Advanced flood inundation modeling skills: Improved boundary conditioning and its interconnection with model parameters
Abstract
Two--dimensional hydrodynamic models are commonly used to simulate the Spatio-temporal distribution of water depth and flood inundation areas. In most studies, roughness coefficients and channel bathymetry information are tuned to improve the performance of model simulations compared to reference data. To better characterize the physics and to enhance the accuracy of these models, we recommend including vertical fluxes and lateral flows as additional boundary conditions in the model setup. We demonstrate that calibration of hydrodynamic models without applying sufficient boundary conditions can result in generating unrealistic parameters which deteriorate the performance of these models. We utilize two well-known 2D hydrodynamic models, LISFLOOD-FP and HEC-RAS 2D for different study areas and flood events and investigate the impacts of this new boundary conditioning approach on the overall performance and the calibrated parameters. Our results show that including lateral flows and vertical fluxes can significantly improve the accuracy of flood extent mapping. The maximum flood extents of Hurricane Harvey simulated by the LISFLOOD-FP and HEC-RAS2D models are improved by 10% and 20%, respectively. We found that including these additional boundary conditions does not necessarily change the calibrated roughness parameters. Including lateral flows and vertical fluxes, however, significantly affect the calibrated offset parameter used to characterize the channel bathymetry. Our results also demonstrate that changing the offset parameter of river bathymetry mostly affects the flood depth distribution and its dynamics while it rarely changes the flood extent maps provided by hydrodynamic models.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFMNH12B..03J