Removal of Turbidity and Assessment of Groundwater Contribution during Riverbank Filtration

Abstract

Indian rivers carry large sediment loads. During monsoon season when river discharge is high, rivers carry suspended particles, silt, clay, and organic substances, causing river water turbidity that is not suitable for drinking. Riverbank filtration (RBF) is a natural process in which one can abstract good quality water from a highly polluted river through abstraction wells. As RBF water is a mixture of riverbank filtered water and groundwater, assuming no turbidity was picked up by the river water as it travelled through the soil matrix to reach the pumping well. A strategy-A was developed in which flow of groundwater contribution toward pumping well was neglected, while simultaneously measuring source water and abstracted water quality; it was observed that the variation of K, the kinetic coefficient, captured the variation in ln(C0), natural logarithm of influent quality, with R2 equal to 0.974. The kinetic coefficient, K, also captured the variation with ln(C0) of turbidity after travel times of 14 days and 21 days, having R2 equal to 0.991 and 0.989, respectively. A strategy-B was developed using the mass balance relationship, wherein the flow of groundwater contribution toward pumping well was utilized in refining the abstracted water quality. It was observed that K, the kinetic coefficient, of strategy-B captured the variation with ln(C0), the natural logarithm of the influent quality of turbidity with R2 equal to 0.972. If both the strategies were to be compared, strategy-B was found to be superior to strategy-A. However, one additional data of measurement on the groundwater turbidity was required for strategy-B. Further, strategy-C was developed with the assumption that there was no turbidity in groundwater, as a low value of turbidity was observed in groundwater. The variation of K, the kinetic coefficient, with ln(C0), the natural logarithm of influent quality of turbidity in strategy-C, showed that when the groundwater turbidity was ignored, the agreement between observed and computed turbidity was adversely affected. The average contribution of groundwater toward pumping well was observed nearly 12.43% and the average contribution of river water toward the pumping well was observed nearly 87.57%, considering the depth of river water. In strategy-C also, the agreement between simulated and observed values of turbidity was satisfactory.