Determination of Effective Discharge Responsible for Sediment Transport in Cauvery River Basin

Abstract

The mechanism of sediment transport is mainly governed by surface water flow within the river basins. Excessive sediment transport plays an important role in reducing the carrying capacity of channel networks, storage capacity of reservoirs/dams. An important task for most of the hydrologists is to determine the reliable stream flow estimate which causes majority of the sediment transport within river basins/stream channels. The transport effectiveness of a stream flow event of particular magnitude in carrying a sediment load is defined as the product of the effect of that event (i.e. sediment transport rate corresponding to the stream flow event) and the frequency with which the event occurs. This approach is famously known as magnitude frequency analysis (MFA). MFA has been widely used to compute “effective discharge” which is considered as the stream flow that is responsible for transportation of majority of the sediments from a river basin or catchment over a long period of time. In MFA, the stream flow at a location is assumed to follow a continuous probability distribution (e.g., Normal, Lognormal, Exponential, Gamma, Generalized Pareto and Poisson) whereas the sediment transport is described by a power law function between stream flow and sediment rate. Subsequently, a transport effectiveness function is constructed by taking product of stream flow distribution with power law function. Finally, the effective discharge can be obtained by maximizing the transport effectiveness function with respect to stream flow. In this paper, effective discharge estimates were determined for 12 stream gauges in Cauvery river basin by fitting appropriate continuous probability distributions (Normal, Lognormal, Exponential, Gamma, Generalized Pareto and Poisson) and assuming power law relationship for sediment transport. Kolmogorov–Smirnov test (KS test) at 1% significance level was tested for fitting probability distributions to daily stream flow data at each of the gauges. Results indicated that all of the above distributions failed to fit stream flow data at all the gauges. However, following the previous literature, the daily stream flow data at every gauge was assumed to follow Log-normal distribution and corresponding effective discharge was determined. Further, recurrence interval was calculated for the effective discharge estimate at the each of the gauge. The results from this study can find use in effective planning and functioning of dams/reservoirs.