Required accuracy of rainfall data for integrated urban drainage modeling

It is standard practice in integrated urban water management to apply mathematical models of the total drainage system for calculating the frequency of occurrence of critical states in the receiving water body. The model input for such computations are long term time-series of rainfall data. However, it is inevitable that those rainfall data measurements deviate from reality. This is a result of inaccuracy of the measurement devices, errors in data transmission, local meteorological effects, etc. In this work we investigate the effect of such uncertainty in the rainfall data on the return period of the occurrence of oxygen depletion in the river due to the drainage of storm water. The errors in the rain data measurements are simulated by means of both stochastic and deterministic models. A comparison of the results obtained from the erroneous data series against the reference data reveals the small effect of random deviation in rain measurements. Only a constant and significant offset of the measured data (greater 20%) has an equally significant effect on the modeling result.