Leveraging Transferrable Bid-Rent Functions to Forecast Land Use in a Small Kansas Region 

Date and Time: 8/30/2022
Location: North Star

Presenter: Colby Brown, , Manhan Group, LLC

  

 

 

PRESENTATION DESCRIPTION

Many small and medium-sized regions around the country are faced with the challenge of working with land use data and projections that do not provide the information needed by state-of-the-practice travel demand models. Even basic trip-based (i.e. four-step) traffic forecasting models increasingly require segmentation of households by income and household size (or auto ownership) in order to perform trip generation, whereas most future land use scenarios only provide the general numbers of housing units added (perhaps by single-family and multi-family categories). In the past, highway-only models have circumvented this challenge by using vehicle-trip generation rates borrowed from ITE or other sources; yet many of these rates have been found to be based on outdated, small-sample traffic count studies, and the limitation of forecasting only cars leaves regions unable to adapt to changing mode preferences and policy priorities. Another popular solution is the use of lookup tables to predict marginal shares of households by income and size, yet this process still requires some forecast of median income and/or average household size, and is generally not sensitive to changes in the demographics of regions as a whole moving forward. Faced with these challenges many regions simply hold household demographics constant in their forecasts, though they know major population shifts are coming.

We present an alternative approach based upon nationally-available bid-rent functions estimated by the author, which can easily be transferred to any county in the USA and used to translate residential land use (i.e. housing units by type) into household demographics for future scenarios. This approach was applied for the Flint Hills MPO as part of a long-range forecast in the greater Manhattan, Kansas region. This project may also serve as a case study of R-Land, an open-source, user-customizable land use modeling software library for R software developed by Manhan Group LLC.

SPEAKER BIOGRAPHY