Model building and analysis tools and recommended practices

A range of tools and recommendations to aid the analysis of urban catchments and the implementation of urban storm water drainage models were developed, documented and tested within the project.

Click on the links below to access additional information and/or documentation of each of the tools.

  • Fractal tools for analysis of urban catchments:these tools can be used to analyse geometric features of different aspects of urban catchments, including sewer layout and land use distribution. They were developed by partners from ENPC during the RainGain project. Additional information (will be available soon): Tutorial kit

  • Automatic Overland Flow Delineation (AOFD) tool:this is a GIS (Geographic Information Systems) tool which, based upon a digital elevation model (DEM) of the study area, generates a 1-dimensional (1D) flow model of the urban surface. In this 1D model the urban surface is discretised as a set of storage nodes (representing ponds) and conduits (representing streets and other pathways through which surface runoff flows. As compared to 2D models of the urban surface, 1D models are significantly faster, thus being suitable of real-time applications. However, 1D models have some limitations, including poor visualisation of flood simulation results and inability to handle flows in flat areas. This tool was initially developed by Imperial College London partners as part of a previous research project (Maksimovic et al., 2009). During the RainGain project this tool was further improved and a detailed tutorial was developed. Additional information: Overview of the AOFD tool; User manual.  To obtain a copy of this tool (i.e. an executable file) and test data, please contact Prof. Cedo Maksimovic ( or Dr Ana Mijic (

  • General recommendations for dealing with open channels and other small surface features in urban pluvial flood simulations:  Recent developments in urban drainage simulation tools include the seamless integration of classical 1 dimensional (1D) models of the sewer network with 2-dimensional (2D) models of the urban surface which enable a more realistic and accurate representation of the urban drainage system, both during and after the occurrence of intense rainfall. Given the growing risk of urban pluvial flooding as well as increasing drives to make urban drainage systems ever more sustainable, it is important that small surface elements are modelled sufficiently accurately, as these play a critical role in the management of overland flows, which in turn affects the overall performance of the drainage system.

Different tools will use different approaches to represent such surface elements, but all are based on the availability of digital elevation models (DEMs). Whereas ten years ago the typical average grid size of DEMs was anything between 2-5 m, recent LIDAR technology has made it possible to create models with a density of around 15 – 20 points per m² (or a corresponding grid size of about 0.25 m).

High resolution DEMs create new opportunities for fully integrating small surface elements, such as road ditches and small trenches, into the 2D model of the rest of the surface, which includes roads, fields, parks, etc. This was not possible using traditional (coarser) DEMs, as their resolution was often coarser than the size of the small surface features. When this was the case (and it may still be in given areas), a better approach may be to model the small surface features in 1D and embed them into the 2D model of the surface. The approach that is adopted to model such elements needs to be carefully selected by the modeller, based upon the resolution and quality of the available data, the simulation tools available to him/her, and the purpose of the modelling exercise.

  • General recommendations for dealing with buildings in 2-dimensional (2D) urban flood simulations: Similar as for the widespread availability of high resolution digital elevation models (DEMs), it is now common to have very detailed GIS background layers of buildings in urban environments. Although this information is very important for 2D flood simulations (given that buildings somehow are "obstacles" for the flood propagation), its high level of detail can turn out to be quite challenging for numerical calculations.

Vector layers of buildings often have a very high point (vertex) density to make sure that even the most complex shapes and corners are well represented.  However, in creating a 2D calculation mesh this can become problematic because the alignment of the mesh with the buildings can cause the mesh to become incredibly complex both in terms of geometry and of number of mesh elements required to cover the buildings’ contours. This can result in exponentially increased simulation times.

It is therefore important that building layers are as much as possible cleaned and simplified (however without losing the essential details) before using them in a 2D model. A number of GIS routines exist which can be used for this purpose (e.g. SimplyPy QGIS plugin), although they may be refined and tailored for specific purposes. It is important to conduct sensitivity tests when using such tools, so as to ensure that a desired level of simplification is achieved, while preserving essential topological and geometrical features.