Hydraulic Modelling Of Unsteady Flow And Flood Routing In Nethravathi River Basin, India

Abstract

River systems need improved hydraulic models in order to better simulate river conditions at different points in time. Flooding is one of the changes in the flow characteristics of a river. Also, Channel roughness co-efficient is found to be a critical factor that is dynamic in nature influencing the flow characteristics of a river. The variation of roughness co-efficient based on the riverbed material and morphology affects the fluvial erosion and deposition altering the channel geometry. However, the parameter has been consistently considered as a constant value for the simulation in the numerical models. On the contrary, various experimental studies and analytical models have revealed roughness co-efficient to be one of the sensitive parameter in simulation of flow. Various models were developed for study to analyze the influences of the bed morphology (Roughness coefficient, the shape of the channel, channel slope) on discharge and water level at various locations of the river. It is also proved to be good means to assess the impacts for its capability to do the sensitivity analysis and to predict flood flow and inundated area. The aim of the present study is to simulate unsteady flow analysis using hydraulic model HEC-RAS for Nethravathi River basin in India for identifying the impact of variability in roughness co-efficient on the river-stage and discharge. Also, it is intended to assess the hydraulic response of the river using flood-routing analysis and to study sensitivity of geometric and computational parameters on model results and stability. A river length of 45 km of the Nethravathi river regime, Karnataka from Uppinangadi and Kumaradhara to Bantwal is considered for the study. HEC-RAS model was used for the simulation of surface water levels and discharge values. Manning’s roughness coefficient and river cross sections were defined for the calibration of observed river stage and discharge data. The model is built to examine the hydraulic response in Nethravathi River basin for a calibration period of 2007 - 2009 and validation period of 2010. The simulated model results provided good correlation between observed and simulated discharge and stage respectively. The error function values during validation are found to be marginally lower in ii comparison to calibration results. Thus, considering the overall values of the error functions for the gauging station, it is found that the model performs reasonably well for unsteady flow analysis. The variation of Manning’s roughness coefficient was observed to affect the river stage and thereby influencing the associated peak discharges. The coefficients of correlation for the developed rating curves showed the best fit for Manning’s n=0.070 flood plain. The peak discharge computation accuracy is approximately 80% in the calibration period and 88% in the validation period. The maximum water level computation accuracy is approximately 93.33% in the calibration period and 97.23% in the validation period. Depth of the flood is found to range between 0.1 m to 14.98 m. Velocity of flow along the whole river reach is found to vary between 0.01 m/s to 7.43 m/s. The maximum depth of flow in channel and floodplain is 7.75 m and 3.37 m respectively at downstream gauging station. However, the maximum velocity in the downstream of the river channel is found vary between 0.28 m/s to 1.71 m/s consequently, the river stage and discharge along the cross-section of flow were disturbed resulting in the flooding of river banks and inundation of low lying areas. Flood inundation map shows the spatial variation of the flood in the floodplains of the Nethravathi basin. Flood water flows over the riverbank in the upstream of the Uppinangadi gauging station, near the Nethravathi-Kumaradhara river confluence and at the meandering section in upstream of the Bantwal gauging station. The Manning’s roughness coefficient, normal depth, time step and θ-weighting parameter were considered to test the model sensitivity. The results from the sensitivity analysis showed that the model is very sensitive to the choice of Manning’s n. Reducing Manning’s n will decrease magnitude of peak and stage, and reduce the total inundation extent. The output is more sensitive to Manning’s n than time step, θ- parameter, cross-section spacing and normal depth. This indicates that the choice of friction coefficient can to some extent overshadow the uncertainties related to insufficient geometry data and the numerical solution.