AGENDA REPORT - Laserfiche WebLink

It should be noted that the travel system in and through the City of Pleasanton is a highly complex interaction of excessive demand on overcapacity roadways with the interplay and balance of regional travel on regional facilities vs. cutting-through the cities of Pleasanton as well as Dublin and Livermore. In this kind of system, it is extremely difficult to simultaneously validate to ITE trip generation, traffic counts, travel surveys, and cut-through estimates. However, the City of Pleasanton model does reasonably and acceptably well on all aspects. Traffic Count Validation The City of Pleasanton provided traffic count information from a variety of source, including automatic street segment counts, monitoring at major intersections, and peak hour intersection volumes from traffic impact studies. Additional detailed traffic count information on state routes was obtained from Caltrans. Traffic validation was determined reasonable and well within the acceptable limits on nine screenlines throughout the City. Screenlines are imaginary lines, often along natural or man-made physical barriers (e.g., rivers, railroad tracks) that have a limited number of crossings. The screenlines should "cut" the entire study area, intercepting all travel across them, thereby eliminating issues about individual route choice. Use of a system of screenlines allows systematic comparison of total model estimated versus observed travel in different parts of the model area. The City of Pleasanton is within 10% on nearly every screenline by direction and all percent RMSEs are well under the target 30% (see Appendix for details). Matrix Estimation Model validation results can be significantly improved by using the traffic counts as one of the inputs to the model process. After estimation of trips using conventional procedures (trip generation, distribution and assignment), the matrix of trip origins and destinations can be factored to more closely replicate the traffic counts. For example, if the conventional model process results in estimated traffic volumes that are 20 percent too high on a freeway segment, the matrix estimation process can adjust all of the origin-destination pairs using that segment by a factor of 0.8. This is a complex iterative process, as every adjustment affects many segments with traffic counts. Matrix estimation can improve the validation statistics, but it adjusts the trip origins and destinations in ways that cannot be easily explained or replicated. The recommended application is to perform matrix estimation to provide a more accurate base year result, but use the conventional model procedures to forecast increments between the base year and future years. Matrix estimation was applied to the Pleasanton Model 2005/6 peak hour trip estimates. Total trip estimates to and from each zone were maintained at the ITE trip generation rates. City of Pleasanton Trauel Demand Forecasting and Micro-Simulation Models Page 12 January 23, 2007