Nonlinear prediction of river flow in different watershed acreage

Abstract

Analysis of river flow is important in water resources management. For this purpose, river flow data at three measurement stations, Lubok Paku at Pahang River, Ladang Victoria at Muda River and Tanjung Tualang at Kinta River in Peninsular Malaysia have been studied using chaos theory. The analysis involved rivers which are located at different watershed acreage. Pahang River watershed are categorized as large, Muda River as medium and small for Kinta River based on watershed area. The analysis shows that the size of a large watershed has high average annual river flow while a small acreage watershed has low average annual river flow. Referring to the matter, analysis and prediction of river flow was carried out using nonlinear prediction method based on chaos theory. This method involves the reconstruction of phase space and prediction. Reconstruction of phase space has been performed by reconstruction of a single variable (river flow data) in m-dimensional phase space. Then, the prediction is done by using local linear approximation method. Overall, the prediction result is closed to agreement for all stations with high correlation coefficient. These results proved that nonlinear prediction method is suitable for river flow prediction of rivers that located in the area of different watershed acreage which is significant with different average annual flow. The results also indicate that the presence of chaotic behavior of the system as well as more than six variables affected the dynamics of the Pahang River, Muda River and the Kinta River.