NITC Research Brings Transportation Impact Assessment into the 21st Century
Learn more about this research by viewing the two-page Project Brief, related presentations, and the full Final Report on the Project Overview page.
As cities aim to promote sustainable, multimodal growth, sometimes the way we go about development review processes can create barriers to achieving the results we want. Some of the methods we have inherited, while still useful, have distinct limitations.
NITC dissertation fellow Kristina Currans took on this challenge in her doctoral research project, Data and Methodological Issues in Assessing Multimodal Transportation Impacts for Urban Development.
The guidelines for evaluating transportation impacts of new development were originally published in 1976 by the Institute of Transportation Engineers (ITE). Decades later, we’re still using essentially the same processes all across the U.S. and Canada, and these methods—which harbor a lack of sensitivity to urban contexts—could use an update.
Currans graduated from Portland State University in the spring of 2017 with a Ph.D. in transportation engineering. She starts work this coming fall as an assistant professor of planning at the University of Arizona.
She will be moderating a session on trip generation at the upcoming Transportation and Communities Summit 2017, this September at Portland State University. A guide to help practitioners use this research, in the form of a one-page technical brief, will also be available at the Summit.
Her final report, Issues in Urban Trip Generation, includes a thorough discussion of alternative approaches to development-level review and provides concrete recommendations to improve practice.
Sometimes, the recommendations save money. For instance, Currans found that separating retail and service land uses into three categories (heavy goods, convenience uses, and everything else) performed nearly as well as the ITE’s more extensive taxonomy.
“ITE has over 170 land uses. If you simplify down to about 3 land uses, that performs nearly as well and is way cheaper. There’s a little bit of information loss, but it’s tiny,” Currans said.
In other instances, Currans’ recommendations yield better accuracy. If land use has been considered too much by ITE, demographic information has been considered too little.
“We cannot ignore demographics. Demographics influence travel behavior but have not been included in models as much as they should be. For example, a grocery store in a low-income area could potentially be overcharged on its development costs,” Currans said.
She also identified shortcomings that can be addressed in ITE’s trip generation data.
When spatial, social, and temporal contexts are ignored, the bias, in many urban cases, is extensive and substantially overestimates demand. This can result in “phantom trips,” vehicle trips which are estimated but don’t actually happen.
“It’s problematic to rely on the ITE data as a single metric. We should be moving toward thinking about impacts in a more balanced fashion,” Currans said.
She demonstrates that planners would be better off evaluating development impacts using multiple factors such as person trips, mode shares, trip lengths and vehicle miles traveled, rather than just ITE’s trip rates.
Currans’ dissertation consists of four separate papers, presented as chapters. The first paper (Chapter Two) evaluates state-of-the-art methods and identifies gaps in the literature.
The second paper evaluates the existing land-use taxonomy, and considers the statistical impacts of the age of data on vehicle trip rates for retail and service land uses—spoiler alert: it’s significant, and ITE’s older data tend to have higher trip rates.
The third explores the use of ITE’s data as a baseline data for other models. The results point to the need to consider demographics in site-level transportation impact analysis, particularly to estimate overall demand.
The fourth paper evaluated the conventional methods (ITE’s Handbook) of urban trip generation, and quantified the potential bias. Results indicate the compounding overestimation of automobile demand may inflate estimation by more than 100 percent, even in contexts where ITE data should be applicable (suburban areas with moderate incomes).
The final chapter discusses implications for practice, including alternative approaches to development-level review.