Author(s)

Sivarajah Mylevaganam^*, Ting Fong May Chui, Jiangyong Hu

Department of Civil and Environmental Engineering, National University of Singapore, Singapore, Singapore.

Under the possible hydrological conditions, with a design hyetograph of 3-month average rainfall intensities of Singapore, multiple regression equations on hydrological processes, specifically on overflow volume, average vertical ex-filtration rate and horizontal flow coefficient, of a soak-away rain garden are established based on simulated results of a mathematical model. The model that is based on Richard’s equation is developed using COMSOL Multiphysics. The regression equation on overflow volume and the regression equation on log of horizontal flow coefficient show a very strong relationship with the independent variables (saturated hydraulic conductivity of the filter media, saturated hydraulic conductivity of the in-situ soil, depth to groundwater table, and surface area of the soak-away rain garden). The coefficients of determination of the fitted equations on overflow volume and log of horizontal flow coefficient were 0.992 and 0.986, respectively. However, the regression equation on average vertical ex-filtration rate has high p-values (p-values > significance level, α = 0.01) for saturated hydraulic conductivity of the in-situ soil and surface area of the soak-away rain garden. Thus, forward stepwise regression was used to develop the best regression equation on average vertical ex-filtration rate with saturated hydraulic conductivity of the filter media and depth to groundwater table. The coefficient of determination of the fitted equation was found to be 0.911. These easy to use regression equations will be of great utility for local mangers in the design of soak-away rain gardens.

COMSOL Multiphysics, Soak-Away Rain Garden, Overflow Volume, Multiple Regression, Ex-Filtration

Mylevaganam, S. , Chui, T. and Hu, J. (2015) Multiple Regression Model of a Soak-Away Rain Garden in Singapore. Open Journal of Modelling and Simulation, 3, 49-62. doi: 10.4236/ojmsi.2015.32006.