There are three (3) approaches to forecasting project traffic for Express Lanes:
Manual Estimation Using Peak Hour Origin-Destination (O-D)
This method uses a manual estimation of the Express Lanes volume by applying a fixed percentage of the Express Lanes share of traffic to future year peak hour Origin-Destination (O-D) volumes. The shares of the Express Lanes can be derived from observed data on existing corridors, such as I-95 Express. The future year O-D volumes are developed through use of observed data and corridor forecasts from a travel demand model.
Travel Demand Model (TDM) Based
- Regional TDM with Dynamic Toll Function or VTTS Curve Assignment
Some Florida travel demand models have embedded highway assignment scripting to specifically estimate Express Lanes traffic. The Southeast Regional Planning Model (SERPM) uses a generalized cost assignment and a logit function that dynamically calculates the toll after each iteration based on the volume to capacity (v/c) ratio in the Express Lanes. The Tampa Bay Regional Planning Model (TBRPM) uses predefined VTTS curve to estimate the probability of user payment given the marginal cost/minute saved, and a toll policy curve to describe how the toll varies by congestion as measured by the volume to capacity ratio. - Regional TDM with Express Lanes Time of Day (ELToD) Assignment Model
ELToD is a traffic assignment tool used in conjunction with a regional travel demand model to split traffic between Express Lanes and general use lanes. The ELToD toll choice model uses travel time savings, costs, reliability, and trip distance to calculate the percentage of travelers choosing the Express Lanes. ELToD estimates the volume of traffic by hour on both the general use and the Express Lanes using a highway trip table from any travel demand model. In addition, it estimates the Express Lanes dynamic toll and congested speeds by hour based on traffic conditions.
Microsimulation Model
This method uses modules within microscopic simulation software packages to dynamically assign traffic to the Express Lanes. Microscopic simulation models use a pricing component to estimate the toll amount based on measured conditions such as travel time savings and speed. An embedded decision model determines the probability of choice to use the Express Lanes given the travel time savings and costs. When using microsimulation models, the decision model should be modified so that it is consistent with ELToD’s toll choice model (Equation 9-1). In addition, the model has to be a Dynamic Traffic Assignment (DTA) model which is based on the regional network. This method requires O-D matrices from a macro- or meso-scopic model as an input to perform the traffic assignment and can be part of a multi-resolution approach.
A core function of Express Lanes forecasting is to accurately account for traveler preferences to choose the Express Lanes. This preference is largely influenced by the model inputs, such as VTTS, VOR, and the dynamic toll amount. Table 9-1 summarizes the model inputs for the various Express Lanes forecasting methods previously described, along with some pros and cons for each method. It also includes the appropriate phase(s) in the project development process where the method is recommended to use. Each project’s needs should be evaluated against the various forecasting methods.
Manual estimation can be used to quickly approximate the anticipated range in traffic projections for an Express Lanes segment. The manual method is typically suited for sketch-level activities on a simplified Express Lanes corridor and is generally not for Project Traffic Forecasting at the PD&E or Design level. Regional travel demand models with customized highway assignment scripting can provide an estimation of Express Lanes demand at the period or daily level. However, the models do not account dynamic pricing fluctuations at the design hour level. In addition, the choice component used to calculate the Express Lanes share typically includes the VTTS and costs, but excludes the VOR. Microsimulation models can provide the sensitivity to dynamic pricing, but require specialized scripting to include the VOR and the additional effort to properly calibrate the existing conditions model. As stated earlier, each project’s needs should be evaluated against the various forecasting methods.
