3.4 Sediment rating curves
Strong correlation is often observed between suspended sediment concentration and discharge in most streams. Compared to large river systems that usually contain an abundance of materials, the suspended load of small mountain streams usually depends on episodic events that transport fine materials from banks and upland areas Thus, suspended sediment concentration depends on both supply of sediments and discharge (Yaksich & Verhoff, 1983; Zhang et al., 2021). Nonetheless, suspended sediment concentration and discharge can be plotted to create what is known as a sediment rating curve. One benefit of developing a sediment rating curve is that it can be applied to interpolate missing data during the observation period (Asselman, 2000). Here, the sediment rating curve is a power function expressed as:
SSC = a×Qb
where SSC is in mg/L, Q is the water discharge (in m3/s), and a and b are fitting coefficients.
When b > 1, the increase in sediment volume per 1 unit of flow volume can be nonlinear. Such situations imply that more than half of the sediment load is carried by high flows that account for less than 15% of the water volume or less than 5% of the period of measurement.
Another characteristic of suspended sediment concentration during high flow is their increased variability. Both factors indicate that more sampling should take place during periods of high discharge. However, one common limitation of suspended sediment data from small mountainous catchments is the lack of measurements conducted during high flows. Storms make access difficult and measurements hazardous. Nevertheless, high-resolution data collected during high flows are essential for the development of good sediment rating curves. Due to the infrequency of high flows and logistical problems, collecting adequate high flow measurements cannot be achieved through hand sampling alone.
Suspended sediment rating curves were determined for Ming Yong glacial catchment using discharge and suspended sediment concentration data between August 2013 and July 2017. The rating curves were fit by nonlinear least-squares curve fitting using the Levenberg-Marquardt (L-M) algorithm produced in Origin 2021b. One of the major limitations of the sediment rating curve is the assumption that the rating coefficients will remain constant (Zhang et al., 2021). As mentioned by Yaksich and Verhoff (1983), small mountainous catchment often results in “event response” streams, which causes large scatter between sediment concentration and discharge and hence poorer relationships between suspended sediment concentration and discharge. The low accuracy of the sediment rating curves may be attributed partly to hysteretic effects, where at a given discharge, the sediment concentration on the rising and falling stage of the hydrograph differs (Khanchoul & Jansson, 2008). Separate rating curves are needed for the rising and falling limbs to account for seasonal variations so as to enhance the accuracy of the estimated suspended sediment concentration (Khanchoul & Jansson, 2008). In this study, the hydrological year is defined as August to July, with the rising stage (February-July) and falling stage (August-January in the next year).