ASSESSMENT OF FLOOD PEAK INCREASE DUE TO URBANIZATION

Abstract

By using the kinematic wave method, the impact of urbanization on the flood peak increase for an overland plane, an idealized drainage basin, and a natural drainage basin are assessed. For the overland plane, explicit expressions for the estimation of time of concentration and the corresponding peak discharge are derived. These expressions are in terms of the physical characteristics of the plane and the rainfall. Marning's and Darcy-Weisbach resistance coefficients are used in the derivation. The expressions that use the Darcy-Weisbach resistance coefficient can be applied to flow regimes from laminar to curbulent. Further, a method for determining the critical rainfall duration that associates with the maximum peak discharge, and equations for determining the maximum peak discharge and the corresponding contributing area are also developed. To further the understanding of the rainfall-runoff process on an overland plane, experiments have been conducted at the Nanyang Technological University (NTU) on concrete and Cow grass surfaces. By analysing the runoff hydrographs, the resistance coefficients for the two surfaces are evaluated. For the grass surface, the analysis includes the infiltration process of Hortonian and saturated overland flows. A comparison of estimates from the time of concentration formulas with the experimental data shows that for the formulas that do not incorporate the effect of rainfall intensity, they are valid for a limited range of rainfall intensities. For the formulas that do incorporate the effect of rainfall intensity, given the appropriate value of resistance coefficient, they offer reasonable agreement with the experimental data. In the assessment of flood peak increase for an overland plane, the individual effects of resistance coefficient, infiltration and depression storage have been evaluated. The evaluation shows that the reduction in resistance coefficient causes the greatest increase in the flood peak, whereas the reduction in depression storage has the least effect on the flood peak increase. For the idealized drainage basin comprising two overland planes and a drainage channel, the assessment of flood peak increase has been carried out by means of the Urbanization-Peak Discharge (UP) model developed in this study. In the assessment, the degree of urbanization is expressed in terms of the percentage of area made impervious, and the percentage area sewered. The impact of urbanization on the flood peak increase is expressed in isopleths, which are lines of equal ratio of flood peak after and before urbanization. Since the peak flow under the undisturbed land use condition is used as a basis to calculate the peak flow ratio, the undisturbed land use condition has a significant influence on the pattern and magnitude of the flood peak increase. The increase in flood peak is greater for storms of shorter recurrence interval, and for storms that are associated with the semi-arid region. For a grass-based basin subjected to the Singapore 2-year rainfall, the increase in flood peak is between two to three-fold. The effects of basin shape and basin scope were found to primarily alter the isopleth pattern, with little influence on the maximum peak flow ratio. For a forest-based basin subjected to the Singapore 2-year rainfall, the increase in flood peak is five times or more. Analysis of rainfall and runoff data of the Upper Bukit Timah Basin shows that urbanization has significant effect on flood peak and flood volume increases, if the Basin is under the dry antecedent soil moisture condition. Further, by developing a computer model of the Upper Bukit Timah Basin using the HEC-1 program or the "BT model", the model results show that the flood peak increase is greater for developed area with higher degree of imperviousness. It also shows that the effect of spatial distribution of urbanized area on the flood peak increase is not great. The results generated by the UP and BT models are similar to each other, and are consistent with the published results from earlier studies, which were based on observed or synthesized flood data.