The aim of the present study was to estimate the performance of slow sand filtration (SSF) facilities, including the time needed for reaching stabilization (maturation), operated with surface water bearing high fecal contamination, representing realistic conditions of rivers in many emerging countries. Surface water spiked with wastewater was infiltrated at different pore water velocities (PWV) and samples were collected at different migration distances. The samples were analyzed for phages and to a lesser extent for fecal bacteria and enteric adenoviruses. At the PWV of 50 cm/d, at which somatic phages showed highest removal, their mean log10 removal after 90 cm migration was 3.2. No substantial differences of removal rates were observed at PWVs between 100 and 900 cm/d (2.3 log10 mean removal). The log10 mean removal of somatic phages was less than the observed for fecal bacteria and tended more towards that of enteric adenoviruses This makes somatic phages a potentially better process indicator than Escherichia coli for the removal of viruses in SSF. We conclude that SSF, and by inference in larger scale river bank filtration (RBF), is an excellent option as a component in multi-barrier systems for drinking water treatment also in areas where the sources of raw water are considerably fecally polluted, as often found in many emerging countries.

The microbial degradation of pharmaceuticals found in surface water used for artificial recharge is strongly dependent on redox conditions of the subsurface. Furthermore the durability of production wells may decrease considerably with the presence of oxygen and ferrous iron due to the precipitation of trivalent iron oxides and subsequent clogging. Field measurements are presented for oxygen at a bank filtration site in Berlin, Germany, along with simplified calculations of different oxygen pathways into the groundwater. For a twodimensional vertical cross-section, oxygen input has been calculated for six scenarios related to different water management strategies. Calculations were carried out in order to assess the amount of oxygen input due to (1) the infiltration of oxic lake water, (2) air entrapment as a result of water table oscillations, (3) diffusive oxygen flux from soil air and (4) infiltrating rainwater. The results show that air entrapment and infiltrating lake water during winter constitute by far the most important mechanism of oxygen input. Oxygen input by percolating rainwater and by diffusive delivery of oxygen in the gas phase is negligible. The results exemplify the importance of well management as a determining factor for water oscillations and redox conditions during artificial recharge.

Large-scale dispersion was studied in an unconsolidated, sandy, glaciofluvial, river-recharged, and confined aquifer in Germany. Groundwater observation wells from a 3.5-km-long transect located in flow direction from the river Oder into a large lowland area (Oderbruch polder) were sampled for noble gases in order to date the groundwater with the tritium and helium (3H-3He) technique. The apparent 3H-3He ages of the groundwater increased from only a few months to >40 years along the flow path. Highest values for initial 3H (sum of 3H and its decay product tritiogenic helium, 3Hetri) were encountered in 2.6-km river distance. Concentrations of 4He in the water increased to 1.1 × 10-7 cm3 STP/g with distance from the river. The initial 3H data enabled an estimation of the longitudinal dispersivity with a simplified one-dimensional transport model. The best fit of modeled and measured initial H data was obtained using a dispersivity of 120 m. Deviations of modeled hydraulic ages and measured apparent 3H- 3He ages for older samples can be explained by dispersive mixing.

The behaviour of residues of phenazone-type pharmaceuticals during bank filtration was investigated at a field site in Berlin, Germany, where bank-filtered water is used for drinking water production. The concentrations of the pharmaceutical residues in the shallow, young bank filtrate (travel times < one month) were correlated to the prevailing hydrochemical conditions at the field site. In addition, their behaviour during passage through an undisturbed sediment core from the lake base at the site (clogging layer) was evaluated in the laboratory. Phenazone, 4-acetylaminoantipyrine (AAA), 4-formylaminoantipyrin (FAA) and 1,5-dimethyl-1,2-dehydro-3-pyrazolone (DP) were eliminated more efficiently under oxic conditions, while 1-acetyl-1-methyl-2-dimethyloxamoyl-2-phenylhydrazide (AMDOPH) was not eliminated at all. The redox conditions and the elimination of the respective pharmaceutical residues displayed strong seasonal variations. Oxic conditions were only encountered close to the shore in winter, when temperatures were low. The column study showed that the elimination is restricted to the uppermost decimetres of the lake base, where oxygen is present. While phenazone elimination is almost complete during aerobic rapid sand filtration in the waterworks, the compounds were found to be more persistent under anoxic field conditions.

Managed aquifer recharge is gaining importance as a practice to bank and treat surface water for drinking water production. Neon (Ne) concentrations were analysed at four different recharge sites in and near Berlin, where groundwater is recharged directly from surface water courses, either by near-natural bank filtration, induced bank filtration or engineered basin recharge. Neon concentrations in excess of saturation (DNe) were used to identify excess air in the infiltrates. Excess air concentrations were around saturation at the near-natural bank filtration site, where river water infiltrates through a permeable river bed into a confined aquifer under completely saturated conditions. At two induced unconfined bank filtration sites, samples generally contained excess air (up to 60% DNe). Highest excess air concentrations (up to 81% DNe) were encountered at the engineered basin recharge site. The degree of water table fluctuations, the water saturation of the sediments in the infiltration zone and the presence of a confining layer affect the formation of excess air. Excess air can only be used to trace bank filtrate or artificially recharged water in a setting where the ambient groundwater in the near vicinity of production wells is not affected by large water-table fluctuations. Nevertheless, excess air concentrations provide valuable additional information on the type of recharge (saturated or unsaturated, degree of water table fluctuations).

The behaviour of residues of antibiotic drugs during bank filtration was studied at a field site in Berlin, Germany, where bank-filtered water is used for the production of drinking water. The neighbouring surface water used for bank filtration is under the influence of treated municipal wastewater. Seven out of 19 investigated antimicrobial residues were found in the surface water with median concentrations between 7 and 151 ng L¡1. Out of the seven analytes detected in the surface water only three (anhydroerythromycin, clindamycin and sulfamethoxazole) were found with median concentrations above their limits of quantitation in bank filtrate with a travel time of one month or less. With the exception of sulfamethoxazole, none of the 19 analytes were present in bank filtrate with a residence time larger than one month or in the water-supply well itself. Sulfamethoxazole found with a median concentration of 151 ng L¡1 in the surface water was the most persistent of all antimicrobial residues. Nevertheless, it was also removed by more than 98% and only found with a median concentration of 2 ng L¡1 in the water-supply well. The degradation of clindamycin and sulfamethoxazole appear to be redox-dependent. Clindamycin was eliminated more effi­ciently under oxic infiltration conditions while sulfamethoxazole was eliminated more rapidly under anoxic infiltration conditions. A slight preference for an improved degradation under oxic (clarithromycin and roxithromycin) or anoxic (anhydroerythromycin) conditions was also observed for the macrolide antibiotics. Nevertheless, all macrolides were readily removable by bank filtration both under oxic and anoxic conditions.

Investigations on the behavior of bulk organics and trace organic compounds in a temperated soil column system are reported. Objective of the research was to assess the importance of temperature for the degradation of bulk and trace organics. The analysis of the bulk organic behavior showed a fast mineralization of easy degradable organic carbon in the first few centimeters of the columns, which does not seem to be temperature-dependent. Along the further infiltration path an influence of the different temperatures on the bioactivity was clearly visible. However, a significant increase of mineralization potential of bulk organic compounds with increasing temperature was shown. The monitoring of the single organic pollutants Iopromide, Sulfamethoxazole and naphthalenedisulfonic acids showed that temperature has an influence on the degradation behavior of the monitored compounds. In most cases higher temperatures increased the mineralization potential.

Cyanobacterial toxins are toxic substances produced by cyanobacteria or blue-green-algae. The can occur in surface waters wordwide and have to be removed sufficiently when using infested surface waters as drinking water source. Bank filtration has been used since 150 years for drinking water (pre-)treatment and utilizes natural elimination processes like sorption and degardation in the sub-surface. During several research projects the German Federal Environmental Agency (Umweltbundesamt, UBA) carried out field investigations and experiments in different scales in order to assess under which conditions secure elimination of microcystins (the most common group of cyanobacterial toxins) takes place. Filtration of cells on the sediment surface is the most prominent process for eliminating the primarily cell-bound toxins. Middle to coarsly grained sands eliminated more than 99.9 % of intracellular toxins within the first 10 centimeters. Elimination of extracellular microcystin during underground passage is mainly due to biodegradation. Reversible adsorption processes do not reduce the total load but lead to longer contact times for extended biodegradation. Laboratory experiments showed that high clay and silt content is crucial for maximum adsorption. However, redox conditions play an important role for degradation rates: under aerobic conditions half lives of less than one day occured frequently whereas anoxic conditions resulted in lag phases of one day and more as well as in half lives of up to 25 days. Field experiments could show that temperature is crucial for degradation velocity under natural conditions.

In Berlin – wie auch in anderen Regionen Deutschlands – wird ein Großteil des Trinkwassers durch Uferfiltration gewonnen. Durch eine Untergrundpassage mit einer Dauer von meist mehreren Wochen erhält es eine Aufreinigung, die den Aufwand der konventionellen Trinkwasseraufbereitung verringert und eine zusätzliche Barriere gegenüber Schadstoffen darstellt (Kühn 2001). Das Ziel eines interdisziplinären Forschungsvorhabens mit dem Titel NASRI (Natural and Artificial Systems for Recharge and Infiltration) war, die Reinigungsprozesse für verschiedenste Substanzen zu ermitteln und Empfehlungen für das zukünftige Wassermanagement in Berlin abzuleiten (Fritz 2003). Aufgabe der Arbeitsgruppe des Umweltbundesamtes war dabei zu klären, wie wirksam Microcystinen (MCYST) als wichtigste Gruppe der Cyanobakterientoxine durch die Bodenpassage eliminiert werden. Im Folgenden werden einige Schlüsselergebnisse berichtet. Für eine ausführliche Ergebnisdarstellung siehe Grützmacher et al. (2006). MCYST sind in der Regel überwiegend (> 90 %) zellgebunden, so dass die physikalische Filtration der Zellen an der Sedimentoberfläche als Eliminationsprozess im Vordergrund steht (Grützmacher et al. 2003). Das extrazelluläre MCYST wird dagegen überwiegend biologisch abgebaut (Lahti et al. 1998, Grützmacher et al. 2005a). Um unter naturnahen Bedingungen Extremfälle für den biologischen Abbau zu simulieren, wurden i) Freilandversuche unter variierenden Redoxbedingungen und ii) Laborsäulenversuche bei unterschiedlichen Temperaturen durchgeführt. Ferner wurde die Freisetzung von MCYST aus sedimentierten Zellen untersucht.