Development and Application of Regression Models for Predicting the Water Quality Performance of Permeable Pavement

The water quality performance of permeable pavement is influenced by many factors. The knowledge of the combined effect of multifactor on the performance of permeable pavement is vital to its design and construction. However, few quantitative relations are available in literature. Regression orthogonal composite design was adopted to develop models to predict the combined effects of multifactor on the performance of surface pavement layer and gravel layer of permeable pavement. The most commonly concerned factors, including rainfall intensity, inflow concentration of total suspended sediment (TSS), gradation of gravels, and thickness of the gravel layer, were selected. The interactions of these factors were also considered. The viability of the models was tested using analysis of variance (ANOVA), and the results showed the models for TSS removal rate of the surface pavement layer and gravel layer, Cd, Cu, Zn, TP, NH4-N, and NOx-N removal rate of the gravel layer were reliable and can be used for prediction purpose. More importantly, an integrated model was developed to predict the overall performance of permeable pavement, and a good performance was achieved (-2.43% to approximately 1.85%) through comparison with the measured values, illustrating its promising application. Then, the integrated model was compiled as modeling tools based on EPA SWMM (modified SWMM) and applied to a campus renovation scenario (three scenarios) assessment. Scenario 1 with higher pollutant removal rates is better than others. The results obtained demonstrated that the modified SWMM can respond to the changes of influencing parameters and will be beneficial for both practitioners and decision makers in permeable pavement design and construction.