UC Davis Institute of Transportation Studies

To help guide applications of more advanced models in policy studies, this paper presents an evaluation of model accuracy and induced demand in an integrated land use and transportation model, the 2000 Sacramento MEPLAN model. The model is currently used by the region's metropolitan planning organization (MPO) for land use projections. The accuracy of the model is assessed with validation tests that show how well the model predicts observed data over a ten-year period that are not used to estimate or calibrate the model. Forecasts are compared to observed 2000 land use and travel data to identify the magnitude of model error resulting from model functional forms and parameter specifications. Forecasts are also used to identify the model's representation of induced demand and to estimate actual induced demand. The model's representation of induced demand includes the change in land use (i.e., development and allocation) and travel (trips, distance, mode choice, and time) that results from new transportation capacity. The results illustrate how validation tests can be used to improve the application of uncertain models in policy studies requiring absolute accuracy such as conformity analysis (emissions budgets) and environmental impact analysis (level of roadway service).