Propulsé sur une orbite polaire en avril 1995 par l’Agence Spatiale Européenne, l’instrument Global Ozone Monitoring Experiment (GOME) est le précurseur d’une nouvelle génération de satellites dédiés à la mesure globale de la composition atmosphérique. Ce spectromètre hyperspectral mesure, entre 240 et 790 nm, à la résolution de 0,2-0,4 nm, la radiance diffusée par l’atmosphère et réfléchie par la surface terrestre et les nuages au nadir du satellite. La technique de spectroscopie d’absorption optique différentielle (DOAS) permet d’en inverser la concentration columnaire de l’ozone et du dioxyde d’azote atmosphériques.<p><p>Les travaux décrits dans cette thèse portent d’une part sur la caractérisation du contenu en information géophysique accessible par ce type de sondage atmosphérique, et d’autre part sur la mise au point des méthodes et algorithmes d’inversion propres à la mission GOME. <p><p>Au cours des premiers chapitres, nous établissons les propriétés pluridimensionnelles de lissage et d’échantillonnage du champ atmosphérique associées à l’observation du rayonnement diffusé. Nous explorons ensuite les problèmes posés par le cycle diurne des oxydes d’azote, ainsi que l’effet des gradients atmosphériques interférant avec le chemin optique. Nous analysons enfin les capacités des réseaux de télédétection de l’Organisation Mondiale Météorologique (OMM) pour le diagnostic des algorithmes et données des systèmes satellitaires.<p><p>Deux chapitres sont ensuite consacrés à la mise au point des méthodes et algorithmes d’inversion DOAS pour le processeur GDP, qui traite de manière opérationnelle les données radiométriques acquises par GOME. Nous abordons successivement le problème de la dépendance en température des sections efficaces d’absorption, l’évaluation du facteur d’amplification géométrique du chemin optique, l’estimation de la colonne fantôme masquée par les nuages, les effets de l’anomalie magnétique sud-atlantique des ceintures de radiation, et l’incidence du dioxyde d’azote troposphérique. Suit un diagnostic systématique de la mise au point du processeur GDP sur base des données globales fournies par les réseaux de l’OMM, ainsi qu’une critique comparée des algorithmes TOMS de la NASA.<p><p>Le dernier chapitre décrit la construction de la première climatologie globale du dioxyde d’azote stratosphérique et de ses variations harmoniques. Développée pour nos études du facteur d’amplification géométrique, cette climatologie composite est issue de l’analyse conjointe des jeux de données complémentaires acquis par différents satellites, par des réseaux au sol et par des ballons stratosphériques.<p> / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

La pollution atmosphérique constitue un risque environnemental majeur dont les effets néfastes sur la santé et sur l’environnement sont déjà clairement démontrés. Toutefois, la mesure d’exposition des populations aux polluants tels que les particules fines et l’ozone troposphérique demeure approximative en raison de la faible densité des stations d’échantillonnage de ces polluants. Peu d’études ont considéré la variation spatiale intra-urbaine dans la modélisation spatiale des concentrations de polluants. Certaines études ont cependant combiné interpolation spatiale et corrélation avec des facteurs locaux. De plus, l’effet du régime météorologique (par exemple l’occurrence d’une inversion de température) sur l’amplitude de ces corrélations n’est pas pris en compte. Cette thèse a donc pour objectif d’évaluer de nouvelles manières de caractériser la distribution spatiale et temporelle des particules fines (PM2.5) et de l’ozone troposphérique (O3) à l’échelle intra-urbaine. Plus particulièrement, nous avons développé un modèle de géostatistique multivariable appelée krigeage avec dérive externe (KED, kriging with external drift) basé sur l’intégration de variables auxiliaires dans le processus d’estimation journalière des PM2.5 et de l’O3. Le krigeage constitue une forme d’interpolation spatiale des données de stations de mesures éparses, alors que la dérive externe mise sur des corrélations entre des conditions locales (axes de transport routier, espaces verts, etc.) et la concentration des polluants atmosphériques. Afin de prendre en compte les variations temporelles, notamment celles reliées aux conditions météorologiques, ces modèles ont été développés par groupes basés sur des conditions synoptiques et six classes d’états établies selon la température, le vent, l’humidité relative et les précipitations, d’après des données météorologiques journalières. Les résultats montrent que l’intégration des variables auxiliaires telles que la densité de la végétation et les zones des activités industrielles locales dans le KED expliquent en partie les variations intra-urbaines des PM2.5 de l’île de Montréal, mais que cet apport est variable selon la classe météorologique. Ainsi, lorsque les corrélations sont très faibles, une interpolation spatiale simple, comme la méthode IDW (inverse distance weighting) est plus exacte que l’interpolation multivariable, alors que pour d’autres conditions synoptiques le KED produit les prédictions les plus certaines. Nous avons pour cette raison proposé un modèle d’interpolation hybride (KED-IDW) s’adaptant aux conditions météorologiques. Nous avons également montré, particulièrement dans le cas de l’O3, que le krigeage avec dérive externe améliore les résultats obtenus par krigeage ordinaire (sans variables auxiliaires). Cette thèse a aussi permis d’évaluer l’apport d’un modèle spatio-temporel (BME, bayesian maximum entropy) dans l’estimation de l’effet à court terme de l’exposition à l’O3 sur les décès à Montréal. Les résultats suggèrent que ce modèle spatio-temporel dans les conditions développées (par ex. basé sur les données de 12 stations de mesures, pour un territoire de 1 310 km2) n’apporte pas de gains significatifs dans l’estimation de l’effet de l’exposition. Dans l’ensemble, cette thèse contribue au progrès de modélisation spatiale empirique des polluants atmosphériques en se fondant notamment sur l’adaptation aux conditions météorologiques et par l’ajout de certains facteurs météorologiques comme prédicteurs. Dans ce contexte, cette thèse ouvre une voie prometteuse pour l’amélioration des estimations de polluants atmosphériques à l’échelle intra-urbaine et de la capacité à évaluer les risques à la santé de la population par une meilleure caractérisation de l’exposition. Mots-clés : pollution de l’air, particules fines, ozone troposphérique, santé, géostatistique, krigeage avec dérive externe, environnement urbain. / Air pollution is a major environmental hazard with clearly demonstrable adverse effects on health and the environment. However, the measurement of populations’ exposure to pollutants such as particulate matter and ground-level ozone remains approximate due to the low density of sampling stations for these pollutants. Moreover, intra-urban spatial variation in the spatial modeling of pollutant concentrations has received little research attention. If anything, some studies have combined spatial interpolation and correlation with local factors; however, they do so without examining the effect of the weather regime (e.g., a temperature inversion) on the magnitude of these correlations. In order to overcome these shortcomings, this dissertation aims to evaluate new ways of characterizing the spatial and temporal distribution of fine particles (PM2.5) and ground-level ozone (O3) at the intra-urban scale. In particular, we developed a multivariable geostatistical model called “kriging with external drift” (KED) based on the integration of auxiliary variables into the process of estimating daily PM2.5 and O3 concentrations. Kriging is a form of spatial interpolation of data from measurement stations that are dispersed, while external drift is based on correlations between local conditions (road transport arteries, green spaces, etc.) and the concentration of atmospheric pollutants. In order to take account of temporal variations, especially those related to weather conditions, we designed these models around six synoptic weather classes based on daily meteorological data (such as temperature, wind, relative humidity and precipitation). v The results show that the integration of auxiliary variables (such as vegetation density and local industrial activity areas) in KED partly explains the intra-urban variations of PM2.5 on the island of Montreal, but that this contribution is variable depending on the weather conditions. Thus, when the correlations are very low, a simple spatial interpolation (such as the inverse distance weighting (IDW) method) is more accurate than multivariable interpolation, whereas for other synoptic conditions KED produces the most certain predictions. For this reason, we proposed a hybrid interpolation model (KED-IDW) that can adapt to different weather conditions. We have also shown, particularly in the case of O3, that KED improves the results obtained from ordinary kriging (without auxiliary variables). This dissertation also allowed to evaluate the contribution of a spatial-temporal model—BME (bayesian maximum entropy)—in the estimation of the short-term effect of exposure to O3 on deaths in Montreal. The results suggest that this spatio-temporal model under the determined conditions (e.g., based on data from 12 measurement stations, for a territory of 1 310 km2) does not offer significant improvements to the estimation of the effect of exposure. Overall, this dissertation contributes to the advancement of the empirical spatial modeling of air pollutants, namely by taking into account the adaptation to weather conditions as well as certain predictive meteorological factors. In this context, the dissertation opens up a promising path for improving the estimation of air pollution at the intra-urban scale and the capacity to assess population health risks through better characterization of exposure.

L'augmentation des concentrations de fond en ozone dans la troposphère depuis le XXéme siècle est responsable de baisses conséquentes des rendements des grandes cultures. Le maïs ne semble pas épargné et les pertes de rendement pour l'année 2000 seraient de l'ordre de 2,2% à 5%. Toutefois, ces estimations sont établies à partir des résultats d'un nombre très faible d'expérimentations, toutes réalisées en chambres d'ozonation à ciel ouvert. Afin de vérifier si ces projections sont véritablement transposables aux champs de maïs cultivé en conditions conventionnelles, des plants de la variété NK Perform ont été mis en culture et exposés en champ à différentes concentrations d'ozone, en conditions semi-contrôlées. Pendant le développement des plantes, des séries d'échantillons ont été prélevées afin de comprendre comment le stress oxydatif, potentiellement induit par l'ozone, pourrait les avoir affectées. Ainsi les niveaux d'activités endoprotéolytiques des feuilles de l'épi (précédemment proposés comme indicateurs du degré de contrainte hydrique) ont été mesurés à l'aide d'une méthode par fluorescence (nouvellement adaptée chez les plantes). Parallèlement, les variations des teneurs en protéines oxydées (groupements carbonyles), de la peroxydation lipidiques et de la teneur en espèces réactives de l'oxygène (ROS) ont également été analysées. Nos résultats mettent en évidence une stimulation des activités endoprotéolytiques en réponse aux niveaux élevés d'exposition à l'ozone, ainsi qu'une différence significative de teneurs en protéines oxydées entre les plants contrôles et les plants les plus exposés. De mêmes les teneurs en ROS et les niveaux de peroxydation lipidique témoignent d'un effet de l'exposition à l'ozone. Toutes ses réponses cellulaires sont également influencées par l'âge des tissus foliaires. Ces résultats semblent abonder dans le sens des modèles en démontrant un impact certain de l'ozone sur le maïs, cependant toutes les analyses menées sur les paramètres de rendement (poids de mille grains, biomasse, teneur en amidon…) ne laisse entrevoir aucune perte, nous obligeant alors à reconsidérer la sensibilité du maïs à l'ozone, généralement admise jusqu'ici / For the past 150 years, background tropospheric ozone concentrations have been increasing constantely to the point where they now affect grain yield in major cereals, such as maize. In 2000, it has been estimated that yield loss was between 2.2% and 5% in this crop. Such estimates have been established from a very low number of experiments, all carried out in open top ozone fumigation chambers. To verify the accuracy of these estimations, we cultivated maize plants and exposed them to various ozone concentrations in the field. During plant development, series of cob-leaf samples have been collected in order to analyze the impact of ozone-induced oxidative stress on various biochemical processes. Thus, we studied changes in leaf endoproteolytic activities (a parameter previously used as a dehydration stress indicator), using a fluorescence-based method newly adapted to plant tissues. Concurrently, changes in protein oxidation levels (carbonyl groups) were analyzed, along with lipid peroxidation and accumulation of reactive oxygen species (ROS).Our results indicate that ozone induced increases in the global level of protein oxidation, endoproteolytic processes and lipid peroxidation, most likely as a result of an over-accumulation of ROS in the leaf tissues. Furthermore, the impact of ozone is enhanced by aging. To some extent, these conclusions agree with those obtained from impact modeling that also show that maize is midly sensitive to ozone. However, because yield was not affected whatsoever in our experiment (1000 grain weight, biomass, starch accumulation), it is our opinion that the general consensus about the sensitivity of maize to ozone should be revised

Dans un contexte de changement climatique, il est important d'étudier les processus de distribution et les tendances des gaz à l'état de traces dans l'atmosphère tels que l'ozone et la vapeur d'eau. L'observatoire atmosphérique du Maïdo, situé à 2160 m, héberge un ensemble d'instruments de mesures atmosphériques in-situ et par télédétection, dont le Lidar1200, pour la mesure de profils verticaux de vapeur d'eau. La validation, l'évaluation et une première exploitation des données du Lidar1200 font l'objet de cette thèse. Des comparaisons avec des radiosondages ont permis de valider les profils de vapeur d'eau du lidar depuis la basse troposphère jusqu'à 22 km d'altitude dès lors qu'un nombre suffisant de données est intégré. À courte échelle de temps, le Lidar1200 est capable de suivre temporellement des structures de l'ordre de la centaine de mètres d'épaisseur jusqu'à une dizaine de kilomètres d'altitude. L'analyse des données montre le potentiel du Lidar1200 de surveillance de la vapeur d'eau à long-terme dans la basse stratosphère avec un ou deux points de mesure par mois. La base de données 2013-2015 apporte des informations sur la variabilité saisonnière de la vapeur d'eau au-dessus de La Réunion. Les données de vapeur d'eau permettent de documenter divers processus atmosphériques dont les échanges stratosphère-troposphère. Des signatures d'intrusion stratosphérique ont été observées sur des observations lidar quasi simultanées et co-localisées d'ozone et de vapeur d'eau pendant la campagne de mesure MALICCA (MAïdo LIdar Calibration Campaign, La Réunion), le 4 avril 2013, suggérant un processus d'intrusion stratosphérique avec plusieurs origines. / In a context of climate change, it is important to study distribution processes and the trends of trace atmospheric gases such as ozone and water vapor. The Maïdo atmospheric observatory, located at 2160 m in height, hosts a set of atmospheric in-situ and remote sensing instruments, including the Lidar1200 providing vertical water vapor profiles. This thesis aims to provide a thorough work of validation, evaluation and analysis of the Lidar1200 data. Comparisons with radiosoundings allowed to validate the water vapor profiles of the lidar from the lower troposphere up to an altitude of 22 km if a sufficient number of data is integrated. At a smaller temporal scale, the Lidar1200 is able to follow temporally structures in the range of a hundred of meters thick up to a dozen of kilometers in altitude. The data analysis show the potential of the lidar in monitoring the water vapor in the lower stratosphere on a long term with one to two points of measurements by month. The database 2013-2015 brings information on the seasonal variability of the water vapor over Reunion Island. The water vapor data allow to study various atmospheric processes including stratosphere-troposphere exchanges. Signatures of stratospheric intrusions were observed on quasi-simultaneous and co-localized ozone and water vapor observations realized by two lidars of the Maïdo Observatory during the measurement campaign MALICCA-1 (MAïdo LIdar Calibration CAmpaign, Reunion Island), on April 4, 2013, suggesting stratospheric intrusions with multiple origins.

Ce travail a eu pour objectif l’étude de la dégradation hétérogène d’hydrocarbures aromatiques polycycliques (HAP) et de polybromo diphényles éthers (PBDE) en présence d’oxydants atmosphériques. La réactivité du benzo[a]pyrène et du dibenzo[a,l]pyrène pour les HAP et du BDE 153 (hexabromé) et du BDE 209 (décabromé) pour les PBDE, a été étudiée lorsqu’ils sont adsorbés sur des particules minérales modèles (des particules de silice) et exposés à l’ozone, au dioxyde d’azote ou aux radicaux hydroxyles. L’extraction et l’analyse de ces composés ont été optimisées. Le suivi de l’évolution de la concentration en composé adsorbé en fonction de son temps d’exposition à l’oxydant a permis de déterminer les paramètres cinétiques de chaque composé, pour chaque oxydant. L’influence de la concentration particulaire et de la structure moléculaire sur la réactivité a ainsi été évaluée et débattue. L’identification des produits de réactions majoritaires a été effectuée pour les réactions des HAP avec le dioxyde d’azote et les radicaux hydroxyles et les réactions des PBDE avec les radicaux hydroxyles. Enfin, l’étude de la génotoxicité des extraits de particules pour différents temps d’exposition aux oxydants a été réalisée. Cette étude regroupe une approche cinétique, une approche mécanistique et une approche toxicologique. / This work deals with the heterogeneous degradation of polycyclic aromatic hydrocarbons (PAH) and polybromo diphenyls ethers (PBDE) with various atmospheric oxidants. The reactivity of Benzo[a]pyrene (BaP) and dibenzo[a,l]pyrene (DalP) for PAH and BDE 153 and BDE 209 for PBDE, had been studied when they are adsorbed on a model of mineral particles (silica particles) and exposed to ozone, nitrogen dioxide and hydroxyl radical. Extraction and analyse of compounds were optimised. For all studied reactions, kinetic parameters were obtained following the remaining concentration of compound for different exposure times. Influence of the particle loading and molecular structure on the reactivity had been evaluated and discussed. Identification of the main degradation products was made for the reaction of PAH with nitrogen dioxide and hydroxyl radical and for the reaction of PBDE with hydroxyl radical. Then, the genotoxicity of the particle extracts was made for different exposure times. This study talks about kinetics, mechanistic and genotoxicity.

An analysis of air quality data from British Columbia has identified the Lower Mainland and surrounding rural areas as one of the regions in Canada where the Canadian Maximum Acceptable Air Quality Objective of 0.082 ppm ozone for one hour is frequently exceeded. Ozone at this level has the potential for affecting crops in the Fraser Valley. Field experiments were undertaken to attempt to evaluate the effect of randomly fluctuating levels of ozone on the yield of two cultivars each of Brassica oleracea L. (broccoli), Phaseolus vulgaris L. (bean), Pisum sativum L. (pea), Daucus carota L. (carrot) in 1985 and on one cultivar each of Solanum tuberosum L. (potato) and Pisum sativum L. (pea) in 1986, using a zonal air pollution system. As there is no current consensus regarding the most appropriate numerical expression of pollutant exposure to use in vegetation response studies, a comparison of various exposure terms was also undertaken. Ozone was added in various proportions to ambient levels between 0700 and 2100 hours (PDT) throughout the growing season. Three levels of ozone addition were used in 1985 and 12 in 1986. In 1985, treatments were assigned to three blocks over which ozone was released. Each block was supplied with different total amounts of supplementary ozone, a fourth block serving as an ambient air control. In 1986, ozone treatments were randomly assigned to four sub-plots on each of the three blocks over which ozone was released, with each block receiving the same total amount of supplementary ozone. Different treatments were achieved by each sub-plot being subject to different rates of release and degrees of mixing. For both years the ozone concentration distributions achieved over the season were approximately log-normal. Additional analysis of the air quality data from the ambient air plot found other types of skewed distributions such as the three parameter Weibull, three parameter gamma and Johnsons SB (four parameter log-normal) provided better descriptions of the data. The distribution providing the best fit depended on the concentration averaging time, the daily time span over which the ozone concentrations are analyzed and the selection criterion used. In 1985, field observations indicated that there were numerous plot to plot differences for disease and soil factors, which were confounded with the ozone treatments applied. Without true replication of the treatments, differentiation between the effects due to ozone and those from abiotic and biotic causes was not possible, and hence no clear conclusions concerning ozone response could be drawn. In 1986, without the confounding of ozone and plot location, significant linear reductions in yield were found for pea and pod fresh weight using the number of days on which the concentration exceeded 25 ppb, during the vegetative growing period (D25²). A significant linear reduction in fresh potato tuber weight was found using the geometric mean of all geometric mean ozone concentrations computed between 1200 and 1259-h for the season (GH12) as the exposure statistic. A significant multiple linear regression was found for pea fresh weight using the D25¹, statistic together with the number of occurrences in which the concentration exceeded 25 ppb for two, three and four consecutive hours (2C25, 3C25 and 4C25 respectively) in an episode; and for pod fresh weight using D25, 2C25 and 3C25 as independent variates. From the results presented it seems clear that ozone at the concentrations dispensed here would have a significant negative impact on crops grown in the Fraser Valley. Based on the 1986 experiment the best case estimate indicates that yield reductions of 28% could be expected for peas and potatoes at 37 ppb ozone (expressed as the season-long 7-h mean, M7). The exposure statistics used in the present study in comparison with the season-long 7 and 12 hour means provided good fits with the data. They are easily calculated from ambient air quality data and present attractive alternatives to those exposure statistics currently in use, for assessing the potential impact of ozone on crops in the Fraser Valley and for use in the setting of air quality standards. / Land and Food Systems, Faculty of / Graduate

No description available.

Climate change and ozone depletion are topical challenges the world over and are both attributed mainly to human activities, particularly emissions of ozone depleting substances (ODSs). One such substance is chlorodifluoromethane (HCFC-22), a cheap, widely used refrigerant with a high global warming potential of 1780. Botswana is a signatory to the Montreal Protocol (MP), which guides international efforts to phase-out HCFC-22 and requires signatories to develop and implement a country-level Hydrochlorofluorocarbon Phase-out Management Plan (HPMP). This study, which used a mixed methods approach, was conducted to evaluate the phase-out of HCFC-22 management strategies and their environmental and socioeconomic implications in Botswana. A census of nine HCFC-22 importing companies was conducted and probability sampling proportional to size was used to select a sample of 159 respondents from the Department of Meteorological Services, HCFC-22 importers, customs officers from 20 purposively selected Botswana entry ports and HCFC-22 consumers from the importing companies. Category-specific respondent questionnaires and interview guides, site visits and assessment of records were used to gather data. Of particular interest were the annual HCFC-22 importation figures for each company, the Botswana Unified Revenue Services and the National Ozone Unit, as well as the level of compliance of the companies’ HCFC-22 phase-out management practices with relevant national regulations, the Botswana HPMP and the MP resolutions. Botswana’s HCFC-22 importers were found to be moderately to highly compliant to nonregulatory elements rather than regulatory elements. Overall, HCFC-22 consumption decreased from the baseline to 10.5% for the first stage (2013-2015), which was slightly more than the 10% reduction expected. A steady decrease in HCFC-22 consumption was noted towards the 35% target for 2020, largely due to awareness-raising initiatives directed at the surveyed stakeholders. Absolute HCFC-22 consumption dropped by approximately 510400 kgs from 2011-2017 or 28072 ozone depleting potential saved. On the downside, gaps were identified in the industry-wide quota-system, data reporting, prevention of illegal ODS trade, service technician training, user knowledge of alternatives and disposal of ODS equipment. The study recommends the use of a planning, policy formulation and implementation framework that integrates and balances three fundamentals, namely, stakeholder involvement, the process and the plan enablers.

The main objective of this study was to determine the applicability of advanced water treatment processes namely granular activated carbon (GAC) adsorption, ultraviolet (UV) light disinfectant and ozone in the management of deteriorating water quality in the Mid-Vaal River system for drinking purposes. Both the scarcity and the deteriorating quality of water in South Africa can be addressed by investigating advanced water treatment processes such as GAC adsorption, UV light disinfectant and ozone. Previously disregarded water resources have the potential to be purified and advanced treatments can improve water quality where conventional water treatments have failed. In addition, advanced treatment processes can be applied to treat used water. The two sampling sites selected for the study, Rand Water Barrage (RWB) and Midvaal Water Company (MWC), are both located in the Middle Vaal Water Management Area with RWB upstream of MWC. RWB uses GAC adsorption and UV light disinfection and MWC uses ozone as pre- and intermediate treatment process steps for water purification. The quality of the source water at both sampling sites was determined by analysing the physical and chemical characteristics as well as the algal and invertebrate compositions of the source water. The physical and chemical water quality variables measured included pH, conductivity, turbidity, dissolved organic carbon (DOC), total organic carbon (TOC), total photosynthetic pigments (TPP), microcystin and geosmin. The source water of both sites was characterised as hypertrophic on account of high chlorophyll concentrations. The water quality of the two sites was distinctly different and a downstream change was observed. The source water of RWB was characterised by high microcystin, geosmin, DOC, TOC and conductivity measurements and dominated by Bacillariophyceae (diatoms) and Cyanophyceae (blue-green bacteria). Problematic species that were present in the source water of RWB included Aulacoseira sp., other unidentified centric diatoms, Pandorina sp., Anabaena sp., Microcystis sp., Oscillatoria sp., Cryptomonas sp., Ceratium sp. and Trachelomonas sp. The source water of MWC was characterised by high pH, turbidity and TPP measurements and was dominated by Chlorophyceae (green algae) and Bacillariophyceae (diatom) species. Problematic algal species that were present in the source water of MWC included Cyclotella sp., Coelastrum sp., Pediastrum sp. and Scenedesmus sp. The source water of MWC was deemed to be of a better quality due to the lower Cyanophyceae concentrations and lower microcystin levels. The invertebrate composition of both sites was similar with Rotatoria as the dominant invertebrate group. The efficacy of GAC adsorption/UV light disinfection/ozonation on restoring the physical and chemical characteristics of the source water at both sampling sites as well as the algal and invertebrate compositions was determined by ascertaining the nature of the change in or the percentage removal of a water quality variable. The potable water of both sites complied with the standards of water intended for domestic use except for the conductivity at RWB that was slightly elevated. The phytoplankton was removed effectively from the source water of both sites but the removal of invertebrates was unsatisfactory. GAC adsorption and filtration proved to be more effective in the removal of TPP, turbidity, DOC, microcystin and geosmin than ozone. Ozone effected an increase in DOC. UV light disinfection had no or little effect on restoring the water quality variables investigated in this study. / M (Environmental Sciences), North-West University, Potchefstroom Campus, 2015

The main objective of this study was to determine the applicability of advanced water treatment processes namely granular activated carbon (GAC) adsorption, ultraviolet (UV) light disinfectant and ozone in the management of deteriorating water quality in the Mid-Vaal River system for drinking purposes. Both the scarcity and the deteriorating quality of water in South Africa can be addressed by investigating advanced water treatment processes such as GAC adsorption, UV light disinfectant and ozone. Previously disregarded water resources have the potential to be purified and advanced treatments can improve water quality where conventional water treatments have failed. In addition, advanced treatment processes can be applied to treat used water. The two sampling sites selected for the study, Rand Water Barrage (RWB) and Midvaal Water Company (MWC), are both located in the Middle Vaal Water Management Area with RWB upstream of MWC. RWB uses GAC adsorption and UV light disinfection and MWC uses ozone as pre- and intermediate treatment process steps for water purification. The quality of the source water at both sampling sites was determined by analysing the physical and chemical characteristics as well as the algal and invertebrate compositions of the source water. The physical and chemical water quality variables measured included pH, conductivity, turbidity, dissolved organic carbon (DOC), total organic carbon (TOC), total photosynthetic pigments (TPP), microcystin and geosmin. The source water of both sites was characterised as hypertrophic on account of high chlorophyll concentrations. The water quality of the two sites was distinctly different and a downstream change was observed. The source water of RWB was characterised by high microcystin, geosmin, DOC, TOC and conductivity measurements and dominated by Bacillariophyceae (diatoms) and Cyanophyceae (blue-green bacteria). Problematic species that were present in the source water of RWB included Aulacoseira sp., other unidentified centric diatoms, Pandorina sp., Anabaena sp., Microcystis sp., Oscillatoria sp., Cryptomonas sp., Ceratium sp. and Trachelomonas sp. The source water of MWC was characterised by high pH, turbidity and TPP measurements and was dominated by Chlorophyceae (green algae) and Bacillariophyceae (diatom) species. Problematic algal species that were present in the source water of MWC included Cyclotella sp., Coelastrum sp., Pediastrum sp. and Scenedesmus sp. The source water of MWC was deemed to be of a better quality due to the lower Cyanophyceae concentrations and lower microcystin levels. The invertebrate composition of both sites was similar with Rotatoria as the dominant invertebrate group. The efficacy of GAC adsorption/UV light disinfection/ozonation on restoring the physical and chemical characteristics of the source water at both sampling sites as well as the algal and invertebrate compositions was determined by ascertaining the nature of the change in or the percentage removal of a water quality variable. The potable water of both sites complied with the standards of water intended for domestic use except for the conductivity at RWB that was slightly elevated. The phytoplankton was removed effectively from the source water of both sites but the removal of invertebrates was unsatisfactory. GAC adsorption and filtration proved to be more effective in the removal of TPP, turbidity, DOC, microcystin and geosmin than ozone. Ozone effected an increase in DOC. UV light disinfection had no or little effect on restoring the water quality variables investigated in this study. / M (Environmental Sciences), North-West University, Potchefstroom Campus, 2015