Ambient ozone, as a secondary air pollutant in the troposphere, is a major threat to human health, plants and the environment. In order to develop effective air quality policy to minimise ozone pollution, it is important to gain a quantitative understanding of the chemical factors that drive tropospheric ozone production. There are a number of limitations and uncertainties in the current models and indirect methods used to estimate chemical ozone production rates. Here, an Ozone Production Rate (OPR) instrument is developed to fulfil the demand of accurately measuring ambient ozone production rates in the atmosphere. This prototype system aims to directly measure the in situ oxidant (Ox: O3 + NO2) production rate p(Ox) in ambient air. This thesis describes the OPR experimental methodology, instrument properties and system characteristics. Two field deployments (London and India) are comprehensively discussed, and correction approaches are implemented to improve measurement accuracy. The field measurement results indicated the measured p(Ox) levels could be used to interpret modelled Ox production rates, and changes in ambient oxidant level. The OPR system could be a useful tool to determine the balance between advection and chemical production in controlling local ozone levels, and hence support ozone control policy.

The studies on physical and chemical properties of airborne particulate matter were carried out in the UK atmosphere. Aerosol samples were collected at the Elm Road Observatory Site and Harwell representing urban background and rural area, respectively. The chemical components of both PM\(_2\)\(_.\)\(_5\) and PM\(_2\)\(_.\)\(_5\)\(_-\)\(_1\)\(_0\) were mainly analysed for carbonaceous compounds, sulphate, nitrate, chloride and oxalate. Size distributions of aerosol components were investigated in conjunction with the study of ammonia gas to stabilise the semivolatile species in atmospheric particles. Concentration composition of carbonaceous compounds (organic and elemental carbon - OC and EC) in particulate matter clearly observed higher value at urban background than those at rural area. The OC/EC minimum ratio of 0.35 was used to distinguish between primary and secondary OC as EC was a good indicator of primary sources. Sulphate and nitrate showed the good relationship, indicating that two species undergo similar formation and removal processes in atmosphere. Chloride was commonly originated from marine aerosol with the weak correlation observed with other major species. Oxalate is highly focused specie in this study as there is rarely available data in the UK. The good correlation and the similar pattern of size distribution between oxalate and sulphate suggest that their formation pathway formed from the same atmospheric processes. The excess ammonia gas supplied during the air sampling could be stablilsed and yielded the higher concentrations of nitrate and chloride which were expected in the forms of ammonium salts. In contrast, oxalate in ambient air was exhibited the discordant behaviour in this experiment. Our trajectory analysis suggests the important precursor sources of oxalate were from continental area associated with the high atmospheric pollutants.

A stochastic backscatter (SB) approach to subgrid-scale (SGS) modelling for large-eddy simulation (LES) of the neutral atmospheric boundary layer (ABL) has previously been shown to reduce excessive velocity shear, as seen with the popular Smagorinsky SGS model, in the under-resolved surface layer. However, previous SB models exhibit unwanted grid-dependency issues, and the range of atmospheric flows tested remains limited. Here, a new SB model is proposed that uses a grid-adaptive filter to control the length-scale, anisotropy and momentum flux of the backscatter fluctuations, independently of the model grid. Model performance is confirmed to be grid-independent in simulations of the neutral ABL, in which an 80% reduction in excessive near-surface velocity shear is achieved. The model is also applied to street canyon flow, where the shear layer that separates the recirculating vortex within the canyon from the external flow is again typically under-resolved in most LES set-ups. The backscatter acts to increase momentum transfer across the shear layer, bringing the simulated vortex intensity significantly closer towards wind-tunnel observations. A passive tracer is also released to model traffic emissions, and the pollutant exchange velocity between the canyon and the external flow is again found in better agreement with wind-tunnel data. This information can be used to improve operational urban dispersion models.

Copy of author's previously published article inserted. / Bibliography: p. 239-253. / xxx, 253, [42] p. : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / This thesis is concerned with the determination of gravity wave characteristics using operational radiosondes. / Thesis (Ph.D.)--University of Adelaide, Dept. of Physics and Mathematical Physics, 1996?

This study combines passive directional dust monitoring methods with ICP-MS analysis, binary mixing modelling and geostatistical modelling to generate dust dispersion maps, enabling a record of Air Pollution Control residue (APC) dust dispersion on and around a hazardous waste landfill site for a complete calendar year. This is therefore, the first study of nuisance dust dispersion on such a scale, using these methods, and the first time these methods have been used in conjunction to accurately visualise dust dispersion data over time. The proportion of APC in fugitive dust at and around the Wingmoor Farm hazardous waste landfill site near Bishop’s Cleeve, Gloucestershire, UK, was quantified and mapped using passive directional samplers, which were deployed both on the site and in farmland to the north and north-east. Samples from 19 monitors, collected fortnightly over the year, were analysed for Absolute Area Coverage (AAC) and Effective Area Coverage (EAC) prior to HF and HNO3 digestion and analysis by ICP-MS. Following geochemical characterisation of the two key “end-members”, background soils (sand and gravel, and clay) and APC, Ca/Fe and Mn/Pb mixing models were developed providing a means of determining the proportion of APC in each sample. Sample proportions derived from each mixing model were mapped using both variograms and kriging, supported through ISATISTM, to build a model of both general dust and APC dispersion over a square kilometre for the year. Meteorological conditions and movements on site were also recorded and their relationships with dust levels in the surrounding area analysed. APC was found at off-site monitoring locations throughout the year, however at far lower levels than previously assumed by prior modelling exercises. The maps presented a visualisation of dust dispersion on- and off-site, demonstrating the viability of this methodology for modelling both fugitive general dust, and, in this case, APC. In combination with the other data collected on-site, factors affecting the dispersion of both general dust and APC were identified. The development of this modelling and mapping method provides a novel, robust and cost-effective technique for both representing and understanding the dispersion of APC, delivering a useful tool for practical application by industry, local government and in the field of Environmental Forensics.

In this thesis, Rossby waves are considered within the one-layer Charney-Hasegawa-Mima (CHM) equation and two-layer quasi-geostrophic (QG) model. They are studied from a wave turbulence (WT) perspective. Since nonlinearity is quadratic, interactions take place between triplets of waves known as triads. A triad is said to be resonant if its wave vectors and frequencies satisfy k1 + k2 - k3 = 0 and w(k1)+w(k2)-w(k3) = 0 respectively. These triads can then be joined together to form resonant clusters of various sizes. The wave vectors can be continuous, in an unbounded domain, or discrete, in a bounded domain. Continuous, otherwise known as kinetic, WT has been extensively studied in the one-layer case. It is known that three quadratic invariants exist and they take part in a triple cascade in k-space. This thesis is interested in �finding quadratic invariants, of which there can be many, in the discrete regime. It begins by considering discrete clusters of resonant triads arising from a Hamiltonian three-wave equation. A cluster consists of N modes forming a total of M connected triads. It is shown that that finding quadratic invariants is equivalent to a basic linear algebra problem, consisting of fi�nding the null space of a rectangular M x N matrix A with entries 1, -1 and 0. An algorithm is then formulated for decomposing large clusters into smaller ones to show how the quadratic invariants are related to topological parts of the cluster. Speci�c examples of clusters arising in the CHM wave model are considered. The second part of this thesis focusses on the large-scale limit of the CHM equation. This limit has been studied the least; however, it would appear to be more relevant since Rossby waves in the ocean are large-scale. Recently a new quadratic invariant, known as semi-action, has been discovered in this limit. Its density is one in the meridional region |ky|< /3kx and zero in the zonal region |ky|> /3kx: As a consequence of the conservation of semi-action, conditions are placed on the triad interactions involving zonal (Z) and meridional (M) modes. In this thesis it is proved directly, without appealing to conservation, that the following triad interactions are prohibited: M -> M +M,M -> Z + Z,Z -> M + Z and Z -> M +M: The cascade directions are studied of the three invariants, the energy, enstrophy and, depending whether the initial spectrum is in the meridional or zonal sector, the semi-action or zonsotrophy respectively. The results are interpreted to explain the formation of unisotropic turbulence with dominating zonal scales. In the �nal part of this thesis, a symmetric form of the two-layer kinetic equation for Rossby waves is derived using canonical variables, allowing the turbulent cascade of energy between the barotropic and baroclinic modes to be studied. It turns out that energy is transferred via local triad interactions from large-scale baroclinic modes to the baroclinic and barotropic modes at the Rossby deformation scale. From there it is transferred into large-scale barotropic modes via a non-local inverse transfer.

The oxidation of primary emitted species such as volatile organic compounds (VOCs) acts as a source of ozone and therefore has detrimental effects on air quality and climate. In order to understand at what rate oxidation is occurring in the troposphere, i.e. the oxidation capacity, an understanding of the contributors to oxidation and possible markers of oxidation are imperative. Formic acid, a ubiquitous trace gas, which contributes significantly to the acidity of precipitation, could, because of its dominant source being suggested to be the ozonolysis of alkenes, be used as a marker for oxidation. A chemical ionisation mass spectrometer (CIMS) with a quadruple mass analyser utilising the I ionisation scheme was therefore deployed in London during both the summer and winter Clean Air for London (ClearfLO) Campaign to measure formic acid in order to asses this possibility. Simultaneous formic and nitric acid measurements with the CIMS, as well as supplementary gas phase data, in the winter ClearfLO campaign indicated the dominance of direct anthropogenic emissions of formic acid at this time and location. Global modelling simulations including a direct anthropogenic emission of formic acid suggested that this source is responsible for up to 30% of total formic acid in certain regions of the northern hemisphere, an area where this acid is most significantly underestimated. This ruled out the possibility of using formic acid as a marker for oxidation as a result of its vast range of sources. The mirrored summer campaign, also simultaneously detecting formic acid and nitric acid, showed that secondary photochemical production was, however, a dominant source in this time and location. The first UK measurements of ClNO2 are reported here – which is a region where modelling studies predict significant ClNO2 production all year round. Concentrations of ClNO2 up to 724 ppt and a mean of 207 ppt were measured in London during the summer ClearfLO campaign. Cl atoms from ClNO2 photolysis were shown to contribute up to 15%, 3% and 26% of the oxidation of alkanes, alkenes and alkynes, respectively over the full day. This will increase the tropospheric ozone forming potential at this time, particularly in the early morning. As well as very limited ClNO2 measurements in Europe, measurements from low NOx sites and airborne measurements are also highly understudied. Measurements of ClNO2 in the rural coastal site of Weybourne and an altitude profile from the FAAM Bae-146 Research aircraft were therefore undertaken. Concentrations of ClNO2 were significantly lower in Weybourne in comparison to the London study with peaks of 80 ppt with night-time average of 15 ppt being observed. The importance of Cl atoms to VOC oxidation at this site was much reduced in comparison to London owing to the higher OH concentrations and lower production of ClNO2 at this site.

L’ozone (O3) est un gaz reconnu comme un des polluants atmosphériques majeurs du XXIème siècle, de par ses propriétés très oxydantes. Dans les plus basses couches de notre atmosphère, ce polluant secondaire est produit par des réactions photochimiques impliquant des précurseurs émis par les activités humaines et la végétation. De nombreuses études expérimentales ont révélées le caractère phytotoxique de l’O3, qui dégrade sévèrement la photosynthèse, réduit le rendement des récoltes et inhibe la croissance des arbres. Afin de quantifier les effets de l’O3 sur les écosystèmes à l’échelle globale, des modèles numériques ont été développés. Dans cette thèse, nous avons principalement développé un nouveau modèle d’impact basé sur le concept dose/réponse au sein du modèle global de végétation ORCHIDEE. L’approche semi-mécaniste utilisée vise à représenter de manière simplifiée les effets du stress oxydatif causés par l’O3 sur la photosynthèse. L’étalonnage des paramètres a été réalisé en utilisant des données physiologiques issues majoritairement d’expériences de fumigation à l’air libre et en chambres à ciel ouvert menées pendant la saison de croissance de plusieurs espèces végétales. Cette étape de calibration a permis de déterminer les paramètres spécifiques à plusieurs types fonctionnels de plantes. La sensibilité du modèle à l’O3 et à son interaction avec le CO2 et le stress hydrique a ensuite été testée. A l’échelle locale, notre modèle d’impact reproduit fidèlement l’allure de la réponse de la photosynthèse observée in situ à l’échelle saisonnière, et l’élévation du CO2 compense partiellement l’impact de l’O3. En outre, le couplage entre la photosynthèse et la conductance stomatique dans ORCHIDEE permet de simuler la limitation des effets de l’O3 par les flux entrants. La fermeture stomatique réduit également la transpiration, ce qui augmente l’humidité du sol et préserve ainsi la végétation d’un stress hydrique accentué. Cependant, les effets observés sur la respiration autotrophe ne sont pas représentés et l’impact sur la surface foliaire est sous-estimé. Afin d’évaluer la performance du modèle, nous avons comparé l’impact simulé sur la productivité nette (NPP) annuelle avec l’effet estimé par les relations empiriques linéaires standards, recommandées par l’UNECE pour évaluer les risques phytotoxiques chez plusieurs types d’écosystèmes. En outre, à l’échelle globale, nous obtenons des résultats similaires aux estimations réalisées avec un autre modèle de végétation munie d’une paramétrisation d’impact différente. Enfin, nous estimons que les concentrations actuelles causent localement une chute de productivité nette totale annuelle allant jusqu’à 11.7% en moyenne dans le nord-est des Etats-Unis, et réduirait le contenu en carbone du sol de 10.9% dans cette région et de 42.5% en Indonésie si la pollution actuelle perdure pendant 50 ans. Ce travail indique que l’impact de l’O3 sur la végétation est non-négligeable dans le modèle ORCHIDEE, et doit être pris en compte dans les études globales du bilan de carbone terrestre. / Ozone (O3) is a gas recognized as a major atmospheric pollutant of the 21st century, due to its strong oxidant properties. As a secondary pollutant, O3 is produced by photochemical reactions between both anthropogenic and biogenic precursors. Numerous experimental studies have highlighted the phytotoxic effects of O3, which severely impairs photosynthesis, reduces crop yields and forest growth. Numerical models are used in order to quantify the impact of O3 at global scale. In this thesis, the major work was to develop a new parameterization based on the dose/response concept and to integrate it in the global vegetation model ORCHIDEE. We used a semi-mechanist approach to represent the effects of oxidative stress induced by O3 on photosynthesis. Parameters were obtained by comparing modelled and observed physiological variables related to free-air and open-top chamber fumigation experiments, carried out during the growing period for different plant species corresponding to several plant functional types in the model. Sensitivity tests were conducted to evaluate the influence of O3 and its interactive effects with CO2 and water-stress on the ecophysiological processes in ORCHIDEE. We reproduced the curvilinear response observed on photosynthesis at the leaf level during a growing season, and the elevation of CO2 concentration partially mitigates the O3 effect. The induced simulated stomatal closure slows down the impact on photosynthesis coupled to conductance by reducing the entering O3 fluxes. It also decreases transpiration and increases soil water content, which protects vegetation from higher water-stress. However, observed changes in autotrophic respiration are not simulated and the impact on foliar surface is underestimated. In order to evaluate the performance of our model, we compared the simulated impact on annual net primary productivity (NPP) with the empirical linear dose/response relationships recommended by UNECE to assess the risk for different types of ecosystems. Moreover, at global scale, we found results similar to those from another land surface model using a different impact relationship. Finally, we estimated that current O3 concentrations cause locally a decrease in total annual net productivity up to 11.7% on average in the north-west of USA, and could decrease the soil carbon content by 10.9% in this region and by 42.5% in Indonesia, if current O3 pollution remains the same for 50 years. This work confirms that impact of O3 on vegetation is non-negligible in the model ORCHIDEE, and must be considered in global carbon budget modelling.

This thesis presents the results from investigations into various features of the dynamics of the middle atmosphere. Wind measurements from the meteor radar on Ascension Island and temperatures measured by the Aura MLS instrument are used to characterise Ultra-Fast Kelvin Waves (UFKW) in the MLT-region. Rainfall rates from TRMM are used as a proxy for latent heat release in an investigation of the excitation of UFKW, and results are compared with predictions of the Kyushu-GCM. Amplitudes and vertical wavelengths are determined, as are the accelerations resulting from wave dissipation. Intra-seasonal oscillations (ISOs) are observed in the wave amplitudes and accelerations, and oscillations with the same periods are observed in the rainfall and MLT-region winds, suggesting that UFKW play a role in carrying the tropospheric ISOs to higher regions. The seasonal and interannual variability of the tidal field over Ascension Island is investigated. Amplitudes, phases and vertical wavelengths of the diurnal and semidiurnal tides are reported on. Our observations of tidal parameters are compared with the predictions of the extended Canadian Middle Atmosphere Model (eCMAM) and the Whole Atmosphere Community Climate Model (WACCM). Correlations between tidal amplitudes, the stratospheric Quasi-Biennial Oscillation (QBO) and the El Nin ̃o Southern Oscillation (ENSO) are discussed. The Hocking (2005) method is adapted to recover monthly-mean gravity-wave momentum fluxes and wind variances either side of the Drake passage gravity-wave hotspot. The ability of the method to recover momentum fluxes over each radar are tested by sampling a series of specified (known) wave fields of increasing complexity with the actual meteor distributions, and comparing the recovered momentum fluxes with the specified values. The analysis is then applied to the real data to obtain climatologies of the MLT-region variances and momentum fluxes for a composite year of the 2008 to 2012 data.

Lightning represents a key interaction with climate through its production of nitrogen oxides (NOx) which lead to ozone production. These NOx emissions are generally calculated interactively in chemistry-climate models but there has been little development of the representation of the lightning processes since the 1990s. In most models the parametrisation of lightning is based upon simulated cloud-top height. The aims of the thesis are: to explore existing schemes, and develop a new process-based scheme, to parametrise lightning; to use a new process-based lightning scheme to give insights regarding the role of lightning NOx in tropospheric chemistry; and to use alternative lightning schemes to improve the understanding of the response of lightning to climate change, and the consequent impacts on tropospheric chemistry. First, a new lightning parametrisation is developed using reanalysis data and satellite lightning observations which is based on upward cloud ice flux. This parametrisation is more closely linked to thunderstorm charging theory. It greatly improves the simulated zonal distribution of lightning compared to the cloud-top height approach, which overestimates lightning in the tropics. The new lightning scheme is then implemented in a chemistry-climate model, the UK Chemistry and Aerosol model (UKCA). It is evaluated against ozone sonde measurements with broad global coverage and improves the simulation of the annual cycle of upper tropospheric ozone concentration, compared to ozone simulated with the cloud-top height approach. This improvement in simulated ozone is attributed to the change in ozone production associated with the improved zonal distribution of simulated lightning. Subsequently, data from a chemistry-climate model intercomparison project (ACCMIP) are used to study the state-of-the-art in lightning NOx parametrisation along with its response to climate change. It is found that the models using the cloud-top height approach produce a very similar response of lightning NOx to changes in global mean surface temperature of +0.44± 0.05 TgNK-1, for a baseline emission of 5 TgN yr-1. However, two models using two alternative lightning schemes produce a weaker and a negative response of lightning to climate change. Finally, simulations in a future climate scenario for year 2100 in the UKCA model were performed with the cloud-top height and the ice flux parametrisations. The lightning response to climate change when using the cloud-top height scheme is in good agreement with the positive response found in the multi-model results of the cloud-top height approach. However, the new ice flux approach suggests that lightning will decrease in future. These opposing responses introduce large uncertainty into the projections of tropospheric ozone and methane lifetime in the future scenario. An analysis of the radiative forcing from these two species also shows the large uncertainty in the individual methane and ozone radiative forcings in the future. Due to the opposite effect that lightning NOx has on methane (loss) and ozone (production) the net radiative forcing effect of lightning in present-day and future is found to be close to zero. However, there is a small positive feedback suggested by the results of the cloud-top height approach, whereas no feedback is evident with the ice flux approach. These results show there are large and crucial uncertainties introduced by lightning parametrisation choice, not only in terms of the actual lightning distribution but also atmospheric composition and radiative forcing. The new ice-based parametrisation developed here offers a good alternative to the widely-used approach and can be used in future to model lightning and develop the understanding of associated uncertainties.