01516080

Component

As traffic congestion increases, the K-factor, defined as the proportion of the 24-hour traffic volume that occurs during the peak hour, may decrease—a behavioral response known as peak spreading. Knowing how K-factors will change affects the estimation of travel demand, and the resultant transportation performance, since the traffic volume during a given hour may affect travel speed and vehicle emissions. This paper describes a model for forecasting peak spreading as measured by change in the K-factor. Data were collected from 52 sites based on continuous count stations in six Northern Virginia counties for the period 1997-2010, where the average annual K-factor decreased by an average of 0.006 (p < 0.01). The model explained 66% of the variation in K-factors based on the percentage change in the jurisdiction’s employment; circuity, i.e., whether the route is radial or circumferential; and for freeways, the 24-hour volume-to-capacity ratio (a surrogate for traffic congestion). An additional 10 sites in the same region, which were not used to calibrate the model, suggested an average error of 18%. (An alternative model, which can be used when a previous K-factor is available, explained 88% of the variation in K-factors in the training set data and had an average error of 12% with the testing set data.) The model has four implications for forecasting peak spreading. First, in addition to site characteristics such as functional class, regional socioeconomic characteristics (e.g., jurisdictional employment growth) affect the K-factor. Second, the K-factor varies more across sites with the time period held constant than across time periods with the site held constant. Third, variables typically available when a 10-year forecast is made can successfully explain variation in peak spreading. Fourth, traffic congestion affects the forecasts—consistent with expectations—but the effect is evident only after other variables are controlled for.

This paper was sponsored by TRB committee ADB40 Transportation Demand Forecasting. Alternate title: Longitudinal Aggregate Model to Forecast Peak Spreading: Northern Virginia Results.

01503729

14-1056

English

17p

2014

Transportation Research Board 93rd Annual Meeting

Digital/other

Figures; References; Tables

Data collection; Peak hour traffic; Pollutants; Speed; Traffic congestion; Traffic estimation; Traffic volume; Travel demand

Northern Virginia

Highways; Operations and Traffic Management; Planning and Forecasting; I72: Traffic and Transport Planning

Transportation Research Board Annual Meeting 2014 Paper #14-1056

TRIS, TRB, ATRI

Jan 27 2014 2:25PM