Das Berliner Trinkwasser wird überwiegend durch induzierte Uferfiltration entlang der Oberflächengewässer gewonnen. Durch die geringen Durchlässigkeiten der Seesedimente findet eine Infiltration nur an den besser durchlässigen Uferzonen statt, und es kommt zu einer Unterströmung der Seen. Durch die Kombination verschiedener Umwelttracer konnte eine starke vertikale Altersdifferenzierung des Uferfiltrats nachgewiesen werden. Die Fließzeiten betragen in den flacheren Grundwasserleiterbereichen einige Monate, in den tieferen Bereichen sogar mehrere Jahre. Das den Abbau redox-sensitiver Substanzen beeinflussende, vorherrschende Redoxmilieu weist ebenfalls eine starke vertikale Differenzierung auf, die Infiltration erfolgt überwiegend anoxisch, und das Uferfiltrat wird mit der Tiefe reduzierender. Da das Oberflächenwasser einen variablen Anteil geklärten Abwassers enthält, konnten einige abwasserbürtige Substanzen (z.B. pharmazeutische Rückstände) in Oberflächenwasser- und im Uferfiltrat nachgewiesen werden. Obwohl der überwiegende Teil pharmazeutischer Rückstände effizient während der Untergrundpassage entfernt wird, erwiesen sich einige Substanzen als äußerst persistent (AMDOPH, Primidon und Carbamazepin).

Berlin relies on induced bank filtration from a broad-scale, lake-type surface water system. because the hydraulic conductivity of the lake sediments is low, infiltration only occurs close to the more permeable shore zones. Using multiple environmental tracer methods, a strong vertical age stratification of the bank filtrate could be shown. travel times are generally long and vary throughout the upper aquifers from a few months near the ground surface to several decades in greater depth. infiltration is mostly anoxic and redox zones were found to be vertically stratified too, becoming more reducing with depth. because berlin’s watercourses contain a proportion of treated municipal sewage a number of wastewater residues, e. g. pharmaceutical residues, were detected in surface water and groundwater. While the majority of the pharmaceutical residues studied were efficiently removed during underground passage, some substances (aMDOPh, primidone and carbamazepine) were found to be very persistent.

Berlin relies on induced bank filtration from a broad-scale, lake-type surface water system. Because the surface water contains treated sewage, wastewater residues are present in surface water and groundwater. Multiple environmental tracers, including tritium and helium isotopes (3H, 3He, 4He), stable isotopes (d18O and d2H) and a number of persistent sewage indicators, such as chloride, boron and a selection of pharmaceutical residues (phenazone-type analgesics and their metabolites, carbamazepine and anthropogenic gadolinium, Gdexcess), were used to estimate travel times from the surface water to individual production and observation wells at two sites. The study revealed a strong vertical age stratification throughout the upper aquifer, with travel times varying from a few months to several decades in greater depth. Whereas the shallow bank filtrate is characterized by the reflection of the time-variant tracer input concentrations and young 3H/3He ages, the deeper, older bank filtrate displays no tracer seasonality, 3H/3He ages of a few years to decades and strongly deviating concentrations of several pharmaceutical residues, reflecting concentrations of the source surface water over time. The phenazone-type pharmaceuticals persist in the aquatic environments for decades. Bank filtration in Berlin is only possible at the sandy lakeshores. In greater water depth, impermeable lacustrine sapropels inhibit infiltration. The young bank filtrate originates from the nearest shore, whereas the older bank filtrate infiltrates at more distant shores. This paper illustrates the importance of using multiple tracer methods, capable of resolving a broad range of residence times, to gain a comprehensive understanding of time-scales and infiltration characteristics in a bank filtration system.

The aim of the study was to calculate mixing proportions of treated wastewater in the surface water and production wells during bank filtration as well as the travel times to observation and abstraction wells. For this purpose, a variety of tracers such as the stable isotopes deuterium (D) and 18O and several wastewater indicators like chloride, EDTA (ethylenediaminetetraacetic acid), boron and the rare earth element (REE) gadolinium (Gd) are used and compared to each other. Time series measurements in the surface water could be traced back in bank filtrates and raw water. Gd-DTPA was found to be a useful sewage indicator, even though it is biodegradable at favourable conditions at very slow rates. The travel times of the bank filtrates were obtained by the analysis of the peak shift in time-series of the tracer. Most tracers were found to be applicable but best results were obtained with the stable isotopes.

Hydrochemical conditions were evaluated at both bank filtration and artificial recharge sites in Berlin. All bank filtration sites show a strong vertical age stratification. Rather than showing a typical redox zoning with more reducing conditions in greater distance from the surface water, the redox zones are horizontally layered, with more reducing conditions in greater depth. This is believed to be an effect of the strongly alternating groundwaterlevels and by the age stratification. The redox conditions are generally more reducing at the bank filtration sites, mainly as a result of the longer travel times and operational differences. Redox conditions at all sites vary seasonally in particular at the artificial recharge site, which is mainly caused by temperature changes.

The redox conditions below an artificial recharge pond in Berlin were largely dependent on seasonal temperature changes of 0-24 °C in the infiltrate. Aerobic conditions prevailed in winter, when temperatures were low, while anaerobic conditions were reached below the pond when temperatures exceeded 14 °C. In contrast to temperature changes, cyclic changes between saturated or unsaturated conditions below the pond had only a minor effect on the redox conditions. However, the intrusion of gaseous oxygen during unsaturated conditions caused a temporary reinforced increase in oxidation of particulate organic matter. The effect of variable redox conditions on the behaviour of a number of pharmaceutically active compounds, namely carbamazepine, phenazone and several phenazone-type PhACs, was investigated. Phenazone is redox sensitive and was generally fully degraded before reaching the first groundwater well, as long as oxygen was present. When conditions turned anaerobic, phenazone was not fully eliminated. 1-Acetyl-1-methyl-2-dimethyl-oxymoyl-2-phenylhydrazide (AMDOPH) and carbamazepine are very persistent drug residues. However, results suggest that AMDOPH may be slightly degradable under aerobic conditions too, but further studies will be needed to verify this statement.

Bank filtration and artificial ground water recharge are important, effective, and cheap techniques for surface water treatment and removal of microbes, as well as inorganic, and some organic, contaminants. Nevertheless, physical, chemical, and biological processes of the removal of impurities are not understood sufficiently. A research project titled Natural and Artificial Systems for Recharge and Infiltration attempts to provide more clarity in the processes affecting the removal of these contaminants. The project focuses on the fate and transport of selected emerging contaminants during bank filtration at two transects in Berlin, Germany. Several detections of pharmaceutically active compounds (PhACs) in ground water samples from bank filtration sites in Germany led to furthering research on the removal of these compounds during bank filtration. In this study, six PhACs including the analgesic drugs diclofenac and propyphenazone, the antiepileptic drugs carbamazepine and primidone, and the drug metabolites clofibric acid and 1-acetyl–1-methyl–2-dimethyloxamoyl– 2-phenylhydrazide were found to leach from the contaminated streams and lakes into the ground water. These compounds were also detected at low concentrations in receiving public supply wells. Bank filtration either decreased the concentrations by dilution (e.g., for carbamazepine and primidone) and partial removal (e.g., for diclofenac), or totally removed PhACs (e.g., bezafibrate, indomethacine, antibiotics, and estrogens). Several PhACs, such as carbamazepine and especially primidone, were readily transported during bank filtration. They are thought to be good indicators for evaluating whether surface water is impacted by contamination from municipal sewage effluent or whether contamination associated with sewage effluent can be transported into ground water at ground water recharge sites.

In Berlin, 70 % of the drinkinq water is derived from bank filtrate or artificially recharged water. Because the surface water system contains elevated proportions of secondary treated municipal sewage, a number of sewage indicators from various sources can be detected in the bank filtrate. An artificial recharge site and a bank filtration site in Berlin Tegel are introduced and compared in terms of their hydrogeological and hydrochemical properties. Because of a permanent clogging layer and the geological properties, travel times are slower at the BF site and the hydrochemical conditions are more reducing. First estimates for the reaction rate constants of oxygen and nitrate are obtained with exponential data fitting. Some of the effects of the different redox conditions on minor substances such as drug residues are highlighted.

Induced by well abstraction, surface water infiltrates into Berlin aquifers and is used for drinking water production. A major advantage of bank filtration is the capability of the subsurface to remove contaminants and save natural groundwater resources. Since a large proportion of the surface water in Berlin originates from treated effluents released by wastewater treatment plants, certain wastewater residues can be traced into the groundwater. A powerful tool to characterise bank filtration systems is the use of wastewater indicators and additional environmental tracers to estimate flow velocities and proportions of bank filtrate in the abstraction wells prior to reactive transport evaluations. Examples for tracer applications at the Berlin system are introduced in this paper. In addition, an overview on results of various studies conducted on contaminant transport and removal during underground passage of the bank filtrate in Berlin is given.