Impact Of Built Environment On The Passenger Flows And Transfer Behavior Of Urban Rail Transit

Faced with the urban sprawl and the increasingly serious urban traffic problems caused by the sustained and rapid economic growth and fast development of urbanization,prioritizing the development of public transport(e.g.urban rail transit),and advocating the optimization of urban structure and the intensive & compact development to form a green,transportation-oriented built environment has become a national strategic need.The passenger flow and travel characteristics of urban rail transit to a great extent reflect the human mobility rhythms and the layout of urban built environment.The station areas are not only the key nodes in public transit networks and important places to affording numerous activity opportunities,but also crucial areas for urban development.It is particularly important to clarify the passenger flows and travel characteristics of urban rail transit and the impacts of built environment on it.However,to the best of the author's awareness,there have been few previous studies that systematically and comprehensively carried out theoretical and empirical researches on the relationship between urban rail transit network passenger flows,station ridership,transfer trips and built environment.To this end,the present thesis founded by the National Natural Science Foundation of China(No.71771049)and the National Key R&D Program of China(No.SQ2018YFB16001805)explored the urban rail transit passenger flows and travel characteristics at network,station and individual levels,and identified the impacts of built environment on urban rail transit passenger flows and travel characteristics for providing new ideas and research framework to understand residents' mobility rhythms,by treating Nanjing urban rail transit as a case study and combining multi-source data,such as the built environment,the smart card data and travel survey.The main content of this thesis is shown as follows:(1)Based on the international general “5Ds” indicators of built environment and in view of the lack of an open platform for environmental data and the difficulty of data acquisition in China,ArcGIS was used to import the base map,spatial correction,SHP output,computational geometry and spatial statistics of urban land use and OpenStreetMap road network to calculate the population distribution density,employment density,land use mix based on mixed entropy index,traffic network characteristics based on network density and number of intersections at both ends of the occupational and residential areas.The POI data including coordinate longitude and latitude was acquired by network crawler technology,and then several methods such as coordinate rectification,coordinate projection,spatial fusion and statistical summary on ArcGIS platform were used to obtain the corresponding built environment data.The research scale of this thesis was also stated in the paper.On the other hand,we also introduced the collection and pretreatment of smart card data and individual trip survey data of urban rail transit.This part provides a data basis for the analysis of urban rail transit passenger flows and trip characteristics,and for exploring the influences of built environment on the passenger flows and trip characteristics.(2)Shift power law was first utilized to fit OD passenger flow of urban rail transit network in different time periods,and then depending on random walking method,the community division of urban rail transit network based on OD passenger flow is carried out.The standard mutual information based on the information theory criterion was also used to compare the community structure at different time periods.From a perspective of average travel distance(ATD)of passenger flows at station level,this paper analyzed the spatial and temporal patterns of human mobility and the results showed that the average travel distances of OD flows based station would be larger when the station was farther from the city center.Using cluster analysis to identify station passenger flow patterns based on its time characteristics,urban rail transit stations in Nanjing were divided into seven types such as employment-oriented,mixed employment-oriented,residential-oriented,etc.According to the survey of transfer trips between urban rail transit and other travel modes,it is found that at urban rail transit stations with higher population density,more intersections,more density of road network and higher proportion of commercial office land,respondents preferred to choose active travel modes such as walking and bicycle to go to and leave from urban rail transit stations,and the walking distances were mainly within 800 meters,while the bicycle distances were mainly within 2 kilometers.(3)Because there is no need for hypothesis and presupposition and it can deal with multi-collinearity more effectively to obtain more accurate and stable results,gradient boosting regression tree(GBRT)model was thus established to explore the impacts of built environment factors such as population density,number of intersections,road network density,land use/land cover,land use mix,number of bus lines and distance to the city center on OD passenger flows,which can model complex and non-linear data and capture the non-linear effects of independent variables.Then,quantile regression model was used to quantify the relationship between the station-based average travel distance(ATD)and the distance to the city center,and to explore the effects of increasing land use mix and improving occupational-residential balance on reducing travel distance at different statistical scales(e.g.2 km,5 km).(4)A multinomial logistic model was established to reveal the relationship between built environment variables(e.g.urban land use/land cover,distances to the city center)and daily passenger flow patterns of urban rail transit stations.Several global and local regression models including spatial lag model and mixed geographical weighted regression model were established.The results indicated that the mixed geographic weighted regression model has the best fitting results for modelling the station ridership.The influences of population density,number of shops,proportion of commercial office land and number of bus lines on daily passenger flow at urban rail transit stations were spatially non-stationary,and the influence magnitudes varied with the change of geographic space.The effects of station attributes and built environment factors on hourly station boardings and alightings were further discussed,so that the impact analyses of station ridership was further extended from unilateral spatial perspective to synchronous perspective of time and space.(5)Travel behavior and mechanism of urban rail transit access and egress trips were explored from two different perspectives,namely travel mode choice and travel distance.Random parameter logistic models,which could capture the unobserved heterogeneity of explanatory variables,were established to examine the association between subjective built environment perception variables,objective built environment factors and travel mode choice when controlling for person,family and travel characteristics.Similarly,random parameter negative binomial regression models were established to explore the relationship between subjective built environment perception variables,objective built environment factors and walking and bicycling distances of transit access and egress trips.All the models could reveal which independent variables may have individual heterogeneity.It is expected to draw more attention of urban policymakers and traffic planners to the relationship between built environment and urban rail transit.With the opportunities of giving priority to the development of public transport and optimizing urban structure,this work is helpful for guiding urban development and urban rail transit planning and management.Lastly,it is hoped that this study can provide scientific theoretical basis and empirical reference for guiding the urban development and renewal,improving the management of urban rail transit system,expanding the attraction of urban rail transit,etc.in practice.