Daily Rainfall Disaggregation Using HYETOS Model for Peninsular Malaysia

摘要

In this paper, we have examined the applicability of single site disaggregation model (HYETOS) based on the Poisson cluster model to disaggregate daily rainfall to hourly data using proportional adjusting procedure. In this study, the modified Bartlett Lewis Model (MBL) is fitted to the hourly rainfall depth from 1970 to 2008 available at the rain gauge station in Petaling Jaya. In addition, the synthetic hourly rainfall is generated by inputting the estimated parameters found based on MBL into the Hyetos model. The nature of occurrence of rainfall in Peninsular Malaysia comprising of very heavy rain during a short period of time contribute to the discrepancy between the synthetic data and the observed and expected data. When the disaggregated synthetic rainfall data is compared with the observed and expected data, it is found that the mean values for the three types of depths are quite closed; however, the synthetic data are quite different from the observed and expected depths when comparison is base on autocorrelation and standard deviation. The model is also validated by considering statistical property that was not used in the fitting procedure such as the extreme values. A comparison is also made between the extreme values based on the disaggregated model and the observed data. A bad fit in the extremes is found at all time scales considered, which are 1,6 and 12 hour levels of aggregation. An underestimation of the disaggregated values is evident at all time scales.