Comparison of sediment transport computations using hydrodynamic versus hydrologic models in the Simiyu River in Tanzania

• Complex and simple concepts for sediment transport give very different results.
• Complex not necessarily more reliable in data scarce areas.
• Sediment transport calculations in data poor areas should be treated with care.

This paper presents the results of a study that compares the sediment routing of the Simiyu River using the hydrologic model, Soil and Water Assessment Tool (SWAT) and the 1D hydrodynamic simulation software for Rivers and Estuaries (SOBEK-RE) model. Routing in SWAT is completed using the simplified Bagnold’s equation and in the SOBEK-RE model is undertaken using the Saint Venant equation. The upstream boundary conditions for the routing modules were derived from the subcatchments sediment yields that were estimated by SWAT using the Modified Universal Soil Loss Equation (MUSLE). The sediment loads extrapolated or interpolated from the sediment rating curve for the catchment outlet were used for calibration and validation purposes. The SWAT model predicted an erosion rate of 2.09 Mt/yr. The total sediment load transported to the main outlet of the catchment simulated by the SWAT and SOBEK-RE models was equal to 2.94 and 2.72 Mt/yr, respectively. Thus the models computed a net erosion in the channels of 0.84 Mt/yr (SWAT) and 0.63 Mt/yr (SOBEK-RE). When comparing the results of the models for the different reaches of the main channel and main tributaries, the models showed different results both in magnitude and in sign (erosion/deposition). However, in a situation where data is scarce (such as grain size, channel geometry), the more complex hydrodynamic model does not necessarily lead to more reliable results.