Multi-Scale Analysis Of Streamflow Using The Hilbert-Huang Transform
chongli di,xiaohua yang,xuejun zhang,jun he,ying mei
DOI: https://doi.org/10.1108/HFF-04-2013-0110
2014-01-01
International Journal of Numerical Methods for Heat & Fluid Flow
Abstract:Purpose - The purpose of this paper is to simulate and analyze accurately the multi-scale characteristics, variation periods and trends of the annual streamflow series in the Haihe River Basin (HRB) using the Hilbert-Huang Transform (HHT).Design/methodology/approach - The Empirical Mode Decomposition (EMD) approach is adopted to decompose the original signal into intrinsic mode functions (IMFs) in multi-scales. The Hilbert spectrum is applied to each IMF component and the localized time-frequency-energy distribution. The monotonic residues obtained by EMD can be treated as the trend of the original sequence.Findings - The authors apply HHT to 14 hydrological stations in the HRB. The annual streamflow series are decomposed into four IMFs and a residual component, which exhibits the multi-scale characteristics. After the Hilbert transform, the instantaneous frequency, center frequency and mean period of the IMFs are obtained. Common multi-scale periods of the 14 series exist, e. g. 3.3a, 4 similar to 7a, 8 similar to 10a, 11-14a, 24 similar to 25a and 43 similar to 45a. The residues indicate that the annual streamflow series has exhibited a decreasing trend over the past 50 years.Research limitations/implications - The HHT method is still in its early stages of application in hydrology and needs to be further tested. Practical implications - It is helpful for the study of the complex features of streamflow.Social implications - This paper will contribute to the sustainable utilization of water resources.Originality/value - This study represents the first use of the HHT method to analyze the multi-scale characteristics of the streamflow series in the HRB. This paper provides an important theoretical support for water resources management.