Lead Presenter Biography

Mark Feldman is a transportation engineer / planner with 17 years of industry experience. Prior to joining CDM Smith in 2018, he worked at Steer, Citilabs and Fehr & Peers. Mark's skills include traffic and revenue forecasting studies for tolled facilities, travel demand model development and forecasting, transit ridership forecasts, traffic impact studies, and statistical modeling including discrete choice and stated preference survey analysis. He has been primarily a project technical leader at CDM Smith for traffic and revenue forecasts of existing and proposed toll roads and managed lanes across the eastern and midwestern United States.

Co-Authors

Presentation Description

CDM Smith was selected to complete a comprehensive traffic and revenue (T&R) study of the proposed Hampton Roads Express Lanes Network (HRELN) on I-64 in the Norfolk, VA Region for summer weekend traffic conditions. The study developed summer weekend traffic T&R estimates to be combined with previous weekday estimates in refining annual T&R forecasts.

Adaptation of the weekday Hampton Roads Transportation Planning Organization (HRTPO) travel model to perform these forecasts included a significant summer weekend traffic count program, analysis of summer weekend traffic patterns, vehicle occupancies and travel speeds from Streetlight and INRIX data, customized summer weekend time periods and trip matrices, and a mixed multinomial logit (MMNL) value of time model, based on a stated preference survey administered specifically for this study.

Time periods for the weekend model were chosen by grouping consecutive periods with similar levels of congestion, similar to how AM peak and PM peak periods are chosen in a typical weekday travel demand model. Friday time periods were similar to an average weekday, with an additional mid-afternoon peak. Saturday and Sunday time periods differed considerably, with the peak levels of congestion occurring in mid-day.

The most complex component of this adaptation was the development of trip matrices for summer weekend time periods. Streetlight data was used to adjust the weekday matrices to account for relative overall levels of travel between weekdays and weekends, and the differences in temporal distribution of travel throughout the day.

Vehicle occupancy counts at several points in the study corridor helped inform adjustments to the vehicle occupancy distribution among passenger cars. Friday occupancies were slightly higher than a typical weekday, and Saturday and Sunday occupancies were considerably higher.

The resultant model was able to test a range of pricing policies with respect to vehicle occupancies, dynamic tolling by direction and time of day, and inclusion / pricing of trucks. The use of explicit summer weekend revenue forecasts from this model produced a significant revenue increase compared to the previous, less refined method of using the weekday forecasts and a global annualization assumption. 

Presentation File