Global sensitivity analysis in flood mapping using HEC-RAS 2D: effects of the floodplain manning coefficient for a dam-break case

Abstract

ABSTRACT The flood map outputs from hydrodynamic models may contain uncertainties due to inaccuracies in the input parameters. On the other hand, the results of dam break models are crucial for performing risk analyses of potential material and social damage. Thus, this study applied a global sensitivity analysis (GSA) to evaluate the effects of the floodplain Manning coefficient on the inundation generated by the hypothetical dam break of a real water dam in HEC-RAS 2D. Three aspects of flood output, the peak discharge, the arrival time and the time of flood damping, were analyzed via variance tests. The Manning coefficient varied from 0.026 to 0.150, resulting in a change of up to 54% in the peak discharge at a distance of 25 km and a delayed arrival time corresponding to 16 minutes in the cross section located 1 km downstream of the dam. However, Manning coefficients between 0.026 and 0.050 show statistically similar results in terms of two aspects (the peak discharge and the arrival time). This study indicated that the GSA and variance tests may contribute to critical analyses of Manning coefficient adoption and help decision-makers develop criteria for planning emergency actions.

Keywords

• Variance tests
• Flood maps
• Dam failure
• 2D model