Analysing the Variability of Streamflow and Suspended Sediment Concentration Using Time Dependent Intrinsic Correlation

Understanding the association between streamflow and sediment concentration time series may help to improve the modeling efforts of sediment transport processes in natural channels. This paper proposes an alternative approach to find such associations in multiple time scales based on Time Dependent Intrinsic Correlation (TDIC) method. The method is applied to the monthly streamflow and TSS concentration data from Kallada river in Kerala state, India, for the period 1980-2007. In this method, first both the streamflow and Total Suspended Sediment (TSS) concentration data are decomposed to different oscillatory modes called Intrinsic Mode Functions (IMFs) of specific periodicity. In the running correlation analysis procedure namely TDIC, the sliding window is adaptively fixed based on instantaneous periods obtained by the Hilbert transform of the obtained IMFs. TDIC analysis showed a strong long range positive association between streamflow and TSS concentration at intra annual to annual modes, which depict the influence of hydrologic dynamics in sediment transport processes at these time scales. For the inter annual modes of IMF3 and IMF4, the association is primarily positive, but there are alterations in the nature of association between the two for shorter time spells, particularly for time series segments of length less than 60 months. For low frequency inter annual mode IMF5, there exists a strong long range negative association between streamflow and TSS concentration, which depicted the influence of human interventions on the hydrologic variability of Kallada river at upstream regions of the gauging station.