MODELING RAIL ACCESS MODE AND STATION CHOICE

Access mode and station choice by commuter rail and subway users are modeled using morning peak-period work trip commuting in the greater Toronto, Ontario, area as a case study. Based on observed station choice behavior, rules for determining access station choice sets for both commuter rail and subway were developed. For commuter rail, the two closest stations on the two closest lines (relative to the worker's place of residence) define the access station choice set. For automobile access to the subway, the five closest subway stations define the choice set. A nested logit model of commuter rail access mode and station and a multinomial logit model of subway automobile access station choice were then developed. Consistent with the findings of other researchers, credible models of access mode and station choice were obtained. Directions identified for further work include testing alternative overall main mode plus access choice structures, properly capturing parking supply and price effects with these models, developing improved representations of the automobile passenger mode, and development of improved network modeling software for dealing with mixed modes of travel.