1. Project Overview

Planning agencies at the local level (cities/counties/multi-jurisdictional transportation agencies) typically develop, maintain, and operate four-step travel demand models to forecast traffic volumes arising from changes in land uses and circulation networks in proposed general plans, specific plans, and major development projects. Local travel models not only provide inputs for the assessment of impacts such as safety, circulation efficiency, noise, and air quality at the local level; but also generate inputs for transit service planning and the design and operation of ramps and intersections between local streets and state highways. Ensuring the quality and applicability of local travel demand models is thus an important issue demanding attention from both local and state transportation agencies. No designs and operation measures can successfully achieve communities’ safety, reliability, performance, flexibility, and productivity without a sensitive and accurate assessment of travel and traffic impacts in the planning stage.

This study proposes to assess local travel models specifically in regard to land use and alternative modes of travel through a fundamental and comprehensive examination on "state of the practice" of modeling, including aspects in data, methodologies, software, and operation practices. This study will identify local modeling in regard to commute and non-commute travel in regard to localized and community-wide impacts and benefits of land use location, density, design and configuration (e.g., mixed land uses, interconnected travel networks, availability of transit and non-motorized facilities, transit-oriented development, and the like).

The catalyst for this research interest is two-fold. First, are standard demand models that were designed for application in infrastructure planning proven successful in evaluating the impacts of land use and demand-oriented policies that deal with sustainable systems? Second, have “alternative” analysis techniques been successfully integrated into standard travel demand models used in transportation planning and assessment for local applications?

2. Terminology

The project scope calls for careful definition of terms, particularly relative to (a) travel demand models, (b) local versus regional applications, and (c) model policy sensitivity (especially with respect to land use characteristics).

2.1 Travel Demand Models
The phrase "travel demand models used at the local level" may be interpreted in several ways. Travel demand models (or travel forecasting models) typically are applications of the standard four-step travel forecasting model system (as applied in software packages such as TransCAD, CUBE, or EMME/2). This model system, in place for over four decades, essentially transforms characteristics of the land use (or activity) system and of transportation networks into link volumes and speeds. It does so via the sequential steps of trip generation, trip distribution, mode choice, and network assignment. The direct influence of land use (activity) characteristics is in the trip generation stage. In some model systems, land use (activity) characteristics may also influence distribution and mode choice results if choice-based models are utilized and reflect socio-economic characteristics, but the primary influence is in trip generation where the frequency of travel is determined. The influence of land use characteristics is also present in model systems where feedback allows for congestion (which via the model system is dependent on the relative location of activities and on characteristics of the transportation network) to influence distribution, mode choice, and time-of-day distributions. If the feedback process incorporates a land use / transportation model, then network performance can influence the land use (activity) characteristics that serve as input to trip generation.

Travel demand models have been enhanced and expanded in numerous ways but most current travel demand model applications in California (if not the whole United States) follow the general four-step convention. In this review, the term “model” will indicate travel demand model or travel forecasting model of the conventional four-step format. The term “alternate model” will refer to various evolving model structures and other analysis techniques directed toward capturing local policy impacts (land use policies, in particular). Other model types, such as land use models, will be explicitly identified.

There are, however, several different project application levels for travel demand models. The first project application corresponds to a "full model application" of the local (or regional) travel demand model system. If transportation and land use changes are anticipated to have significant impact beyond the spatial extent of the project, then a conventional four-step application is performed. For projects with more "localized" impacts, the traffic impact study with often utilized an abbreviated model system. Here, the standard trip generation model is often replaced with ITE (or other) trip rates based on land uses. The remaining steps of the model are often based on "engineering judgment" regarding the directional distribution of project trips as well as the modal distributions. The focus of these studies is often designing access strategies that minimize the impact of the proposed project when traffic is assigned to the local network. The mitigation strategies are not limited to site access but also include distribution, mode, and time-of-day options as well as traffic control options. The requirements that determine the level of modeling vary significantly from one area to another.

2.2 Local versus Regional Applications
Regarding the term "local level", the Caltrans RFP for this project refers to "city or county" level models, which indicates (travel demand) models that are developed by city or county agencies and/or applied at the city or county level. Such local models, however, are evolving into what might be termed "local regional models" where regional travel demand models that were developed for regional applications have been adapted by local agencies for local applications. Typically, this means adding detail (e.g., number of zones, network density, or enhanced sub-models) in the local area, and perhaps aggregating outlying areas. For example, the model developed and applied by Southern California Association of Governments (SCAG) for regional forecasting has been adapted and modified by the Orange County Transportation Authority (OCTA) to create the OCTAM model. This model covers the same geographical area (the developed portions of the six-county SCAG region) but with a significant increase in detail within Orange County and the immediately adjacent parts of Los Angeles County. Further, cities within Orange County have adapted and modified OCTAM for municipal applications. For example, Irvine's travel forecasting model starts with OCTAM as a base and adds significant levels of detail in and spatially adjacent to the City. Local travel demand models, in an increasing number of applications, are thus essentially regional models.

The applications of these models vary in temporal and spatial scale as well as in the level of impact on the designated land use and transportation systems. While a parallel Caltrans-funded research effort is focused explicitly on regional integrated land use/economic/transportation models, this effort will focus exclusively on local transportation models.

Local travel demand models provide data required as inputs to micro-simulation model analyses and infrastructure and control system design. These models also provide input in Caltrans planning, investment, and operations, including the review of the effects of local land use projects, general and specific plans, and other transportation system elements on State transportation systems (PS-12 RFP). The applications of interest include identifying the use of alternate models in local modeling, in general, and in traffic impact studies, in particular, and defining gaps that exist between the applications of these alternate models and the conventional travel demand models, especially in regard to impact assessment and mitigation.