Influence of Road Surface Distresses on Pavement Roughness Considering Spatial Autocorrelation Effects

This paper investigates pavement distresses to determine which has the most influence on pavement roughness. For this purpose, the stepwise regression method is used. In addition, the spatial autocorrelation of the data on pavement roughness as a new weight vector parameter considering the geographical location is evaluated and introduced into the regression method. The pavement distress and International Roughness Index (IRI) dataset of the road network managed by Iran’s Road Maintenance and Transportation Organization, collected from 2020 to 2022, is used to select the best predictor variables for each road type in the regression. The results show that using spatial autocorrelation in the method reduces the residual standard errors and increases reliability because of improved residuals’ normality in comparison with the method without considering the spatial autocorrelation. The residuals’ standard errors of the regression considering the spatial autocorrelation are decreased by about 30%, 27%, 25%, and 20% for main roads, secondary roads, freeways, and highways, respectively. Also, the sensitivity analysis reveals that IRI is more affected by increasing the patching area and rutting value for secondary roads, while growing alligator cracks have the most influence on IRI reduction for main roads.