Within the last 50 years, the world’s annual production of organic chemicals, such as pesticides, solvents, cleansing agents and drugs has increased from 7.5 Megatons to 300 Megatons, and more and more of these compounds are detected at trace concentrations in the aquatic environment, including the groundwater (Scheffer and Schachtschabel, 2002). This demonstrates that anthropogenic pollution of freshwater resources is one of the most challenging problems of the modern world, which has to be faced by the human race in the coming decades (UNEP, 2002).

Since groundwater is a major source for drinking water and also compensates agricultural and industrial water demands, artificial groundwater recharge will be increasingly important for effective water resource management, safety and continuity in water supply (Martijn, 1998). Artificial recharge of groundwater includes a variety of techniques, such as riverbank filtration (e.g., Sontheimer, 1980; Kühn and Müller, 2000; Ray et al., 2002), aquifer storage and recovery (ASR, Pyne, 1995; Herczeg et al., 2004; Dillon et al., 2005), deep well injection (e.g., Stuyfzand, 1998) and infiltration ponds (Asano, 1992; Bouwer, 2002). All over the world, numerous artificial recharge applications already exist or are currently under investigation. The following list gives a brief overview of the broad range of environmental and technical problems for which artificial recharge of groundwater has already been applied or planned:

• Artificial recharge of pre-treated Rhine water in the coastal dune area west of Amsterdam prevents further sea water upconing that resulted from excessive groundwater extraction for drinking water supply (van Breukelen et al., 1998).

• ASR of treated imported water is considered to avoid further land subsidence in the city of Lancaster (California), which was induced by excessive groundwater production since the 1920s, causing increased erosion, flooding and structural damages (Phillips et al., 2002).

• Storm and reclaimed water ASR in South Australia is intended to store excess water in the wet season for the supply of irrigation water in the dry season (Dillon et al., 1999; Vanderzalm, 2004).

• Deep well injection of surface water in the Netherlands is proposed to recharge confined aquifers in order to compensate the drawdown of groundwater tables due to increasing drinking water demands (Stuyfzand, 1998).

• So-called soil-aquifer treatment (SAT), i.e., quality improvement of pre-treated wastewater by infiltration in recharge ponds and its subsequent downgradient recovery through extraction wells is very popular in many dry regions of the world (e.g., Idelovitch and Michail, 1984; Fox, 2002). The feasibility of this wastewater reuse technique was demonstrated in the prominent Flushing Meadows Project installed west of Phoenix, USA in 1967 (Bouwer et al., 1980).

• Riverbank filtration and ponded infiltration has been used for pre-treatment of surface water for more than hundred years in Europe (Kühn and Müller, 2000; Hiscock and Grischek, 2002). For instance, about 16% of the drinking water in Germany is provided by riverbank filtration and ponded infiltration (Kühn and Müller, 2000), which is applied mainly in densely populated urban areas, e.g., the City of Berlin, where about 70% of the drinking water production depends on these techniques (Massmann et al., 2004a).

During infiltration and subsurface flow, the artificially recharged water usually undergoes quality changes. In general, they result from the interaction of physical (filtration, advection and dispersion), chemical (sorption, precipitation/dissolution) and biological (biodegradation) processes (e.g., Jacobs et al., 1988; von Gunten et al., 1991; Bourg and Bertin, 1993, 1994; von Gunten and Zobrist, 1993; Hiscock and Grischek, 2002). The geochemical changes in a hydrogeological system are often to a large extent triggered and determined by microbial mediated redox reactions (e.g., Eckert and Appelo, 2002; Massmann et al, 2004b). These redox reactions are driven by the biodegradation of dissolved and/or sediment-bound organic matter and involve the consumption of so-called terminal electron acceptors (TEA), such as O₂, NO₃ ^-, manganese- and iron-(hydro)oxides, SO₄ ^2- and dissolved inorganic carbon (e.g., Lovley and Phillips, 1988; McMahon and Chapelle, 1991; Chapelle and Lovley, 1992; Lovley et al., 1994; Chapelle et al., 1995). Typically, the consumption of the TEA occurs in a sequential order constrained by thermodynamic principles (e.g., Champ et al., 1979; Stumm and Morgan, 1996; Postma and Jakobsen, 1996; Christensen et al., 2000). With decreasing of Gibbs free energy, this redox sequence theoretically starts with aerobic respiration and is subsequently followed by denitrification, manganese-, iron- and sulphate reduction and finally methanogenesis (Figure 1.1a). In a groundwater flow system, this leads to the formation of spatially distinct redox zones along the flow direction (Champ et al., 1979; Matsunaga et al., 1993; von Gunten and Zobrist, 1993; Bjerg et al., 1995; Amirbahman et al., 2003). This is schematically shown in Figure 1.1b.

Figure 1 a. (adapted from van Breukelen (2003)): Theoretical sequence of terminal electron acceptor (TEA) processes related to the biodegradation of organic matter in hydrogeochemical systems. RRS = reduced redox species. b. Schematic illustration of spatially distinct redox zones (O₂=aerobic-, NO₃=denitrifying-, Mn=Mn-reducing- , Fe=Fe-reducing- and SO₄=sulfate reducing conditions) along the ground water flow direction during riverbank filtration.

The redox reactions often induce additional reactions such as precipitation and dissolution of minerals, ion exchange or surface complexion reactions due to the consumption and production of protons and other reactants (e.g., Eckert and Appelo, 2002). Besides water quality changes, these reactions can also change the hydrogeological properties of the aquifer such as porosity and hydraulic conductivity (Vandevivere and Baveye, 1992; Rinck-Pfeiffer et al., 2000) as well as the geochemical properties such as mineral reactivity (Postma and Jakobsen, 1996; Larsen and Postma, 2001) and sorption capacity (Dzombak and Morel, 1990, Stumm and Morgan, 1995). Thereby, the latter is of high importance, since it can affect the mobility of trace metals such as Cd, Ni, Zn or As (e.g., Jacobs et al., 1988; Appelo and de Vet, 2002).

Natural degradation processes (i.e., biodegradation and sorption) can help to remove dissolved (trace) organic compounds as well as pathogenic bacteria and viruses, and thus are of importance for the protection of ground- and drinking water quality in artificial recharge systems (Bosma et al., 1996). However, the nature and reaction rates of these processes are strongly determined by physical and chemical conditions. For instance, the biodegradability of many organic micropollutants, such as pharmaceutically-active compounds (PhAC’s, Holm et al., 1995; Khan and Rorije, 2002; Ternes et al., 2004 and references therein), pesticides (Agertved et al. 1992; Tuxen et al., 2000; Broholm et al., 2001a,b) or adsorbing organic halogens (AOX, Ziegler et al., 2002) and other halogenated organic compounds (Bouwer and McCarty, 1983; Gupta et al., 1996; Bosma et al., 1996) is affected by locally prevailing redox conditions. Moreover, the fate of pathogenic microorganisms is often influenced by water quality parameters such as pH, redox state, ionic strength and divalent cation concentration (Schijven, et al., 2000; Zhuang et al., 2003).

A major aim of the research on artificial recharge of groundwater is to understand the potential key factors that control the water quality changes occurring between recharge/infiltration and recovery. The influence of those factors can be very different, depending on field site characteristics such as recharge water quality, composition of aquifer matrix, subsurface residence time or hydraulic conditions. In order to assess these factors and quantify their influence on the water quality changes at a specific field site, a complex understanding of the interactions of hydraulic, hydrogeochemical and microbial processes is necessary. Since these (non-linear) interactions are generally not intuitive (Amos et al., 2004), the identification and quantification of key processes by interpreting collected field data is difficult and often cannot be clarified with confidence. Interpretation is even trickier when the hydraulic and hydrochemical boundary conditions such as recharge/recovery rate and recharge water quality are highly transient. Unfortunately, such conditions are the rule rather than the exception in artificial recharge systems (e.g., von Gunten et al., 1991, 1994; Bourg and Bertin, 1994; Bouwer, 2002; Vanderzalm, 2004). Thus, for a thorough analysis of such complex hydrogeochemical systems, simultaneous high-resolution monitoring of both hydraulic and hydrochemical parameters in space and time is fundamental.

Typically, the acquisition of spatially and temporally dense information, especially in the field, is quite expensive and so records are generally sparse. Integrated mathematical modelling is becoming increasingly popular to (i) fill the gaps between the measured data and (ii) assist in analysing and understanding complex environmental systems (Prommer and Barry, 2005). In the past decade, multi-component reactive transport modelling has been used to support the identification of the relevant hydrogeochemical processes in artificial recharge systems both at laboratory (Matsunaga et al., 1993; Amirbahman et al., 2003) and field scale (Valocchi et al., 1981; Lensing et al., 1994; Doussan et al., 1997; Parkhurst and Petkewich, 2002; Appelo and de Vet; 2002; Prommer and Stuyfzand, 2005). However, most of these modelling studies only regarded steady-state hydraulic and/or hydrogeochemical boundary conditions and also only addressed a sparse set of possible hydrogeochemical interactions. It appears that there is still a lack of reactive transport models applied to more complicated field situations, which, as already pointed out, are the rule rather than the exception.

Moreover, the importance of redox conditions for the fate of organic micropollutants was highlighted in a number of field site investigations (e.g., Holm et al., 1995; Agertved et al. 1992; Broholm et al., 2001a,b; Bosma et al., 1996; Grünheid et al., 2005; Massmann et al., 2005a,b). Studies incorporating this knowledge into process-based, field scale modelling frameworks would be extremely helpful in supporting these findings and evaluating their importance also in relation to the transient hydraulic and hydrochemical boundary conditions. To date, such mechanistic modelling studies do not seem to exist in the literature.

Conclusively, for a sound understanding of how the interacting processes, especially under transient boundary conditions, affect the water quality changes during artificial recharge of groundwater, there is an urgent need for more detailed field site and modelling investigations.

In this work, two specific artificial recharge sites were investigated for the relevant hydraulic and hydrogeochemical processes. The research was conducted by three separate studies described in Chapter 2, 3 and 4. The specific objectives were as follows:

• Providing a process-based modelling framework that integrates and interprets the hydrogeochemical data colleted during a reclaimed water ASR experiment at Bolivar, South Australia. Thereby, the focus was to understand the fate of the various forms of organic and inorganic carbon, and the role of biomass dynamics.

• Understanding the impact of transient saturated and unsaturated conditions directly below an artificial recharge pond on the dynamics of hydrogeochemical changes during infiltration. This field site investigation was carried out at a recharge pond in Berlin, Germany.

• Providing a process-based modelling framework to clarify the seasonal degradation behaviour of the redox-dependent pharmaceutical residue phenazone within the aquifer surrounding the investigated recharge pond.

• Exploring the capabilities of multi-component reactive transport modelling to interpret field observations in artificial recharge systems, and to reveal and quantify the complex non-intuitive interactions of the controlling processes in those systems.

The different studies of this work were carried out with the help of (i) field site methods, i.e., monitoring of hydraulic and hydrochemical data as well as measurements of the aquifer properties and (ii) modelling methods, i.e., hydraulic and multi-component reactive transport modelling of the investigated hydrogeochemical systems. The specific methods used in this thesis are described in detail in the following chapters.

This thesis comprises this introduction, three separate pieces of research work, synthesis and appendices. The general introduction (Chapter 1) gives a brief overview of artificial groundwater recharge, fundamental processes that occur in these systems, the state-of-the-art research background, remaining problems and the scope of this thesis. The main research work is described in Chapters 2, 3 and 4. Written as manuscripts for publication in peer-reviewed journals, each of these chapters can be read independently as a stand-alone piece of research, including introduction, methodology, results and discussion, and conclusions. They are as follows:

Chapter 2 is published as:

Greskowiak, J., Prommer, H., Vanderzalm, J., Pavelic, P., Dillon, P. (2005), Modelling of carbon cycling and biogeochemical changes during injection and recovery of reclaimed water at Bolivar, South Australia, Water Resources Research, 41(10), W10418, doi:10.1029/2005WR004095.

Chapter 3 is published as:

Greskowiak, J., Prommer H., Massmann, G., Johnston, C. D., Nützmann, G., Pekdeger, A. (2005), The Impact of variably saturated conditions on hydrogeochemical changes during artificial recharge of groundwater, Applied Geochemistry, 20(7), 1409-1426.

Chapter 4 is published as:

Greskowiak, J., Prommer H., Massmann, G., Nützmann, G. (2006), Modeling seasonal redox dynamics and the corresponding fate of the pharmaceutical residue phenazone during artificial recharge of groundwater, Environmental Science and Technology, doi: 10.1021/es052506t.

In all three cases I have carried out the research work described and authored the paper manuscripts. The co-authors have played advisory or supervisory roles. In chapters 2-4 a limited amount of repetition of introductory information can be found, which was necessary for the autonomy of the papers.

Chapter 5 summarises the conclusions of this thesis and provides recommendations for future research.

Appendix A: Conference paper supporting the findings of Chapter 2:

Greskowiak J., Prommer H., Vanderzalm J. L., Pavelic P., Dillon P. J. (2005), Quantifying Biogeochemical Changes during ASR of Reclaimed Water at Bolivar, South Australia, Proceedings of 5 th International Symposium on Management of Aquifer Recharge, Berlin, June 2005, in press.

Appendix B: Conference paper extending the results found in Chapter 3:

Greskowiak J., Massmann G., Nützmann G., Prommer H. (2005), Hydrogeochemical Changes of Seepage Water during Artificial Recharge of Groundwater in Berlin, Germany, Proceedings of 5 th International Symposium on Management of Aquifer Recharge, Berlin, June 2005, in press.

Appendix C: Additional Figures and Tables, which in the original manuscripts (Chapters 2-4) were declared as not shown.