Application of cellular automata approach for fast flood simulation

Abstract

The increasing pluvial flooding in many urban areas of the world has caused tremendous damage to societies and has drawn the attention of researchers to the development of a fast flood inundation model. Most available models are based on solving a set of partial differential equations that require a huge computational effort. Researchers are increasingly interested in an alternative grid-based approach called Cellular Automata (CA), due to its computational efficiency (both with respect to time and computational cost) and inherent parallel nature. This paper deals with the computational experiment with a new CA method for modelling 2D pluvial flood propagation. A Digital Elevation Model (DEM) comprising square grids forms the discrete space for the CA setup. Local rules are applied in the von Neumann Neighbourhood for the spatio-temporal evolution of the flow field. The proposed model is applied to a hypothetical terrain to assess its performance. The results from the CA model are compared with those of a physically based 2D urban inundation model (UIM). The CA model results are comparable with the results from UIM model. The advantages of low computational cost of CA and its ability to mimic realistic fluid movement are combined in a novel and fast flood simulation model.