How trees affect urban air quality: It depends on the source

Abstract

Large-eddy simulation (LES) is used to systematically analyse the impacts of trees on air quality in idealised street canyons. The LES tree model includes radiation, transpiration, drag and deposition effects. The superposition of background concentrations and local emissions is used to construct realistic urban scenarios for fine particulate matter (PM2.5) and nitrogen oxides (NOx). Both neutral and convective atmospheric conditions are considered to assess the importance of buoyancy effects and the role of tree shading and transpiration. Tree impact on local air quality is shown to be driven by the balance between the rate at which they actively remove pollutants from the air (deposition) and the way in which they alter the transport of pollutants within and out of the street canyon (dispersion). For pollutant species or street types where the concentration field is dominated by background levels (such as PM2.5), deposition will generally dominate and thus local air quality will improve. For pollutants and street types where local emission sources dominate (e.g. NOx on a busy road), the dispersion effects of trees become more prominent and can lead to elevated concentrations where mixing or exchange is significantly inhibited. Mixing in the convective simulation is more vigorous than in the neutral simulation which results in substantial differences in in-canyon flow fields and exchange velocities, highlighting the importance of incorporating thermal effects when studying urban trees. Increased residency times, and thus deposition, under neutral conditions suggest that trees can have amplified effects under conditions conducive of poor air quality. For the cases considered, trees largely act to improve air quality with the exception of localised hotspots. The competing effects of trees — specifically deposition versus altered exchange with the atmosphere — are also incorporated in a simple integral model that predicts whether or not the air quality will improve. The model matches well with LES predictions for both PM2.5 and NOx and can serve as a simple tool for urban design purposes.