Uncertainty exists in statewide travel demand forecasting and land use models. Both input data and adopted model parameters can vary from their true values because of model misspecification, imperfect input information, and innate randomness of events. The aim of this paper is to study the sensitivity of VMT outputs of an integrated statewide travel demand and land use forecasting model system for Indiana to plausible amounts of variations on travel model parameters and input data. The Land Use in Central Indiana 2 (LUCI2) model is used in conjunction with the statewide travel demand model to predict travel demand over time through various scenarios. Results indicate that VMT outputs are most sensitive to trip distribution function parameters and trip production rates. Population growth rates and trip assignment parameters don't have a high degree of influence on VMT outputs.
Abstract