Maximizing the spatial representativeness of NO2 monitoring data using a combination of local wind-based sectoral division and seasonal and diurnal correction factors

摘要

This article describes a new methodology for increasing the spatial representativeness of individual monitoring sites. Air pollution levels at a given point are influenced by emission sources in the immediate vicinity. Since emission sources are rarely uniformly distributed around a site, concentration levels will inevitably be most affected by the sources in the prevailing upwind direction. The methodology provides a means of capturing this effect and providing additional information regarding source/pollution relationships. The methodology allows for the division of the air quality data from a given monitoring site into a number of sectors or wedges based on wind direction and estimation of annual mean values for each sector, thus optimising the information that can be obtained from a single monitoring station. The method corrects for short-term data, diurnal and seasonal variations in concentrations (which can produce uneven weighting of data within each sector) and uneven frequency of wind directions. Significant improvements in correlations between the air quality data and the spatial air quality indicators were obtained after application of the correction factors. This suggests the application of these techniques would be of significant benefit in land-use regression modelling studies. Furthermore, the method was found to be very useful for estimating long-term mean values and wind direction sector values using only short-term monitoring data. The methods presented in this article can result in cost savings through minimising the number of monitoring sites required for air quality studies while also capturing a greater degree of variability in spatial characteristics. In this way, more reliable, but also more expensive monitoring techniques can be used in preference to a higher number of low-cost but less reliable techniques. The methods described in this article have applications in local air quality management, source receptor analysis, land-use regression mapping and modelling and population exposure studies.