Citation

Hankey S, Lu T, Mondschein A, Buehler R. J. Transp. Health 2017; 7: 149-159.

Copyright

(Copyright © 2017, Elsevier Publishing)

DOI

10.1016/j.jth.2017.08.009

PMID

unavailable

Abstract

A number of recent studies have made progress on specific components of monitoring and modeling bicycle and pedestrian traffic. However, few efforts merge the goals of collecting traffic counts and developing spatial models to meet multiple objectives, e.g., tracking performance measures and spatial modeling for use in exposure assessment. We used estimates of bicycle and pedestrian Annual Average Daily Traffic (AADT) from a comprehensive traffic monitoring campaign in a small community to develop direct-demand models of bicycle and pedestrian AADT. Our traffic monitoring campaign (101 locations) was designed specifically to capture spatial variability in traffic patterns while controlling for temporal bias. Lacking existing counts of cyclists and pedestrians, we chose count sites based on street functional class and centrality (a measure of trip potential). Our direct-demand models had reasonable goodness-of-fit (bicycle R2: 0.52; pedestrian R2: 0.71). We found that aspects of the transportation network (bicycle facilities, bus stops, centrality) and land use (population density) were correlated with bicycle and pedestrian AADT. Furthermore, spatial patterns of bicycle and pedestrian traffic were different, justifying separate monitoring and modeling of these modes. A strength of our analysis is that we conducted counts at a representative sample of all street and trail segments in our study area (Blacksburg, Virginia; ~5.5% of segments) - an advantage of monitoring in a small community. We demonstrated that it is possible to design traffic monitoring campaigns with multiple goals (e.g., estimating performance measures and developing spatial models). Outputs from our approach could be used to (1) assess land use patterns that are correlated with high rates of active travel and (2) provide inputs for exposure assessment (e.g., calculating crash rates or exposure to other hazards). Our work serves as a proof-of-concept on a relatively small transportation network and could potentially be extended to larger urban areas.

Language: en

Keywords

Bicycle and pedestrian planning; Facility-demand model; Non-motorized transport