The COVID-19 pandemic, originating in Wuhan, China, prompted widespread policy interventions globally, including lockdowns, transport shutdowns, and travel restrictions, significantly impacting human mobility and social activity. The study of human movement became crucial for tracking infections, predicting pandemic waves, and evaluating intervention effectiveness. While existing research correlated human movement with pandemic dynamics, a systematic understanding of how human movement patterns varied and recovered remained limited, especially regarding long-term group mobility changes in public transport. Most studies presented aggregated analyses or surveyed attitudes and preferences, lacking systematic tracking of long-term group mobility. A significant modal shift from public transport to private vehicles presented a challenge to carbon emission reduction and sustainable transportation development, making the recovery of public transport use a key area of inquiry. This study bridges this gap by analyzing the recovery of group mobility in Kunming's subway system from November 2019 to September 2020. The restoration of city life after lockdown interventions can be viewed from supply and demand perspectives. On the supply side, the reopening of public transport signifies system resilience. From the demand side, full recovery implies a return to pre-pandemic behavior or adaptation to new norms. This 'demand-side' recovery relates to behavioral resilience in psychology, encompassing positive adaptation after negative events. Travel behavior during the recovery period is affected by various factors: service facilities, social context, transport supply, and individual psychological states. Individual differences, including socioeconomic background and preferences, influence the spatial and temporal diversity of travel behavior. This study defines travel behavior resilience as the co-evolution of transport supply and demand, achieving a temporary equilibrium. The study aims to measure whether and how public transport use recovers to pre-pandemic mobility levels by focusing on a longitudinal analysis of frequent subway users in Kunming, a representative city with a developing transit system.

The literature review highlights the existing research on human mobility and its correlation with the COVID-19 pandemic's spread and dynamics across various spatial scales. Studies focused on using mobility data to track infection, identify super-spreading events, simulate imported cases, predict pandemic waves, and assess intervention effectiveness. However, systematic tracking of long-term group mobility changes, particularly concerning public transport, was lacking. Several studies analyzed specific cities, often making predictions during the pandemic, but lacked comprehensive measurement of variation and recovery in human movement. Aggregated analyses of transport data were common, while some studies surveyed behavior, attitudes, and preferences. The literature acknowledges a significant modal shift from public transport to private vehicles, posing challenges to sustainable transportation development. The existing research identified gaps in understanding the recovery of public transport use during and after a pandemic, necessitating a longitudinal analysis to track group mobility differences and evaluate the recovery to pre-pandemic levels. This study utilizes previous research on engineering resilience (supply-side) and social-ecological resilience (demand-side) to better understand travel behavior resilience in the context of the pandemic.

The study utilized transit trip records from the Kunming subway system between November 2019 and September 2020. Data was collected from smartcards and virtual payment platforms. Trips from November 1st to 28th, 2019 were used to identify frequent travelers (those using the subway at least 3 days a week for 4 consecutive weeks). These frequent travelers formed the study's focus. The researchers tracked these individuals' trip records in September 2020 to assess mobility recovery. A total of 16,403 subjects, generating over 3.5 million trip records, were included in the longitudinal analysis. Subjects were classified into four groups: commuters, elderly (above 60 years old), students, and others. The study measured several mobility indicators, including the proportion of subjects who traveled, traveled days, number of trips, total trip distance, activity space (defined by the minimum convex polygon of visited stations), and number of stations visited. Weekly data allowed the researchers to track the three phases of public transport use during the pandemic: drastic reduction, rapid growth, and stabilization. The researchers compared the pre-pandemic mobility level with the post-pandemic levels to quantify the degree of reduction and recovery. Statistical analyses, including t-tests, were conducted to compare mobility indicators across groups and weeks. The study also analyzed the revisiting tendency during the pandemic by ranking stations based on visit frequency and calculating the proportion of trips generated at each station. Regression analysis examined the correlation between recovery duration and pre-pandemic mobility levels for different groups. Formulas for calculating rates of change were provided for each mobility indicator.

The study identified three phases in the recovery of public transport use in Kunming: drastic reduction, rapid growth, and stabilization. The proportion of subjects traveling decreased faster initially than other mobility indicators, highlighting the immediate impact on travel willingness. The recovery rate of subjects traveling returned to the pre-pandemic level before other mobility indicators. Activity space and stations visited recovered slower than other indicators, confirming Hypothesis 1 and 2. The study found a strengthened revisiting tendency during the pandemic, with trips concentrated on fewer frequently visited stations, supporting Hypothesis 3. Travel behavior resilience varied significantly across groups, with commuters showing greater resilience compared to the elderly, confirming Hypothesis 4. Commuters' mobility indicators increased earlier and sustained for a longer period. The elderly exhibited a slower recovery rate and longer reduction and recovery periods. The elderly tended to decrease travel frequency but not the size of their activity space, indicating a different recovery pattern compared to commuters. The study also analyzed the correlation between recovery duration and pre-pandemic mobility levels, indicating that individuals with higher pre-pandemic mobility levels had longer recovery durations. The recovery durations of trips, total distance, and activity space were longer for the elderly than for commuters. Table 1 summarizes the mobility indicators across the three phases for each group. Table 2 shows the t-test results for commuting trips. Figure 3 presents mobility indicators by groups over time. Figure 4 shows t-test analyses and trip distribution by station rank. Figure 5 demonstrates the correlation between recovery duration and pre-pandemic mobility levels.

The findings address the research question by demonstrating the heterogeneous nature of travel behavior resilience during the COVID-19 pandemic. The slow recovery of public transport use, even in a city with a single pandemic wave, highlights the need for long-term interventions. The substantial differences in recovery across groups, especially between commuters and the elderly, emphasize the importance of tailoring policies to specific needs and vulnerabilities. The intensified revisiting tendency and reduced destination diversity suggest a need for strategies to stimulate intra-urban mobility and facilitate city restoration. The study offers valuable insights for urban studies, transport management, and social science, particularly in understanding the long-term impact of pandemic restrictions on public transport usage and urban mobility. The Kunming experience, despite its relative isolation from multiple waves of infection, provides a relevant case study for cities employing stringent zero-COVID policies, offering valuable comparative insights.

This research offers a novel measurement of travel behavior resilience based on group mobility patterns, demonstrating the slow and heterogeneous recovery of public transport use in Kunming during COVID-19. The findings underscore the prolonged impact of pandemic restrictions on urban mobility and the need for targeted, group-specific interventions in transport management and city restoration. Future research could develop quantitative methods to define the re-equilibrium between travel demand and supply in cases where pre-pandemic levels are not reached and investigate the effects of public transport recovery on traffic congestion and carbon emissions.

The study's primary limitation is the 10-month data collection period, which could be affected by seasonal variability. A longer dataset would strengthen the analysis. The focus on Kunming, a city with a single pandemic wave, might limit the generalizability of the findings to cities experiencing multiple waves or differing pandemic management strategies. The study focuses on frequent public transport users, potentially neglecting the experience of infrequent users. Future research should explore the recovery patterns of infrequent users and incorporate other transport modes to offer a more comprehensive picture.