Urban Flood Modelling using 1D Channel Network. Case Study of Nimes, France

Abstract

Floods are one of the leading natural hazards which cause immense damage to properties as well as to human lives. Due to increasing climate impacts, the occurrence of flood events is also increasing. Amongst the different types of floods, flash floods are of severe type due to their intensity and short warning time. In urban areas, flash floods can be even more destructive, as their economic cost is more. The measures to combat floods can be categorized into structural and non-structural. Numerical modelling figures among the non-structural measures for the flood management. A wide variety of different flood simulation models are available for analyzing the flood propagation. For engineers, the interest is normally, in knowing the spatial and temporal history of flood depths and velocities and their peaks. To that end, Saint Venant equation based models are a potent tool for the engineers. In urban areas, numerical flood modeling is rendered difficult due to heterogeneous milieu, consisting of number of large and small obstacle e.g. roads, streets, buildings, parks, parking areas. Nowadays, the bi-dimensional Saint Venant equation models are preferred due to their superior ability to predict the flow direction in channel-floodplain case i.e. strong lateral variation in topography. They, however, need large amount of data and observations to successfully run and validate the model. In addition, computational mesh building is laborious and in the absence of suitable pre-processing tools, is difficult to manage. In addition, many city centers have the configuration of a network of streets meeting at crossroads which lend itself better to a 1d flow simulation using 1d Saint Venant equation model supplemented by suitable junction models. In this study, flood modelling in the French city of Nimes, is carried out using BASEMENT model which is a freeware using interconnected 1d channel network supplemented by available junction models. The aim is to evaluate the capacity of a freeware model in predicting an urban flash flood through the 1d approach using the in-built functionalities. The model results were compared with the observed flood marks and also with the result of the 2d model. The model results were comparable to the 2d model results. The model needs to be expanded to include more junction flow models to predict the complex flow pattern at the vi crossroads as well as to efficiently incorporate the complex crossroad configurations with 4 or more streets. We also require more detailed urban topographical datasets to test and evaluate the coupled flood modeling approach whereby large open spaces can be modeled in 2d and streets in 1d.