EN中文

Journal of Resources and Ecology >

2025 , Vol. 16 >Issue 1: 265 - 282

DOI: https://doi.org/10.5814/j.issn.1674-764x.2025.01.024

Ecotourism

Spatio-temporal Changes and Influencing Factors of the Travel Network on China’s National Day Holiday under COVID-19

  • DONG Yaojia , 1, 2, 3 ,
  • WANG Fuyuan , 1, 2, * ,
  • WANG Kaiyong 1, 2
Expand
  • 1. Institute of Geographic Sciences and Natural Resources Research, Key Laboratory of Regional Sustainable Development Modeling, Chinese Academy of Sciences, Beijing 100101, China
  • 2. China Association for the Promotion of Administrative Divisions and Regional Development, Beijing 100044, China
  • 3. College of Resources and Environment, University of Chinese Academy of Sciences, Beijing 100049, China
*WANG Fuyuan, E-mail:

DONG Yaojia, E-mail:

Received date: 2024-05-13

  Accepted date: 2024-09-12

  Online published: 2025-01-21

Supported by

National Natural Science Foundation of China(42271252)

National Natural Science Foundation of China(42230510)

National Natural Science Foundation of China(42371253)

Theme Academic Activities of National Social Science Foundation Academic Societies(22STA021)

Fold

Abstract

COVID-19 has tremendously impacted the travel habits of urban residents. This study used the “node- line segment-network” hierarchy and Baidu migration big data to conduct a comparative analysis of the travel network patterns of urban people on China’s National Day holiday before and during the COVID-19 outbreak (i.e., 2019 and 2021). The results revealed four interesting patterns: (1) In 2021, there was an increase in the aggregation coefficient and access to the intra-provincial linkage network, indicating a higher degree of the travel agglomeration of residents under the pandemic. (2) In 2021, the “hub-and-spoke” pattern on the national scale, the “core-edge” pattern of the intra-provincial scale and the “rhomboidal” structure of the extra-provincial scale were more contracted, aggregated, and low-value. (3) The coverage of the urban advantageous association decreased while the total numbers of urban clusters and single provincial clusters increased, reaching 25 and 16, respectively. This indicates that the pandemic intensified the effect of administrative boundaries as a barrier. (4) The primary determinants of movement during the pandemic were urban competition, policy control, administrative boundary barriers, and the travel intentions of residents.

Key words： resident travel; pandemic; network characteristics; influencing factors; National Day holiday

Cite this article

DONG Yaojia , WANG Fuyuan , WANG Kaiyong . Spatio-temporal Changes and Influencing Factors of the Travel Network on China’s National Day Holiday under COVID-19[J]. Journal of Resources and Ecology, 2025 , 16(1) : 265 -282 . DOI: 10.5814/j.issn.1674-764x.2025.01.024

1 Introduction

With rapid informatization and globalization, many regions, urban clusters, and core urban nodes have been transformed into important flow space units, and the coexistence of local space and flow space has contributed to the transformation of the territorial spatial structure from hierarchical to networked characteristics (Fang, 2021). The “spatial and temporal compression” effect brought on by socio-economic development, transportation, and communication technologies has greatly enhanced the mobility of residents, and led to normalized, scaled, and dynamic travel by urban residents and movement of the population (Li et al, 2020a). Outbreaks of public health incidents affect human spatio-temporal behavior and flow space patterns, causing changes in urban resident travel networks. So it is important to explore the travel of urban resident during holidays under the influence of the pandemic to better understand the spatial movement patterns of urban residents during the holidays, identify the factors influencing the urban resident travel network in China under the pandemic, and further delineate the spatial and temporal heterogeneity of urban flow space during public health emergencies.
Currently, population movement studies based on urban spatial connections are predominantly divided into three distinct categories. First, most studies focus on the aspects of population flow and its spatial distribution from the perspective of population geography. The existing studies have mainly analyzed the population census through the manual collection of data, and explored the spatial pattern characteristics, evolutionary laws, and mechanisms of population flow under inter-provincial (Deng et al., 2014), city-county (Qi et al., 2017) or urban clusters (Ma and Zhang, 2020; Zhou and He, 2022). The second aspect that is being explored is tourism flow networks from the perspective of tourism geography. These studies mostly focus on the spatial association and coupling paths of small-scale territories such as road flows and scenic flows (Cao et al., 2021; Cong et al., 2021), the characteristics of tourism flow networks and the resilience of single or multiple urban clusters (Xu et al., 2018; Cheng et al., 2021), and the exploration of inter-provincial tourism flow networks based on modified gravity models (Zhou and Wang, 2020), which reflect the trends of multiple scales and diversified cases. The third aspect is urban resident travel networks from the perspective of transportation geography. This has recently become a research hotspot due to rapid developments in spatiotemporal big data, technologies, and tools in the era of big data for examining large scale spatial geography and social behavior to reveal the fine-scale individual activity footprints and group travel paths (Wang et al., 2013; Wei et al., 2018; Shen and Chai, 2012; De Montis et al., 2011). The spatiotemporal patterns of population movement around different means of transportation, such as air (Jin, 2001), rail (Dai et al., 2005; Jiao et al., 2016), and road (Chen et al., 2015), in addition to specific travel times, can facilitate the interpretation of the flow pattern and urban network structure, thereby yielding better results.
The impact of public emergencies on population mobility is a hot topic of research. Numerous studies have shown that major public health events, like the spread of various infectious diseases, interact very closely with urban mobility (Sattenspiel and Dietz, 1995; Balcan et al., 2009; Balcan et al., 2010), which in turn affects the processes of urban development and regional integration. From the vantage point of early viruses, the Severe Acute Respiratory Syndrome (SARS) virus was effectively contained through government-imposed isolation measures (Peng et al., 2003; Twu et al., 2003). However, the significant reduction in travel volumes adversely affected the Chinese tourism industry (Zhu et al., 2003). The distinct spatial and temporal characteristics, as well as the transmission patterns of SARS, were studied in Beijing (Cao et al., 2010) and Guangzhou (Cao et al., 2008). A spatial display model was subsequently developed to depict the H1N1 influenza transmission network among urbanized regions, with the aim of enhancing resilience against urban public health threats (Mao and Bian, 2010). From a pandemic perspective, COVID-19 has profoundly influenced economic and social development since 2020, notably altering the travel behavior of residents (Wang et al., 2022). In terms of the spatial pattern, Liu et al. (2020) investigated the dynamic distribution and prediction of migrating populations at various scales using multi-source geo-temporal big data, with a focus on the migrating population from Wuhan during the pandemic as a case study. Regarding trigger conditions, Farzanegan et al. (2021) discovered that countries with high international tourism flows tended to have greater numbers of confirmed COVID-19 cases and an increasing number of deaths. In terms of influencing factors, Zhao et al. (2022) identified short-term shifts in the spatial and temporal distribution patterns of urban populations by assessing the impact of population movement control as a policy measure to curb the novel coronavirus pandemic. In terms of prevention, control and recovery, Xue et al. (2020) summarized the positive role of geography in decision making, planning, and emergency responses using technologies such as big data and Location-Based Services (LBS) for early prevention and control of the pandemic. Athanasopoulos et al. (2023) proposed a new scenario probability statistical method for predicting the tourism recovery path after a pandemic based on a large survey of tourism experts.
Overall, the existing studies on the impacts of disturbances caused by public health emergencies on the travel of urban residents are gradually increasing in number, but these studies have mostly explored urban clusters and individual cities. To the best of our knowledge, urban resident travel networks and their characteristics during pandemics have not yet been explored from a national multi-scale perspective. There is also a need to assess the influencing factors both qualitatively and quantitatively. In addition, studies on the travel of urban residents during specific periods using big data have predominantly concentrated on the Spring Festival period. There is limited research on other holiday periods, such as the National Day holiday, within the context of the ongoing pandemic. Moreover, there are no comprehensive studies that encompass the overall movement patterns at the national level, or the movements within and between provinces and cities. To address this knowledge gap, the popular migration data of residents were procured from the Baidu Maps platform for the National Day holidays in two years (October 1-7, 2019, and 2021). The data were divided into overall, intra-provincial linkage, extra-provincial linkage, and urban clusters to identify the urban resident travel networks. Then, a research path was established from the overall pattern, the multi-dimensional spatial characteristics were examined and an impact factor analysis was performed. This analysis path compared the changes in national node attributes, scale structure, and urban network connections before and after the pandemic, based on the logical relationships of nodes (cities), lines (city associations), and groups (regions). In addition, the factors that brought about the changes in mass travel were analyzed.

2 Methods, variable explanations, and data sources

2.1 Methods

Using the relatively mature “S-dimensions” index method proposed by N. Limtanakool, the study focused on network equilibrium, relative advantages, and symmetry by examining three dimensions: network strength, structure, and symmetry.

2.1.1 Node structure and a potential measurement index

To reveal the characteristics of the city hierarchy in the network as a whole, this study introduces the concept of entropy value (EI), which characterizes the degree of urban equilibrium, and it was calculated as:
$E I=-\sum_{i=1}^{I} \frac{Z_{i} \ln Z_{i}}{\ln I}$
where Zi denotes the ratio of the linkage intensity of city i to the sum of the linkage intensities of all the cities in the network. I denotes the number of cities in the network. The value of EI is from zero to one. When EI is zero, the difference in city hierarchy in the network is the largest; and when the value increases infinitely close to one, then all of the links are becoming equal in strength, and the network tends to be balanced. This index is generally used to illustrate the level of difference in the city hierarchy in the whole urban resident travel network.
To reveal the node’s (city’s) position potential in the network at a fixed point, the Weighted Degree Centrality (WDCi) (Li et al., 2020a) and Weighted Dominance Degree (DITi) (Yang and Xie, 2020) were chosen as the two indicators, and they were calculated according to formula (2) and formula (3):
$W D C_{i}=D_{i}^{\alpha} \times\left(T_{i} / D_{i}\right)^{(1-\alpha)}$
where Di is the total degree value of city i, which is defined as the sum of the outgoing and incoming degrees of city i in the network; α is the assignment parameter, which was assigned a value of 0.5; and Ti is the sum of the inflow and outflow population scales of city i.
$D I T_{i}=\frac{T_{i}}{\sum_{j=1}^{J} \frac{T_{j}}{J}}$
where J is the total number of cities in the network (n=366), and since the direction is considered, i≠j. WDCi refers to the absolute dominance of city i in the network, and the larger the value, the higher the absolute dominance level of city i in the whole network. DITi refers to the relative dominance of city i in the network, which characterizes the level of the relative dominance of each city in the whole network, and the values range is [0, ∞). A value of 0 represents no dominance, and when DITi is greater than one, then the dominance of the city is higher than the average of the travel network.

2.1.2 City linkage strength measures

According to Yang et al. (2018), city linkage strength mainly characterizes the importance of any pair of city-linked edges in the network from the edge perspective, and it was calculated as:
$R S L_{i j}=\frac{t_{j j}}{\sum_{i=1}^{I} \sum_{j=1}^{J} t_{i j}}$
where tij is the total scale of urban resident travel between cities i and j, and i≠j. RSLij represents the dominance of the city linkage, which characterizes the proportion of the strength of a linkage between two cities to the total strength of linkages in the whole travel network. Its value ranges from zero to one, and the closer to one, the higher the proportion of the total linkage between the pairs of city i and city j, and higher the dominance.

2.1.3 Identifying the structure of the urban cluster

An association detection and identification algorithm was designed based on edge weights, direction, and modularity optimization. Gephi was used to measure the local clustering characteristics of the National Day holiday in 2019 and 2021 (Meo et al., 2008), which determined the distribution of the spatial organization structure of the national travel network. The higher the value of “modularity”, the clearer the structure of the urban cluster.

2.2 Variable descriptions and model selection

2.2.1 Explanatory variable selection and descriptions

This study selected the inter-city inflow scale index during the National Day holiday in 2021 as the dependent variable. According to the classical gravity model, the strength of inter-city association is related to the attractiveness of the destination city (such as economic scale, population size, etc.) and the cost of the inter-city association (such as transportation connectivity, travel time, etc.). Referring to relevant research (Zhang et al., 2020), this study selected 11 types of explanatory variables from three major categories of data (urban economy, population, and connectivity) to construct an impact index system for inter-city travel. Among them, GDP and per capita GDP are used to characterize the scale of urban economic development; urbanization rate and the proportion of tertiary industry represent the scale of urbanization; permanent population and total number of employees represent the scale of the urban population and employment positions; richness of tourism resources represents the attractiveness of the cities to tourists during long holidays; and administrative level indicates the administrative resources of the cities. The factors of transportation connectivity include whether there is an airport, whether there is a high-speed rail, and highway passenger volume as three variables representing the impacts of aviation, high-speed rail, and highways on inter-city travel, respectively. To eliminate dimensional differences, all variables except for the dummy variables, ordinal variables, and proportional variables were log-transformed (Table 1).
Table 1 Selection of variables and descriptive statistics for the factors influencing intercity travel
Variable
type		 Variable name Variable description Average
value		 Standard
error
Dependent variable Intercity travel scale Intercity inflow scale index (ln) 12.61 0.92
Independent variable GDP Regional gross domestic product (ln) 7.57 0.99
GDP per capita Per capita GDP (ln) 10.94 0.60
Population End-of-year resident population (ln) 447.68 387.48
Practitioners Total scale of employment (ln) 12.72 0.92
Urbanization rate Proportion of urban population to total population (%) 0.62 0.14
Proportion of tertiary industry Proportion of the tertiary industry’s output value to the Gross Domestic Product (%) 49.31 8.02
Abundance of tourism resources Tourism resource abundance degree (0, 1) based on the natural breaks method for the different levels 0.38 0.24
Aviation Division characterization, whether there is a flight (0, 1) 0.58 0.49
High-speed rail Whether high-speed rail is available (0, 1) 0.75 0.43
Highway Highway passenger traffic volume (ln) 7.29 1.20
Administrative hierarchy Administrative level of cities (municipalities directly under the Central Government =5, sub-provincial/ separately planned cities=4, provincial capitals=3, general prefecture-level cities=2, cities/counties directly governed by provinces=1) 2.14 0.55

2.2.2 Determination of the influencing factors

Model setting. Spatial econometric models are often used to explore the factors influencing the travel of residents, among which the Spatial Lag Model (SLM) and Spatial Error Model (SEM) (Zhao et al., 2022) are more suitable for studying spatial cross-sectional data. The SLM model has a spatial transmission mechanism for the local and neighboring explanatory variables, which generates a technology diffusion effect, and it was calculated as:
$Y=\rho W Y+X \beta+\varepsilon, \quad \varepsilon \sim N\left[\sigma^{2} I\right]$
where X is the (n×k)-dimensional matrix of explanatory variables; Y is the explanatory variable; W is the spatial weight matrix; and ρ is the spatial autoregressive coefficient. If the coefficient is significant, then there is a spatial correlation between the explanatory variables. β is the matrix of parameters to be estimated, and ε is the matrix of random disturbance terms.
The SEM model represents a spatial transmission mechanism in the presence of neighboring ground error terms, and the explanatory variables were calculated as:
$Y=X \beta+\mu, \quad \mu=\rho W \mu+\varepsilon, \quad \varepsilon \sim N\left[\sigma^{2} I\right]$
where μ is the error term matrix and the meanings of other indicators are the same as in formula (5).

2.3 Data sources

The data used in this study were derived from the “Baidu Migration” data available in Baidu Maps (Yang and Xie, 2020). The “Baidu Migration” is based on location-based service (LBS) big data for calculation and analysis, and it is presented in the Baidu Migration view as the percentage of the population moving into and out of the cities. The data illustrate the integration of the spatial nodes of population movement at different time scales and the intensity and directionality of movement between different nodes. This data source has several advantages. First, it is more suitable to use cities as nodal units for larger-scale population flow research and urban resident travel network analysis. Second, accurate and real-time population flow is one of the core elements that characterize the urban network structure. Third, with GPS positioning and trajectory tracking technology, “Baidu Migration” big data are characterized by clarity, simple collection, and fast updating, so it has better timeliness compared to the population data represented in traditional statistical yearbooks.

3 Spatio-temporal variations in the urban resident travel networks in urban agglomerations

3.1 Urban nodes

3.1.1 Node level and travel size structure

By calculating the out-degrees and in-degrees of 366 cities on the National Day holiday, the mean value of network degree centrality at the national scale in 2021 decreased by 18.5% compared to that in 2019, reaching only 290.31. This indicated that the pandemic was relatively consistent, although the average number of directly associated cities for each city decreased if the weighting was not considered. The SPSS correlation test revealed that the correlation coefficient (R²) exceeded 0.9 in both 2019 and 2021, indicating that the degrees of outbound and inbound travel of residents between the cities of the nation were significantly and positively correlated both before and after the pandemic. Although the pandemic led to a reduction in high-frequency contacts, it did not stop the flow of residents during the National Day holiday, and the relatively balanced inflows and outflows directly indicated that high-frequency or low-frequency flows were maintained between the cities during the National Day holiday.
As for the entropy value, the ODL entropy values of the six regions of the country (northeast, north, east, south, central, northwest, and southwest) all showed increasing trends from 2019 to 2021 (Figure 1), indicating that the balance of the urban resident travel network was further strengthened despite the reduced scale of travel between cities. This was especially notable with the Beijing-Shanghai High-Speed Railway and Beijing-Guangzhou High-Speed Railway that traverse multiple regions, greatly shortening the spatiotemporal distances between regions. Among them, the national growth rate had the highest rate of change, with an equilibrium coefficient of 0.91, indicating an overall trend toward balance.
Figure 1 Comparison of the urban resident travel network structure indicators between the six major regions of China on the National Day holidays in 2019 and 2021
When considering the weights, the WDC of the whole nation and each of the six regions before and after the pandemic were heterogeneous with respect to their evolution in terms of time and spatial distribution. The rates of decrease in WDC were in the order of: East China (92%) > South China (91%) > North China (57%) > China (53%) > Central China (44%) > Northeast China (29%) > Southwest China (13%), which meant that the stronger the competitiveness of a region, the greater the drop and level above the national average. For instance, the average absolute potential decreases were the greatest in regions such as East China and South China. Guangdong Province in South China is a major province for labor export, and it was hit the hardest after the outbreak of the pandemic. In terms of weighted dominance, the DIT values for the national scale and the six regions were higher in 2021 than before the pandemic in 2019, suggesting that the pandemic reduced the travel distance and scale of the National Day holiday, leading residents to prefer traveling within their regions, and thereby enhancing the average relative position of the urban resident travel network. At the same time, the pandemic changed the relative ranking of the regions. The East China region eclipsed the South China region in 2021 to become the foremost economically developed area, with both boasting robust transportation infrastructure. The Central China region also outperformed the national average, ascending to third place. This achievement could be attributed to cities such as Wuhan, Changsha, and Zhengzhou within the Central China region. These cities serve as central transportation hubs, with their high-speed railways and highway networks undergoing continuous enhancement. This progress facilitates short-distance travel within and between the Central China region and its surrounding areas, aligning with post-pandemic travel trends. The relative positions of Northwest, Northeast and Southwest China were lower than the national weighted dominance average before and after the pandemic, because they contain relatively low-grade cities. In addition, the absolute potential rank (WDC) and relative potential rank (DIT) of the core nodes within each region before and after the pandemic still exhibited a positive correlation, but the correlations of changes between WDC and DIT were negative, indicating significant advantages within each region.

3.1.2 Spatial distribution characteristics

The “spatiotemporal compression” effect brought about by the rapid development of transportation networks has led to rapid transformation of the urban nodes from “space of place” to “space of flow”. Considering the national-level perspective (Figure 2), compared to 2019 when only the weighted centralities of mega-cities such as Shenzhen, Beijing, Chengdu, and Shanghai were high, the weighted dominance in 2021 increased significantly in the high-value areas, which are clustered in the core city groups such as the Yangtze River Delta, Pearl River Delta, Beijing-Tianjin- Hebei, Chengdu-Chongqing, middle reaches of the Yangtze River, Guanzhong Plain, and Shandong Peninsula. These areas are expanded in a belt-like manner along the Beijing-Shanghai High-Speed Railway, the Ningbo-Hangzhou High-Speed Railway, the Guangzhou-Dongguan-Shenzhen High-Speed Railway, the Longhai-Lanxin Railway, and the Jiaozhou-Jinan-Kunming Passenger Dedicated Line. The reason for the increase in DIT in most cities was that the spatial mobility range was more extensive before the pandemic, so the relative advantages of the city nodes were less important. After the pandemic, as constrained by the administrative boundaries, the urban network during the National Day holiday often reflected small-scale internal trips. This led to an increase in the DIT of most of the core local cities, such as the city nodes of Yangtze River, Chengdu and Chongqing, while the Guanzhong Plain, Shandong Peninsula, and Central Plains urban cluster showed relatively high increases in their relative advantages, resulting into a denser piecewise cluster distribution (Figure 2c). In 2021, cities such as Chengdu, Changsha, Wuhan, Xi’an, and Zhengzhou, which serve as tourist and leisure destinations as well as transportation hubs, experienced the largest improvements in DIT. Their relative advantages became more prominent, with respective changes of 5.92, 4.24, 4.13, 3.83, and 2.87.
Figure 2 Spatial patterns of weighted node dominance of the urban resident travel networks in China

Note: Hong Kong, Taiwan and Macao were excluded from this study due to difficulties in obtaining data.

3.2 City network linkages

3.2.1 Overall spatial pattern of urban resident travel networks

The network analysis tool Gephi was used to analyze the overall attributes of the urban resident travel networks during six periods and at the three scales of the national, extra-provincial and intra-provincial travel networks during the National Day holiday in 2019 and 2021 (Table 2). The results show significant differences in the travel flow of urban residents at the different scales. The traffic of intra-provincial connections accounted for 67.29% of the national travel network in 2019, which increased to 70.03% in 2021, and the average daily intra-provincial trips during the National Day holiday in 2021 were more than 10 times greater than those of extra-provincial connections. Initially, due to the pandemic, the urban resident travel network in 2021 was mainly characterized by the predominance of short-distance travel from one city to the surrounding neighboring cities. An analysis of the topological properties of the network in each period revealed that the national urban resident travel routes reached 3.7969×108 in 2019 with a network density of 0.197, which declined to 16504 in 2021 due to the pandemic, with a network density of only 0.152, indicating a reduction in the resident travel density. The average value of the national paths in 2021 was slightly higher than that in 2019, indicating that the distances of some paths were on the rise. The average clustering coefficients of the nationwide, inter-provincial, and intra-provincial links in 2021 were higher than those in 2019, with the average clustering coefficient of intra-provincial links in 2021 being as high as 0.844, indicating a higher degree of urban resident travel clustering and closer links between urban nodes after the pandemic.
Table 2 Dimensional attributes of the urban resident travel network on the National Day holiday
Time period Range Average travel length Lines Network density Average path length Average clustering coefficient
2019 National 1973 37969 0.197 1.746 0.678
2021 National 1361 16504 0.152 1.807 0.752
2019 Provincial contact 10400 4847 0.141 2.023 0.736
2021 Provincial contact 6091 2345 0.108 1.924 0.844
2019 Out-of-Province contact 727 33122 0.178 1.671 0.452
2021 Out-of-Province contacts 452 14159 0.139 1.711 0.519
Based on the Natural Breaks method, the scale of travel by urban residents was divided into five levels. The urban resident travel network during the National Day holiday in 2019 (Figure 3a, 3b) displayed an obvious “hub-and-spoke” structure consisting of a star-shaped topology rooted at the Hub connected to the spokes that were responsible for transmitting and managing data of the travel of urban residents. This formed a polycentric pattern of “strong in the east weak in the west, and balanced in the north and south”, but the overall pattern remained stable in 2021. Compared to 2019, the high-value associated flows in 2021 were dominated by shorter-distance trips, and there was a significant contraction in the radial network of major city nodes across the country. The first tier network (red line segment) in 2021 was dominated by regional travel flows such as Guangzhou-Foshan, Xi’an-Xianyang, Shenzhen-Dongguan, Shenzhen-Huizhou, Shanghai-Suzhou, and Beijing-Longfang, but it lacked the 2019 counterparts of Beijing-Shijiazhuang, Chengdu-Chongqing, Guangzhou-Ganzhou, Shanghai- Ningbo, and other longer-distance cross-regional flows, as well as Huizhou-Meizhou, Qingdao-Yantai, Foshan-Yunfu, Wuxi-Taizhou and other long-distance flows between the non-capital cities in the province. The overall expansion pattern of the resident travel network at the second and third levels in 2021 was like a “hub-and-spoke” structure, which was highly overlapping with the second-level network in 2019, indicating that the travel scale and travel area had shrunk significantly due to the pandemic. However, the average daily travel scale was on the rise in some regions, such as some of the fourth-tier networks in Guangxi, Gansu, and Ningxia, and some of the fifth-tier networks in Xinjiang, Qinghai, and Tibet in 2019. All of these increased by one tier in 2021. This pattern showed that during the pandemic period, residents were more inclined to travel within the province, and the number of trips decreased. The resident travel network in some remote areas shifted from a small- world network to a scale-free network.
Figure 3 Spatial patterns of China’s National Day holiday travel networks in 2019 and 2021

Note: Hong Kong, Taiwan, and Macao were excluded from this study due to difficulties in obtaining data.

To further explore the agglomerating effect of the pandemic on the urban resident travel network, it was divided into two dimensions: intra-provincial linkages and extra- provincial linkages. In 2019 and 2021, the extra-provincial linkages (Figures 3c, 3d) exhibited a “rhomboidal structural framework” of “Big Four and Little Two” networks in the high-value linkage flow. An “axis-spoke network” with Beijing, Shanghai, Chengdu-Chongqing, Guangzhou-Shenzhen (as the “Big Four”), and Xi’an, and Wuhan (as the “Little Two”) as the core hubs had formed. The city network association of these six nodes covers almost all 366 urban nodes with a “Pareto distribution”. This is a power-law probability distribution, also referred to as the “80-20” rule, which signifies that 80% of societal wealth is controlled by 20% of the population. Furthermore, this distribution suggests that the urban resident travel network had scale-free characteristics.
Compared to 2019, the feature of “eastern travel network is dense and western travel network is sparse” was more prominent in 2021, and the “rhomboidal structural framework” was highly overlapping with the scope of the travel network in national urban clusters, mostly for short-distance cross-province travel within the Yangtze River Delta and Beijing-Tianjin-Hebei urban clusters. The high-value linkage pairs within the “rhomboidal structural framework” network were significantly reduced, and the reduction in the external linkages in the Hunan Province was most significant. This may be attributed to the large-scale pandemic spillover in Zhangjiajie in August, and the lagging effect of the residents’ willingness to travel reduced the scale of external travel in and around Hunan Province.
Intra-provincial linkage trips in both 2019 and 2021 (Figures 3e, 3f) were centered on major cities across the country, forming a local neighborhood connection network within the urban cluster. This indicated that the city trips with the main purpose of travel and leisure and the short-distance family visits during the National Day holiday mainly occurred within the provincial area radiated by the urban cluster. Compared to 2019, the largest reduction in 2021 was for the scale of intra-provincial travel in the regional core cities, where contact flows related to Nanjing, Yangzhou, Xiamen, and Harbin were significantly reduced and diminished. This could be attributed to the pandemic in these areas in August-September, which led to a reduction in intra-provincial, inter-city travel in the corresponding areas due to strict control measures. At the same time, in the remote provincial areas such as Xinjiang, Tibet, and Qinghai, a significant increase in the high-value flow of the spoke network was observed, mostly for the radiative pull from the core regional cities to the surrounding areas, and the overall travel pattern was more balanced. This pattern reflected a shift in the preference of residents from long-distance travel to short-distance travel due to the enforcement of pandemic control measures.

3.2.2 Analysis of city linkage advantage degree

For comparative analysis, based on the intra- and extra-provincial linkages in 2019 and 2021, the advantageous association (RSL) of the travel of urban residents during the National Day holiday could be divided into four levels. For both the intra-provincial and extra-provincial linkages, the advantageous association (RSL) of the cities during the National Day holiday constituted multi-spoke patterns with the main linkages between the central city in the region and its neighboring cities. When the advantageous association (RSL) level of a city was higher, the node centrality of the spoke pattern was more prominent, and the coverage of the advantageous association of the travel of urban residents during the National Day holiday showed the sequence of national urban cluster > regional urban cluster > local urban cluster (Figure 4).
Figure 4 Spatial patterns of the advantageous association of intra- and extra-provincial city linkages during China’s National Day holiday (levels one to four)

Note: Hong Kong, Taiwan, and Macao were excluded from this study due to difficulties in obtaining data.

From a local perspective, the advantageous association (RSL) of extra-provincial linkages in 2019 reflected a large number of cross-regional trips led by the core cities of national urban clusters and radiating to the surrounding central cities (Figures 4a, 4b). Compared with 2019, the pattern in 2021 mostly belonged to the fourth level of advantageous association (blue line segment) between the nodes of local urban clusters, which were dominated by short-distance travel. Unlike the “diamond-shape” structure of the travel network formed by the advantageous association (RSL) of the first and second levels in 2019, the volume of extra-provincial trips in the first and second levels decreased drastically in 2021, with the advantageous association (RSL) existing only in four regions, namely Beijing, Shanghai, Chengdu-Chongqing, and Guangzhou-Shenzhen (Figure 4c). This shift shows that there was still some rigid demand for travel between major cities in the context of pandemic prevention and control in China and the increasingly severe global pandemic situation in 2021.
The advantageous association network of extra-provincial linkages was clearer during the National Day holiday in 2019, from the fourth level increment to the first level, and the high-level flow of city advantageous association showed a strong overlap with the network scope of national urban clusters (Figures 4e, 4f). By 2021, the first and second levels of advantageous association of the cities increasingly did not cross to outside of the urban clusters (Figure 4g), but the travel flow concentrated in the Yangtze River Delta, Beijing-Tianjin-Hebei, Chengdu-Chongqing, and other core urban clusters had decreased. The most significant drop was in the urban clusters of the Yangtze River Delta, Shandong Peninsula and the west coast of the Strait, which was also related to the short outbreak of the local pandemic in Nanjing, Yangzhou, Yantai and Xiamen. A comparison of the intra-and extra-provincial linkages revealed that the coverage of the travel network in 2021 after the pandemic was gradually becoming less extensive. The advantageous association lines were reduced after the pandemic, the high-level of advantageous association in areas where there had been outbreaks of local pandemic was significantly reduced, and the proximity effect of cyberspace became more pronounced.

3.3 Characteristics of urban clusters

To reflect the spatial clustering effect of travel by urban residents in 2019 and 2021 during the National Day holiday, the urban cluster structure was categorized based on modularity. Based on the principle of relatively close urban resident travel links within clusters and relatively sparse urban resident travel links between clusters, the urban clusters were divided as shown in Figure 5. The structure of the urban resident travel network can be divided into neighboring clusters composed of adjacent provinces and single provincial clusters. They were divided into 20 clusters in 2019, which included nine single provincial clusters, five clusters composed of two adjacent provinces, and four multi-cross- provincial clusters containing three provinces or more. One of the groups covering three provinces was centered on Xi’an and Taiyuan, and it covered 20 cities, such as several cities in Shaanxi Province and Shanxi Province, and Qingyang City in Gansu Province. The scope of the three clusters containing more than three provinces represented the spread and extension of cross-provincial or intra-provincial urban clusters, such as the Pearl River Delta, Chengdu- Chongqing, and Yangtze River Delta urban clusters, with strong spillover effects.
Figure 5 Spatial patterns of city grouping during China’s National Day holiday

Note: The numbers in the figure represent the category of the cluster the area belong to. Hong Kong, Taiwan, and Macao were excluded from this study as their data could not be acquired.

In 2021, The structure of the urban resident travel network was divided into 25 clusters, which included 16 single provincial clusters, eight clusters composed of two neighboring provinces, and one cluster containing three neighboring provinces, namely Jilin, Liaoning, and Inner Mongolia. Compared to 2019, the total number of clusters and single-province clusters had increased, and the agglomeration effect was reflected in the smaller areas. Due to the stronger siphoning effect of the province where a regional center city is located, the neighboring cities of the relatively weaker provinces formed neighboring province clusters with it. While the cross-multi-provincial clusters had been significantly reduced in places where the cross-provincial clusters shrunk, the overlapping range with provincial administrative regions had increased. They mostly showed an axis-spoke travel pattern centered on the core city of the urban cluster and closely linked to other cities within their respective provincial clusters, which is also consistent with the aforementioned travel characteristics of the intra-provincial linkages.

4 Factors influencing the impact of COVID-19 on resident travel during National Day

For analyzing the influencing factors, the relevant findings in Section 3 include: 1) The number of travel routes showed a notable decline in 2021, the clustering coefficient rose, and post-pandemic travel among residents became more provincialized and of lower value; 2) Regions that witnessed substantial reductions in both intra-provincial and inter-provincial travel networks were associated with recent local outbreaks. Rigorous control measures coupled with delayed impacts on the residents’ travel intentions kept these cities at a minimal travel level; and 3) In 2021, multi-provincial group tours were subdivided into multiple groups, further intensifying the barrier effects of administrative divisions due to the pandemic. Based on these findings, we can infer that the alterations in intercity travel among residents following the pandemic were profoundly influenced by four primary factors: urban competitiveness, pol-icy control, administrative boundary barriers, and the residents’ travel intentions and risk perception.

4.1 City competitiveness

Urban competitiveness serves as a pivotal catalyst for inter-city travel among residents. A profound correlation exists between urban competitiveness and the potency of urban allure, encompassing elements such as economic scale and population size, in conjunction with the associated costs for travel between cities, including transportation connectivity and travel time. Consequently, based on the nationwide urban economic, population, and connectivity data for 2021, an index system was formulated to analyze the factors influencing urban travel. In the results pertaining to the SLM model and SEM model fit (Table 3), all variables except urbanization rate, high-speed rail, and administrative rank were positively correlated with the inflow size index. In the SEM model, GDP per capita, number of employed persons, rate of urbanization, proportion of tertiary industries, and the air and road passenger traffic are significant and include the main characteristic indicators of “city competitiveness”. On the economic side, the fitted coefficient of GDP per capita was 0.192, which surpassed the 5% level of significance. This suggests that the post-pandemic economic status of various cities exerted a certain influence on the scale of public travel, thereby indicating high economic resilience. Concurrently, cities with a larger economic base typically have greater numbers of industries, enterprises, and job opportunities. This attracts a significant influx of people, which in turn promotes the city’s economic growth and competitiveness. The significance level of employees in terms of population is 1%, indicating that this correlation is significant. The abundant labor force in cities with large populations has a strong positive correlation with the scale of mass travel, indicating that the more employees, the stronger their willingness to travel, and the larger the scale of travel. This virtuous cycle enhances the competitiveness of cities. The urbanization rate showed a negative effect that was related to the change in preference of urban residents to travel to the nearby small and medium-sized cities on the National Day holiday in 2021, despite the fact that the share of tertiary industries and urbanization rate both surpassed the 10% level of confidence. The severe and inconsistent pandemic prevention and control methods in various cities and scenic areas had significant impacts on administrative rank and the abundance of tourism resources, making the coefficient insignificant, so the resources could not be used to an advantage. The negative, non-correlation of administrative rank further corroborated the role of administrative barriers that intensified during the pandemic. Aviation did not contribute to connectivity during the National Day holiday in 2021, since the average number of daily flights operated on mainland routes during the National Day holiday was 14.57 less in 2021 than in 2020. This may be mainly attributed to the public's preference for short-distance travel during the pandemic, which implies an increase in travel efficiency and a reduction in transportation costs, thereby strengthening the economic ties and personnel exchanges between cities. This synergistic effect contributed to the enhancement of the overall competitiveness of these cities.
Table 3 SEM and SLM model fitting outcomes for the volume of public traffic during the National Day holiday
Explanatory variables Explained variable: size of inflows
SLM model SEM model
Per capita GDP 0.191** 0.192**
(0.0897) (0.0897)
Resident population 0.000181 0.000169
(0.000247) (0.000249)
Practitioner 0.325*** 0.329***
(0.107) (0.107)
Urbanization rate -0.718* -0.725*
(0.424) (0.426)
Tertiary industry share 0.0110* 0.0108*
(0.00625) (0.00624)
Abundance of tourism resources 0.005 70 0.045 30
(0.004 58) (0.004 56)
Airline -0.186* -0.188**
(0.0956) (0.0954)
High-speed rail -0.0133 -0.0110
(0.119) (0.119)
Highway 0.118*** 0.117***
(0.0450) (0.0453)
Administrative level -0.0956 -0.0919
(0.117) (0.118)
Constant term 5.018*** 4.981***
(1.294) (1.300)
Spatial lag term 0.00133***
(0.000428)
Spatial error term 0.00300***
(0.00102)
Sample size 297 297

Note: *, **, *** mean that the significant levels are 10%, 5%, 1%, respectively.

4.2 Policy control

In China and around the world, the outbreak of the pandemic caused a serious public health emergency (Zhao et al., 2022). The crisis has made the internal reinforcement of urban risk and the external transmission of population movement more evident. The crisis has also highlighted the serious challenges faced by the nation, urban agglomerations, and high-density cities in the process of urbanization, in addition to the necessity for the constraints of population movement control during the spread of the virus. Policy control proved to be an effective measure to combat the spread of the pandemic (TNCPERET, 2020), and it is a major factor that affects the movement of the population. Figure 6 clearly demonstrates that the influence of ongoing pandemic management tended to flatten the national travel scale curve for the National Day holiday in 2021. The overall total travel scale decreased significantly compared to 2019, with a 64% yearly reduction in total travel. The total average daily travel decreased from 12067 in 2019 to 4291 in 2021, which also showed a reduction compared to 2020. This resulted from the introduction of the Delta mutant strain into the country in June 2021, followed by sudden local outbreaks and spillover across the country, which resulted in further strict policy controls on population movement and associated factors.
Figure 6 Urban resident travel volume on the National Day holidays in 2019 and 2021
Data from the Ministry of Culture and Tourism indicates that, as of August 4, 2021, a total of 1152 A-level tourist attractions in 20 provinces, autonomous regions, and municipalities nationwide had temporarily suspended their operations. These attractions refrained from accepting tour groups originating from medium and high-risk areas, and also ceased to organize excursions for tourists from these regions. Additionally, inter-provincial travel activities were halted in the provinces, autonomous regions, and municipalities where medium and high-risk areas are present. Consequently, the public transportation and tourism sectors have been subjected to a multitude of restrictions, resulting in a reduction in travel volume. Figure 7 illustrates that the pre-pandemic travel volume (in 2019) only approximated the normal average on October 3rd. However, the pandemic caused a shift in the turning point of the daily migration scale for National Day 2021, moving it to around October 2nd, after which the level remained below normal. The daily migration scale curve for 2021 exhibited a “high before low” characteristic, reflecting a significantly low level overall as influenced by the pandemic.
Figure 7 Spatial pattern of top linkages of the urban agglomerations in China in 2019 and 2021
A crucial role was played by the stringent mobility control policies enacted at all levels of government in China, such as slowing down urban operations by controlling population movement, running public opinion campaigns, encouraging people to work and run schools from home, and reducing outdoor activities, to contain the spread of the pandemic, especially at the beginning of each round of the pandemic outbreaks throughout the country. For example, Zhejiang Province employed a strategy of “big data analysis + grid-based investigation” to rigorously enforce “source control + hard isolation + precise intelligent control”. As of July 2021, the province has amassed 1.3 billion data points, accurately identified over 2 million key individuals associated with the pandemic and implemented control measures for approximately 9.3337 million people. Big data research has indicated that more than 95% of confirmed cases were detected proactively. This finding indirectly corroborates the significant reduction in urban travel during the National Day holiday following the outbreak.

4.3 Administrative boundary barriers

As human-defined boundaries determined by government agencies based on political, administrative, and management purposes (Yin et al, 2017), administrative boundaries often do not fully overlap with the invisible boundaries of urban influence with respect to urban planning, traffic management, and resource allocation (Chen and Wang, 2019). Therefore, the relevant cross-provincial urban agglomerations are good products for following such spatial boundaries of social ties, political activities, and the interactions of residents. However, in the post-pandemic era, the “top-down” control of administrative boundaries has been particularly significant, as the existence of tangible boundaries through zoning, combined with the implementation of routine prevention and control measures, has reduced the travel of residents across the administrative boundaries into other provinces and municipalities. In addition,d the scope of their activities and interactions with the urban environment have been restricted to those specific provinces and municipalities.
The comparison in Figure 7 illustrates the premier linkage pattern of China’s seven largest urban agglomerations, which demonstrates that the high-ranking linkage flow in the spatial pattern was relatively increased in 2021. Most of these links are between nearby cities acting as regional integration hubs, and the lower the premier linkage intensity value, the more distant the links are from each other. Note that the central axis of Chengdu-Chongqing is “collapsing in the middle”. Chengdu’s dominance in the Chengdu- Chongqing urban cluster is highlighted. The interactions between provincial capitals in the middle reaches of the Yangtze River urban cluster are weakened, and the province’s self-formed axis-spoke network pattern is more obvious. Additionally, the high values of the first linkage intensity in the Shandong Peninsula urban cluster are mostly associated flows between Qingdao and its neighboring cities. Overall, the pandemic worsened the effect of administrative boundaries as a barrier, so the economy of the administrative district became more prominent, and the administrative boundaries influenced resident movement due to policy controls.

4.4 Residents’ willingness to travel and their perception of risk

The viral variant continued to spread throughout the National Day holiday, and the outbreak resurfaced and was repeated in nearby places, impacting the travel and tourism of local citizens. Using comprehensive review data sourced from the TripAdvisor website, a significant reduction in tourist destinations and corresponding reviews between 2019 and 2020 was observed. In 2019, there were 382 tourist destinations visited by travelers, accumulating a total of 8499 reviews. Conversely, in 2020, the number of tourist destinations plummeted to 84, with the volume of reviews diminishing to 583. This stark contrast underscores a notable decline in the residents’ propensity for travel.
In addition, surveys have confirmed that due to the gradual fragmentation of legal holidays and the promotion of “local holidays” in some regions, residents are more willing to travel shorter distances (Li, 2022) and prefer short-term leisure travel. This is consistent with the conclusion of this study that the urban resident travel network during the National Day holiday tends to be more provincial based on big data visualization, which indicates that peripheral travel is favored by the public for its distinctive features related to a quick decision time, small travel radius, and short single-trip timeframe.
According to Meituan’s “2021 National Day Holiday Consumption Trend Report”, individuals prefer intra-provincial travel during the National Day holiday, with only 22.4% of users opting to travel across the country and 77.6% of users opting to travel locally. The National Day holiday’s dominant correlation pattern in this study also supports a spatial “proximity effect” of the priority connections within the province, which can reflect the steep rise in the size of the peripheral travel market. Simultaneously, it also draws attention to the willingness to travel in the context of pandemic prevention and control as the primary factor that has influenced the movement of residents. Proximity travel in the local area is also influenced by the residents’ growing sensitivity to perceived risks. A survey conducted by Xu et al. (2022a) found that 96% of residents would explore the relevant pandemic prevention policies of their destinations before traveling, and the increase in residents choosing travel modes with low risk was 0.469 times, indicating that the recurrent uncertainty of the pandemic and the normalization of pandemic prevention policies have caused residents to pay more attention to the safety of their travel modes rather than prioritizing factors such as speed of travel and accommodation. In the places affected by the pandemic, the movement of people has slowed down to some extent, which has affected the magnitude and network pattern of the National Day holiday, according to the travel psychology and risk perception of the locals.

5 Discussion, Conclusions and future directions

5.1 Discussion and conclusions

This study presents a comparative analysis of urban resident travel network patterns during the National Day holidays in China, in the pre- and post-pandemic periods. The findings indicate that the pandemic has further enhanced the balance of urban resident travel networks. The correlation between the absolute potential of core nodes and their relative potential is positive, while the correlation of changes between WDC and DIT is negative, aligning with previous studies (Li et al. 2020a). The DIT of most core cities increased in 2021, with maximum improvement in the transportation hubs such as Chengdu, Changsha, and Wuhan, and it showed a decreasing influence from east to west. Urban linkages exhibited a significant reduction in travel routes compared to 2019, along with an increase in aggregation coefficients, which are consistent with the reports of reduced intercity travel due to COVID-19 (Li et al. 2021). This study differentiated between the intra-provincial and inter-provincial scales, revealing increased accessibility within provincial networks and reduced inter-provincial networks, indicating greater post-pandemic urban travel agglomeration. High-value linkage flows were dominated by shorter distance travel in 2021, with a significant reduction in extra-provincial links, suggesting a trend toward provincialized and low-value travel.
The urban travel network shows a spatial pattern of national urban clusters > regional urban clusters > local urban clusters during the National Day holiday. After the pandemic outbreak in 2021, the coverage area shrank, reducing the advantageous associations in those areas and constraining population flow. In 2021, there were 25 urban clusters. Cross-provincial clusters with stronger development potential before the pandemic were more likely to gather into larger clusters, like those identified by Li et al. (2020b). The extent of group travel within the provinces was curtailed due to provincial economic conditions and pandemic control measures, reflecting heightened administrative barriers. The decline in cross-provincial travel supports findings by Neuburger and Egger (2021), indicating that inter-provincial group travel is infrequent, while intra-provincial agglomeration is rising.
The primary factors impacting travel during the pandemic included urban competitiveness, policy control, administrative boundaries, and the residents’ travel intentions, and this list overlaps with the components identified in previous research (Li et al., 2021). This study incorporated policy control into the analysis. The spatial econometric model revealed a significant positive correlation between urban travel scale and urban competitiveness data. Standardized policy control has heightened the residents’ risk perception, reduced their willingness to travel, and reduced urban migration. These effects align with the findings of Xu et al. (2022b). The turning point for travel size on National Day holiday in 2021 was closer to October 2 than in 2019, and total urban travel was 64% lower than in 2019. Compared to a study by Li et al. (2021), which found a 46.0% reduction in travel for May Day in 2020, this decline demonstrates a stronger shock to the urban resident travel networks due to the mutated virus. Administrative boundaries serve as the primary medium for policy control, and they influence urban travel patterns.

5.2 Implications and measures

This study presents a comprehensive research framework, which includes an overall pattern analysis, multi-dimensional feature comparison, and factor analysis. This framework is based on the hierarchical comparison of “node-line segment-network” to examine the hierarchical structure, scale, and dynamic travel flow patterns of mass travel in the two years preceding and following the 2020 pandemic outbreak. Additionally, by integrating data related to the urban economy, population, and connectivity, this study identified four primary factors that influenced China’s National Day mass travel during the pandemic. Based on the above findings, the following three policy implications for travel management during urban public health crises in China are suggested: 1) Implementing QR codes for health, travel, and payments, and adopting measures like flexible leave, consumption vouchers, and upgraded service standards to improve travel safety and boost consumer confidence; 2) Developing local and regional travel incentives and enhancing tourism through technology, such as online scenic spots and virtual tours; and 3) Using predictive tools to manage passenger flow can aid in congestion control and improve travel experiences and marketing strategies.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Athanasopoulos G, Hyndman R J, Kourentzes N, et al. 2023. Probabilistic forecasts using expert judgment: The road to recovery from COVID-19. Journal of Travel Research, 62(1): 233-258.

DOI PMID

[2]
Balcan D, Colizza V, Gonçalves B, et al. 2009. Multiscale mobility networks and the spatial spreading of infectious diseases. Proceedings of the National Academy of Sciences of USA, 106(51): 21484-21489.

[3]
Balcan D, Gonçalves B, Hu H, et al. 2010. Modeling the spatial spread of infectious diseases: The global pandemic and mobility computational model. Journal of Computational Science, 1(3): 132-145.

PMID

[4]
Cao F D, Huang Z F, Huang Rui, et al. 2021. Spatial correlation and coupling path between expressway flow and scenic tourist flow in Jiangsu Province. Economic Geography, 41(1): 232-240. (in Chinese)

[5]
Cao Z D, Wang J F, Gao G G, et al. 2008. Spatial pattern and differentiation of SARS pandemic in Guangzhou. Geographical Research, 27(5): 1139-1149, 1226. (in Chinese)

[6]
Cao Z D, Zeng D J, Zheng X L, et al. 2010. The characteristics and spatiotemporal transmission of SARS in Beijing. Science China Earth Sciences, 40(6): 776-788. (in Chiese)

[7]
Chen W, Xiu C L, Ke W Q, et al. 2015. Hierarchical characteristics of urban networks in China from the perspective of multiple traffic flows. Geographical Research, 34(11): 2073-2083.

[8]
Chen Y F, Wang K Y. 2019. A study on the influence mechanism of urban spatial expansion: A case study of Hangzhou. Geographical Research, 38(2): 221-234. (in Chinese)

[9]
Cheng X L, Fang Y L, Su X Q, et al. 2021. Study on spatial distribution characteristics of tourism flow network structure of five major urban agglomerations in East China. Progress in Geography, 40(6): 948-957. (in Chinese)

[10]
Cong L, Li S Y, Hong J X, et al. 2021. Spatial structure of tourism flow network in China’s red tourism scenic spots. Journal of Arid Land Resources and Environment, 35(12): 188-194. (in Chinese)

[11]
Dai T Q, Jin F J, Wang J E. 2005. Spatial interaction and urban association network evolution: Taking intercity railway passenger flow in the 1990s as an example. Progress in Geography, 24(2): 80-89. (in Chinese)

[12]
De Montis A, Caschili S, Chessa A. 2011. Time evolution of complex networks: Commuting systems in insular Italy. Journal of Geographical Systems, 13(1): 49-65.

[13]
Deng Y, Liu S H, Cai J M, et al. 2014. An empirical study on the evolution of the spatial pattern of population in China. Acta Geographica Sinica, 69(10): 1473-1486. (in Chinese)

[14]
Fang C. 2021. The construction of urban agglomerations and metropolitan circles in China under the new development pattern. Economic Geography, 41(4): 1-7. (in Chinese)

[15]
Farzanegan M R, Gholipour H F, Feizi M, et al. 2021. International tourism and outbreak of Coronavirus (COVID-19): A cross-country analysis. Journal of Travel Research, 60(3): 687-692.

[16]
Jiao J J, Wang J E, Jin F J, et al. 2016. Research on the influence of high-speed railway on urban network structure—Based on the analysis of railway passenger train. Acta Geographica Sinica, 71(2): 265-280. (in Chinese)

[17]
Jin F J. 2001. Research on the development of air passenger flow network and its regional system in China. Geographical Research, 20(1): 31-39. (in Chinese)

[18]
Li F Y. 2022. Analysis of travel intention of small-town residents in the post-pandemic era. China Market, (33): 36-39. (in Chinese)

[19]
Li T, Li Y, Dai L, et al. 2021. Characteristics and influencing factors of inter-city travel during the May Day short holiday under the influence of COVID-19. Geographical Research, 40(11): 325-3241. (in Chinese)

[20]
Li T, Wang J E, Gao X C. 2020a. Intercity travel network heterogeneity between weekdays and holidays in China. Acta Geographica Sinica, 75(4): 833-848. (in Chinese)

[21]
Li T, Wang J E, Huang J. 2020b. Inter-city travel patterns and network characteristics of Chinese urban agglomerations during National Day Holiday based on Tencent migration data. Journal of Geo-Information Science, 22(6): 1240-1253. (in Chinese)

[22]
Liu Z, Qian J L, Du Y Y, et al. 2020. Estimation model of the multilevel spatial distribution of interregional migration population based on multi-source spatiotemporal big data: A case study of people who migrated from Wuhan during the COVID-19 pandemic. Journal of Geo-Information Science, 22(2): 147-160. (in Chinese)

[23]
Ma X, Zhang Z G. 2020. Spatial evolution characteristics and trends of the Yangtze River Delta based on population flow. Urban Planning Forum, (5): 47-54. (in Chinese)

[24]
Mao L, Bian L. 2010. Spatial-temporal transmission of influenza and its health risks in an urbanized area. Computers, Environment and Urban Systems, 34(3): 204-215.

[25]
Meo P D, Ferrara E, Fiumara G, et al. 2008. Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory & Experiment, (10): 155-168.

[26]
Neuburger L, Egger R. 2021. Travel risk perception and travel behaviour during the COVID-19 pandemic 2020: A case study of the DACH region. Current Issues in Tourism, 24(7): 1003-1016.

[27]
Peng P W H, Wong D T, Bevan D, et al. 2003. Infection control and anesthesia: Lessons learned from the Toronto Sars outbreak. Canadian Journal of Anesthesia, 50(10): 989-997.

[28]
Qi W, Zhao M F, Liu S H. 2017. Calculation and evolution of regional types of floating population at county and city level in China from 1982 to 2010. Acta Geographica Sinica, 72(12): 2131-2146. (in Chinese)

[29]
Sattenspiel L, Dietz K. 1995. A structured pandemic model incorporating geographic mobility among regions. Mathematical Biosciences, 128(1): 71-91.

[30]
Shen Y, Chai Y W. 2012. Research on commuting resilience of urban residents based on GPS data: A case study of suburban megacities in Beijing. Acta Geographica Sinica, 67(6): 733-744. (in Chinese)

[31]
TNCPERET (The Novel Coronavirus Pneumonia Emergency Response Epidemiology Team). 2020. The epidemiological characteristics of an outbreak of 2019 Novel Coronavirus Diseases (COVID-19)—China. China CDC Weekly, 2(8): 113-122. (in Chinese)

PMID

[32]
Twu S J, Chen T J, Chen C J, et al. 2003. Control measures for severe acute respiratory syndrome (SARS) in Taiwan. Emerging Infectious Diseases, 9(6): 718-720.

[33]
Wang B, Zhen F, Xi G L, et al. 2013. Research on network information geography based on Weibo user relationship: A case study of Sina Weibo. Geographical Research, 32(2): 380-391. (in Chinese)

[34]
Wang X, Pei T, Li K, et al. 2022. Analysis of changes in population’s cross-city travel patterns in the pre-and post-pandemic era: A case study of China. Cities, 122: 103472. DOI: 10.1016/j.cities.2021.103472.

[35]
Wei Y, Xiu C L, Wang Q, et al. 2018. Analysis of the phenomenon and imbalance of rich people’s clubs in the migration network during the Spring Festival travel rush in China. Human Geography, 33(2): 124-129. (in Chinese)

[36]
Xu C, Zheng Y Y, Guan M Q, et al. 2022a. Analysis of travel characteristics and influencing factors of urban residents in the context of pandemic prevention and control. Journal of People’s Public Security University of China (Natural Science Edition), 28(2): 31-37. (in Chinese)

[37]
Xu M, Huang Z F, Cao F D, et al. 2018. Structural characteristics and influencing factors of tourism flow network based on online booking data analysis: A case study of the Yangtze River Delta. Economic Geography, 38(6): 193-202. (in Chinese)

[38]
Xu Y, Peng P, Lu F, et al. 2022b. Spatial and temporal characteristics of tourist travel behavior before and after the COVID-19 outbreak: A case study of Hong Kong, China. Geographical Research, 41(12): 3199-3213. (in Chinese)

[39]
Xue B, Xiao X, Su F, et al. 2020. Academic response and prospect of geography in the early prevention and control of COVID-19. Scientia Geographica Sinica, 40(10): 1593-1600. (in Chinese)

[40]
Yang H, Dijst M, Witte P, et al. 2018. The spatial structure of high speed railways and urban networks in China: A flow approach. Tijdschrift Voor Economische En Sociale Geografie, 109(1): 109-128.

[41]
Yang M, Xie Z Y. 2020. Impact of COVID-19 pandemic prevention and control on population mobility in China: An empirical study based on Baidu Map migration big data. Population Research, 44(4): 74-88. (in Chinese)

[42]
Yin J, Soliman A, Yin D, et al. 2017. Depicting urban boundaries from a mobility network of spatial interactions: A case study of Great Britain with geo-located Twitter data. International Journal of Geographical Information Science, 31(7): 1293-1313. (in Chinese)

[43]
Zhang W, Derudder B, Wang J, et al. 2020. An analysis of the determinants of the multiplex urban networks in the Yangtze River Delta. Journal of Economic and Human Geography, 111(2): 117-133.

[44]
Zhao Z Y, Han Z H, Wei Y, et al. 2022. Effects of population movement control on COVID-19 in China. Acta Geographica Sinica, 77(2): 426-442. (in Chinese)

[45]
Zhou C S, He X. 2022. Migration of population factors and countermeasures in the Pearl River Delta urban agglomeration. Planners, 38(6): 20-26. (in Chinese)

[46]
Zhou H L, Wang F Y. 2020. Spatial network structure of inter-provincial tourist flows in China based on modified gravity model. Geographical Research, 39(3): 669-681. (in Chinese)

[47]
Zhu Y B, Ge Q S, Wei X A, et al. 2003. Study on the impact of SARS on the number of inbound tourists in China. Geographical Research, 22(5): 551-559. (in Chinese)

Options
Outlines

/

〈 〉