Abstract:

Metro railway is emerging as the optimal choice for residents' daily travel in many metropolitan areas in China, which profoundly affects the spatial-temporal characteristics of the travel mode of residents. Studying the spatio-temporal characteristics of metro traffic flow is of great practical significance for optimizing metro traffic layout and relieving urban traffic pressure. Reported studies have afforded significant achievements in the spatial-temporal characteristics of metro traffic flow; however, there are very few studies on the overall perspective of the station network, and most of them fail to further analyze population differences in using metro systems. In view of these gaps in the existing research, this study has two main objectives: 1) to describe the characteristics of metro traffic from the overall perspective of metro station network, with parameters of traffic flows within the station network, average travel distance (time), and their aggregation features; and 2) to analyze the differentiation of population groups near metro stations for understanding population differences in metro traffic flow. The Origin-Destination (OD) analysis method was used to calculate traffic flow characteristics between different metro stations, including the average travel cost (distance/time) and travel aggregation distance interval. Our results lead to the following inferences. 1) The cumulative proportion of passengers in metro stations along with travel distance show an "S" curve function feature, the average travel cost of metro stations in Guangzhou is approximately 14.04 km (20.48 min), and increase by ~4 km (~5 min) and ~13 km (~10 min) from the central area to the inner and outer suburbs. The average weekend travel costs are slightly higher than that of weekdays: the average travel cost on weekend ranged from 0.03 km (0.06 min) less in the central area to 0.32 km (0.49 min) and 0.64 km (0.77 min) more in the inner and outer suburbs, respectively. Three types of metro stations—those serving external transportation, business offices, and public units—have higher average travel costs. 2) The distance interval of passenger flow concentration in metro stations tends smoothly from the center to the periphery. The travel aggregation distance interval in the entire city is 8.55-26.61 km, with 71.88% passengers aggregated within this travel distance range. The travel aggregation distance interval in the central area and the inner and outer suburbs are 2.74-19.23, 7.49-25.23, and 24.3-46.73 km, respectively, with 78.87%, 71.81%, and 56.55% passengers, respectively. The spatial distributions on weekend and weekdays are almost the same, only with a narrow range of travel aggregation distance interval and smaller proportion of passengers on weekends than weekdays. The station types of external transportation, leisure tourism, and residential areas have higher travel aggregation distance interval and passenger proportion. 3) Population groups like women, elderly, people with college-level education or above, office staff, business people, and service personnel are the main components of metro passengers. The average metro travel distance of different population groups is differentiated, as groups with small travel demand and weak travel ability have short average travel distance, while groups with large travel demand generally have a long average travel distance. Our research contributes to the existing literature mainly in two aspects. 1) Average travel cost, knee point analysis, and travel aggregation distance interval were used to depict the spatial-temporal characteristics of metro traffic flow from the overall perspective of metro station network; this is a supplement and expansion of existing literature mainly based on stations or lines. 2) The differentiation of different population groups helped further optimize the planning, design, and line arrangement of metro systems.

Key words: residents' travel mode, metro traffic, OD analysis, population differentiation, traffic flow, Guangzhou