Abstract：Approximately one-fifth of Perth's population is aged 60 or older. Projections suggest that this proportion will continue to increase as a result of the large number of children born after the World War II (1946)(1947)(1948)(1949)(1950)(1951)(1952)(1953)(1954)(1955)(1956)(1957)(1958)(1959)(1960)(1961)(1962)(1963)(1964). Access to and accessibility around train stations for the aging population is and will become a more important issue as the elderly population continues to grow. The aim of the paper is to develop and apply a new measure of accessibility to train stations at a fine spatial scale, justified by the special circumstance of the elderly using a case study in Perth, Western Australia. Intercept surveys are used to collect data on factors affecting train station accessibility for patrons aged 60 years or older, at seven highly dispersed train stations. Overall accessibility is measured separately using a composite index based on three travel modes (walk-and-ride, park-and-ride and bus-and-ride). The results illustrate that key variables, such as distance from an origin to a station, walking or driving route directness, land-use diversity, service and facility quality, bus connection to train stations, all affect the accessibility to train stations for the elderly. This implies that improvements to these factors will improve accessibility for this population group.

A train station catchment area delineates the spatial territory from which the users of a train station are drawn. The size and shape of this catchment can be influenced by a variety of factors, such as the transport network, the location of stations and the service quality they offer, as well as the land use density and diversity in the transport corridor. Although numerous studies have been conducted to understand the size of catchment areas, limited research has focused on determining the spatial boundary (shape) of train station catchments. This paper develops a framework for deriving a spatial boundary of a Park and Ride (PnR) catchment area by incorporating the Huff model and Geographic Information Systems (GIS) technologies. The approach is staged, firstly determining the PnR station choice as a function of the attractiveness of a train station and the cost of access between the origin (such as a suburb) and the destination of a trip (such as the Perth CBD). Linear referencing method is then applied to redefine the origins to train stations based on the derived station choice probability. Finally, the spatial boundary of a catchment area is determined according to the adjusted origins, using GIS technologies. The model outputs were evaluated against licence plate survey of station users, where the Kappa coefficient (0.74) and overall accuracy (0.88) statistic suggested that the model's results are robust. The paper then shows how catchment area data can be used to better manage travel demand and plan design solutions aimed at increased accessibility to train stations.

No abstract

A B S T R A C TWe usually assume that each commuter is an efficient traveller, which means they maximize trip utility. From a spatial optimization perspective, a commuter might therefore choose the nearest station to reach their destination. However, based on a survey at seven train stations in Perth, Western Australia, only between 30 and 80 percent of commuters choose the nearest station to their origin. Many factors could affect this travel behaviour. From a logistic regression model, five factors were found to be significant (pvalue <0.05), indicating that commuters are more likely to choose the non-nearest station for longer commutes, while traveling further away from origins and destination if the chosen stations are at, or near, the end of train lines (captive stations). If the chosen stations are along the train line (non-captive stations), longer distance, longer wait times and lower costs from the chosen station to a destination were found to be significant. The results of the study are important for public transport policy makers to understand transit choice behaviours. Therefore public transport policies such as adjustments of travel fees and improving station service and facilities, could be developed.

A train station catchment area can be generated in two ways: by surveying train users or by modelling methods. The former method is time consuming and labor intensive. In this paper attention was given to develop methods that automatically generate station catchment areas. This study's aim was to compare two modelling methods: the Euclidean distance transform / Voronoi diagram generation method and a location-allocation method for automatically generating station catchment areas. A case study of the Perth Metropolitan area, in Western Australia, was used to implement these two methods. The results from these two methods are consistent and the methods demonstrate robustness for understanding the nature of station catchment areas and provide useful insights for public transport planning in Perth.