Groundwater recharge and agricultural contamination John-Karl Böhlke

Abstract Agriculture has had direct and indirect effects on the rates and compositions of groundwater recharge and aquifer biogeochemistry. Direct effects include dissolution and transport of excess quantities of fertilizers and associated materials and hydrologic alterations related to irrigation and drainage. Some indirect effects include changes in water–rock reactions in soils and aquifers caused by increased concentrations of dissolved oxidants, protons, and major ions. Agricultural activities have directly or indirectly affected the concentrations of a large number of inorganic chemicals in groundwater, for example NO3–, N2, Cl, SO42–, H+, P, C, K, Mg, Ca, Sr, Ba, Ra, and As, as well as a wide variety of pesticides and other organic compounds. For reactive contaminants like NO3–, a combination of chemical, isotopic, and environmental-tracer analytical approaches might be required to resolve changing inputs from subsequent alterations as causes of concentration gradients in groundwater. Groundwater records derived from multicomponent hydrostratigraphic data can be used to quantify recharge rates and residence times of water and dissolved contaminants, document past variations in recharging contaminant loads, and identify natural contaminant-remediation processes. These data indicate that many of the world’s surficial aquifers contain transient records of changing agricultural contamination from the last half of the 20th century. The transient agricultural groundwater signal has important implications for longterm trends and spatial heterogeneity in discharge. Résumé L’agriculture a eu des effets directs et indirects sur la recharge et la composition des nappes et sur la biogéochimie des aquifères. Les effets directs sont la dissolution et le transport de quantités excessives d’engrais

Received: 14 August 2001 / Accepted: 4 October 2001 Published online: 22 January 2002 © Springer-Verlag 2002 J.-K. Böhlke (✉) US Geological Survey, 431 National Center, Reston, VA 20192, USA e-mail: [email protected] Fax: +1-703-648-5274 Hydrogeology Journal (2002) 10:153–179

et des produits associés et des modifications hydrologiques liées à l’irrigation et au drainage. Certains des effets indirects sont des modifications des réactions eau-roche dans les sols et dans les aquifères, causées par des concentrations croissantes d’oxydants, de protons et d’ions majeurs dissous. Les activités agricoles ont affecté directement ou indirectement les concentrations d’un grand nombre de composés minéraux dans les eaux souterraines, comme par exemple NO3–, N2, Cl, SO42–, H+, P, C, K, Mg, Ca, Sr, Ba, Ra, As, de même qu’une grande variété de pesticides, de produits de dégradation et d’autres composés organiques. Pour les contaminants réactifs comme NO3, une combinaison d’approches analytiques de traceurs chimiques, isotopiques et environnementaux peut être nécessaire pour analyser les changements des entrées dues à des modifications causant des gradients de concentration dans les eaux souterraines. Les chroniques de nappes fournies par des données hydrostratigraphiques à composantes multiples peuvent être utilisées pour quantifier les valeurs de recharge et les temps de séjour de l’eau et des contaminants dissous, pour établir les variations passées d’apports de charges contaminantes et pour identifier les processus naturels de remédiation des contaminations. Ces données indiquent que de nombreux aquifères à la surface du globe contiennent et sont en train de transporter des témoignages transitoires provenant des variations de la contamination agricole au cours de la dernière moitié du 20ème siècle. Le signal transitoire d’origine agricole dans les nappes présente des implications importantes en ce qui concerne les tendances à long terme et l’hétérogénéité spatiale de l’écoulement. Resumen La agricultura tiene efectos directos e indirectos en las tasas y composiciones de la recarga y de la biogeoquímica de las aguas subterráneas. Entre los primeros, cabe citar la disolución y transporte de fertilizantes y materiales asociados que han sido aplicados en exceso, así como las alteraciones hidrológicas relacionadas con el riego y drenaje. Entre los segundos, se incluye los cambios en la interacción entre aguas y roca, tanto en suelos como en acuíferos, como resultado del aumento de la concentración de oxidantes disueltos, protones e iones mayoritarios. Las actividades agrícolas han afectado directa o indirectamente a las concentraciones de un gran número de compuestos químicos inorgánicos en las aguas subterráneas; entre éstos, destacan el nitrato, nitróDOI 10.1007/s10040-001-0183-3

154

geno, cloruro, sulfato, hidrogeniones, fósforo, carbono, potasio, magnesio, calcio, estroncio, bario, radio, arsénico, y una enorme variedad de pesticidas, productos de degradación y otros compuestos orgánicos. Para los compuestos reactivos, como el nitrato, puede ser necesaria una combinación de métodos analíticos químicos, isotópicos y de trazadores naturales de cara a resolver las modificaciones en la entrada a partir de alteraciones posteriores como causas de los gradientes de concentración en el acuífero. Los registros de aguas subterráneas basados en datos hidro-estratigráficos multicomponente pueden ser utilizados para cuantificar la recarga y los tiempos de residencia del agua y de contaminantes disueltos, para documentar variaciones pasadas en las cargas de contaminantes de la recarga, y para identificar los procesos naturales de contaminación y remediación. Estos datos indican que muchos acuíferos someros del mundo contienen y proporcionan registros transitorios de contaminación agrícola variable desde la segunda mitad del siglo XX. Las marcas transitorias de contaminación agrícola en las aguas subterráneas son importantes en relación con las tendencias a largo plazo y con la heterogeneidad espacial de la descarga. Keywords Agriculture · Contamination · Groundwater recharge

Introduction Agriculture has had profound effects on the rates and compositions of groundwater recharge. In many areas of the world, the major-element chemical loads of water entering surficial (unconfined) aquifers in the last several decades have been dominated by constituents derived directly or indirectly from agricultural practices and additives. Irrigation and drainage have altered groundwater fluxes and flow patterns. Agricultural contaminant loads in recharging groundwater have resulted in well-known societal problems related to drinking-water quality (Fan and Steinberg 1996) and ecological effects of groundwater discharge to surface-water bodies (Howarth et al. 2000). In addition, agricultural contaminants have caused substantial changes in groundwater geochemistry and water-rock interactions, which have received somewhat less attention. Global trends indicate that agricultural effects on the hydrochemical cycle will continue to be important topics of research in the future (Tilman et al. 2001). Many studies indicate that agricultural practices have resulted in nitrate (NO3–) contamination of groundwater, with concentrations in shallow aquifers commonly exceeding the drinking-water maximum contaminant level (MCL) of 10 mg/L as N (714 µmol/L) in the USA (USEPA 1986) or the maximum admissible concentration (MAC) of 50 mg/L as NO3– (806 µmol/L) in Europe (EC 1980). Recent reviews of NO3– contamination in the USA include comparisons of regional N inputs on land with corresponding surface-water N discharge loads, Hydrogeology Journal (2002) 10:153–179

which are at least partly related to N fluxes in groundwater (Howarth et al. 1996; Smith et al. 1997; Goolsby et al. 1999), as well as statistical analyses of groundwater occurrence and aquifer susceptibility to NO3– contamination (Nolan 2001; Burkart and Stoner 2001). In addition to NO3–, other reviews address groundwater susceptibility to contamination by other agricultural contaminants with potential health or ecological implications such as pesticides (Hallberg 1989; Barbash and Resek 1996; Flury 1996; Kolpin et al. 1998; Kladivko et al. 2001) and phosphorus (Simms et al. 1998; Sharpley et al. 2000). A major part of the work on P and pesticides has been directed at surface-water runoff rather than groundwater recharge, in part because those contaminants are relatively likely either to be degraded rapidly in the subsurface (pesticides) or adsorbed strongly to solid particles (P). The contrasting transport behavior of chemicals like NO3–, P, and various pesticides makes the design of management strategies for minimizing their effects on ecosystems complicated (Heathwaite et al. 2000). Some other inorganic constituents of groundwaters that are present in agricultural additives include Cl, K, Ca, Mg, S, and a variety of minor elements. Agricultural sources of some of these elements can dominate natural sources locally. Furthermore, agricultural effects on the recharge fluxes of various ions like NO3– and H+ can cause changes in weathering rates and ion-exchange equilibria in the subsurface, thereby altering indirectly the concentrations of other constituents in groundwater. These indirect effects have important implications for geochemical studies of water–rock interactions and can represent sources or sinks for a variety of problematic contaminants such as nutrients and toxic trace elements. The general purpose of this paper is to summarize some effects of agriculture on the composition of recharging groundwater and to illustrate empirical approaches to quantifying those effects within the saturated zone. Because of the magnitude of the overall subject, and because other reviews have treated many of the important issues, this paper deals selectively with a few subtopics that have been featured recently in multidisciplinary groundwater research projects, including: 1. chemical and isotopic signatures of agricultural contamination involving mainly inorganic constituents; 2. use of environmental tracers in the saturated zone to determine recharge rates, chemical fluxes, and historical variations in agricultural recharge on time scales of years to decades; 3. interactions between agricultural contaminants and aquifer geochemistry; and 4. implications of recharge contamination history for discharge contaminant responses. The phrase “agricultural contaminants” here refers to chemicals whose concentrations in groundwater seem to be higher than they would be in the absence of agricultural activities, although they might not necessarily come directly from applied artificial substances. Use of the DOI 10.1007/s10040-001-0183-3

155

word “contamination” is meant to imply that human activities have caused an increase in the flux or concentration of a constituent, but not necessarily that the change has been harmful (Freeze and Cherry 1979). The major contaminant considered in this paper is NO3–, which is the most abundant, mobile, and persistent agricultural contaminant in many shallow groundwaters.

Selected Agricultural Contaminants in Recharging Groundwater Nitrate Correlations between agricultural land use and high NO3– concentrations in groundwater have been documented at least since the 1970s (Hallberg 1986; Follett 1989; Burt et al. 1993; Hamilton et al. 1993; Spalding and Exner 1993; Follett and Hatfield 2001), but the relations between groundwater NO3– contamination and N sources at the land surface are complicated. A variety of important sources of N in agricultural soils could be converted to NO3– and incorporated in groundwater recharge (Stevenson 1982; Keeney 1986; Meisinger and Randall 1991; USNRC 1993; Jordan and Weller 1996; Burkart and James 1999; Goolsby et al. 1999) (Fig. 1). The trends and magnitudes of major N sources illustrated in Fig. 1 are considered to represent large-scale regional features of the Mississippi River basin, USA, but all of these features vary widely among smaller areas. The most dramatic regional trend is exhibited by the fertilizer curve, which reflects a global increase in the use of N fertilizers since the middle of the 20th century, when artificial N fixation became an economical large-scale industrial process. In addition to N fertilizers and manures, N released from pre-existing soil organic matter or crop

Fig. 1 Estimated magnitudes of various non-point sources of N within the Mississippi River basin, USA, during the second half of the twentieth century (Goolsby et al. 1999; W. Battaglin, USGS, personal communication, 2001). Atmospheric deposition refers to the sum of wet+dry NO3– and wet NH4+. Soil N mineralization rates were estimated from (1) a constant fraction (2%) of the estimated amount of organic N in soils (Goolsby et al. 1999), and (2) the amount of N in annual crop residues (Burkart and James 2002) Hydrogeology Journal (2002) 10:153–179

residues (N “mineralization”) can be one of the largest potential sources of N in recharge generally (Paul et al. 1997; Burkart and James 1999; Goolsby et al. 1999; Burkart and James 2001), but it is difficult to quantify. Acceleration of soil organic N mineralization and oxidation caused by land clearing, plowing, drainage, and other agricultural practices provide large amounts of leachable NO3–, either annually or in large pulses at times of land-use change. The soil mineralization curves in Fig. 1 represent different assumptions about the controls of soil N-release rates and might not represent accurately many field situations where fertilization and cropping practices have changed with time (Stevenson 1982; Paul et al. 1997). Fixation of atmospheric N2 by leguminous crops such as alfalfa, soybeans, and peanuts is another major potential source of N in agricultural recharge that could dominate locally and be a significant fraction of the total regionally (Fig. 1). Atmospheric N deposition, including re-deposition of ammonia lost from agricultural land, can be an important N source regionally over mixed land uses, but it is a relatively minor source at the field scale where soil and fertilizer sources are concentrated. With sufficient field data, the amounts of excess N available for transport below the root zone to the water table from combined sources locally can be estimated by using deterministic models of N cycling in agricultural soils (Knisel et al. 1983; Molina and Richards 1984; Leonard et al. 1987; Wagenet and Hutson 1989; Shaffer et al. 1991). Reasonable estimates of NO3– concentrations in recharge also can be derived from watershed mass-balance calculations by making various assumptions (Hall and Risser 1993; Puckett et al. 1999). Variables affecting the relation between N loads at the land surface and recharging NO3– fluxes at the water table include the application rate, timing, and form of N in the fertilizer or manure; N losses by volatilization of species like ammonia gas; crop management and tillage practices; and local climatic factors. In areas with high agricultural N loading, regional-scale analyses indicate that NO3– concentrations in shallow groundwater can be positively correlated with the thickness of the unsaturated zone, efficiency of soil drainage, and recharge rate (Hamilton and Helsel 1995; Burkart et al. 1999; Nolan 2001), in part because shallow water tables and poor drainage are relatively likely to promote denitrification (microbial reduction of NO3– to N2 gas) or inhibit nitrification (microbial oxidation of NH4+ to NO3–) during infiltration. Despite the potential for variability, studies of various types in different areas indicate that the magnitudes of groundwater NO3– recharge rates are commonly about 10–50% of fertilizer N application rates beneath heavily fertilized well-drained fields (Hallberg 1986; Bouwer 1989; Meisinger and Randall 1991; Böhlke and Denver 1995). Some watershed-scale agricultural nutrient-balance studies have yielded similar estimates of excess fertilizer N use (Barry et al. 1993; Trachtenberg and Ogg 1994; Puckett et al. 1999). NO3– recharge fluxes in many cases are similar in magnitude to the rates of N removal in harvested crops. DOI 10.1007/s10040-001-0183-3

156

Evidence for fertilizer N as an important source of groundwater contamination has been developed from several different types of evidence, some of which are discussed in later sections. For example, long-term (decade-scale) monitoring studies have documented systematic increases in NO3– concentrations in springs and wells that are correlated with regional or local increases in the rates of N fertilizer use (Hallberg 1986; Hallberg and Keeney 1993). Monitoring NO3– concentrations and loads in soil lysimeters and subsurface drains for periods of years also has indicated rapid responses of shallow groundwaters to changes in fertilizer-N application rates (Baker and Johnson 1981; Hallberg et al. 1986). In areas lacking long-term monitoring data, groundwater dating and analysis indicate changes in past NO3– recharge rates that are correlated with fertilizer-use history (see below). NO3– is also correlated with other fertilizer components (e.g., Cl, Mg, Ca) and with pesticides in groundwater (Hallberg 1986; Denver 1989; Hamilton et al. 1993; Kolpin et al. 1994; Hamilton and Helsel 1995). Measurements of N isotope ratios have supported fertilizer or manure sources of NO3–-N in some agricultural groundwaters, although fractionations occurring before and after groundwater recharge commonly complicate those interpretations (see below). Experiments with 15N-enriched fertilizers reveal considerable variability in the transmission of fertilizer N to the water table, depending on soil characteristics, recharge mechanisms, and fertilizer form and application rates (Chichester and Smith 1978; Walters and Malzer 1990; Baker and Timmons 1994; Porter 1995; Normand et al. 1997; Haynes 1999; Wilkinson and Blevins 1999; Randall and Mulla 2001). The isotope-tracer experiments indicate that some field conditions lead to rapid (within-year) direct transmission of some fertilizer N to the water table when application rates are high, especially in areas with preferential hydrologic flow paths in the unsaturated zone; however, most experiments indicate substantial dilution of the labeled fertilizer N signal by unlabeled new or pre-existing soil N before recharge. Major forms of N applied to crops to increase productivity include urea [CO(NH2)2], ammonia (NH3), ammonium nitrate (NH4NO3), and various animal manures. In the nineteenth century and early twentieth century, natural nitrate mineral fertilizers were imported from areas like the Atacama Desert of Chile (Ericksen 1983). Except for manures, the N fertilizers used since the middle of the 20th century have been largely artificial, being produced by inorganic processes from atmospheric N2. Although NO3– is generally not the major form of N added to agricultural fields, it is usually the major form of N (other than dissolved atmospheric N2) in agricultural recharge. Concentrations of aqueous NH4+ in recharging groundwater are commonly less than a few µmol/L even where the dominant forms of soluble N fertilizer are reduced (urea, ammonia) and where NO3– concentrations are in the order of hundreds of µmol/L. High NO3–/NH4+ ratios in recharge result from high rates of nitrification relative to those of NH4+ transport through aerated agriHydrogeology Journal (2002) 10:153–179

cultural soils and unsaturated zones. Higher concentrations of NH4+ in groundwater occur where NH4+-rich wastewater is applied locally with an exceptionally high recharge rate, and in anoxic organic-rich saturated soils and aquifers. Many occurrences of NH4+ at concentrations of tens to hundreds of µmol/L in groundwater can be attributed to anaerobic degradation of organic matter within the saturated zone by processes similar to those occurring in recent sediments (Berner 1971). Concentrations of NO3– in uncontaminated groundwater are relatively low in most agricultural regions. Concentrations cited as averages or typical threshold values for non-agricultural or natural recharge include <30 µmol/L (Hamilton et al. 1993), <50 µmol/L (Pionke and Urban 1985), and <140–210 µmol/L (Hallberg and Keeney 1993); however, these values vary greatly from place to place and it is difficult to generalize. In particular, groundwaters in arid regions commonly have relatively high concentrations of NO3– of apparently natural origin, derived from weathering of N-bearing rocks, degradation of organic matter in soils, or atmospheric deposition. For example, NO3– concentrations of several hundred µmol/L and more are common in groundwaters that recharged throughout Holocene time in arid regions of Africa (Heaton et al. 1983; Edmunds and Gaye 1997; Edmunds 1999), Australia (Barnes et al. 1992), and western North America (McMahon and Böhlke 2001). High NO3–concentrations do not necessarily imply high recharge fluxes of NO3– in arid regions with low water-recharge rates; however, they could represent higher fluxes during wetter climates in the past, and they could be associated with high fluxes locally in areas with high recharge rates. Large quantities of NO3– accumulated in the unsaturated zone by natural processes over thousands of years or more can be incorporated rapidly in recharging groundwater as a result of irrigation, flooding, or other hydrologic disturbances (Densmore and Böhlke 2000). Other Agricultural Contaminants Commonly Associated with Nitrate Because microbial oxidation of NH4+ releases protons (H+), nitrification of reduced N from fertilizers and manures generates acidity along with NO3– in soils that lack phases capable of buffering pH: (1) Dolomite [CaMg(CO3)2] is commonly applied to agricultural land to provide Ca and Mg for plant growth and to neutralize acid soils: (2) Nitrification of reduced N from fertilizers and manures in soils containing imported or naturally occurring carbonate minerals can result in unnaturally high Ca and Mg concentrations in recharging groundwater: (3) DOI 10.1007/s10040-001-0183-3

157

Because of reactions like these, groundwaters recharged beneath fertilized fields in parts of the mid-Atlantic coastal plain in the US are best described as Ca–Mg–NO3– type waters, in marked contrast with the normal Ca, Nabicarbonate, and Ca,Na-sulfate types that are recharged in non-agricultural areas (Denver 1989; Hamilton et al. 1993). Even with dolomite addition, recharging groundwaters commonly have pH values around 5, and the C from the dolomite might not be recharged with the Ca and Mg (Böhlke and Denver 1995), presumably having been released to the atmosphere in the form of CO2. Hamilton et al. (1993) identify positive correlations of NO3– with Mg, Ca, Sr, Ba, K, Cl, and specific conductance (total ionic strength), but not with Na, Si, SO42–, or alkalinity (mainly bicarbonate). Agricultural application of K as a plant nutrient commonly results in Cl contamination of recharging groundwater, because the K commonly is applied in the form of KCl. The fate of K from KCl applications is complicated in part by ion exchange with clay minerals, so that K/Cl ratios in agricultural recharge are generally less than 1, and evidence for agricultural K enrichment over background concentrations is sporadic. Phosphate is sorbed strongly onto solid phases, including Fe and Al oxides in soils, and P concentrations in recharging groundwater generally do not reflect the large amounts of P applied to agricultural fields, especially in manures. Sulfate can be added to fields with NH4+ or Ca, but the movement of SO42– in the subsurface also is retarded by sorption. Because of interactions with solid phases in the unsaturated zone, contamination of recharging groundwater with agricultural K, P, and SO42– is most likely in areas with preferential flow paths (e.g., macropores or fractures); long histories of anthropogenic loading; high recharge rates; and coarse-grained surficial deposits lacking oxides, clays, and other phases with charged surfaces and exchange sites (Johnson and Cole 1980; Simms et al. 1998). Some minor elements added to agricultural fields currently or in the past include As in pesticides and poultry and swine feeds, Cu in feeds, and Br in soil fumigants, among many others. According to Welch et al. (2000), agricultural As use in the US peaked around the 1960s and resulted in substantial accumulations of As in soils but generally did not cause significant enrichment above background levels in recharging groundwater. Release of Br to pore water in soils can occur as a result of degradation of soil fumigants like ethylene dibromide (EDB) and dibromo1,2-dibromo-3-chloropropane (DBCP) (Pignatello 1986; Deeley et al. 1991). Radioactive elements such as Ra in the U decay series commonly are associated with P fertilizer components, which are derived largely from marine deposits that are enriched in U; however, precipitation and removal of CaSO4 during the wet process of phosphoric acid production tend to remove Ra and Po relative to U, Th, and Pb, which tend to remain in the phosphoric acid (Guimond 1990; Roessler 1990). Thus, depending on the product and the details of the process, the P fertilizers can have widely varying Ra Hydrogeology Journal (2002) 10:153–179

concentrations and Ra/U ratios. The complementary CaSO4 product might be a source of Ra if used as a soil conditioner or source of Ca or SO42–. Potential indirect effects of infiltration of agriculturally contaminated water with high ionic strength or acidity include increasing weathering rates and solubilities of minerals and higher release rates of sorbed or included trace elements in soils. Parkhurst and Plummer (1993) illustrate by mass-balance calculations some effects of sorption and ion exchange in the transmission of cations, including NH4+, K+, Ca2+, and Mg2+, from agricultural additives through agricultural soils in Delaware, USA. Whereas major agricultural cations such as these are capable of displacing a variety of other cations from exchange sites in soils, the major agricultural anions (NO3– and Cl–) are somewhat less efficient at displacing other anions. Sulfate and especially phosphate are more likely to do this, because of their relatively high affinities for exchange sites (Johnson and Cole 1980). For example, high rates of agricultural P loading might release problematic anions like arsenate from soils (Welch et al. 2000). Groundwater concentrations of Sr and Ra several times higher than background levels could be caused directly by addition of Sr- and Rabearing fertilizer components like carbonate or phosphate, or they could result indirectly from the effects of fertilizer components on rates of ion exchange or dissolution in agricultural soils. In eastern Maryland, USA, Sr with an isotopic composition similar to that of Srrich agricultural groundwater was identified in dolomite additives, but not in sufficient concentrations to account for the anomalous groundwater Sr abundances. One possible explanation is that dolomite used in the past had higher Sr concentrations; another is that much of the Sr in recharge was released from soils by reaction with other major agricultural contaminants (Böhlke and Horan 2000). Similarly, Kozinski et al. (1995) and Szabo et al. (1997) report a correlation between Ra and NO3– contamination in agricultural areas of the coastal plain of New Jersey, USA, and suggest that the groundwater Ra might have been derived largely from soils or aquifer materials subjected to enhanced rates of Ra leaching by ion exchange with agriculturally enriched cations like H+, Ca2+, and Mg2+. Organic compounds that can be recharged beneath agricultural land along with NO3– include a wide variety of herbicides, fungicides, and insecticides used on crops and soils, as well as hormones, antibiotics, and other feed additives that appear in manures from commercial livestock. Major pesticides that are detected in shallow groundwater beneath agricultural land include triazine and chloroacetanilide herbicide compounds (e.g., atrazine, cyanazine, simazine, alachlor, metolachlor, and prometon). Most of these compounds are degraded rapidly in soils, and their transport through the unsaturated zone is retarded by sorption, so that typically less than a few per cent of applied pesticides are recharged with groundwater (Rao and Alley 1993; Barbash and Resek 1996). In major surveys of agriculDOI 10.1007/s10040-001-0183-3

158

tural recharge areas in the US, the concentrations of these compounds in most groundwater samples were less than 0.01 µg/L, almost all were less than 1 µg/L, and almost all were less than maximum contaminant levels or lifetime health-advisory levels (Barbash et al. 1999). Statistical analyses of detection frequencies in groundwater for different sets of pesticides indicate positive correlations with relative degradation half lives in aerobic soils (Barbash et al. 1999) and negative correlations with sorption coefficients on organic C (Kolpin et al. 1998).

Approaches for Identifying and Quantifying Agricultural Nitrate Contamination in Recharge Isotopic Characterization of Agricultural Nitrate Sources The isotopic composition of NO3– in recharging groundwater depends on the source materials and reactions during its formation and subsequent biogeochemical alterations. Isotopic data have been used since the 1970s to argue that a substantial amount of NO3– in aquatic environments is from agricultural sources (Kohl et al. 1971; Kreitler 1975; Gormly and Spalding 1979; Kreitler 1979; Heaton 1986), but the evidence is sometimes difficult to interpret. Nitrate reagents and fertilizers produced artificially from atmospheric N2 and O2 generally have δ15N and δ18O values near those of the atmospheric gas sources, 0.0 and +23.5 ‰ for N and O, respectively (δiE= Rx/Rst–1 in parts per thousand, where iE=15N or 18O, Rx= 15N/14N or 18O/16O of the sample, and R =15N/14N of atst mospheric N2 or 18O/16O of Vienna Standard Mean Ocean Water). If transmitted directly to the water table, those nitrate fertilizer values would be distinguishable from the isotope ratios in some other NO3– sources (Kendall and Aravena 1999); however, it is unusual for those values to be encountered in NO3– recharging beneath fields receiving excess N fertilizers. Generally, δ15N values in groundwater NO3– beneath fertilized cropland are somewhat higher (averaging between about +2 and +6 ‰) and δ18O values are considerably lower (limited data from about 0 to +10 ‰) than nitrate fertilizer values (Kreitler 1979; Wolterink 1979; Böhlke and Denver 1995; Fogg et al. 1998; Kendall 1998; Kendall and Aravena 1999). These differences occur in part because: 1. fertilizer N mixes with other N reservoirs in soils that commonly have δ15N values higher than 0 ‰; 2. nitrate is not the major form of N fertilizer used in most agricultural areas, and reduced forms of N such as urea and ammonia acquire low δ18O values from H2O when nitrified microbially to produce NO3–; and 3. N transformations in soils fractionate the isotopes of N–O compounds and obscure the source signatures (Kreitler 1975; Kreitler 1979; Heaton 1986; Hübner 1986; Amberger and Schmidt 1987; Kendall and Aravena 1999) (Fig. 2). Hydrogeology Journal (2002) 10:153–179

Fig. 2 Some processes affecting the amount and isotopic composition of N in recharging groundwater. The boxes give representative isotopic differences between applied N and recharging N in fields receiving excess fertilizer and manure-N applications (δ15N could be ±2‰ or more). Arrows indicate some potential effects on the isotopic composition of applied N moving through the system: volatilization of NH3 leaves behind NH4+ with higher δ15N; nitrification produces NO3– with lower δ15N; denitrification leaves NO3– with higher δ15N

Transformations that could result in relatively high δ15N values in recharging NO3– include: 1. volatilization of NH3 to the atmosphere, which leaves soil NH4+ relatively enriched in 15N; and 2. denitrification, which leaves residual NO3– relatively enriched in 15N (Fig. 2). Experiments with 15N tracers commonly indicate that where excess fertilizer N applications result in substantial increases in NO3– leaching rates from soils, much of the excess fertilizer N is transformed into soil or plant organic N before being re-released and leached as NO3– (Walters and Malzer 1990; Baker and Timmons 1994; Normand et al. 1997; Haynes 1999). Nitrate derived from soil organic-N mineralization in the absence of fertilizer applications can also have δ15N values of around +2 to +6 ‰ (Kreitler and Jones 1975; Heaton 1986; Kaplan and Magaritz 1986; Fogg et al. 1998). Because fertilizer-N can be diluted and fractionated isotopically in soils, and because natural organic sources of NO3– in some types of soils have δ15N values near those of fertilizers, it is commonly difficult to prove from stable-isotope measurements whether or not N in recharging NO3– was fertilizer N on the land surface (Keeney 1986). A more distinctive agricultural isotopic signature commonly results from manure spreading, which typically yields NO3– in soils and recharging groundwater with δ15N>+8‰, commonly between about +10 and +25‰ (Kreitler 1975; Gormly and Spalding 1979; Wolterink 1979; McMahon and Böhlke 1996; Fogg et al. 1998; Fig. 2). The relatively high and variable δ15N values of NO3– recharged beneath manure-spreading areas DOI 10.1007/s10040-001-0183-3

159

apparently result largely from isotopic fractionation before nitrification (e.g., by volatilization of NH3 gas), and therefore might be affected by the handling of the material as well as by climate, soil chemistry (especially pH), and other environmental variables (Kreitler 1975). Despite the potential for variability from any source, a fairly consistent contrast commonly exists between the δ15N values of NO3– recharged beneath fields receiving artificial fertilizers and those receiving manures, and this contrast is evidence that the anthropogenic N components are not overwhelmed by natural components. Another locally distinctive isotopic signature can result from nitrogen fixation by leguminous crops, which produces organic matter with relatively low δ15N values near 0‰ or slightly negative (Hübner 1986; Bergersen et al. 1988; Peoples and Herridge 1990). Anomalously low δ15N values of around 0±1 ‰ in groundwater NO3– from some agricultural sites in the US have been interpreted as a result of mineralization, nitrification, and recharge of N from alfalfa residues (Böhlke et al. 2002; J.K. Böhlke, unpublished data). If true, this would imply that mineralization of the organic N source, nitrification, and transport of NO3– to the water table did not have a large net effect on the δ15N value of the N. Atmospheric NO3– in soils and groundwater recharge also can have δ15N near zero, but it can be identified by relatively high δ18O values (as high as 70–80‰) (Amberger and Schmidt 1987; Böhlke et al. 1997; Kendall and Aravena 1999), and by anomalous non-mass-dependent variations of δ17O and δ18O (Galanter et al. 2000; Michalski and Thiemens 2000), which have not been observed in other terrestrial NO3– sources. Accounting for Denitrification in the Saturated Zone Most agricultural contaminants are reactive under various conditions in the subsurface. Reactions need to be quantified if the effects of historical variations and natural remediation processes are to be resolved from either temporal or spatial data sets. For NO3–, the most important natural remediation reaction in the saturated zone is denitrification, which occurs where groundwater flow carries NO3– into contact with reduced phases in an aquifer: (4)

(5) (6) Where groundwater flow paths beneath agricultural watersheds have been followed downgradient through redox transitions, denitrification typically occurs within a sequence of microbially-mediated reactions, including O2 reduction, denitrification, Fe3+ reduction, SO42– reduction, and methanogenesis, in accordance with therHydrogeology Journal (2002) 10:153–179

modynamic principles. Concentrations of NH4+ commonly increase along groundwater flow paths after NO3– concentrations decrease. Reduction of NO3– to NH4+ could occur in some environments, but it appears that denitrification generally can account for the major NO3– losses in aquifers (see below). Much of the NH4+ encountered downgradient from NO3– in aquifers is probably released from organic C–N compounds by de-amination during anaerobic degradation, for example during SO42– reduction or methanogenesis. Denitrification can be detected in groundwaters by several different types of analysis, including: 1. δ15N and δ18O values of NO3– that increase exponentially with decreasing concentration, with the change in δ15N about twice the change in δ18O (Amberger and Schmidt 1987; Mariotti et al. 1988; Böttcher et al. 1990; Fig. 3 a); 2. decreases in the ratios of NO3– to various conservative constituents like Cl; 3. increases in the concentrations of reaction products like HCO3– or SO42– in solution (Vogel et al. 1981; Kölle et al. 1985; Trudell et al. 1986); and 4. changes in the isotopic compositions of the product species that can be related to the isotopic compositions of reacting aquifer phases. These indicators might be only qualitative or semiquantitative in many situations because: 1. isotopic fractionation factors attributed to denitrification are not always the same; 2. initial ratios of reactive and conservative species can vary in recharge; and 3. other processes in aquifers can alter the isotopic compositions of S and C species. In contrast, a rapidly growing number of studies indicates that denitrification in the saturated zone can be quantified by analyses of dissolved gases, including N2, if sampled groundwater parcels represent closed N systems and atmospheric sources of N2 can be resolved. The concentration of N2 produced by denitrification (e.g., 357 µmol/L from NO3– at the MCL value of 714 µmol/L) typically is less than the concentration of atmospheric N2 incorporated in recharging groundwater (e.g., 650 µmol/L in equilibrium with humid air at 10 °C and 1 atm total pressure). Commonly, the concentration of “excess N2” produced by denitrification in groundwater is estimated by comparing the measured concentrations of Ar and N2 with those expected from atmospheric sources, assuming the Ar is a conservative tracer (Vogel et al. 1981; Wilson et al. 1990; Dunkle et al. 1993; Böhlke and Denver 1995; Martin et al. 1995; McMahon and Böhlke 1996; Johnston et al. 1998; Modica et al. 1998; McMahon et al. 1999; Fig. 3b). More than two unknown quantities affect the total concentration of N2, including equilibrium recharge temperature, amount of excess air incorporated, fractionation of excess air, producDOI 10.1007/s10040-001-0183-3

160

Fig. 3a–c Approaches for detecting denitrification and for reconstructing the initial concentration and isotopic composition of NO3– in denitrified groundwaters. The illustrated curves are only representative, because they depend on variable isotope-fractionation factors as well as initial concentrations and isotope ratios. a Relation between reaction progress and the isotopic composition of reactant NO3– and product N2. Representative fractionation factors are indicated by α (Rproduct/Rreactant) and ε (1000×[α–1]), where R is 15N/14N or 18O/16O and ε(N)≈2×ε(O). Reaction progress (ξ) is given by the fractional loss of reactant NO3– (1–C/C°) in a sample, where C and C° are the measured and initial NO3– concentrations, respectively. b Relation between Ar and N2 concentrations in groundwaters. N2 from denitrification is given by the concentration of N2 in excess of atmospheric sources acquired by equilibration with air at the water table and by dissolution of excess air entrained during recharge events. c Relation between Ar/N2 ratios and δ15N[N2] values in groundwaters undergoing denitrification. The two curves indicate the reaction paths followed by two different parcels of water containing different initial NO3– concentrations (714 and 1428 µmol/L; or 10 and 20 mg/L as N) Hydrogeology Journal (2002) 10:153–179

tion of excess N2 from denitrification, amount of degassing, and fractionation during degassing; therefore, measurements of Ar and N2 cannot yield unique estimates of the N2 yield from denitrification. Nevertheless, in some field areas the recharge temperatures are relatively constant, excess air is relatively unfractionated, and degassing has not occurred. In these situations, the concentrations of excess N2 can be determined reasonably well from Ar and N2 analyses, as suggested by Vogel et al. (1981). Blicher-Mathiesen et al. (1998) considered the problem of estimating concentrations of excess N2 from Ar and N2 concentrations in shallow groundwaters that were partially degassed, in which case it was necessary to assume that all of the waters had the same Ar concentrations when recharged. In principle, the number of assumptions needed for these calculations is reduced if the concentrations of other atmospheric noble gases (Ne, Kr, and Xe) are known (Wilson et al. 1990). Analyses that include Ne, Ar, and N2 indicate variability in both recharge temperatures and the quantities of excess air in recharge and yield relatively precise estimates of potential denitrification N2 enrichments in the absence of degassing and excess air fractionation (Plummer et al. 2001; Verstraeten et al. 2001). Increasingly complicated situations can be evaluated with the addition of Kr and Xe data (Aeschbach-Hertig et al. 1999; Ballentine and Hall 1999; Stute and Schlosser 1999; Aeschbach-Hertig et al. 2000), although these options have not been explored fully in denitrification studies. Given the concentrations of NO3– and of excess N2 attributable to denitrification in a groundwater sample, the initial NO3– concentration and the degree of reaction progress can be calculated by mass balance (Fig. 3a). Similarly, the initial isotopic composition of NO3– that has been denitrified can be determined from analyses of residual NO3– and product N2 in a groundwater sample, if the sample represents a closed N system and does not contain large concentrations of intermediate species like NO2– and N2O. Some exceptions to these conditions have been reported (Wilson et al. 1990; BlicherMathiesen et al. 1998; Böhlke et al. 2000); nevertheless, a growing number of studies indicates that the initial N isotopic composition of the NO3– from recharge commonly is well preserved in the reactant NO3– and product N2 in partially to completely denitrified groundwaters (Vogel et al. 1981; Böhlke and Denver 1995; Böhlke et al. 2002). The δ15N value of the excess N2 produced by denitrification is relatively low initially, then gradually increases as reaction proceeds; thus, the total N2 in a closed system will follow a curved path in δ15N versus Ar/N2 with progressive reaction (Fig. 3c). The δ15N value of the excess non-atmospheric N2 in a sample at any point in the reaction progress is derived from measurements of the total N2 by a linear mixing calculation. Mass-balance calculations for the N isotopes provide supporting evidence that denitrification (rather than NO3– reduction to NH4+, for example) is the major NO3– reducing process in the aquifers. For example, if the total isotopic compositions of combined NO3– and non-atmoDOI 10.1007/s10040-001-0183-3

161

spheric N2 are constant through a series of samples from the same recharge area but with different degrees of denitrification, it is relatively unlikely that another N species such as NH4+ was a major reaction product. The isotopic evidence is important because dissolved NH4+ and NO3–-N will not constitute a closed system if the NH4+ is more strongly sorbed (Ceazan et al. 1989). Isotopic analyses of aquifer solid N and aqueous NH4+ may indicate if the NH4+ could have been derived from sedimentary organic matter locally (e.g., McMahon et al. 1999). Besides being used to reconstruct initial recharge characteristics, the evaluation of denitrification products also may be used to identify aquifer characteristics promoting natural remediation of NO3– contamination. For example, although denitrification is commonly coupled with organic carbon oxidation (Eq. 4), there is abundant evidence that sulfide and other reduced Fe phases provide electrons for denitrification in a variety of aquifers [Eqs. (5) and (6)], even in some places where organic carbon is more abundant (Kölle et al. 1985; Böttcher et al. 1991; Postma et al. 1991; Böhlke et al. 2002). Geochemical modeling of groundwaters containing agricultural contamination is complicated if the initial conditions have changed as a result of changing agricultural practices on time scales that are similar to those of the reactive transport flow paths. To address this problem, Postma et al. (1991) used variations in specific conductance to estimate variations in NO3– initial concentrations along flow paths in an application of forward geochemical modeling. In contrast, Böhlke and Denver (1995) selected data from samples with similar ages but different flow paths for inverse mass-balance calculations. Estimating Recharge, Chemical Fluxes, and Reaction Rates from Groundwater-Age Gradients Groundwater recharge rates and velocities are needed to determine fluxes and reaction rates of agricultural contaminants. The recharge rate of groundwater is defined as the volumetric flux of water across the water table per unit area, commonly expressed in units of m/year. A variety of techniques has been used to estimate recharge rates, including evaluations of soil–moisture balance, analyses of water-level changes during recharge events, and assessments of stream baseflow. Some of these methods have the advantage of yielding information about recharge on short (<1 year) time scales, but they are commonly relatively indirect or imprecise and they do not yield corresponding information about groundwater residence times or long-term contaminant histories. In contrast, relatively direct estimates of recharge rates can be derived from analyses of groundwater-age gradients to obtain vertical velocities, which also yield information about longer time scales within the aquifer but with correspondingly less resolution at sub-annual time scales. For the important recent time scale for documenting changes in agricultural practices between about 1940 Hydrogeology Journal (2002) 10:153–179

and the present, and for recharge rates of around 0.1 m/year or more, groundwater age gradients commonly are derived from measurements of environmental tracers such as 3H, 3He, CFCs, SF6, and 85Kr (Schlosser et al. 1989; Solomon et al. 1993; Ekwurzel et al. 1994; Schlosser et al. 1998; Loosli et al. 1999; Plummer and Busenberg 1999; Busenberg and Plummer 2000). Except for 3He, which is formed by radioactive decay of 3H, these are globally distributed anthropogenic environmental tracers whose concentrations have varied over the last 50 years in the atmosphere and whose concentrations in a groundwater sample are related to the time of recharge of the sample (Cook and Herczeg 1999). The resolution of the groundwater dating methods based on these tracers is limited by a variety of factors, including analytical uncertainties, contamination or degradation of the tracers, varying transport behavior of dissolved gases and water in the unsaturated zone and saturated zone, gas exchange across the water table, and spatial or temporal variations in the recharge rate. Where the vertical groundwater-age gradient can be determined, the recharge rate (r, in m/year) is estimated from the tangent to the age–depth curve, or vertical velocity, at the water table (Vv°, in m/year), multiplied by the connected porosity (n, dimensionless): (7) Given r, the recharge fluxes of agricultural contaminants and other chemicals (F, in mol/m/year) are determined from the measured concentrations (C, in mol/L): (8) Chemical fluxes derived in this way can be compared directly with input data such as land-surface loading (e.g., Fig. 1) or soil mass-balance data to determine the fractions of various sources needed to account for a contaminant in recharge. In simple water-table aquifers receiving distributed recharge, age gradients are commonly approximated by linear or logarithmic functions of depth (Vogel 1967; Zuber 1986; Cook and Böhlke 1999). One type of hypothetical aquifer situation is illustrated in Fig. 4a, where it is assumed that the hydraulic properties of the aquifer do not vary systematically with depth, that recharge occurs uniformly across the top, and that the change in elevation of the water table is small compared with the aquifer thickness. In this case, the groundwater age–depth profile is approximately logarithmic (Fig. 4b), and the recharge rate is given by: (9) where Z is the total thickness of the saturated zone, and ti is the age of a groundwater parcel at depth zi below the water table. The recharge rate is also given by: (10) where τ is the mean age or residence time of groundwater in the aquifer (e.g., 32 years in Fig. 4a). Figure 4b includes the vertical distributions of some of the tracers DOI 10.1007/s10040-001-0183-3

162

Fig. 4a, b Groundwater flow paths, groundwater age distributions, and well configurations in a hypothetical surficial (unconfined) aquifer. a Hypothetical groundwater flow system and selected well-screen locations. Flow lines (solid) and isochrons (dotted) were calculated by assuming homogeneous aquifer properties and evenly distributed recharge with the parameters listed on the figure. Symbols and bars labeled A–F represent the vertical intervals over which the different types of well are open to the aquifer. b Vertical profiles of groundwater age and concentrations of NO3– and selected atmospheric environmental tracers corresponding to the flow system in a. The groundwater-age curve is from Eq. (9). NO3– concentrations are in mmol/L from Böhlke and Denver (1995; see Fig. 8a), with linear extrapolation to 0 in 1850 and leveling off at 1.2 in 2000. Other concentration curves indicate representative groundwater environmental-tracer distributions in the mid-Atlantic region of North America, as sampled in the year 2000: CFC-12 and SF6 as atmospheric mixing ratios in parts per trillion by volume; 3H in tritium units (10–18×H); , where λ is the 3H decay constant (0.0558/year)

used to date young groundwater as they would occur in the hypothetical aquifer shown in Fig. 4a in the absence of vertical mixing. Groundwater age data from some representative surficial aquifers in the eastern USA (Fig. 5) could be approximately consistent with logarithmic age–depth profiles, though they lack important information in deep parts of the systems, where ages are more difficult to resolve because of mixing or non-detectable concentrations of the environmental tracers (more than 40–50 years old). Application of Eq. (9) to these data sets yields average recharge rates that range from about 0.05–0.45 m/year. Coincidentally, the mean groundwater residence times derived from Eq. (10) are approximately the same at all four sites, because the apparent recharge rates are roughly proportional to the estimated aquifer pore volumes. The shallow parts of the profiles also indicate local spatial variations in recharge rates of at least a factor of two and perhaps more at each site. Local variations in the apparent recharge rates at the Princeton site in Minnesota, USA, are attributed in part to focused recharge caused by surface runoff into shallow topographic depressions, which also resulted in spatial variations in the penetration depth of recent agricultural contaminaHydrogeology Journal (2002) 10:153–179

Fig. 5 Relation between groundwater age and depth below the water table in four watershed study areas in agricultural regions of the USA. Data points indicate apparent ages derived from CFC concentrations in groundwater samples collected from unconsolidated surficial (unconfined) aquifers at Edgewater, Maryland (O’Connell et al. 1997a); Locust Grove, Maryland (Dunkle et al. 1993; Böhlke and Denver 1995); Princeton, Minnesota (Delin et al. 2001; Böhlke et al. 2002); and southern New Jersey (Szabo et al. 1996). Shading highlights trends from the different study areas. Age–depth curves and residence times were calculated by using Eqs. (9) and (10) with the parameters listed on the right side of the figure (Z=aquifer thickness, n=porosity, r=recharge rate, and τ=mean age or residence time)

tion (Delin et al. 2001). A larger data set from the Delmarva Peninsula in the mid-Atlantic coastal plain of the USA (Dunkle et al. 1993) indicates a range in Vv° from less than 0.2 to about 3 m/year, possibly as a result of variations in topography, soil permeability, or agricultural drainage and tillage practices. DOI 10.1007/s10040-001-0183-3

163

Data from the Edgewater site in Maryland (Fig. 5) are interpreted to include a discontinuity in the average age gradient at about 0.5–1.0 m below the water table that corresponds to a decrease in average hydraulic conductivity, which is related in part to the vertical distribution of weathering and macropores in the coastal-plain sediments (O’Connell et al. 1997a; O’Connell 1998). This discontinuity results in a two-level flow system in which the shallower part has relatively high rates of recharge and discharge and low mean residence time (months to years), whereas the deeper part has lower rates of recharge and discharge and higher mean residence time (decades, as indicated in Fig. 5). Vertical zonation of recharge rates and solute movement within the saturated zone occurs in a variety of natural settings, including colluvial deposits or weathering zones overlying consolidated fractured-rock aquifers (Schnabel et al. 1993), and it may have some features in common with artificial subsurface drainage (Skaggs et al. 1994). Vertical zonation of flow systems with contrasting residence times can be associated with seasonal variations in the transport of agricultural contaminants from recharge to discharge (Schnabel et al. 1993; O’Connell et al. 1997b; Fenelon and Moore 1998). Analyses of recharge rates, groundwater velocities, and chemical and isotopic mass balances of N, C, S, and other elements can be used to derive information about both the stoichiometries and the rates of biogeochemical reactions in aquifers that are useful in evaluating natural contaminant-remediation potential. The rate of a reaction like denitrification can be derived from the relation between reaction progress and groundwater age if the reaction zone is gradual enough to span a range of measurable ages and if the reaction progress (ξ, stoichiometric amount of reaction from the initial condition) can be defined for each sample within the zone. For example, if ∆ξ is defined as a small incremental decrease in the concentration of a reactant like NO3–, then incremental apparent reaction rates are given by: (11) or (12) where ∆ξ/∆t is the difference between the progress of the reaction for two samples with different ages, t1
(Böttcher et al. 1989; Böhlke et al. 2002) have been estimated this way. Accounting for Groundwater Mixing and Unsaturated-Zone Transport Wells with short screens at discrete depths below the water table yield the least ambiguous age-gradient data for estimating recharge rates and chemical fluxes (Figs. 4b and 5), but many groundwater dating studies must rely on samples from wells with large open intervals that yield mixtures of waters of different ages. In principle, it should be possible to derive age gradients and recharge rates from analyses of mixed samples by making assumptions about the shape of the age-frequency distributions of the mixtures (Vogel 1967; Zuber 1986; Cook and Böhlke 1999). Alternatively, mixed samples can yield information about both the residence time and the shape of the age-frequency distribution if data for several different tracers are evaluated simultaneously or if long-term tracer records are available. Several different types of sampling situation are illustrated in Fig. 4a, where A, B, and C represent wells with small screened intervals open at the water table and at depths corresponding to one third and two thirds of the thickness of the saturated zone, respectively; E represents a well with a screened interval extending through the whole saturated zone; and D and F represent wells with long screened intervals spanning the upper third and lower third of the aquifer, respectively. The different positions and lengths of the well screens would result in dramatically different age distributions of pumped groundwaters (Fig. 6a) with mixed environmental-tracer concentrations (Fig. 6b, c, d). To produce the curves in Fig. 6, the concentrations of 3H and 3H° (3H°=3H+3He, or initial 3H) and the equilibrium atmospheric mixing ratios of SF6 and CFC-12 in mixed groundwaters were calculated by summing the contributions of annual or biannual increments: (13) where j is one of the constituents, t is an interval of time (typically a year or month), Xt is the volume fraction of water in the mixture that recharged at time t (from Fig. 6a), Cx,t is the concentration of j in recharge at time t (from Fig. 4b), and the sum is over all values of t in the mixture. Equilibrium atmospheric mixing ratios are used as concentration units for the gas tracers to facilitate comparison with global atmospheric tracer records. These units correspond to volumetric (or molar) fractions in dry air mixtures and are derived from aqueous concentrations by accounting for recharge temperature, solubility, recharge elevation, and excess air (Plummer and Busenberg 1999; Busenberg and Plummer 2000). The curves in Fig. 6 correspond to simple age-distribution models (f(Xt)) with different mean ages or residence times (Zuber 1986; Cook and Böhlke 1999), and the labeled points indicate the specific values for the aquifer and well configurations depicted in Fig. 4a. For example, in each diagram points representing wells A, B, DOI 10.1007/s10040-001-0183-3

164

Fig. 6a–d Age–frequency distributions and environmental-tracer concentrations in groundwater samples collected in the year 2000. a Age–frequency distributions of groundwaters discharged from wells A–F (shown in Fig. 4a) calculated for 1-year increments. Annual integration of the shaded area under each curve should yield a total of 100% of the discharge from the well; arrows for idealized short-screen wells A, B, and C point toward 100%, indicating that the discharge has a single age to within a year. The area labeled “G” indicates the portion of the water from well E that would contain NO3– if the groundwater below the level of well B had no NO3–. b, c, d Concentrations of selected groundwater-dating tracers for hypothetical samples from wells A–F (see Fig. 4a), illustrating approaches for estimating groundwater ages and residence times from single mixed samples. The mean ages of groundwaters pumped from the wells (in years) would be approximately 0 for well A, 13 for well B, 35 for well C, 6 for well D, 32 for well E, and 67 for well F. The points representing individual wells (A–F) are on curves that represent similar forms of age–frequency distributions but different mean ages (Zuber 1986; Cook and Böhlke 1999) Hydrogeology Journal (2002) 10:153–179

and C are on a curve that represents discrete groundwater parcels of all ages (“piston-flow” or “pipe-flow” model, PF), whereas point E is on a curve that represents “exponential mixtures” (EM) with a range of mean ages. Point F is on a hybrid model curve representing groups of flow lines that have exponential age distributions in distant recharge areas, resulting in “exponential-piston mixtures” (EPM). In addition to the situation represented by well F (Fig. 6a), variations of the EPM age-distribution model have also been applied to discharge areas that are separated from recharge areas by intervals with no recharge, and to the reverse situation where a tracer like 3H has a substantial piston-flow transit time in the unsaturated zone and then enters an exponential flow system and discharges as at well E (referred to as PEM, or “piston-exponential mixtures”). Comparison of multiple environmental-tracer data can indicate if any simple models like these are applicable, or independent information about aquifers and wells can indicate which is more likely to be. Mean ages or model residence times can then be used to determine recharge rates from Eq. (10). In contrast with the hypothetical models applied to the homogeneous aquifer represented by Fig. 4a, various combinations of CFCs, 3H, 3He, and 85Kr data from wells in karst, fractured-rock, and heterogeneous unconsolidated aquifers commonly indicate that samples are bimodal mixtures in which the older component is essentially tracer-free (recharged before about 1950) and the young component contains the tracers (Plummer et al. 1998; Loosli et al. 1999; Plummer et al. 2000). In areas with relatively little agricultural contamination in recharge before 1950, the old components in these types of bimodal mixtures would dilute the concentrations of NO3– and other agricultural contaminants in the young components roughly in proportion to the dilution of the atmospheric tracers; however, the ages of the old components and the history of contamination in recharge before 1950 are commonly not known. Apparent age discordance among the different dating tracers in groundwater also results from differential transport through the unsaturated zone, which also affects the transit times of agricultural contaminants. An example is where 3H and other dissolved constituents in infiltrating water do not keep pace with relatively insoluble atmospheric gas tracers that are transported more rapidly to the water table in unsaturated-zone air. In this case, unsaturated-zone travel times of water and agricultural chemicals can be derived by comparing distributions of 3H and gas tracers in the saturated zone (Solomon et al. 1995; Zoellmann et al. 2001). This approach is complicated in areas with relatively thick or impermeable unsaturated zones where the atmospheric gas tracers themselves do not have modern concentrations at the water table (Cook and Solomon 1995; Johnston et al. 1998; Plummer et al. 2000). If water transit times in the unsaturated zone are relatively long, it is possible that 3H and recent agricultural contaminants will not have recharged at all except locally by preferential flow. In such cases, where piston flow through the matrix is a good approxiDOI 10.1007/s10040-001-0183-3

165

mation of unsaturated-zone transport, substantial records of past agricultural contamination can be found above the water table. Unsaturated-zone age gradients and recharge rates since the 1950s are sometimes derived from the vertical distributions of bomb-produced 3H and 36Cl in unsaturated-zone pore waters (Phillips et al. 1988) and for longer time scales from distributions of stable Cl or other species (Herczeg and Edmunds 1999). If temporal variations are not well preserved in a vertical profile because of mixing or heterogeneous flow, cumulative fluxes may be estimated from accumulated masses of the same constituents.

Reconstructing Groundwater Records of Agricultural Contamination History Groundwater Stratigraphy in the Saturated Zone Changes in the concentrations of agricultural contaminants like NO3– in monitoring wells, lysimeters, or drains provide direct evidence of recharge responses to changes in agricultural practices. In the absence of monitoring records, groundwater records (or “archives”) of agricultural contamination history can be reconstructed from arrays of wells that are sampled for analyses of groundwater age and composition. Surficial (unconfined) aquifers receiving distributed recharge can be subjected to groundwater stratigraphic analysis to determine deposition rates (recharge rates) of water and dissolved constituents as well as past changes in the recharge characteristics by using some of the same general principles that are applied commonly in stratigraphic analysis of sedimentary rocks (Cook and Böhlke 1999). In many unconsolidated aquifers, vertical dispersion is sufficiently limited that annual-to-decadal layers of recharged groundwater with different compositions are not homogenized after entering the saturated zone. Some of the evidence for this is from well-preserved vertical 3H gradients in groundwater layers that crossed the water table after the 3H bomb peak in precipitation, and from concordance between the various groundwater dating techniques that would yield discordant apparent ages if the samples were mixed (Robertson and Cherry 1989; Solomon et al. 1993; Ekwurzel et al. 1994; Reilly et al. 1994; Szabo et al. 1996; Cook and Solomon 1997; Plummer and Busenberg 1999; Busenberg and Plummer 2000). Whereas groundwater ages (travel times from the water table) and age gradients are the quantities used to determine vertical velocities and recharge fluxes, the recharge dates are the more relevant quantities for comparing the groundwater contaminant records to land-use history. Early dating studies documented the distribution of NO3– with respect to the presence or absence of 3H, or the position of the 3H bomb peak, to demonstrate major changes in the recharge concentrations of agricultural NO3– (Howard 1985; Libra et al. 1987; Postma et al. 1991). Subsequently, groundwater recharge dates derived from atmospheric gas tracers and 3H/3He ratios have been used to develop more detailed NO3– recharge histoHydrogeology Journal (2002) 10:153–179

Fig. 7a, b Vertical sections showing groundwater ages and initial NO3– concentrations in surficial aquifers in the mid-Atlantic coastal plain, USA, sampled in 1990–1992. a Fairmount site in Delaware (Dunkle et al. 1993; Hamilton et al. 1993); b Locust Grove site in Maryland (Dunkle et al. 1993; Böhlke and Denver 1995). Groundwater ages were estimated from CFC analyses. All but one of the Locust Grove samples were oxic and undenitrified. Shading indicates groundwaters with initial NO3– concentrations higher than the drinking-water maximum contaminant level (MCL) of 10 mg/L as N (714 µmol/L)

ries (Dunkle et al. 1993; Böhlke and Denver 1995; Johnston et al. 1998; Modica et al. 1998; Plummer et al. 2000; Verstraeten et al. 2000; Böhlke et al. 2002). An important complication in the application of groundwater stratigraphy to the agricultural record is that increasingly deeper groundwaters in a single vertical profile are likely to have been recharged at increasing distances away from the well site, in which case it is difficult to separate the effects of changes in agricultural practices from the effects of different contributing recharge areas to different parts of the profile. An example of a spatially heterogeneous section is presented in Dunkle et al. (1993) for an unconsolidated coastal-plain aquifer in Delaware, in which denitrification was limited by high O2 concentrations (Fig. 7a). The distribution of ages indicates relatively uniform recharge rates across the section (varying by about a factor of 2), whereas NO3– concentrations indicate that a large-scale contaminant plume was recharged beneath agricultural land. Relatively uncontaminated groundwater below and above DOI 10.1007/s10040-001-0183-3

166

Fig. 8a–d Reconstructed groundwater record of recharge characteristics beneath agricultural land in coastal-plain sediments at the Locust Grove site in Maryland from about 1940 to 1992, based on analyses of samples collected in 1990–1992 (Böhlke and Denver 1995) (includes data from Fig. 7b). a Comparison of NO3– concentrations and recharge fluxes in groundwater with N-fertilizer loading rates. Measured NO3– concentrations were adjusted to remove effects of denitrification in some samples. Recharge years were assigned to individual samples by CFC dating (Dunkle et al. 1993). The “LOWESS smooth” curve is a locally-weighted moving average (Cleveland 1979). The fertilizer-loading curve represents 30% of the county-level sales data divided by the area of fertilized land in the county, dissolved in 0.25 m/year of recharge. The vertical axes are related to each other quantitatively by the recharge rate. b Relation between groundwater initial NO3– concentrations and isotopic compositions (both adjusted to remove effects of denitrification). c Relation between initial NO3– and Mg concentrations. d Relation between initial NO3– and Cl concentrations

the agricultural plume were recharged beneath forest and mixed land uses in upgradient and downgradient areas, respectively. This example illustrates that although agricultural NO3– is commonly described as a “non-pointsource” contaminant, it exhibits “point-source” attributes within surficial aquifers at the watershed scale. In contrast, a relatively homogeneous vertical section is presented for a predominantly agricultural recharge area at the Locust Grove site in Maryland (Fig. 7b). In addition to sampling vertical profiles in recharge areas, it is also possible to acquire groundwater records of agricultural contamination from discharge areas by dating and analysis of samples collected at shallow depths in horizontal arrays across streams and riparian wetlands where groundwater flow lines turn upward. Böhlke and Denver (1995) incorporate analyses of shallow discrete samples of discharging groundwater from beneath small streams (Fig. 7b) with data from wells in recharge areas to develop a NO3– recharge history for a surficial aquifer in Maryland. Other studies have investigated the use of distributed discharge samples alone to develop records of NO3– concentrations in recharge (Modica et al. 1998). This procedure is attractive because it is simpler and cheaper to sample shallow old groundwater in the discharge area than deep old groundHydrogeology Journal (2002) 10:153–179

water in the recharge area, but it can be complicated by relatively large amounts of mixing near the discharge area and relatively large uncertainty about the recharge areas and land uses represented by individual samples. Also, discharge samples do not yield directly the recharge rates and groundwater fluxes of water and dissolved species. Some examples of groundwater stratigraphic records of agricultural recharge are summarized in Fig. 8 and Fig. 9. In both of these examples, dissolved gas data are used to reconstruct initial NO3– concentrations and isotopic compositions in denitrified samples, and data from recharge areas and discharge areas are combined. In coastal-plain sediments at the Locust Grove site in Maryland (watershed-scale study), the average concentration of NO3– in recharge increased by a factor of about 4–5 between 1950 and 1990 (Fig. 8a), while the N isotopic composition of the NO3– remained relatively constant at around 2–5‰ (Fig. 8b) (Böhlke and Denver 1995). The increase in the concentration of NO3– is correlated with increases in the concentrations of Mg and Cl (Fig. 8c, d), which are attributed to dolomite and potash applications on the fertilized fields. The concentrations of dissolved constituents are combined with the average recharge rate of water (0.25 m/year, derived from groundwater dating) to obtain a record of chemical recharge fluxes for comparison with published fertilizer loading rates (Fig. 8a). The comparison indicates that the changing recharge rate of NO3– was equivalent to about 20–35% of the changing regional application rate of fertilizer N throughout the period represented by the groundwater record. DOI 10.1007/s10040-001-0183-3

167

Fig. 9a, b Reconstructed groundwater record of recharge characteristics beneath agricultural land in glacial outwash deposits at the Princeton site in Minnesota from about 1950 to 1994, based on analyses of samples collected in 1993–1994 (modified from Böhlke et al. 2002). General features and approaches are similar to those of Fig. 8. Shading and arrows highlight trends that are related to major periods of different crop types. a Comparison of NO3– concentrations and recharge fluxes in groundwater with N-fertilizer loading rates. The fertilizer curves represent 20% of the stateor county-level averages from 1950–1992, then 20% of the recorded local application rates in 1992–1994, dissolved in 0.15 m/year of recharge. Farming consisted of continuous alfalfa in the 1980s (to which fertilizer-N applications were lower than indicated by the regional curves) and continuous corn in the 1990s (with nearaverage fertilizer-N application rates). b Relation between groundwater initial NO3– concentrations and isotopic compositions (both adjusted to remove effects of denitrification). c Relation between initial NO3– and Cl concentrations

In glacial outwash sand deposits at the Princeton site in Minnesota (field-scale study), groundwater data indicate that the average concentration of NO3– increased during the 1980s, when alfalfa was grown continuously, and increased further during the early 1990s, when corn was grown continuously (Böhlke et al. 2002; Fig. 9a). The change in crops apparently resulted in a shift from relatively low δ15N values (down to about 0 ‰) and high Cl/NO3– ratios to relatively high δ15N values (averaging around 4 ‰) and low Cl/NO3– values in recharge (Fig. 9b, c). These changes are rationalized on the basis of different fertilizer formulations and N sources for the crops: alfalfa received substantial amounts of KCl but little or no N fertilizer and yielded crop residue with low δ15N as a potential source of NO3–, whereas corn received large amounts of reduced N fertilizer along with KCl and yielded isotopically fractionated N as a source of NO3–. The average recharge rate of NO3– in the early 1990s was equivalent to about 20% of the rate of fertilizer-N application to the corn, dissolved in 0.15 m/year of recharge. Groundwaters that recharged in the 1980s during the alfalfa-growing period illustrate the importance of nitrogen fixation and mineralization as a source of Hydrogeology Journal (2002) 10:153–179

NO3– in recharge beneath crops that do not receive large amounts of N fertilizer. Groundwater monitoring in a different location at the Princeton site indicates that Cl recharged during alfalfa and corn production was gradually displaced downward in the aquifer when grass was grown without major applications of KCl (Landon et al. 1997) (Fig. 10). A similar trend was observed for NO3–, but recharge of NO3– persisted slightly longer than that of Cl, and denitrification probably disturbed part of the pattern at depth. LowCl water infiltrated within a year to the water table, then moved down into the upper part of the aquifer at a rate apparently somewhat higher than indicated by the average recharge rate at the site (Fig. 5). The rapid rate of downward displacement of the Cl contours was probably affected by dispersion and by local variations in hydraulic properties or recharge. A variety of other techniques, including stable-isotope measurements of precipitation recharge, application and recovery of chemical tracers, unsaturated-zone moisture profiles, and water-balance measurements, all confirm that transit times of water and dissolved constituents through the unsaturated zone to the water table are of the order of weeks to months at the Princeton site (Landon et al. 2000; Delin et al. 2001). Systematic changes in the isotopic composition of recharging NO3– resulting from changes in agricultural practices might be difficult to detect owing to the variety N sources and isotopic fractionations in soils (Figs. 1 and 2). In some cases, variations in the activities of soil N-cycling processes cause substantial variations in the isotopic composition of recharging NO3– on seasonal or shorter time scales (Ostrom et al. 1998); however, seasonal variations in chemical and isotopic compositions of recharging water in the unsaturated zone and at shallow water tables are typically smoothed within the saturated zone, except where recharge is highly focused. Mineralization of preexisting soil N presumably would reduce the magnitude of interannual isotopic responses to changing sources. Nevertheless, observed interannual to interdecadal isotopic responses of recharging NO3– to changes in N sources at the DOI 10.1007/s10040-001-0183-3

168

Fig. 10 Multiport monitoring record of a NO3– and b Cl recharging beneath the Princeton site in Minnesota (modified from Landon et al. 1997). Each point represents a groundwater sample from a specific date and depth from a single multiport (B-70). The field was planted with alfalfa from 1981–1989, corn in 1990, and grass from 1991–1995

land surface (Fig. 9b) indicate that the amount of pre-existing soil N available annually for nitrification and recharge is limited in some cases. Thus, widespread regional increases in the ratio of applied fertilizer-N:manure-N over the last several decades (Fig. 1) could be responsible for subtle increases in the average initial δ15N values of NO3– with groundwater age (Verstraeten et al. 2000). Groundwater stratigraphic studies of variations in pesticide recharge rates are complicated by pesticide retention in the unsaturated zone and by difficulties accounting quantitatively for degradation products, which commonly are present at higher concentrations than the parent compounds in recharging groundwater. Pesticide degradation also is known to continue within the saturated zone. For example, Denver and Sandstrom (1991) report that the ratios of atrazine to desethylatrazine (DEA, the major degradation product) decreased with relative groundwater age at a site in Delaware, indicating that atrazine degradation occurred under oxic conditions within the aquifer over a period of years to decades. Nevertheless, the sum of the concentrations of atrazine and its major degradation products was directly related to NO3– concentrations and inversely related to groundwater age, as estimated qualitatively from dissolved silica concentrations (Denver and Sandstrom 1991). When combined with subsequent CFC analyses (Dunkle et al. 1993), those data are consistent with increasing application rates of both N and atrazine in the watershed in the second half of the 20th century. In contrast, a groundwater dating study in California, USA, indicates that concenHydrogeology Journal (2002) 10:153–179

Fig. 11 Relation between NO3– concentration in a water-supply well and time at a site near Cedar Falls, Iowa (modified from Schaap 1999). When sampled in 1998, the apparent age of the well water indicated by concentrations of CFC-11 and CFC-12 was approximately 22 years (assuming the sample was unmixed) and the δ15N value of the NO3– was 5‰. The long-term NO3– monitoring record is adjusted by 22 years to obtain a translated record of NO3– recharge history, for comparison with the regional fertilizer-N sales record (the right and left scales are not matched quantitatively). Regression of the adjusted record yields a baseline NO3– concentration of about 300 µmol/L when fertilizer sales were nil (Schaap 1999)

trations of a nematicide (DBCP) and its major degradation product peaked between 1950 and 1980 and then decreased, whereas concentrations of NO3– increased relatively steadily from about 1950 to 1990 (Burow et al. 1999). The discordance between the age-NO3– and ageDBCP relations is consistent with a ban on the use of DBCP in California after 1977, when N fertilizer use continued to increase. Though the amounts of pesticides and degradation products in these aquifers are small compared with the amounts applied in the time interval represented by the samples, studies like these indicate that groundwater records of pesticide recharge are retrievable in some environments. A hybrid approach to reconstructing recharge records is to combine long-term chemical monitoring data for wells or springs with more recent information about groundwater ages from the same wells or springs. If the source areas and flow systems have not been disturbed, it might be possible to apply a constant age offset to the chemical record of the well or spring to derive a chemical record of the corresponding recharge. A problem with this approach is that the interpretation of groundwater ages from monitored springs and wells (commonly water-supply wells with long open intervals) is likely to be complicated by mixing of waters of different ages (Figs. 4a and 6a). Nevertheless, the approach has been applied with some apparent success to relate land-surface N loading trends to delayed NO3– discharge trends. For example, Schaap (1999) shows that a trend of increasing NO3– concentrations in a water-supply well near Cedar Falls, Iowa, USA, between 1975 and 1995 is similar to a trend of increasing regional N fertilizer use between 1955 and 1975, consistent with the 22-year apparent CFC age of the well water sampled in 1998 (Fig. 11). DOI 10.1007/s10040-001-0183-3

169

Denitrification was not quantified in the study, but the presence of O2 and the relatively low δ15N value of the NO3– in 1998 indicate that it was probably not a major factor. The relatively simple (unmixed) interpretation of the CFC data is justified in part by the limited screened interval and distant recharge area of the well (Schaap 1999), which limit the distribution of ages in the pumped water; however, a more realistic assessment might be made by assuming a partially mixed age distribution (e.g., analogous to well F or to a well screened between points B and C in Fig. 4a). Katz et al. (2001) compare land-surface N loading records with NO3– concentration records for springs in the Suwanee River basin in Florida, USA, that had apparent ages of around 20 years, by applying several models of potential age distributions, including bimodal mixtures. Dissolved-gas analyses indicate that the dated samples were not altered substantially by denitrification. Effects of Unsaturated-Zone Transport Whereas groundwater stratigraphic analysis can provide a relatively robust record of recharge history, that record is not directly comparable with historical records of land-surface practices if transit times in the unsaturated zone are substantial. Most groundwater records of anthropogenic contamination have been retrieved in areas where water and solutes are transmitted through the unsaturated zone relatively quickly (a few years or less); however, other situations exist where much of the record still resides within the unsaturated zone. In the recent anthropogenic time scale, these situations have been investigated by evaluating groundwater samples for concordance between the distributions of 3H (moving with the water in the unsaturated zone) and some of the atmospheric gas tracers (moving with gas in the unsaturated zone). For example, comparisons between 3H, 3He, 85Kr, CFC, and SF6 concentrations in shallow coastal-plain aquifers in the eastern USA indicate that unsaturatedzone transit times are short in comparison to the average transit times in the saturated zone (several decades) (Dunkle et al. 1993; Ekwurzel et al. 1994; Szabo et al. 1996). In contrast, Johnston et al. (1998) interpret apparent age discordance between groundwater CFC-11 and 3H/3He data to be a result of delayed transport of CFC-11 through a 25-m-thick unsaturated zone in Ontario, Canada. The estimated residence time of water in the unsaturated zone at that site was about 10–15 years, which also presumably delayed the arrival of agricultural NO3– at the water table. Zoellmann et al. (2001) use apparent age discordance between 3H and SF6 in groundwaters at a site in Germany to resolve transit times of water in the unsaturated zone and transit times in the saturated zone to wells. In that paper, a flow model incorporating both sets of travel times (each of which is on the order of a few years, on average) is used to generate age-frequency distributions for calculating tracer concentrations in mixed groundwater samples, then the same model is used to rationalize Hydrogeology Journal (2002) 10:153–179

and predict changes in NO3– concentrations at water-supply wells. The model yields an age distribution within the saturated zone for SF6 that is roughly consistent with the exponential age-frequency distribution calculated for well E (Figs. 4a and 6a). In contrast, the combined transport in the unsaturated zone and saturated zone yields a groundwater age distribution for 3H that is more like the EPM (or PEM) distribution of well F. Relatively complete records of 3H deposition and recent agricultural NO3– contamination have remained within the unsaturated zone in some areas where the transit times of infiltrating water to the water table are long in comparison to the input history. A well-developed example of this type of situation in the English Chalk is summarized by Burt and Trudgill (1993), based in part on a study by Young et al. (1976). Vertical profiles of 3H and NO3– concentrations in unsaturated-zone pore waters at depths of 0–40 m are interpreted as a 30year record of precipitation infiltration with limited dispersion. Variations in the concentration of NO3– are correlated with documented changes in local agricultural practices; some of the highest concentrations apparently represent infiltration at times of major grassland plowing events. Similar approaches also have been applied to NO3– movement on longer time scales in more arid regions. For example, Edmunds and Gaye (1997) give evidence for widespread NO3– and Cl enrichment in the unsaturated zone in the southern Sahara, with vertical profiles 10–40 m deep that represent infiltration records over periods of 50–400 years. Hartsough et al. (2001) describe pore-water profiles from Nevada, USA, that are interpreted as representing infiltration of Cl and NO3– during the last 15,000 years. Effects of Irrigation and Drainage Irrigation and drainage have important effects on the distribution and rate of recharge of agricultural constituents and their subsequent preservation and distribution in surficial aquifers. Regional analyses indicate that NO3– concentrations are commonly relatively high in groundwaters beneath irrigated areas, possibly because of relatively rapid infiltration and leaching (Spalding and Exner 1993; Kolpin 1997). High rates of recharge and low groundwater residence times in some irrigated aquifers might also overwhelm biogeochemical remediation capacities or rates, thereby depressing redox fronts and enabling deeper penetration of O2 or NO3– than would occur in non-irrigated aquifers. Irrigation effects depend in part on whether surface water or groundwater is the major source of irrigation water. A simplistic view of some differences between these types of systems is illustrated in Fig. 12, where pumping a groundwater irrigation water-supply well pulls down young water that is removed from the vicinity of the well, whereas water-table mounding beneath a leaky surface-water irrigation supply canal pushes down young water that subsequently moves away with the regional flow system. Chemical and isotopic effects of irriDOI 10.1007/s10040-001-0183-3

170

Fig. 12a–c Hypothetical effects of irrigation and drainage on groundwater (GW) age profiles in surficial aquifers. a Irrigation by surface water, resulting in localized canal leakage and producing groundwater sheets or plumes with anomalous internal age gradients. The effect of canal leakage is drawn by analogy to some documented effects of point-source artificial recharge (DunkleShapiro et al. 1999) and line-source riverine recharge (Deak et al. 1996). b Irrigation by groundwater pumping from lower parts of the surficial aquifer, resulting in depressed isochrons and groundwater mixing near the pump. c Subsurface drainage, resulting in localized discharge and possible reduction in the rate of recharge to parts of the aquifer beneath the drain

gation depend in part on local contrasts between the compositions of the irrigation water and local precipitation-derived groundwater. Recharge of irrigation water pumped from deep parts of aquifers can be an important cause of apparent age discordance among different groundwater dating tracers if, for example, the irrigation water is relatively old and has low pre-bomb 3H or high bomb-peak 3H. During recharge, the old groundwater reequilibrates with atmospheric gases, including CFCs, He isotopes, and SF6, subsequently yielding a much younger age from the gas tracers than implied by the 3H. This effect would be difficult to distinguish from similar discordant effects of long transit times of precipitation recharge in the unsaturated zone. Isotopic signatures of surface-water irrigation include relatively high or low meteoric δ2H and δ18O values from river water originating at higher or lower elevations Hydrogeology Journal (2002) 10:153–179

than native groundwater, as well as deviations from meteoric isotope values caused by evaporation in the rivers or irrigation systems (IAEA 1981). Leakage water from irrigation canals may lack constituents derived from agricultural soils or fertilizers that are present in distributed recharge beneath agricultural fields. Plumes or sheets of water entering an aquifer from leaky irrigation canals and ditches might result in groundwater-age gradients that are highly variable and complex. At a watershed scale, relatively low average vertical velocities of distributed recharge beneath the fields might give way to higher apparent vertical velocities at depth, where plumes of canal recharge are present (Fig. 12a). This pattern is the opposite of the one more commonly expected of distributed recharge to a homogeneous aquifer (Fig. 4a). For example, in western Nebraska, USA, Verstraeten et al. (2000) identify irrigation water diverted from the North Platte River by an evaporated H and O isotope signature and relatively low average NO3– concentration. Contamination of the irrigated alluvial aquifer with NO3– requires substantial distributed recharge beneath agricultural fields, whereas local and regional NO3– dilution and disturbance of groundwater-age gradients are attributed to plumes of irrigation-canal leakage water. In contrast, Plummer et al. (2000) identify irrigation recharge in the eastern Snake River Plain aquifer of Idaho, USA, by a relatively low-elevation meteoric-water H and O isotope signature without evidence of evaporation. Minor spatial variations in groundwater NO3– concentrations are correlated in part with the relative contributions of new surface-water irrigation (low NO3–) and recycled groundwater irrigation (higher NO3–). Plummer et al. (2000) suggest that NO3– concentrations in the aquifer are generally kept low, despite high agricultural N application rates, by of the overall dilution effect of the imported Snake River water. Groundwater irrigation in the South Platte River alluvial aquifer in Colorado, USA, recycles large amounts of NO3– and also causes substantial homogenization of groundwater-age gradients and dissolved O2 gradients, depending on proximity to pumps (McMahon and Böhlke 1996; Fig. 12b). In the three studies cited, there is no consistent isotopic evidence of evaporation that could be attributed to the irrigation distribution systems. The strong seasonality of irrigation recharge can cause substantial deviations from normal regional patterns of water-table atmospheric-gas equilibration temperatures, which affect the evaluation of groundwaterdating tracers and excess N2 from denitrification. Temperatures of equilibration between atmospheric gases and recharging groundwaters are typically similar to mean annual ground temperatures, which are generally similar to mean annual air temperatures where water tables are not more than several tens of meters deep (Stute and Schlosser 1999). Recharge temperatures estimated from concentrations of Ne, Ar, and N2 in the North Platte River alluvial aquifer in western Nebraska are higher than the mean annual air temperature by about five degrees Celsius on average. and by as much as ten degrees Celsius DOI 10.1007/s10040-001-0183-3

171

down-gradient from leaky irrigation canals (Verstraeten et al. 2001). The anomalous equilibration temperatures are attributed to the fact that most of the annual recharge occurs in a relatively short period during the warm irrigation months, as indicated by water-level records in wells. Drainage effects depend in part on whether the drainage is on the surface or in the subsurface, and whether it is accompanied by water-level controls (Skaggs and Van Shilfgaarde 1999). Important effects of uncontrolled subsurface drainage typically include reductions in surface runoff carrying sediment-bound P, and increases in the recharge and discharge rates of NO3– (Hallberg et al. 1986; Skaggs et al. 1994). Lowering the water table with subsurface drains increases the thickness of the unsaturated zone (Fig. 12c), thereby potentially increasing the rate of nitrification relative to that of denitrification in soils. Subsurface drains also promote rapid discharge of NO3– in shallow groundwater soon after recharge, thereby possibly reducing the flux of NO3– to deeper parts of aquifers and also enabling the NO3– to bypass denitrifying environments in the saturated zone before discharging.

Interactions of Agricultural Contaminants in Recharge with Other Toxic Chemicals in Aquifers Indirect effects of agricultural recharge on groundwater chemistry include water–rock interactions within the saturated zone, some of which involve constituents with potential human and environmental health effects. Two major common features of recharging agricultural groundwater that contribute to such problems are the much higher concentration of potential oxidants represented by NO3– and the greater capacity for ion exchange represented by concentrations of H+, K+, Ca2+, Mg2+, and other cations. For example, Szabo et al. (1997) suggest that high concentrations of aqueous Ra that commonly exceed the drinking-water MCL of 5 picocuries/L might be maintained in groundwaters within the saturated zone by exchange of agriculturally enriched cations for Ra on aquifer solids. This suggestion is supported by subequal measured activities of 224Ra (t1/2=3.7 days) and 228Ra (t1/2=5.8 years) (Szabo et al. 2001), which would not be maintained in groundwater more than a few years old if these Ra isotopes were separated from their Th parent isotopes entirely within the unsaturated zone before recharge. High 224Ra concentrations caused indirectly by agricultural activity and from other sources are likely to be missed by some types of procedures for sampling and analysis of radionuclides in groundwater (Focazio et al. 2001; Szabo et al. 2001). Increasing the rate of recharge of electron acceptors in the form of NO3– could increase in the same proportion the rates, progress, or cumulative effects of microbial oxidation of sulfide minerals, organic carbon, or other electron donors, thereby potentially increasing the dissolved concentrations of associated toxic chemicals. Hydrogeology Journal (2002) 10:153–179

Assuming denitrification is the major NO3– transformation reaction, the flux of electron acceptors attributable to recharge of NO3– at the MCL of 714 µmol/L is 2.55 times the flux of electron acceptors attributable to recharge of air-saturated water containing 350 µmol/L of dissolved O2 (10 °C at sea level). Kölle et al. (1985) describe a situation in Germany where an increase in SO42– concentrations threatened to close drinking-water wells in an agricultural area. Results of biogeochemical studies indicate that the increase in SO42– was caused indirectly by an increase in the recharge rate of agricultural NO3–, which was denitrified by microbial oxidation of iron-sulfide minerals in the aquifer. If oxidation of pyrite (FeS2) were the only source of SO42– in an aquifer, then natural recharge of O2 at a concentration of 350 µmol/L could account for 187 µmol/L of SO42–: (14) whereas agricultural NO3 at 1–3 times the MCL would be responsible for an additional 476–1,428 µmol/L of SO42–, according to Eq. (5). Kölle et al. (1985) show that the sum of recharged SO42– plus SO42– produced by denitrification of agricultural NO3– was high enough to exceed the drinking-water limit (2,600 µmol/L in Germany at the time of the study), especially at times when grassland was initially converted to cropland, resulting in relatively high NO3– recharge rates. Another potential water-quality problem associated with enhanced sulfide oxidation by agricultural NO3– results from the fact that sulfide minerals in aquifers contain other trace elements, such as As, Zn, Ni, Cu, Cd, and Se, that could be released to solution. Acidity generated by pyrite oxidation in non-calcareous sediments could promote solubility and movement of some of the released metals. Farther downgradient from oxidation zones where SO42– is produced, the SO42– can be reduced subsequently by organic carbon (Frind et al. 1990). This redox cycle could result in re-precipitation and concentration of associated trace elements with new sulfide minerals formed in the zone of SO42– reduction. Thus, increasing the recharge rate of NO3– could increase indirectly the rate of migration and accumulation of a variety of redox-sensitive trace elements at the redox boundaries (e.g., similar to some “roll-front”-type mineral deposits). Rabenhorst and Fanning (1989) describe mid-Atlantic coastal-plain glauconitic “greensands” in which oxidative weathering released Zn and Ni from pyrite, possibly accompanied by Cd from another sulfide phase and by Zn, Ni, and Cu from glauconite. Recharge of agricultural NO3– has increased substantially the rates of oxidative weathering reactions, including pyrite oxidation, in the mid-Atlantic region, USA (Böhlke and Denver 1995; O’Connell 1998). Van Beek et al. (1989) describe a sandy aquifer containing agriculturally contaminated groundwater in which As concentrations exceeded the drinking-water limit in a narrow interval where SO42– concentrations were also high –

DOI 10.1007/s10040-001-0183-3

172 Fig. 13 Relation between depth below the water table and groundwater concentrations of a NO3–, b SO42–, and c As, in an aquifer beneath agricultural fields receiving manure applications (modified from Van Beek et al. 1989). Vertical dashed lines indicate concentrations that have been considered limiting for human consumption

(Fig. 13). Van Beek et al. (1989) conclude that recharging agricultural NO3– moved down in the aquifer until encountering sulfide minerals, where denitrification was coupled with sulfide oxidation, producing SO42–. Large amounts of As were released to solution when the sulfide minerals were oxidized; then the As was re-precipitated in a SO42– reduction zone farther downgradient. The observed As concentrations are considered to be higher as a result of denitrification than they would have been as a result of natural O2 reduction in the absence of the agricultural NO3–. Another result of sequential SO42– production and reduction in agriculturally contaminated aquifers is that the distribution of the SO42– with groundwater age sometimes resembles the pattern that might otherwise be attributable to atmospheric-acid deposition. Some aquifers in the northeastern USA apparently contain records of a peak in atmospheric SO42– deposition and recharge in the 1960s and 1970s followed by a subsequent decrease in more recent years (Robertson et al. 1989; Busenberg and Plummer 1996). Perhaps coincidentally, a peak in SO42– concentrations attributed to denitrification occurs in groundwaters with recharge dates in the 1960s and 1970s at the Princeton, Minnesota, site (Böhlke et al. 2002). Situations like these are generally resolvable with S-isotope analyses (Robertson et al. 1989; Böhlke et al. 2002), because 34S/32S ratios generally are higher in atmospheric SO42– than in sedimentary sulfide minerals, though this is not always the case. A potential benefit of high groundwater NO3– concentrations is oxidative degradation of problematic organic compounds, such as chlorinated solvents, aromatic hydrocarbons, or petroleum. Although these reactions are known to occur and have been tested locally for remediation purposes (Major et al. 1988; Burland and Edwards 1999), their general relevance with respect to non-pointsource NO3– in agricultural recharge depends in part on hydraulic interactions between groundwaters containing the complementary contaminants. Cozzarelli et al. (1999) report possible oxidative degradation of aromatic gasoline hydrocarbons by reduction of NO3– in agriculHydrogeology Journal (2002) 10:153–179

tural groundwaters in New Jersey. Mixing of groundwaters contaminated by agricultural NO3– and hydrocarbons is limited in many situations unless the groundwaters are disturbed.

Implications of Agricultural Contaminant Recharge History for Groundwater Discharge One of the important implications of the long-term recharge history of agricultural contamination preserved in groundwater is that it represents a transient but persistent potential source of NO3– in discharge to wells, springs, and surface waters. For the NO3– recharge history and groundwater stratigraphy shown in Fig. 4, hypothetical wells with short screened intervals exhibit delayed but relatively rapid responses to the downward movement of the 1960s–1970s NO3– “front” (Fig. 14, wells B and C). Wells with longer screens exhibit more gradual responses, in some cases damped and in some cases delayed, as increasing fractions of water containing high NO3– concentrations are discharged over time (wells D, E, and F). The response curve for well F is also applicable to a system with a substantial transit time in the unsaturated zone; that is, the age distribution obtained for well F would also be obtained from well E if the transit time in the unsaturated zone were approximately 35 years. Responses would differ in flow systems that were disturbed by irrigation or drainage (Fig. 12). In aquifers with reactive electron donors, the progress of agricultural NO3– might be retarded or, more likely, stopped abruptly at a particular depth where a lithologically-controlled redox transition occurs. For a well that is screened across a redox transition in a hydraulically homogeneous aquifer, the NO3– trend would be similar to the corresponding trend for the portion of the well above the redox transition but diluted by the denitrified water coming from below the transition (Fig. 14, well G). Some of the age distributions represented in Figs. 6a and 14 might be applicable not only to samples from wells but also to discharges in springs or streams, and they DOI 10.1007/s10040-001-0183-3

173

Fig. 14 Relation between time and hypothetical concentrations of NO3– in groundwater discharge. Each curve corresponds to a groundwater mixture derived from one of the wells (A–F) shown in Fig. 4a, with the age–frequency distribution shown in Fig. 6a and the NO3– recharge history shown in Figs. 4b and 8a (to the year 2000). After the year 2000, the concentration of NO3– in recharge is assumed to be zero

therefore might indicate important types of responses of surface waters to the common agricultural contamination history preserved in surficial aquifers. In particular, groundwater flow models commonly indicate that the logarithmic (exponential) age distribution is a good approximation for discharge from unconsolidated aquifers to small stream networks. The aquifer design and response curve given for well E are based on the example of a small stream at the Locust Grove site in Maryland whose baseflow NO3– concentration is substantially less than that of young groundwater in the watershed (Böhlke and Denver 1995; Figs. 5 and 8). The difference between the recharge flux and discharge flux of NO3– can be explained as a result of mixed discharging groundwaters of different ages and initial agricultural NO3– concentrations in the absence of denitrification (Böhlke and Denver 1995). Given a recharge history, age-frequency distribution, and configuration of denitrification zones, there is a possibility of forecasting discharge responses to proposed changes in the recharge fluxes. To illustrate this procedure, the curves in Fig. 14 include calculated responses to an abrupt cessation of NO3– recharge in the year 2000. In some cases, cessation of contamination in recharge results in a quick initial downward response but a long recovery (e.g., wells D and E), because the discharge includes both young and old water. In other cases, the initial downward response is delayed and the discharge concentration continues to increase for some time after the source is reduced (e.g., wells B, C, and F), because the discharge does not include young water. More realistic representations of age distributions, reactive barriers, and future contamination scenarios could be used to assess management options for effects on water-supply or surface-water quality. Another important implication of stratified agricultural contamination for groundwater/surface-water interacHydrogeology Journal (2002) 10:153–179

Fig. 15 Simplified representation of groundwater flow from an upland agricultural recharge area to a riparian wetland (modeled after Böhlke et al. 2002). Arrows indicate directions of flow as determined from gradients in hydraulic heads and groundwater ages (ranging from about 0–50 years). Lateral variations in the concentrations of agricultural contaminants discharging upward beneath the riparian wetland correspond in part to vertical variations in the ages and concentrations of NO3– in the recharge area, as indicated by representative groundwater ages in years: 1 young groundwater with low-NO3– non-agricultural recharge; 10 moderately young water with high-NO3– agricultural recharge; 30 moderately old water with low-NO3– agricultural recharge (result of denitrification or variations in historical input); 50 old water with non-agricultural or low-NO3– agricultural recharge

tions is related to spatial variations in discharge chemistry. Where groundwater flow lines converge upward in a discharge area, vertical gradients in groundwater ages and compositions from the recharge area are tilted over to become horizontal gradients (Tóth 1963; Winter 1976). Empirical studies prove that such gradients exist near small streams even where discharge is highly focused (Böhlke and Denver 1995; Modica et al. 1998; Böhlke et al. 2002). Horizontal gradients in the age and composition of groundwater entering a discharge area from below introduce substantial variability in the distributions of both conservative and reactive chemicals. For example, in a broad area of groundwater discharge beneath a riparian wetland, contaminant fluxes, ratios of stable constituents, and biogeochemical reaction rates exhibit horizontal gradients that are the result of historical variations in the vertical fluxes of contaminants across the water table in the recharge area (Fig. 15). These types of gradients complicate studies of hydrologic and geochemical processes beneath discharge areas, requiring multicomponent analytical approaches and detailed sampling in both the vertical and horizontal directions. In addition, where the discharge area is substantially smaller than the recharge area, horizontal age gradients are relatively abrupt, and vertical velocities are large in comparison with those in the recharge area. High discharge velocities needed to match lower recharge velociDOI 10.1007/s10040-001-0183-3

174

ties might offset some of the natural NO3– remediation potential of denitrifying environments in riparian or stream-bottom sediments. Furthermore, reactive shallow riparian environments are commonly bypassed by NO3–bearing groundwaters that follow deep flow paths from broad agricultural recharge areas to focused discharge areas in streams (Böhlke and Denver 1995). Recharge and discharge of agricultural contaminants might also interfere with large-scale studies of continental weathering rates, oceanic loading rates, and geochemical provenance of dissolved constituents in rivers. For example, Semhi et al. (2000) propose that carbonate dissolution by nitric acid generated by nitrification of N fertilizers has caused a significant increase above natural levels in the dissolved fluxes of bicarbonate, Ca, and Mg from the Garonne River basin in France. Artificial drainage has altered substantially the relative magnitudes of surface-water chemical loads delivered by surface runoff and groundwater discharge (Skaggs et al. 1994). Böhlke and Horan (2000) call attention to the potential impact of agriculturally enhanced recharge of Sr on regional studies of Sr-isotope geochemistry in groundwaters and rivers. Chemicals enriched in groundwater recharge either directly or indirectly by agricultural practices have varying influence on surface-water compositions during low baseflow, high baseflow, and stormflow conditions.

Conclusions Common agricultural practices have caused substantial increases in the groundwater recharge fluxes and concentrations of N, Cl, Ca, Mg, and a variety of other elements in the last half of the 20th century. These increases have resulted in a transient agricultural chemical signal in aquifers that have groundwater residence times of decades or more. Some elements are added to the hydrosphere directly as components of fertilizers or other additives, whereas some are associated indirectly with fertilizer loadings or land disturbance through enhanced rates of leaching of natural sources in soils, owing to physical and biological changes and to increased acidity or ionic strength of agricultural recharge. In addition, high fluxes of NO3– in recharging groundwater have resulted in unnaturally high concentrations of chemical species released from aquifer materials by oxidation reactions in the saturated zone, because the electron demand of NO3– in agriculturally contaminated recharge is commonly several times higher than that of dissolved O2 in uncontaminated recharge. Indirect agricultural effects on groundwater composition are of geochemical significance and are useful as tracers of agricultural groundwater, and some have achieved levels locally that could cause important environmental problems (e.g., SO42–, Ra, and As). Although uncertainty and debate still exist about the sources and pathways of NO3– in groundwater recharge, several different types of evidence indicate that excessive applications of artificial fertilizers and manures Hydrogeology Journal (2002) 10:153–179

have contributed substantially to the recharge rate of NO3– in many parts of the world: 1. annual to decadal monitoring of drainage water and shallow groundwater compositions indicates rapid responses to changing application rates; 2. groundwater dating and stratigraphic analysis of surficial (unconfined) aquifers yield long-term contamination patterns that are correlated locally with known cropping practices and fertilization histories; and 3. regional and local mass-balance calculations indicate excess quantities of N available for recharge. In addition, the N-isotope ratios in recharging NO3– are commonly different beneath fields with mainly manure, fertilizer, or N2 fixation as annual N sources, indicating that the anthropogenic sources were not overwhelmed by turnover of natural N sources in the soils. This conclusion is supported by rapid changes in the isotopic composition of recharging NO3– in response to changes in land use in some areas. Agricultural studies on the behavior of N and other important constituents in soils and crops have not been coupled universally with studies that follow the progress of those constituents below the root zone, in groundwater recharge, and through aquifers. Regional statistical analyses relating both ends of this continuum (from land surface to groundwater discharge) provide important clues about processes within the continuum, but not enough to guide local understanding or decision-making. Filling this gap requires detailed information about local geologic and hydrologic conditions controlling transport and reaction in the subsurface. In addition to the major constituents commonly associated with agricultural contamination, more work is needed on the indirect effects that agricultural groundwater infiltration and recharge have on water–rock interactions such as leaching, weathering, and biogeochemical transformations in both soils and aquifers, including dissolution and precipitation of toxic or environmentally sensitive elements that are not necessarily included in agricultural additives. Combining groundwater dating with other chemical and isotopic analyses of discrete samples within aquifers provides relatively direct assessments of several recharge issues that are important in agricultural settings: 1. groundwater recharge rates, including spatial variability and irrigation losses; 2. recharge fluxes of non-point-source contaminants for comparison with agricultural loadings; 3. historical variations in recharge of agricultural chemicals for comparison with changes in land-use practices; 4. natural remediation processes in aquifers; and 5. residence times of water and contaminants in aquifers that influence discharge responses to change. Most of the same principles and limitations that govern groundwater dating are relevant to retrieval of groundwater stratigraphic records of agricultural contamination, including effects of dispersion, potential degradation, DOI 10.1007/s10040-001-0183-3

175

variable transport properties and transit times in the unsaturated zone and saturated zone, and the benefits of multiple analytical data sets. Contaminant trends revealed by groundwater stratigraphy have focused mainly on the last half of the twentieth century, largely because the environmental tracers most commonly used for dating groundwaters (e.g., CFCs, 3H, 3He, and SF6) are most useful for waters recharged during that period of time. This is an important time scale in studies of agricultural effects, because rapid increases in fertilizer use and the recharge rates of agricultural contaminants also occurred during that period in many areas of the world; however, it is important to recognize that earlier recharge characteristics remain largely undocumented. In some areas, 14C dating has yielded evidence for substantial recharge concentrations of NO3– over time-scales of 103–104 years, in some cases almost certainly as a result of natural processes. But a major problem in reconstructing the history of anthropogenic changes in recharge chemistry is the lack of convenient environmental tracers for time scales of 50–1,000 years, when agricultural practices underwent major changes, including forest clearing and plowing, conversion of grassland to cropland, and introduction of imported fertilizers. Studies involving groundwater 4He accumulation rates (Solomon 1999) or cosmogenic isotopes like 32Si (t ≈140 years) (Morgenstern 1999) or 39Ar 1/2 (t1/2=269 years) (Loosli et al. 1999) have some potential to provide information about agricultural effects on recharge before the widespread use of artificial fertilizers. Future applications of environmental tracers to agricultural contamination studies should include more sophisticated applications of multiple-tracer analyses to systems with substantial transit times in both the unsaturated zone and the saturated zone, and to groundwater mixtures in discharges from wells and springs. Only recently has it been practical to obtain analyses of three or more different dating tracers in a single sample. Advances in understanding of complex systems are expected to occur as long-term trends in multiple-tracer data are developed at individual sites, and by the application of more realistic transport simulations for multiple tracers, including agricultural contaminants. Acknowledgements Many of the recent studies of groundwater recharge and agricultural contamination described in this paper were made possible by conceptual and analytical advances and assistance in groundwater dating and stable isotopes by E. Busenberg, T.B. Coplen, L.N. Plummer, K. Revesz, P. Schlosser, and others. Field studies were motivated in part by collaborations with G. Delin, J. Denver, B. Katz, M. Landon, P. McMahon, M. O’Connell, K. Prestegaard, M. Tuttle, I. Verstraeten, R. Wanty, and others. Laboratory assistance was provided by many people, including G. Casile, T. Coplen, M. Doughten, J. Hannon, S. Mroczkowski, H.P. Qi, J. Wayland, and P. Widman. Funding was provided by the National Research Program, National Water Quality Assessment Program, Toxic Substances Hydrology Program, and Cooperative Studies Program of the US Geological Survey; and by the US National Science Foundation, US Environmental Protection Agency, and US Department of Agriculture. Manuscript reviews by R.A. Alexander, M.R. Burkart, J.M. Denver, and L.N. Plummer are much appreciated. Hydrogeology Journal (2002) 10:153–179

References Aeschbach-Hertig W, Peeters F, Beyerle U, Kipfer R (1999) Interpretation of dissolved atmospheric gases in natural waters. Water Resour Res 35:2779–2792 Aeschbach-Hertig W, Peeters F, Beyerle U, Kipfer R (2000) Paleotemperature reconstruction from noble gases in ground water taking into account equilibration with trapped air. Nature 405:1040–1044 Amberger A, Schmidt H-L (1987) Natürliche Isotopengehalte von Nitrat als Indikatoren für dessen Herkunft. Geochim Cosmochim Acta 51:2699–2705 Baker JL, Johnson HP (1981) Nitrate-nitrogen in tile drainage as affected by fertilization. J Environ Qual 10:519–522 Baker JL, Timmons DR (1994) Fertilizer management effects on leaching of labeled nitrogen for no-till corn in field lysimeters. J Environ Qual 23:305–310 Ballentine CJ, Hall CM (1999) Determining paleotemperature and other variables by using an error-weighted, nonlinear inversion of noble gas concentrations in water. Geochim Cosmochim Acta 63:2315–2336 Barbash JE, Resek EA (1996) Pesticides in Ground Water. Ann Arbor Press, Ann Arbor, 588 pp Barbash JE, Thelin GP, Kolpin DW, Gilliom RJ (1999) Distribution of major herbicides in ground water of the United States. US Geological Survey Water-Resources Investigations Report 98–4245, 57 pp Barnes CJ, Jacobsen G, Smith GD (1992) The origin of high-nitrate groundwaters in the Australian arid zone. J Hydrol 137: 181–197 Barry DA, Goorahoo D, Goss MJ (1993) Estimation of nitrate concentrations in groundwater using a whole farm nitrogen budget. J Environ Qual 22:767–775 Bergersen FJ, Peoples MB, Turner GL (1988) Isotopic discriminations during the accumulation of nitrogen by soybeans. Aust J Plant Physiol 15:407–430 Berner RA (1971) Principles of chemical sedimentology. McGraw–Hill, New York, 240 pp Blicher-Mathiesen G, McCarty GW, Nielsen LP (1998) Denitrification and degassing in groundwater estimated from dissolved dinitrogen and argon. J Hydrol 208:16–24 Böhlke JK, Denver JM (1995) Combined use of groundwater dating, chemical, and isotopic analyses to resolve the history and fate of nitrate contamination in two agricultural watersheds, Atlantic coastal plain, Maryland. Water Resour Res 31:2319– 2339 Böhlke JK, Ericksen GE, Revesz K (1997) Stable isotope evidence for an atmospheric origin of desert nitrate deposits in northern Chile and southern California. Chem Geol 136:135– 152 Böhlke JK, Horan M (2000) Strontium isotope geochemistry of ground waters and streams affected by agriculture, Locust Grove, Maryland. Appl Geochem 15:599–609 Böhlke JK, Smith RL, Revesz KM, Yoshinari T (2000) Quantifying reaction progress and isotope fractionation effects in denitrification: Integration of results for nitrate, nitrite, nitrous oxide, and nitrogen gas in shallow contaminated ground water at Cape Cod, Massachusetts (Abstract). EOS, Trans Am Geophys Union 81:S186 Böhlke JK, Wanty R, Tuttle M, Delin G, Landon M (2002) Denitrification beneath the recharge area and discharge area of a transient agricultural ground-water nitrate plume in Minnesota. Water Resour Res (in press) Böttcher J, Strebel O, Duynisveld WHM (1989) Kinetik und Modellierung gekoppelter Stoffumsetzungen im Grundwasser eines Lockergesteins-Aquifers. Geol Jahrb 51:3–40 Böttcher J, Strebel O, Duynisveld WHM, Frind EO (1991) Reply to comment on “Modeling of multicomponent transport with microbial transformation in groundwater: the Fuhrberg case”. Water Resour Res 27:3275–3278 Böttcher J, Strebel O, Voerkelius S, Schmidt H-L (1990) Using isotope fractionation of nitrate-nitrogen and nitrate-oxygen for DOI 10.1007/s10040-001-0183-3

176 evaluation of microbial denitrification in a sandy aquifer. J Hydrol 114:413–424 Bouwer H (1989) Linkages with ground water. In: Follett RF (ed) Nitrogen management and ground water protection. Elsevier, Amsterdam, pp 363–372 Burkart MR, James DE (1999) Agricultural-nitrogen contributions to hypoxia in the Gulf of Mexico. J Environ Qual 28:850–859 Burkart MR, James DE (2002) Agricultural nitrogen trends in the Mississippi Basin, 1949–1997. Agricultural production systems of the 21st century: challenges and opportunities in plant nutrient management, ASA Special Publication (in press) Burkart MR, Kolpin DW, Jaquis RJ, Cole KJ (1999) Agrichemicals in ground water of the midwestern USA: Relations to soil characteristics. J Environ Qual 28:1908–1915 Burkart MR, Stoner JD (2001) Nitrogen in groundwater associated with agricultural systems. In: Follett RF, Hatfield JL (eds) Nitrogen in the environment: sources, problems, and management. Elsevier, Amsterdam, pp 123–145 Burland SM, Edwards EA (1999) Anaerobic benzene biodegradation linked to nitrate reduction. Appl Environ Microbiol 65: 529–533 Burow KR, Panshin SY, Dubrovsky NM, Vanbrocklin D, Fogg GE (1999) Evaluation of processes affecting 1,2-dibromo-3-chloropropane (DBCP) concentrations in ground water in the eastern San Joaquin Valley, California: analysis of chemical data and ground-water flow and transport simulations. US Geological Survey Water-Resources Investigations Report 99–4059, 57 pp Burt TP, Heathwaite AL, Trudgill STE (1993) Nitrate: processes, patterns, and management. Wiley, New York, 444 pp Burt TP, Trudgill ST (1993) Nitrate in groundwater. In: Burt TP, Heathwaite AL, Trudgill ST (eds) Nitrate. processes, patterns, and management. Wiley, New York, pp 213–238 Busenberg E, Plummer LN (1996) Concentrations of chlorofluorocarbons and other gases in ground water at Mirror Lake, New Hampshire. In: US Geological Survey Water Resources Investigations Report 94–4015, pp 151–158 Busenberg E, Plummer LN (2000) Dating young groundwater with sulfur hexafluoride: natural and anthropogenic sources of sulfur hexafluoride. Water Resour Res 36:3011–3030 Ceazan ML, Thurman EM, Smith RL (1989) Retardation of ammonium and potassium transport through a contaminated sand and gravel aquifer: the role of cation exchange. Environ Sci Technol 23:1402–1408 Chichester FW, Smith SJ (1978) Disposition of 15N-labelled nitrate applied during corn culture in field lysimeters. J Environ Qual 7:227–233 Cleveland WS (1979) Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 74:829–836 Cook PG, Böhlke JK (1999) Determining timescales for groundwater flow and solute transport. In: Cook PG, Herczeg A (eds) Environmental tracers in subsurface hydrology. Kluwer, Dordrecht, pp 1–30 Cook PG, Herczeg AL (1999) Environmental tracers in subsurface hydrology. Kluwer, Dordrecht Cook PG, Solomon DK (1995) The transport of atmospheric trace gases to the water table: implications for groundwater dating with chlorofluorocarbons and krypton-85. Water Resour Res 31:263–270 Cook PG, Solomon DK (1997) Recent advances in dating young groundwater: chlorofluorocarbons, 3H/3He and 85Kr. J Hydrol 191:245–265 Cozzarelli IM, Herman JS, Baedecker MJ, Fischer JM (1999) Geochemical heterogeneity of a gasoline-contaminated aquifer. J Contam Hydrol 40:261–284 Deak J, Deseo E, Böhlke JK, Revesz K (1996) Isotope hydrology studies in the Szigetköz region, northwest Hungary. Isotopes in Water Resources Management, International Atomic Energy Agency 1, pp 419–432 Deeley GM, Reinhard M, Stearns SM (1991) Transformation and sorption of 1,2-dibromo-3-chloropropane in subsurface samples collected at Fresno, California. J Environ Qual 20:547–556 Hydrogeology Journal (2002) 10:153–179

Delin GN, Healy RW, Landon MK, Böhlke JK (2001) Effects of topography and soil properties on recharge at two sites in an agricultural field. Journal of the American Water Resour Assoc 36:1401–1416 Densmore JN, Böhlke JK (2000) Use of nitrogen isotopes to determine sources of nitrate contamination in two desert basins in California. In: Reichard EG, Hauchman FS, Sancha AM (eds) Interdisciplinary perspectives on drinking water risk assessment and management. Int Assoc Hydrol Sci Publ 260:63–73 Denver JM (1989) Effects of agricultural practices and septicsystem effluent on the quality of water in the unconfined aquifer in parts of eastern Sussex County, Delaware. Delaware Geological Survey Report of Investigations 45, 66 pp Denver JM, Sandstrom MW (1991) Distribution of dissolved atrazine and two metabolites in the unconfined aquifer, southeastern Delaware. US Geological Survey Water-Resources Investigations report 91–4034, pp 314–318 Dunkle SA, Plummer LN, Busenberg E, Phillips PJ, Denver JM, Hamilton PA, Michel RL, Coplen TB (1993) Chlorofluorocarbons (CCl3F and CCl2F2) as dating tools and hydrologic tracers in shallow ground water of the Delmarva Peninsula, Atlantic Coastal Plain, United States. Water Resour Res 29:3837– 3860 Dunkle-Shapiro SD, LeBlanc D, Schlosser P, Ludin A (1999) Characterizing a sewage plume using the 3H-3He dating technique. Ground Water 37:861–878 EC (European Community) (1980) EC Directive relating to the quality of water intended for human consumption. Official Journal of the European Community L229, 80/778/EEC [available in Department of the Environment Circular 20/82, HMSO, London, 1982] Edmunds WM (1999) Groundwater nitrate as a paleoenvironmental indicator. In: Armannsson H (ed) Geochemistry of the earth’s surface. Balkema, Rotterdam, pp 35–38 Edmunds WM, Gaye CB (1997) Naturally high nitrate concentrations in groundwaters from the Sahel. J Environ Qual 26: 1231–1239 Ekwurzel B, Schlosser P, Smethie WMJ, Plummer LN, Busenberg E, Michel RL, Weppernig R, Stute M (1994) Dating of shallow ground water: comparison of the transient tracers 3H/3He, chlorofluorocarbons, and 85Kr. Water Resour Res 30:1693– 1708 Ericksen GE (1983) The Chilean nitrate deposits. Am Sci 71: 366–374 Fan AM, Steinberg VE (1996) Health implications of nitrate and nitrite in drinking water: an update on methemoglobinemia occurrence and reproductive and developmental toxicity. Regul Toxicol Pharmacol 23:35–43 Fenelon JM, Moore RC (1998) Transport of agrichemicals to ground and surface water in a small central Indiana watershed. J Environ Qual 27:884–894 Flury M (1996) Experimental evidence of transport of pesticides through field soils – a review. J Environ Qual 25:25–45 Focazio MJ, Szabo Z, Kraemer TF, Mullin AH, Barringer TH, dePaul VT (2001) Occurrence of selected radionuclides in ground water used for drinking water in the United States: a reconnaissance survey, 1998. US Geological Survey, WaterResources Investigations report 00–4273, 39 pp Fogg GE, Rolston DE, Decker DL, Louie DT, Grismer ME (1998) Spatial variation in nitrogen isotope values beneath nitrate contamination sources. Ground Water 36:418–426 Follett RF (1989) Nitrogen management and ground water protection. Developments in agricultural and managed-forest ecology 21. Elsevier, New York, p 395 Follett RF, Hatfield JL (2001) Nitrogen in the environment: sources, problems, and management. Elsevier, Amsterdam Freeze RA, Cherry JA (1979) Groundwater. Prentice-Hall, Englewood Cliffs, 604 pp Frind EO, Duynisveld WHM, Strebel O, Boettcher J (1990) Modeling of multicomponent transport with microbial transformation in groundwater: the Fuhrberg case. Water Resour Res 26:1707–1719 DOI 10.1007/s10040-001-0183-3

177 Galanter M, Sigman DM, Levy H, Böhlke JK, Lipshultz F, Steig E (2000) Controls on the oxygen isotopic composition of atmospherically-derived nitrate (Abstract). EOS, Trans Am Geophys Union 81:F191 Goolsby DA, Battaglin WA, Lawrence GB, Artz RS, Aulenbach BT, Hooper RP, Keeney DR, Stensland GJ (1999) Flux and sources of nutrients in the Mississippi–Atchafalaya River basin. Topic 3 report for the integrated assessment on hypoxia in the Gulf of Mexico. NOAA Coastal Ocean Program Decision Analysis Series No 17, NOAA Coastal Ocean Program, Silver Spring, 130 pp Gormly JR, Spalding RF (1979) Sources and concentrations of nitrate-nitrogen in ground water of the central Platte region, Nebraska. Ground Water 17:291–301 Guimond RJ (1990) Radium in fertilizers: the environmental behavior of radium, International Atomic Energy Agency, Technical Reports Series 310 1990, pp 113–128 Hall DH, Risser DW (1993) Effects of agricultural nutrient management on nitrogen fate and transport in Lancaster County, Pennsylvania. Water Resour Bull 29:55–76 Hallberg GR (1986) Overview of agricultural chemicals in ground water. Proc Conf Agricultural Impacts on Ground Water, Omaha, Nebraska. National Water Well Association, pp 1–63 Hallberg GR (1989) Pesticide pollution of groundwater in the humid United States. Agric Ecosyst Environ 26:299–367 Hallberg GR, Baker JL, Randall GW (1986) Utility of tile-line effluent studies to evaluate the impact of agricultural practices on groundwater. Agricultural Impacts on Ground Water. National Water Well Association, Dublin, OH, pp 298–328 Hallberg GR, Keeney DR (1993) Nitrate. In: Alley WM (ed) Regional ground-water quality. Van Nostrand Reinhold, New York, pp 297–322 Hamilton PA, Denver JM, Phillips PJ, Shedlock RJ (1993) Waterquality assessment of the Delmarva Peninsula, Delaware, Maryland, and Virginia: effects of agricultural activities on, and the distribution of, nitrate and other inorganic constituents in the surficial aquifer. US Geological Survey Open-File Report 93-40, 149 pp Hamilton PA, Helsel DR (1995) Effects of agriculture on groundwater quality in five regions of the United States. Ground Water 33:217–226 Hartsough P, Tyler SW, Sterling J, Walvoord M (2001) A 14.6-kyr record of nitrogen flux from desert soil profiles as inferred from vadose zone pore waters. Geophys Res Lett 28:2955– 2958 Haynes RJ (1999) Fate and recovery of 15N-labelled fertilizer urea applied to winter wheat in spring in the Canterbury region of New Zealand. J Agric Sci 133:125–130 Heathwaite AL, Sharpley AN, Gburek WJ (2000) A conceptual approach for integrating phosphorus and nitrogen management at watershed scales. J Environ Qual 29:158–166 Heaton THE (1986) Isotopic studies of nitrogen pollution in the hydrosphere and atmosphere – a review. Chem Geol 59:87– 102 Heaton THE, Talma AS, Vogel JC (1983) Origin and history of nitrate in confined groundwater in the western Kalahari. J Hydrol 62:243–262 Herczeg AL, Edmunds WM (1999) Inorganic ions as tracers. In: Cook PG, Herczeg AL (eds) Environmental tracers in subsurface hydrology. Kluwer, Dordrecht, pp 31–77 Howard KWF (1985) Denitrification in a major limestone aquifer. J Hydrol 76:265–280 Howarth R, Anderson D, Cloern J, Elfring C, Hopkinson C, Lapointe B, Malone T, Marcus N, McGlathery K, Sharpley A, Walker D (2000) Nutrient pollution of coastal rivers, bays, and seas. Ecological Society of America, 15 pp Howarth RW, Billen G, Swaney D, Townsend A, Jaworski N, Lajtha K, Downing JA, Elmgren R, Caraco N, Jordan T, Berendse F, Freney J, Kudeyarov V, Murdoch P, Zhao-Liang Z (1996) Regional nitrogen budgets and riverine N & P fluxes for the drainages to the North Atlantic Ocean: natural and human influences. Biogeochemistry 35:75–139 Hydrogeology Journal (2002) 10:153–179

Hübner H (1986) Isotope effects of nitrogen in the soil and biosphere. In: Fritz P, Fontes JC (eds) Handbook of environmental geochemistry, vol 2. The terrestrial environment B. Elsevier, Amsterdam, pp 361–425 IAEA (International Atomic Energy Agency) (1981) Stable isotope hydrology. IAEA, Technical Report Series 210, Vienna, 337 pp Johnson DA, Cole DW (1980) Anion mobility in soils: relevance to nutrient transport from forest ecosystems. Environ Int 3:79–90 Johnston CT, Cook PG, Frape SK, Plummer LN, Busenberg E, Blackport RJ (1998) Ground water age and nitrate distribution within a glacial aquifer beneath a thick unsaturated zone. Ground Water 36:171–180 Jordan TE, Weller DE (1996) Human contributions to terrestrial nitrogen flux: assessing the sources and fates of anthropogenic fixed nitrogen. BioScience 46:655–664 Kaplan N, Magaritz M (1986) A nitrogen-isotope study of the sources of nitrate contamination in groundwater of the Pleistocene coastal plain aquifer, Israel. Water Res:131–135 Katz BG, Böhlke JK, Hornsby HD (2001) Timescales for nitrate contamination of spring waters, northern Florida, USA. Chem Geol 179:167–186 Keeney DR (1986) Sources of nitrate to ground water. CRC Crit Rev Environ Control 16:257–304 Kendall C (1998) Tracing nitrogen sources and cycles in catchments. In: Kendall C, McDonnell JJ (eds) Isotope tracers in catchment hydrology. Elsevier, Amsterdam, pp 519–576 Kendall C, Aravena R (1999) Nitrate isotopes in groundwater systems. In: Cook PG, Herczeg AL (eds) Environmental tracers in subsurface hydrology. Kluwer, Boston, pp 261–297 Kladivko EJ, Brown LC, Baker JL (2001) Pesticide transport to subsurface tile drains in humid regions of North America. Crit Rev Environ Sci Technol 31:1–62 Knisel WG, Foster GR, Leonard RA (1983) CREAMS: A system for evaluating management practices. In: Schaller FW, Bailey GW (eds) Agricultural management and water quality. Iowa State University Press, Ames, pp 178–199 Kohl DH, Shearer GB, Commoner B (1971) Fertilizer nitrogen – contribution to nitrate in surface water in a corn belt watershed. Science 174:1331–1334 Kölle W, Strebel O, Böttcher J (1985) Formation of sulfate by microbial denitrification in a reducing aquifer. Water Supply 3:35–40 Kolpin DW (1997) Agricultural chemicals in groundwater of the midwestern United States: relations to land use. J Environ Qual 26:1025–1037 Kolpin DW, Barbash JE, Gilliom RJ (1998) Occurrence of pesticides in shallow ground water of the United States: initial results from the National Water-Quality Assessment Program. Environ Sci Technol 32:558–566 Kolpin DW, Burkart MR, Thurman EM (1994) Herbicides and nitrate in near-surface aquifers in the midcontinental United States, 1991. US Geological Survey Water-Supply Paper 2413, 34 pp Kozinski J, Szabo Z, Zapecza OS, Barringer TH (1995) Natural radioactivity in, and inorganic chemistry of, ground water in the Kirkwood–Cohansey aquifer system, southern New Jersey, 1983–1989. US Geological Survey Water-Resources Investigations Report 92–4144, 130 pp Kreitler CW (1975) Determining the source of nitrate in ground water by nitrogen isotope studies. University of Texas, Bureau of Economic Geology Report of Investigations 83, 57 pp Kreitler CW (1979) Nitrogen-isotope ratio studies of soils and groundwater nitrate from alluvial fan aquifers in Texas. J Hydrol 42:147–170 Kreitler CW, Jones DC (1975) Natural soil nitrate: the cause of the nitrate contamination of ground water in Runnels County, Texas. Ground Water 13:53–61 Landon MK, Delin GN, Komor SC, Regan CP (2000) Relation of pathways and transit times of recharge water to nitrate concentrations using stable isotopes. Ground Water 38:381–395 DOI 10.1007/s10040-001-0183-3

178 Landon MK, Delin GN, Nelson KJ, Regan CP, Lamb JA, Larson SJ, Capel PD, Anderson JL, Dowdy RH (1997) Water-quality and hydrogeologic data used to evaluate the effects of farming systems on ground-water quality at the Management Systems Evaluation Area near Princeton, Minnesota, 1991–1995. US Geological Survey Open-File Report 97–22, p 25 Leonard RA, Knisel WG, Still AD (1987) GLEAMS: Ground-water loading effects of agricultural management systems. Trans Am Soc Agric Eng 30:1403–1417 Libra RD, Hallberg GR, Hoyer BE (1987) Impacts of agricultural chemicals on groundwater quality in Iowa. In: Fairchild DM (ed) Ground water quality and agricultural practices. Lewis, Chelsea, pp 185–217 Loosli HZ, Lehmann BE, Smethie WMJ (1999) Noble gas radioisotopes: 37Ar, 85Kr, 39Ar, 81Kr. In: Cook PG, Herczeg AL (eds) Environmental tracers in subsurface hydrology. Kluwer, Boston, pp 379–396 Major DW, Mayfield CI, Barker JF (1988) Biotransformation of benzene by denitrification in aquifer sand. Ground Water 26:8–14 Mariotti A, Landreau A, Simon B (1988) N isotope biogeochemistry and natural denitrification process in groundwater – application to the chalk aquifer of northern France. Geochim Cosmochim Acta 52:1869–1878 Martin GE, Snow DD, Kim E, Spalding RF (1995) Simultaneous determination of argon and nitrogen. Ground Water 33:781– 785 McMahon PB, Böhlke JK (1996) Denitrification and mixing in a stream–aquifer system: effects on nitrate loading to surface water. J Hydrol 186:105–128 McMahon PB, Böhlke JK (2001) Radiocarbon ages, paleorecharge conditions, and nitrogen transport in the central High Plains Aquifer. 4th Int Symp Applied Isotope Geochemistry, Pacific Grove, CA, June 2001, Program and Abstracts, pp 42–44 McMahon PB, Böhlke JK, Bruce BW (1999) Denitrification in marine shales in northeastern Colorado, USA. Water Resour Res 35:1629–1642 Meisinger JJ, Randall GW (1991) Estimating nitrogen budgets for soil-crop systems. In: Follet RF, Keeney DR, Cruse RM (eds) Managing nitrogen for groundwater quality and farm profitability. Soil Science Society of America, Madison, p 85 Michalski G, Thiemens MH (2000) Mass independent fractionation in nitrate aerosols (abstract). EOS, Transactions of the American Geophysical Union 81, F120 Modica E, Buxton HT, Plummer LN (1998) Evaluating the source and residence times of ground-water seepage to streams. Water Resour Res 34:2797–2810 Molina JAE, Richards K (1984) Simulation models of the nitrogen and carbon cycle in the soil–water–plant system, NCSWAP; guide for the preparation of input data files and execution of NCSWAP. University of Minnesota, Department of Soil Science, Soil Series 116, St Paul Morgenstern U (1999) Silicon-32. In: Cook PG, Herczeg AL (eds) Environmental tracers in subsurface hydrology. Kluwer, Dordrecht, pp 498–502 Nolan BT (2001) Relating nitrogen sources and aquifer susceptibility to nitrate in shallow ground waters of the United States. Ground Water 39:290–299 Normand B, Recous S, Vachaud G, Kengni L, Garino B (1997) Nitrogen-15 tracers combined with tensio-neutronic method to estimate the nitrogen balance of irrigated maize. Soil Sci Soc Am J 61:1508–1518 O’Connell ME (1998) Processes affecting the discharge of nitrate from a small agricultural watershed over a range of flow conditions, Maryland coastal plain. PhD Thesis, University of Maryland O’Connell ME, Böhlke JK, Prestegaard KL (1997a) Low recharge rates indicated by CFC age gradients in a low-conductivity aquifer, Maryland coastal plain (Abstract). Geological Society of America Abstracts with Programs 29, A76 O’Connell ME, Böhlke JK, Prestegaard KL (1997b) Processes affecting discharge of nitrate from a small agricultural watershed Hydrogeology Journal (2002) 10:153–179

at different flow conditions, Maryland coastal plain (Abstract). EOS, Transactions of the American Geophysical Union 78, S168 Ostrom NE, Knoke KE, Hedin LO, Robertson GP, Smucker AJM (1998) Temporal trends in nitrogen isotope values of nitrate leaching from an agricultural soil. Chem Geol 146:219–227 Parkhurst DL, Plummer LN (1993) Geochemical models. In: Alley WM (ed) Regional ground-water quality. Van Nostrand Reinhold, New York, pp 199–225 Paul EA, Paustian K, Elliott ET, Cole CV (1997) Soil organic matter in temperate agroecosystems. CRC Press, Boca Raton, 414 pp Peoples MB, Herridge DF (1990) Nitrogen fixation by legumes in tropical and subtropical agriculture. Adv Agron 44:155–223 Phillips FM, Mattick JL, Duval TA, Elmore D, Kubik PW (1988) Chlorine 36 and tritium from nuclear weapons fallout as tracers for long-term liquid and vapor movement in desert soils. Water Resour Res 24:1877–1891 Pignatello JJ (1986) Ethylene dibromide mineralization in soils under aerobic conditions. Appl Environ Microbiol 51:588–592 Pionke HB, Urban JB (1985) Effect of agricultural land use on groundwater quality in a small Pennsylvania watershed. Ground Water 23:68–80 Plummer LN, Busenberg E (1999) Chlorofluorocarbons. In: Cook PG, Herczeg AL (eds) Environmental tracers in subsurface hydrology. Kluwer, Dordrecht, pp 441–478 Plummer LN, Busenberg E, Böhlke JK, Nelms DL, Michel RL, Schlosser P (2001) Groundwater residence times in Shenandoah National Park, Blue Ridge Mountains, Virginia, USA: a multi-tracer approach. Chem Geol 179:93–111 Plummer LN, Busenberg E, Drenkard S, Schlosser P, Ekwurzel B, Weppernig R, McConnell JB, Michel RL (1998) Flow of river water into a karstic limestone aquifer – 2. Dating the young fraction in groundwater mixtures in the Upper Floridan aquifer near Valdosta, Georgia. Appl Geochem 13:1017–1043 Plummer LN, Rupert MG, Busenberg E, Schlosser P (2000) Age of irrigation water in ground water from the eastern Snake River Plain aquifer, south-central Idaho. Ground Water 38: 264–283 Porter LK (1995) Depleted 15N carryover, leaching and uptake for three years of irrigated corn. J Contam Hydrol 20:209–226 Postma D, Boesen C, Kristiansen H, Larsen F (1991) Nitrate reduction in an unconfined sandy aquifer: water chemistry, reduction processes, and geochemical modeling. Water Resour Res 27:2027–2045 Puckett LJ, Cowdery TK, Lorenz DL, Stoner JD (1999) Estimation of nitrate contamination of an agro-ecosystem outwash aquifer using a nitrogen mass-balance budget. J Environ Qual 28:2015–2025 Rabenhorst MC, Fanning DS (1989) Pyrite and trace metals in glauconitic parent materials of Maryland. Soil Sci Soc Am J 3:1791–1797 Randall GW, Mulla DJ (2001) Nitrate nitrogen in surface waters as influenced by climatic conditions and agricultural practices. J Environ Qual 30:337–344 Rao PSC, Alley WM (1993) Pesticides. In: Alley WM (ed) Regional ground-water quality. Van Nostrand Reinhold, New York, pp 345–382 Reilly TE, Plummer LN, Phillips PJ, Busenberg E (1994) The use of simulation and multiple environmental tracers to quantify groundwater flow in a shallow aquifer. Water Resour Res 30:421–433 Robertson WD, Cherry JA (1989) Tritium as an indicator of recharge and dispersion in a groundwater system in central Ontario. Water Resour Res 25:1097–1109 Robertson WD, Cherry JA, Schiff SL (1989) Atmospheric sulfur deposition 1950–1985 inferred from sulfate in groundwater. Water Resour Res 25:1111–1123 Roessler CE (1990) Control of radium in phosphate mining, beneficiation, and chemical processing. The environmental behavior of radium. International Atomic Energy Agency, Technical Reports Series 310, vol 2, pp 269–279 DOI 10.1007/s10040-001-0183-3

179 Schaap BD (1999) Concentrations and possible sources of nitrate in water from the Silurian-Devonian aquifer, Cedar Falls, Iowa. US Geological Survey Water Resources Investigations Report 99–4106, 19 pp Schlosser P, Shapiro SD, Stute M, Aeschbach-Hertig W, Plummer LN, Busenberg E (1998) Tritium/3He measurements in young groundwater: chronologies for environmental records. Isotope techniques in the study of environmental change. International Atomic Energy Agency, Vienna, pp 165–189 Schlosser P, Stute M, Sonntag C, Münnich KO (1989) Tritiogenic 3He in shallow groundwater. Earth Planet Sci Lett 94:245–256 Schnabel RR, Urban JB, Gburek WJ (1993) Hydrologic controls in nitrate, sulfate, and chloride concentrations. J Environ Qual 22:589–596 Semhi K, Suchet PA, Clauer N, Probst J-L (2000) Impact of nitrogen fertilizers on the natural weathering-erosion processes and fluvial transport in the Garonne basin. Appl Geochem 15: 865–878 Shaffer MJ, Halvorson AD, Pierce FJ (1991) Nitrate leaching and economic analysis package (NLEAP): model description and application. In: Follett RF, Keeney DR, Cruse RM (eds) Managing nitrogen for groundwater quality and farm profitability. Soil Science Society of America, Madison, 285 pp Sharpley A, Foy B, Withers P (2000) Practical and innovative measures for the control of agricultural phosphorus losses to water: an overview. J Environ Qual 29:1–9 Simms JT, Simard RR, Joern BC (1998) Phosphorus loss in agricultural drainage: historical perspective and current research. J Environ Qual 27:277–293 Skaggs RW, Breve MA, Gilliam JW (1994) Hydrologic and water quality impacts of agricultural drainage. Cri Rev Environ Sci Technol 24:1–32 Skaggs RW, Van Shilfgaarde J (1999) Agricultural drainage. American Society of Agronomy, Agronomy Monograph 38, Madison, pp 1328 Smith RA, Schwartz GE, Alexander RB (1997) Regional interpretation of water-quality monitoring data. Water Resour Res 33:2781–2798 Solomon DK (1999) 4He in groundwater. In: Cook PG, Herczeg A (eds) Environmental tracers in subsurface hydrology. Kluwer, Dordrecht, pp 425–439 Solomon DK, Poreda RJ, Cook PG, Hunt A (1995) Site characterization using 3H/3He ground-water ages, Cape Cod, MA. Ground Water 33:988–996 Solomon DK, Schiff SL, Poreda RJ, Clarke WB (1993) A validation of the 3H/3He method for determining groundwater recharge. Water Resour Res 29:2951–2962 Spalding RF, Exner ME (1993) Occurrence of nitrate in groundwater – a review. J Environ Qual 22:392–402 Stevenson FJ (1982) Origin and distribution of nitrogen in soil. In: Stevenson FJ (ed) Nitrogen in agricultural soils. American Society of Agronomy, pp 1–42 Stute M, Schlosser P (1999) Atmospheric noble gases. In: Cook PG, Herczeg AL (eds) Environmental tracers in subsurface hydrology. Kluwer, Dordrecht, pp 349–377 Szabo BJ, Rice DE, McLeod CL, Barringer TH (1997) Relation of distribution of radium, nitrate, and pesticides to agricultural land use and depth. US Geological Survey, Water-Resources Investigations Report 96–4165A, 107 pp Szabo Z, dePaul VT, Kraemer TF, Parsa B (2001) Occurrence of radium-224 in water of the unconfined Kirkwood-Cohansey aquifer system, southern New Jersey, and comparison to that of radium-226 and radium-228. US Geological Survey Water Resources Investigations Report Szabo Z, Rice DE, Plummer LN, Busenberg E, Drenkard S, Schlosser P (1996) Age dating of shallow ground water with chlorofluorocarbons, tritium/helium 3, and flow path analysis, southern New Jersey coastal plain. Water Resour Res 32:1023–1038 Tilman D, Fargione J, Wolff B, D’Antonio C, Dobson A, Howarth R, Schindler D, Schlesinger WH, Simberloff D, Swackhamer D (2001) Forecasting agriculturally driven global environmental change. Science 292:281–284 Hydrogeology Journal (2002) 10:153–179

Tóth J (1963) A theoretical analysis of groundwater flow in small drainage basins. J Geophys Res 68:4795–4812 Trachtenberg E, Ogg C (1994) Potential for reducing nitrogen pollution through improved agronomic practices. Water Resour Bull 30:1109–1118 Trudell MR, Gillham RW, Cherry JA (1986) An in-situ study of the occurrence and rate of denitrification in a shallow unconfined sand aquifer. J Hydrol 83:251–268 USEPA (US Environmental Protection Agency) (1986) Maximum contaminant levels (subpart B of part 141, National primary drinking-water regulations), US Code of Federal Regulations, Title 40, parts 100–149, revised July 1, 1986. US Environmental Protection Agency, Washington, DC, pp 524–528 USNRC (US National Research Council) (1993) Soil and water quality: an agenda for agriculture. Committee on Long-Range Soil and Water Conservation, Board on Agriculture, National Research Council. National Academy Press, Washington, DC, 487 pp Van Beek CGEM, Hettinga FAM, Straatman R (1989) The effects of manure spreading and acid deposition upon groundwater quality in Vierlingsbeek, the Netherlands. Int Assoc Hydrol Sci Publ 185:155–162 Verstraeten IM, Böhlke JK, Kraemer TF (2000) Groundwater/surface-water interactions and sources of nitrogen, radon, and uranium in a irrigated alluvial aquifer, Nebraska. Tracers and modeling in hydrogeology. Int Assoc Hydrogeol Publ 262: 525–531 Verstraeten IM, Steele GV, Böhlke JK, Kraemer TF, Wilson KE, Carnes AE (2001) Selected field and analytical methods and analytical results in the Dutch Flats area, western Nebraska, 1995–99. US Geological Survey Open-File Report 00–413, 53 pp Vogel JC (1967) Investigation of groundwater flow with radiocarbon. Isotopes in hydrology. International Atomic Energy Agency, Vienna, pp 355–368 Vogel JC, Talma AS, Heaton THE (1981) Gaseous nitrogen as evidence for denitrification in groundwater. J Hydrol 50:191–200 Wagenet RJ, Hutson JL (1989) LEACHM: leaching estimation and chemistry model. A process based model of water and solute movement, transformations, plant uptake, and chemical reactions in the unsaturated zone. Water Resources Institute, Cornell University Walters DT, Malzer GL (1990) Nitrogen management and nitrification inhibitor effects on nitrogen-15 urea: II. nitrogen leaching and balance. Soil Sci Soc Am J 54:122–130 Welch AH, Westjohn DB, Helsel DR, Wanty RB (2000) Arsenic in ground water of the United States: occurrence and geochemistry. Ground Water 38:589–604 Wilkinson DH, Blevins DW (1999) Observations on preferential flow and horizontal transport of nitrogen fertilizer in the unsaturated zone. J Environ Qual 28:1568–1580 Wilson GB, Andrews JN, Bath AH (1990) Dissolved gas evidence for denitrification in the Lincolnshire limestone groundwaters, eastern England. J Hydrol 113:51–60 Winter TC (1976) Numerical simulation analysis of the interaction of lakes and ground water. US Geological Survey Professional Paper 1001, 45 pp Wolterink TJ (1979) Identifying sources of subsurface nitrate pollution with stable nitrogen isotopes. US Environmental Protection Agency, EPA-600/4–79–050, 150 pp Young CP, Hall ES, Oakes DB (1976) Nitrate in groundwater – studies on the Chalk near Winchester, Hampshire. Water Research Centre Technical Report 31 Zoellmann K, Kinzelbach W, Fulda C (2001) Environmental tracer transport (3H and SF6) in the saturated and unsaturated zones and its use in nitrate pollution management. J Hydrol 240:187–205 Zuber A (1986) Mathematical models for the interpretation of environmental radioisotopes in groundwater systems. In: Fritz P, Fontes JC (eds) Handbook of environmental geochemistry, vol 2. The terrestrial environment B. Elsevier, Amsterdam, pp 1–59 DOI 10.1007/s10040-001-0183-3