In situ denitrification on nitrate rich groundwater in South Africa

Journal Title
Journal ISSN
Volume Title
Stellenbosch : Stellenbosch University
ENGLISH ABSTRACT: South Africa is a water scarce country and in certain regions the quantity of surface water is insufficient to provide communities with their domestic water needs. In many arid areas groundwater is often the sole source of water. This total dependence means that groundwater quality is of paramount importance. A high nitrate concentration in groundwater is a common cause of water being declared unfit for use and denitrification has been proposed as a potential remedy. In many areas of South Africa nitrate levels exceed the recommended maximum concentration of 40 mg/L NO3- as N. Concentrations of 100 mg/L NO3- as N or even greater than200 mg/L NO3- as N are found in various places. Water with nitrate concentrationsexceeding 40 mg/L NO3- as N, belongs to the category of “dangerous” drinking waterquality (“purple”, i.e. Class IV) according to DWA (1996, 1998) water quality guidelines. Concentrations in this range have been reported in case studies to cause conditions like methaemoglobinaemia (“blue baby syndrome”), spontaneous abortions, stomach cancers and livestock deaths. The purpose of the study includes laboratory experiments to compare the denitrification efficiency, reaction rates and reaction mechanisms between woodchips, biochar and a mixture of woodchips and biochar. Further work included modelling of denitrification using the PHREEQC-2 1D reactive transport model. Field implementation of a denitrification technique was tested at a site which previously experienced some NH4NO3 spills, to determine the lifespan of the woodchips used during the experiment based on available data. The underlying intended purpose of this research is to contribute to the wellbeing of rural South Africans in areas where groundwater is plentiful, but elevated nitrate levels prevent the use of this water. The purpose of the laboratory experiment was to establish the efficiency of carbon sources and compare their rates, sorption properties and processes by which they react. Laboratory experiment consisted of three leaching columns containing two layers of building sand on either side of a carbon containing layer. The carbon containing layers were made of about 600g of woodchips, biochar and woodchip and biochar mixture respectively. Parameters analysed from the effluent from the columns included NO3-, NO2-, SO42-, NH4+, Alkalinity, DOC (dissolved organic carbon content) and Phosphate. The purpose of the field experimental work was to install a barrier containing a cheaply available carbon source to treat groundwater and to monitor changes with time in order to determine the efficiency and life span of carbon source used for the experiment. Experimental work was done at a site in Somerset West (South Africa) that had experienced spills in the past from agrochemical storage factory premises. Somerset West normally receives about 568 mm of rain per year. It receives most of its rainfall during winter; it thus has a Mediterranean climate. It receives the lowest rainfall (10 mm) in February and the highest (96 mm) in June. The “reactor”/ tank with dimensions- 1,37m height, 2.15m diameter used for the experiment was slotted for its entire circumference by marking and grinding through the 5mm thick plastic material. The top section was left open to allow for filling and occasional checking of filled material during the experiment. The tank was packed with Eucalyptus globulus woodchips which was freely available at the site. Concentrations of groundwater nitrate at the site were well over what could be expected in any naturally occurring groundwater systems, and would result only by major anthropogenic activities in unconfined aquifer areas of South Africa. Nitrate levels in monitoring boreholes at the site ranged from about 20 mg NO3--N/L at background boreholes up to about 600 mg/L NO3--N. Woodchips used to denitrify groundwater in the field experiment were sampled after 27 months and 35 months of being active in the treatment zone. Various depths of samples were collected namely the top section, bottom of the tank and a full core sample of the tank. Main results from the laboratory studies showed that biochar on its own as a carbon source for nitrate removal would not be viable, however, the presence of biochar in the mixture of woodchips and biochar increased the rate of denitrification. Biochar on its own was able to remove some nitrate, but results showed incomplete denitrification and limited reactivity. The results also confirmed that different processes were in play, while the redox reaction of denitrification was taking place in woodchips and biochar and woodchip mixtures, the biochar treatment followed a physical process and had only a small percentage of incomplete denitrification. This was confirmed by sulphate reduction and increased alkalinity in the woodchips and biochar and woodchip mixture treatments. Rates deduced from the data also showed that the woodchip and biochar mixture would take a shorter period to affect total denitrification. Main results from the field work showed that nitrate was totally removed at the treatment zone and surrounding boreholes, and even sulphate and NH4+ were removed during the experiment. This shows that the woodchips were successful in affecting denitrification for 35 months. Data also shows that boreholes further downstream from the tank had reduced NO3-, SO42- and NH4+ levels. This would relate to higher permeability flow paths possibly present on the downstream side of the treatment zone. This became evident when pumping boreholes during sampling and noting that upstream boreholes had to be allowed for a recovery period, while downstream boreholes could be pumped continuously for 30 minutes without any reduced yields. This shows that not only did the treatment zone work at removing nitrate, but migration of excess available carbon from the tank may have further treated nitrate rich areas on the site. During monitoring on the site, woodchips were sampled and analysed for their components at time period 27 months and 35 months of the experiment respectively. Results showed that woodchips were considerably more degraded than a) woodchips of the same species of tree that had undergone natural degradation on the floor and b) un-degraded woodchips of the same tree species. Comparing data from the two time series samples, a rate of woodchip degradation could be calculated. Using the available biodegradable carbon for the woodchips based on its composition, a barrier lifespan could be determined. The results of calculations show that the barrier would be effective for at least another 6.9 years from the period of the last sampling date. A total lifespan of about 10 years can thus be estimated. These calculations are tree species composition specific and rate specific. PHREEQC-2 modelling was used to estimate the use of carbon in the experiment by adding incremental moles of carbon to the influent composition. Saturation indices from PHREEQC-2 showed that mineral phases of iron may precipitate from solution during the experiment. Experimental data were plotted against results of intermittent carbon reactions in PHREEQC-2 and it was found that initial rates in the experiment were higher and agreed with up to 100mg/L of carbon consumption when a 24 hour residence time was used while later stages agreed with about 37.5 mg/L carbon consumption, where a 72 hr residence time was used. It was concluded that biochar and woodchips combined are more effective than woodchips on their own at denitrifying groundwater. Also woodchips successfully denitrified groundwater at the Somerset West site for 35 months, with added removal of sulphate and NH4+. Barrier life span calculations show that the barrier could remain active for an additional 6.9 years which relates to a total period of about 10 years of denitrification should the rates remain constant. It was concluded that nitrate removal and barrier lifespan would be extended by testing variable lignin content in different tree species prior to use in a denitrification barrier as lignin is unlikely to degrade in an anaerobic environment. It was recommended that implementation or field test should be done using a biochar and woodchip mixture. Improved results may be achieved by analysing wood or plant material for comparative lignin content, cellulose content and hemicellulose contents. Wood types or plant species with higher lignin content would be more resistant to degradation in anaerobic conditions.
AFRIKAANSE OPSOMMING: Suid Afrika is ‘n waterarm land en in sommige gebiede is die beskikbare bogrondse water te min om aan plaaslike gemeenskappe se huishoudelike vereistes te voorsien. Grondwater is in hierdie gebiede die enigste bron van drinkwater. Dit is dus baie belangrik dat die grondwatergehalte van sodanige aard is dat dit met die minimum behandeling geskik sal wees vir mens en dier. Dit is egter so dat hoë nitraat– vlakke in baie gevalle die algemene rede is waarom grondwater ongeskik verklaar word vir huishoudelike gebruik. As gevolg hiervan word in-situ denitrifikasie van grondwater voorgestel as ‘n moontlike oplossing vir hierdie probleem. In baie gebiede in Suid Afrika is nitraat vlakke hoër as die aanbevole maksimum nitraat vlak van 40mg/L NO3- as N. Nitraat vlakke van ongeveer 100 mg/L en meer as 200 mg/L is al in verskeie gebiede gevind. Grondwater met nitraat vlakke hoër as 40 mg/L word as “gevaarlik” beskou volgens die WGO en Departement van Waterwese se riglyne. Gekonsentreede vlakke in hierdie reeks is deur gevallestudies aangemeld wat metheamoglobinaemia, maagkanker en spontane aborsies en sterftes by lewende hawe veroorsaak. Die doel van hierdie studie was om 1) laboratorium eksperimente uit te voer om die doeltreffendheid van ’n koolstofbron vir denitrifiekasie, die reaksietempo en die meganismes te vergelyk tussen hout stukkies, houtskool en ‘n mengsel van houtskool en houtstukkies 2) Modellering van denitrifikasie in PREEQC-2 te doen, 3) Veld toepassing van ’n denitrifikasie tegniek, die eerste van sy soort in Suid Afrika wat getoets was, op ’n area wat voorheen besoedel was. 4) om die lewensduur van die houtstukkies wat in die eksperiment gebruik word te bepaal. Die onderliggende doel van die navorsing is om ’n positiewe bydra te lewer vir ver afgeleë Suid Afrikaaners waar daar volop grondwater beskikbaar is, maar hoë nitraat die gebruik van die water verhoed. Laboratorium eksperiment bestaan uit drie kolomme waarvan twee lae bousand aan beide kante van die koolstof laag geplaas is vir loging. Elke koolstof laag het bestaan uit 600g houtstukkies, houtskool, asook ’n kombinasie van die twee onderskeidelik. Die uitvloeisel is ontleed vir NO3-, NO2-, SO42-, NH4+, alkaliniteit, DOC (opgeloste organiese koolstof) en fosfaat. Die veld eksperiment is gedoen in ’n besmette area (Somerset Wes, Suid Afrika) waar agrichemikalië voorheen in ’n fabriek gestoor was. Die hoogste reënval word gedurende die winter maande in hierdie area ontvang en Somerset Wes het ‘n gemiddeld van 568 mm per jaar. Dit het ’n Meditereense klimaat waar die laagste reënval in Februarie 10mm en hoogste reënval in Junie 96mm is. Die reaktor/ tenk met hoogte 1.37m en deursnee van 2.15m is gebruik vir die eksperiment. Die omtrek van die 5mm dik plastiek tenk was verdeel in gelyke dele waarvan elk sowat 2mm breed was. Die boonste gedeelte van die tenk was onbedek sodat die houdstukkies daarin geplaas kon word asook om gereelde opvolg besoeke toe te laat. Die tenk is opgevul met Eucalyptus globulus houtblokkies wat vrylik in die area beskikbaar is. Daar is bevind dat die konsentrasie van grondwater nitraat ver bo die natuurlike nitraat vlakke in grondwater bronne was, en kan vergelyk word met areas waar antropogeniese aktiwiteite op onbeperkte ondergrondse waterbronne, plaasvind. Nitraat vlakke het gewissel van ongeveer 20 mg/L tot 600mg/L by boorgate. Monsterneming van houtstukke wat gebruik was in die veld eksperiment het na ‘n tydperk van onderskeidelik 27 maande en 35 maande plaasgevind. Monsters is op verskeie dieptes geneem, dit het die boonste gedeelte, die onderste gedeelte en ’n volledige profiel van die tenk ingesluit. Hoofresultate van die laboratoriumstudie het getoon dat die houtskool op sy eie nie effektief was as ’n koolstofbron nie, alhoewel dit die effektiwiteit van die houtstukkies verbeter het waar dit in die mengsel was. Houtskool het ’n klein hoeveelheid nitraat verwyder, maar resultate dui daarop dat dit onvoltooide denitrifikasie veroorsaak het en beperkte reaktiwieteit getoon het. Die resultate dui ook aan dat daar definitief verskillende meganismes van reaksie plaasgevind het. Terwyl houtstukkies en die houtstukkie- houdskool mengsel via redoks reaksies gedenitrifiseer het, het houdskool ’n meer fisiese proses gevolg, en ’n klein persentasie nitraat nie totaal verwyder is nie. Dit was bevestig deur sulfaat reduksie en ’n verhoging van alkaliniteit in die houtstukkies en houtstukkies en houtskool mengsel. Reaksietempo resultate van die navorsing toon dat houtstukkies en houtskool mengsel in ’n verkorte periode wel totale nitraat verwydering getoon het. Hoofresultate van die veld toepassingseksperiment dui daarop dat nitraat totaal verwyder was in die reaksiesones en boorgate rondom die reaksiesones. Sulfaat en NH4+ was ook in die tydperk verwyder. Dit dui aan dat houtstukkies suksesvol aangewend is as ’n koolstofbron vir denitrifikasie gudurende die 35 maande wat die eksperiment gemoniteer was. Die data dui ook aan dat boorgate, verder stroomaf van die tenk, verminderde NO3-, NH4+, en SO42- vlakke getoon het. Dit dui op ’n area van hoër deurlaatbaarheid aan die stroomaf gedeelte van die reaksiesones. Dit het duidelik geword gedurende die grondwater monsternemingsprosedure waartydens boorgate aan die stroomaf gedeelte van die studie area vir 30 minute aanhoudend gepomp kon word sonder om minder water te lewer, terwyl boorgate wat stroomop geleë is gepomp kon word met ’n herstel periode omdat water langer geneem het om die boorgat te bereik. Dit dui aan dat die reaksiesones suksesvol gewerk het, en ook dat ongereageerde opgeloste koolstof wel van die reaksiesones af kon versprei in die stroomaf rigting. Monsterneming van houtstukke het plaasgevind na sowat 27 en 35 maande, en ontleding van uiteensettings produkte van houtstukke was gedoen. Resultate dui aan dat die houtstukke uit die reaksiesones aansienlik meer afgebreek het in vergelyking met 1) houtstukke wat onder natuurlike toestande afgebreek het en 2) natuurlike Eucalyptus globulus monster. ’n Reaksietempo vir afbreek van houtstukke kon bepaal word deur die data van elke monster van dieselfde posisie te vergelyk vir die verandering in kosentrasie oor die tydperk. Die bio-afbreekbare koolstof kon bepaal word met die beskikbare data vir die houtstukke. Die lewensduur van die reaksiesones kon bepaal word met die reaksietempo en beskikbare bio-afbreekbare koolstof vlakke. Die resultate dui aan dat die reaksiesones vir ten minste nog 6.9 jaar effektief kan wees na die 35 maande tydperk. Dit dui aan dat hierdie eksperiment vir ongeveer 10 jaar suksesvol kan opereer voordat houtstukke in die reaksiesones nie verder kan afbreek nie. Hierdie bepaling is egter spesie spesifiek en reaksie spesifiek. Gevolgtrekkings van hierdie werk is 1) die houtskool en houtstuk mengsel behandeling meer effektief is as ‘n koolstofbron vir denitrifikasie vergeleke met net houtskool. 2) Houtskool alleen is nie effektief as ’n koolstofbron vir denitrifikaie nie. 3) Houtstukke het suksesvol grondwater gedenitrifiseer vir ’n tydperk van 35 maande, met addisionele verwydering van sulfaat en NH4+. Houtstukke wat gedurende hierdie projek gebruik was kan vir nog 6.9 jaar suksesvol denitrifiseer. Dit was aanbeveel dat toekomstige werk dalk die gebruik van die houtstuk en houtskool mengel in die veld toets. Vir beste resultate met houtstukke, word dit aanbeveel dat ’n verskeidenheid van plantspesies ontleed word. Plantmateriaal of houttipes met hoër lignin vlakke sal minder afbreekbaar wees, en dalk ’n beperkte lewensduur hê in vergeleke met plantspesies wat laer lignin vlakke het.
Thesis (PhD)--Stellenbosch University, 2015.
Groundwater -- South Africa, Domestic water needs, Nitrate rich groundwater, UCTD