Study between short- and long-reach flow simulations with 1d hydrodynamic model in a floodplain environment

Abstract

Bangladesh is a floodplain deltaic country. There are around 700 rivers in the country with an extensive seasonally flooded floodplain. For proper management of these rivers and the natural resources of the floodplains, it is important to understand the hydrodynamic behavior of the rivers. The Atrai River is such a river where river-floodplain interaction is very prominent. Thus, in this study the Atrai River is chosen as the study river and the hydrodynamic parameters of this river are simulated with the HEC-RAS one-dimensional (1D) model. Moreover, in our country both long- and short-reach HEC-RAS models are being used extensively. However, there has been no study to compare the performances of these models. In this context, the main focus of this study was to setup two short- and one long-reach HEC-RAS 1D models for the Atrai River – the setup with 106 km reach of the river is regarded as the long-reach model and the two setups with 6 km reach each in the upstream and downstream of the river are regarded as the short-reach models. The main objective of this study was to compare the performance of these models with both steady and unsteady flow modules. For this study, all necessary data were collected from LGED, BWDB, USGS website and Google Earth. Hydrodynamic parameters were simulated for the flood season of the years 1998, 2006, 2014 and 2018 using different boundary conditions. The long-reach model was calibrated for 2013 and validated for 2016. The calibrated Manning’s roughness co-efficient ‘n’ was 0.025 for the main channel and 0.030 for the floodplain. The long-reach model results were evaluated based on six performance evaluation criteria (RMSE, MAE, RMAE, Relative bias, Co-efficient of determination, Efficiency index). The RMSE values for the unsteady flow module of the years 1998, 2006, 2014 and 2018, it was found that the long-reach model performed best at 2014. Finally, a comparative analysis of the model results was done to identify the suitability of the short- and long-reach models for hydrodynamic simulation in the floodplain environment. According to the results of the study for different boundary conditions, the performance of short-reach model was found better in the floodplain environment than the long-reach model. The possible reason behind it is that the fluctuation of water level due to river-floodplain interaction is more significant in the long-reach model, which cannot be captured well. Moreover, the unsteady flow module was found to produce better result than the steady flow module in both short- and long-reach models. Another finding was that the long-reach model with storage areas performed better in earlier years of 1998 and 2006, but the model without storage areas performed better in recent years of 2014 and 2018. The possible reason can be the increasing settlement or construction of embankments or roads surrounding the river. The study has broadened the understandings of hydrodynamic simulations with 1D model in floodplain environment. The results can help water development authorities, hydrodynamic modelers and water professionals in choosing model reach while quantifying flood parameters and designing hydraulic interventions at different locations of a river in floodplain environment.