the city, connection-distribution performance directly reflects the “last kilometer time-space compression” effect, not only associated with high-speed railway hub itself, but also its relationship with high-speed railway stations and the city center, contact, and so on. Therefore, research on high-speed railway connection needs to be carried out from the urban scale. Based on this, taking the Yangtze River Economic Belt as an example, this paper studies the connection-distribution performance of high-speed railway hub from the perspective of urban geography. The objectives of this study were to first of all to build evaluation system and model for urban rail connection-distribution performance, then measure the connection-distribution performance and analyze the spatial differentiation characteristics of the high-speed railway, and reveal the influencing factors and mechanism of the performance of the high-speed connection-distribution system. The proposed high-speed interchange system is based on the principle of “last kilometer of time-space compression” effect. The high-speed railway connection-distribution performance measurement index system and the theory of traffic geography affecting connection performance, have strong theoretical significance. At the same time, this research focused on high-speed connection, fully responding and serve the strategic requirements of “the Yangtze River Economic Belt is not well connected in various modes of transportation, and the construction of comprehensive transportation hub needs to be strengthened” put forward in “the planning”, so as to provide scientific basis for building a smoother construction path of the Yangtze River Golden Waterway, reflecting a certain practical value.

2 Research design

2.1 Overview of the research area

The Yangtze River Economic Belt covers 11 provinces and municipalities of Shanghai, Jiangsu, Zhejiang, Anhui, Jiangxi, Hubei, Hunan, Chongqing, Sichuan, Yunnan and Guizhou, with an area of about 205.23 km2, accounting for 21.4% of the total area and exceeding 40% of both population and economy of China. This region represents the largest population, largest industry, and most complete river economic belt system in the world. Promoting the coordinated development of the upper, middle and lower reaches, and the interaction and cooperation between the East, West and Central China are important development goals of the Yangtze River Economic Belt. As the “eight vertical and eight horizontal” passenger dedicated lines, the Shanghai-Wuhan-Chengdu high-speed railway and the Shanghai-Kunming high-speed railway are the important trunk lines of the “eight horizontal” passenger dedicated lines, which straddle the three regions of East, West and Central, play a significant role in promoting the coordinated development of the region. Therefore, this study takes the station cities (prefecture-level cities/autonomous prefectures and municipalities directly under the central government) of the Shanghai-Wuhan-Chengdu high-speed railway and the Shanghai-Kunming high-speed railway as the research object, with a total of 37 high-speed railway station cities (Figure 1). In particular, seven prefecture-level cities/autonomous prefectures, including Chuzhou, Huanggang, Xiaogan, Zhuzhou, Liupanshui, Qianxinan and Qiannan, were not included in the study because the main high-speed railway station failed to pass the prefecture level city or the prefecture government.
Figure 1 Map showing the Yangtze River Economic Belt

2.2 High-speed railway connection-distribution performance measurement model

2.2.1 Principle of “last kilometer of time-space compression” effect of high-speed railway connection system

Before high-speed railway, the period of inter-city travel was long, while trips within cities are especially short. However, due to the “time-space compression” effect of high-speed railway, the traffic time between cities along the high-speed railway was greatly shortened, which completely overturned the other time feelings of “inside and outside” (Wang et al., 2009). Taking Beijing-Tianjin high-speed railway as an example, the inter-city travel time between Beijing and Tianjin has been shortened to 0.5 h, while the time consumed for going towards or leaving the high-speed railway station can reach even as long as 2 h under the background of an extremely large and scattered city (Niu et al., 2015). Taking the Wuhan-Guangzhou high-speed railway as an example, passengers travelling to Wuhan University to enjoy cherry blossoms remain stuck in traffic jams for 2-3 h after arriving in Wuhan (Wang, 2010). As a result, the connection between urban traffic and high-speed railway hub takes too much time, thus weakening the significance of the long-distance transport at high speed (Niu et al., 2016).
It can be seen that the faster the high-speed railway, the smaller its proportion of in the whole travel time and the greater the sensitivity of passengers to the length of “Off route subsidiary time” (Wang et al., 2011), which highlights the importance of the connection between high-speed railway and urban traffic, that is, to fully realize the effect of “last kilometer of time-space compression” of high-speed railway hub. To achieve this goal, it is not only separated from the comprehensive connection between high-speed railway stations and urban public transport (Lu et al., 2018), which reflects the convenient access and seamless transfer provided by the “node function” of the high-speed railway hub (Wang et al., 2011), but also closely related to the accessibility of urban public transport collection and distribution network system of high-speed railway hub. The public transport collection and distribution network system of high-speed railway hub is mainly composed of one or more combinations of urban rail transit, bus rapid transit, conventional bus, taxi and private car (He et al., 2014; He and Guo, 2014b). Its core function is to realize the rapid distribution of passenger flow in high-speed railway hub (He and Guo, 2014a). Only to ensure high quality of accessibility, namely, hub stations outside organizations, effective separation of all kinds of traffic road systems and exchange with high-grade roads of the city are required within the scope of services that facilitate passenger flow distribution (Peek et al., 2006; Zheng et al., 2007; Wang et al., 2010). Therefore, the effect of “last kilometer of time-space compression” of high-speed railway is realized through the “transfer convenience” of the connection system of high-speed railway hub and the “accessibility” of public transport collection and distribution network system, that is, the performance of connection-distribution.
Figure 2 shows, before the opening of the high-speed railway, inter-city journey was difficult because of the long-distance travel, while intra-city travel was short and easy, and the inter-city time was much longer than the intra-city time, that is, A>B. After the opening of the high-speed railway, because of inter-city time is greatly shortened due to the “time-space compression” effect of high-speed railway, namely, A’<A. Due to the “space-time compression” effect, passengers’ “inside and outside” experience based on time has been completely overturned (Wang et al., 2009). The inter-city time became less than the intra-city time (“A”<“B”), and psychological feelings, namely, high iron “Off route subsidiary time” sensitivity increased, and the sensitivity is limited by high-speed railway connection-distribution performance. If the performance is good, the effect of “last kilometer of time-space compression” can be achieved for high-speed railway, while maintaining low sensitivity to “Off route subsidiary time”. If the performance is poor, the effect of “last kilometer of time-space compression” of high-speed railway will be inhibited, and the sensitivity to “Off route subsidiary time” of high-speed railway will be strong.
Figure 2 Schematic diagram of “last kilometer of time-space compression” effect of high-speed railway
As shown in Figure 2, the “last kilometer of time-space compression” effect of high-speed railway needs to be realized through three subsystems, namely, pedestrian zone (I), transfer area (II), and collection and distribution area (III). Among them, pedestrian zone (I) and transfer area (II) reflect the “transfer convenience” effect of the connection system, while collection and distribution area (III) reflects the “accessibility” effect. Specifically, (1) the subsystem of pedestrian zone (I) is expressed as “transfer degree”, which mainly refers to the area where passengers from outbound high iron gate move to the transfer area (II). The connection performance of the transfer is indicated by the change of the walking distance and transfer time. (2) The subsystem of transfer area (II) is expressed as “connection degree”, which mainly includes the shuttle traffic area, and the shuttle’s connectivity is shown by performance indicators of the number of public transportations, types of connection traffic, operating time of public transport, departure frequency, and capacity. (3) The last subsystem is collection and distribution area (III) is characterized by “accessibility”, which mainly covers the passengers from the shuttle after transit to the destination in the city. The collection and distribution performance of accessibility is expressed through the urban traffic integration degree, congestion index, and other indicators.
Further comparing the findings of superior (Mj) and inferior (Mi) systems, (1) in the pedestrian zone (I) subsystem, high transfer degree excellent connection performance, and closer transfer to outbound high iron gate, reflecting Mj, correspond to shorter amount of time (aj) on foot to the transfer area in the outbound high iron gate, whereas a longer walking time (ai) reflects Mi. (2) In the transfer area (II) subsystem, the high-speed railway hub (Mj) has connecting traffic such as subway, bus and taxi. At the same time, high amount and departure frequency of various shuttle transportation types, good compatibility between public transportation operation time and high iron gate, and large capacity (subway) to achieve the goal of high connectivity and optimized performance. On the contrary, Mi is characterized by low connectivity, low performance of the transferring system, which increase the waiting time (bi). (3) In the collection and distribution area (III) subsystem, Mj achieves the running speed of metro connection, or improved urban road traffic network and small urban congestion index with fast-running taxis and set up fast lines or bus lanes set up to ensure the fast bus running, such that the transfer from the high-speed transfer area to the destination in the city can be achieved at a short time (cj). Finally, the goal of high accessibility and high performance was achieved. On the contrary, Mi is characterized by low accessibility, poor performance of the connection system with long driving times between the subway and urban road traffic network (ci).
In summary, after the opening of the high-speed railway, the intra-city time (B’) showed differences due to the different advantages and disadvantages of the connection-distribution system. City j with high-speed railway hub (Mj) time Bj°=(aj+bj+cj), optimal performance and relatively short travelling time, realized the “last kilometer of time-space compression” effect, and city i with high-speed railway hub (Mi) time Bi°=(ai+bi+ci), poor performance, and long travelling time, suppressed “last kilometer of time-space compression” effect. Finally, after the opening of the high-speed railway, if the high-speed railway connection-distribution performance is good, the total travel time of passengers (A°+ Bj°) is far less than the total travel time before the opening of the high-speed railway (A+B), and the sensitivity of passengers to the “Off route subsidiary time” of the high-speed railway is weak. If high-speed railway connection-distribution performance is poor, the total travel time of passengers (A°+ Bi°) is not significantly improved compared with the total travel time before the opening of the high-speed railway (A+B), and passengers have strong sensitivity to the “Off route subsidiary time”.

2.2.2 Construction of high-speed railway connection-distribution performance measurement model

(1) Performance measurement index system of high-speed railway connection-distribution
The “last kilometer of time-space compression” effect of high-speed railway is realized through high-speed railway connection-distribution performance. So, this study evaluated the connection-distribution performance of high-speed railway from three dimensions: accessibility (λ1), connectivity (λ2), and transfer degree (λ3). Among them, accessibility (λ1) represents the accessibility of high-speed railway public transport collection and distribution network, which is measured by integration index (X1) and congestion index (X2), reflecting the driving time between the transfer area of high-speed railway and the city destination. Integration index is calculated by the spatial syntax software depthmap1.0 to reflect the accessibility of urban internal roads. The degree of connectivity (λ2) represents the degree of completion of high-speed railway connection traffic. It is measured by five indicators, including the number of public transportation (X3), type of connection traffic (X4), operating time of public transportation (X5), departure interval (X6), and capacity of passengers (X7), reflecting the waiting time of high-speed railway transfer areas. The degree of transfer (λ3) represents the convenience of the high-speed railway to connect with traffic, and represents the layout characteristics of the high-speed railway station planning. The average transfer walking distance (X8) and transfer time (X9) are used to measure the walking time between the stop and transfer areas of the high-speed railway. The closer the average walking distance and the shorter the transfer time, the closer the station, the closer the upper and lower passenger points, and the better the transfer streamline. On the other hand, the farther the station, the farther the upper and lower the passenger points, and the worse the transfer line. Finally, a performance evaluation system of high-speed railway connection with three dimensions and 9 factors was constructed (Table 1).
Table 1 Performance evaluation index of urban high-speed railway connection system
Target layer System layer Index layer Direction Weight Index acquisition and calculation
Urban high-speed railway connection performance Accessibility (λ1) Integration (X1) + 0.0466 ${{I}_{i}}=\frac{m\left[ lo{{g}_{2}}\left( \left( \frac{m+2}{3} \right)-1 \right)+1 \right]}{(m-1)|\bar{D}-1|}$
The integration value reflects the degree of agglomeration between the high-speed railway station and all other spatial units in the city. The greater the integration degree, the closer the distance between the high-speed railway station and all other spatial units in the city, and the fewer obstacles between them, the higher the traffic convenience, the better the accessibility. Ii is the degree of integration, and m is the number of unit Spaces in the urban system.$\bar{D}$ is average depth value (Hiller et al., 1984)
Congestion index (X2) - 0.0211 Motor vehicle travel time/motor vehicle free flow (unblocked) travel time
connectivity (λ2) Number of public transportation/lines(X3) + 0.0338 $A=\sum{({{a}_{i}}\times {y}')}$
where ai is the traffic quantity of the ith mode of traffic connection, and y’ represents the weight
Type of connection traffic/type (X4) + 0.0381 Six types of coach, rail transit (metro and light rail), bus, taxi, social vehicles, and non-motor vehicles
Operating time of public transport/h (X5) + 0.0169 It mainly takes the daily operation time of public transportation (rail transit and bus) and obtains data through field research
Departure interval /min (X6) - 0.0042 $F=\sum{\left( \frac{\sum{{{f}_{ij}}}}{n}\times {z}' \right)}$
where fij is the departure frequency of line i of j traffic types, (i=1, 2, 3, …, n), and z° represents the weight
Capacity/person (X7) + 0.0677 $C=\sum{({{c}_{ij}}\times {{a}_{i}}\times {z}')}$
where cij is the carrying capacity of line i of j traffic types, ai is the number of traffic lines of type i traffic connection, and z° represents the weight
Transfer
degree (λ3)		 Average walking distance to transfer/m (X8) - 0.0169 $L=\frac{\Sigma {{l}_{i}}}{n}$
where li represents the distance on foot that passengers need to travel from the exit of the high-speed railway station to the type i connection station (i=1, 2, 3, …, n)
Transfer time /min (X9) - 0.0042 $T=\frac{{{T}_{1}}+{{T}_{2}}}{2}$, ${{T}_{1}}=\frac{\Sigma ({{a}_{i}}+{{b}_{i}}+{{c}_{i}}+{{d}_{i}})}{n}$
where T1 is the average transfer time of subway, and T2 is the average transfer time of bus (common bus and BRT). ai is for the time of ticket purchase of subway line i; bi refers to the time used for the security check of subway line i; ci refers to the time for the subway gate entry of the i; di is the waiting time of the i subway

Note: “+” means positive indicator, and “-” means negative indicator

(2) Grey relational TOPSIS evaluation method
The grey relational TOPSIS method integrates the grey relational analysis method and TOPSIS method, which can objectively evaluate the proximity between alternatives and ideal solutions and is widely used (Chen et al., 2004; Cao et al., 2014; Liu et al., 2017). This study comprehensively applied the grey relational TOPSIS method to measure the difference between the measurement index values and ideal values of the urban high-speed railway connection-distribution system and calculated the connection-distribution performance of the urban high-speed railway system in the Yangtze River Economic Belt (T). The weight of each measurement index was determined using the entropy method (Table 1). The connection value (T) of the urban high-speed railway system calculated using the grey relational TOPSIS method is between 0 and 1. The closer the T value to 1, the closer the connection-distribution performance of urban high-speed railway to the optimal level, and the better the connection-distribution performance. On the contrary, the closer the T value to 0, the closer the connection-distribution performance of urban high-speed railway to the worst level, and the poorer the performance.

2.3 Random Forest model of high-speed railway connection-distribution performance

This study identified the key influencing factors of the connection-distribution performance of the high-speed railway in cities in the Yangtze River Economic Belt by the Random Forest model, and analyzed the influencing mechanism of the spatial differentiation of the connection performance among the cities. The Random Forest model was proposed by Breimanin 2001 as a type of machine learning algorithm based on classification trees. Its basic principle is to use a bootstrap heavy sampling method to extract multiple samples from an original sample and perform decision tree modeling of each bootstrap sample. Because randomness is introduced in the process of decision tree generation, the overfitting phenomenon is not likely to occur. Thus, it is a natural nonlinear modeling tool with good tolerance for high-speed interchange of random factors during the process of measuring the connection-distribution performance of high-speed railway hub (Breiman, 2001; Fang et al., 2011). Owing to its distinct and unique advantages in algorithm, the Random Forest model can be used for clustering, discrimination, regression, and survival analysis. Meanwhile, it can also measure the importance of each variable according to its contribution to the prediction, so as to explain the role of each spatial variable in the high-speed railway connection-distribution performance. In this study, the Random Forest package in R software was used to fit the Random Forest model of high-speed railway connection-distribution performance. High-speed railway connection is a complex system, which involves many factors, such as the scale of high-speed railway station, the level of social and economic development of the city, and the relationship between station and urban space. Based on the existing research results, this study selected 12 variables from three dimensions, namely, scale of high-speed railway station, urban development level and relationship between high-speed railway station and city, as explanatory variables for the Random Forest model of high-speed railway connection performance (Table 2).
Table 2 Variable description and expected impact
Primary variable Secondary variables Variable explanation and assignment description Expectations
Scale of high- speed railway station Number of platforms and tracks (Number) The number of completed platforms and tracks of high-speed railway station +
Concentration of station (Concentration) Passenger flow of the station/urban population of the municipal district (Lin, 2011) +
Times of shuttle service (Times) Daily number of daily shuttles (times/day) +
Station passenger flow (Flow) Daily passenger flow during peak period of high-speed railway station (excluding important holidays) (ten thousand people) +
Station building area (S-area) Construction area of high-speed railway station (m2) +
Urban development level City level (Level) According to the notice of the State Council on the adjustment of the standard for the classification of city size, cities are divided into five grades: super city, megacity, large city, medium-sized city and small city, with values of 5, 4, 3, 2 and 1 respectively +
Urbanization rate (Rate) Urban population/total population (%) +
Proportion of tertiary products (Proportion) Value added of tertiary industry in urban GDP (%) +
Built-up area (B-area) Built-up area of the city where the high-speed railway station is located (km2) +
GDP 37 cities in the Yangtze River Economic Belt in 2018 (one hundred million yuan) +
Relationship between high-speed railway station and city Distance to the city center (Distance) The straight-line distance (km) between the high-speed railway station and the city center, where the city center is defined as the traditional commercial center or historical center of the city, is measured represented by public buildings or public spaces recognized by the public, such as People’s Square in downtown Shanghai and Guanqian street in downtown Suzhou -
Location of high-speed railway station (Location) According to “the direct distance between the high-speed railway station and the city center/the urban built-up area”, the high-speed railway station is divided into urban center stations (<0.5), edge stations (0.5-1.5), and peripheral stations (>1.5) (Zhao et al., 2015), with values of 3, 2, and 1, respectively +

2.4 Data collection and processing

In this study, the data are mainly collected through field investigation and panel data. (1) The field investigation mainly collects data of comprehensive transportation facilities of high-speed railway hub, which is carried out in two stages. The first stage is a formal preliminary research (from March 4 to 10, 2019). Suzhou north station, Nanjing south railway station, Nanchang west railway station and Shanghai Hongqiao station are selected as the pre-investigation objects. Through field investigation, preliminary drafting, correction and improvement, the relevant contents of data to be collected for official investigation are prepared. The second stage is an official investigation (from May 15 to July 16, 2019), which is divided into two groups to survey 37 MTR stations of Shanghai-Wuhan-Chengdu and Shanghai-Kunming high-speed railways at the same time. The research content included three aspects: The first concerns the number of public transportations, type of connection transportation, operating time of public transport, departure time interval, passenger capacity and other relevant data. Among them, the number of public transportations is mainly investigated by rail transit (subway and light rail), bus line (bus and BRT), and tourism line. The types of connection traffic mainly include coach, rail transit (metro and light rail), bus, taxi, social vehicles, and non-motor vehicles. The operation time of public transport is mainly the daily operation time of the rail transit and bus connection transportation with the longest operation time. The departure time interval and passenger capacity mainly calculated the weighted average sum of rail transit lines, coach lines, and bus lines. The second aspect is on station building area, times of shuttle service, the number of platform and track, station passenger flow and other relevant data of the high-speed railway hub. The third is on the walking distance to transfer, and transfer time and other relevant data, which select the time intervals from 9:00 am to 12:00 am and 14:00 pm to 16:00 pm on weekdays as survey periods for the average walking distance required by the transfer and for the transfer and connection traffic. The exit gate of the high-speed railway station was set as the starting point of the measurement, and the connection traffic station was set as the terminal point. Walking speed refers to the forward speed of most passengers. (2) Panel data mainly involves urban socio-economic development data and urban internal traffic network data. The data of the socio-economic development of the cities with the high-speed railway stations were derived from the 2017 statistical yearbook of urban and rural construction and the 2018 statistical yearbook and bulletin of 37 cities. The internal traffic network information of high-speed railway stations was vectorized from the urban traffic map of 37 stations in the Yangtze River Economic Belt (2018). In addition, existing research found that the congestion index released by Amap is basically consistent with the results obtained from the construction of road traffic operation index model (Liu, 2018), which is of scientificity to measure urban traffic accessibility. Therefore, the traffic congestion data were obtained from the traffic analysis report of major Chinese cities in 2018 released by AutoNavi based on more than 700 million users. The statistics of Amap cover 100 major cities in China, but only 17 cities could be considered in line with the objectives of this research. To account for missing data, on the basis of minimizing errors, data were extrapolated according to municipal district areas, the population of permanent residents, and motor vehicle ownership. Screening was performed and data of 20 cities with missing data were referred from those closest to them in the 100 cities. The congestion index for cities with missing data was referred from that of their closest cities, such as regarding the congestion indexes of Lu’an and Quzhou cities as reference for those of Lianyungang and Jinhua, respectively.

3 Spatial differentiation of the connection-distribution performance of urban high-speed railway hub in the Yangtze River Economic Belt

3.1 Performance grade difference is an “olive-shaped” pattern

This study uses the measurement model of high-speed railway connection-distribution performance to calculate the connection-distribution performance of 37 cities in the Yangtze River Economic Belt. The performance is divided into four grades by the natural discontinuity method (Jenks): high quality, good, medium and general. The performance values of high-quality grade are 0.464-0.574, and those of good grade, medium grade and general grade are 0.391-0.463, 0.347-0.390 and 0.296-0.346 respectively.
Figure 3 shows that in the high-speed railway connection-distribution system of 37 cities in the Yangtze River Economic Belt, the number of high-speed railway stations with high-quality performance and general performance is small, and the number of high-speed railway stations with good performance and medium performance is large. In other words, it presents an “olive-shaped” grade pattern, which conforms to the characteristics of a normal distribution (Figure 3). Specifically, the number of medium performance high-speed railway stations is the largest, up to 15, accounting for 40.54% of the total, mainly including Jiaxing south station, Jinhua station, Huzhou station, and Xiangtan north station. The number of high-speed railway stations with good performance was next to the medium, with 10 stations, accounting for 27.03% of the total, mainly including Suzhou north railway station, Wuxi east railway station, and Kunming south railway station. The number of high-speed railway stations with general performance and high-quality performance is equivalent, 7 and 5 respectively, accounting for 18.92% and 13.51% of the total. The number of general performance and quality performance HSR stations was the same, with 7 and 5 stations respectively, accounting for 18.92% and 13.51% of the total. Among them, Anshun west railway station (0.296), Shaoyang north railway station (0.333) and Qujing north railway station (0.334) rank the last 3 in general performance. The high quality performance grade includes Shanghai Hongqiao station (0.574), Hangzhou east station (0.552), Nanjing south station (0.551), Chongqing north station (0.536), and Changsha south station (0.519), and these five station cities rank the top 5 in all connection performance rankings. From the perspective of the difference between the highest and lowest connection-distribution performances of high-speed railway, Shanghai Hongqiao station with the highest performance had 1.431 times the average performance, while Anshun west station with the lowest performance was only 73.82% of the average performance, and the difference between the two is nearly one. Therefore, the spatial difference is significant.
Figure 3 Performance level distribution of the high-speed rail connection-distribution system in the Yangtze River Economic Belt

3.2 Regional differentiation of performance shows the distribution of “high in the east but low in the west, and high in the north but low in the south”

The ArcGIS 10.5 Geostatistical Analyst module was used to generate the overall trend surface analysis chart of high-speed railway connection-distribution performance of the Yangtze River Economic Belt (Figure 4), in which the green line represents the east-west direction and the blue line represents the north-south direction. The connection-distribution performance of urban high-speed railway in the Yangtze River Economic Belt generally presents an unbalanced spatial distribution pattern of “high in the east and low in the west” and “high in the north and low in the south”.
Figure 4 Overall trend of high-speed railway connection-distribution performance in the Yangtze River Economic Belt
From the viewpoint of direction, the eastern region of the Yangtze River Economic Belt (including Shanghai, Jiangsu, and Zhejiang provinces) had the highest performance (0.439), followed by the central region (including Anhui, Jiangxi, Hubei, and Hunan) (0.398), and the western region (including Chongqing, Sichuan, Guizhou, and Yunnan), which had the lowest connection performance (0.381). This corresponds to a spatial differentiation pattern of “high in the east and low in the west”. Among them, the performance grades of high-speed railway connections in the eastern region were high-quality, good, and medium, accounting for 36.36%, 36.36%, and 27.28% respectively. There were no high-speed railway stations with general performance. The performance grade of high-speed railway connection in the central region was medium, accounting for 55.56%, while the other three grades were relatively small. In the western region, the performance grade of high-speed railway connection was mainly good and general, accounting for 37.5%, while the proportions of high-quality and medium were relatively small.
From the north and south direction, north of the Yangtze River Economic Belt (including the Shanghai-Wuhan-Chengdu high-speed railway covering 15 HSR stations in Shanghai, Jiangsu, Anhui, Hubei, Chongqing, and Sichuan) showed high performance (0.438), and the southern region (including the Shanghai-Kunming high-speed railway covering 22 HSR stations in Zhejiang, Jiangxi, Hunan, Guizhou, and Yunnan) showed low performance value (0.376) was low, with a spatial pattern of “high in the north but low in the south”. Among them, the performance grades of high-speed railway connections in the northern region were mainly high-quality, good, and medium, while the proportion of general grade was very small at only 6.67%. The performance grades of high-speed railway connections in the southern region were medium and general, accounting for 50% and 27.27%, respectively.

3.3 Urban agglomeration differentiation of performance shows a pattern of “high in the core areas and low in the marginal areas”

The Yangtze River Economic Belt mainly covered five urban agglomerations: Yangtze River Delta, the middle reaches of the Yangtze River, Chengdu and Chongqing, Central Yunnan and Central Guizhou. From the perspective of inter cluster differentiation, there was a significant difference in the connection-distribution performance of high-speed railway hubs in the five urban agglomerations of the Yangtze River Economic Belt (Figure 5). Yangtze River Delta Urban Agglomeration and Chengdu-Chongqing Urban Agglomeration were the high-performance value aggregation areas, with average values of 0.448 and 0.447 respectively. Seven of the top ten high-speed railway stations were located in these two urban agglomerations. The middle reaches of the Yangtze River, Central Guizhou and Central Yunnan urban agglomerations were low performance aggregation areas, with average values of 0.383, 0.377 and 0.367 respectively, 16.97%, 18.83% and 22.07% lower than the Yangtze River Delta Urban Agglomerations.
Figure 5 Spatial distribution of the connection-distribution performance of the high-speed railway hubs in the Yangtze River Economic Belt
From the perspective of inter-city differentiation within urban agglomerations, the performance of high-speed railway hub connection-distribution in the core cities of the five urban agglomerations was higher than that in the marginal cities, showing a distribution pattern of “high in the core areas and low in the marginal areas” (Figure 5). According to the city’s comprehensive economic strength, Shanghai, Wuhan, Chongqing, Guizhou and Kunming of the Yangtze River Delta, the middle reaches of the Yangtze River, Chengdu and Chongqing, Central Guizhou and Central Yunnan were selected as the core cities, while Jiaxing, Fuzhou, Suining, Anshun and Qujing were the representatives of the marginal cities of each urban agglomeration. The performance value of the core city in the Yangtze River Delta Urban Agglomeration (Shanghai Hongqiao station) was 0.575, which was 62.94% higher than that of the marginal city (Jiaxing south station); the performance value of core cities (Hankou station) of urban agglomeration in the middle reaches of the Yangtze River was 0.464, 38.14% higher than that of marginal cities (Fuzhou east station); the performance value of core cities (Chongqing station) of Chengdu-Chongqing Urban Agglomeration was 0.536, which was 51.86% higher than that of marginal cities (Suining station); The performance value of core cities (Guiyang station) of Central Guizhou Urban Agglomeration was 0.421, which was 25.71% higher than that of marginal cities (Qujing station); The performance value of core cities (Kunming station) in Central Yunnan Urban Agglomeration was 0.438, which is 47.60% higher than that of marginal cities (Anshun station). It can be seen that the performance values of core cities and marginal cities of urban agglomeration were significantly different.

3.4 Spatial variation of each index value of performance is obvious

(1) Level of accessibility. 1) High-speed railway stations with a high degree of integration are concentrated in the eastern region, whereas those with a low degree of integration are distributed in the central and western regions (Figure 6a1). Figure 6a1 shows that, the internal accessibility of some cities in the eastern region is good and cities with high integration value are mainly distributed in Shanghai (0.175), Suzhou (0.165), Nanjing (0.135) and Hangzhou (0.128), while the internal accessibility of cities in the western region is relatively poor, and the city with the lowest integration value is Tongren city, which is only 0.011. 2) It can be seen from Figure 6a2 that high-speed railway stations with high congestion index are concentrated in national and regional central cities. For example, the congestion index of Chongqing as a municipality directly under the central government and Kunming, Changsha, Hangzhou and Chengdu as provincial capitals all exceed 1.5, that is, the actual travel time is more than half of the ideal travel time. This is mainly because the higher the city level, the larger the built-up area and the larger the population size, the more urban traffic is likely to be congested, which will have a certain negative impact on the connection-distribution performance of high-speed railway hub.
Figure 6 Spatial distribution of the indicators of the connection-distribution performance of the high-speed railway hub in the Yangtze River Economic Belt
(2) Level of connectivity. 1) The number of public transport lines and the capacity of passengers show a high spatial correlation (Figures 6b1 and 6b5). High-speed railway stations with more public transport lines and passenger capacity are mostly located in the city center and fringe areas, whereas those with less public transport lines and passenger capacity are mainly located in the peripheral areas of the city. Among them, Shanghai Hongqiao railway station, Hangzhou east railway station, and Chongqing north railway station have the largest number of public transport lines in the Yangtze River Economic Belt, which is 26 for all above mentioned stations. Accordingly, these three high-speed railway hubs also have the largest passenger capacity. On the contrary, Shaoyang north railway station has the least number of public transport lines at only one, and the smallest passenger capacity. 2) Figure 6b2 shows that, high-speed railway stations with the most connections are mainly in provincial capitals and eastern developed cities, with six types of connections, such as coach, rail transit (metro and light rail), bus, taxi, social vehicles, and non-motor vehicles. On the contrary, high-speed railway stations with the smallest transportation connections are mainly distributed in general prefecture-level cities in the central and western regions, especially most of them lack subway connections. 3) High-speed railway stations with long operation hours are mainly concentrated in the regional high-speed railway hub cities. The long operation hours of connections could be mainly attributed to the large passenger flow and number of shuttles. In contrast, high-speed railway stations with short operation time are mainly distributed in low-grade high-speed railway stations with small passenger flow and few shuttles, such as Qujing north railway station and Enshi station (Figure 6b3). 4) high-speed railway stations with long departure interval are mainly distributed in general prefecture-level cities in the central and western regions. There are fewer types of high-speed railway connection traffic and public transport lines, so the departure interval is long. High-speed railway stations with short departure interval are concentrated in the densely distributed areas of eastern cities and towns. There are many types of high-speed railway connection traffic, and some high-speed railway stations are equipped with inter-city buses, resulting in short departure interval (Figure 6b4).
(3) Level of transfer. 1) High-speed railway stations with a long walking distance are concentrated in the regional HSR hub city (Figure 6c1). The HSR stations with large building areas are associated with long walking distances from the exit to the connection area. High-speed railway stations with short transfer walking distance are mainly distributed in the cities with general high-speed railway stations of low grade. High-speed railway stations with small building areas are associated with relatively short transfer walking distance. 2) There is a high spatial correlation between the transfer time and departure interval. The longer the departure interval, the shorter the transfer time (Figure 6c2). High-speed railway stations with long transfer time are mainly distributed in general prefecture-level cities in the central and western regions. The long transfer time can be attributed to the imperfect types of high-speed railway connections, few public transport lines, and long departure interval. In the eastern densely populated areas, there are many types of connection traffic, some inter-city buses are open, and the average interval of departure is short; thus, the average transfer time is longer.

4 Factors and mechanisms influencing the connection-distribution performance of urban high-speed railway in the Yangtze River Economic Belt

The Random Forest model of high-speed railway connection-distribution performance mainly involves two user-defined key parameters: ntree and mtry. To determine the parameter mtry, in the case of a large number of decision trees (ntree=1000), the higher the precision of the test parameter mtry, the smaller the deviation error of the influencing factor, and the higher its practicality. The measurement results show that when ntree=1000 and mtry=4, the accuracy reaches the highest (85.83%). Therefore, in this study, ntree=1000 and mtry=4 were selected as setting parameters for the Random Forest model. The Random Forest package in the R language platform was adopted to analyze the key driving factors and the mechanism of high-speed railway connection-distribution performance of the Yangtze River Economic Belt, with the values of high-speed railway connection-distribution performance of 37 cities as dependent variables and 12 variables in three dimensions (the grade scale of high-speed railway stations, the level of urban development and the relationship between stations and cities) as independent variables.

4.1 Key driving factors for urban high-speed railway connection-distribution performance

In this study, the two indexes of IncMSE and IncNodePurity in the Random Forest model of high-speed railway connection performance were used to measure the importance of factors (Breiman, 2001). Among them, IncMSE refers to the increase of the estimation error of the Random Forest model relative to the original error after the random value of the variable is set. The larger the IncMSE value is, the more important the variable; IncNodePurity refers to the heterogeneity influence of each variable on the observed values of each node in the classification tree. The larger the IncNodePurity value is, the more important the variable is. In this study, the importance of 12 driving factors was evaluated by the above two indicators (Figure 7). Among them, Figure 7a is the analysis result based on IncMSE, and Figure 7b is the analysis result based on IncNodePurity. The results show that the order of the importance degree of factors based on different indexes is roughly the same, and the difference of the order of the importance degree of some factors is 1-2 bits, with little difference. Therefore, in order to avoid redundancy, this study only analyzes the importance of each driver factor based on the analysis results of IncMSE. The results based on IncMSE show that the level of urban development is the key driving factor for the performance of high-speed railway connection in the Yangtze River Economic Belt, and the GDP, urbanization rate and city level of the top three in the order of importance of factors are all variables at the level of urban development. In addition, scale of high-speed railway station is also an important factor affecting the connection-distribution performance of urban high-speed railway. Its core variables of the “number of shuttle bus service” and “station passenger flow” rank fourth and fifth respectively in importance.
Figure 7 The ranking of the variable importance of the connection-distribution performance of the high-speed railway hub in the Yangtze River Economic Belt
In order to further analyze the impact of key driving factors on the connection-distribution performance of urban high-speed railway hub, this study drew the effect curves of the top four factors (GDP, urbanization level, city level and times of shuttle service) (Figure 8). In Figure 8, the abscissa corresponded to the factor value, and the ordinate corresponded to the contribution of the factor to the performance value, that is, the greater the ordinate value, the higher the performance. The action mechanism of key driving factors on performance is as follows:
Figure 8 The influence of key factors on the connection-distribution performance of the high-speed railway hub in the Yangtze River Economic Belt
(1) GDP was the most critical driving factor for the connection-distribution performance of high-speed railway hub, and its impact degree presented an “Γ-shaped” increasing Type trend (Figure 8a). That was, when GDP was less than 100 billion yuan, the performance rose sharply with the increase of GDP, reaching the peak when GDP was 100 billion yuan; When GDP was greater than 100 billion yuan, the growth rate of performance value became smaller and tended to be stable as a whole. The number of high-speed railway stations with GDP less than 100 billion yuan accounted for 75% in the Yangtze River Economic Belt. There was a large gap in performance value among the stations with GDP less than 100 billion yuan. And the performance grades were mainly medium and general. Among them, GDP of the city where Anshun west station was located was the smallest, only 80.246 billion yuan. Accordingly, the performance value of Anshun west station was also the smallest. The number of high-speed railway stations with GDP more than 100 billion yuan in the Yangtze River Economic Belt was small, accounting for 25%. The gaps in the performance values were weak among the stations with GDP more than 100 billion yuan, and the performance grades were mainly good and high-quality. Among them, when GDP was 1171.51 billion yuan, the performance value appeared the first peak, namely, Nanjing south railway station (0.551); when GDP reached 3013.39 billion yuan, the performance value reached the maximum, that is, Shanghai Hongqiao station (0.574).
(2) Urbanization level was the sub-key driving factor for the connection-distribution performance of high-speed railway hub. With the improvement of urbanization level, the performance value gradually increased (Figure 8b). Among them, the top five cities with urbanization rate were Shanghai, Nanjing, Wuhan, Changsha and Hangzhou, and their corresponding performance values were ranked 1st, 3rd, 7th, 6th and 2nd respectively. The performance grade was mainly high-quality. When the urbanization rate reached the maximum value of 0.876, the performance value also reached the maximum value, that was, Shanghai Hongqiao station (0.574). The cities with the lowest urbanization rate were Qujing, Huzhou, Anshun, Tongren and Pingxiang respectively, and their corresponding performance values were ranked at the lowest 3rd, 18th, 1st, 6th and 5th, respectively. The performance grade was mainly general. Qujing was the city with the lowest urbanization rate (0.3568), and its performance value was only 0.335.
(3) City level was an important driving factor for the connection-distribution performance of high-speed railway hub, and its influence degree showed a ladder upward trend (Figure 8c). When the city level≤2, that was, medium-sized and small cities, the performance value increased slowly with the rise of city level; when 2<city grade≤5, that was, big cities, supercities and megacities, the performance value increased significantly with the rise of city level. The medium-sized and small cities in the Yangtze River Economic Belt including Anshun and Enshi totalled 21, accounting for 56%. The performance value of medium-sized and small cities was low, and the performance grades were mainly medium and general. Major cities, supercities and megacities in the Yangtze River Economic Belt included 16 cities such as Suzhou, Shanghai and Chongqing, accounting for 44%. The performance value of major cities, supercities and megacities was high, and the performance grade was mainly high-quality and good.
(4) Times of shuttle service was the main driving factor for the connection-distribution performance of high-speed railway hub, and its influence degree fluctuated and increased (Figure 8d). When times of shuttle service was less than 63, the performance value gradually increased with the increase of times of shuttle service, and reached the first peak when times of shuttle service was 63; when 63<the number of sending shuttle buses≤91, the performance value decreased gradually with the increase of times of shuttle service, and reached the bottom when times of shuttle service was 91; then, when times of shuttle service was more than 91, the performance value increased sharply with the increase of times of shuttle service. There were only 9 high-speed railway stations in the Yangtze River Economic Belt with times of shuttle service was being less than 63 a day, accounting for 21% of the total. The performance value was low, and the performance grades were mainly general and medium, including Fuzhou east station, Anshun west station, etc. There were few high-speed railway stations with times of shuttle service between 63 and 91, including 7 high-speed railway stations such as Xinyu north station and Huzhou station, and the performance grade was still general and medium. There were a large number of high-speed railway stations with times of shuttle service more than 91, including 20 high-speed railway stations such as Nanjing south railway station and Shanghai Hongqiao, accounting for 51% of the total. When the times of shuttle service reached 520 (Nanjing south railway station), the performance reached the peak.

4.2 Key driving factors for the connection-distribution performance of high-speed railway

Using the stochastic forest model, the importance of single scale factors of high-speed railway connection-distribution performance of four grades (high-quality, good, medium, and general) was measured and ranked. The results show that the key driving factors of high-speed railway connection-distribution performance differ in grade, that is, the key driving factors of performance of the four grades are different. Specifically speaking:
(1) The three key driving factors for the high-quality connection-distribution performance of high-speed railway are GDP, urbanization level, and distance to the city center (Figure 9a). GDP had the greatest impact on the high-quality performance, with an importance index of 14.03. So, the level of urban development had the greatest impact on the high-quality performance, with four of the top six ranked factors in importance ranking belonging to the level of urban development. It shows that for high-speed railway stations with high-quality performance, the higher the urban development level, the better the connection-distribution performance.
Figure 9 The variable importance of the connection-distribution performance of the high-speed railway hub in the Yangtze River Economic Belt
(2) The three key driving factors for the good connection-distribution performance of high-speed railway are the station building area, times of shuttle service, and distance to the city center (Figure 9b). Station building area had the greatest impact on the performance of good grade, with an importance index of 7.33. Therefore, the grade scale of high-speed railway station had the greatest influence on the performance of good grade, and four of the top six ranked by the importance of factors belong to the grade scale level of high-speed railway station. In other words, for high-speed railway stations with good connection-distribution performance, the higher the grade and the larger the scale of high-speed railway stations, the better the connection-distribution performance.
(3) The key driving factors of the connection-distribution performance of medium high-speed railway are the station building area and built-up areas (Figure 9c). Station building area had the greatest impact on the connection-distribution performance of medium high-speed railway, with an importance index of 9.26. In summary, both the scale of high-speed railway stations and urban development conditions had significant impact on the performance of medium grade. Among the top five factors in the order of importance, two are subordinate to the scale of high-speed railway stations, and two are subordinate to the level of urban development. For high-speed railway stations with medium performance, the higher the scale of high-speed railway stations and the higher the level of urban social and economic development, the better the connection-distribution performance of high-speed railway.
(4) The key driving factors of the connection-distribution performance of general grade HSR are the passenger flow of the station and distance to the city center (Figure 9d). Station passenger flow had the greatest impact on the connection-distribution performance of the general grade, with an importance index of 6.79. It shows that the scale of high-speed railway stations and the relationship between stations and cities are the main influencing factors of the connection-distribution performance of general high-speed railway, while the level of urban development has a weak influence on the connection-distribution performance of general high-speed railway. For high-speed railway stations with general performance, the higher the scale of high-speed railway stations and the closer they are to the city center, the better the performance of high-speed railway stations.

4.3 Influence mechanism of urban high-speed railway connection-distribution performance

Based on the results of the identification of key driving factors and the analysis of the mode of action of high-speed railway connection-distribution performance in the Yangtze River Economic Belt, this study constructed the framework of the mechanism of impact of high-speed railway connection-distribution performance in the Yangtze River Economic Belt (Figure 10). In general, the connection-distribution performance of high-speed railway in the Yangtze River Economic Belt is influenced by the interaction of three dimensions, namely, the level of urban development, scale of high-speed railway stations, and relationship between stations and cities. Among them, the level of urban development with GDP, urbanization rate, and city level as the representative indicators and the scale of high-speed railway station with passenger flow and times of shuttle service as the representative indicators are the key factors affecting the connection-distribution performance of high-speed railway. Specifically speaking:
Figure 10 The influence mechanism of the connection-distribution performance of the high-speed railway hub in the Yangtze River Economic Belt
(1) Urban development level. The three indicators of GDP, urbanization rate, and city level depict the social and economic development level of a city from the perspective of economic development level, social progress status, and population agglomeration scale, respectively, which are the concentrated embodiment of the comprehensive strength of a city. The results show that the higher the GDP, urbanization rate, and city level, the higher the performance of high-speed railway connection-distribution, and vice versa. That is, the level of urban social and economic development has a positive impact on the performance of high-speed railway connection. The level of urban social and economic development affects the performance connection-distribution of high-speed railway from two aspects: on one hand, the level of urban social and economic development represents the capital adequacy of the city to invest in the construction of high-speed railway hub, suggesting economic strength, high-speed shuttle transportation system, more abundant construction funds, more perfect construction, and high connectivity; on the other hand, the weaker the city’s economic strength, the more scarce the construction resources, the lower the connectivity, and the lower the performance. The level of urban social and economic development represents the degree of improvement of a city’s comprehensive transportation system. The more mature the city and the higher the level of urban development, the more perfect the urban transportation infrastructure construction. As a result, the integration degree and connectivity as well as the performance value would be higher.
(2) Scale of high-speed railway station. The passenger flow and the times of shuttle service are two important indexes that represent the scale of high-speed railway stations, and are also the core indexes that distinguish general high-speed railway stations from hub high-speed railway stations. The results show that the higher the passenger flow of the station and the higher the times of shuttle service, the higher the performance value. The connection-distribution performance of hub type stations is much better than that of general type stations. The scale of high-speed railway stations represents the degree of demand for HSR. The higher the high-speed station rating scale, the larger number of passengers. The optimization would also be improved, resulting in more public transport lines, longer operation time, lower departure interval, more passengers and shorter transfer time, which improves the connection degree and transfer degree of high-speed railway hub, so the performance value is large. Conversely, the performance value is small.
(3) Relationship between station and city. The location of the high-speed railway station is the key index of the spatial relationship between the high-speed railway station and city. The results show that the shorter the distance between the high-speed railway station and the city, the higher the performance. The performance of high-speed railway connection-distribution in urban center stations, edge stations and peripheral stations decreases successively. The shorter the distance between the high-speed railway station and the city, the higher the amount of original urban transportation infrastructure that the high-speed railway connection transportation system can rely on, the shorter the running time, and the greater the performance value. Conversely, the smaller the performance value. However, compared with the urban development level and the grade scale of the high-speed railway station, the influence degree of the relationship between the location of the high-speed railway station and the distance from the city center on the connection performance is relatively weak. In the order of importance of 12 factors based on IncMSE, the location of the HSR station and distance from downtown areas only ranked 7th and 9th, respectively.
The key driving factors of connection-distribution performance under different grades are not completely consistent and there are obvious differences. The advantages and obstacles of high-quality, good, medium, and general high-speed railway performance to the transportation system are different. This means that when formulating optimization and upgrading strategies and related policies for high-speed railway hubs, it is necessary to take measures based on “class” and “station” conditions to realize the targeted and effective upgrading and improvement.

5 Conclusion and discussion

As the implementation of the national strategy of the Yangtze River Economic Belt, “the comprehensive three-dimensional transport corridor planning of the Yangtze River Economic Belt” clearly pointed out that in accordance with the requirements of the “last kilometer of time-space compression” of high-speed railway hub, we should speed up the construction of comprehensive transportation hubs, strengthen the organic connection of various transportation modes, improve the operation efficiency of the comprehensive transportation system, and solve the “last kilometer” problem effectively. As the key link of the comprehensive three-dimensional transportation corridor of the Yangtze River Economic Belt, the connection system, especially the high-speed railway connection system, has the effect of “last kilometer of time-space compression”, which has become an urgent task for the planning and construction of the comprehensive three-dimensional transportation corridor of the Yangtze River Economic Belt. Therefore, the research on the measurement and evaluation of the performance of the high-speed railway hub system responds to and serve national policies, which is the requirement of building the Yangtze River smoother Golden Waterway, so it has very important practical significance.
The principle of the “last kilometer of time-space compression” effect of the high-speed railway hub is explained theoretically. The “last kilometer of time-space compression” effect of high-speed railway hub is not only related to the high-speed railway hub itself, but also related to the relationship between the high-speed railway station areas and the urban centers. The connection performance of the high-speed railway hubs should be analyzed at the urban scale, that is, the connection performance of high-speed railway hub in a broad sense. It includes not only the comprehensive connection system between high-speed railway hub and urban public transport in the station area (that is the connection performance of high-speed railway hub in a narrow sense) through “transfer convenience”, but also the public transport collection and distribution network system of high-speed railway hub outside the station area through “accessibility”. Therefore, the “last kilometer of time-space compression” effect of high-speed railway hub is quantified through the connection-distribution performance.
The measurement of the connection-distribution performance of high-speed railway hub is constructed. Since the “last kilometer of time-space compression” effect of high-speed railway hub is represented by the connection-distribution performance of high-speed railway hub, it is mainly reflected in two aspects: One is “fast speed”, that is, the operation speed of connecting transportation, which directly reflects the “last kilometer of time-space compression” effect of high-speed railway hub. The other is “convenience”, that is, the convenience of connecting transportation, which indirectly reflects the “last kilometer of time-space compression” effect of the high-speed railway hub. Therefore, the connection performance of high-speed railway hub is measured and evaluated from the two aspects of “fast speed” and “convenience”. Specifically, “fast speed” is measured by the dimension of “accessibility”, which reflects the collection-distribution performance; and “convenience” is mainly measured by the two dimensions of “connection” and “transfer”, which reflects the performance of transfer convenience. As a whole, the high-speed railway hub connection-distribution performance evaluation system is constructed from three dimensions and comprised of nine factors.
This study identifies the spatial differentiation of connection-distribution performance of high-speed railway hub in the Yangtze River Economic Belt, and reveals its influence mechanism. According to the natural discontinuity method, the connection-distribution performance of 37 high-speed railway hubs in the Yangtze River Economic Belt can be divided into four categories: high-quality, good, medium and general. The spatial differentiation characteristics can be identified from four levels, that is, the performance grade difference exhibits the “olive-shaped” pattern; the regional differentiation of performance shows an unbalanced pattern of “high in the east but low in the west, and high in the north but low in the south”; the urban agglomeration differentiation of performance shows a pattern of “high in the core and low in the marginal areas”; the factor level also has obvious spatial differences. The performance of connection collection and distribution of urban high-speed railway hub is influenced not only by the macro level factors, but also by the micro level factors. Specifically, GDP, urbanization rate, city level, station passenger flow and times of shuttle service are the key driving factors affecting the connection-distribution performance of high-speed railway hub. And the above determinants have grade differences, that is, there are significant differences in the key factors among the high-quality, good, medium and general high-speed railway hubs.
High-speed railway station is an important urban transportation hub, which has two functions of node and place. According to Bertolini’s “node-place” model, promoting the coordinative coupling development of node function and place function in high-speed railway hub is the core to realize the sustainable development of high-speed railway hub. This study makes the systematic and in-depth analysis on the node functions of high-speed railway stations based on the connection-distribution performance of high-speed railway hub. Comprehensively considering the two functions of node function and place function, exploring the impact and catalysis of high-speed railway station on urban function development will be the focus of high-speed railway hub research in the further research. As the high-speed railway platform design will also have a certain impact on the connection performance, which belongs to the micro research field of architecture. Therefore, it is necessary to implement cross-disciplinary study, absorb relevant theories and methods in the field of architecture, and take the platform design elements into account in the high-speed railway connection performance research in the future.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Bertolini L, 1996. Nodes and places: Complexities of railway station redevelopment. European Planning Studies, 4(3): 331-345.

DOI

[2]
Bertolini L, 1999. Spatial development patterns and public transport: The application of an analytical model in the Netherlands. Planning Practice & Research, 14(2): 199-210.

[3]
Breiman L, 2001. Statistical modeling: The two cultures. Statistical Science, 16(3): 199-231.

DOI

[4]
Cao X Z, Zeng G, 2014. The mode of transformation and upgrading based on the methods of entropy weight and TOPSIS in case of Wuhu economic and technological development zone. Economic Geography, 34(4): 13-18. (in Chinese)

[5]
Cao Y, Li S T, Sun W et al., 2019. Research on spatial planning strategy of Luoyang Longmen high-speed railway hub area based on restructuring action. Urban Development Studies, 26(3): 15-20. (in Chinese)

[6]
Carreira R, Patrício L, Jorge R N et al., 2013. Towards a holistic approach to the travel experience: A qualitative study of bus transportation. Transport Policy, (25): 233-243.

[7]
Chen M F, Tzeng G H, 2004. Combining grey relation and TOPSIS concepts for selecting an expatriate host country. Mathematical & Computer Modelling, 40(13): 1473-1490. (in Chinese)

[8]
Deng H B, Lu L, Yu H, 2018. Spatial evolution of air-rail integrated transport hub area: The case of Shanghai Hongqiao hub. Human Geography, 33(4): 130-136. (in Chinese)

[9]
Duan J, Yin M, 2013. Study on convenience degree of spatial change in high-speed rail site in Yangtze River Delta region. Science China, 43(2): 201-207. (in Chinese)

[10]
Duan J, Yin M, 2014. Study on the relationship between planning layout of high-speed railway stations and transfer space convenient level: An empirical study of Yangtze River Delta region. City Planning Review, 38(10): 44-50. (in Chinese)

[11]
Fang K N, Wu J B, Zhu J P et al., 2011. A review of technologies on random forests. Statistics & Information Forum, 26(3): 32-38. (in Chinese)

[12]
He X Z, Guo X C, 2014a. Layout method for high-speed passenger railway terminal transfer facility based on transfer space. Modern Urban Research, (4): 97-102. (in Chinese)

[13]
He X Z, Guo X C, 2014b. Transport planning of high-speed railway hub. Planners, 30(2): 53-58. (in Chinese)

[14]
He X Z, Guo X C, Zhang X H, 2014. High-speed railway passenger distributing patterns and development strategies. Urban Transport of China, 12(1): 41-47. (in Chinese)

[15]
Hiller B, Hanson J, 1984. The Social Logic of Space. Cambridge: Cambridge University Press, 46-48.

[16]
Jin F G, 1993. A study of coexistent development of transport and economic connection. Economic Geography, 13(1): 76-80. (in Chinese)

[17]
Li S Q, Lu J Y, Mao H L, 2019. Research on site selection and connecting traffic system of high-speed railway station in China. Journal of Human Settlements in West China, 34(6): 94-101. (in Chinese)

[18]
Lin C H, 2011. Research of impact factors of high-speed railways hub area development in China. Urban Planning International, 26(6): 72-77. (in Chinese)

[19]
Liu H Y, Lei L, 2017. Comprehensive evaluation on development level of regional smart industry in China. Economic Geography, 37(2): 106-113. (in Chinese)

[20]
Liu Q, 2018. Study on multi-level road traffic congestion index based on big data. Chongqing Jiaotong University. (in Chinese)

[21]
Loo B P Y, Lam W W Y, 2012. Geographic accessibility around health care facilities for elderly residents in Hong Kong: A microscale walkability assessment. Environment and Planning B: Planning and Design, 39(4): 629-646.

DOI

[22]
Lu L, Deng H B, 2019. Progress and prospect of the node-place model and its application. Scientia Geographica Sinica, 39(1): 12-21. (in Chinese)

DOI

[23]
Lu P Y, Wang B, 2018. The importance of integrated transport in fostering the formation of the Guangdong-Hong Kong-Macao Greater Bay area. Progress in Geography, 37(12): 1623-1632. (in Chinese)

DOI

[24]
Niu Y, Wang D G, 2015. Perceptual evaluation of connection between urban transport and HSR station by traveler: A case study of Suzhou. Human Geography, 30(6): 98-105. (in Chinese)

[25]
Niu Y, Wang D G, 2016. Characteristics and modes of connection system between urban transportation and high-speed railway station: Case study of Suzhou and Shanghai. Tourism Tribune, 31(3): 106-113. (in Chinese)

[26]
Peek G J, Bertolini L, Jonge H D, 2006. Gaining insight in the development potential of station areas: A decade of node-place modeling in the Netherlands. Planning Practice and Research, 21(4): 443-462.

DOI

[27]
Song W J Shi, Y J, Zhu Q et al., 2016. Evaluation on planning of high-speed railway station area based on node-place model in Yangtze River Delta area. Economic Geography, 36(10): 18-25. (in Chinese)

[28]
Teng J, Shen B, Fei X et al., 2013. Designing feeder bus lines for high-speed railway terminals. Systems Engineering-Theory & Practice, 33(11): 2937-2944. (in Chinese)

[29]
Wang H, Long H, 2009. Effect of high-speed railway network on spatial structure of urban agglomeration. City Planning Review, (4): 41-44. (in Chinese)

[30]
Wang H Y, 2010. Heat cherry tourism by high-speed rail induced cold reflection, how tourism of Wuhan develop[EB/OL]. http://hb.qq.com/a/20100329/001723.htm, 2010-03-29. (in Chinese)

[31]
Wang J, Zeng J, Liu C, 2010. A study of transit interchanges between high-speed railway and the city bus. Urban Planning Forum, (6): 68-71. (in Chinese)

[32]
Wang J X, Lin C H, 2011. High-speed railway and its impacts on the urban spatial dynamics in China: The background and analytical framework. Urban Planning International, 26(1): 16-23. (in Chinese)

[33]
Wu W H, 2010. Research on basic characteristics and system framework of urban passenger transport development. Macroeconomics, (4): 3-31. (in Chinese)

[34]
Xu H N, Lai X, 2019. Research and practice of traffic connection strategy of high-speed railway hub based on location and function analysis. Traffic & Transportation, (7): 119-123. (in Chinese)

[35]
Xu W Y, Huang Z Y, 2013. On the interchange system at high-speed railway station. Urban Mass Transit, 16(1): 82-86. (in Chinese)

[36]
Yu X G, Wang L, 2015. Background of the Yangtze River Economic Belt development strategy and geography interpretation of its innovative development. Progress in Geography, 34(11): 1368-1376. (in Chinese)

DOI

[37]
Zhang X H, Guo X C, Du X C et al., 2015. Travelling characteristics of intercity railway passengers and the analysis of the interchange transportation system. Modern Urban Research, (6): 2-7. (in Chinese)

[38]
Zhao Q, Chen G W, 2015. Impact of location on land use development around high-speed railway stations. Urban Transport of China, 13(3): 17-23. (in Chinese)

[39]
Zheng D G, Du B D, 2007. Looking for the balance between transport value of node and functional value of city. Urban Planning International, 22(1): 72-76. (in Chinese)

Options
Outlines

/

〈 〉