Dynamic aspects of suspended-sediment-concentration recession curves

Abstract

Understanding and modeling suspended sediment concentration (C) dynamics in river channels are essential for many theoretical as well as practical purposes. The sediment rating curve approach that relates C with river discharge (Q) is among the most followed modeling approaches, mainly because of its simplicity and effectiveness. Although studies in the past have suggested the possibility of C-Q relationship for a basin not being unique, how the relationship evolves with time is not yet fully understood. In this study, we limit our analysis to recession periods when both C and Q continuously decrease with time and show that the coefficient of C-Q power-law equation changes across recession events. Furthermore, we propose to express -dC/dt as a function of C for recession periods: -dC/dt=fCe. The exponent e is assumed to be related to static catchment characteristics and the coefficient f to dynamic factors, namely catchment-water storage and sediment availability condition. Assuming the dynamic factors controlling f to evolve slowly with time, we propose that it is possible to predict f for a recession event considering only past C data. Once f is estimated, we can predict C for anytime during the recession event. The advantage of the proposed model is that it can predict future suspended sediment concentration at a river cross section during a recession event. We applied the proposed model to 35 US Geological Survey basins and found the 25th, 50th, and 75th values of NSE, respectively, to be 0.55, 0.64, and 0.78, and the corresponding values for the rating curve model - 0.05, 0.01 and 0.20. The results here essentially highlight the importance of recognizing the dynamic nature of suspended sediment concentration-discharge relationship.