REMOVAL OF DISSOLVED ORGANIC CARBON (DOC) DURING RIVERBANK FILTRATION

Abstract

Riverbank filtration (RBF) has been extensively in use, throughout the world to supply safe and wholesome drinking water, in places of fluctuating surface water quality. The benefits of RBF in attenuating the various contaminants have been extensively studied and reported. Although RBF has been in use in India since, a long time, the water quality improvements due to bank filtration has been studied and reported only since 2005. The studies from across the world has shown that attenuation capacity of any bank filtration capacity is highly site specific. Attenuation capacity of RBF with respect to organics has been gaining momentum, in the recent years due to the stringent regulations which would be coming into effect with respect to DBPs. The most practical and the cost effective solution of control of DBPs is to control the source water quality. In developing countries like India, removing DBPs after its formation is not feasible. Hence the emphasis has to be on treatment schemes that would prevent or reduce the DBP formation. RBF schemes help in attaining the above said benefit. Several studies have shown the effectiveness of bank filtration in removing organics and trace organic compounds and thereby reducing the cost of treatment (Amiri et al. 2004; Börnick et al. 2001; Grünheid et al. 2005; Heberer et al. 2004; Kumar et al. 2012; Maeng et al. 2011; Ray et al. 2003; Ray et al. 2002; Schoenheinz et al. 2003; Worch et al. 2002). However, the results of these studies cannot be directly applied to Indian conditions due to the highly site specific nature of bank filtration mechanisms as well as due to the difference in operating conditions. Furthermore, most of the studies involving organics removal during bank filtration elsewhere have concentrated on rivers where DOC is < 6 mg/L. However in India most of the rivers in the urban areas have DOC > 6 mg/L. This may cause significant redox changes in the aquifer which may result in anaerobic conditions. Anoxic biodegradation is much slower than oxic biodegradation. Hence the effect of higher organic contamination on effectiveness of RBF to remove organics has to be investigated. Moreover, in Europe and the United States field tests are conducted at temperatures ranging from 5-25 °C, where in India temperatures vary from 5-40 °C. The observed water temperatures at Mathura ranged from 10-35 C. Higher temperature occurring predominantly affects biodegradation, adsorption and oxygen vi concentration in river water and hence the removal mechanism. Furthermore, the possibility of high DOC resulting in nutrient limitation condition for DOC degradation in the aquifer also needs to be explored. Moreover the Indian practice of placing the well in the riverbed is unique (Kumar et al., 2012) and the effects of this practice on the attenuation benefits of bank filtration have not been studied. One of the major disadvantages of this practice is that the travel time is relatively short and accordingly only limited purification in terms of organics and microorganisms can be achieved. Moreover RBF is, relatively untested in monsoon climates, i.e. locations dominated by strong seasonal rains followed by a prolonged dry season. Keeping these factors in view, it was proposed to evaluate the removal of dissolved organic carbon (DOC) during riverbank filtration. DOC, UV absorbance at 254nm and colour are kept as the main monitoring parameters keeping in view the stringent water quality regulations which may come into effect with regard to disinfection by products such as THMs, HAAs and other chloro-organics. Research objectives and scope The objective was to assess the potentials of bank filtration with regard to attenuation of organics where the surface water is polluted having distinct disagreeable visual properties. Work scheme In order to attain the objectives delineated in the first chapter the work was carried out in three different phases 1. Phase I included: (a) The isotopic studies to ensure that the water in the radial well is bank filtrate. Determination of mixing ratios (mixture of river and groundwaters) using the isotope data (δ18O and 2H), chloride mass balance and conductivity methods, (b) Sampling and analysis of water samples collected from (i) common source (River Yamuna at Mathura), (ii) the radial well (with RBF and without pre-chlorination), and (iii) the conventional treatment plant (without RBF and with pre-chlorination), and Chlorination experiments with the river and riverbed-filtered waters to evaluate the maximum THM yields, 2. Phase II included experiments and studies carried out to assess effect of biodegradation on the removal of organics during RBF. 3. Phase III included the experiment conducted to assess the effect of adsorption on the removal of organics during RBF. Surface water and groundwater systems are two primary sources to any RBF well. In the present study, isotopic and water quality characterization of the Yamuna River, along with the radial well water and groundwater reveals that River is the main feeding source to the radial well during all seasons. The average contribution of the river in the radial well was about 60-67% calculated by four different methods. Isotopic studies also revealed that River is one of the recharging sources for ground water table. Assessment of bank filtration potential of River Yamuna The various water quality parameters studied conclusively prove the efficacy of riverbank filtration in providing water of adequate quality even when the source water is polluted. The efficacy of RBF was evaluated based on the removal efficiency of parameters like turbidity, coliforms and organics. Results from monitoring indicate that the radial well at Mathura provides water with turbidity less than 1 NTU throughout the period of monitoring although the river water turbidity fluctuated significantly during the monitoring period. Reduction in turbidity varied from 68-99 % with an average removal rate of 91 %. The total as well as fecal coliorms showed significant reduction post bed filtration. The of surrogate parameters of organics like DOC, UV absorbance and colour also showed significant reduction post bed filtration. Comparison of bed filtration to conventional treatment Bed filtration was much more effective in reducing the turbidity, coliforms, organics in water compared to as indicated by the appreciably reduced values of turbidity and coliforms along with the surrogate parameters of organics like DOC, UV-abs at 254nm and colour. The log reduction of coliforms along with percentage reduction of surrogate parameters of organics and turbidity during the non monsoon and monsoon viii season for both the bed filtrate and conventionally treated water is tabulated in Table 1. Table 1: Comparison of removal of turbidity, coliforms and organics Parameter Bed Filtrate Conventionally Treated water Non Monsoon Monsoon Non Monsoon Monsoon Turbidity 67-98(90) * 84-99(94) 28-97(72) 80-90(88) Total coliforms 0.3-4.0(2.0) 0.2-0.4(0.3) 3.0-5.6(4.7) 3.9-7.6(6.2) Fecal coliforms 0.3-3.4(1.9) 0.2-0.4(0.3) 3.0-7.0(4.9) 3.9-8.1(6.4) DOC 26-86 (48) 22-57(40) 1-30(11) 3-19(7) UV absorbance 37-70(52) 48-80(58) 4-48(4) 5-78(25) Colour 34-81(70) 44-98(65) 1-90(45) 48-95(65) *Value inside the parenthesis denotes average reduction Analysis of DBPs Contrary to expected high values, Yamuna water, bed-filtrate, and RBF treated water were found to have very low values of total THMs. Summation of THM species in Treated Water (i.e. pre-chlorinated) amounted to 30.5 μg/L, which is lesser than the guidelines suggested by EU and USEPA. Considering the high dose of Cl2 applied at the treatment plant on the date of sampling (i.e. 40 mg/L) this appeared to be quite low (Kumar et al. 2012). Chlorination experiments carried out indicated that concentrations of THM species were found to be either low or less than detectable. AOX concentration was found to be low in the river water (23.5 μg/L) and bed-filtrate (17.5 μg/L), but it was found to be as high as 268 μg/L in the pre-chlorinated treated water. Although onetime assessment of either THMs or AOX precludes formation of any significant conclusions with regard to DBP formation in Yamuna water the from the results of the above study clearly indicates that bed filtration results in significant reduction of DBPs irrespective of the group of DBPs formed. Comparison of removal of turbidity, DOC and pathogens in different RBF sites with the present data The average removal rate of turbidity in the present study (91 %) is slightly lesser than the minimum removal rate documented even though the final turbidity at all instances was below the US Environmental Protection Agency(EPA) recommended value of 1 NTU. The coliforms removal in the radial well at Mathura is comparable to the removal rates reported in sites across the world. The log removals are slightly lesser than the values reported by (Weiss et al. 2005) and (Schijven et al. 2003). The lesser values are due to the influx of coliforms into the well during the heavy flow conditions during monsoon season. The average removal for DOC documented in the various BF sites throughout the world covers a wide range from 14 % to 84 % but usually lies between 39 % and 50 % (Grützmacher et al. 2009). The removal rates for the present study (40-48 %) also fall well within this range. Phase II: Biodegradation studies The objective of this study was to evaluate the role of biodegradation in organic removal/transformations during bank filtration under a variety of concentration ranges. The experiments also aimed to compare the feasibility of using batch BDOC tests in place of test filter experiments for determining the first order rate constants for biodegradation. The BDOC tests were conducted in both oxic and anoxic ranges to determine the extent of biodegradation. The test filter experiments were carried out in different concentration ranges and using different source waters and the results were compared. The kinetic data for both DOC and UV absorbace for both the test filters and BDOC experiments were fitted by a two term model (shown by dashed lines) of the form C=Co e-kt +C1, where k is the biodegradation coefficient and C1 represents the refractory fraction of UV absorbance or DOC. Comparison of rate constants and removal rates from test filter experiments and BDOC tests indicate that the rate constants for degradation obtained for the different source waters from the test filters were higher than that obtained from the BDOC experiments. Consequently the half life period of DOC for different source waters in the test filter also were lower. The percent removal of DOC was also higher in the case of test filter for all the concentration ranges studied except or Yamuna water in the lower concentration range. The higher degradation rate and the removal of DOC in the test filter in comparison to the BDOC experiments with the approximately same contact period suggests that biosorption onto the biofilm is a predominant factor influencing the removal of organics. The change in character of source waters during biodegradation was assessed by studying the SUVA values. In all the source waters studied, it was noted that the SUVA values did not show any particular trend before and after the x biodegradation studies. The decrease or increases noted in all the cases post the test filter experiments and BDOC tests were too erratic to conclusively predict a change in character of source waters. Both the test filter experiments and BDOC tests simulating biodegradation during riverbank filtration clearly indicate that under favourable conditions, bank/bed filtration can result in appreciable attenuation of organics, irrespective of the influent concentration. Phase III: Adsorption studies Adsorption is an important process whose role in the attenuation of organics cannot be neglected. Adsorption studies were carried out to quantify the effect of adsorption in attenuation of organics during bank filtration. The objective of this study was also to examine the possibility of using batch experiments to predict the retardation factors of DOC for different aquifer materials. The adsorption experiments were conducted in three different concentration ranges of DOC using three different aquifer materials. The results of the adsorption experiments indicate that the batch tests experiments seem to be appropriate for rapid determination of retardation factor for various aquifer materials in different ranges of DOC. The behavior of the different source waters were studied by comparing their respective breakthrough curves (BTCs). In order to assess the combined effects of both adsorption and biodegradation experiments were conducted with columns filled with different aquifer materials operated in recirculation mode. The removal in the column experiments would be due to combination of biodegradation and adsorption. The percentage removal of DOC in all concentration ranges varied 38-58 % which were comparable to the removal those observed in field. The biodegradation rates obtained from the biodegradation studies conducted were used to model the combined effects of adsorption and biodegradation in the breakthrough curves for the columns. The adsorption capacity /unit weight of the aquifer material derived from the columns run in recirculation mode was then compared to the adsorption capacity/unit weight derived from the batch tests to quantify the removal due to adsorption and biodegradation. The results indicate that although biodegradation seems to the predominant mechanism for attenuation of organics during bank filtration the effect of adsorption cannot be neglected. The percentage removal due to biodegradation ranged from 50-86% where as the percentage removal due to adsorption ranged from 14-50%. Conclusions Results from the present study suggest that, even though riverbank/bed filtration does not completely eliminate the need for water treatment, use of RBF significantly reduces the organic contamination which in turn may result in the reduction of the concentrations of chloro-organics in the treated water. The results of turbidity and total and fecal coliform also show that RBF process is very effective in removing particles in surface water suggesting that it is an effective water treatment process for reducing the potential of microbial contamination in drinking water. The results from the monitoring also indicate that bed filtration is effective in attenuating the concentration of DBPs irrespective of the group of DBPs formed. From the present study it can be concluded that riverbank/bed filtration is an effective alternative to pre-chlorination of polluted waters even though the need for further treatment is not eliminated. Furthermore the data from the monitoring program coupled with that obtained from the biodegradation experiments suggests that biodegradation is the primary sustainable mechanism for removal of organics during bank filtration. From the present study it can be concluded that test filter experiments can be used effectively to simulate the biological degradation occurring during riverbank/bed filtration. BDOC experiments can be used as a rapid way of assessing the general biological lability of different samples and the direct affect of biodegradation on the DOC structure in place of test filter experiments. The results from adsorption studies indicate that batch tests can be used for rapid determination of retardation factors in case of sandy soils. The retardation factors determined from batch and column studies can be used in transport modeling. Biodegradation attributes to about 50-86% of removal of organics and rest is attributed to adsorption. Adsorption is a significant removal mechanism of organics during sub surface transport