Identifying Urban Design Details Affecting the Spatial Variability of Micro-Scale Ambient Nitrogen Dioxide for the Auckland CBD Area

Abstract

The physical properties of urban areas have the potential to play a significant role in determining local air quality. The complex interplay of the urban design properties (intersections, bus stops, height-to-width ratios) with meteorology and emissions affect the temporal and spatial patterning of pollutants. A land-use regression model traditionally examines the relationship between ambient concentrations and urban design properties typically have meso- scales, where only general identification of properties possible. Few studies have created a micro-scale model in order to resolve this issue, with Auckland, NZ, providing an excellent test-bed due to the high variety of urban design features present. The aim of this research is to examine the association between urban design and near-surface concentrations of ambient nitrogen dioxide (NO2) within the Auckland central business district (CBD). A dense network of passive diffusion tubes was used to examine the spatial trends in concentrations over a six-week period. This data, along with examination of previous data sets and a multivariate land-use regression (LUR) model, was used to characterise the NO2 concentration as a result of the surrounding micro-scale urban environment. Urban design variables were collected via observations/counts, satellite images, and GIS datasets. Focus was on identifying site-specific urban design elements and locating sites with consistently elevated concentrations (‘hotspots’). The results demonstrate extensive spatial variability in NO2, with large differences observed over small distances, however trends in consistent site ranking were observed. The developed LUR model revealed seven urban design variables (‘number of lanes’, ‘sum of bus stops within 100m’, ‘sum of bus lanes within 100m’, awning present’, ‘vegetation present within 25m’, ‘building footprint within 100m’, ‘distance toward the next set of traffic lights’) that influenced NO2. Of those the more consistent determinants of air quality were the variables number of bus stops and proximity to traffic lights. A generalised version of the LUR model, constructed using historic and new data sets at both meso- and micro-scales, showed the greatest potential for model generalisation and transferability. The identification of micro-scale design elements can offer assistance in mitigating urban air pollution, with options such as judicious use of vegetation in ‘hotspot’ sites or by encouraging more environmentally-friendly transport alternatives within CBD areas.