Multiscale Analysis of Suspended Sediment Concentration Data from Natural Channels Using the Hilbert-Huang Transform

Abstract

The sediment transport processes in natural channels are governed by numerous physical processes having specific time scales. Identification and selection of multiple time scales associated with sediment concentration time series data may help to improve the modeling efforts of sediment transport processes in natural channels. Owing to the nonlinear and non-stationary character of the considered hydrologic time series datasets, this paper first applies the Complete Ensemble Empirical Mode decomposition with Adaptive Noise (CEEMDAN) algorithm to decompose time series of biweekly streamflow and Total Suspended Solid (TSS) concentration data from Basantpur station in Mahanadi river for the period July 1975 to February 1980 to understand the multiscaling behavior of sediment transport. Then the resulting orthogonal modes namely Intrinsic Mode Functions (IMFs) are subjected to Normalized Hilbert Transform (NHT) to obtain instantaneous frequencies and amplitudes. The statistical significance test performed upon IMF components showed that the semi-annual and near annual modes are statistically significant for both the series while inter annual modes are also significant for TSS concentration time series. The cross correlation analysis between the IMFs of the two series for the entire time domain identifies the link between streamflow and sediment load transport, with the highest correlation (0.85) between inter-annual modes of the series. The Hilbert spectral representation of both time series showed high intermittency and frequency modulation for the lower modes, which depicted the nonlinear and non-stationary characteristics of the datasets. Further, it is observed that the dominant frequency (i.e., concentration of high amplitudes) shows a time varying behavior. This study also proves the fractal character of the TSS time series which is the primary suspect for multiscaling behavior and the turbulence observed in the lower modes of TSS concentration time series.