Jörg E. Drewes, Christiane Hoppe, Gretchen Oldham, John McCray, and Ken Thompson

OBJECTIVES:

The objectives of this project were to (1) quantify the degree of removal of key water quality constituents such as total organic carbon (TOC), unregulated organic micropollutants, and nutrients (ammonia, nitrate, and phosphorus) in full-scale riverbank filtration (RBF) systems as a function of geo-hydrological, operational, and water quality conditions; (2) understand at the mechanistic level the boundary conditions and relevant transport parameters for the removal of unregulated organic chemicals during RBF; (3) and summarize findings regarding removal potential and limitations in recommendations for design and operation of RBF systems.

BACKGROUND:

RBF is a natural process that has been used for public and industrial water supply in Europe for more than 100 years and in the United States for 50 years. The process has been shown to be effective in moderating the peak concentrations of various contaminants present in river or lake water. Today, water quality considerations have taken an increasingly important role in the expansion of riverbank filtration in the United States.

HIGHLIGHTS:

Findings of this study demonstrated that RBF systems have the capability to reduce and eliminate many contaminants of concern. In order to take full advantage of attenuation processes occurring in subsurface systems, proper retention times are needed. Most RBF installations will exhibit a dynamic change of redox conditions from oxic in the early phase of infiltration to anoxic during subsequent travel. This reduced environment will favor denitrification and with moderate loads to the system (river water nitrate concentrations of less than 5 mg-N/L), RBF can usually achieve nitrate concentrations of less than 2 mg-N/L in the RBF treated water. RBF, however, is not a reliable barrier for phosphate retention.

Where RBF systems are constantly exposed to organic matter and potentially also the presence of trace organic compounds, microorganisms will adapt to the type of organic matter present in the system and use this TOC as the primary substrate. The amount of BDOC present in river water is important for the removal of trace organic compounds. A well-adopted system can promote the removal of biodegradable trace organic compounds and first-order decay constants determined for targeted compounds in this study can be adopted to assess the removal under various site conditions.

APPROACH:

The first phase consisted of a comprehensive literature review to identify removal efficiencies of existing RBF systems as well as site characterization efforts of field RBF sites selected for this study. The second phase of the study consisted of a sequence of controlled experiments at the laboratory scale to evaluate the role of site specific and operational conditions on removal efficiencies for water quality parameters of interest using various column systems differing in soil type, loading rate, redox conditions, stage of acclimatization, and feed water qualities. These efforts were paralleled through water quality monitoring efforts at three full-scale RBF installations. These facilities were operated by the City of Aurora, Colo., at the South Platte River, the Louisville (Ky.) Water Company, at the Ohio River, and the City of Cedar Rapids, Iowa, at the Cedar River. The sites represented various groundwater travel times, distances from river to well, pumping rates, and hydraulic conductivities. The final phase of the study consisted of efforts directed to develop contaminant transport modeling approaches and recommendations for design and operation of RBF systems.

RESULTS/FINDINGS:

Monitoring efforts at three full-scale RBF sites revealed TOC concentrations ranging from 3 to 10 mg/L in the source water. Regardless of this range, temperature, and flow variations in the source waters, RBF treatment at all three sites resulted in TOC concentrations varying between 1 and 3.5 mg/L with standard deviations of less than 0.4 mg/L. All sites provided conditions favorable for achieving denitrification. This reduced environment resulted in nitrate concentrations of less than 2 mg-N/L in the RBF treated water. All full-scale RBF facilities also exhibited dissolution of manganese reaching levels in excess of the secondary MCL in the RBF treated water, which does require appropriate post-treatment. During RBF, phosphate is retained in the subsurface through metal precipitation and complexation, and field monitoring at two sites suggested that the adsorptive capacity in the initial phase of infiltration might have been already exhausted. Thus, RBF does not represent a reliable barrier to retain phosphate. Significant removal to below the detection limit was observed for most of the trace organic compounds under oxic redox conditions after a travel time of 6 days. The removal for the same compounds under anoxic conditions was not as efficient but continued during longer retention times (>20 days) in the subsurface to concentration levels also below the detection limit. Only chlorinated flame retardants and antiepileptic drugs did not exhibit removal under either condition.

IMPACT:

According to the EPA Long Term 2 Enhanced Surface Water Treatment Rule (LT2), bank filtration is defined as a "water treatment process that uses one or more pumping wells to induce or enhance natural surface water infiltration and to recover that surface water from the subsurface after passage through a river bed or bank(s)". It is noteworthy that RBF systems investigated in this study do fulfill the definition of bank filtration under the LT2 and therefore qualify for Cryptosporidium removal credits, but not all bank filtration sites under the LT2 might achieve water quality improvements (regarding TOC, nitrogen, and trace organic chemicals) as described in this study.

RESEARCH PARTNER:

City of Aurora Utilities

PARTICIPANTS:

• Louisville Water Company

• Cedar Rapids Water Department