Evaluating the Comprehensive Development Level and Coordinated Relationships of Urban Multimodal Transportation: A Case Study of China’s Major Cities

Abstract

Urban multimodal transportation effectively meets the diversified travel demand of residents. However, it also generates extensive development problems such as traffic congestion, exhaust emissions and low operational efficiency. Therefore, there is an urgent need in urban sustainable development to achieve the coordinated and stable development of various modes of transportation. In this study, we took 36 major cities in China as the research object; measured the comprehensive development level of urban multimodal transportation; used the coupling coordination degree model (CCDM) to research the coordinated development relationship among buses, rail transit, and taxis; and clarified the shortcomings of the coordinated development of multimodal transportation. The results show that the comprehensive development of urban multimodal transportation in China has shown a significant upward trend from 2016 to 2020, with an average annual growth rate of about 7.36%. There are significant differences in the development levels of multimodal transportation in different cities. In addition, the relationship among buses, rail transit, and taxis in the major cities in China presents a state of uncoordinated development. Therefore, the relevant departments of cities should optimize the allocation of transportation resources, in terms of infrastructure construction and operation, according to these development levels and coordination of multimodal transportation.

1. Introduction

The rapid development of urban motor vehicles has met people's demand for convenient travel, and people’s travel distances and travel times have become longer. However, this has also led to extensive development problems, such as traffic congestion, exhaust emissions, and low operating efficiency. These problems are seriously restricting the green and sustainable development of the use of urban resources and harming the environment and the economy. For example, Fan et al. [1] concluded that transportation accounts for about 32% and 27% of the global CO₂ and greenhouse gas production, respectively. The 2021Q1 China Major Cities Traffic Analysis Report showed that 1.66% of the 361 cities monitored were in a congestion state, 59% had a slow traffic state, and 39.34% a smooth traffic state, based on Amap traffic big data [2]. Moreover, the overall traffic situation still needs to be improved and optimized. Today, a single mode of urban transportation can no longer satisfy the travel demand of residents. Meanwhile, urban residents’ transportation modes are diversified, and include private car travel, public transportation, taxi (including cruise cars and online car hailing) travel, and sharing bicycles [3]. People choose their travel mode according to their travel preferences.

Urban multimodal transportation covers a variety of transportation modes, including buses, rail transit, and taxis. Among these, bus transport is one of the major transportation modes of urban multimodal transportation, having the characteristics of flexibility, convenience, and wide coverage. Furthermore, as an important mode of urban public transportation, rail transit is developing rapidly in all major cities, having the advantages of high capacity, high efficiency, and low energy consumption. Taxis, meanwhile, represent a public transportation mode that is necessary to meet the different travel needs of residents. According to the 2017 National Household Travel Survey [4], taxi trips and public transportation trips accounted for 8.5% and 13.2%, respectively. In recent years, the development of buses, rail transit, and taxis has been significant in the major Chinese cities. By the end of 2020, China had 1.5 times as many urban bus lines as in 2016 and 7.1 times as many new-energy taxis as in 2016. Meanwhile, 43 cities had opened urban rail transit lines, and the number of vehicles assigned to them was, in 2020, 2.1 times that in 2016 [5,6].

At present, the different modes of transportation in different countries are showing a trend of being developed singly. The coupling relationship between multimodal transportation is not consistent, resulting in a serious waste of resources and economic problems. For example, the gathering and distributing capacities of multiple modes of transportation are not being matched at traffic connections, due to the long access and egress times of public transportation [7], which further leads to problems such as low traffic efficiency, difficult transfers, and long waiting times for transfers. Therefore, the coordinated development of urban multimodal transportation has become an important development direction for future urban transportation. In 2013, the European Commission first proposed the Sustainable Urban Mobility Plans (SUMPs) [8], which advocated the synergistic development of different transport modes and the shift to sustainable mobility. The Singapore Land Transport Master Plan 2040 [9] states that the goal of land transport development in Singapore in 2040 is to create more convenient and well-connected land transport. Furthermore, the State Council of China issued the Outline of Building a Transport Power in 2019 [10]. The outline stated that the relatively independent development of various transportation modes would be transformed into a more integrated development. More innovation-driven development would be achieved through a move away from relying on traditional factors towards building a modern comprehensive transportation system that would be safe, convenient, efficient, green, and economical. The synergistic and stable development of such an integrated transportation system would require the mutual support of different transportation modes so as to achieve seamless interchanges and coordinated development among the various transportation modes, thereby improving the travel efficiency for urban residents, and meeting the diverse travel demand [10].

In the process of urban development, the coordinated development of multimodal transportation is conducive to the sustainable development of the regional economy and society [11]. For example, there is a correlation between urban multimodal transportation and the regional economy. Transportation provides trade opportunities and reduces production costs for integrated regional economic development [12]. For tourist cities, transportation has a significant impact on tourist destination development. Coordinated transport systems can improve the accessibility of specific locations for tourists [13]. The construction of urban multimodal transportation can also have an impact on commercial complexes, whose development requires the consideration of traffic flow, routes, layouts, and regional distribution [14]. It is important to understand the hidden relationship between urban multimodal transportation development and urban land use [15], in order to coordinate transportation and land use, and modify existing urban land planning schemes, so as to alleviate urban traffic congestion and other problems according to local traffic supply and demand [16].

Previous studies mainly focused on the development status of urban multimodal transportation and the coordination relationship between multimodal transportation and the economy, tourism, urban planning, etc. However, there is less relevant literature on the coordination among buses, rail transit, and taxi systems. Urban buses, rail transit, taxis, and other modes of transportation interact with each other in the development process, and their coordinated development is also related to the sustainable development of urban transportation. Accordingly, our research focuses on the following two areas: (1) What is the development level of multimodal transportation in major cities in China? (2) What is the coordinated development relationship between multimodal transportation in major cities in China?

Based on these research questions, we constructed an evaluation index system of the development level of urban multimodal transportation, and measured the development levels of urban buses, rail transit, and taxis. Then, we used the coupling coordination degree model (CCDM) to measure the coordinated development relationship among buses, rail transit, and taxis. Here, we focus on discussing the shortcomings and key influencing factors that restrict the development of multimodal transportation. It is of great theoretical significance and practical value to scientifically understand the development level of urban multimodal transportation and to clarify the coordination of urban multimodal transportation. Furthermore, it will help relevant departments to formulate more scientific and reasonable traffic planning schemes, improve the management of urban traffic, contribute to the healthy and sustainable development of the multimodal transportation system, and clarify the various multimodal transportation problems existing in various cities.

The rest of the paper is arranged as follows. Section 2 reviews the relevant literature. Section 3 describes the study area and data sources. Section 4 describes the methodology. Section 5 shows the analysis of the results and presents the discussion. Section 6 summarizes and puts forward suggestions.

2. Literature Review

2.1. Current Status of Urban Multimodal Transportation

At present, urban multimodal transportation includes buses, rail transit, private cars, taxis (online cars), and bike sharing [17]. Each mode of transportation is suitable for a specific market and spatial environment [18]. Bus transport is an important part of urban public transport, providing travel services with fixed stops and planned routes. Furthermore, bus services are usually synchronized with rail transit, serving as spurs or necessary complements. Rail transit is a relatively independent mode, with a fixed operation track [18]. In addition, many countries have implemented policies to reduce private car travel and encourage travelers to change their travel patterns to achieve the goal of “carbon peak and carbon neutral”. As an alternative to private cars, car sharing combines the characteristics of private and collective traditional transportation [17], and has become a recognized sustainable transportation mode around the world. In addition, the bicycle is also considered to be a green alternative to motorized travel and is part of multimodality [3]. The government should encourage the coordinated development of multimodal transportation in order to satisfy the different travel demands of residents, that is, multiple modes of transportation with efficient connections and coordination, which will further promote the sustainable development of urban transportation.

The development of multimodal transportation is closely related to local policy. Different countries and cities have formulated a series of policies. Barros et al. [19] used ordinary least squares (OLS) models to investigate the relationship between public transport market share and value added tax (VAT) in Europe, and they found that tax policies had an impact on public transport sustainability. Measures such as reducing the VAT rate on public transport fares can expand the public transport market share. Poland requires that cities with more than 50,000 inhabitants should replace at least 30% of their buses by 2028 with zero-emission buses [20]. Lee et al. [21] used the Institutional Analysis and Development (IAD) framework to identify institutional barriers in the implementation of transportation projects in Seoul, Korea. They found that an imbalance between traffic and spatial development in urban planning would lead to negative social impacts, such as poor quality of life for residents. Therefore, institutional coordination between departments should be ensured to further promote the integration of urban traffic and spatial planning. The State Council of China issued The Guiding Opinions on the Priority Development of Public Transport in Cities in 2013, which particularly emphasized and highlighted the key position and role of public transport in the whole urban transport system [22]. This looked at improving public transport conditions, strengthening the comprehensive connections between public transport and other transportation modes, the construction of a low-carbon urban public transport system, etc. Since 2015, Zhoushan, China, has implemented an urban green transportation development policy, strengthened the construction of the urban road transportation system, improved the infrastructure of new-energy vehicles, optimized the development of urban multimodal transportation, and achieved the green development of urban transportation [23]. Beijing, China, has been active in implementing driving restrictions and license plate restrictions since 2007. Relevant studies have found that driving and license plate restriction policies could reduce the number of motor vehicles and increase public passenger volumes by 20–30%, which would be conducive to promoting the green development of urban transportation [24].

2.2. Evaluation of Urban Multimodal Transportation Development

With the development of urban transportation systems, people have begun to construct index systems to evaluate the development of various transportation modes. Jin et al. [25] evaluated the development level of green transportation in Xuzhou based on the Driving forces–Pressures–State–Impacts–Responses (DPSIR) conceptual model and the analytic hierarchy process (AHP)-entropy method. This work focused on the factors influencing the development of green transportation and the construction of an index system, mainly based on aspects of social and economic development, resource utilization, environmental protection, and relevant policies. The study found that Xuzhou should pay more attention to the number of bus signs, vehicle pollutant emissions, per capita GDP, public recognition of green travel, and other indicators. Yang et al. [26] mainly considered the significance of “Internet + transportation” integrated development and constructed an evaluation index system for the urban transportation development level looking at five aspects: road infrastructure, intelligent travel, intelligent transportation, government influence on transportation, and transportation travel safety. They found that Nanjing showed excellent performance in terms of government influence on transportation, and Suzhou had better transportation information development. Awasthi [27] selected indicators from four categories: economic, environmental, social, and technical, and used a multicriteria decision making model to assess the sustainability of a transport project in Luxembourg. They found that the creation of a new tramway in the city center of Luxembourg was evaluated as having the highest score for the indicator in the technical category.

Evaluating the level and performance of public transportation development is useful for providing decision support to relevant departments. Gruyter [28] selected indicators to measure the level of urban public transport sustainability development and performance in four dimensions: economic, social, environmental, and system effectiveness; after normalizing the indicator values, they used a distance-to-reference-based approach. They found that cities in developed Western countries (Western Europe, North America, and Oceania), public transport sustainability performs better in terms of environmental and social indicators, but less well in terms of system effectiveness and economic indicators. Public transport sustainability in Asian and Latin American cities performed better in terms of economic and system effectiveness, and worse for social and environmental indicators. The overall performance of Eastern Europe was higher. Some developing cities in Asia, Latin America, and Africa should learn from developed cities in Europe and Asia to develop urban public transportation and non-motorized transportation if they want to maintain a sustainable level of urban transportation and avoid environmental losses [29]. Dawda et al. [30] used Data Envelopment Analysis to evaluate the public transport system in terms of aspects such as operations, route design, and cost efficiency in Surta, India. They found that for public transport systems, the number of buses should be shifted in route design planning, and bus route lengths should be reduced, to improve cost benefits. Zhang [31] used the structural entropy-TOPSIS model to explore the overall public transportation performance of Wuhan by constructing a public transportation priority performance evaluation index system, based on four dimensions: overall public transportation development level, infrastructure construction, service level, and policy support. The study concluded that Wuhan should focus on the harmonious development of the above four dimensions, improve public transportation, prioritize infrastructure construction, and improve public service levels.

2.3. Studies on the Coordinated Development of Urban Multimodal Transportation

The coordinated development of the interconnected urban multimodal transportation network is of great significance to the overall layout and planning of urban transportation. Hu [32] used the urban transport network measurement model to quantitatively analyze comprehensive coordination of the bicycle transportation network, the conventional public transportation network, the bus rapid transit transportation network, and the rail transportation network in Hangzhou, China. It was found that the overall development of Hangzhou’s transportation was fine, and the coordinated evolution of the multi-layer urban transportation network was high. Georgiadis et al. [33] explored the performance of 34 multimodal public transport networks worldwide by applying Data Envelopment Analysis and bootstrapping techniques. They found that, in developing the performance of bus and railway transit systems, they are more coexisting rather than just cooperating in a competitive relationship. Lin [34] constructed a public transport criteria matrix to study the performance of public transport networks in the cities of Stonnington, Bayswater, and Cockburn in terms of the level of public transport infrastructure, public transport service, economic efficiency, and sustainability. They found that Bayswater and Cockburn need to focus on their public transport infrastructure, public transport services, and economy, while Stonnington needs to strengthen its sustainability, public transport services, and economy.

In fact, the transportation capacity, service quality, and coordination degree of each subsystem in an urban multimodal transportation network are different. For example, Xue [35] used the coupling coordination degree model to study the coupling coordination status of the four subsystems in Chongqing: the general railway network, the high-speed railway network, the municipal regional network, and the urban rail network. It was found that the coupling degree of the four transportation subsystems was high, but the coordination degree was normal, and needed further improvements in the form of road network connection construction. Yu [36] used the multi-objective coordination optimization model to evaluate the operational organizational efficiency of Chengdu East Railway Station’s comprehensive passenger transport hub. It was found that, during the peak passenger flow period (19:00 to 20:00), the coordination degree of the subway’s transport capacity was high, while that of both buses and taxis was poor. It was deduced that measures such as shortening the operating time interval of Subway Lines 2 and 7 and improving the efficiency of taxi departures could improve the coordination of multimodal transportation at the hub. Yue [37] took Nanjing as an example to study the coordination degree of an urban multimodal transportation network and found that various networks were in an unbalanced distribution state and the sharing rate of the bus network was low. Relevant departments, they stated, could improve the attraction of transfers by optimizing the construction of bus facilities. For the sustainable development of urban transportation, it would be helpful to design the overall framework of the urban multimodal transportation network, considering accessibility and mobility (time, cost, and mode selection) [38].

The coordinated development of the urban public transport infrastructure should make people's travel more convenient [39]. Therefore, the development of urban multimodal transport requires the construction of a reasonable and effective transport infrastructure, which could in turn promote the coordination of the economy and comprehensively improve transport [40]. For example, the public transport infrastructure generates many benefits, including economic, social, and environmental. It could effectively improve local GDP and contribute to the improvement of urban government fiscal revenue [41]. However, the development of this infrastructure is not coordinated, resulting in adverse effects in many areas. The coordination level of public transport infrastructure was found to be low in terms of economic, social, and environmental benefits in Beijing, Tianjin, Shanghai, and Chongqing, which would affect urban development. Therefore, these cities need to further improve such coordination [42]. Moreover, He [43] found that an increase in the economic benefits of the public transport infrastructure in 13 cities in Jiangsu Province was conducive to an increase in social benefits, while environmental benefits were less dependent on economic and social benefits. In addition, the coordinated development of the urban transportation infrastructure with the economy, the resources, and the environment would be conducive to the sustainable development of a city. Chakraborty et al. [44] designed a framework and used the state of Maryland as a modeling area to measure the association between traffic patronage, land use, and socioeconomic variables, and the results found that land use type, accessibility, income, and density were significant influences on traffic patronage.

To sum up, previous scholars have mainly concentrated on the status of urban multimodal transportation, the evaluation of urban multimodal transportation, the coordinated development of the urban traffic network, and the public transport infrastructure. However, few studies have explored the issues of the development level and coordinated relationships of buses, rail transit, and taxis. In this paper, we establish an index evaluation system for the development of urban multimodal transportation, so as to estimate the trends in the development levels of buses, rail, and taxis in 36 major cities in China from 2016 to 2020. The coupling coordination status among buses, rail transit, and taxis is estimated, so as to clarify the coordinated development trend based on the CCDM. This will provide theoretical support and a decision-making basis for the coordinated development of urban multimodal transportation in China.

3. Study Area and Data Sources

We chose 36 major cities in China as the study area (ignoring Hong Kong, Macao, and Taiwan, we used the 31 provincial capitals plus Dalian, Qingdao, Ningbo, Xiamen, and Shenzhen). According to the Evaluation Index System for City Public Transportation Development Performance [45], set out by the national standards of China, we scientifically selected indexes of bus and rail transit to effectively calculate the development level of urban public transport. Furthermore, as an important transportation mode, the development level of taxis is also particularly important for the development of urban multimodal transportation. Scientifically selected evaluation indexes of the development level of buses, rail transit, and taxis can be used to reasonably measure the development status of multimodal transportation, and thus make the existing shortcomings clear.

Referring to previous studies, it was found that the construction of transportation infrastructure is an important factor affecting the urban transportation system [46]. For example, a scientific and reasonable road network layout and route planning can help buses, rail transit, and taxis run efficiently, and sufficient operating vehicles can also meet residents’ different travel demands. Moreover, operating index data, such as passenger capacity, operating mileage, and average travel speed, clearly reflect the development status of a transportation mode in a certain period. Therefore, we selected the evaluation index system to measure the development levels of buses, rail transit, and taxis based on the two aspects of infrastructure and operations. Moreover, we referred to existing relevant research results and China’s National Report on Urban Passenger Transport Development and constructed the index system according to systematic, comprehensive, scientific, and operability principles. Therefore, we selected 22 secondary indexes representing the development level of buses, 20 secondary indexes representing the development level of rail transit, and 18 secondary indexes representing the development level of taxis. See

Table 1. Evaluation index system for the development level of urban multimodal transportation.

Target Layer	First-Class Index	Second-Class Index
Buses	Infrastructure status	Operating line length (km); Number of operating lines (units); Operating line length per 10,000 people (km/10,000 persons); Number of operating lines per 10,000 people (units/10,000 persons); Year-on-year growth rate of number of operating lines (%); Length of bus lane (km); Bus lane ratio (%); Operating line length of BRT[bus rapid transit] (km); Terminal area (10,000 m2); Year-on-year growth rate of terminal area (%); Terminal area per bus (m2/ standard unit); Operating vehicles (units); Year-on-year growth rate of operating buses (%); Number of new-energy operating vehicles (units); Rate of new-energy operating vehicles (%); Number of BRT operating vehicles (units); Number of buses per 10,000 people (standard units/10,000 persons)
	Operational indexes	Number of business operators (units); Passenger capacity (transported people per year); Operating mileage (km); Annual passenger capacity per bus (transported people per year/unit); Passenger capacity per unit operating mileage (person-times/km)
Rail transit	Infrastructure status	Operating line length (km); Number of operating lines (units); Operating line length per 10,000 people (km/10,000 persons); Number of operating lines per 10,000 people (units/10,000 persons); Number of stations (units); Number of transfer stations (units); Number of transfer stations / Number of stations (%); Number of attached vehicles (units); Subway (units); Light rail transit (units); Number of attached vehicles per 10,000 people (units/10,000 persons)
	Operational indexes	Operating mileage (10,000 vehicle km); Inbound volume (transported people per year); Passenger capacity (transported people per year); Daily passenger capacity (transported people per year); Maximum passenger capacity of the line (10,000 person-times); Maximum passenger capacity of the station (transported people per year); Average travel speed (km/h); Passenger-kilometers (10,000 passenger-km); Intensity of passengers (transported people per year /km)
Taxis	Infrastructure status	Operating vehicles (units); Year-on-year growth rate of operating vehicles (units); Battery-powered electric vehicles (units); Proportion of battery-powered electric vehicles (%); Gasoline vehicles (units); Ethanol gasoline vehicles (units); Natural gas vehicles (units); Dual-fuel vehicles (units); Number of taxis per 10,000 people (units/10,000 persons)
	Operational indexes	Operating mileage (10,000 kms); Year-on-year growth rate of operating mileage (%); Annual operating mileage per taxi (10,000 kms); Passenger capacity (transported people per year); Year-on-year growth rate of passenger capacity (%); Total number of vehicles carrying passengers (10,000 times); Passenger mileage (10,000 kms); Mileage utilization rate (%); Average number of passengers per trip (transported people per year)

Note: The above data were taken from the 2016–2020 National Report on Urban Passenger Transport Development [5,6,47,48,49] produced by the Ministry of Transport, the 2017–2021 China Statistical Yearbook [50] and the website of the China Association of Metros, https://www.camet.org.cn/ [51].

for details.

4. Methodology

4.1. Calculate the Comprehensive Development Level of Multimodal Transportation

Previous studies have used different evaluation methods to measure the weights and the comprehensive evaluation scores of indexes [52]. Among them, the entropy weight method (EWM) effectively solves the problem of information overlap among multiple indexes and has relatively high reliability. A single method may lead to excessive differences in evaluation results when calculating index weights. Therefore, we introduce the coefficient of variation method (CVM) based on the EWM. Then, we use the combination of the EWM and CVM to calculate the weights and comprehensive evaluation scores of the indexes of urban multimodal transportation (bus, rail transit, and taxi). This combination weighting method weakens the influence of outliers in the index data and overcomes the equalization problem of weighting by the CVM alone, which makes the evaluation results more objective and reasonable. The steps are given below.

(1) Use the CVM to calculate the weights of the second-class indexes of urban multimodal transportation (bus, rail transit, and taxi).

The CVM uses information entropy to calculate the weight of each index based on the information differences of each index. As a completely objective evaluation method, the CVM better explains the measurement results of indexes and effectively avoids subjective judgments of researchers. The specific measurement steps are as follows:

Step 1: Data standardization. In order to eliminate the differences in units and orders of magnitude between the different types of indexes and make heterogeneous indexes homogeneous, we standardize the original data. First, the matrix formed from all the observed values of the second-class indexes is denoted by $X={\left(X_{ij}\right)}_{m\times n}$. $X_{ij}$ represents the index $j\left(j=1,2,\dots ,J\right)$ (bus indexes, $J=22$; rail transit indexes, $J=20$; taxis indexes, $J=18$) and value of the city $i\left(i=1,2,\dots ,I;I=36\right)$. The matrix formed from the standardized data is denoted by $Y={\left(Y_{ij}\right)}_{m\times n}$. Since indexes can be positive or negative, the specific standardization formulas are as follows [53]:

$$\mathrm{Positive} \mathrm{indexes}: Y_{ij}^{\prime }=\frac{X_{ij}-\min X_{ij}}{\max X_{ij}-\min X_{ij}}$$

(1)

$$\mathrm{Negative} \mathrm{indexes}: Y_{ij}^{″}=\frac{\max X_{ij}-X_{ij}}{\max X_{ij}-\min X_{ij}}$$

(2)

where $Y_{ij}$ ($Y_{ij}^{\prime }$ and $Y_{ij}^{″}$) represents the standardized value of index $j$ for city $i$ and where $\max X_{ij}$ and $\min X_{ij}$ are the maximum and minimum values of an index.

Step 2: Calculate the weights of the second-class indexes. Information entropy was first proposed by the American mathematician Shannon. With the help of the concept of entropy from physics, it is used to measure the disorder of information and the complexity and balance between systems. The larger the information entropy, the more complex the system will be, and the smaller the information entropy, the simpler the system will be [54]. The calculation of the information entropy $e_j$ and entropy method weight of index $j$ is as follows [52]:

$$\begin{array}{l}{e}_j=-\frac{1}{\ln 36}{\displaystyle \sum _{i=1}^{36}{P}_{ij}\times \ln P_{ij}\left(0\le e\le 1\right)}, P_{ij}=Y_{ij}/{\displaystyle \sum _{i=1}^{36}{Y}_{ij}}\\ A_j=\left(1-e_j\right)/{\displaystyle \sum _{j=1}^n\left(1-e_j\right)}\end{array}$$

(3)

where $P_{ij}$ is the value of index $j\left(j=1,2,\dots ,J\right)$ for city $i\left(i=1,2,\dots ,I;I=36\right)$ as a proportion of the total value of index $j\left(j=1,2,\dots ,J\right)$ across all cities. $n\left(n=1,2,\dots ,N\right)$ represents the number of indexes when calculating the second-class indexes for the bus, rail transit, and taxi development levels (for buses, $N=22$, for rail transit, $N=20$, for taxis, $N=18$).

(2) Use the CVM to calculate the weights of the second-class indexes of urban multimodal transportation (bus, rail transit, and taxi).

To avoid a situation where the final evaluation result is too different due to using a single method of calculating the index weights and to overcome the defect of equal distribution of the index weights obtained by the EWM, the CVM is introduced to ensure that the final weights for each index are more objective and reasonable. $Y={\left(Y_{ij}\right)}_{m\times n}$ is used to denote the dimensionless data matrix after normalization. The steps of the CVM used to calculate the index weights are as follows [55]:

Step 1: Calculate the mean and standard deviation of each index after standardization:

$$\overline{Y_j}=\frac{1}{m}{\displaystyle \sum _{i=1}^m{Y}_{ij}};s_j=\sqrt{\frac{1}{m}{{\displaystyle \sum _{i=1}^m\left(Y_{ij}-\overline{Y_j}\right)}}^2}$$

(4)

Step 2: Calculate the coefficient of variation:

$$v_j=\frac{s_j}{\left|\overline{Y_j}\right|}$$

(5)

Step 3: Calculate the weights of the indexes using the coefficient of variation:

$$w_j=\frac{v_j}{{\displaystyle \sum _{j=1}^n{v}_j}}$$

(6)

(3) Use the combination of the two weighting methods to calculate the combination weights ${\widehat{w}}_j$ of the second-class indexes:

$${\widehat{w}}_j=\lambda A_j+\left(1-\lambda \right)w_j$$

(7)

where ${\widehat{w}}_j$ is the combination weight of second-class index $j$. $A_j$ is the weight of second-class index $j$ calculated using the EWM. $w_j$ is the weight of second-class index $j$ calculated using the CVM. $\lambda$ is the preference coefficient, $\lambda \in \left(0,1\right)$. Without considering special conditions, the two kinds of weighting methods are generally considered to be equally important [56]. Thus, $\lambda =0.5$.

(4) Calculate the comprehensive evaluation score of the development level of urban multimodal transportation (bus, rail transit and taxi).

Firstly, based on the combination weights of the second-class indexes, we calculate the comprehensive evaluation scores of the first-class indexes of the development level of urban multimodal transportation:

$$\begin{array}{l}{B}_{i1}={\displaystyle \sum _{j=1}^{17}{\widehat{w}}_j\times Y_{ij}},B_{i2}={\displaystyle \sum _{j=18}^{22}{\widehat{w}}_j\times Y_{ij}}\\ R_{i1}={\displaystyle \sum _{j=23}^{33}{\widehat{w}}_j\times Y_{ij}},R_{i2}={\displaystyle \sum _{j=34}^{42}{\widehat{w}}_j\times Y_{ij}}\\ T_{i1}={\displaystyle \sum _{j=42}^{51}{\widehat{w}}_j\times Y_{ij}},T_{i2}={\displaystyle \sum _{j=52}^{60}{\widehat{w}}_j\times Y_{ij}}\end{array}$$

(8)

where $B_{ik}$ represents the comprehensive evaluation score for index $k\left(k=1,2\right)$ for buses in city $i\left(i=1,2,\dots ,I;I=36\right)$, with $B_{i1}$ representing the buses’ infrastructure status indexes and $B_{i2}$ representing the buses’ operational indexes. $R_{ik}$ represents the comprehensive evaluation score for index $k\left(k=1,2\right)$ for rail transit in city $i\left(i=1,2,\dots ,I;I=36\right)$, with $R_{i1}$ representing the rail transit infrastructure status indexes and $R_{i2}$ representing the rail transit operational indexes. $T_{ik}$ represents the comprehensive evaluation score for index $k\left(k=1,2\right)$ for taxis in city $i\left(i=1,2,\dots ,I;I=36\right)$, with $T_{i1}$ representing the taxis’ infrastructure status indexes and $T_{i2}$ representing the taxis’ operational indexes.

Secondly, based on the comprehensive evaluation scores for the first-class indexes of the development level of urban multimodal transportation, we use the combination weighting method to calculate the comprehensive scores for the development levels of buses, rail transit, and taxis, as follows:

$$B_i={\displaystyle \sum _{k=1}^2{\widehat{w}}_B}\times B_{ik},R_i={\displaystyle \sum _{k=1}^2{\widehat{w}}_R}\times R_{ik},T_i={\displaystyle \sum _{k=1}^2{\widehat{w}}_T}\times T_{ik}$$

(9)

where ${\widehat{w}}_B,{\widehat{w}}_R,{\widehat{w}}_T$ represent the combination weights of the first-class indexes for buses, rail transit, and taxis, respectively. $B_i$ represents the comprehensive evaluation score for the buses’ development level, $R_i$ represents the development level of rail transit and $T_i$ represents the development level of taxis. The higher the comprehensive evaluation score, the higher the development level, for each mode of transport and each city, and vice versa.

Finally, the comprehensive evaluation score for urban multimodal transportation, $U_i$, is calculated as follows:

$$U_i={\tilde{w}}_B\times B_i+{\tilde{w}}_R\times R_i+{\tilde{w}}_T\times T_i$$

(10)

where $\tilde{w}$ represents the combination weights for target layer indexes (buses, rail transit, taxis). ${\tilde{w}}_B,{\tilde{w}}_R,{\tilde{w}}_T$ represents the combination weights for the development levels of buses, rail transit, and taxis, respectively. The larger the value of $U_i$, the higher the overall development level of the urban multimodal transportation, and vice versa.

4.2. Measure the Coupling Coordination Degree of Multimodal Transportation

The coupling degree is derived from a physical concept and refers to the degree of interaction between two or more systems [57]. Based on this, the coupling coordination degree further reflects the degree of coordinated development between systems, not only measuring whether each system has a good development level, but also the interaction between the systems [58,59,60]. High coupling coordination among the urban bus, rail transit, and taxi systems would indicate that they were coordinated and promote each other. Low coupling coordination would mean they were mutually restrictive and influencing each other’s development. Next, based on the CCDM, we quantitatively measure the coupling degree and the coupling coordination degree among the three transportation systems of buses, rail transit, and taxis. The calculation steps are as follows:

Step 1: The coupling degree $C$ is calculated as follows [60]:

$$\begin{array}{l}{C}_i{}^{B-R-T}={\left\{\frac{B_i\times R_i\times T_i}{{\left[\left(B_i+R{+}_i{T}_i\right)/3\right]}^3}\right\}}^{1/3}, C_i{}^{B-R}={\left\{\frac{B_i\times R_i}{{\left[\left(B_i\times R_i\right)/2\right]}^2}\right\}}^{1/2}\\ C_i{}^{B-T}={\left\{\frac{B_i\times T_i}{{\left[\left(B_i\times T_i\right)/2\right]}^2}\right\}}^{1/2},C_i{}^{R-T}={\left\{\frac{R_i\times T_i}{{\left[\left(R_i\times T_i\right)/2\right]}^2}\right\}}^{1/2}\end{array}$$

(11)

where $C_i{}^{B-R-T}$,$C_i{}^{B-R}$,$C_i{}^{B-T}$,$C_i{}^{R-T}$ represent the coupling degree among buses, rail transit, and taxis (B-R-T), buses and rail transit (B-R), buses and taxis (B-T), and rail transit and taxis (R-T), respectively. The larger the value of $C$, the stronger the interaction and mutual influence, and vice versa.

Step 2: The coupling coordination degree is calculated as follows [60]:

$$D_i^{B-R-T}=\sqrt{C_i^{B-R-T}\times H_i^{B-R-T}}$$

(12)

$$D_i^{B-R}=\sqrt{C_i^{B-R}\times H_i^{B-R}},D_i^{B-T}=\sqrt{C_i^{B-T}\times H_i^{B-T}},D_i^{R-T}=\sqrt{C_i^{R-T}\times H_i^{R-T}}$$

(13)

where $H_i$ is the comprehensive evaluation index of urban multimodal transportation in city $i$. $D_i^{B-R-T}$ is the coupling coordination degree of B-R-T, with $H_i{}^{B-R-T}=\alpha B_i+\beta R_i+\chi T_i$. $D_i^{B-R}$ is the coupling coordination degree of B-R, with $H_i{}^{B-R}=\alpha B_i+\beta R_i$. $D_i^{B-T}$ is the coupling coordination degree of B-T, with $H_i{}^{B-T}=\alpha B_i+\chi T_i$. $D_i^{R-T}$ is the coupling coordination degree of R-T, with $H_i{}^{R-T}=\beta R_i+\chi T_i$. $\alpha ,\beta ,\chi$ are undetermined coefficients, with $\alpha +\beta +\chi =1$. We refer to the values of $\alpha ,\beta ,\chi$ determined by previous studies [61].

In order to accurately judge the development stage of the coupling coordination relationship of urban multimodal transportation, we refer to previous studies, and use the criteria shown in

Table 2. Classification of coupling coordination degree.

$\mathit{D}$	Classification
$0.8<D\le 1.0$	Superior balanced development
$0.6<D\le 0.8$	Favorably balanced development
$0.5<D\le 0.6$	Barely balanced development
$0.4<D\le 0.5$	Slightly unbalanced development
$0.2<D\le 0.4$	Moderately unbalanced development
$0.0\le D\le 0.2$	Seriously unbalanced development

[62].

5. Results

5.1. The Spatial and Temporal Characteristics of the Comprehensive Development Level of Multimodal Transportation

We calculated the comprehensive evaluation scores for the development levels of buses, rail transit, and taxis in 36 Chinese cities from 2016 to 2020, as well as the comprehensive evaluation scores for the overall development level of multimodal transportation, as shown in Figure 1.

On the whole, the comprehensive development level of urban multimodal transportation in China has shown an obvious upward trend in the five years from 2016 to 2020, with an average annual growth rate of about 7.36%. This indicates that the development of such transportation in China has improved well in recent years. The infrastructure of buses, rail transit, and taxis has gradually been improved, and the operational capacity has been greatly improved, satisfying the different travel demands and preferences of residents.

Specifically, the comprehensive evaluation score for buses was the highest, followed by that for rail transit, and then that for taxis, from 2016 to 2020. In 2020, the urban bus passenger capacity was 1.2 times that of rail transit, the taxi passenger capacity was 3.2 times that of rail transit, and the bus operating line length was 52 times that of rail transit. To some extent, this shows that buses, as a mature public transport mode, have been relatively better in terms of operation management, operation technology, infrastructure planning and construction, and other aspects, compared with rail transit and taxis. Therefore, buses show a relatively good level of development. However, in terms of growth rate, that of China's urban bus development level was only 10.6%. Looking at rail transit development, it required more infrastructure in the urban environment, and the rail transit construction and development cycle is relatively long, consuming manpower, material resources, and finances. By 2020, 32 out of 36 major cities in China had rail transit in operation. Some cities started late and still need time to develop their rail transit. However, the growth in rail transit development was fast over 2016–2020, up to 30%. To some extent, this shows that rail transit, as an emerging, green, and high-capacity public transport mode, has won the favor of more travelers. Compared with 2016, the number of rail transit stations and passenger capacity had increased by 88.2% and 8.4%, respectively, by 2020. At the same time, the number of cities with rail transit in operation has gradually increased, and the coverage of the rail transit infrastructure has also increased, allowing it to carry more passengers, and showing a good trend of development. In 2020, the taxi passenger capacity was only 37% of the rail transit passenger capacity. This is significantly lower, and there is still a big gap in the overall development level of taxis compared with buses and rail transit. However, the overall taxi development level grew rapidly from 2016 to 2020, by 21.2%. Therefore, the smooth development of the bus system and the positive development of the rail transit system and taxi system will play a crucial role in improving the development of urban multimodal transportation in China.

Figure 2 shows the comprehensive development level of buses in 36 cities in China from 2016 to 2020. On the whole, Guangzhou, Chongqing, Beijing, Shanghai, and Harbin had the highest level of bus development from 2016 to 2020. In November 2018, under guidance and support from the government, Guangzhou Bus Group completed the electrification of its buses, a green development, further improving the level of green travel in response to the Outline of Building a Transport Power. In 2020, the area of bus stations in Guangzhou was 4.1 times that in 2016, and the number of new-energy vehicles was 3.4 times that in 2016. The coverage of bus infrastructure in Guangzhou was further expanded. At the same time, in 2020, the length of the BRT operation line, the number of BRT operating vehicles, the number of new-energy operating vehicles, and other indexes put Guangzhou at the forefront of the 36 cities. In 2020, the number of business operators in Chongqing put it first among the 36 cities, and the operating line length and number of operating lines were 2.5 times and 2.4 times the average level, showing a good level of bus infrastructure construction. Beijing ranked first among the 36 cities in terms of the length of the bus lanes and passenger capacity. The number of new-energy operating vehicles in Shanghai and the operating mileage put it at the forefront of the 36 cities. For cities with a better development level of buses, this is the key factor for maintaining a good level of bus infrastructure to ensure its stable development. Meanwhile, Nanchang, Taiyuan, Lhasa, Shijiazhuang, and Haikou were the five lowest-ranked cities in terms of bus development from 2016 to 2020, with Haikou the lowest. By the end of 2020, Haikou had not opened any BRT routes or put BRT vehicles into use, and its numbers of new-energy bus vehicles and bus lanes were only 10.1% and 2.1% of Guangzhou’s. A series of infrastructure construction problems, such as with bus resource allocation, route length planning, and route operation, led to the low development level of Haikou’s bus system. Shijiazhuang, Lhasa, and Taiyuan did not plan to open BRT routes or put BRT vehicles into use in 2020. In 2020, the bus mileage and passenger capacity in Lhasa were only 8.3% and 12% of the average levels, respectively. These cities having a low level of bus development should focus on a “bus priority” development strategy, optimize the layout of local bus routes, pay attention to the delivery and use of new-energy vehicles and BRT vehicles, and further improve the development level of their urban bus systems.

In terms of the growth rate of bus development in the 36 cities, the cities with the fastest average growth rates were Lhasa, Nanchang, and Zhengzhou, of which Lhasa had the highest at 29%. In 2020, operating line length, the number of operating lines, and the number of new-energy operating vehicles in Lhasa were 2.7 times, 2 times, and 3 times those in 2016, respectively. The operating mileage increased by 8.8% compared with 2016. Although the average growth rate of Lhasa's bus development was fast, the development level was still low and there is still a lot of space for development. It will be necessary to pay attention to the planning and construction of bus lines and stations, and to open BRT vehicles and routes. The growth rate of bus development in Zhengzhou was 14.6%. In 2018, the operating line length of its BRT was 8.9 times that in 2017, representing breakthrough growth. For cities with rapid growth in bus development, the key to maintaining steady growth is to stabilize the index factors that promote development. Haikou, Wuhan, and Nanning had the slowest average growth rates, with Haikou having the lowest at -7.8%, and also the lowest development level. For example, in 2020, the number of operating lines, length of bus lanes, and number of bus vehicles in Haikou decreased by 2.7%, 61.1%, and 7.8%, respectively, compared to in 2019. Haikou’s bus passenger capacity has been in continuous decline since 2017. Compared with 2016, by 2020 its bus passenger capacity had decreased by 58.3%. The relevant departments will need to improve the local bus infrastructure construction, looking at the aspects of bus route planning, layout, and vehicles.

The growth rates of the comprehensive evaluation scores for bus development fluctuated greatly in Tianjin and Ningbo between 2016 and 2020. In 2020, the bus operating mileage and passenger capacity in Tianjin decreased by 12.5% and 43.7%, respectively, compared with 2019, while its operating line length increased by 6.3% compared with 2019. For the cities where the growth rates of bus development fluctuate greatly, the relevant departments should determine the factors restricting bus development and make some adjustments accordingly. The growth rate of bus development was relatively stable in Dalian, Lanzhou, and several others. The comprehensive evaluation score for Lanzhou's bus development level maintained steady growth from 2016 to 2020. Its operating line length in 2020 was 1.2 times that in 2019, and 1.5 times that in 2018. For cities with relatively stable growth in bus development, the relevant departments will need to maintain this stable growth, by continuing to optimize and improve the bus infrastructure, and operating level.

When it comes to urban rail transit, the duration of operations is not uniform across the cities studied. In order to make a clearer comparison of the variation in the comprehensive development of urban rail transit from 2016 to 2020, only 24 of the cities were selected as the research objects, all of which operated rail transit from 2016 to 2020. Figure 3 shows the changes in the comprehensive development of rail transit in these 24 Chinese cities from 2016 to 2020. On the whole, Beijing, Shanghai, Guangzhou, and Shenzhen had a good level of rail transit development during 2016–2020. Of these, Beijing had the best level with long operating times, perfect rail transit infrastructure, a mature operation mode and management experience, and an effective policy support environment. It was clearly required in the Beijing City Master Plan (2016–2035) [63] that the comprehensive benefits of rail transit and transportation hubs should be fully utilized, and the integrated planning of rail transit stations and use of the surrounding land should be strengthened. By the end of 2020, Beijing’s rail transit operating mileage was ahead of other cities, with the number of stations and transfer stations 3.5 times and 5.1 times the average levels, respectively. In 2020, Shanghai’s rail transit ranked first in terms of operating line length, number of stations, number of attached vehicles, passenger capacity, passenger-kilometers, and some other indexes, with excellent overall rail transit infrastructure construction. In the developing cities of the Guangdong-Hong Kong-Macao Greater Bay Area, Guangzhou and Shenzhen are actively building an integrated public transport system with urban rail transit at its core, further optimizing the structure of urban transport and alleviating urban congestion. Among these cities, Guangzhou surpassed Beijing in rail transit passenger capacity in 2020, at 5.04 times the average level. However, overall, there was still a large gap between its rail transit infrastructure construction and Beijing’s. Nanchang, Hefei, Harbin, and Fuzhou all had a relatively poor development level of rail transit from 2016 to 2020. Among them, Fuzhou had the lowest level. By the end of 2020, the length of rail transit operating lines, number of assigned vehicles, and operating mileage in Fuzhou were at only 28.5%, 27.1%, and 24.4% of the average levels, respectively. In Harbin, the number of attached vehicles and operating mileage were only 14.2% and 13.3%, respectively, of the average levels in 2020. Hefei only opened rail transit operations in 2016. For cities with a low development level of rail transit, the local departments should continue to focus on the layout and optimization of infrastructure, such as the planning of the opening of rail transit operating lines, route construction, stations, and vehicle configuration, so as to attract more passengers and meet the daily travel needs of city residents.

From the perspective of the growth rate of rail transit development in the 24 cities, Qingdao, Nanning, and Hefei had the highest average growth rates from 2016 to 2020. Qingdao had the highest at 30.3%. In 2017, the operating mileage and inbound volume in Qingdao were 2.9 times and 5.8 times those in 2016, respectively, and the number of attached vehicles increased by 75.5% compared with 2016. It is necessary to maintain a good infrastructure and operating level. Nanning and Hefei both had newly opened tracks in 2016; thus, they started late but grew fast. In 2020, the operating line length, operating mileage, and passenger capacity in Nanning were 3.4 times, 22.3 times, and 30.5 times the figures in 2016, respectively. In 2020, the operating line length and number of attached vehicles in Hefei were 4.6 times and 5.5 times, respectively, those in 2016. However, the comprehensive development levels of the rail transit in Hefei and Nanning were still relatively low. Therefore, the relevant local departments still need to pay close attention to the gap between their rail transit infrastructure construction levels and those of other cities, so as to optimize and improve their infrastructure as much as possible. Furthermore, the average growth rate of the rail transit development level was relatively low in Ningbo, Beijing, and Dalian from 2016 to 2020. Among these, Ningbo had the lowest rate of −5.1%. The operating line length, number of attached vehicles, and number of operating lines in Ningbo did not change from 2016 to 2018, and the overall growth rate was poor. Due to the impact of COVID-19, the inbound volume, passenger capacity, and daily passenger capacity in Beijing in 2020 decreased by 42.03%, 42.02%, and 42.3%, respectively, compared with 2019, but the overall development trend for Beijing’s rail transit was positive.

Figure 4 shows the comprehensive development level of taxis in the 36 cities in China from 2016 to 2020. On the whole, Beijing, Changchun, Harbin, Shenzhen, and Shanghai had good levels of development in terms of taxis from 2016 to 2020. Among them, Beijing taxis showed the highest level of development. In 2020, the number of operating vehicles and gasoline vehicles in Beijing led the country, and the number of taxis per 10,000 people was 1.7 times the average level. The overall development trend of taxis was positive. The number of taxis per 10,000 people in Changchun was 1.7 times the average level, and in terms of the taxi mileage utilization rate, it ranked first among the 36 cities. The number of ethanol gasoline vehicles, taxi passenger capacity, and taxi mileage utilization rate in Harbin were all good, with taxi passenger capacity 2.01 times the average level. The number of battery-powered electric vehicles in Shenzhen was 8.4 times the average, and the proportion of new-energy vehicles was high in 2020. The operating mileage and passenger capacity in Shanghai were 2.9 times and 2.04 times the average levels, respectively. It is necessary for taxi development to maintain a high level of basic operating vehicles. Cities with a relatively weak development level for taxis included Nanchang, Haikou, Lhasa, Hohhot, and Ningbo. In 2020, the number of taxis operating in Ningbo increased by only 3.6%, while the number of passenger miles decreased by 91.6% compared with 2016. The number of battery-powered electric vehicles in Hohhot was zero in 2020. Thus, development and investment in new-energy taxis will help Hohhot to improve its development level. In 2020, the numbers of taxis operating in Lhasa and Haikou were only 7.1% and 14.9% of the average levels, respectively. For cities with low development of taxis, the local governments should focus on updating and increasing vehicles or giving priority to the use of new-energy vehicles, so as to promote the improvement in the local taxi industry.

Looking at the average growth rates of the taxi development in the 36 cities, those with the fastest growth rates were Zhengzhou, Nanning, Kunming, Chengdu, and Chongqing. In 2020, the number of battery-powered electric vehicles in Zhengzhou achieved a large coverage area, despite the number of such vehicles in 2016 having been zero, and the total number of vehicles carrying passengers was 4.8 times that in 2016. Meanwhile, Zhengzhou introduced a subsidy policy for the renewal of battery-powered electric vehicles, greatly promoting the proportion of such taxis. Compared with 2016, the total number of vehicles carrying passengers in Nanning in 2020 was 4.5 times that in 2016. The number of battery-powered electric vehicles and the total number of vehicles carrying passengers in Kunming in 2020 were 23.08 times and 4.9 times those in 2016, respectively. Cities with slower growth rates were Lhasa, Taiyuan, Tianjin, Urumqi, and Haikou. Lhasa had the lowest rate of -11.7%. Due to the severe impact of COVID-19, the number of taxis operating in Lhasa in 2020 decreased by 39.5% compared with 2016. The overall development level of Lhasa’s taxis was poor, and the proportion of new-energy taxis was very low. The transportation department there needs to expand the share of new-energy vehicles based on specific local traffic conditions. The number of taxis operating in Taiyuan did not change from 2017 to 2020, and taxi passenger mileage in 2020 decreased by 85.6% compared with that in 2016. During 2016-2020, the number of battery-powered electric vehicles in Tianjin did not expand, and the electrification of taxis there was extremely low. Passenger capacity and passenger mileage in 2020 decreased by 77.2% and 97%, respectively, compared with 2016. For cities where the average growth rate of taxi development is relatively slow, it is necessary to prioritize the use of new-energy vehicles, to add or update vehicles operating as taxis, and to improve taxi service facilities.

5.2. Coupling Coordination Relationship of Multimodal Transportation

The results for the urban coupling coordination degrees for B-R-T, B-T, B-R, and R-T are shown in the Appendix A (

Table A1. Coupling coordination degree among B-R-T in 36 cities in China. Note: “-” represents no final data result for the corresponding cities.

	2016	2017	2018	2019	2020
Beijing	0.717	0.708	0.715	0.732	0.718
Shanghai	0.71	0.698	0.698	0.713	0.73
Guangzhou	0.669	0.673	0.654	0.707	0.721
Shenzhen	0.598	0.602	0.613	0.61	0.64
Chongqing	0.59	0.563	0.581	0.605	0.65
Chengdu	0.456	0.471	0.509	0.539	0.618
Changchun	0.451	0.484	0.526	0.523	0.535
Wuhan	0.489	0.489	0.511	0.515	0.481
Xi’an	0.475	0.464	0.497	0.509	0.531
Tianjin	0.477	0.457	0.479	0.475	0.457
Shenyang	0.421	0.423	0.45	0.46	0.471
Nanjing	0.447	0.455	0.448	0.459	0.47
Dalian	0.427	0.45	0.45	0.457	0.464
Zhengzhou	0.374	0.393	0.435	0.456	0.507
Harbin	0.434	0.439	0.452	0.446	0.424
Hangzhou	0.429	0.427	0.46	0.441	0.486
Qingdao	0.317	0.348	0.41	0.403	0.421
Changsha	0.379	0.376	0.375	0.401	0.45
Hefei	0.33	0.345	0.38	0.384	0.4
Kunming	0.348	0.348	0.393	0.38	0.409
Ningbo	0.374	0.339	0.336	0.356	0.376
Nanchang	0.296	0.326	0.313	0.356	0.346
Fuzhou	0.293	0.29	0.298	0.329	0.353
Nanning	0.3	0.311	0.355	0.325	0.355
Guiyang	-	0.32	0.343	0.375	0.359
Xiamen	-	0.337	0.338	0.363	0.391
Shijiazhuang	-	-0.282	0.304	0.297	0.3
Urumqi	-	-	0.347	0.347	0.331
Lanzhou	-	-	-	0.403	0.417
Jinan	-	-	-	0.322	0.329
Hohhot,	-	-	-	0.305	0.301
Taiyuan	-	-	-	-	0.296
Haikou	-	-	-	-	-
Lhasa	-	-	-	-	-
Xining	-	-	-	-	-
Yinchuan	-	-	-	-	-

,

Table A2. Coupling coordination degree among B-T in 36 cities in China.

	2016	2017	2018	2019	2020
Chongqing	0.601	0.611	0.614	0.653	0.727
Guangzhou	0.619	0.624	0.59	0.663	0.682
Beijing	0.617	0.602	0.619	0.646	0.646
Shanghai	0.62	0.6	0.598	0.623	0.646
Harbin	0.602	0.565	0.591	0.608	0.584
Shenzhen	0.574	0.566	0.58	0.577	0.598
Changchun	0.498	0.497	0.549	0.536	0.549
Lanzhou	0.456	0.476	0.506	0.534	0.547
Shenyang	0.471	0.481	0.499	0.518	0.516
Xi’an	0.497	0.461	0.49	0.505	0.523
Chengdu	0.445	0.446	0.471	0.505	0.575
Zhengzhou	0.425	0.426	0.478	0.496	0.554
Guiyang	0.504	0.446	0.439	0.484	0.479
Urumqi	0.487	0.473	0.46	0.468	0.459
Tianjin	0.488	0.456	0.472	0.476	0.449
Wuhan	0.468	0.453	0.467	0.471	0.437
Dalian	0.411	0.439	0.458	0.481	0.489
Hangzhou	0.451	0.438	0.469	0.432	0.469
Qingdao	0.382	0.392	0.446	0.443	0.455
Hefei	0.414	0.4	0.421	0.42	0.426
Xiamen	0.41	0.34	0.407	0.434	0.461
Taiyuan	0.437	0.393	0.412	0.405	0.389
Changsha	0.395	0.373	0.378	0.402	0.43
Nanjing	0.391	0.382	0.379	0.4	0.411
Kunming	0.385	0.348	0.402	0.399	0.43
Jinan	0.373	0.341	0.383	0.417	0.441
Xining	0.377	0.357	0.384	0.387	0.427
Yinchuan	0.361	0.386	0.355	0.372	0.418
Fuzhou	0.365	0.322	0.326	0.355	0.399
Hohhot,	0.333	0.345	0.356	0.367	0.339
Ningbo	0.351	0.32	0.316	0.348	0.378
Nanning	0.339	0.346	0.349	0.305	0.341
Nanchang	0.306	0.337	0.301	0.375	0.347
Lhasa	0.296	0.276	0.315	0.326	0.351
Shijiazhuang	0.326	0.289	0.303	0.315	0.307
Haikou	0.32	0.273	0.257	0.268	0.277

,

Table A3. Coupling coordination degree among B-R in 36 cities in China. Note: “-” represents no final data result for the corresponding cities.

	2016	2017	2018	2019	2020
Beijing	0.769	0.752	0.773	0.78	0.751
Guangzhou	0.765	0.749	0.746	0.775	0.778
Shanghai	0.756	0.734	0.759	0.771	0.78
Shenzhen	0.638	0.615	0.616	0.618	0.651
Chongqing	0.617	0.571	0.604	0.623	0.642
Chengdu	0.504	0.523	0.579	0.606	0.669
Wuhan	0.511	0.506	0.544	0.546	0.494
Xi’an	0.494	0.478	0.53	0.531	0.543
Nanjing	0.49	0.494	0.531	0.521	0.522
Changchun	0.431	0.464	0.522	0.515	0.522
Tianjin	0.487	0.462	0.494	0.491	0.487
Hangzhou	0.463	0.452	0.496	0.472	0.513
Dalian	0.454	0.478	0.482	0.48	0.485
Zhengzhou	0.395	0.403	0.484	0.497	0.521
Shenyang	0.414	0.412	0.457	0.464	0.476
Changsha	0.399	0.391	0.399	0.423	0.46
Kunming	0.373	0.388	0.421	0.411	0.422
Harbin	0.393	0.41	0.423	0.409	0.382
Qingdao	0.32	0.36	0.441	0.421	0.437
Ningbo	0.419	0.372	0.385	0.393	0.4
Hefei	0.313	0.336	0.386	0.394	0.399
Fuzhou	0.286	0.289	0.31	0.34	0.35
Nanning	0.313	0.321	0.357	0.357	0.369
Nanchang	0.298	0.35	0.328	0.374	0.363
Xiamen	-	0.37	0.339	0.361	0.388
Guiyang	-	0.296	0.336	0.357	0.343
Shijiazhuang	-	0.28	0.302	0.303	0.298
Urumqi	-	-	0.328	0.324	0.303
Lanzhou	-	-	-	0.366	0.385
Hohhot,	-	-	-	0.316	0.299
Jinan	-	-	-	0.317	0.315
Taiyuan	-	-	-	-	0.265
Haikou	-	-	-	-	-
Lhasa	-	-	-	-	-
Xining	-	-	-	-	-
Yinchuan	-	-	-	-	-

and

Table A4. Coupling coordination relationship among R-T in 36 cities in China. Note: “-” represents no final data result for the corresponding cities.

	2016	2017	2018	2019	2020
Beijing	0.76	0.766	0.75	0.764	0.752
Shanghai	0.75	0.757	0.734	0.743	0.759
Guangzhou	0.629	0.645	0.631	0.682	0.699
Shenzhen	0.578	0.622	0.642	0.631	0.668
Chongqing	0.547	0.504	0.522	0.535	0.585
Wuhan	0.485	0.505	0.517	0.524	0.508
Changchun	0.436	0.503	0.508	0.521	0.539
Chengdu	0.419	0.443	0.479	0.508	0.606
Xi’an	0.431	0.449	0.467	0.486	0.524
Nanjing	0.46	0.488	0.443	0.459	0.476
Tianjin	0.452	0.453	0.469	0.455	0.431
Dalian	0.412	0.43	0.408	0.407	0.414
Hangzhou	0.371	0.387	0.412	0.417	0.474
Shenyang	0.38	0.382	0.394	0.396	0.419
Changsha	0.34	0.36	0.347	0.377	0.456
Zhengzhou	0.301	0.348	0.345	0.376	0.442
Harbin	0.335	0.358	0.355	0.344	0.338
Ningbo	0.352	0.325	0.31	0.325	0.347
Kunming	0.284	0.31	0.353	0.328	0.372
Hefei	0.273	0.303	0.331	0.335	0.377
Qingdao	0.247	0.291	0.343	0.34	0.369
Nanning	0.246	0.264	0.36	0.312	0.351
Nanchang	0.282	0.288	0.308	0.317	0.326
Fuzhou	0.232	0.259	0.256	0.29	0.309
Xiamen	-	0.3	0.268	0.295	0.325
Shijiazhuang	-	0.279	0.306	0.273	0.295
Guiyang	-	0.231	0.257	0.292	0.264
Urumqi	-	-	0.261	0.262	0.248
Lanzhou	-	-	-	0.339	0.337
Hohhot,	-	-	-	0.233	0.264
Jinan	-	-	-	0.234	0.238
Taiyuan	-	-	-	-	0.265
Haikou	-	-	-	-	-
Lhasa	-	-	-	-	-
Xining	-	-	-	-	-
Yinchuan	-	-	-	-	-

). Figure 5 shows the coupling coordination degree for B-R-T and its average growth rate for 31 of the major cities in China from 2016 to 2020. Just 19.35% of cities were in the balanced development stage, while 80.65% were in the unbalanced development stage. On the whole, there is uncoordinated development between buses, rail transit, and taxis in the major cities of China.

Beijing, Shanghai, Guangzhou, Shenzhen, Chongqing, Chengdu, and Changchun were all in the first quadrant. During 2016–2020, the coupling coordination of B-R-T in Shenzhen, Guangzhou, Shanghai, and Beijing fell into favorably balanced development. Relying on a favorable policy support environment, Shenzhen, Guangzhou, Shanghai, and Beijing implemented the policy of “giving priority to the development of public transport” [22]. They were at the forefront of the country in terms of buses, rail transit vehicle configuration, line number development, and other aspects, and optimized the delivery and use of new-energy taxis. Moreover, they performed well in terms of bus, rail transit and taxi operating mileage, passenger capacity, and other aspects, thus achieving good coordinated development of their public transport. For example, the bus operating line length, rail transit operating line length, and number of taxis in Shenzhen in 2020 were 1.99 times, 2.06 times, and 1.50 times the average levels, respectively. However, the coupling coordination relationships among B-R-T in Chongqing, Chengdu, and Changchun were all at the level of barely balanced development. The average growth rate of the coupling coordination degree of B-R-T in Chengdu was high, at 7.98%. Chongqing, Chengdu, and Changchun still need to improve their construction of public transport infrastructure and expand their public transport operations so as to enhance and develop this coupling coordination relationship.

The proportion of cities in the second quadrant was 61.29%. The coupling coordination degree of B-R-T in these cities was in the stage of unbalanced development. However, their average growth rates were greater than zero. Among them, the average growth rates of the coupling coordination degree of B-R-T in Zhengzhou, Qingdao, and Xiamen were 7.94%, 7.59%, and 5.14%, respectively. The average growth rates of the coupling coordination degree of B-R-T in Dalian, Nanjing, and Ningbo were 2.12%, 1.28%, and 0.33% respectively. There is still room for improvement in the public transport infrastructure index and the operating index.

Wuhan, Tianjin, Harbin, Hohhot, and Urumqi were in the third quadrant, where the coupling coordination relationship of B-R-T was in the stage of balanced development, and the average growth rate was less than 0. Because of the impact of COVID-19, the coupling coordination score for B-R-T in Wuhan decreased in 2020. Due to the lagging development level of taxis in Tianjin, the coupling coordination relationship of B-R-T was poor. Hohhot and Urumqi were slow in their development due to the late opening of rail transit transportation in those cities, which affected the coordinated development of public transportation. Local transportation departments should pay attention to these shortcomings in the development of public transportation in these cities.

Figure 6 shows the mean and the average growth rate of the coupling coordination degree of B-T in the 36 cities from 2016 to 2020. On the whole, this relationship was inferior in the 36 major cities, with 77.8% of the cities showing unbalanced development. Furthermore, the relationship was unevenly developed in each region. Chongqing, Guangzhou, Beijing, Shanghai, Shenzhen, Changchun, and Lanzhou were in the first quadrant, with coupling coordination relationships for B-T higher than the average level. Among them, Chongqing, Guangzhou, Beijing, and Shanghai showed favorably balanced development. The average growth rate of the coupling coordination of B-T in Chongqing was the highest, at 4.96%. This better coordination relationship may be closely related to its mature policy support environment. For example, in the Outline of Chongqing’s Comprehensive Three-dimensional Transportation Network Planning (2021-2035) [64], it was proposed that, by 2035, Chongqing should form a “123 travel traffic circle” (1 means one hour traffic circle, 2 means two hours of free travel within the city area, and 3 means three hours of coverage of major cities in China), to build a strong traffic situation in the city. In Beijing, Shanghai, and Guangzhou, the bus operating line length, operating lines, station facilities, and other construction levels were all acceptable. At the same time, new-energy taxis proliferated. Beijing, Shanghai, and Guangzhou were leading the way in terms of bus and taxi passenger capacity and mileage.

There were more cities in the second and third quadrants, accounting for 77.7% of the total, and here the coupling coordination degree of B-T was in the stage of unbalanced development. However, for the cities in the second quadrant, the average growth rate of the coupling coordination degree of B-T was greater than 0. This indicated an upward trend year by year. For example, Zhengzhou and Chengdu had changed from slightly balanced development to barely balanced development, showing a trend of continuous positive development. The average growth rates of the coupling coordination degree of B-T in Zhengzhou and Chengdu were 6.98% and 6.73%, respectively. This was mainly due to their response to the policy of “priority development of urban public transport”, in which they increased the construction of public transport stations and supporting facilities, such that the development level of buses was growing at a faster rate. In Nanning and Hohhot, the average growth rates of the coupling coordination degree of B-T were also relatively low, at 0.53% and 0.56%, respectively. For the cities in the third quadrant, the average growth rate of the coupling coordination degree of B-T was less than 0. Among these cities, Wuhan was an important core city hit by the COVID-19 epidemic in 2020, which seriously affected the growth of indicators such as bus passenger capacity, operating mileage, and average annual operating mileage. Compared with 2019, the passenger capacity and operating mileage of buses decreased by 58.7% and 40.4%, respectively. At the same time, the passenger capacity and operating mileage of Wuhan taxis decreased by 39.4% and 34%, respectively, compared with 2019, leading to the uncoordinated development of buses and taxis.

Harbin was in the fourth quadrant, with a coupling coordination relationship of B-T of barely balanced development from 2016 to 2020. Harbin had good performance in terms of the number of bus operating lines, bus passenger capacity, taxi passenger capacity, taxi mileage utilization rate, and other indicators. Simultaneously, it had a good resource endowment and initially formed a transportation network in the urban circle. However, the average growth rate of the coupling coordination degree of B-T in Harbin was less than 0, at −0.65%. Similarly, the bus mileage and passenger volume decreased by 22.5% and 54.3%, respectively, compared with 2019, due to the impact of COVID-19 in 2020, and the development level of buses showed a downward trend, further affecting the coordinated development between buses and taxis.

As of the end of 2020, Haikou, Lhasa, Xining, and Yinchuan had not yet put rail transit into operation, and Taiyuan only did so in 2020. Thus, there is no growth rate data for these cities. Therefore, Figure 7 shows the means and average growth rates of the coupling coordination degrees of B-R in 31 cities in China during 2016–2020. Generally, 77.4% of these cities showed balanced development, while 22.6% showed unbalanced development.

Shanghai, Guangzhou, Chongqing, Shenzhen, Xi'an, Nanjing, and Chengdu were in the first quadrant. In terms of the coupling coordination relationship of B-R, Shanghai, Guangzhou, Chongqing, and Shenzhen showed favorably balanced development. Xi’an, Nanjing, and Chengdu showed barely balanced development. For example, in 2020, the lengths of the bus operating lines and rail transit operating lines in Chengdu were 2.13 times and 6.18 times those in 2016, respectively. In addition, Shanghai, Guangzhou, Chongqing, and Shenzhen had good economic endowments and advantages in terms of the allocation of resources to, construction of infrastructure for, and operation of urban bus and rail transit, and so the coordinated development of B-R was in a good state. There were more cities in the second quadrant, accounting for 54.8% of all cities. Here the coupling coordination degree of B-R was in the stage of unbalanced development. Shijiazhuang, Xiamen, and Guiyang began rail transit operations in 2017, and Lanzhou in 2019. This relatively short time of operation meant that the infrastructure construction was not perfect, and the development level of rail transit was poor, resulting in a state of unbalanced development between buses and rail transit. However, the average growth rate of the coupling coordination degree of B-R in these cities was greater than 0, and the coordination relationship of B-R continued to develop in a good direction. For example, the average growth rates of the coupling coordination degree of B-R in Qingdao, Zhengzhou, and Hefei were high from 2016 to 2020, at 8.57%, 7.41%, and 6.39%, respectively. The coordinated development of B-R showed an obvious upward trend. Among these, the coordination relationship of B-R in Zhengzhou presented the following characteristics: moderately unbalanced development in 2016, slightly unbalanced development during 2017-2019, and barely balanced development in 2020.

Jinan, Ningbo, Harbin, Urumqi, and Hohhot were in the third quadrant. The coupling coordination of B-R showed unbalanced development in these cities, and their average growth rates of coupling coordination were less than zero. Since Urumqi, Jinan, and Hohhot had only opened rail transit operations in 2018 and 2019, respectively, there were still some deficiencies in construction in terms of rail line length and vehicle configuration. Wuhan and Beijing were in the fourth quadrant, with barely balanced development and favorably balanced development in their B-R coupling coordination, respectively. Compared with 2016, the buses’ operating line length and rail transit’s operating line length in Wuhan had increased by 70.09% and by 2.15 times, respectively, by 2020. In 2020, Beijing ranked first among the 36 cities in terms of bus lane length and passenger capacity, and its number of rail transit stations and number of transfer stations were 3.5 times and 5.1 times the average level, respectively, showing a good level of bus and rail transit infrastructure construction. However, the average growth rates of the coupling coordination of B-R in Wuhan and Beijing were less than 0. One of the reasons for this was that the number of rail transit inbound stops and the passenger traffic in Wuhan and Beijing had decreased by 49.27% and 49.27%, and 42.03% and 42.02%, respectively, in 2020 compared with 2019.

Figure 8 shows the means and average growth rates of the coupling coordination degrees of R-T in 31 major cities in China from 2016 to 2020. Overall, 77.42% of the cities showed unbalanced development of R-T, reflecting a poor condition.

Shanghai, Guangzhou, Shenzhen, Chongqing, Wuhan, and Changchun were in the first quadrant, showing balanced development, and an average growth rate in the coupling coordination degree greater than 0. The coupling coordination relationship of R-T in Shanghai, Guangzhou, and Shenzhen was in the stage of favorably balanced development from 2016 to 2020. These cities had good economic and resource endowments, advantages, and good performance in terms of rail transit infrastructure construction, rail transit operations, the launch and use of taxi vehicles, taxi operations, and other aspects. For example, Shanghai ranked first among the cities in terms of the rail transit lines in operation, number of stations, number of assigned vehicles, passenger capacity, passenger-kilometers, and other indicators in 2020. Moreover, the mileage and passenger capacity of taxis in Shanghai were 2.9 times and 2.04 times the average levels, respectively. The coupling coordination relationships of R-T in Wuhan, Chongqing, and Changchun showed barely balanced development from 2016 to 2020. Among them, the average growth rate of the coupling coordination degree of R-T in Changchun was high, at 5.59%. The relationship changed from slightly unbalanced development in 2016 to barely balanced development from 2017 to 2020, showing an obvious trend of improvement.

There were 64.52% cities in the second quadrant. These cities were in the stage of unbalanced development. Specifically, Guiyang opened its rail transit service in 2018, and the overall infrastructure and operating level were poor. Moreover, the number of battery-powered electric taxis in Guiyang was 0 in 2020, and investment in new-energy vehicles was less. These are key factors affecting the coordinated development of R-T. However, in the second quadrant, the average growth rate of the coupling coordination degree of R-T was greater than 0. Among the cities in that quadrant, Hohhot, Qingdao, and Nanning had higher average growth rates of the coupling coordination degree of R-T, at 13.30%, 10.83%, and 10.71%, respectively. The average growth rates of the coupling coordination degree of R-T in Harbin, Dalian, and Nanjing were slightly greater than 0, but still indicated relatively slow growth. In the third quadrant, the coupling coordination relationship of R-T was in the stage of unbalanced development and the average growth rate was less than 0. Due to the low level of overall development of rail transit in Urumqi and Lanzhou, which only opened in 2018 and 2019, respectively, the development of local rail transit and taxis was not coordinated.

Finally, Beijing was in the fourth quadrant, with a coupling coordination relationship of R-T implying favorably balanced development from 2016 to 2020. In 2020, the rail transit operating line length in Beijing was 1.39 times that in 2016, and the number of taxis per 10,000 people in the city was 1.7 times the average level. The overall development trend of R-T was fine. However, the average growth rate in the coupling coordination degree of R-T in Beijing was less than 0. This was due to the impact of COVID-19, which meant that the comprehensive score for Beijing's rail transit development level in 2020 had decreased by 9.92% compared with that in 2019, mainly due to the rail transit operating indicators. Relevant departments should focus on the impact of emergencies on transport operations.

6. Discussion and Conclusions

It is of great practical significance for the sustainable development of urban resources, the environment, and the economy to scientifically understand the development level of urban multimodal transportation and clarify the coordination of the development of the different modes of urban transportation. In this paper, 60 factors affecting the development level of urban multimodal transportation were selected, including 22 second-class indexes of the bus development level, 20 for rail transit, and 18 for taxis. Then, the comprehensive development level of urban multimodal transportation was calculated. Based on the CCDM, we studied the coordinated development relationships among buses, rail transit, and taxis, and analyzed the differences in these across different cities.

Firstly, overall, the comprehensive development level of China's urban multimodal transportation showed a significant upward trend from 2016 to 2020, with an average annual growth rate of about 7.36%, indicating that the development of China's urban multimodal transportation is in a good stage of improvement. We concluded that the development level of buses was better than that of rail transit, which was better than that of taxis, in China's major cities from 2016 to 2020. Secondly, from the perspective of the development level of urban multimodal transportation, there were significant differences in the development levels of multimodal transportation in different cities. Cities with better development levels of buses, rail transit, and taxis were mostly big cities and first-tier cities (e.g., Beijing, Shanghai, Guangzhou, Shenzhen), while cities with poorer development levels were mostly third- and fourth-tier cities (e.g., Haikou, Lhasa). This is consistent with the findings of Fitzová et al. [65] that in Czech Republic, big cities with high population density had more efficient public transportation operations than small cities. For example, the cities with more efficient transport systems were Prague and Brno, which are two big cities in Czech Republic. Gruyter et al. [28] also found that some cities (e.g., Ankara, Budapest, Cracow) in Eastern Europe had a higher level of public transport development compared to cities (e.g., Bogota, Brasilia) in Latin America. Ye et al. [66] also found that the construction and operation efficiency of rail transit in the first-tier cities of Beijing, Shanghai, and Guangzhou were good. Finally, from the perspective of coordinated relationship, buses, rail transit, and taxis show uncoordinated development in the major cities of China. For example, in terms of the coupling coordination of B-R-T, 80.65% of cities were in the stage of unbalanced development. In terms of the coupling coordination relationships of B-T, B-R, and R-T, unbalanced development was found in 77.8%, 77.4%, and 77.42% of cities, respectively. Only the coupling coordination relationship of B-R-T, B-T, B-R, and R-T in the first-tier cities (Beijing, Shanghai, Guangzhou, Shenzhen) were in a balanced development stage.

To summarize, we propose that the development levels and coordinated development relationships of multimodal transportation in China be improved in terms of the following aspects. Firstly, the large differences among urban buses, rail transit, and taxis in China were mainly reflected in traffic infrastructure and operating conditions. Therefore, the traffic departments need to balance the local economic resources against the improvements needed, and adjust the measures taken to address local conditions and the local traffic, so as to achieve reasonable public transport planning to meet the diverse travel needs. For cities with a poor development level of multimodal transportation, optimization of network design and vehicle dispatching can be adopted to improve the level of transportation services and competitiveness; for cities with a good development level of multimodal transportation, operators need to improve management levels and introduce advanced technologies to provide passengers with a better travel experience, leading to an increase in passenger flow [67]. Secondly, China’s major cities showed an uncoordinated development between urban multimodal transportation. Coordinating the relationship among buses, rail transit, and taxis is conducive to improving urban capacity, alleviating traffic congestion, and promoting sustainable urban development. Cities need to identify the specific problems, analyze their transportation infrastructure and operational indexes, determine the restrictive factors, and strengthen their infrastructure. Thirdly, in the face of emergencies (such as COVID-19), the impact on different transportation modes was not consistent. For example, public transportation was most affected by the epidemic shock [68]. Therefore, transportation departments should make planning adjustments in accordance with national policies and specific conditions, so as to stabilize the development level of the urban multimodal transportation. For example, the government implements targeted traffic interventions to promote the perceived safety of passengers using public transportation after an emergency, thereby maintaining a positive image of public transportation [69].

There are some limitations in the research process of this paper, which can be resolved in future work. Firstly, we only focused on three transportation modes: buses, rail transit, and taxis to explore the development level of multimodal transportation and its coordinated development relationships, without considering private cars, bicycles, walking, and other transportation modes. Future work should consider as many diverse modes of transportation as possible. Secondly, we constructed the evaluation index system for the development level of urban multimodal transportation based on the infrastructure status and operational indexes in China’s major cities. We mainly chose the indexes by considering the internal factors of transportation modes; however, different urban multimodal transportation development levels will be subject to the external environment, such as local area condition, the urban development environment, capital investment, land use guarantees, and policy environment. Future research should consider establishing evaluation index systems accounting for regional differences. In addition, we could further explore the factors influencing the coordination relationships of multimodal transportation in cities in China in future study.