Estimating Transit Accessibility with an Alternative Method Evidence from Broward County, Florida

Scientists have attempted to measure accessibility in several ways—the gravity-based measure being the most widely used. A typical gravitybased model estimates accessibility on a zonal basis as being a function of the sum of total opportunities weighted by the distance, time, and cost needed to travel from the origin zone to those dispersed opportunities. The model includes a parameter that represents the distance-decay relationship and takes an exponential form. Unfortunately, most scientists have arbitrarily chosen the value of the distance-decay parameter instead of estimating it from field survey data. Also, a typical model does not have any parameter attached to the socioeconomic variables. This study uses distance-decay parameters estimated with the use of survey data in Sacramento County, California, to estimate transit accessibility to jobs in Broward County, Florida. Assuming that transferability of distance-decay parameters is possible from one geographic area to another, it then explores such transferability of parameters from Sacramento County to Broward County by analyzing the spatial distribution of transit accessibility and compares the effectiveness of estimated transit accessibility with the traditional transit accessibility measure—proportion of a geographic unit covered by 1/4-mi buffer from a transit route. Results indicate that accessibility indices estimated by using the method presented in this paper reflect what one would expect in reality—much better than what a simple 1/4-mi transit buffer would produce. The paper explores the fact that the distance-decay parameters estimated in one geographic unit are transferable to another. It advances knowledge of the accessibility measuring method that would help solve long-standing debate on what parameters to use for distance-decay and socioeconomic variables going into the accessibility model. Future research needs to focus on validating such transferability of distance-decay parameters from one study area to another