Stochastic streamflow analysis and simulation using disaggregation model

Abstract

Synthetic hydrological series is beneficial for the water resources planning, management and design purposes such as evaluating water supply management decision and reservoir design. Stochastically generated data can be used in any fields of study that use historical data. It would produce a better result for any application where the result depends on the time series pattern. Therefore, this study examines three stochastic disaggregation models that are capable of reproducing statistical parameters especially means and standard deviation of historical data series. The objectives of this study are to test, identify and confirm the best stochastic disaggregation model in generating synthetic data series. Stochastic Analysis Modeling and Simulation (SAMS-2000) is used to generate the synthetic hydrological data series. The method is applied for single site cases and comparison is made between three disaggregation models namely Valencia and Schaake model, Mejia and Rouselle model, and Lane model. The simulations of monthly streamflow are carried out for five stations from Kedah, Perak and Selangor. The comparison of results shows that Valencia and Schaake (VLSH) model is the most satisfactory and robust model that preserves both monthly and annual statistical parameters of the historical data sequences. This is true for both untransformed and transformed series. Therefore, it is recommended to use VLSH model for simulation of water resources in study area.