Seasonal Variations of Manning’s Coefficient Depending on Vegetation Conditions in Tärnsjö, Sweden

Abstract: Hydrological models are used widely and they demand for multiple input variables and observations. One of those variables is Manning’s roughness coefficient. In the current literature the variability of the coefficient poses an unknown uncertainty. This study examines a small river channel located in central Sweden, and aims to determine the variability and uncertainty of the roughness coefficient during diverse vegetation conditions within the channel. During multiple field visits to the location, slope, water level and cross-section examination is performed. With numerical simulation, discharge and roughness coefficients are obtained. With the hydraulic model (HEC-RAS), stage-discharge rating curves are produced and extrapolation is applied to obtain high flows. Manning’s roughness coefficients and their uncertainties are assessed by two different approaches. Determining the coefficient in a simplified sensitivity analysis by using Manning’s equation and calibrating HEC-RAS while applying Mean absolute error (MAE) calculation. The calculated roughness coefficients presents higher range when using Manning’s equation (summer vegetation conditions – 0.2, winter vegetation conditions – 0.095). On the contrary MAE provides values closer to each other (summer – 0.15, winter – 0.11). The obtained results indicate a high variance between summer and winter vegetation conditions, producing 38 cm water level differences during high flows using Manning’s equation and 6 cm difference using the calibration of the model in HEC-RAS. These results confirm that the roughness coefficient cannot be assumed to be constant throughout different seasons as had been assumed widely when applying hydrological modelling. Throughout the study innovative approaches and methods (e.g. back-calculating from Manning’s equation and calibrating HEC-RAS based on observed water levels) are used in order to determine the consequences of ignoring the variability of the roughness coefficient. Due to the study, one can derive that vegetation needs to be considered in having an important impact on the varying roughness coefficient value and it cannot be left as a constant value within hydrological models.