Aggregation simulation model of flow and rainfall series

Abstract

Synthetic hydrology series is useful for evaluating the consequences of water supply management decisions and reservoir design. The main objective of this study is to identify and confirm the best model in flow and rainfall simulation. The study covers the application of aggregation and disaggregation methods for flow and rainfall stochastic simulation. In general, the application of various periodic models for the flow simulation was mostly successful. The application of disaggregation models was found to yield sufficient performance and competitive to the periodic models. It has been proven that the transformation does not always guarantee improvement in the candidate models performance. The Periodic Autoregressive of Order One (PAR (1)) model is the best performer for the monthly and annual flow simulation using periodic models for both untransformed and transformed series. The Valencia and Schaake (VLSH) model is the robust model from disaggregation group for the monthly and annual flow simulation. Simulation for monthly and annual rainfall series shows that the VLSH model is the best performer to produce sufficient results for both untransformed and transformed series. The results from this study are based on investigation from graphs and frequency analysis. The outcome of study has potential to assist the water engineers and consultant in making decisions for the operation of the water resources systems. It is suggested that the rainfall simulation should be applied in water resources planning because observed flow series are subjected to disturbance due to development.