Representativeness and application of long-term trace gas and photolysis measurements for evaluating local air quality

Abstract

Networks of long-term measurements of trace gases are critical for understanding spatio-temporal trends in air pollutants. This data is used to assess long-range and trans-boundary transport of emissions, quantify effects on public health, develop mitigation strategies and examine the impact of implemented policy changes. As part of the European Monitoring and Evaluation Programme (EMEP), the UK operates two “super sites” which have provided a suite of co-located measurements for this purpose. These supersites have been running for decades, and are located in rural background conditions, with the intention of being representative of the north and south of the country. A Monitor for AeRosols and Gases in ambient Air (MARGA; Metrohm Applikon, NL) has been included in these sites’ measurements for over a decade. However its gaseous measurements of nitric acid (HNO3) have been demonstrated to include potential artefacts from other oxidised reactive nitrogen species (NOy), such as dinitrogen pentoxide (N2O5). This interference has not yet been formally quantified. Other NOy measurements at either site are infrequent. Nitryl chloride (ClNO2) in particular was first measured in the UK in 2012, and has been measured only sporadically since. Meteorological variables are similarly measured in networks to provide locally representative data, which are utilised in atmospheric chemistry and chemical transport models. Photolysis reactions are key drivers of atmospheric chemistry, initiating many reaction routes via the production of reactive radical species. As such, accurate estimation of photolysis rate constants (or photolysis frequencies; j-values) are imperative for understanding subsequent reactions and predicting accurate pollutant concentrations. Photolysis rate constants are highly influenced by local meteorology (e.g. clouds, aerosols), but capturing the spatio-temporal variability of these changing conditions is challenging, and often computationally costly. Consequently, modelled j-values are often parameterised or determined for unrepresentative local conditions, and results are not validated beyond model conception. Some studies apply adjustment factors to these model results to account for local conditions, but these have not yet been standardised nor explored. Part of this PhD research presents a systematic analysis of a measurement-driven adjustment factor (MDAF) to adjust clear-sky or cloud-free modelled j-values to capture changes in the local meteorology. MDAFs were derived from the ratios of j-values from both filter- and spectral radiometer measurements and clear-sky estimates from the Tropospheric Ultraviolet and Visible radiative transfer model (TUV). MDAFs were examined in terms of space (3 UK sites), time resolution (hourly to annual averages), photolysis reactions (12 studied), optical inlet used (4-π sr and 2-π sr) and qualitative impact on model chemical schemes. MDAFs derived from j(NO2) were found to be seasonally similar around the UK, but specific to local environments at higher time resolutions, demonstrating the importance of local j-value measurements. Downwelling (2-π) MDAFs demonstrated a slight increase with solar zenith angle (SZA), which was amplified when measurements of upwelling j(NO2) were considered (4-π). Increased surface albedo (snow cover) resulted in approximately 36% lower downwelling compared with 4-π MDAF, but the difference was negligible at other times. Derivations of MDAF for the 12 different atmospheric photolysis reactions were grouped using hierarchical cluster analysis (HCA). The groupings of the photolysis reactions were found to be driven by the extent to which a species photodissociates at longer (UVA) wave-lengths. MDAFs derived from j(NO2) measurements were deemed an applicable reference for local adjustment of the j-values for other photodissociations at wavelengths >350 nm. For j-values of photodissociations at shorter wavelengths, adjustment using MDAFs based on a reference of j(O1D) resulted in lower total error. The presence of clouds had a greater influence on reducing cloud-free model results of j(NO2) (approx. 45%). Shorter wavelengths, such as those required for the photolysis rate constant j(O1D), are scattered more readily in clear skies, and thus resulted in a lower magnitude difference (20%). The other part of this PhD investigated atmospheric composition at the two UK supersites, by assessing the impact of the relocation of the southern EMEP supersite from Harwell to Chilbolton Observatory, and deploying an iodide chemical ionisation mass spectrometer (I – CIMS) to measure NOy species at the northern supersite (Auchencorth Moss). Meteorological normalisation was used on a concatenated time series of pollutant concentrations pre- and post-relocation from Harwell to Chilbolton Observatory, to identify any resulting effects of the move on these time series. Of all the species considered, only nitrogen oxides (NOx) and ammonia (NH3) had a step change in concentration, both increasing. The additional contributing sources at Chilbolton Observatory were identified. As a consequence, the long-term time series of NOx and NH3 should be considered to be restarted following the relocation, and the new site not strictly representative of the wider area it is intended to be. The aim of the CIMS study at Auchencorth Moss was to measure HNO3 and N2O5 to quantify the interference in co-located MARGA measurements, as well as to contribute the first Scottish ClNO2 measurements. The challenges of this study, and future work required is discussed. This PhD research has demonstrated a new potential application of meteorological normalisation for air quality site relocations, which will become more pertinent in future years where background sites will on occasion need to be relocated due to local development. Furthermore, this study has emphasised the importance of measuring local photolysis rate constants to account for highly variable local conditions. It provides discussion around making existing measurements standardised and accessible, so as to make more frequent model validation or implementation of MDAF-like metrics easier, and to improve modelled estimations of local photolysis rate constants without significantly increasing computational cost. This PhD research explores the ongoing need to measure both atmospheric chemical components and photolysis rate constants to understand changes in the atmosphere as pollutant emission abatement policies are implemented under real local conditions.