272f. / Submitted by Ana Hilda Fonseca (anahilda@ufba.br) on 2013-04-03T15:02:31Z No. of bitstreams: 1 TESE DE DOUTORADO_ELIZABETH DA ROCHA COUTO.pdf: 5379213 bytes, checksum: abcb42aa71cfdf41403c0028e4fa7652 (MD5) / Approved for entry into archive by Ana Hilda Fonseca(anahilda@ufba.br) on 2013-04-23T15:33:05Z (GMT) No. of bitstreams: 1 TESE DE DOUTORADO_ELIZABETH DA ROCHA COUTO.pdf: 5379213 bytes, checksum: abcb42aa71cfdf41403c0028e4fa7652 (MD5) / Made available in DSpace on 2013-04-23T15:33:05Z (GMT). No. of bitstreams: 1 TESE DE DOUTORADO_ELIZABETH DA ROCHA COUTO.pdf: 5379213 bytes, checksum: abcb42aa71cfdf41403c0028e4fa7652 (MD5) Previous issue date: 2011 / CAPES / O desenvolvimento desse estudo parte da motivação de aplicar esforços técnicos e científicos visando avaliar as transformações do SO2 e NO2 na atmosfera da área de influência do Pólo Industrial de Camaçari. Este estudo teve como principal objetivo caracterizar a atmosfera da área estudada através da deposição úmida e em relação ao SO2, NO2 e suas transformações na fase gasosa e particulada <2 μm. As campanhas de amostragem foram feitas em três estações da Rede de Monitoramento do Ar da Cetrel S.A: Gravatá, em Camaçari; Escola, em Dias D´Àvila e Lamarão, em Lamarão do Passé. As concentrações de espécies inorgânicas foram determinadas na precipitação da região e as contribuições para a formação da precipitação ácida foram identificadas. Foram medidos pH, condutividade e os principais constituintes inorgânicos da chuva (Na+, K+, Ca2+, Mg2+, NH4+, SO42-, NO3- e Cl-). Esta se caracteriza por pH entre 3,8 e 7,6 e elevado número de eventos ácidos (pH<5,6), com maior frequência de acidez na chuva amostrada na estação Lamarão. Cl-, Na+ e NH4+ são os íons predominantes na deposição apenas úmida das três áreas monitoradas no período 2006-2007. A incorporação do spray marinho na atmosfera da área de influência do Pólo Industrial de Camaçari neutraliza parte da acidez das chuvas, freqüentemente em baixos percentuais, com significância até 25% de neutralização, associada à direção e velocidade dos ventos oriundos do oceano. Os níveis de amônio comparados aos de SO42-nss (non sea salt) encontrados na chuva daquela área, no período de 2006 a 2008, são suficientes para neutralizar também por esse meio uma parcela significativa da acidez da chuva da região. Correlações significativas e comuns a todas as estações foram encontradas para os pares de íons: Na+ x Cl-, Na+ x Mg2+ e Mg2+ x Cl-, indicando uma origem comum para esses íons e associada às massas de ar oceânicas transportadas para a região, principalmente, quando a predominância dos ventos era de Leste e Sudeste. Entre os compostos inorgânicos em fase gasosa medidos na atmosfera da área estudada, predomina amônia, cujos níveis médios em 2008 são comparáveis a outras áreas com influência industrial (150 nmol m-3 na Estação Gravatá;133 nmol m-3 na Estação Escola e 193 nmol m-3 na estação Lamarão). Entre os ácidos fortes determinados na atmosfera, HCl predomina em função de suas emissões locais como poluente primário com faixa de concentração nas três estações com valores superiores àqueles reportados para outras áreas industriais brasileiras. Isso também ocorre de forma semelhante para HNO3 e em menor escala para H2SO4. Em todos os casos, a atmosfera em Lamarão apresenta concentrações mais altas para os ácidos, inclusive para o H2SO4, embora a estação Gravatá seja o local que apresenta concentrações mais altas do precursor (SO2). Isso pode ser explicado pela presença de maior concentração de partículas de ferro naquela atmosfera, favorecendo a formação de H2SO4 por ação catalítica. Em relação à fase particulada <2 μm, amostrada no período de 2008, as massas de ar naquela região são moderadamente enriquecidas por sais de / Salvador

Os processos de remoção de espécies químicas da atmosfera têm sido estudados atualmente utilizando modelos numéricos, na tentativa de compreender melhor, os processos de transferências de gases e material particulado (sejam elas naturais ou antropogênicas) intra-reservatórios na atmosfera e seus efeitos na dinâmica do tempo e clima. Neste estudo, foi utilizado o modelo RAMS para simular a estrutura vertical das nuvens que se desenvolvem na região amazônica, em conjunto ao modelo de remoção B. V. 2, para os processos de remoção úmida que ocorrem tanto dentro quanto abaixo da nuvem, além das condições atmosféricas locais da região da Bacia Amazônica para, assim, simular a transferência das espécies químicas da atmosfera para a hidrosfera dentro do escopo do projeto LBA. Dentro deste projeto, foram realizadas campanhas intensivas de medições, como a LBA/DRY-TO-WET e LBA/SMOCC (setembro a novembro de 2002) na região de Rondônia. No período das campanhas, foram realizadas medições das concentrações dos gases amônia, ácido nítrico e dióxido de enxofre, além das espécies inorgânicas solúveis em água, como amônio, nitrato e sulfato, entre outros. Estas concentrações de gases e partículas, bem como os parâmetros meteorológicos obtidos durante as campanhas, realizadas durante o período de transição entre as estações seca e chuvosa na região, foram utilizados como dados de entrada para ambos os modelos, onde foram escolhidos alguns eventos específicos. Com intuito de melhor representar o espectro de gotículas de nuvens no modelo de remoção, foram utilizadas a função de distribuição de Levine & Schwartz, 1982 e funções gama ajustadas aos dados observados em distribuição de gotículas de nuvem obtidas em vôos efetuados durante o experimento. Conseqüentemente, este trabalho visou à simulação da concentração na água de chuva de três espécies químicas (SO42-, NO3- e NH4+) removidas da atmosfera pelo evento de precipitação, comparando-as às composições químicas da água de chuva observadas experimentalmente, em dois eventos selecionados (9 e 10 de outubro de 2002). Simulações atmosféricas com o RAMS apresentaram resultados bastante satisfatórios conseguindo representar aspectos microfísicos das nuvens que se desenvolvem na região amazônica com bastante fidelidade. Os resultados da modelagem dos processos de remoção mostraram uma boa concordância com os observados, principalmente para o sulfato (que em alguns casos a quantidade encontrada na água de chuva pela simulação foi 97% da observada) em ambos os eventos, quando a altura da nuvem foi considerada mais realista para região (16 km). Além disso, observou-se que o espectro de gotículas de nuvem utilizado foi um parâmetro importante nos resultados. Os resultados mostraram ainda, uma predominância dos processos que ocorrem dentro da nuvem, sendo estes responsáveis por cerca de 80% a 97% da concentração da espécie química encontrada na água de chuva, corroborando a literatura. Com isso, ficou evidente a complexidade das interações e transferências entre os reservatórios atmosfera / hidrosfera através dos processos de remoção de poluentes, ressaltando assim, a importância dos estudos sobre este assunto. / The scavenging processes of chemical species have been studied using numerical modeling in order to understand the gases and particulate matter intra-reservoir transferences (natural or anthropogenic) which affect weather and climate. In this study RAMS model was used in turn to simulate cloud vertical structure formed over Amazonian area working together to B.V.2 scavenging model. The last model was used to simulate the in- and below-cloud scavenging processes, besides the local atmospheric conditions within the LBA Project. In this Project, there were evaluated many measurements of LBA/DRY-TO-WET and LBA/SMOCC (September to November) Campaigns at Rondonia State. During the Campaigns, ammonia, nitric acid and sulfur dioxide gases were evaluated and their respective particulate matter, ammonium, nitrate and sulfate, among others, as well as rainwater chemistry. These concentrations and meteorological parameters were also obtained, during the transition from dry to wet season, and used as input data to the both modeling, where some events were chosen. With the intention of modeling improvement, cloud droplet spectra were used from Levine & Schwartz, 1982 and gamma functions, according to each case and based on the droplet distribution obtained from flight collected data during the field Campaign. Consequently, this work simulated the rainwater concentrations of three chemical species (SO42-, NO3- e NH4+) scavenged from atmosphere by the precipitation event and compared to the observed data of two selected events (9 and 10 October 2002). RAMS atmospheric simulations presented satisfactory results which showed detailed cloud microphysics processes of Amazonian region. The modeling results show good agreement of observed data, mainly to sulfate, reaching 97% of the observed sulfate for both events, when the cloud height was considered more realistic for the region (16 km). Besides, the cloud droplet spectra were an important parameter to the modeling. The results also showed that the in-cloud process is responsible by 80% to 97% of the chemical species found in rainwater. Additionally, it was clear that the complexity of the interaction and intra-reservoir transferences through the scavenging processes and their importance.

京都大学 / 0048 / 新制・課程博士 / 博士(地球環境学) / 甲第18628号 / 地環博第123号 / 新制||地環||25(附属図書館) / 31528 / 京都大学大学院地球環境学舎環境マネジメント専攻 / (主査)教授 藤井 滋穂, 教授 梶井 克純, 准教授 田中 周平 / 学位規則第4条第1項該当 / Doctor of Global Environmental Studies / Kyoto University / DFAM

Atmospheric mercury deposition, known to be a major source of mercury to aquatic and terrestrial environments, was studied at an urban site in Orlando, FL. Precipitation sampling was conducted from September 2003 to May 2006 at a Mercury Deposition Network site located on the University of Central Florida campus. Weekly rainfall and mercury wet deposition data were gathered from this site, which provided the framework of data for this study. Historical mercury wet deposition data from several sites in Florida were used to develop a regression model to predict mercury deposition at any location in Florida. Stormwater runoff from a 2-acre impervious surface at this study area was monitored during the spring and summer of 2005. Runoff water quality was analyzed to characterize mercury dry deposition. Atmospheric monitoring was also conducted during this period to study the interaction of atmospheric constituents on wet and dry deposition patterns. Spatial and seasonal trends for the entire state suggest 80% of Florida's rainfall and mercury deposition occur during the wet season. A strong linear correlation was established between rainfall depth and mercury deposition (R2 = 0.8). Prediction equations for the entire state, for both wet and dry seasons, were strongly correlated with measured data. The results of two unique methods to quantify dry deposition were similar at this site during this study period. Runoff monitored at this site contained significant levels of mercury, primarily in particulate form (58%). The vast majority of particulate mercury was flushed from the surface during storm events, while significant dissolved fractions remained. Runoff mercury concentrations were consistently higher than rainfall mercury, suggesting dry deposition accounted for 22% of total mercury in runoff. Atmospheric monitoring at this location showed gaseous elemental mercury was the dominant form (99.5%) followed by reactive gaseous mercury (0.3%) and particulate mercury (0.2%). Comparison of the contributions of wet and dry deposition suggested 80% of total mercury deposition was wet deposited during this study, while dry deposition accounted for the remaining 20%. Statistical correlations revealed rainfall scavenging of reactive gaseous mercury was the main factor controlling dry deposition.

No description available.

I investigate the links between mercury use, release, deposition, and population exposure in Tampa Bay, with the focus of identifying levers for reducing population mercury exposures. To achieve this, I investigated the trends in mercury use and release by products and processes in the Tampa Bay area using a Material Flow Analysis. Analysis of USEPA National Emissions Inventory data over time (1999 - 2008) identified relevant air source emission categories, and explored and compared state and regional trends in mercury emissions. To understand source contributions to wet deposited mercury in the Tampa Bay area, I analyzed trends in mercury deposition data from the National Atmospheric Deposition Program, Mercury Deposition Network, and the 2001 Bay Regional Atmospheric Chemistry Experiment. I also collected wet deposition samples for mercury and trace metals in the Tampa Bay area during a 6-month campaign at a site at the University of South Florida (USF) campus. Samples were analyzed using Cold Vapor Atomic Fluorescence Spectrometry (CVAFS) for mercury, and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) for trace metals analysis. Concentration data were analyzed for source contributions using HYSPLIT back-trajectory meteorology-based modeling to assess source locations, and the Positive Matrix Factorization (PMF) statistical receptor model to apportion the deposition data by source type. To explore the factors influencing fish consumption behaviors of the local angler population, I analyzed population surveys collected previously from fisher-folks along the Hillsborough River, in Hillsborough County, Florida. Results from the mercury inventory indicate that mercury releases from industrial sources and dental facilities were the most important sources of mercury to the Tampa Bay area. Furthermore, the solid-waste pool was the most important direct sink in the domain, with air emissions an important indirect sink. Emissions inventory data indicated that coal-fired power plants were the largest contributors of mercury emissions in the Tampa Bay area. Medical and municipal waste incineration also accounted for significant fractions of total mercury releases to the domain. Emissions from sources in Hillsborough County accounted for a significant portion of mercury emissions in the region and state. Measurement data indicated that event mercury concentration was only very weakly correlated with event precipitation depth, with both studies showing agreement with this phenomenon. Back-trajectory simulations reveal that high mercury concentration events were often from air masses with recent trajectories over Florida land (6 and 24 hr), and with previous high precipitation depth events over the trajectory in the long term (72 hr). The statistical PMF results indicate the importance of coal burning power plant emissions, medical and municipal waste incineration, and agrochemicals on mercury in wet deposition in the Tampa Bay area. Changes were observed between the 2001 and 2012 data, including greater mercury concentrations in 2012, and the removal of medical waste incineration as a mercury source in the 2012 model results. Together with local emissions inventory data, these results suggest that sources local to the Tampa Bay area and in Florida likely contribute substantially to mercury deposition in the region. Finally, population survey data suggests that mercury exposure risks are poorly understood by the fishing population in Hillsborough County. Taken together, these results suggest that policies targeting mercury emissions control, particurlarly for coal-fired power plants and municipal waste processing, and fish consumption education may be instrumental to the protection of susceptible populations.

In Canada, more than 35 million kilograms of pesticide active ingredients are applied annually on agricultural land with about 84% of this mass applied in the Canadian Prairies Provinces. Pesticide residues may become airborne through post-volatilization losses from land, water or vegetative surfaces, via wind-eroded soil, and by spray drift during pesticide application. Once in the air, pesticides may be dispersed and transported as parent molecules or as degradation products. Currently used and legacy pesticide air concentrations were measured in the agricultural region of the Prairie Provinces and the northern subarctic and arctic regions of Canada in 2005 and 2007. More intensive air sampling was conducted across the province of Manitoba in 2008 and 2009. Separate wet versus dry atmospheric deposition samples and wetland water samples were also taken in these years. In general, pesticides were detected in the monitored environmental media as mixtures with the frequency and concentrations detected being largest for pesticides that were applied on-site. Although the most commonly used herbicides in western Canadian agriculture were frequently present in the air in the regions where they are applied, they appeared infrequently and at low levels in the Canadian Subarctic, Arctic, and remote sites far from agricultural pesticide applications. Results also indicated that the concentrations of legacy pesticides such as lindane (γ-HCH) and its manufacturing by-product prior to 1971 (α-HCH) continued to decrease over time in the Canadian environment. Air concentrations of pesticides in agricultural regions, and subsequent wet and dry atmospheric deposition, contaminated surface water following periods of pesticide application. Dry deposition contributed 12–51% of the total deposition. Calculations were performed to predict wetland water column pesticide concentrations based on total atmospheric deposits alone. The estimated concentrations were closest to actual concentrations for MCPA and predictions were also reasonable for a range of other herbicides, but a source other than atmospheric deposition was clearly relevant to the detection of clopyralid in the wetland water-column. Individual herbicide concentrations did not exceed Canadian water quality guidelines for the protection of aquatic life.

An extensive investigation with parallel field andlaboratory exposures has been conducted to elucidateatmospheric corrosion and metal runoff processes on copper andzinc used for roofing applications. Detailed studies have beenperformed to disclose the effect of various parameters on therunoff rate including: surface inclination and orientation,natural patination (age), patina composition, rain duration andvolume, rain pH, and length of dry periods inbetween rainevents. Annual and average corrosion rates and runoff rateshave been determined consecutively during urban field exposuresin Stockholm on naturally patinated copper and zinc of varyingage and patina composition. The corrosion rate was found todecrease with time, amounting to 6.7 g Cu/(m2.y) and 5.0 gZn/(m2.y) after 48 weeks of exposure, whereas the runoff ratewas relatively constant with time on a yearly basis, being 1.3g/(m2.y) and 3.1 g/(m2.y) for copper and zinc, respectively.The annual runoff rate was found to be significantly lower thanthe corresponding corrosion rate for both copper and zinc.Somewhat higher runoff rates of copper were determined fromnaturally green-patinated copper (&gt;40 years old, 2.0g/(m2.y)) compared to brown-patinated copper (1 year old). Themain reasons are specific environmental conditions combinedwith characteristics of the patina layer, which increase themagnitude of dissolved species flushed from the surface duringthe first flush volume of a rain event. No intrinsic effect ofpanel age on the runoff rate was seen for naturally patinatedzinc. However, differences in prevailing environmentalconditions during the initial exposure period and, hence,differences in formation rate and surface coverage of thecorrosion patina, resulted in variations in runoff rate. Thisinitial difference remained also during prolonged exposureperiods and was referred to as a memory effect. Model roof investigations and laboratory studies showedsurface orientation and inclination to have a detrimentaleffect on the runoff rate with high runoff rates from surfacesof low inclination from horizon and surfaces exposed towardsthe wind direction. Based on fieldexposures and literature data, a correlationwas established between the runoff rate and the prevailingSO2-concentration. The runoff rate increases with increasingSO2 level for exposure sites of similar annual precipitationquantities (500-1000 mm/y). A rain device, using artificialrain, was shown to successfully simulate outdoor rain events ofvarying intensity and pH and result in realistic runoff ratesof both copper and zinc. The device was used to monitor changesin metal concentration and quantity of runoff water duringindividual rain events. High metal concentrations are found inthe initial rain volume flushing the surface (first flush),which decreased to rather constant metal concentrations duringthe subsequent rain volume (steady-state). The magnitude offirst flush depends primarily on environmental conditions priorto a rain event and the characteristics of the corrosionpatina. The metal concentration in runoff water increases withrain acidity, decreases with rain intensity and increases withlength of the dry period preceding a rain event. A comparison between instantaneous corrosion rates,monitored by electrochemical impedance spectroscopy using a2-electrode set-up, and runoff rates during a continuous rainevent was performed for naturally patinated copper panels.Corrosion rates were found to be approximately 10 (brownishpatina) and 25 times (greenish patina) lower than correspondinginstantaneous runoff rates. A schematic description of the first flush and steady-stateregion of the runoff process was established. The magnitude ofthe concentration during first flush is primarily affected byprevailing environmental conditions prior to a rain event,while rain pH and intensity primarily affect the concentrationduring steady-state. <b>Key words:</b>atmospheric corrosion, corrosion rate, runoffrate, copper, zinc, field study, laboratory study, roof, firstflush, rain quantity, rain intensity, rain pH, dry and wetdeposition, corrosion and runoff process.

<p>An extensive investigation with parallel field andlaboratory exposures has been conducted to elucidateatmospheric corrosion and metal runoff processes on copper andzinc used for roofing applications. Detailed studies have beenperformed to disclose the effect of various parameters on therunoff rate including: surface inclination and orientation,natural patination (age), patina composition, rain duration andvolume, rain pH, and length of dry periods inbetween rainevents. Annual and average corrosion rates and runoff rateshave been determined consecutively during urban field exposuresin Stockholm on naturally patinated copper and zinc of varyingage and patina composition. The corrosion rate was found todecrease with time, amounting to 6.7 g Cu/(m2.y) and 5.0 gZn/(m2.y) after 48 weeks of exposure, whereas the runoff ratewas relatively constant with time on a yearly basis, being 1.3g/(m2.y) and 3.1 g/(m2.y) for copper and zinc, respectively.The annual runoff rate was found to be significantly lower thanthe corresponding corrosion rate for both copper and zinc.Somewhat higher runoff rates of copper were determined fromnaturally green-patinated copper (>40 years old, 2.0g/(m2.y)) compared to brown-patinated copper (1 year old). Themain reasons are specific environmental conditions combinedwith characteristics of the patina layer, which increase themagnitude of dissolved species flushed from the surface duringthe first flush volume of a rain event. No intrinsic effect ofpanel age on the runoff rate was seen for naturally patinatedzinc. However, differences in prevailing environmentalconditions during the initial exposure period and, hence,differences in formation rate and surface coverage of thecorrosion patina, resulted in variations in runoff rate. Thisinitial difference remained also during prolonged exposureperiods and was referred to as a memory effect.</p><p>Model roof investigations and laboratory studies showedsurface orientation and inclination to have a detrimentaleffect on the runoff rate with high runoff rates from surfacesof low inclination from horizon and surfaces exposed towardsthe wind direction.</p><p>Based on fieldexposures and literature data, a correlationwas established between the runoff rate and the prevailingSO2-concentration. The runoff rate increases with increasingSO2 level for exposure sites of similar annual precipitationquantities (500-1000 mm/y). A rain device, using artificialrain, was shown to successfully simulate outdoor rain events ofvarying intensity and pH and result in realistic runoff ratesof both copper and zinc. The device was used to monitor changesin metal concentration and quantity of runoff water duringindividual rain events. High metal concentrations are found inthe initial rain volume flushing the surface (first flush),which decreased to rather constant metal concentrations duringthe subsequent rain volume (steady-state). The magnitude offirst flush depends primarily on environmental conditions priorto a rain event and the characteristics of the corrosionpatina. The metal concentration in runoff water increases withrain acidity, decreases with rain intensity and increases withlength of the dry period preceding a rain event.</p><p>A comparison between instantaneous corrosion rates,monitored by electrochemical impedance spectroscopy using a2-electrode set-up, and runoff rates during a continuous rainevent was performed for naturally patinated copper panels.Corrosion rates were found to be approximately 10 (brownishpatina) and 25 times (greenish patina) lower than correspondinginstantaneous runoff rates.</p><p>A schematic description of the first flush and steady-stateregion of the runoff process was established. The magnitude ofthe concentration during first flush is primarily affected byprevailing environmental conditions prior to a rain event,while rain pH and intensity primarily affect the concentrationduring steady-state.</p><p><b>Key words:</b>atmospheric corrosion, corrosion rate, runoffrate, copper, zinc, field study, laboratory study, roof, firstflush, rain quantity, rain intensity, rain pH, dry and wetdeposition, corrosion and runoff process.</p>

Measurement of chemical composition of precipitation is important both to understand acidification of terrestrial and aquatic ecosystems and neutralization process in the atmosphere. Such data are scarce in the Mediterranean region. In this study, chemical composition of daily, wet-only, 387 number of rain water samples collected between 1991 and 1999 were investigated to determine levels, temporal variation and long-term trends in concentrations of major ions and trace elements between 1991 and 1999. Samples had already been collected and some of the analysis had been completed. The anions SO42-, NO3- and Cl- were analyzed by HPLC coupled with UV-VIS detector, NH4+ was analyzed by colorimetry and H+ ion was analyzed by pH meter. The major ions and trace metals were analyzed by using Atomic Absorption Spectrometry (AAS) and Graphite Furnace Atomic Absorption Spectrometry (GFAAS). In this study complete data set were generated by analyzing samples that had not been previously analyzed for major ions and trace elements with Inductively Coupled Plasma with Optical Emission Spectrometry (ICP-OES). Statistical tools were used to determine the distribution of the pollutants. The rain water data tends to be log-normally distributed since data show large variations due to meteorological conditions, physical and chemical transformations and air mass transport patterns. The median pH of the rain water was found to be 5.29, which indicates that the rain water is not strongly acidic. This case is not a result of lacking of acidic compounds but rather indicates extended neutralization process in rain water. Eastern Mediterranean atmosphere is under the influence of three general source types: (1) anthropogenic sources, which are located to the north and northwest of the basin brings low pH values to the region (SO42-, NO3- ions) / (2) a strong crustal source, which is dried and suspended local soil and air masses transported from North Africa transport which have high pH values (Ca2+, Al, Fe ions) and (3) a marine source, which is the Mediterranean Sea itself (Na+, Cl- ions). In the region, the main acid forming compounds are H2SO4 and HNO3 whereas / CaCO3 and NH3 are responsible for the neutralization process. To describe the level of pollutant concentrations and the factors that affect their variations in rain water / ion compositions, neutralization of acidity, short and long-term variability of ions and elements, their time trend analysis and wet deposition fluxes were investigated briefly. Positive matrix factorization (PMF) was used to determine components of ionic mass in the precipitation. In Antalya Station the rain water has five factors: free acidity factor, crustal factor, marine factor, NO3- factor and SO42- factor. Potential Source Contribution Function (PSCF) and trajectory statistics were used to determine source regions generating these components. NO3- has potential source regions of Western Mediterranean countries and North Africa, whereas SO42- has additional southeasterly trajectory components of Israel and south east of Turkey.