Lead Presenter Biography

Mr. Schiffer has over 39 years of experience in transportation planning. He has served for the past 5 years as President of FuturePlan Consulting in Tallahassee, Florida. He specializes in travel demand modeling, long-range transportation plans, travel behavior and origin-destination travel surveys, site impact traffic studies, and forecasting multi-modal corridor travel demand. He has held leadership roles and volunteer service positions for the Transportation Research Board (TRB), the Institute of Transportation Engineers (ITE), and the American Planning Association (APA). His experience encompasses transportation planning studies in 28 states and commonwealths for national, statewide, regional, municipal, subarea, and private sector clients. Rob is also an affiliate of Metro Analytics and Adjunct Faculty for the Department of Urban & Regional Planning at Florida State University.

He has worked on 8 different NCHRP studies, including Principal Investigator on 3 of these, and Co-Author of TRB Transportation Research Circular E-C075. Rob was Conference Chair of the 2011 TRB Transportation Planning Applications Conference; Technical Chair of the 2009 TRB Transportation Planning Applications Conference; Technical Chair of the 2018 TRB Tools of the Trade Conference; Planning Committee Member for 2016 and 2022 TRB Tools of the Trade Conferences; and Planning Committee Member for 2022 TRB Statewide Model Peer Exchange. He is a Member of TRB Committee AEP15-Transportation Planning Analysis and Application.

Rob received his MS in Urban & Regional Planning from Florida State University in 1984 and a
BA in Geography & Urban Studies from Memphis State University (now University of Memphis) in 1982.

Co-Authors

Presentation Description

Since the 1960s, much effort has been expended on the development, calibration, and/or validation of travel demand forecasting models. The norm has been to develop an updated model every 5 to 10 years, for a recent base year that often coincides with the Decennial Census. Once a model is deemed valid, forecasted land use and demographic growth is combined with committed transportation projects to identify needs for 20 to 25 years in the future.

What has rarely been done is a comparison of past traffic forecasts to recent traffic counts. Part of the problem in the past was technological… file storage was limited, and travel demand modeling software was fluid enough to make older files difficult to access. This process also required the passage of 20 to 25 years so that traffic counts would become available for previous model horizon years.

This research study, funded in part by a small grant from the American Planning Association’s Transportation Planning Division, looks at traffic forecasts produced in the late 1980s and 1990s for horizon years ranging from 2005 to 2020. While the author was directly responsible for most of these forecasts, file storage from this time generally consisted of floppy disks which are no longer supported by most personal computers. Most files were eventually trashed or left behind with prior employers. Fortunately, the author has maintained an extensive library of printed reports from MPO Long-Range Transportation Plans and highway corridor studies to provide a source for model comparisons.

While the practice of travel demand modeling has advanced greatly since the 1980s and 1990s, some of the same issues are still relevant to today’s state-of-the-practice models. All models still require future assumptions on land use, demographics, and committed transportation projects. This study identifies some cases where traffic forecasts were incredibly close to recent traffic counts though many show a significant variation between projected traffic estimates and actual counts. There are a number of explanations for this variability, including demographic forecasts, assumed transportation projects, external trips, model validation error, and even the variability of traffic counts. Demographic forecasts cannot predict future economic events such as the Great Recession or the Covid pandemic. Thus, the recent emphasis on uncertainty in travel demand models is clearly warranted.

This presentation will describe numerous comparisons between traffic forecasts and counts, along with related model assumptions and potential causes of inaccuracies. Attendees will learn about conditions where forecasts were most accurate, along with considerations on assumptions most likely to impact the longevity of travel demand forecasts. 

Presentation File