Temporal and spatial variability of suspended sediment rating curves for rivers draining to the Ethiopian Rift Valley

Quantification of suspended sediment dynamics is a proxy for the intensity of soil erosion processes in a catchment. However, direct measurement of suspended sediments is difficult. Long-term-based discharge-suspended sediment relationships often fail to provide sufficiently accurate results due to the temporal variations in sediment supply. This study evaluates temporal variations in suspended sediment-discharge relations for six stations in two selected catchments within the Gamo Highlands in the southern Ethiopian Rift Valley. During a three-year period (2018-2020), discharge was monitored at a 10-min interval using a TD-diver, while 1938 samples of suspended sediment concentration were taken at specific events. Observed suspended sediment concentration varied between 0.04 and 130 kg m(-3) for discharge ranging between quasi-zeros to 339 m(3) s(-1). The results show that pooled annual datasets lead to unsatisfying sediment rating curves and large errors in estimated sediment load. However, stratification of the data into three rainy season periods strongly improved the performance of the sediment rating curves. The development of a vegetative cover throughout the rainy season largely controls sediment supply from hillslopes to the river channel and hence the shape of the sediment rating curves. Furthermore, localized bank erosion and gully head dynamics also lead to important interannual changes in sediment rating curves within one river system, as well as between different rivers. Thus, attention should be given to possible errors due to high temporal and spatial variability while using discharge-suspended sediment relations to quantify sediment load.