In Virginia, the Hampton Roads hurricane evacuation plan uses an Abbreviated Transportation Model (ATM) to combine socioeconomic and behavioral characteristics data into a geographic model that can estimate the number of people and vehicles that will be evacuating each jurisdiction under various storm scenarios. Based on estimates of the notification lead times required to clear each area, a Traffic Control Plan (TCP) is used to facilitate evacuation management using ramp controls and lane reversal techniques.
On I-64/I-664, ramp controls are designated to restrict access to Interstates as needed to influence route choice and balance demand across the network. Under more severe conditions lane reversal (contra-flow). is implemented on I-64 and evacuating traffic can use both sides of the Interstate to travel westward from the Hampton Roads Bridge-Tunnel (HRBT) to the I-64/I-295 interchange, a distance of 73 miles.
The overall goal of the study was to identify bottlenecks or other problem areas that may develop as a result of evacuation activites. Micro-simulation software (VISSiM) was used to model the impacts of lane management and quantify the impacts of different storm categories (Cat 1-5) on traffic conditions. Data collected was drawn directly from the ATM and TCP. Where data was not available or incomplete, but required for modeling, a number of assumptions and simplifications were made. (See note below)
The assessment indicated that lane reversal is warranted for any hurricane predicted to make landfall as a Category 4 or 5 storm, and is strongly recommended for any Category 3 storm. In addition, the study found that with lane reversal, increasing ramp metering rates will reduce ramp queuing and allow more efficient use of available mainline capacity. The authors noted, however, that if a Category 4 hurricane is expected, ramp metering would most likely not be effective at improving Interstate traffic flow. Under these conditions significant ramp queues could extend back onto the arterial network and cause additional congestion and gridlock.
The accuracy of the analysis was limited to the authors' acceptance of the assumptions input into the model regarding: background traffic levels, ramp load balancing, and the temporal distribution of traffic.