DEVELOPMENT OF A TRAFFIC MODELING SYSTEM FOR DETOUR PLANNING ON THE DOWNTOWN SEATTLE TRANSIT PROJECT

A major transit subway project is being constructed in downtown Seattle, Washington. A 1.3-mi electric-bus tunnel and associated surface-street improvements are in the final design phases and initial tunnel construction has begun; the expected completion date is 1990. The Downtown Seattle Transit Project (DSTP) was initiated by Metro Transit, the city of Seattle, and UMTA to help relieve existing traffic congestion in downtown Seattle and to provide capacity for growth. The tunnel will have three underground stations as well as combined station and staging areas at each end of the alignment. Both cut-and-cover and tunnel boring construction techniques will be utilized on the project. One of the greatest consequences of such a major construction project in a central business district (CBD) can be the adverse impacts on CBD traffic. An important task for project planners has thus been to assess the likely impacts of construction on traffic and to develop traffic maintenance plans that will best facilitate the tunnel construction and keep traffic impacts to a minimum. An innovative and complex traffic modeling system has been developed to aid in this task. Based on three existing traffic planning software programs (LINKOD, MINUTP, and TRANSYT-7F), a modeling chain has been developed that provides a systematic means for assessing the impacts of street closures, detours, and other traffic restrictions; identifies potential hot spots; and facilitates the development of traffic control plans to mitigate these impacts. The development and calibration of this modeling system, which has several innovative features likely to be of interest to other traffic modelers, are described. The modeling system organizes the analysis of traffic maintenance schemes as well as provides an ongoing tool for helping to design the longer-range (design-years) traffic improvements. Also included in the paper is a discussion of the effort involved in developing the modeling system and some suggestions for further research to improve the system for future applications.