Comparing Land Use Forecasting Methods: Expert Panel Versus Spatial Interaction Model

Problem: Legal requirements and good planning practice dictate that land development induced by major highway investments be forecasted. Two forecasting methods, the first qualitative and based on expert judgment and the second quantitative and based on formal spatial interaction models, are often presented as equivalent. Purpose: We aim to extract lessons about the strengths and weaknesses of the two methods from a case study of a controversial highway, the Intercounty Connector (ICC), in the suburbs north of Washington, DC. Methods: We compare forecasts of induced development obtained using both methods and judge their reasonableness against the empirical literature. Results and conclusions: The two methods gave dramatically different results. The subjective judgment of experts predicted small impacts, on average, compared to a simple spatial interaction model. Also, subjectively forecasted impacts were limited to lands near the new facility, while modeled impacts rippled out across a much larger area. The subjective method seemed to give too little weight to accessibility effects and too much to zoning constraints, while a simple spatial interaction model seemed to do the opposite. Takeaway for practice: Where time, budget, or data limitations preclude the development of state-of-the-art integrated land use and transportation models, we conclude based on this case study that the best approach is to combine simple models and expert judgment. Expert panels can be used to check model inputs against local knowledge and to adjust outputs in light of factors otherwise unaccounted for. Conversely, model outputs can be used to check expert opinion for inconsistency with known land use–transportation relationships. Research support: None.