Thesis (MSc Food Sc)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: The increased consumption of fresh produce has been shown to be related to increases in foodborne disease outbreaks and these have in many cases been ascribed directly to carry-over of pathogens from contaminated irrigation water. In South Africa, rivers are the main source of irrigation water but many have been found to be unsuitable for irrigation of fresh produce because of the unacceptably high levels of faecal contamination. The main aim of this study was to do a baseline evaluation of the microbiological quality of the Plankenburg and Eerste Rivers and to determine which bacterial contaminants are present. Two sampling sites were selected for the Plankenburg (Plank-1 and -3) and one for the Eerste River (Eerste-1). The microbiological analysis included aerobic colony count (ACC), aerobic and anaerobic sporeformers, Staphylococcus, Salmonella, Listeria, enterococci, coliforms, faecal coliforms and E. coli using standard methods. The faecal contamination levels for both rivers exceeded the DWAF and WHO guidelines of <1 000 E. coli per 100 mL water for irrigation of fresh produce intended to be consumed raw. The Plankenburg River sites always had higher coliform contamination levels (1 200 - 13 000 000 MPN per 100 mL water) than the Eerste River site (230 - 79 000 MPN per 100 mL water). There was also a high incidence of index organisms including Salmonella, Staphylococcus, Listeria and endosporeformers. The isolation of intestinal enterococci suggested the presence of potential pathogens that can cause disease outbreaks. The baseline data also showed large variations in microbial loads over the 15 month study with the faecal coliform counts ranging for Plank-1 from 1 200 to 7 000 000 MPN.100mL-1, Plank-3 from 10 to 460 000 MPN.100mL-1 and Eerste-1 from 28 to 79 000 MPN.100mL-1. The water temperatures at all three sites ranged from 12.1° to 21.7°C with COD values in most cases below 100 mg.L-1. As the baseline study showed large variations in microbial loads over the 15 month study period an assessment using the Colilert-18 system of the weekly, daily and hourly variations, for 6 weeks over a period of 4 months was conducted at site Plank-2. This site was specifically used as it is an irrigation source point for nearby fresh produce farmers and is about 2 km further downstream from an informal settlement. The weekly variation trend for total coliforms (TC) showed a decrease over the entire sampling period with the highest count of 3 200 000 MPN.100 mL-1 during the warmer period. The E.coli (Ec) counts showed a similar trend with the highest count of 440 000 MPN.100 mL-1 also in March. The daily variation trends were the same for both the TC and Ec and counts found to increase from Monday to Thursday followed by a decrease to Sunday. The highest counts were on Thursday with average TC and Ec counts of 1 900 000 and 160 000 MPN.100 mL-1, respectively. The hourly variation trends were similar for both TC and Ec with counts increasing from 06h00 to 12h00 followed by a decrease to 18h00. The increases in TC and Ec counts found during the weekly, daily and hourly variation trend studies clearly suggests that the 15 month sampling that was done once a month on Mondays at 08h00 could be considered an underestimation of the contamination levels of the Plankenburg and Eerste Rivers. The overall weekly variation trend for the water temperature showed a decrease over the sampling period while the daily and hourly variation trends showed an increase from 06h00 to 18h00. The overall weekly trend for pH differed from that of the temperature with an increase over the sampling period. The analysis of covariance showed no correlation (p < 0.05) between the physico-chemical (temperature and pH) and the microbial variables (TC and Ec). Therefore it was concluded that temperature and pH had no direct impact on either the total coliform or E. coli counts. Both the Plankenburg and Eerste Rivers were found to be unsuitable for the irrigation of fresh produce intended to be consumed raw due to the high levels of faecal contamination that exceeded DWAF and WHO guidelines. Irrigation with such water could pose a health risk because of presence of potential pathogens that could be carried-over to fresh produce. / AFRIKAANSE OPSOMMING: Die toenemende gebruik van vars produkte hou direk verband met die toename in voedseloordraagbare siektes. Alte dikwels kan dit toegeskryf word aan die teenwoordigheid van patogene in besproeiingswater. In Suid Afrika is riviere die hoofbron van besproeiingswater maar dit is al gevind dat meeste ongeskik is vir gebruik as besproeïngsbron as gevolg van die onaanvaarbare hoe vlakke van fekale besmetting. Die hoofdoel van hierdie studie was om ‘n basislyn evaluasie van die mikrobiologiese kwaliteit van die Plankenburg en Eerste Riviere te doen en ook vas te stel watter bakteriese kontaminante teenwoordig is. Twee bemonsteringpunte is geselekteer vir die Plankenburg (Plank- 1 en -3) en een vir die Eerste Rivier (Eerste-1). Mikrobiologiese analises met standaard metodes het die volgende ingesluit: aërobe kolonie telings (AKT), aërobe en anaërobe spoorevormers, Staphylococcus, Salmonella, Listeria, enterococci, koliforms, fekale koliforms en E. coli met gebruik van standaard metode. Die fekale besmettingsvlakke vir beide riviere het die DWAF en WHO leistreep van <1 000 E. coli per 100 mL water vir besproeiing van vars produkte wat rou geëet kan word oorskry. Die Plankenburg Rivier bemonsteringspunte het in alle gevalle ‘n hoër kolivorm besmettingsvlak (1 200 - 13 000 000 MPN per 100 mL water) as die Eerste Rivier punt (230 - 79 000 MPN per 100 mL water) gehad. Daar was ook ‘n hoër voorkoms van indeksorganismes insluitend Salmonella, Staphylococcus, Listeria en endosporevormers. Die voorkoms van ingewand enterococci was ‘n addisionele aanduiding van die voorkoms van patogene wat ernstige gesondheidsrisikos vir die verbruiker kan inhou. Die basislyn data het groot variasies in die mikrobe vlakke oor die 15 maand van studie getoon. Die faecal koliforms vir Plank- 1 het gewissel van 1 200 tot 7 000 000 MPN.100mL-1, vir Plank-3 van 10 tot 460 000 MPN.100mL-1 en vir Eerste-1 van 28 tot 79 000 MPN.100mL-1. Die water temperature het gewissel van 12.1° tot 21.7°C met die CSB waardes in meeste gevalle minder as 100 mg.L-1. Aangesien daar sulke groot variasies in mikrobe ladings oor die 15 maande tydperk voorgekom het, is die Colilert-18 sisteem gebruik om die weeklikse, daaglikse en uurlikse variasies vas te stel vir 6 weke oor ‘n periode van 4 maande by die Plank-2 bemonsteringspunt. Daar is spesifiek op die bemonsteringspunt gefokus omdat dit as ‘n besproeiingsbron gebruik word deur groente produsente. Dit is ook gelee ongeveer 2 km stroomaf van ‘n informele nedersetting. Die weeklikse variasies in totaal koliforms (TC) het ‘n afname oor die hele bemonsteringsperiode getoon, met die hoogstes telling van 3 200 000 MPN.100 mL-1 gedurende die warmer tydperk. Die E.coli (Ec) tellings het ‘n soortgelyke neiging getoon, met die hoogste telling van 440 000 MPN.100 mL-1 ook in Maart. Die daaglikse neigings was dieselfde vir beide die TC en Ec en die tellings het vermeerder van Maandag tot Donderdag, met ‘n afname tot Sondag. Die hoogste telling was op Donderdag met gemiddelde TC and Ec tellings van 1 900 000 and 160 000 MPN.100 mL-1, respektiewelik. Die uurlikse variasie profiel was soortgelyk vir beide TC and Ec met tellings wat vermeerder het van 06h00 tot 12h00 gevolg deur ‘n afname tot 18h00. Die toename in TC en Ec getalle soos vasgestel gedurende die weeklikse, daaglikse en uurlikse variasie het duidelik getoon dat die bemonsterings wat een maal per maand op Maandae om 08h00 gedurende die 15 maande tydperk uitgevoer is, tot ‘n erg onderskatting van die besmettings vlakke in die Plankenburg en Eerste Riviere gelei het. Die algehele weeklikse variasies vir die water temperatuur het ‘n verlaging oor die bemonsteringstydperk getoon terwyl die daaglikse en uurlikse variasie neigings ‘n verhoging van 06h00 tot 18h00 getoon het. Die weeklikse neigings vir pH het van die van die temperatuur verskil. Die analises van kovariante het geen korrelasie (p < 0.05) tussen die fisiese-chemiese (temperature en pH) parameters en die mikrobe veranderlikes (TC en Ec) getoon nie. Dus is daar afgelei dat temperatuur en pH geen direkte impak op die totale kolivorm of E. coli tellings gehad nie. Die data van die studie het duidelik getoon dat water van beide die Plankenburg en Eerste Riviere nie geskik is vir gebruik vir besproeiing van vars produkte wat rou geëet gaan word nie. In beide gevalle het die fekale besmettingsvlakke die DWAF en WHO leistreep oorskry. Besproeiing met sulke water hou ‘n gesondheidsgevaar in as gevolg van die teenwoordigheid van potensiële patogene wat oorgedra kan word na vars produkte.

Thesis (MSc Food Sc)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: Many South African farmers are forced to use water from nearby rivers for crop irrigation, since it is the most affordable and sometimes only source of water available to them. During this research project, a baseline study was performed on a farm irrigating fresh produce with water obtained from the Eerste River. The baseline study was done over a five month period, at six preselected sampling points, to determine the microbial and physico-chemical parameters of the water so a baseline could be established to compare the results to when the ultraviolet (UV) apparatus was installed (February 2013). Aerobic colony count (ACC), total coliforms (TC) and Escherichia coli (E. coli) were tested for during the microbiological study, while the physico-chemical analysis comprised of temperature, pH, conductivity, chemical oxygen demand (COD), alkalinity and total soluble solids (TSS). The UV treatment study was also performed over a five month timeline, at eight different sampling points (original six sampling points, with additional sampling points before and after UV). The same microbiological tests were performed during the UV treatment study, but turbidity and percentage ultraviolet transmittance (% UVT) were performed additionally during physico-chemical analysis. During the baseline study ACC, TC and E. coli counts as high as 9 600 cfu.mL-1, 13 799 MPN.100 mL-1 and 2 098 MPN.100 mL-1 were isolated at the river (Sampling Point 1), respectively. While performing the UV treatment study ACC, TC and E. coli counts as high as 142 000 cfu.mL-1, 241 960 MPN.100 mL-1 and 6 867 MPN.100 mL-1 were isolated at the river, respectively. As a result it was concluded that the Eerste River was mostly unsuitable for irrigation of fresh produce that are consumed raw. The higher counts in the river, during the UV treatment study might be attributed to the increase in rainfall that occurred in the sampling months (March to July 2013). The counts as measured at the point of irrigation are considered of greater importance, since the counts present in the river might still decrease to below the guideline levels after passing through sand filters and the addition of hydrogen peroxide (current mode of treatment) or after passing through the UV in the UV treatment study. The ACC, TC and E. coli counts during the baseline study were as high as 8 800 cfu.mL-1, 24 196 MPN.100 mL-1 and 85 MPN.100 mL-1 at the point of irrigation (Sampling Point 6), respectively. After hydrogen peroxide addition average logreductions ranging between 0.65 and 1.13 were seen, but reduction was never constant. The counts at the point of irrigation remained more or less constant compared to the river due to contamination that occurred at the sand filters, making the water unsuitable for irrigation of fresh produce in terms of ACC and TC counts. In the UV treatment study ACC, TC and E. coli counts were as high as 35 000 cfu.mL-1, 10 462 MPN.100 mL-1 and 63 MPN.100 mL-1 at the point of irrigation (Sampling Point 8), respectively. Average log-reductions in the range of 0.90 to 1.25 were achieved, but it was inconsistent. After treatment with chlorine and re-sanding of the sand filters, no further contamination occurred and counts decreased to below guideline limits, making the water safe for irrigational use in terms of all of the microbiological parameters. Not only is UV treatment more effective in reducing microbiological counts than H2O2, it is also relatively less expensive in the long term. Hydrogen peroxide treatment of water amounts to a very high capital expense every month, whereas UV may seem expensive when starting up, but the monthly operating cost thereafter is marginally less than for H2O2. It is of great importance to farmers to find a treatment that would reduce the counts in the river water to below the guideline limits required for safe irrigation since pathogens can be carried over from water onto fresh produce, resulting in an increase in produce-associated foodborne outbreaks and loss of consumer trust. / AFRIKAANSE OPSOMMING: Menigte Suid-Afrikaners is afhanklik van nabygeleë riviere om hulle oeste te besproei aangesien dit meestal die mees bekostigbare en soms enigste bron tot hul beskikking is. Tydens hierdie projek is ‘n grondslag sowel as ‘n UV behandelingsmetode studie uitgevoer op ‘n plaas wat vars vrugte en groente besproei met water water wat hul vanuit die Eersterivier verkry. Die grondslagstudie is oor ‘n tydperk van vyf maande uitgevoer by ses voorafgekose punte. Dit is gedoen om die mikrobiologiese sowel as chemiese parameters van die water te bepaal sodat ‘n grondslag beskikbaar kon wees om met resultate te vergelyk wat met behulp van die ultravioletmasjien verkry is (in Februarie 2013 geïnstalleer). Tydens die mikrobiologiese studie is daar vir aerobiese koliform tellings (ACC), totale koliforme (TC) en Escherichia coli (E. coli) getoets. Tydens die chemiese analise is temperatuur, pH, konduktiwiteit, chemiese suurstof benodiging, alkaliniteit en totale oplosbaie vastestowwe in die water getoets. Die UV behandelingsmetode studie is ook oor ‘n tydperk van vyf maande uitgevoer, met twee addisionale toetspunte by. Presies dieselfde mikrobiologiese analises as wat tydens die grondslag studie uitgevoer is, is tydens die UV behandelingsmetode studie uitgevoer, maar vir die chemiese analise het turbiditeit en persentasie ultraviolet transmissie van die water bygekom. Gedurende die grondslag studie was ACC, TC and E. coli tellings so hoog as 9 600 cfu.mL-1, 13 799 MPN.100 mL-1 en 2 098 MPN.100 mL-1 onderskeidelik uit die rivier geïsoleer (Punt 1). Tydens die UV behandelingsmetode studie was ACC, TC en E. coli tellings so hoog as 142 000 cfu.mL-1, 241 960 MPN.100 mL-1 en 6 867 MPN.100 mL-1 onderskeidelik by die rivier geïsoleer. Gevolglik is daar afgelei dat die Eersterivier se water meestal ongeskik is om te gebruik vir die besproeiing van vars groente en vrugte wat rou geëet word sonder dat enige verdere behandeling plaasvind. Die hoër tellings wat tydens die UV behandelingsmetode in die rivier sigbaar was kan hoofsaaklik toegeskryf word aan die toename in reënval in daardie tyd (Maart tot Julie 2013). Tellings soos gemeet by die punt van besproeiing is wel van groter belang as die wat aangeteken is by die rivier; aangesien die tellings wat in die rivier aangeteken is steeds kan afneem tot onder aanvaarbare hoeveelhede soos in die standaarde uiteengesit, want die water moet steeds deur sandfilters beweeg en word ook huidiglik deur waterstofperoksied behandel tydens die die grondslagstudie of beweeg deur die UV apparaat in die UV behandelingsmetode studie. Die ACC, TC en E. coli tellings soos gemeet by die besproeiingspunt (Punt 6) was so hoog as 8 800 cfu.mL-1, 24 196 MPN.100 mL-1 en 85 MPN.100 mL-1, onderskeidelik. Na waterstofperoksied byvoeging was die gemiddelde log-reduksies sigbaar, tussen 065 en 1.13, maar afnames was nooit konstant nie. Die tellings by die punt van besproeiing het ongeveer konstant gebly in vergelyking met die tellings wat by die rivier aangeteken is; moontlik as gevolg van die hoë kontaminasie vlakke in die sandfilters. Kontaminasie van sandfilters het veroorsaak dat die water ongeskik was vir die gebruik van besproeiing van vars groente as gevolg van die hoë ACC en TC vlakke. Tydens die UV behandelingsmetode studie is ACC, TC en E. coli tellings so hoog as 35 000 cfu.mL-1, 10 462 MPN.100 mL-1 en 63 MPN.100 mL-1, onderskeidelik aangeteken (Punt 8). Gemiddelde log-reduksies tussen 0.90 tot 1.25 was verkry, maar behandeling en afnames in tellings was nie konstant nie. Nadat die sandfilters met chloor behandel is en die sand daarin vervang is, het geen verdere kontaminasie by die punt voorgekom nie. Nadat al die voorafgenoemde behandelings afgehandel is, het die tellings tot laer as die van die standaarde gedaal en dus was die water nou veilig om te gebruik vir besproeiingsdoeleindes in terme van die mikrobiologiese parameters. Die UV behandelingsmetode is nie net meer effektief in die verlaging van mikrobiologiesese tellings as waterstofperoksied nie, dis ook heelwat goedkoper in die langtermyn. Waterstofperoksied behandeling van water lei tot ‘n baie hoë kapitale onkoste per maand, terwyl UV baie duur mag voorkom in die beginfase, maar die maandelikse kostes is aansienlik laer as die van waterstofperoksied en maak sodoende op daarvoor. Dit is van uiterste belang vir boere om ‘n water behandelingsmetode te vind wat die hoë tellings in die rivier sal afbring tot laer as Suid-Afrikaanse en Kanadese riglyne; aangesien patogene oorgedra kan word van vars vrugte en groente. Laasgenoemde kan tot ‘n drastiese toename in vars voedsel geassosieerde siektes en gevolglik ‘n afname in die vertroue wat ‘n kliënt in ‘n produk plaas, lei.