Riverbank Filtration: Understanding Contaminant Biogeochemistry and Pathogen Removal

In Europe, riverbank filtration (RBF) has been the primary mode of drinking water production for many cities located along major rivers such as the Danube in Central Europe (from Austria to Black Sea), Rhine and Elbe in Germany, Lot and Seine in France, and Rhine in the Netherlands, as well as along rivers in Austria, Switzerland, Slovenia, and Spain. Many of these systems have been operating for well over a century. Lake bank filtration is also common in many European countries, including Finland, where wells are placed close to natural lakes or artificial reservoirs for drinking water production. In the United States, RBF systems are also used for drinking water production. They have been operating in many cities located along the Columbia, Missouri (including the sub-basins of Platte and North Platte), Mississippi (including the sub-basins of Des Moine, Minnesota, Cedar, and Illinois), Ohio (including the Wabash River basin), Colorado, Rio Grande, Russian, and Connecticut River basins for nearly half a century.

Engineers and scientists are faced with the problem of the behavior and fate of contaminants during the infiltration of river waters to groundwater. Sontheimer [1], e.g., found that induced bank filtration resulted in the elimination of some of the contaminants and could thus be considered as the first treatment step for the production of drinking water from river water. If rivers that naturally lose water to aquifers are contaminated, e.g., from outlets of sewage treatment plants, then the groundwater may become contaminated, too (e.g., Schwarzenbach et al. [2]). The quality of the water, which is freshly infiltrated from the river, revealed to be strongly dependent on the residence time and the mixing rate of river water and deeper groundwater. In a few infiltration systems, some contaminant compounds could partly be eliminated from the aquatic environment, and the quality of river water has somewhat improved (e.g., von Gunten and Lienert [3]). Today’s concerns are broader than looking at the water quality. The state of the groundwater/surface water ecosystem and its evolution in the future is of interest. In many floodplains, the state of riverbeds and banks is far from being natural (e.g., Brunke and Gonser [4]; Bencala [5]). Various forms of land use, such as hydropower generation, flood prevention, or the need of arable land, require a safe and defined bank line and a channeled riverbed. These requirements are in conflict with the needs of the vegetation and animals at the bank and in the bed of rivers (riparian zone).

High-quality drinking water is a resource in high demand. Because of easy access and high productivity, alluvial aquifers can supply large quantities of drinking water. Water is pumped in the aquifer rather than directly in the river because it is generally assumed that filtration through the porous geological medium improves the quality of the water. Bank filtration either is the only treatment (usually in small communities) or serves as a pre-treatment (mostly in large urban areas) before the water is distributed to the consumer.

Both groundwater and surface water are extensively used as sources of potable water. Groundwater may become contaminated with viral, bacterial and protozoan pathogens from domestic wastewater, e.g., improperly designed or malfunctioning septic systems, leaking sewer lines, land application of wastewater or its mixing with infiltrated surface water. Surface water may be contaminated with pathogenic microorganisms from discharges of treated and untreated wastewater and by manure run-off from agricultural land.

Bank filtration is a water treatment process [1,2] that makes use of surface water that has naturally infiltrated into groundwater via the riverbed or bank(s) and is recovered via a pumping well. Infiltration is typically enhanced by the hydraulic gradient imposed by a nearby pumping water supply or other well(s). Bank filtrate is water emanating from a pumping well that originated nearby as surface water and traveled through the subsurface, mixing to some degree with other groundwater. Through bank filtration, microbial pathogens, fecal indicator organisms, and other surrogates are removed by contact with the aquifer materials. The removal process performs most efficiently when groundwater velocity is slow and the aquifer is comprised of granular materials with open pore-space for water flow around the grains. In these granular porous aquifers, the flow path is very tortuous, thereby providing ample opportunity for the organism to come into contact with and attach to a grain surface. Although detachment can occur, it typically occurs at a very slow rate. Organisms typically remain attached to a grain for long periods. When groundwater velocity is exceptionally slow or when little or no detachment occurs, the organism will become inactivated before it can enter a well. Thus, removal of microorganisms by bank filtration relies on attachment to the soil/aquifer particles and inactivation.

The behavior of natural and synthetic toxins during bank filtration is commonly investigated either in controlled laboratory experiments or through observations at bank filtration sites. In laboratory batch and column experiments a range of parameters can be varied to assess processes potentially responsible for the elimination of substances. However, hypotheses gained from such experiments need to be verified for field situations. Observations at bank filtration sites are limited to the conditions encountered at such sites, and they are highly dependent on the occurrence of the agents to be assessed in the surface water. For example, cyanotoxins usually occur only during a few weeks of the year. Between the two extremes of laboratory experiments and field observations, controlled field experiments offer an intermediate approach that may substantially improve the understanding of processes relevant in natural systems.

In Germany, riverbank filtration as a natural treatment process for the production of drinking water has been used for more than 100 years [1,2,3]. Nowadays, the major raw water resource for the drinking water supply in Germany is groundwater (about 64%), whereas bankfiltrated (or infiltrated) water has a portion of about 16% [3,4]. Compared to this, direct abstraction of river water is of minor importance (less than 1%). In many cases, mostly along larger rivers, a clear distinction between bankfiltrated water and groundwater is difficult, and the raw water used for drinking water production is bankfiltrated water blended with groundwater.

The quest for pure water began in prehistoric times. Recorded knowledge of water treatment is found in Sanskrit medical lore and in Egyptian inscriptions. Pictures of apparati to clarify liquids (both water and wine) have been found on Egyptian walls dating back to the fifteenth century B.C. Boiling of water, the use of wick siphons, filtration through porous vessels, and even filtration with sand and gravel, as means to purify water, are methods that have been prescribed for thousands of years. In his writings on public hygiene, Hippocrates (460–354 B.C.) directed attention principally to the importance of water in the maintenance of health, but he also prescribed that rain water should be boiled and strained. The cloth bag that he recommended for straining became known in later times as “Hippocrates’ sleeve.”

There are a number of practical reasons for studying groundwater recharge via the riverbed, so-called riverbed infiltration. The most important are quantitative and qualitative reasons:

Due to Hungary’s natural endowments, the public utility water supply is fundamentally (more than 90%) based on subsurface water resources (Figure 1). This is in line with World Health Organization recommendations, since underground waters are much more protected than surface waters because of the overlying geological formations.

Riverbank filtration (RBF) is a significant part of drinking water production in Hungary. Around 40% of the public water supply is based on RBF. It means that about 1,300 Ml/d is abstracted from bank filtration wells. In Hungary RBF areas are mainly along the Danube River, where large sand-gravel alluvial deposits have good hydraulic conductivity.

As of January 2000, the Romanian population was 22,455,500, with 12,297,000 of the inhabitants living in cities and towns and 10,158,500 living in villages. The population is clustered in 263 cities and towns and 15,779 villages and smaller entities (communes). All cities and towns, as well as a smaller percentage of villages, have centralized water supply systems. Approximately 11.3 million inhabitants in cities and towns and 3.4 million inhabitants in villages have centralized drinking water systems (see Appendix for details).

The NATO Advanced Research Workshop primarily focused on three issues: (a) hydrogeology and contaminant biogeochemistry issues for riverbank filtration, (b) removal of pathogens by riverbank filtration, and (c) national experiences in riverbank filtration. The following conclusions were drawn.