Enhancing Mode Shares on High-Occupancy Toll Lanes – Applicable to Transit Planning
Corresponding Author: Amir Shahpar, AECOM
Presented By: Amir Shahpar, AECOM
Abstract
High-Occupancy Toll lanes (HOT) are a relatively new approach to alleviating traffic congestion on major freeways in metropolitan areas including I-95 and I-495 in Northern Virginia. HOT lanes are being added to the I-66 corridor as part of the 2015 Constrained Long Range Plans (CLRP). The addition of HOT lanes to the regional transportation networks pose critical modeling challenges to traditional travel demand models. The Metropolitan Washington Council of Government (MWCOG) travel demand model uses a two-step process to estimate traffic volume on HOT lanes. AECOM found that the current model significantly overestimates the share of High-Occupancy Vehicles (HOV3) demand on the HOT lane. The current model estimates show above 50 percent of HOV3 vehicles on I-95 HOT lanes compared to the observed share of 30 percent. AECOM concluded that the mode choice model irrationally shifts trips from single-occupancy vehicle (SOV) or transit to HOV3. The MWCOG trip generation and distribution models are stratified by household income level and the model uses different value-of-time (VOT) to convert tolls to minutes for distributing trips. However, the highway skims procedure uses a single VOT to convert tolls to minutes for all income levels. Thus, the resulting skims do not reflect that high income travelers are more likely to pay tolls to save time compared to low income travelers. AECOM developed a solution by preparing separate highway skims for each income level using different income based VOT meaning the best path for a given origin-destination pair can vary based on traveler’s income level. These income based highway skims were used as inputs to the mode choice model and income based mode choice results prepared for all trip purposes. A multi-user multi-class highway assignment was developed to assign the trip tables and the results showed significant improvements compared to the original model. The revised model estimates show 23 percent and 33 percent of HOV3 vehicles on I-66 and I-95 HOT lanes respectively. AECOM used the revised model to estimate transit ridership elasticity to changes in toll values for a proposed Gainesville-Haymarket extension of Virginia Express Railway which runs along I-66 corridor in Northern Virginia.