Elsevier

Journal of Hydrology

Volume 344, Issues 1–2, 30 September 2007, Pages 55-69
Journal of Hydrology

River sediment and flow characteristics near a bank filtration water supply: Implications for riverbed clogging

https://doi.org/10.1016/j.jhydrol.2007.06.031Get rights and content

Summary

Riverbed clogging is an important issue related to the sustainable exploitation of riverbank filtration well fields. In this research, several complementary field techniques are employed to assess the current state and possible evolution of riverbed clogging at a site in the Saint John River, New Brunswick. The study is conducted in regions of the riverbed that have previously been identified as allowing recharge to the semi-confined aquifer that has been used since 1955 to supply water to the City of Fredericton. Flow velocity measurements, video imaging, and suspended sediment and bed sediment analyses conducted during the low flow (summer) period indicate that part of the recharge area closest to the well field, about 20% of the total area, is affected by bed armoring with cobbles and boulders. Consistent with previous studies, with increasing distance from the riverbank the sediment size decreases and the armor layer disappears. Previous research indicates that turbulent impacting of fine particles into the voids between the cobbles and boulders of the armor layer may reduce infiltration by up to 95%; however, the suspended sediment load in the river is mainly composed of organic matter, and the measured concentrations of suspended sediment (up to 3 mg/L) are not considered high enough to create such large reductions in infiltration. Additionally, the mineral fraction of the suspended sediment would not be expected to settle under the calculated average shear velocity of 0.012 m/s. Other sources of particulate matter, such as the degradation of aquatic vegetation on the riverbed, may be more significant with respect to riverbed clogging; however, annual peak flows may also create bed shear stresses that serve to limit long-term clogging effects.

Introduction

Riverbank filtration (RBF) systems consist of well fields that draw water from an aquifer that is hydraulically connected to surface waters. Pumping water induces a continuous flow of surface water towards the wells through the riverbank and/or the riverbed. Typical aquifers used for RBF consist of alluvial sand and gravel deposits with a hydraulic conductivity greater than approximately 10 m/day.

The benefits of RBF are multiple and include a reduction of turbidity, total coliform, microbial contaminants, natural organic matter (Tufenkji et al., 2002, Wang et al., 2002), and organic contaminants (Tufenkji et al., 2002). RBF is effective in removing some pesticides and some pharmaceuticals (Kuehn and Mueller, 2000) and is at least as efficient in removing disinfection by-product precursors as bench-scale conventional treatment (Weiss et al., 2003). Various authors also report that RBF is highly effective in reducing the risk of Giardia and Cryptosporidium contamination of drinking water when flow path length and filtration time are sufficient (Berger, 2002, Hiscock and Grischek, 2002, Gollnitz et al., 2003).

Some of the disadvantages of RBF include the difficulty of preventing river water from infiltrating the aquifer in instances of severe river contamination, the geochemical reaction of the infiltrate with aquifer materials that may raise the aqueous concentrations of Fe2+, Mn2+, As, NH4+, CH4, Ca2+ and HCO3-, and clogging of the riverbed (Stuyfzand et al., 2004). Hubbs (2003), for example, has demonstrated that riverbed clogging may decrease the specific capacity of RBF wells by as much as 50–75% during the first three to five years of exploitation.

During RBF, clogging of the riverbed is considered unavoidable (e.g. Wang et al., 2002, Schubert, 2004) because river water containing particulate matter infiltrates the riverbed during flow of water towards the well. Two types of clogging may occur; mechanical clogging due to the deposition of inorganic or organic fine-grained particles in the riverbed, and chemical/biological clogging resulting from geochemical or biological reactions.

Clogging is both a desirable and an undesirable process. Clogging may increase the efficiency of natural filtration since it is believed to have comparable characteristics to the “Schmutzdecke” in engineered slow-sand filtration systems: rich in organic matter, with high concentrations of microbes (Ray, 2002b, Tufenkji et al., 2002, Schubert, 2004). However, it may be detrimental because it leads to a decrease in the hydraulic conductivity of the riverbed sediments. The yield of RBF systems can therefore be severely affected by clogging, which is a problem for the sustainability of the drinking water supply (Tufenkji et al., 2002, Schubert, 2004, Caldwell, 2005).

Although riverbed clogging is known to be a factor in the sustainability of RBF water supplies, there have been relatively few direct investigations of the sediment characteristics and river flow conditions in the immediate area of riverbed filtration. The objective of this research was to characterize riverbed characteristics, and the implications of these for riverbed clogging, at an operating RBF water supply. This was accomplished by using a combination of field investigations and analyses that consider velocity profiles, suspended sediment, and bed sediment characteristics. The study provides useful insight into the extent of mechanical clogging at the investigated site and illustrates an approach that is applicable for other RBF sites.

Section snippets

Study site

The Saint John River flows about 700 km from its headwaters in northern Maine, USA, to the Bay of Fundy, New Brunswick, Canada. As shown in the inset of Fig. 1, the City of Fredericton is located in the lower reaches of the river, where the river width is approximately 700 m and the slope is relatively mild (∼0.1%). To supply potable water to a population of approximately 50,000, the City of Fredericton currently relies on eight vertical wells located near the south shore of the Saint John River

Field investigations

Field investigations were carried out between July and October 2005 as shown in Fig. 5. Data were collected during periods of relatively low flow because ice formation and break up, and high river discharge, precluded fieldwork during the winter and spring periods. The study was therefore conducted after high spring flows, which introduce fine sediments into the system through bank erosion and exposure due to the movement of overlying coarse material, had receded.

Field work focused mainly on

Bed material and suspended sediment analyses

All bed material samples were sieved to obtain cumulative size-frequency curves. The samples were individually oven dried for 24 h and then passed through sieves having openings ranging from 50 mm to 0.075 mm. A visual inspection of the samples indicated most of them did not contain much fine material and most samples could therefore be directly sieved after drying. However, for some samples the proportion of fines was large enough to require additional preparation before sieving. In these cases,

Riverbed material

Fig. 8 shows the different sediment zones within the two main riverbed windows, based on visual classification of the video recordings obtained in July and August. The map should be considered qualitative, especially in areas close to zone boundaries.

Four zones were distinguished in the western riverbed recharge area. The surface of the southern zone was a mosaic of moss and sediment in the gravel size range. In the video recording it was not possible to determine the type of sediment the moss

Discussion

The Saint John River near Fredericton is a dynamic system in terms of flow regime. The river discharge and stage experience considerable variations seasonally, and also daily due to upstream dam control. The Mactaquac hydroelectric dam was constructed in 1967 and is located approximately 15 km upstream of the study area. The limited storage capacity of the Mactaquac reservoir means that seasonal variations in river level and discharge closely follow natural patterns (e.g. Fig. 5a). Annual peak

Conclusions

This study was undertaken to assess the characteristics of riverbed sediment and flow, and the implication of these for riverbed clogging – one of the major problems occurring during riverbank filtration. Data from previously-identified infiltration areas on the bed of the Saint John River were collected during the annual low flow period.

Bed material sampling and video imaging support the presence of an armor layer in a portion of the infiltration area during the summer period. The pattern of

Acknowledgements

The authors would like to acknowledge the financial support of the Canadian Water Network (a federally funded Network of Centres of Excellence), the Canada Research Chairs Program, the New Brunswick Innovation Foundation Research Technicians Initiative, and the University of New Brunswick. Dennis Connor, Ken Noftell, Don Goodine, Andres Rodriguez, Jason Veino and Adrian Thompson all provided valuable assistance during various stages of field data collection. Melissa Dawe and Serban Danielescu

References (34)

  • K.M. Hiscock et al.

    Attenuation of groundwater pollution by bank filtration

    J. Hydrol.

    (2002)
  • T.A. Al et al.

    River – connected aquifers: geophysics, stratigraphy, hydrogeology and geochemistry

  • Amskold, L., 2005. Seasonal variation of redox zonation along an infiltration pathway in a river-recharged aquifer:...
  • APHA, AWWA, WEF, 1992. Standard methods for the examination of water and wahstewater. 18th ed. Published jointly by the...
  • Berger, P., 2002. Removal of Cryptosporidium using bank filtration. In: Proceedings of the NATO Advanced Research...
  • K.E. Butler et al.

    Delineating recharge to a river valley aquifer by riverine seismic and EM methods

    J. Environ. Eng. Geophys.

    (2004)
  • Butler, K.E., Nadeau, J.-C., MacQuarrie, K.T.B., Dawe, M.R., Hunter, J.A., Parrott R., 2006. Hydrostratigraphy and...
  • Caldwell, T.G., 2005. Factors significant to yield in riverbank filtration systems. MEng Thesis. University of...
  • A.B. Cunningham et al.

    Effects of sediment-laden flow on channel bed clogging

    J. Irrig. Drain.

    (1987)
  • Dawe, M.R., 2006. Assessing infiltration velocities during riverbank filtration. MScE Thesis. University of New...
  • GEMTEC Limited, 1992. Exploration for, and construction of, municipal well PW7. Unpublished report, Project File...
  • GEMTEC Limited, 1999. Construction and test pumping of production well PW8. Unpublished report, File 213.07R03,...
  • Goldschneider, A.A., 2006. River sediment and flow characteristics near a bank filtration water supply: implications...
  • W.D. Gollnitz et al.

    RBF as a microbial treatment process

    J. AWWA

    (2003)
  • Hubbs, S.A., 2003. Plugging in riverbank-filtration systems: Evaluating yield-limiting factors. In: Program and...
  • Hubbs, S.A., 2004. Processes involved in mechanical clogging of riverbank filtration systems. NATO Advanced Research...
  • P.Y. Julien

    Erosion and Sedimentation

    (1995)
  • Cited by (0)

    View full text