Roughness Parameter Estimation for flood numerical simulation using Differential Evolution

Abstract

A methodology to estimate parameters necessary to carry out numerical simulations of flood phenomena is presented, that may be useful for detecting flood-prone areas. Geospatial information contained in different databases is used as inputs, numerical simulation tools of hydrodynamic flooding phenomenon by solving the shallow water equations, and stochastic optimization algorithms. The objective is to find certain parameters of the simulation model that have a high uncertainty degree through evolutionary algorithms, comparing the simulations carried out with satellite images that monitor the behavior of rivers and streams in areas that may be susceptible to flooding. In this work, the Manning roughness coefficients were determined according to the soil usage identified in a synthetic example, in order to evaluate the usefulness and viability of the methodology proposed.