Thesis (MSc)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: Domestic rainwater harvesting (DRWH) refers to the collection and storage of rainwater for domestic purposes and in an effort to achieve the Millennium Development Goals (MGD), the South African government has started an initiative where DRWH tanks are financed in sustainable housing schemes in an aim to provide an additional water source directly to households. Although many provinces, including parts of the Eastern Cape and KwaZulu Natal, have been using harvested rainwater as a potable water source, there are a limited number of studies indicating the quality of harvested rainwater in South Africa. However, many studies, internationally, have indicated that while the practice of harvesting rainwater is gaining popularity, rainwater quality is not within potable standards (Chapter 1). During the first phase of the study, rainwater samples were collected from the Kleinmond Housing Scheme (Western Cape, South Africa). From a cluster of 411 houses, the DRWH tanks connected to 29 houses were selected for monitoring the microbial and physico-chemical properties of harvested rainwater. Drinking water guidelines stipulated by SANS 241 (2005), DWAF (1996), ADWG (NHMRC and NRMMC, 2011) and WHO (2011) were used throughout the study to monitor the quality of rainwater. Eight sampling sessions were then conducted from March to August 2012, during a high and low rainfall period. Overall, the physico-chemical parameters of the rainwater samples were within the respective drinking water guidelines. However, the microbiological analysis verified results obtained in international studies, and showed that the indicator bacteria numbers present in the DRWH samples exceeded the stipulated guidelines (Chapter 2 and 3). Species specific primers were also used to routinely screen for the virulent genes, aggR, stx, eae and ipaH found in Enteroaggregative E. coli (EAEC), Enterohaemorrhagic E. coli (EHEC), Enteropathogenic E. coli (EPEC) and Enteroinvasive E. coli (EIEC), respectively, in the rainwater samples. The virulent pathogenic E. coli genes were then detected in 3% (EPEC and EHEC) and 16% (EAEC) of the 80 rainwater samples collected routinely during the sampling period from ten DRWH tanks (Chapter 3). Bacterial isolates selected during the high rainfall period (June to August 2012), as well as PCR assays performed on total genomic DNA extraction from the rainwater samples, confirmed the presence of numerous pathogenic bacteria including Legionella spp. Klebsiella spp. and Shigella spp. Yersinia spp. were also isolated and detected for the first time in DRWH tanks (Chapter 4). Based on the results obtained in the first phase and as many studies have indicated the poor quality of rainwater, the second phase of the project was aimed at designing and monitoring point of use treatment systems. Three polyethylene DRWH tanks (2000 L) were installed at the Welgevallen Experimental farm, Stellenbosch University, South Africa. Various treatment systems, such as activated carbon and slow sand filtration, solar pasteurization and a combined activated carbon/PVA nanofibre filtration column, were then intermittently connected to the three DRWH tanks during the high rainfall period (June to October 2013). Results for slow sand filtration and activated carbon filters indicated that the biological layer that had developed on the filtration media had not matured and for this reason chemical and microbial parameters were not reduced to within drinking water guidelines. A polyvinyl (alcohol) (PVA) nanofibre membrane without activated carbon in a column filtration system was analysed and results indicated that this system was also not effective in reducing the microbial numbers to within drinking water guidelines. Lastly, by utilising a PVA nanofibre membrane with activated carbon in a column filtration system, one litre of potable water was produced and all heterotrophic bacteria, E. coli and total coliform counts were reduced to zero and were within drinking water guidelines (Chapter 5). For the solar pasteurization system (Chapter 6), at treatment temperatures of greater than 72°C, all heterotrophic bacteria, E. coli and total coliforms were reduced to zero and were within drinking water guidelines. However, PCR assays confirmed the presence of Yersinia spp., Legionella spp., and Pseudomonas spp., at temperatures greater than 72°C. Results for chemical analysis also indicated all cations were within the international and national drinking water guidelines, with the exception of iron, aluminium, lead and nickel, which were detected in the pasteurized rainwater samples and were above the respective guidelines. It is hypothesised that these elements could have leached from the stainless steel storage tanks of the pasteurization system and it is therefore recommended that the storage tank of the pasteurization system be manufactured from an alternative material, such as a high grade polymeric material, which is able to withstand the high temperatures yet will not negatively influence the quality of harvested rainwater. / AFRIKAANSE OPSOMMING: Reënwater versameling vir huishoudelike gebruik verwys na die versameling en berging van reënwater vir huishoudelike doeleindes. In 'n poging om die Millennium Ontwikkelingsdoelwitte (MOD) te bereik het die Suid-Afrikaanse regering ‘n inisiatief begin om finansiële bystand aan huishoudings te verleen vir die implementering van reënwater tenke, in ‘n poging om ‘n addisionele water bron direk aan huishoudings te verskaf. Hoewel baie provinsies, insluitend dele van die Oos-Kaap en KwaZulu-Natal, reënwater gebruik as ‘n drinkbare water bron, is daar 'n beperkte aantal studies beskikbaar oor die gehalte van versamelde reënwater in Suid-Afrika. Baie studies internasionaal het egter al aangedui dat, alhoewel die praktyk van die versameling van reënwater besig is om in gewildheid toe te neem, die kwaliteit van reënwater nie binne drinkbare standaarde is nie (Hoofstuk 1). Tydens die eerste fase van die studie is reënwater monsters geneem van die Kleinmond Behuisings Skema (Wes Kaap, Suid Afrika). Van ‘n kompleks van 411 huise, is reënwater tenke van 29 huise geselekteer vir die monitering van die mikrobiese en fisiese-chemiese eienskappe van versamelde reënwater. Drinkwater riglyne soos gestipuleer deur SANS 241 (2005), DWAF (1996), AWDG (NHMRC en NRMMC, 2011) en WHO (2011) was regdeur die studie gebruik om die kwaliteit van reënwater te monitor. Reënwater monsters is tydens agt geleenthede geneem vanaf Maart tot Augustus 2012, tydens die hoë en lae rëenval periode. Oor die algemeen was die fisiese-chemiese parameters van die reënwater monsters binne die onderskeie riglyne vir drinkwater. Mikrobiese analises het egter die resultate van internasionale studies bevestig en het aangedui dat die getal indikator bakterieë teenwoordig in the reënwater die gestipuleerde riglyne oorskry (Hoofstuk 2 en 3). Spesies spesifieke inleiers was gebruik om die virulensie gene aggR, stx, eae en ipaH, teenwoordig in onderskeidelik Entero-aggregatiewe E. coli (EAEC), Entero- hemorragiese E. coli (EHEC), Entero-patogeniese E. coli (EPEC) and Entero-indringende E. coli (EIEC), in die reënwater monsters op te spoor. Die virulente pathogeniese E. coli gene was in 3% (EPEC en EHEC) en 16% (EAEC) van die 80 reënwater monsters, geneem van 10 reënwater tenke tydens die studie, waargeneem (Hoofstuk 3). Bakteriese isolate geselekteer tydens die hoë reenval periode (Junie tot Augustus 2012), so wel as PKR (polymerase ketting reaksie) ontledings uitgevoer op genomiese DNS wat van die reënwater monsters ge-ekstraheer was, het die teenwoordingheid van verskeie patogeniese bakterieë, insluitend Legionella spp., Klebsiella spp. en Shigella spp., bevestig. Dit is ook die eerste keer wat Yersinia spp. in versamelde reënwater waargeneem is (Hoofstuk 4). Die doel van die tweede fase van die projek was om water behandeling sisteme by die punt van gebuik te ontwerp en te monitor aangesien die resultate wat gedurende die eerste fase verkry is, sowel as vele ander studies, aangedui het dat die kwaliteit van versamelde reënwater swak is. Drie poli-etileen reënwater tenke (2000 L) was geïnstalleer op die Welgevallen Eksperimentele plaas van die Universiteit van Stellenbosch, Suid Afrika. Verskeie water behandeling sisteme, soos geaktiveerde koolstof en stadige sand filters, sonkrag pasteurisasie en ‘n gekombineerde geaktiveerde koolstof/PVA nanovesel filtrasie kolom, was met tye gekonnekteer aan verskeie reënwater tenke gedurende die hoë reënval periode (Junie to Oktober 2013). Resultate van die stadige sand en geaktiveerde koolstof filters het aangedui die biologiese laag nog nie ver genoeg ontwikkeld het nie en daarom was die chemiese en mikrobiese parameters nie verminder to binne die riglyne vir drinkwater nie. ‘n Polyviniel (alkohol) nanovesel membraan sonder geaktiveerde koolstof binne ‘n kolom filtrasie sisteem, was geanaliseer en resultate het aangedui dat hierdie sisteem ook nie die mikrobiese getalle verminder het tot drinkwater standaarde nie. Laastens, deur gebruik te maak van ‘n PVA nanovesel membraan met geaktiveerde koolstof in ‘n kolom filtrasie sisteem, was een liter drinkbare water geproduseer met alle heterotrofiese bakterieë, E. coli en totale koliform getalle verminder tot nul en was binne die riglyne vir drinkwater (Hoofstuk 5). Vir die sonkrag pasteurisasie sisteem (Hoofstuk 6), by behandelings temperature van groter as 72°C, was alle heterotrofiese bakterieë, E. coli en totale koliform getalle verminder tot nul en was binne die riglyne vir drinkwater. Maar PKR ontledings het die teenwoordigheid van Yersinia spp., Legionella spp., en Pseudomonas spp. by temperature groter as 72°C bevestig. Resultate van die chemiese analises het ook aangedui dat alle katione binne die drinkwater riglyne was, met die uitsondering van yster, aluminium, lood en nikkel. Hierdie elemente was in die gepasteuriseerde reënwater monsters waargeneem met konsentrasies wat die onderskeie riglyne oorskry het. Dit word gehipoteseer dat hierdie elemente uit die vlekvrye staal van die bergings tenk van die pasteurisasie sisteem kon uitgeloog het. Daarom word dit aanbeveel dat die bergings tenk van die pasteurisasie sisteem geproduseer moet word van ‘n alternatiewe materiaal, soos ‘n hoë graad polimeriese material, wat hoë temperature kan weerstaan maar nie ‘n negatiewe invloed sal hê op die kwaliteit van versamelde reënwater nie.

Activated carbon (AC) systems and reverse osmosis (RO) systems are commonly used point-of-use (POU) water filtration systems as the last barrier to remove trace-level contaminants in tap water to protect human health. Limited studies have been done to evaluate trace-level manganese and uranium removal in tap water. Additionally, undesired microbial growth in POU systems may reduce treatment efficiencies of POU systems and limited studies have been done to evaluate microbial growth potential in POU systems. The overall research objective of this study was to systematically evaluate the removal of metals and assimilable organic carbon in POU systems. AC systems were operated to 200% of their designed treatment capacities and RO systems were operated for three weeks. The results indicated that AC systems were generally ineffective to remove metals in drinking water, while metals were effectively removed in RO systems. The results showed that calcium and magnesium were not effectively removed in AC systems with removal efficiencies of less than 1%. Various factions of iron were removed with its removal efficiencies in AC systems ranged between 61% and 84%. Copper was effectively removed in AC systems with removal efficiencies greater than 95%, which was possibly related to its low influent concentration in drinking water (<30 μg/L). Both manganese and uranium were ineffectively removed from AC systems. Different from AC systems, RO systems were consistently effective to remove all metals. Calcium, magnesium, iron, and copper were all removed with removal efficiencies greater than 98%, while removal efficiencies of manganese and uranium in RO systems were above 95%. Assimilable organic carbon was effectively removed from all AC and RO systems and high variability of AOC removal efficiencies were observed, which may be attributed to the heterogenicity of biofilm and microbial growth in POU systems. The new knowledge generated from this study can help improve our understanding of emerging contaminant removal in POU systems and develop better strategies for the design and operation of POU systems to remove emerging contaminants in drinking water and mitigate their health risks to humans.

In the rural areas of Madagascar only 29% of the population has access to clean water and 10% has access to improved sanitation. It has been estimated that environmental risk factors, such as inadequate access to clean drinking water and proper sanitation, are responsible for 94% of the diarrheal disease burden. This study was focused on testing a point-of-use technology called the Tulip filter, which is a siphon ceramic candle filter impregnated with silver. The purpose is to assess its feasibility for implementation in rural regions of Madagascar through a laboratory study performed at the University of South Florida. The study tested the Tulip filters for turbidity, total coliform, and E. coli removal for various types of water. Each filter processed synthetic water classified as control (tap water, <1 >NTU), low turbidity (5-7 NTU), medium turbidity (25-35 NTU), or high turbidity (60-80 NTU). Approximately once every 100 L the filters processed pond water to test coliform removal. Furthermore, the pthesis size distribution was measured to analyze the effectiveness of filter to remove various pthesis sizes. Two of the seven Tulip filters tested had some quality control issues with the glue connecting the ceramic candle to the plastic cap and failed at 350 L. Of the functioning filters, the turbidity removal ranged from 93% to 98% with none of the 779 samples taken from 4 filters above the WHO recommended 5 NTU for drinking water. The log removal of total coliforms was about 3.90 to 4.16 and achieved an average of 1 CFU/100mL of E. coli in the filtered water. WHO guidelines consider water with 1-10 CFU/100 mL a "low risk" and all but one of the working filters had E. coli and total coliform concentrations within, or below, this range for all samples (n=20 for each filter). The filters also showed an average of 96% removal of pthesiss of all size ranging from 0.5 to 10 µm. This study finds that the Tulip filter is an appropriate of point-of-use technology that enables rural areas access to "low risk" water, at a low cost and with minimal maintenance. This study also reinforces the importance of adding silver or another biocide to ceramic filters because pthesiss of sizes up to 10 µm are able to pass through the filter. This is particularly a problem because pathogens can range from 0.01 µm to 100 µm.

Currently 884 million people worldwide are living without access to an improved source of drinking water (WHO/UNICEF, 2011). Piped-water on premises is the ultimate goal of World Health Organization (WHO) due to the ability to treat all of the water and distribute it safely in pressurized pipes. However, Household Water Treatment and Safe Storage (HWTS) is an option for improving the quality of drinking water where that infrastructure is not yet developed, especially where there is a risk of recontamination between point of collection and point of use (Clasen, 2006). This study analyzed one such HWTS, the ceramic pot water filter. The study compared the hydraulic properties of the FilterPure (FP) and Potters for Peace (PFP) ceramic pot filters through a thirteen-month field study in the Dominican Republic and laboratory studies at the University of South Florida. In the field study 55 filters were tested for first hour flow rate and hydraulic conductivity. Eight first hour flow rate tests were conducted in the field on one month intervals during months 7- 13. FP filters had an average first hour flow rate of 553 ml/hr and PFP Filters had a first hour flow rate of 395 ml/hr. No significant change in first hour flow rate was observed over time in FP filters. PFP experienced an average increase of 31 ml/hr per month during the seven-month testing period. Falling head tests were conducted on four filters in the laboratory and the flow rate was modeled to determine hydraulic conductivity. Hydraulic conductivity values for FP filters ranged from k = 0.0495 - 0.0831 cm/hr and for PFP filters ranged from k = 0.0136 - 0.0389 cm/hr. Eight out of 29 (26%) Potters for Peace filters in the field had first hour flow rates of less than 250 ml/hr by month nine of the study and had to be replaced and removed from the study. In total 24 of 55 (44%) filters (8 FP and 16 PFP) had to be removed from the study due to several reasons discussed in this thesis.