Requirements and best practices for traffic noise model validation

Abstract

The exercise of validating a predictive model typically consists of modeling an existing real-world condition and predicting some resulting quantity of interest, and then comparing the predicted results to actual measured values for the same condition within a specified margin of error. In the case of road traffic noise models this would include modeling the characteristics of the existing roadway condition (geometry of the roadway network; number, vehicle type and speed of vehicles using the roadways; pavement types; intervening terrain; etc.) and comparing the resulting predicted noise level to actual noise measurements for the same conditions. In this paper the authors will provide a concise background of the validation purpose and process as well as a review of agency requirements and best practices for traffic noise model validation procedures.