Integrating a finite element model with a geographic information system to model urban stormwater flow

For building site development, it is desirable to be able to predict the depth, direction and magnitude of stormwater flow in order to prevent flood damage. A two dimensional implicit finite element model was developed for stormwater flow prediction by using the depth-averaged form of the continuity and momentum equations for flow on a two dimensional surface. The finite element model computes with reasonable accuracy the depth of flow, and the magnitude and direction of the velocity field. There are no stability constraints on the time step length, and results are computed to first order accuracy in time. The numerical solution has been validated by adjusting the eddy and numerical viscosities in the momentum and continuity equations to match the experimental data from a 3/4 scale physical model of flow into a curb inlet on a street. An eddy viscosity coefficient of about 3 and numerical coefficient of 0.3 for the supercritical flow were the minimum necessary values for numerical stability. The finite element model was also integrated with a geographic information system (GIS), Arc/Info, for storm water prediction over urban terrain. An inventory of the urban terrain and storm drainage system for part of the City of Asheville, North Carolina was used as case study of the integrated model. GIS is used to select a drainage area and access all of the information associated with that area such as contours, street side lines, buildings and terrain specifications. Results from the finite element flow analysis are stored in the GIS database as new items and they are then available for GIS graphical operations, such as mapping the flowfield and depth contours. Although finite element modeling has been applied to two-dimensional shallow water flow computation for more than a decade, this is the first integration of finite element modeling with GIS to compute the flow field and water depth of stormwater flow over in an urban land surface.