Groundwater is an important source of fresh water on earth. A source that is also affected by climate change. Climate change is expected to influence both the groundwater quality and the groundwater level, where elevated levels of iron and manganese in the groundwater are to be expected. In Sweden guideline values are set for iron and manganese in drinking water, which means that waterworks that use groundwater as raw water must treat the water when the guidline values are exceeded. A common treatment step to remove iron and manganese is abiotic oxidation with a chemical oxidant. A relatively new method is biological oxidation where microorganisms oxidize the metals. Biological oxidation has proven to be a faster process than chemical oxidation and better adapted to variations in raw water quality. In this work the possibility of biological oxidation of iron and manganese at smaller groundwater plants with day-to-day production breaks has been studied. The study also concern how seasonal variations in raw water quality might affect the oxidation ability of the microorganisms. During the study a pilot filter for biological oxidation was constructed where the efficiency of the filter and the raw water quality was measured during a one-year test. After a short start-up time for iron oxidation the pilot filter reduced iron in the raw water up to 92 % and after a longer start-up time for manganese oxidation the pilot filter reduced manganese from the raw water up to 97 % during the rest of the test period. The results show that biological oxidation is possible at smaller groundwater plants despite production breaks. The results also show that biological oxidation is not affected by the vaiations i raw water quality that was measured during the test period. Furthermore, the results show that biological oxidation is equally efficient as the chemical oxidation, at the investigated groundwater plant, in terms of removing iron and even more efficient in terms of removing manganese. In addition the biological oxidation do not require pH adjustment or addition of any chemicals. The results indicate that biological oxidation of iron and manganese is a more favorable alternative to chemical oxidation in smaller groundwater plants on the impact of climate change on raw water quality.

Made available in DSpace on 2016-12-23T14:04:29Z (GMT). No. of bitstreams: 1 Waldileia Pereira Leal.pdf: 3453316 bytes, checksum: 63c0a56552f4b09c97956f4c0d9f711c (MD5) Previous issue date: 2013-07-05 / O 2,4-D é o segundo herbicida mais utilizado no Brasil, podendo chegar à água de abastecimento público. Devido à limitação do tratamento convencional na remoção de microcontaminantes e a capacidade do carvão ativado em removê-los, este estudo investigou a aplicação de carvão ativado pulverizado (CAP) na remoção do 2,4-D por processos de adsorção associado ao tratamento convencional utilizando água bruta do Rio Santa Maria da Vitória em equipamentos de jartest. Os diagramas de coagulação foram construídos através de estudo de tratabilidade de água utilizando o sulfato de alumínio e o policloreto de alumínio como coagulantes. Foram realizados ensaios de adsorção com carvões derivados da casca de coco (CAP 1) e de pinnus (CAP 2) que foram caracterizados e avaliados na sua capacidade de remover o 2,4-D. A remoção de matéria orgânica natural também foi determinada pela medição de carbono orgânico dissolvido (COD) e absorvância UV 254 nm. Para a quantificação e detecção dos analitos, ácidos 2,4-D, 2,4,5-T e o 2,4-DCP, foi desenvolvida e validada uma metodologia analítica em cromatografia líquida de alta eficiência (CLAE). Através dos dados de adsorção obtidos, o CAP 1 apresentou maior capacidade de remoção do 2,4-D e foi associado ao tratamento convencional, sendo adicionado quinze minutos antes, um minuto antes e um minuto após a adição do sulfato de alumínio. Os resultados indicaram que a adição do coagulante não influenciou na capacidade de adsorção do CAP 1 que apresentou-se eficiente na remoção do herbicida, atendendo ao padrão de potabilidade de água vigente / The 2,4-D is the second most widely used herbicide in Brazil, reaching the public water supply. Due to the limitation of conventional treatment in removing micro-contaminants and capacity of activated carbon to remove them, this study investigated the application of powdered activated carbon (PAC) in the removal of 2,4-D in adsorption process associated with conventional treatment using raw water from the Rio Santa Maria da Vitória in equipment jartest. The diagrams of coagulation were constructed through treatability study of water using aluminum sulfate and polyaluminum chloride as coagulants. Tests of adsorption with carbons derived from coconut shells (PAC 1) and pinus (PAC 2) were characterized and evaluated in its ability to remove 2,4-D. The removal of natural organic matter was also determined by measuring the dissolved organic carbon (DOC) and UV absorbance at 254 nm. To detection and quantification of analytes, acids 2,4-D, 2,4,5-T and 2,4-DCP an analytical methodology for high performance liquid chromatography (HPLC/DAD) was developed and validated. Through the adsorption data, the PAC 1 showed greater capacity for removal of 2,4-D and was associated with conventional treatment, fifteen minutes prior, one minute prior to or one minute after the addition of the aluminum sulfate. The results indicated that the addition of coagulant had no effect on adsorption capacity of the PAC 1 which showed to be being effective for the removal of herbicide, given the drinking water quality standards force

Water quality in the Vaalharts region in the Northern Cape Province, South Africa, decreased over the past few years and there was a need for the microbiological and physico-chemical assessment. This problem was identified through discussions with Vaalharts Water (Vaalharts Water User Association) in 2010 when the issue of the impact of deteriorating water quality on drinking water production was raised. It was thus important to investigate concerns of the water users association pertaining to water quality issues. The aim of this study was to assess the physico-chemical and microbiological quality of household water in the Vaalharts irrigation scheme. The main residential areas were Hartswater, Pampierstad, Jan Kempdorp and Warrenton. Faecal coliforms were detected in the raw water of all the drinking water distribution systems during 2011 and 2012. No faecal coliforms were detected in the household water during 2011. This was a very positive result, because not only did the household water comply with the SANS 241 (2011) standard (0 CFU/100ml), but the purification processes were successful by removing all the E. coli’s from the raw water. However, during March 2012 faecal coliforms were detected in the household water of Jan Kempdorp (191CFU/100ml). This could be due to point pollution and possible breakage of faecal coliforms in the distribution system. Low amounts of total coliforms were detected in the raw water of some of the drinking water distribution systems. This could be due to high amounts of other colonies (pink and purple) growing on the m-Endo agar which suppress the growth of the metallic green sheen (total coliform) colonies. The total coliform numbers complied with the SANS 241 (2011) standard of ≤10 CFU/100ml at most of the distribution systems, except for Hartswater during July 2011 (14CFU/100ml) and Warrenton during March 2012 (256 CFU/100ml). Heterotrophic plate count bacteria were very high in the household water of some of the distribution systems during 2011 and 2012 which exceeded the SANS 241 (2011) standard of ≤1000 CFU/ml. A large number of pigmented (yellow, orange, pink) and non-pigmented (white) colonies were isolated on R2A agar. This can be an indication of some failure in treatment processes. Other microbiological parameters that were tested such as faceal streptococci, Clostridia, Pseudomonas aeruginosa and fungi did not indicate any danger, but there were high levels of total anaerobic bacteria in the raw water during 2011 and 2012. A high level of anaerobic bacteria was detected in the household water of Hartswater during July 2011. Clostridia were also present in the household water of some of the distribution systems during 2011 and 2012. Sequencing results of the mdh, lacZ and uidA genes indicated that one of the isolates was identified as Enterobacter cloacae and the other isolates were E. coli. Four of the isolates were identified as Escherichia coli O104:H4. This is a pathogenic strain and raised concern. The physicochemical parameters that were measured complied with the SANS 241 (2011) standards during 2011 and 2012, but some of the parameters increased gradually from 2011 to 2012. Statistical analysis indicated that physico-chemical parameters had an influence on microbiological parameters and that deteriorating raw water may have an impact on drinking water quality. Another concern currently is that there is no SANS 241 (2011) for faecal streptococci, Clostridia, Pseudomonas aeruginosa, fungi and anaerobic bacteria. These are all opportunistic pathogenic bacteria and consuming water with high levels of these bacteria may cause health problems. This study indicated good progress in the treatment processes of the distribution systems over the two years. This may be due to the feedback given to Vaalharts Water during this study regarding the water quality of the residential areas. The physico-chemical and microbiological results of the present study indicated possible biofilm formation in the distribution systems. This may have impacts on the drinking water quality of the distribution systems. It was also evident that deteriorating raw water sources may have an impact on drinking water production. / Thesis (MSc (Environmental Sciences))--North-West University, Potchefstroom Campus, 2013.

Water quality in the Vaalharts region in the Northern Cape Province, South Africa, decreased over the past few years and there was a need for the microbiological and physico-chemical assessment. This problem was identified through discussions with Vaalharts Water (Vaalharts Water User Association) in 2010 when the issue of the impact of deteriorating water quality on drinking water production was raised. It was thus important to investigate concerns of the water users association pertaining to water quality issues. The aim of this study was to assess the physico-chemical and microbiological quality of household water in the Vaalharts irrigation scheme. The main residential areas were Hartswater, Pampierstad, Jan Kempdorp and Warrenton. Faecal coliforms were detected in the raw water of all the drinking water distribution systems during 2011 and 2012. No faecal coliforms were detected in the household water during 2011. This was a very positive result, because not only did the household water comply with the SANS 241 (2011) standard (0 CFU/100ml), but the purification processes were successful by removing all the E. coli’s from the raw water. However, during March 2012 faecal coliforms were detected in the household water of Jan Kempdorp (191CFU/100ml). This could be due to point pollution and possible breakage of faecal coliforms in the distribution system. Low amounts of total coliforms were detected in the raw water of some of the drinking water distribution systems. This could be due to high amounts of other colonies (pink and purple) growing on the m-Endo agar which suppress the growth of the metallic green sheen (total coliform) colonies. The total coliform numbers complied with the SANS 241 (2011) standard of ≤10 CFU/100ml at most of the distribution systems, except for Hartswater during July 2011 (14CFU/100ml) and Warrenton during March 2012 (256 CFU/100ml). Heterotrophic plate count bacteria were very high in the household water of some of the distribution systems during 2011 and 2012 which exceeded the SANS 241 (2011) standard of ≤1000 CFU/ml. A large number of pigmented (yellow, orange, pink) and non-pigmented (white) colonies were isolated on R2A agar. This can be an indication of some failure in treatment processes. Other microbiological parameters that were tested such as faceal streptococci, Clostridia, Pseudomonas aeruginosa and fungi did not indicate any danger, but there were high levels of total anaerobic bacteria in the raw water during 2011 and 2012. A high level of anaerobic bacteria was detected in the household water of Hartswater during July 2011. Clostridia were also present in the household water of some of the distribution systems during 2011 and 2012. Sequencing results of the mdh, lacZ and uidA genes indicated that one of the isolates was identified as Enterobacter cloacae and the other isolates were E. coli. Four of the isolates were identified as Escherichia coli O104:H4. This is a pathogenic strain and raised concern. The physicochemical parameters that were measured complied with the SANS 241 (2011) standards during 2011 and 2012, but some of the parameters increased gradually from 2011 to 2012. Statistical analysis indicated that physico-chemical parameters had an influence on microbiological parameters and that deteriorating raw water may have an impact on drinking water quality. Another concern currently is that there is no SANS 241 (2011) for faecal streptococci, Clostridia, Pseudomonas aeruginosa, fungi and anaerobic bacteria. These are all opportunistic pathogenic bacteria and consuming water with high levels of these bacteria may cause health problems. This study indicated good progress in the treatment processes of the distribution systems over the two years. This may be due to the feedback given to Vaalharts Water during this study regarding the water quality of the residential areas. The physico-chemical and microbiological results of the present study indicated possible biofilm formation in the distribution systems. This may have impacts on the drinking water quality of the distribution systems. It was also evident that deteriorating raw water sources may have an impact on drinking water production. / Thesis (MSc (Environmental Sciences))--North-West University, Potchefstroom Campus, 2013.

La filtration membranaire est une technique séparative utilisée fréquemment comme procédé permettant de retenir et d'extraire les microorganismes présents dans un fluide. Le mécanisme de sélectivité classiquement admis dans ce procédé est l'exclusion par la taille. Cependant, nos travaux ont permis de mettre en évidence un transfert de microorganismes à travers la structure poreuse des membranes au cours d'opérations de filtration, alors que les dimensions des cellules vivantes en suspension sont supérieures au diamètre moyen des pores de la membrane, entrainant une diminution du taux de récolte des microorganismes et une contamination de la phase perméat. Les caractéristiques morphologiques et nanomécaniques des cellules bactériennes sélectionnées pour nos travaux ont été observées et les propriétés du matériau membranaire modèle ont été mesurées. Grâce à ces informations, le transfert des bactéries aux premiers instants d'une filtration frontale a pu être évalué et le rôle de différents paramètres opératoires appliqués (type de souche bactérienne, composition physico-chimique de fluide filtré, PTM) a pu être commenté. Ainsi, nos résultats ont permis de préciser les mécanismes de sélectivités appliqués aux bactéries à Gram positif et à Gram négatif qui différent du fait des caractéristiques structurales de la paroi bactérienne (épaisseur et élasticité de la couche de peptidoglycane). Enfin, l'évolution du transfert de cellules vers la phase perméat a également été suivie et la mise en place du dépôt bactérien colmatant à la surface de la membrane a été observée. Le rôle de ce dépôt structuré sur les variations de débit et de transfert a donc pu être mis en évidence pour les trois modèles bactériens sélectionnés. Nos résultats ont permis de définir des conditions critiques (physiques, chimiques et biologiques) pour lesquelles le transfert de cellules bactériennes par déformation est amplifié au cours d'une filtration membranaire frontale. / Membrane filtration is a separation technique commonly used as a method for removing and extracting microorganisms present in a fluid. The selectivity mechanism is size exclusion. However, our work has highlighted transfer of microorganisms through the porous membrane structure during filtration operations, even if the size of the living cell in suspension is greater than the average pore size of the membrane, resulting in lowering of the accumulation rate of the microorganisms and thus contamination of the permeate phase. Morphological and nanomechanical characterization of selected bacteria cells used for our work were performed and the properties of the model membrane were analyzed. With this information, the transfer of bacteria dunring the first moments of the dead-end filtration has been evaluated and the role of operating parameters (type of bacteria strains, physicochemical composition of filtered fluid, TMP) has been studied. Thus, our results have clarified the selectivity mechanisms applied to Gram-positive and Gram-negative bacteria strains because of different structural characteristics of the bacterial cell wall (thickness and elasticity of the peptidoglycan layer). Finally, evolution of cells transfer to the permeate phase has also been followed, and the bacterial deposit clogging the membrane was observed. The role of the structured deposit on the variations in flow and transfer has been demonstrated for the three selected bacterial models. Our results have defined critical conditions (physical, chemical and biological) where the transfer of bacterial cells is amplified by deformation during a frontal membrane filtration.

Poly- och perfluorerade alkylsubstanser (PFAS) är syntetiskt tillverkade kemikalier som omfattar cirka 4700 individuella ämnen. PFAS har använts sedan 1950-talet i ett antal applikationer som brandsläckningsskum, impregneringsmedel samt textilier, på grund av sina unika egenskaper att vara både vattenlöslig (hydrofila) och vattenavstötande (hydrofoba). Dessa egenskaper gör att göra att PFAS kan transporteras långa vägar. Dessutom är ämnets kol-fluorbindning en av det starkaste kemiska bindningar som gör dem mycket stabila och därför svårnedbrytbara i miljön. Många PFAS misstänkts vara skadliga och kan ansamlas i både djur och människor. Ett antal PFAS har visat sig ge negativa effekter på reproduktion, immunsystem, sköldkörtelhormoner och levern vid djurförsök. Vissa tas lättare upp av kroppen men det är oklart vilken påverkan exakt alla 4700 PFAS-ämnen har. Dessa potentiellt skadliga föroreningar upptäcks på flera ställen i Sverige och har redan resulterat i allvarliga problem för dricksvattenförsörjningen. Dagens reningsmetoder avlägsnar inte PFAS och i kombination med toxiska egenskaper och hög bioackumuleringspotential kan det bli farligt för både människor och miljö. Därför regleras summan av 11 PFAS-ämnen (summa-11) med en åtgärdsgräns på 90 ng/L för vattenverk i Sverige. Vissa vattenverk har tvingats sättas ur bruk på grund av för höga PFAS-koncentrationer i producerat dricksvatten. Kunskapen om källor till PFAS är fortfarande låg men utredning av flera potentiella källor pågår.  I detta arbete har ett antal PFAS-ämnen analyserats uppströms Glan, den sjö i Norrköping kommun som försörjer över 90 % av invånarna med dricksvatten. Syftet är att kartlägga och karaktärisera PFAS i Norrköpings råvattentäkt Glan med tillflöden för att få en ökad förståelse för potentiell påverkan på dricksvattenproduktion. Kartläggningen skedde genom provtagningar uppströms Borgs råvattenintag. Det samlades också in data från andra aktörer längre uppströms i vattensystemet längs Motala Ström, Stångån samt även i Glan. Provtagning skedde också vid Bråvalla som är ett sedan tidigare välkänt förorenat område i direkt anslutning till Glan. All data analyserades sedan med ett Kruskal-Wallis test. Studien visar 1) av 21 analyserade PFAS-ämnen detekterades PFHxS, PFOS, PFPeA, PFHxA, PFHpA, PFOA och PFBA i flest provpunkter 2) Både PFOS och PFOA hade högst koncentration i Motala ström av de tillflöden till Glan som undersökts 3) PFBA, visade högst koncentration i andra tillflöden än Motala Ström, Hällestaån, Ysundaviken, till Glan med okänd källa 4) Av alla inflöden till Glan uppmättes högst halter av summa-11 i Motala ström som också utgör det största tillflödet till Glan 5) Våtmarken vid Bråvalla är en potentiell PFAS-källa till Glan då koncentrationen i dessa provpunkter är mycket höga i jämförelse med övriga provpunkter. Däremot indikerar inte resultaten på att Bråvalla är en källa i dagsläget men kan bidra till framtida komplikationer för Borgs råvattentäkt pga. klimatförändringar. / Per- and polyfluoroalkyl substances (PFAS) are synthetically manufactured chemicals comprising about 4700 individual substances. PFAS has been used since the 1950s in a number of applications such as firefighting foam, impregnating agents and textiles, due to its unique properties of being both water-soluble (hydrophilic) and water-repellent (hydrophobic). These properties allow PFAS to be transported long distances. In addition, the substance's carbon-fluorine bond is one of the strongest chemical bonds that makes them very stable and therefore difficult to degrade in the environment. Many PFASs are suspected to be harmful and can accumulate in both animals and humans. A number of PFASs have been shown to have adverse effects on reproduction, immune systems, thyroid hormones and the liver in animal experiments. Some are more easily absorbed by the body, but it is unclear what effect exactly all 4700 PFAS substances have. These potentially harmful pollutants are discovered in several places in Sweden and have already resulted in serious problems for the drinking water supply. Today's purification methods do not remove PFAS and in combination with toxic properties and high bioaccumulation potential, it can be dangerous for both humans and the environment. Therefore, the sum of 11 PFAS substances (summa-11) is regulated with a limit value of 90 ng/L for waterworks in Sweden. Some waterworks have been forced out of service due to too high PFAS concentrations in produced drinking water. Knowledge of sources for PFAS is still low, but investigation of several potential sources is ongoing. In this study, a number of PFAS substances have been analyzed upstream of Glan, the lake in Norrköping municipality that supplies more than 90% of the inhabitants with drinking water. The aim is to map and characterize PFAS in Norrköping's water catchment Glan with inflows to gain an increased understanding of the potential impact on drinking water production. The mapping took place through sampling upstream of Borg's raw water intake. Data were also collected from other actors further upstream in the water system along Motala Ström, Stångån and also in Glan. Sampling also took place at Bråvalla, which is a previously well-known polluted area in direct connection to Glan. All data were then analyzed with a Kruskal-Wallis test. The study shows 1) of 22 analyzed PFAS substances, PFHxS, PFOS, PFPeA, PFHxA, PFHpA, PFOA and PFBA were detected in most test points 2) Both PFOS and PFOA had the highest concentration in Motala stream of all the inflows examined to Glan 3) PFBA showed the highest concentration in inflows other than Motala stream, Hällestaån, Ysundaviken, to Glan with unknown source 4) Of all inflows to Glan, the highest levels of summa-11 were measured in Motala stream which also constitutes the largest inflow to Glan 5) The wetland at Bråvalla is a potential PFAS source of Glan as the concentration in these test points is very high compared to other test points. However, the results do not indicate that Bråvalla is a source at present, but may contribute to future complications for Borg's raw water source due to climate change.

Norrvatten är en av Sveriges största dricksvattenproducenter. Med ett ökande behov av en större produktionskapacitet och nya regleringar, planerar dem att förbättra sin reningsprocess. En aspekt som kräver omfattande undersökning är val av reningsmetod för att nå den nya gränsen från Livsmedelsverket för fyra PFAS kemikalier. PFAS är en grupp kemikalier som är mänskligt producerade sedan 1950-talet, vilka är persistenta och farliga för människors hälsa. Deras stabilitet i miljön leder till att de ackumulerar i naturen och hamnar i mat och dricksvatten. För att uppnå de nya regleringarna måste Norrvatten minska sin PFAS4 koncentration från cirka 5.7 ng/L till 4 ng/L. PFAS-separationen är planerad att genomföras via antingen aktivt kol eller en kombination där flotation även implementeras. Aktivt kol har en bevisad möjlighet att separera och destruera PFAS4, men nyligen har indikationer på en potential hos flotation uppenbarats. Dock är forskningen kring dess effektivitet begränsad vilket medför ett behov av att utvärdera dess användbarhet. Denna rapport syftar till att förse Norrvatten med ett underlag på effektiviteten och gynnsamheten med flotation inför beslutet att implementera flotation. Av den anledningen undersöktes separationseffektiviteten som uppnås i en befintlig flotationsbassäng i Norrvattens vattenverk. Genom att mäta PFAS4 koncentrationen i inflödet, utflödet och slammet kunde en borttagningseffektivitet på 19-38% bestämmas för olika belastningar. Slamkoncentrationen mättes till ungefär 200-700 ng/L. Felkällor kunde identifieras i mätningarna eftersom massbalansen inte stämde överens trots att mätosäkerheten från analysen adderades. Dessutom gjordes en omanalys av två prover som skilde sig från det förväntade resultatet, vilket gav avsevärda skillnader i resultatet. Trots att några felkällor kunde identifieras kunde inte en fullständig förklaring till avvikelserna fastställas. Trots detta kunde slutsatser dras att den uppmätta separationseffektiviteten och slamkoncentrationen kunde förse ett korrekt intervall. Dessutom indikerade resultaten att flotation inte kan appliceras enskilt för att nå koncentrationsmålet utan bör kombineras med kompletterande kolfilter f ̈or att nå gränsvärdet. Flotation ger upphov till en mer koncentrerad ström av PFAS som kräver behandling för att eliminera kemikalierna från miljöns kretslopp. Flera potentiella behandlingsmetoder identifierades för hur slammet kan hanteras. Jämförelsen inkluderade aktivt kolfilter, jonbytare, membran, oxidationsprocesser, sonolys, förbränning, deponi, jordrening, flotation, skumfraktionering, behandling på reningsverk och återanvändning för jordförbättring. Metoderna jämfördes utifrån bevisad robusthet, kostnad och applicerbarhet. Ett möjligt slamhanteringsalternativ visade sig vara ett sekundärt steg med flotation, förbränning av det uppkoncentrerade slammet och behandling av permeatet med jonbytare. Med den föreslagna slamhanteringen kunde implementering av flotation utvärderas baserat på maximal PFAS destruktion, ekonomi, koldioxidutsläpp och energiförbrukning. Det visade sig att energiförbrukningen är större för processen som inkluderar flotation men koldioxidutsläppen är mindre. Dock är det möjligt att avgränsningarna för koldioxid kan ha gynnat flotationsprocessen eftersom utsläpp från reaktivering av kolfilter inkluderades men inte utsläpp från förbränning eller energiproduktion. PFAS destruktionen blev ungefär 1-2% större för processen med exklusivt kolfilter eftersom flotationsprocessen innehåller ytterligare en ström som släpper ut PFAS efter behandling med jonbytare. Det mest ekonomiska alternativet visade sig bero på koncentrationsgränsen där implementering av flotation var dyrare för en PFAS4 koncentration på 4 ng/L i dricksvattnet, men billigare vid en gräns på 3 ng/L. Slutligen upptäcktes det att volymen av sekundärt slam som beräknades att skickas till förbränning har en stor påverkan på driftkostnaderna. Därför bör optimeringar genomföras för att minimera den sekundära slamvolymen om flotation implementeras. / Norrvatten is one of the largest drinking water producers in Sweden. With the need to increase their production capacity and new regulations, they are planning to improve their treatment process. One aspect that requires extensive investigation, is the treatment approach to reach the new limit from the Swedish Food Agency of four PFAS chemicals. PFAS is a group of chemicals that have been humanely produced since the 1950s, which are persistent and hazardous to human health. Their environmental stability causes them to accumulate in nature and ends up in foods and drinking water. To fulfill the new regulations, Norrvatten has to decrease the PFAS4 concentration in their drinking water from approximately 5.7 ng/L to 4 ng/L. The removal is planned to be achieved through either activated carbon filters or a combination where dissolved air flotation is included. Activated carbon has a proven separation and destruction possibility for PFAS4 but recently, DAF has appeared to be a promising alternative. However, there is limited research on its efficiency, which establishes a need for investigations on the potential of DAF for PFAS removal. To navigate Norrvatten through the decision between exclusively applying GAC filters or implementing a combination with DAF, this thesis intends to provide a basis of its efficiency and favorability. Therefore, this thesis investigated the removal efficiency that was obtained in one existing DAF basin in Norrvattens waterworks. By measuring the PFAS4 concentration in the inlet, outlet, and sludge flow, a removal efficiency of 19-38% was acquired for different flow levels. In addition, the sludge concentration was measured to approximately 200-700 ng/L in the water phase. Presence of sources of errors could be identified in the measurements as the mass balance did not add up despite the addition of the measurement uncertainty in the analysis. Also, two samples were analyzed twice as the results differed from the expected results, which provided significant variations. Although some sources of error were identified, a full explanation for the differing results could not be disclosed. However, through some consistency, it was possible to conclude that the removal efficiency and sludge concentration could provide an accurate interval to illustrate the reality. In addition, the results indicate that DAF cannot be applied alone to reach the concentration goal, but has to be combined with additional GAC filters to reach the target. DAF provides a separation of PFAS into a sludge stream that requires treatment to remove the chemicals from the environmental cycle. To investigate how the sludge could be managed, several treatment methods were identified. The comparison included GAC, AIX, membranes, oxidation processes, sonolysis, incineration, landfill, soil cleaning, DAF, foam fractionation, treatment at a wastewater facility, and reuse for soil improvement. The methods were compared based on proven robustness, cost, and applicability. It was concluded that a viable alternative was to transfer the sludge to a second stage of DAF, incinerate the concentrated sludge, and treat the permeate with AIX. From the sludge management proposal, the implementation of DAF was analyzed based on maximum PFAS destruction, economics, carbon emissions, and energy requirements. It was found that the energy requirement was larger for the process including DAF but the carbon emissions are smaller. However, the delimitations on the carbon emissions may have been in favor of the DAF process as the GAC reactivation emissions are included but emissions due to incineration or energy production are excluded. The PFAS destruction was approximately 1-2% larger for the process that excluded DAF as an additional stream containing PFAS would be released into the environment after AIX, which is not present when only GAC is implemented. The economic out-come depended on the concentration limit, where implementation of DAF was more expensive for a treatment target of 4 ng/L in the drinking water and more economical for 3 ng/L. Lastly, it was found that the volume of the concentrated sludge that would be sent for incineration had a large impact on the operational cost. Therefore, if DAF is implemented optimizations are of relevance to minimize the volume of that stream.

Browning of boreal freshwaters due to an increased export of terrestrial dissolved organic carbon (DOC) has been observed for some decennia. Drivers include recovery from acid deposition and changing climate and land cover. Lake Mälaren provides the Swedish capital Stockholm with drinking water and an increased future browning of the lake could demand more treatment to produce acceptable drinking water. Knowledge of what can be expected in a changing climate is therefore needed. The aim of this study was to evaluate; (1) The performance of the GWLF model simulations of hydrology and DOC export for 13 catchments draining into the Galten and Ekoln basins in lake Mälaren; (2) How simulated DOC export from the 13 catchments would change in response to climate scenarios RCP2.6, 6.0 and 8.5; and (3) If climate change impact on DOC export to lake Mälaren potentially will demand more treatment of the lake water in the future to produce acceptable drinking water to the Stockholm area. All data used were open access and include temperature, precipitation, discharge, total organic carbon (TOC) and water colour, land cover and soil types. In comparison to other studies, good model performance was found when simulating daily streamflow, baseflow, surface runoff, daily DOC loads and monthly DOC loads. Model simulations of DOC concentrations were less certain, but for some catchments similar results were obtained as in other studies. Increased air temperature resulted in higher simulated soil temperatures and a longer growing season. An increase in both annual precipitation and evapotranspiration resulted in only slight increase in simulated annual streamflow. There were, however, large seasonal impacts on streamflow with higher winter flows and lower spring flood. Annual DOC loads increased, mainly due to increased DOC concentrations over the whole year, as the annual streamflow did not change greatly. Increased winter streamflow and DOC concentrations resulted in large increases in DOC loads. The impact was larger for the higher emission scenarios. DOC loading to Galten and Ekoln increased in all future scenarios during winter, with a likely impact on DOC concentrations in the eastern basins as well. Increased loads to Ekoln are likely to impact the drinking water production of Stockholm more direct than the increases in Galten, as the Ekoln water has a shorter transit time to the drinking water withdrawals. The GWLF model shows much promise in predicting DOC concentrations and loads to lake Mälaren in a changing climate. To estimate the effects of DOC export to lake Mälaren on future drinking water production, further work also needs to be done on in-lake transport and processes. / Sjöar och vattendrag på nordliga breddgrader har under flera decennier blivit allt brunare. Orsaken är att löst organiskt kol, Dissolved Organic Carbon (DOC), som bland annat utgörs av humusämnen, transporteras till ytvatten från omkringliggande mark. De vanligaste förklaringarna till detta är en återhämtning från tidigare försurning orsakad av sulfatdeposition, klimatförändringar och en ändrad markanvändning. Mälaren är den primära dricksvattenkällan för Stockholm och om vattnet fortsätter bli brunare kan det innebära att reningsprocessen behöver ändras för att undvika produktion av hälsofarliga biprodukter i dricksvattnet. Det är därför viktigt att ta reda på hur ett ändrat klimat kan påverka export av DOC till Mälaren. Målet med den här studien var att undersöka tre frågor kopplade till framtida DOC i Mälaren. Den första var hur väl GWLF-modellen kunde användas på 13 avrinningsområden i Mälaren. Modellen simulerar vattnets väg från nederbörd till vattendrag samt hur DOC bryts ned och transporteras ut från marken. Den andra frågan var hur simulerad nedbrytning och export av DOC skulle påverkas av tre olika klimatscenarion, Representative Concentration Pathways (RCP), som baseras på tre olika nivåer av utsläpp av växthusgaser. Dessa var RCP2.6, 6.0 och 8.5, där högre nivåer av utsläpp av växthusgaser är kopplade till en högre siffra. Den tredje och sista frågan var att utvärdera om klimatförändringar kommer påverka hur mycket DOC som når Mälaren och ifall detta kan väntas påverka framtida dricksvattenproduktion i området. All data som har använts för att genomföra studien har varit så kallad “Öppna data”, som är fritt för alla att ladda ned, från både svenska, europeiska och andra internationella databaser. Data som använts är temperatur, nederbörd, vattenflöde, vattenkemi, vegetationstäcke och jordarter.  Modellen visade ett tillfredsställande resultat i att simulera dagligt vattenflöde och både daglig och månatlig DOC-export. Den var mer osäker på att simulera koncentrationerna av DOC, men i flera områden lyckades den prestera liknande som i många andra studier. Alla framtida klimatscenarier visade en ökning av nederbörd, luft- och marktemperatur, avdunstning och antal växtdagar. Det syntes en större påverkan vid högre utsläppscenarion och effekterna ökade för RCP6.0 och 8.5 mot slutet av detta sekel. För RCP2.6 kunde en återhämtning ses mot slutet av seklet på grund av en minskning av utsläpp av växthusgaser. Den högre årsnederbörden resulterade inte i motsvarande ökning i årsvattenflöde på grund av att även avdunstningen ökade. Simuleringen visade en säsongsbunden effekt på DOC-export med en ökning under vintermånaderna men ingen skillnad under sommaren. Koncentrationer av DOC ökade däremot under hela året, dock mer under vintern. Årsexport av DOC från de undersökta avrinningsområdena till de två bassänger i Mälaren som mottar vattnet, Galten och Ekoln, ökade för alla framtida klimatscenarion. Ökningen var ett resultat av ökad vinterexport då det under sommarmånaderna antingen inte fanns någon skillnad eller en liten minskning av DOC-export. Detta påverkar sannolikt även de östra bassängerna där Stockholm tar sitt dricksvatten. Ökad DOC-export till Ekoln skulle troligen påverka Stockholms dricksvattenproduktion mer eftersom bassängerna har ett direkt vattenutbyte. GWLF-modellen som användes här visade potential i att simulera DOC-export till Mälaren i ett ändrat klimat. För att uppskatta framtida påverkan av DOC-export på dricksvattenproduktionen kring Mälaren behöver även processer som sker i sjön undersökas.

A tecnologia de tratamento com dupla filtração encontra-se em estágio relativamente elevado de desenvolvimento, porém, ainda são poucos os dados de operação em escala real. Este trabalho teve como objetivo avaliar um sistema de dupla filtração em escala real (ETA2) para tratamento de água, do serviço autônomo de água e esgoto da cidade de São Carlos - SP, que trata água bruta proveniente do Ribeirão do Feijão e é constituída de três unidades de filtração cada uma contendo um filtro ascendente e um descendente. Em uma primeira etapa foram levantadas as medidas das unidades de dupla filtração, caracterizado o material filtrante e avaliadas as condições operacionais da ETA 2. Posteriormente, foi estudada a situação geral da ETA por meio de diversos parâmetros (turbidez, pH, expansão da areia na lavagem, contagem de partículas, de coliformes totais e Escherichia. coli), para verificar os principais problemas. Em seguida, realizou-se uma comparação da ETA2 com uma Instalação Piloto de Dupla Filtração. Este trabalho permitiu concluir principalmente que: i) devido às baixas taxas de filtração de unidades filtrantes da ETA2, excessivas dosagens de coagulante e falta de controle de lavagens, geralmente foi baixa a produção efetiva de água (em torno de 90%); ii) o desempenho, algumas vezes insatisfatório das unidades de filtração ascendente e descendente, deveu-se principalmente à ineficiência da mistura de coagulante na água bruta, falta de areia nos filtros ascendentes e à diferença de granulometria nos filtros descendentes; iii) o desenvolvimento da perda de carga nos filtros ascendentes da ETA2 foi mais acentuado na Instalação Piloto devido à mistura de seus materiais filtrantes. / Water treatment by means of a double filtration system is in a relatively high stage of development, however, there are still few data on real scale operations. The aim of this work was to evaluate a double filtration system in a real scale (ETA2) for the treatment of water, of the autonomous service of water and sewer of the city of São Carlos - SP, which treats water from the Ribeirão do Feijão (name of a local river), and is constituted of three filtration units, each a contain a direct up-flow sand filter, and a rapid filter. In the first stage, the sand of the filters was characterized, and the operational conditions of the ETA2 were evaluated. Next, the general situation of the ETA was studied by means of several parameters (turbidity, pH, expansion of the sand during washing process, counting of particles, total coliforms and Escherichia coli), in a order to detect the main problems. Finally, a confrontation between the ETA2 an a double filtration pilot Plant was carried through. This study made infer that: I) due to the low filtration rates of the ETA2`s filtration units, excessive doses of coagulant and lack of control of the washings, the actual production of water was low (around 90%); ii) the sometimes insufficient performance of the direct up-flow sand filtration and rapid filtration systems was mainly caused by an inefficient mixing of the coagulant with the natural water, lack of sand in direct up-flow filters; iii) the building up of the lossof cargo in the direct up-flow san filtration in filters of the ETA2 was more accentuated than in the Pilot Plant due to the mixture of the filtration materials.

Em instalação piloto, de 06/1998 a 10/1999, utilizando-se água natural de superfície com baixas turbidez e cor, foram avaliados dois sistemas de dupla filtração, com pré-tratamento em filtração direta ascendente em areia com taxas de 200 a 360 m3/m2.dia (sistema S01) e em filtração direta ascendente em pedregulho com taxas de 80 a 190 m3/m2.dia (sistema S02) e com tratamento final em filtração rápida descendente com taxas de 115 a 480 m3/m2.dia. As dosagens do coagulante (produto comercial de sulfato de alumínio) foram definidas utilizando-se filtro de laboratório de areia, havendo indicações de coagulação no mecanismo de adsorção, com neutralização parcial de carga. Foram executadas descargas de fundo intermediárias nas unidades de filtração direta ascendente, com significativa redução na perda de carga no meio granular de pedregulho. A turbidez, cor aparente, Fe e Mn totais dos efluentes finais dos sistemas atenderam às Normas Brasileiras para produção de água potável. No sistema S02 o crescimento da perda de carga no meio granular foi menos intenso na unidade de filtração direta ascendente do que no filtro rápido descendente, ocorrendo o inverso no sistema S01, no qual, em alguns ensaios, o filtro rápido descendente mostrou-se desnecessário. As carreiras de filtração foram de 27 a 88 h no sistema S01 e de 14 a 35 h no sistema S02. Em determinados casos, a floculação predominou na unidade de filtração direta ascendente em pedregulho, prejudicando a qualidade do seu efluente. O sistema S02 mostrou-se mais vantajoso que o sistema S01, exceto quando, neste, a filtração rápida descendente foi desnecessária. / Based on me investigation carried out in pilot plant, from June 1998 to October 1999, using natural surface raw water with low turbidity and color, two systems of two-stage filtration have been evaluated. As a pretreatment, the first system (S01) used direct up-flow sand filtration (with filtration rates from 200 to 360 m3/m2.day), and the second system (S02) used direct up-flow roughing filtration (with filtration rates from 80 to 190 m3/m2.day), both with final treatment based on rapid filtration (with filtration rates from 115 to 480 m3/m2.day) The coagulant doses (commercial product of aluminum sulfate) have been established through the use of laboratory scale sand filter, presenting signs of coagulation in the adsorption mechanism, with partial charge neutralization. Intermediate bottom discharges in the direct up-flow filtration units have resulted in significant reduction in thee headloss in gravel granular medium. The turbidity, apparent-color, total iron and manganese concentrations of the final effluents of the systems have attended the Brazilian Standards for drinking water production. The increase of headloss in gravel granular medium in system S02 was smaller in me direct up-fiow filtration unit than in the rapid filter, the opposite occurring in system S01, where some tests have proved lhe rapid filter redundant. The filter runs time were 27 to 88 hours in system S01 and 14 to 35 hours in system S02. In certain cases, flocculation was predominant in the direct up-flow roughing filtration unit, jeopardizing the effiuent quality. System S02 has proved advantageous in relation to system S01, except when in system S01 the rapid filter was redundant.