Detection of Sediment Trends Using Wavelet Transforms in the Upper Indus River

Sediment load trends play a key role in modelling either river morphology or reservoir sedimentation. In this study, suspended sediment concentration (SSC) series of four representative gauging stations of the Upper Indus Basin (Yugo, Dainyor, Bunji and Besham Qila) were selected and updated from a vast network of hydro-meteorological stations being operated and maintained by Water and Power Development Authority (WAPDA) of Pakistan from the 1960s up to now. The temporal variations in the series were analysed using the wavelet transform (WT) method. The WT method disclosed the temporal and frequency information for trend estimation analysis by decomposing data on several levels. The results of the combined methods, WT and Mann-Kendall (MK) trend tests, revealed that the annual sediment time series, available since the 1960s for some stations, exhibited a statistically insignificant trend due to statistically significant intra-annual (monthly) shifts. Generally increasing trends in the winter months and decreasing trends in summer months for major sub-catchments of Upper Indus Basin (UIB) were detected. However, the study also proved that the identified intra-annual or monthly shifts in upper sub-catchments were being neutralized as the sediment progressed downstream. This study of variations in sediment trends was required for constructing sediment budgets and sustainable operations of existing and planned future water storage along the tributaries and the main stem of the Upper Indus River.