A multiple regression model for urban traffic noise in Hong Kong.

This article describes the roadside traffic noise surveys conducted in heavily built-up urban areas in Hong Kong. Noise measurements were carried out along 18 major roads in 1999. The measurement data included L10, L50, L90, Leq, Lmax, the number of light vehicles, the number of heavy vehicles, the total traffic flow, and the average speed of vehicles. Statistical analysis using the analysis of variance (ANOVA) and Tukey test (p<0.05) reveals that the total traffic flow and the number of heavy vehicles are the most significant factors of urban traffic noise. Multiple regression was used to derive a set of empirical formulas for predicting L10 noise level due to road traffic. The accuracy of these empirical formulas is quantified and compared to that of another widely used prediction model in Hong Kong--the Calculation of Road Traffic Noise. The applicability of the selected multiple regression model is validated by the noise measurements performed in the winter of 2000.