Abstract
Cities intimately intermingle people and air pollution. It is very difficult to monitor or model neighbourhood-scale pollutant transport explicitly. One computationally efficient way is to treat neighbourhoods as patches of porous media to which the flow adjusts. Here we use conceptual arguments and large-eddy simulation to formulate two flow regimes based on the size of patches of different frontal-area density within neighbourhoods. One of these flow regimes distributes pollutants in counter-intuitive ways, such as producing pollution ‘hot spots’ in patches of lower frontal-area density. The regimes provide the first quantitative definition of the ‘urban background’, which can be used for more precisely targeted pollution monitoring. They also provide a conceptual basis for further research into neighbourhood-scale air-pollution problems, such as parametrizations in mesoscale models, and the transport of fluid constituents in other porous media.