The emergence and expansion of urban areas would change local ecosystem considerably, especially the water ecosystem. Urban water ecosystem has always remained important players in social development, with its performance in transport, water supply, environmental regulation and purification function. The deterioration of water ecosystem compelled government managers to take measures to settle the subsequent environmental problems, based on insights into hydraulic characteristic in urban area.

Urban flood should be devoted more attention on account of its effect on social property and ecological issue. Coupling hydrodynamic equation and convective-diffusion equation could describe the pollutants evolvement process. Despite advance in computing resources, large scale flood simulation with two dimensional shallow water equations in urban areas remains time consuming when applying finer grids that describe the feature of buildings and other structures. Grid coarsening that resamples from or takes the average of the fine grid data could be the simplest way to improve the modeling performance since the computing load has been significantly reduced. However, the detailed information of fine grid would be lost during the aggregation procedure such that the modeling results may not reflect the reality of flow propagation properly. 

This paper intends to solve the dilemma that how to capture the building affection to flood while enhance computation efficiency of the model. The shallow water equations was modified by introducing extra parameters to account for the reduction in storage and conveyance due to the existence of buildings. Considering the anisotropy in study area, a volumetric porosity and an interface porosity are introduced to represent building coverage and space ratios within each grid cell and on the cell boundary. To reflect the key feature of grids separated by a building inside, one single coarse grid cell may have two or more porosity parameters applied to calculate the flow path. The modified shallow water equations would be coupled with convective-diffusion equation to simulate the transport and distribution of pollutants in flood events. We applied the new model to two case studies and the results show the proposed model has significant improvement in both the computing cost and the accuracy.