A GRASS GIS module for 2D superficial flow simulations


Urban flooding is a growing public concern in megacities such as Mexico City, due to the combined effect of an intense rising in both precipitation and levels of exposure. In this sense, the accurate assessment of urban flood impacts requires the development of fast modelling tools that increase our knowledge on the main processes governing the generation and propagation of floods. This paper presents a new hydrodynamic model named Itzï, designed to simulate superficial flows in two dimensions, by means of a pseudo-2D and partial acceleration numerical scheme on a raster grid. The model has been implemented within a GRASS GIS environment, through a module written in Python that manages inputs (e.g. rainfall) in a raster format with dynamic variables in time and space. The model is validated using two different analytic test cases for the full shallow water equations, for which results from the numerical tool compare very well, with RMSE of 0.03 and 0.003 meters, respectively. Secondly, a test related to an urban setting is also employed for its verification; in this case, we employ a hypothetic flood event in the region of Greenfield, in the city of Glasgow in the UK; for which hydraulic data is available, along with a high resolution Digital Elevation Model (2m). Moreover, in order to compare the performance of our model against other well-established models, we also implement the “acceleration” model of LISFLOOD-FP. Encouraging results are found in this comparison, as nearly identical results with both models are produced. Therefore, the numerical tool here presented represents a solid step forward, towards the development of smart urban flood management strategies, which rely in open-source modern technologies enabling better decision-making within the context of climate change and urbanization.

In Proceedings of the 12th International Conference on Hydroinformatics