Forecasting Transit Walk Accessibility: A Regression Model Alternative to the Buffer Method

This paper describes a methodology for estimating transit walk accessibility that overcomes the problems associated with barriers and uneven distribution of population, and a methodology for forecasting transit walk accessibility for a future year given forecast population and employment data, transit route information, and street configuration type. Comparison of the results with those from the traditional buffer method as well as network ratio methods that consider actual walk distance along streets show that both the buffer method and network ratio methods tend to overestimate transit walk accessibility. Regression analysis also showed that the new transit walk accessibility measure was a stronger predictor of transit use than that produced using the buffer method. The methodologies may be applied to transit planning, urban design for sustainable development, and long-range transit demand modeling. INTRODUCTION AND BACKGROUND A critical factor in transit planning transit mode share analysis is the level of accessibility to transit services by population and employment. “Transit accessibility” here refers to the ability of residents and workers to reach transit facilities, including bus stops and/or rail stations. Transit accessibility is affected by many factors including safe, pleasant, and comfortable streets for walking to transit facilities, parking facilities for cars and bicycles, handicap access, and so on. The majority of transit users access transit systems by walking; and walking distance is an important factor in the choice of transit use. According to a transit onboard (TOB) survey conducted in 2000 in Southeast Florida as part of a regional travel characteristics study, 79.8% of TRB 2003 Annual Meeting CD-ROM Original paper submittal – not revised by author.