Process-Based and Probabilistic Quantification of Co and Ni Mobilization Risks Induced by Managed Aquifer Recharge.

Managed aquifer recharge (MAR) is an increasingly used water management technique that enhances water availability while commonly generating water quality benefits. However, MAR activities may also trigger adverse geochemical reactions, especially during the injection of oxidant-enriched waters into reducing aquifers. Where this occurs, the environmental risks and the viability of mitigating them must be well understood. Here, we develop a rigorous approach for assessing and managing the risks from MAR-induced metal mobilization. First, we develop a process-based reactive transport model to identify and quantify the main hydrogeochemical drivers that control the release of metals and their mobility. We then apply a probabilistic framework to interrogate the inherent uncertainty associated with adjustable model parameters and consider this uncertainty (i) in long-term predictions of groundwater quality changes and (ii) in scenarios that investigate the effectiveness of modifications in the water treatment process to mitigate metal release and mobility. The results suggested that Co, Ni, Zn, and Mn were comobilized during pyrite oxidation and that metal mobility was controlled (i) by the sediment pH buffering capacity and (ii) by the sorption capacity of the native aquifer sediments. Both tested mitigation strategies were shown to be effective at reducing the risk of elevated metal concentrations.

Sustaining Irrigation Supplies through Immobilization of Groundwater Arsenic In Situ.

Geogenic arsenic (As) in groundwater is widespread, affecting drinking water and irrigation supplies globally, with food security and safety concerns on the rise. Here, we present push-pull tests that demonstrate field-scale As immobilization through the injection of small amounts of ferrous iron (Fe) and nitrate, two readily available agricultural fertilizers. Such injections into an aquifer with As-rich (200 ± 52 µg/L) reducing groundwater led to the formation of a regenerable As reactive filter in situ, producing 15 m3 of groundwater meeting the irrigation water quality standard of 50 µg/L. Concurrently, sediment magnetic properties were markedly enhanced around the well screen, pointing to neo-formed magnetite-like minerals. A reactive transport modeling approach was used to quantitatively evaluate the experimental observations and assess potential strategies for larger-scale implementation. The modeling results demonstrate that As removal was primarily achieved by adsorption onto neo-formed minerals and that an increased adsorption site density coincides with the finer-grained textures of the target aquifer. Up-scaled model simulations with 80-fold more Fe-nitrate reactants suggest that enough As-safe water can be produced to irrigate 1000 m2 of arid land for one season of water-intense rice cultivation at a low cost without causing undue contamination in surface soils that threatens agricultural sustainability.

Molybdenum Mobility During Managed Aquifer Recharge in Carbonate Aquifers.

The mobility of molybdenum (Mo) in groundwater systems has received little attention, although a high intake of Mo is known to be detrimental to human and animal health. Here, we used a comprehensive hydrochemical data set collected during a multi-cycle aquifer storage and recovery test to study the mechanisms that control the mobility of Mo under spatially and temporally varying hydrochemical conditions. The model-based interpretation of the data indicated that the initial mobilization of Mo occurs as a sequence of reactions, in which (i) the aerobic injectant induces pyrite oxidation, (ii) the released acidity is partially buffered by the dissolution of dolomite that (iii) leads to the release of Mo with highly soluble sulfurized organic matter prevailing between the intercrystalline spaces of the dolomite matrix or incorporated in dolomite crystals. Once released, Mo mobility was primarily controlled by pH-dependent surface complexation reactions to the sediments and, to a lesser extent, the capture by iron sulfides (FeS). In the studied system, Mo mobilization could be effectively mitigated by reducing or eliminating pyrite oxidation, which decreases the likelihood of dolomite dissolution and associated Mo release.

Molybdenum Release Triggered by Dolomite Dissolution: Experimental Evidence and Conceptual Model.

The injection of oxygenated water into anoxic aquifers during managed aquifer recharge (MAR) can cause the mobilization of metal(loid)s. Here, we study the processes controlling MAR-induced molybdenum (Mo) release in dolomitic aquifers. Sequential chemical extractions and energy dispersive X-ray spectroscopy combined with scanning electron microscopy point to an association of Mo with easily soluble sulfurized organic matter present in intercrystalline spaces of dolomites or directly incorporated within dolomite crystals. The easily soluble character was confirmed by a batch experiment that demonstrated the rapid mobilization of Mo, dissolved organic carbon, and sulfur. The type and time of batch solution contact with the sulfurized organic matter impacted the release of Mo, as demonstrated by a 36% increase in Mo concentrations when shaking was intensified. Based on the experimental results, a conceptual model for the release of Mo was formulated, where (i) the injection of oxygenated water causes the oxidation of pyrite in the aquifer matrix, and (ii) the associated release of protons (H+) induces the dissolution of dolomite as a buffering reaction, which (iii) enhances the accessibility of the injectant to intercrystalline and incorporated sulfurized organic matter within dolomite, causing the release of Mo.

In situ arsenic immobilisation for coastal aquifers using stimulated iron cycling: Lab-based viability assessment.

Arsenic (As) is one of the most harmful and widespread groundwater contaminants globally. Besides the occurrence of geogenic As pollution, there is also a large number of sites that have been polluted by anthropogenic activities, with many of those requiring active remediation to reduce their environmental impact. Cost-effective remedial strategies are however still sorely needed. At the laboratory-scale in situ formation of magnetite through the joint addition of nitrate and Fe(II) has shown to be a promising new technique. However, its applicability under a wider range of environmental conditions still needs to be assessed. Here we use sediment and groundwater from a severely polluted coastal aquifer and explore the efficiency of nitrate-Fe(II) treatments in mitigating dissolved As concentrations. In selected experiments >99% of dissolved As was removed, compared to unamended controls, and maintained upon addition of lactate, a labile organic carbon source. Pre- and post experimental characterisation of iron (Fe) mineral phases suggested a >90% loss of amorphous Fe oxides in favour of increased crystalline, recalcitrant oxide and sulfide phases. Magnetite formation did not occur via the nitrate-dependent oxidation of the amended Fe(II) as originally expected. Instead, magnetite is thought to have formed by the Fe(II)-catalysed transformation of pre-existing amorphous and crystalline Fe oxides. The extent of amorphous and crystalline Fe oxide transformation was then limited by the exhaustion of dissolved Fe(II). Elevated phosphate concentrations lowered the treatment efficacy indicating joint removal of phosphate is necessary for maximum impact. The remedial efficiency was not impacted by varying salinities, thus rendering the tested approach a viable remediation method for coastal aquifers.

Mobilization of Arsenic and Other Naturally Occurring Contaminants during Managed Aquifer Recharge: A Critical Review.

Population growth and climate variability highlight the need to enhance freshwater security and diversify water supplies. Subsurface storage of water in depleted aquifers is increasingly used globally to alleviate disparities in water supply and demand often caused by climate extremes including floods and droughts. Managed aquifer recharge (MAR) stores excess water supplies during wet periods via infiltration into shallow underlying aquifers or direct injection into deep aquifers for recovery during dry seasons. Additionally, MAR can be designed to improve recharge water quality, particularly in the case of soil aquifer treatment and riverbank filtration. While there are many potential benefits to MAR, introduction of recharge water can alter the native geochemical and hydrological conditions in the receiving aquifer, potentially mobilizing toxic, naturally occurring (geogenic) contaminants from sediments into groundwater where they pose a much larger threat to human and ecosystem health. On the basis of the present literature, arsenic poses the most widespread challenge at MAR sites due to its ubiquity in subsurface sediments and toxicity at trace concentrations. Other geogenic contaminants of concern include fluoride, molybdenum, manganese, and iron. Water quality degradation threatens the viability of some MAR projects with several sites abandoning operations due to arsenic or other contaminant mobilization. Here, we provide a critical review of studies that have uncovered the geochemical and hydrological mechanisms controlling mobilization of arsenic and other geogenic contaminants at MAR sites worldwide, including both infiltration and injection sites. These mechanisms were evaluated based on site-specific characteristics, including hydrological setting, native aquifer geochemistry, and operational site parameters (e.g., source of recharge water and recharge/recovery cycling). Observed mechanisms of geogenic contaminant mobilization during MAR via injection include shifting redox conditions and, to a lesser extent, pH-promoted desorption, mineral solubility, and competitive ligand exchange. The relative importance of these mechanisms depends on various site-specific, operational parameters, including pretreatment of injection water and duration of injection, storage, and recovery phases. This critical review synthesizes findings across case studies in various geochemical, hydrological, and operational settings to better understand controls on arsenic and other geogenic contaminant mobilization and inform the planning and design of future MAR projects to protect groundwater quality. This critical review concludes with an evaluation of proposed management strategies for geogenic contaminants and identification of knowledge gaps regarding fate and transport of geogenic contaminants during MAR.

Enhancing Roxarsone Degradation and In Situ Arsenic Immobilization Using a Sulfate-Mediated Bioelectrochemical System.

Roxarsone (ROX) is widely used in animal farms, thereby producing organoarsenic-bearing manure/wastewater. ROX cannot be completely degraded and nor can its arsenical metabolites be effectively immobilized during anaerobic digestion, potentially causing arsenic contamination upon discharge to the environment. Herein, we designed and tested a sulfate-mediated bioelectrochemical system (BES) to enhance ROX degradation and in situ immobilization of the released inorganic arsenic. Using our BES (0.5 V voltage and 350 µM sulfate), ROX and its metabolite, 4-hydroxy-3-amino-phenylarsonic acid (HAPA), were completely degraded within 13-22 days. In contrast, the degradation efficiency of ROX and HAPA was <85% during 32-day anaerobic digestion. In a sulfate-mediated BES, 75.0-83.2% of the total arsenic was immobilized in the sludge, significantly more compared to the anaerobic digestion (34.1-57.3%). Our results demonstrate that the combination of sulfate amendment and voltage application exerted a synergetic effect on enhancing HAPA degradation and sulfide-driven arsenic precipitation. This finding was further verified using real swine wastewater. A double-cell BES experiment indicated that As(V) and sulfate were transported from the anode to the cathode chamber and coprecipitated as crystalline alacranite in the cathode chamber. These findings suggest that the sulfate-mediated BES is a promising technique for enhanced arsenic decontamination of organoarsenic-bearing manure/wastewater.

Controlling Arsenic Mobilization during Managed Aquifer Recharge: The Role of Sediment Heterogeneity.

Managed aquifer recharge (MAR) enhances freshwater security and augments local groundwater supplies. However, geochemical and hydrological shifts during MAR can release toxic, geogenic contaminants from sediments to groundwater, threatening the viability of MAR as a water management strategy. Using reactive transport modeling coupled with aquifer analyses and measured water chemistry, we investigate the causal mechanisms of arsenic release during MAR via injection in the Orange County Groundwater Basin. Here, injection water is oxygenated, highly purified recycled water produced by advanced water treatment. Injection occurs via a well screened at several depth intervals ranging from 160-365 m, allowing recharge into multiple confined horizons (zones) of the aquifer system. However, these zones are characterized by varying degrees of prior oxidation due to historic, long-term infiltration from the overlying aquifer. The resulting sediment geochemical heterogeneity provides a critical control on the release (or retention) of arsenic. In zones with prior oxidation, As mobilization occurs via arsenate desorption from Fe-(hydr)oxides, primarily associated with shifts in pH; within zones that remain reduced prior to injection, As release is attributed to the oxidative dissolution of As-bearing pyrite. We find that As release can be attributed to various geochemical mechanisms within a single injection well owing to geochemical heterogeneity across the aquifer system.

Model-Based Analysis of Reactive Transport Processes Governing Fluoride and Phosphate Release and Attenuation during Managed Aquifer Recharge.

In water-scarce areas, the reclamation of wastewater through advanced water treatment and subsequent reinjection into depleted aquifers is an increasingly attractive water management option. However, such injection can trigger a range of water-sediment interactions which need to be well understood and quantified to ensure sustainable operations. In this study, reactive transport modeling was used to analyze and quantify the interacting hydrogeochemical processes controlling the mobilization of fluoride and phosphate during injection of highly treated recycled water into a siliciclastic aquifer. The reactive transport model explained the field-observed fluoride and phosphate transport behavior as a result of the incongruent dissolution of carbonate-rich fluorapatite where (i) a rapid proton exchange reaction primarily released fluoride and calcium, and (ii) equilibrium with a mineral-water interface layer of hydrated dibasic calcium phosphate released phosphate. The modeling results illustrated that net exchange of calcium on cation exchange sites in the sediments postbreakthrough of the injectant was responsible for incongruent mineral dissolution and the associated fluoride and phosphate release. Accordingly, amending calcium chloride into the injectant could potentially reduce fluoride and phosphate mobilization at the study site. Insights from this study are broadly applicable to understanding and preventing geogenic fluoride mobilization from fluoride-bearing apatite minerals in many other aquifers worldwide.

Biodegradability of legacy crude oil contamination in Gulf War damaged groundwater wells in Northern Kuwait.

During the 1991 Gulf War, oil wells in the oil fields of Kuwait were set aflame and destroyed. This resulted in severe crude oil pollution of the countries only fresh water aquifers. Here, for the first time the natural attenuation and biodegradation of the persisting groundwater contamination was investigated to assess potential processes in the aquifer. Biodegradation experiments were conducted under aerobic and multiple anaerobic conditions using microcosms of the contaminated groundwater from Kuwait. Under the conditions tested, a portion of the total petroleum hydrocarbon (TPH) component was degraded, however there was only a slight change in the bulk concentration of the contaminant measured as dissolved organic carbon (DOC), suggesting the presence of a recalcitrant pollutant. Changes in the associated microbial community composition under different reduction-oxidation conditions were observed and known hydrocarbon degraders identified. The results of this study indicate that lingering contaminant still persists in the groundwater and is recalcitrant to further biodegradation, which presents challenges for future remediation plans.

Deoxygenation Prevents Arsenic Mobilization during Deepwell Injection into Sulfide-Bearing Aquifers.

Coal seam gas (CSG) extraction generates large volumes of coproduced water. Injection of the excess water into deep aquifers is often the most sustainable management option. However, such injection risks undesired sediment-water interactions that mobilize metal(loid)s in the receiving aquifer. This risk can be mitigated through pretreatment of the injectant. Here, we conducted a sequence of three push-pull tests (PPTs) where the injectant was pretreated using acid amendment and/or deoxygenation to identify the processes controlling the fate of metal(loid)s and to understand the treatment requirements for large-scale CSG water injection. The injection and recovery cycles were closely monitored, followed by analysis of the observations through reactive transport modeling. While arsenic was mobilized in all three PPTs, significantly lower arsenic concentrations were observed in the recovered water when the injectant was deoxygenated, regardless of pH adjustment. The breakthrough of arsenic was commensurate with molybdenum, but distinct from phosphate. This allowed for the observed and modeled arsenic and molybdenum mobilization to be attributed to a stoichiometric codissolution process during pyrite oxidation, whereas phosphate mobility was governed by sorption. Understanding the nature of these hydrochemical processes explained the greater efficiency of pretreatment by deoxygenation on minimizing metal(loid) mobilization compared to the acid amendment.

Identifying and Quantifying the Intermediate Processes during Nitrate-Dependent Iron(II) Oxidation.

Microbially driven nitrate-dependent iron (Fe) oxidation (NDFO) in subsurface environments has been intensively studied. However, the extent to which Fe(II) oxidation is biologically catalyzed remains unclear because no neutrophilic iron-oxidizing and nitrate reducing autotroph has been isolated to confirm the existence of an enzymatic pathway. While mixotrophic NDFO bacteria have been isolated, understanding the process is complicated by simultaneous abiotic oxidation due to nitrite produced during denitrification. In this study, the relative contributions of biotic and abiotic processes during NDFO were quantified through the compilation and model-based interpretation of previously published experimental data. The kinetics of chemical denitrification by Fe(II) (chemodenitrification) were assessed, and compelling evidence was found for the importance of organic ligands, specifically exopolymeric substances secreted by bacteria, in enhancing abiotic oxidation of Fe(II). However, nitrite alone could not explain the observed magnitude of Fe(II) oxidation, with 60-75% of overall Fe(II) oxidation attributed to an enzymatic pathway for investigated strains: Acidovorax ( A.) strain BoFeN1, 2AN, A. ebreus strain TPSY, Paracoccus denitrificans Pd 1222, and Pseudogulbenkiania sp. strain 2002. By rigorously quantifying the intermediate processes, this study eliminated the potential for abiotic Fe(II) oxidation to be exclusively responsible for NDFO and verified the key contribution from an additional, biological Fe(II) oxidation process catalyzed by NDFO bacteria.

Model-Based Analysis of Arsenic Immobilization via Iron Mineral Transformation under Advective Flows.

Recent laboratory studies have demonstrated that coinjection of nitrate and Fe(II) (as ferrous sulfate) to As-bearing sediments can produce an Fe mineral assemblage containing magnetite capable of immobilizing advected As under a relatively wide range of aquifer conditions. This study combined laboratory findings with process-based numerical modeling approaches, to quantify the observed Fe mineral (trans)formation and concomitant As partitioning dynamics and to assess potential nitrate-Fe(II) remediation strategies for field implementation. The model development was guided by detailed solution and sediment data from our well-controlled column experiment. The modeling results demonstrated that the fate of As during the experiment was primarily driven by ferrihydrite formation and reductive transformation and that different site densities were identified for natural and neoformed ferrihydrite to explain the observations both before and after nitrate-Fe(II) injection. Our results also highlighted that when ferrihydrite was nearing depletion, As immobilization ultimately relied on the presence of magnetite. On the basis of the column model, field-scale predictive simulations were conducted to illustrate the feasibility of the nitrate-Fe(II) strategy for intercepting advected As from a plume. The predictive simulations, which suggested that long-term As immobilization was feasible, favored a scenario that maintains high dissolved Fe(II) concentration during injection periods and thereby converts ferrihydrite to magnetite.

Quantifying Reactive Transport Processes Governing Arsenic Mobility after Injection of Reactive Organic Carbon into a Bengal Delta Aquifer.

Over the last few decades, significant progress has been made to characterize the extent, severity, and underlying geochemical processes of groundwater arsenic (As) pollution in S/SE Asia. However, comparably little effort has been made to merge the findings into frameworks that allow for a process-based quantitative analysis of observed As behavior and for predictions of its long-term fate. This study developed field-scale numerical modeling approaches to represent the hydrochemical processes associated with an in situ field injection of reactive organic carbon, including the reductive dissolution and transformation of ferric iron (Fe) oxides and the concomitant release of sorbed As. We employed data from a sucrose injection experiment in the Bengal Delta Plain to guide our model development and to constrain the model parametrization. Our modeling results illustrate that the temporary pH decrease associated with the sucrose transformation and mineralization caused pronounced, temporary shifts in the As partitioning between aqueous and sorbed phases. The results also suggest that while the reductive dissolution of Fe(III) oxides reduced the number of sorption sites, a significant fraction of the released As was rapidly scavenged through coprecipitation with neo-formed magnetite. These secondary reactions can explain the disparity between the observed Fe and As behavior.

Numerical Modeling of Arsenic Mobility during Reductive Iron-Mineral Transformations.

Millions of individuals worldwide are chronically exposed to hazardous concentrations of arsenic from contaminated drinking water. Despite massive efforts toward understanding the extent and underlying geochemical processes of the problem, numerical modeling and reliable predictions of future arsenic behavior remain a significant challenge. One of the key knowledge gaps concerns a refined understanding of the mechanisms that underlie arsenic mobilization, particularly under the onset of anaerobic conditions, and the quantification of the factors that affect this process. In this study, we focus on the development and testing of appropriate conceptual and numerical model approaches to represent and quantify the reductive dissolution of iron oxides, the concomitant release of sorbed arsenic, and the role of iron-mineral transformations. The initial model development in this study was guided by data and hypothesized processes from a previously reported,1 well-controlled column experiment in which arsenic desorption from ferrihydrite coated sands by variable loads of organic carbon was investigated. Using the measured data as constraints, we provide a quantitative interpretation of the processes controlling arsenic mobility during the microbial reductive transformation of iron oxides. Our analysis suggests that the observed arsenic behavior is primarily controlled by a combination of reductive dissolution of ferrihydrite, arsenic incorporation into or co-precipitation with freshly transformed iron minerals, and partial arsenic redox transformations.

Reactive transport of iomeprol during stream-groundwater interactions.

The transport and biochemical transformations of the iodinated X-ray contrast medium (ICM) iomeprol were studied at the stream/groundwater interface. During a one-month field experiment piezometric pressure heads, temperatures, and concentrations of redox-sensitive species, iomeprol and 15 of its transformation products (TPs) were collected in stream- and groundwater. The data set was analyzed and transformation processes and rates identified by comparing conservative and reactive transport simulations. ICM and TP transformations were simulated as a cometabolic process during organic carbon degradation. Using iomeprol/TPs ratios as calibration constrain mitigated the uncertainties associated with the high variability of the ICM wastewater discharge into the investigated stream. The study provides evidence that biodegradation of ICM occurs at the field-scale also for predominantly denitrifying conditions. Under these anaerobically dominated field conditions shortest simulated half-life (21 days) was in the same range as previously reported laboratory-determined half-lives for aerobic conditions.

Model-Based Analysis of Arsenic Retention by Stimulated Iron Mineral Transformation under Coastal Aquifer Conditions.

A multitude of geochemical processes control the aqueous concentration and transport properties of trace metal contaminants such as arsenic (As) in groundwater environments. Effective As remediation, especially under reducing conditions, has remained a significant challenge. Fe(II) nitrate treatments are a promising option for As immobilization but require optimization to be most effective. Here, we develop a process-based numerical modeling framework to provide an in-depth understanding of the geochemical mechanisms controlling the response of As-contaminated sediments to Fe(II) nitrate treatment. The analyzed data sets included time series from two batch experiments (control vs treatment) and effluent concentrations from a flow-through column experiment. The reaction network incorporates a mixture of homogeneous and heterogeneous reactions affecting Fe redox chemistry. Modeling revealed that the precipitation of the Fe treatment caused a rapid pH decline, which then triggered multiple heterogeneous buffering processes. The model quantifies key processes for effective remediation, including the transfer of aqueous As to adsorbed As and the transformation of Fe minerals, which act as sorption hosts, from amorphous to more stable phases. The developed model provides the basis for predictions of the remedial benefits of Fe(II) nitrate treatments under varying geochemical and hydrogeological conditions, particularly in high-As coastal environments.

Model-based integration and analysis of biogeochemical and isotopic dynamics in a nitrate-polluted pyritic aquifer.

Leaching of nitrate from agricultural land to groundwater and the resulting nitrate pollution are a major environmental problem worldwide. Its impact is often mitigated in aquifers hosting sufficiently reactive reductants that can promote autotrophic denitrification. In the case of pyrite acting as reductant, however, denitrification is associated with the release of sulfate and often also with the mobilization of trace metals (e.g., arsenic). In this study, reactive transport modeling was used to reconstruct, quantify and analyze the dynamics of the dominant biogeochemical processes in a nitrate-polluted pyrite-containing aquifer and its evolution over the last 50 years in response to changing agricultural practices. Model simulations were constrained by measured concentration depth profiles. Measured (3)H/(3)He profiles were used to support the calibration of flow and conservative transport processes, while the comparison of simulated and measured sulfur isotope signatures acted as additional calibration constraint for the reactive processes affecting sulfur cycling. The model illustrates that denitrification largely prevented an elevated discharge of nitrate to surface waters, while sulfate discharges were significantly increased, peaking around 15 years after the maximum nitrogen inputs.

Sulfate reduction accelerates groundwater arsenic contamination even in aquifers with abundant iron oxides.

Groundwater contamination by geogenic arsenic is a global problem affecting nearly 200 million people. In South and Southeast Asia, a cost-effective mitigation strategy is to use oxidized low-arsenic aquifers rather than reduced high-arsenic aquifers. Aquifers with abundant oxidized iron minerals are presumably safeguarded against immediate arsenic contamination, due to strong sorption of arsenic onto iron minerals. However, preferential pumping of low-arsenic aquifers can destabilize the boundaries between these aquifers, pulling high-arsenic water into low-arsenic aquifers. We investigate this scenario in a hybrid field-column experiment in Bangladesh where naturally high-arsenic groundwater is pumped through sediment cores from a low-arsenic aquifer, and detailed aqueous and solid-phase measurements are used to constrain reactive transport modelling. Here we show that elevated groundwater arsenic concentrations are induced by sulfate reduction and the predicted formation of highly mobile, poorly sorbing thioarsenic species. This process suggests that contamination of currently pristine aquifers with arsenic can occur up to over 1.5 times faster than previously thought, leading to a deterioration of urgently needed water resources.

Sources of ammonium enriched in groundwater in the central Yangtze River Basin: Anthropogenic or geogenic?

The occurrence of excessive ammonium in groundwater threatens human and aquatic ecosystem health across many places worldwide. As the fate of ammonium in groundwater systems is often affected by a complex mixture of transport and biogeochemical transformation processes, identifying the sources of groundwater ammonium is an important prerequisite for planning effective mitigation strategies. Elevated ammonium was found in both a shallow and an underlying deep groundwater system in an alluvial aquifer system beneath an agricultural area in the central Yangtze River Basin, China. In this study we develop and apply a novel, indirect approach, which couples the random forest classification (RFC) of machine learning method and fluorescence excitation-emission matrices with parallel factor analysis (EEM-PARAFAC), to distinguish multiple sources of ammonium in a multi-layer aquifer. EEM-PARAFAC was applied to provide insights into potential ammonium sources as well as the carbon and nitrogen cycling processes affecting ammonium fate. Specifically, RFC was used to unravel the different key factors controlling the high levels of ammonium prevailing in the shallow and deep aquifer sections, respectively. Our results reveal that high concentrations of ammonium in the shallow groundwater system primarily originate from anthropogenic sources, before being modulated by intensive microbially mediated nitrogen transformation processes such as nitrification, denitrification and dissimilatory nitrate reduction to ammonium (DNRA). By contrast, the linkage between high concentrations of ammonium and decomposition of soil organic matter, which ubiquitously contained nitrogen, suggested that mineralization of soil organic nitrogen compounds is the primary mechanism for the enrichment of ammonium in deeper groundwaters.