EFFICACY OF RIVERBANK FILTRATION IN HILLY AREA

Abstract

Riverbank filtration (RBF), a natural filtration process, is in practice in several parts of the world. The quality of the filtered water mainly depends on the aquifer characteristics and quality of the source water i.e. rivers or lake water. It has proved to be successful in many cities such as Dresden, Berlin, Budapest, Louisville, Haridwar, Nanital, Mathura, Ahmedabad, etotal coliform. RBF is a process where the subsurface water collected is directly under the influence of surface water. During the transport process, the contaminants from surface water are attenuated by a combined effect of filtration, degradation, sorption, dilution with ambient groundwater etotal coliform. The success of RBF in the middle and lower courses of the rivers in many parts of mainly Europe, North America and Asia has motivated to investigate RBF systems in the hilly/mountainous regions. Mountainous areas, in general, are remote areas and are difficult to access. Rivers in the mountainous areas are not polluted but have a very narrow discontinuous aquifer deposits with high flow velocities. Field investigations were carried out at four sites along the river Alaknanda and its tributaries at an elevation ranging from 551-769 m above mean sea level in Uttarakhand, India namely Srinagar, Karnaprayag, Agastyamuni and Satpuli for the purpose of assessing bank filtration systems. The hydro geology and water quality of the mountainous/hilly RBF systems were investigated from 2010-2013. To understand the process, (i) columns packed with same grading of aquifer material were operated at different flow rates in the Environmental Engineering Laboratory, IIT Roorkee and (ii) columns packed with various materials were continuously fed with the Elbe River water at Flügelweg, Dresden, Germany under ambient conditions. The river water was pumped to an overhead tank maintained at a constant level. Columns were operated under controlled conditions. Results obtained there from suggesting the effect of the operating parameters or conditions on the performance of an RBF system. RBF systems for reasonably clean rivers during non-monsoon and for high-suspended loadings in monsoon were investigated through hydro geological test such as sieve analysis from the sediment/aquifer materials, infiltration test, pumping tests in two phases (short and long duration). The first pumping test was carried out just after the drilling and commissioning of the production wells. The second set of experiments was carried out after operating pumps for one and half years. The temperatures at different depths were measured in the monitoring iv wells to study the nature of the interaction between the two sources i.e. river and groundwater. Double ring infiltration test was conducted to study the surface water infiltration rate during floods. The qualities of the river water, production well water, and ground water were assessed to evaluate the performance of the RBF. Water samples from production wells, hand pumps and from the River Alkananda at Srinagar and Karnaprayag, the Eastern Nayar at Satpuli and the Mandakini River at Agastyamuni were collected and analyzed for water quality parameters including stable isotopes (δ18O and δ2H) over a period of fourteen months in 2012-13. The mixing ratios of the groundwater to the river water in the bank filtrate were first established by analyzing stable isotopes δ18O and δ2H in water samples. Water qualities of the bank filtrates were compared with the drinking water standards. Factors influencing filtration of suspended solids and entrapment of coliform were investigated through column studies. Observations from the hydro geological investigations suggest the suitability of these sites for RBF. Hydraulic conductivity ranged from 4.0×10-3 to 6.8×10-4 m/s. The depths of the aquifers were more than 20 m below ground level and the grain size distribution was in the range of fine silt to coarse boulders. However, the gravel and sand were predominant. The infiltration rates were found to be in the range of values of existing established RBF sites. The maximum extractable safe yield of the wells at each location was determined to be between 1200 and 4000 L/min. The mean travel time was found to be less than one month at Satpuli and more than 10-12 months at Srinagar and Agastyamuni. Data, however, was inadequate to estimate precisely (i) travel time and (ii) seasonal variation in the mixing ratio. Analysis of stable isotope data suggests that the production well at Satpuli is connected to the river round the year. At Srinagar, bank filtrate is isotopically similar to the river water but chemically close to the ground water i.e. water has high mineral content including nitrate. A few leaching experiments with sediment/rock samples of that area were performed. Minerals were leached on soaking or stirring the distilled water with the sediment/aquifer material and a few phyllite rocks for less than 15 mins. The leaching experiments explain mineralization of the river water during its passage through the aquifers. Some of the phyllite rocks from the Pauri road leached nitrate as much as 500-2000 mg/kg of their mass. Nitrate concentration in many samples was more than 100 mg/L. Leaching of soil/rock samples indicates the geogenic v source of nitrate. δ15N isotope analysis of nitrate rich water from leaching experiments, and well water further substantiated the geogenic source of nitrate. Mineralization is maximum at Srinagar and minimum at Satpuli this is because of the short travel time between the river and production well. Total coliforms were detected once or twice in a year during monsoon. Fecal coliforms, however, were always less than the detectable numbers. Although the turbidity of the river water in monsoons was as high as 2000 NTU, in water of production well it was always less than 5 NTU. Water from the production well in Agastyamuni has up to 30-40 % of the river water. Since the production well in Karnaprayag has so far not been connected to the supply networks, there is no flux of river water into the well. As a result riverbank filtrate is less than 10% of the river water. At Satpuli, the production well draws the bank filtrate largely for most part of the year. Nearly, 90 % of the abstracted water is from the Eastern Nayar River. The temperature profile in the monitoring wells also suggests the minimum exchange of the river and subsurface water at Karnaprayag and maximum at Satpuli site. To assess the bank filtration under a broad range of operational conditions, two sets of column experiments were conducted with the sediment/aquifer materials collected from the bank filtration site. Columns packed with Srinagar aquifer material were operated at a flow rate from close to 0.91 to 12.60 m/d for 72 days. The experiments suggest that there is no effect of flow rates on turbidity removal. Turbidity of all the samples was less than 5 NTU. Coliform removal, however, was sensitive to the flow rates. Coliform was not detected in the water filtered at a rate of 1 mL/min. The hydraulic conductivity of the aquifer material in the column was reduced during operation, and the maximum reduction was in the column operated at 20 mL/min. In other words, filter material is clogged if operated continuously. Another set of four columns packed with materials of different effective sizes were operated for 31 days. Water was fed by gravity from a tank maintained at a constant level. The flow rate of the column was different due to the difference in material properties and head loss. One of the columns was filled with glass beads of uniform size representing coarser material or erosive conditions. Filtrate turbidity was always less than 5 NTU. The minimum flow rate was in the column filled with aquifer material and fine loamy soil at the top simulating the pre and post flooding condition when river beds are clogged. The maximum coliform removal occurred in the latter scenario. It was also observed that increasing the head results in de-clogging in the vi column filled with coarser materials (glass beads). The rise in the head at other columns(C-I to C-III) was not sufficient to remove the infiltrated solids. In glass beads column (coarser material), the entrapped solids were also removed when the flow was increased by increasing the head. However, in other columns solids deposited at the interface could not be dislodged by gravity flow suggesting that particles have entered deep into the filter material. These indicate the potential for clogging and de-clogging that can occur simultaneously during monsoon and non-monsoon in coarse materials but not in fine filter materials. The production wells by the side of the rivers can be effective in removing turbidity and coliform at pumping rates ranging from 490-4000 L/min. The water from the wells does not have removable impurities like turbidity and coliform other than minerals. The high velocities of the rivers in hilly areas are likely to reduce the probability of clogging of the aquifer. The RBF system is sustainable in the hilly/mountainous areas. The thesis has been organized into seven chapters. Chapter-1 defines the objective of the research. A brief literature and methodology adopted are presented in Chapter-2 and Chapter-3 respectively. The hydro geolpgy investigations along with the description of the four RBF locations in Uttarakhand, India form the subject matter of the Chapter-4. Observations on water quality monitoring at the four sites and the analysis of the data form the subject matter of the Chapter-5. Column experiments generating different scenario of RBF conditions have been described in the Chapter-6. Conclusion, limitation of the work and future scope of work are included in the Chapter-7.