eHurricane Evacuation modeling - strategy evaluations and methodology enhancements
Metadata[+] Show full item record
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Hurricane evacuation has attracted renewed emphasis since hurricane Katrina in 2005. Every coastal state is establishing their evacuation guidelines and searching new methods to improve evacuations. In this dissertation, first, hurricane evacuation of the Hampton Roads region of Virginia is investigated using large-scale regional mesoscopic traffic simulation models. Fourteen evacuation scenarios consisting of various combinations of storm categories and traffic control strategies are evaluated. The evaluation of scenarios provided information on the clearance time, average travel times, bottleneck locations, and congestion durations. The major findings from scenario evaluations include: (1) The differences in participation rates (100% versus 70%) did not impact the clearance times in a Category 1 storm evacuation, but have significant impact in a Category 3 storm evacuation, (2) The status (open or close) of a critical tunnel crossing, the Monitor-Merrimac Memorial Bridge-Tunnel did not have impact on the evacuation performance in Category 1 and 2 storm. However, opening the tunnel would improve the performance in a Category 3 storm, (3) The clearance times derived from simulations can be used to determine when to issue evacuation orders for various storm intensities, and (4) The bottleneck locations and durations identified for each evacuation scenario can be used to allocate the limited traffic monitoring equipment during an evacuation. The second focus of this dissertation is to investigate the impact of assumptions made regarding evacuee route choice on evacuation performance estimates. In the hurricane evacuation literature, very few studies have documented the realistic route choice behavior of evacuees during a hurricane. Due to this lack of realistic route choice behavior data, modelers make assumptions about the route choice behavior and traffic assignment. User-equilibrium traffic assignment has been extensively used in past evacuation studies. In this dissertation, realistic route choice behavior was determined by evaluating findings of a few published studies. The impact of route choice behavior on evacuation performance, especially travel times, is then investigated using the regional simulation model of the Hampton Roads region. The analysis found that the user-equilibrium traffic assignment significantly underestimates the travel times during an evacuation. The extent of underestimation of evacuation travel times depends on the total evacuation demand (a function of storm intensity), and the percent of evacuees willing to use en-route information to seek alternate routes when facing congestion. For the three en-route percentages reported in the literature i.e., 30%, 50%, and 70%, the UE travel times were 58%, 42%, and 33% lower than actual travel times realized in a Category 1; 94%, 71%, and 57% lower in a Category 2; and 90%, 69%, and 54% lower in a Category 3 evacuation. These findings illustrate the need to collect real-world data on evacuee route choice in order to build accurate evacuation models. The third focus of dissertation is to propose a procedure to assess the benefit of adding additional intermediate crossovers on a contra flow facility. Contra flow operation in which the direction of traffic on one or more travel lanes is reversed in order to increase the capacity of a road network is becoming a critical component of the evacuation plans of coastal states. Several coastal states have a contra flow plan in place for evacuation, however only a few states have intermediate crossovers between the origin and termination points. The impact of intermediate crossovers on network performance has not been well investigated in previous research. This dissertation investigates the benefits of having intermediate crossovers between regular and contra flow lanes. Based on the investigation, it can be concluded that adding intermediate crossovers did improve network performance for medium and high evacuation demand situations. Adding intermediate crossovers for low demand situations did not improve the network performance and thus any considerations for intermediate crossovers for the low demand evacuations must be based solely on providing access to road-side services (gas, food, and others). For high and medium demand situations and for the road network studied in this section, a 28% improvement in the average travel time was observed by deploying four intermediate crossovers out of the 44 potential crossover locations. The iterative elimination procedure proposed in this dissertation is the first attempt in the literature to provide a systematic approach to determine the critical intermediate crossover locations within reasonable computation times.
Access is limited to the campus of the University of Missouri--Columbia.