Document Type

Report

Publication Date

3-2017

Subjects

Commuting -- Oregon, Cycling -- Oregon, Pedestrians -- Oregon, Automatic data collection systems, Transportation -- Planning -- Statistical methods

Abstract

Estimates of vehicle miles traveled (VMT) drive policy and planning decisions for surface transportation. No similar metric is computed for cycling and walking. What approaches could be used to compute such a metric on the state level? This report discusses three such approaches, identifies the advantages and disadvantages of each, and applies them to Washington State. The first approach employs travel survey data. The second approach is sample-based using pedestrian and bicycle count data. The third approach is an aggregate demand model approach using demographic data combined with count data. Due to data limitations, none of these methods could be properly implemented on the statewide level. Despite the data limits, the methods were implemented for one county (King County) in order to compare findings. The travel survey method estimated the lowest bicycle and pedestrian miles traveled (BMT and PMT), and the sample-based method estimated the highest. The travel survey method is useful for a statewide measure, but it does not provide the detail needed for facility-level estimates. For bicyclists, the sample-based method is appropriate if volumes are desired at the facility level. For pedestrians, the aggregate model might be more appropriate, because of the more dispersed nature of pedestrian travel. Each method has strengths and weaknesses, and each helps us understand bicycle and pedestrian travel in different ways.

For this reason, the project team recommends improving both statewide travel survey data and pedestrian and cyclist traffic count data which feed these methods. Travel survey data should be collected statewide with oversampling for non-motorized travelers. Pedestrian and cyclist traffic counts should be expanded to include a continuous counting program in addition to the short-duration count program. After the continuous count program is in place, short-duration counts should be chosen using a stratified random sampling approach. For example, the sampling frame could consist of all road and path segments in the state divided by region (Coast Range, Puget Lowland, Cascades, Eastern Washington), by urbanity (rural, urban), by facility type (highways/arterials, local/collector roads, paths), and by whether the location is on a bridge or not. To increase sites sampled, the short-duration count program could also be rotated, with each location being counted every three years instead of every year. Better data will allow the state to quantify bicycling and walking at both the state level and facility level to inform decision-making, facility design and planning, and safety analysis.

Description

This is a final report, NITC-RR-708, from the NITC program of TREC at Portland State University, and may be found online at http://nitc.trec.pdx.edu/research/project/708.

The project brief associated with this project can be accessed at: http://archives.pdx.edu/ds/psu/19534.

DOI

10.15760/trec.163

Persistent Identifier

http://archives.pdx.edu/ds/psu/19533

Estimating_Walking_and_Cycling_Presentation-SellingerTRB.pdf (7210 kB)
TRB Presentation

APA_Poster2FINAL.pdf (5107 kB)
APA Poster

APBP_Poster2.pdf (5161 kB)
APBP Poster

Estimating_Walking_and_Cycling_Presentation-Nordback-Denver-V2short.pdf (5979 kB)
Presentation

Olympia2015v5short.pdf (3734 kB)
Olympia Presentation

WSDOT_PMT-BMT_Presentation-PhaseII.pdf (4446 kB)
Phase II Presentation

Nordback_WSDOT_report.pdf (1775 kB)
Nordback Sellinger Report 2014

Share

COinS