Parameter uncertainty in rainfall-runoff modeling for polder system design

An approach is developed for incorporating the uncertainty of parameters fo r estimating runoff in the design of polder systems in ungaged watersheds. Monte Carlo Simulation is used to derive a set of realizations of streamflo w hydrographs for a given design rainstorm using the U. S. Soil Conservatio n Service (SCS) unit hydrograph model. The inverse of the SCS curve number, which is a function of the antecedent runoff condition in the SCS model, i s the random input in the Monte Carlo Simulation. Monte Carlo realizations of streamflow hydrographs are used to simulate the performance of a polder flood protection system. From this simulation the probability of occurrence of flood levels for a particular hydraulic design may be used to evaluate its effectiveness. This approach is demonstrated for the Pluit Polder flood protection system for the City of Jakarta, Indonesia. While the results of the application indicate that uncertainty in the antecedent runoff conditi on is important, the effects of uncertainty in rainfall data, in additional runoff parameters, such as time to peak, in the hydraulic design, and in t he rainfall-runoff model selected should also be considered. Although, the SCS model is limited to agricultural conditions, the approach presented her ein may be applied to other flood control systems if appropriate storm runo ff models are selected.