Journal of Hydrology - California Water Science Center … ·...

Journal of Hydrology 409 (2011) 157–171

Contents lists available at SciVerse ScienceDirect

Journal of Hydrology

journal homepage: www.elsevier .com/ locate / jhydrol

Subsurface transport of orthophosphate in five agricultural watersheds, USA

Joseph L. Domagalski a,⇑, Henry M. Johnson b

a US Geological Survey, California Water Science Center, 6000 J Street, Placer Hall, Sacramento, CA 95819, United Statesb US Geological Survey, Oregon Water Science Center, 2130 SW 5th Avenue, Portland, OR 97201, United States

a r t i c l e i n f o

Article history:Received 5 November 2010Received in revised form 18 May 2011Accepted 6 August 2011Available online 16 August 2011This manuscript was handled by LaurentCharlet, Editor-in-Chief, with the assistanceof Jiin-Shuh Jean, Associate Editor

Keywords:AgricultureHydrochemistryGroundwater protectionGroundwater/surface-water relationsHydrochemical modelingOrthophosphate

0022-1694/$ - see front matter Published by Elsevierdoi:10.1016/j.jhydrol.2011.08.014

Abbreviation: Ortho P, orthophosphate.⇑ Corresponding author. Tel.: +1 916 278 3077; fax

E-mail address: [email protected] (J.L. Domagalski).

s u m m a r y

Concentrations of dissolved orthophosphate (ortho P) in the unsaturated zone, groundwater, tile drains,and groundwater/stream water interfaces were assessed in five agricultural watersheds to determine thepotential for subsurface transport. Concentrations of iron oxides were measured in the aquifer materialand adsorption of ortho P on oxide surfaces was assessed by geochemical modeling. Attenuation of orthoP in these aquifers was attributed primarily to sorption onto iron oxides, and in one location onto clayminerals. Only one location showed a clear indication of phosphorus transport to a stream from ground-water discharge, although groundwater did contribute to the stream load elsewhere. Subsurface ortho Pmovement at a site in California resulted in a plume down gradient from orchards, which was attenuatedby a 200 m thick riparian zone with natural vegetation. Iron oxides had an effect on phosphorus move-ment and concentrations at all locations, and groundwater chemistry, especially pH, exerted a major con-trol on the amount of phosphorus adsorbed. Groundwater pH at a site in Maryland was below 5 and thatresulted in complete sequestration of phosphorus and no movement toward the stream. Geochemicalmodeling indicated that as the surfaces approached saturation, groundwater concentrations of ortho Prise rapidly.

Published by Elsevier B.V.

1. Introduction

Nutrient transport (nitrogen and phosphorus compounds) fromagricultural fields to surface and groundwater is one of the mostserious environmental problems throughout the world (Salvia-Cas-tellvi et al., 2005; National Research Council, 2005; Diaz et al.,2004). Eutrophication, or excessive growth of algae, of surfacewater bodies impairs their use for recreation and domestic con-sumption and limits their use by native biota. Where surface wateris chlorinated and used for drinking water, increased levels of algaecontribute to the formation of trihalomethanes and haloaceticacids (Scully et al., 1988; Oliver and Shindler, 1980), which havebeen shown to be both carcinogenic and mutagenic in laboratorytoxicity studies (Boorman, 1999). Recently, Sprague and Lorenz(2009) reported an increase in total phosphorus concentrationsin streams in the central United States between 1993 and 2003.The increase in total phosphorus was significantly correlated withan increase in fertilizer use. No significant change in total phospho-rus was found in streams in the western or eastern United States.

In natural systems, phosphorus is mainly cycled through aplant-soil dynamic. However in agricultural systems, phosphorus

B.V.

: +1 916 278 3071.

is removed at harvest and must be replenished with either mineralfertilizer or manure. Soils may be over applied with respect tophosphorus because many management programs are tied tonitrogen control or crop needs and phosphorus management isnot always considered (Sharpley et al., 2003). This issue is commonin areas using manure, which typically has a N:P ratio lower thanneeded by most crops. Application of manure to meet N require-ments often results in the application of P exceeding crop demand.Surface water contamination with phosphorus due to runoff frommanured fields (Klausner et al., 1976; Hergert et al., 1981; Allenand Mallarino, 2008) and agriculture in general (Baker et al.,1975; Burwell et al., 1977; Sharpley and Syers, 1979; Vollenweider,1968) is a well-studied topic.

It is known that phosphorus sorbs to soil particles. Phosphorussorbs onto oxides of iron and aluminum (Hemwell, 1957; Zhangand Huang, 2007), clay minerals (Parfitt, 1978) and to calcium car-bonate (Cole et al., 1953). It has been assumed that phosphorustransport to groundwater is negligible because of sorption andtherefore earlier studies on phosphorus transport have been fo-cused solely on surface water pathways (Baker et al., 1975; Burwellet al., 1977; Sharpley and Syers, 1979; Vollenweider, 1968). Atten-uation of phosphorus movement and plant availability by interac-tions onto soil particles has been recognized by previous research(Gunjigake and Wada, 1981; Anderegg and Naylor, 1988; Parfitt,1978). More recent studies (Devau et al., 2009), for example, have

http://dx.doi.org/10.1016/j.jhydrol.2011.08.014

mailto:[email protected]

http://dx.doi.org/10.1016/j.jhydrol.2011.08.014

http://www.sciencedirect.com/science/journal/00221694

http://www.elsevier.com/locate/jhydrol

158 J.L. Domagalski, H.M. Johnson / Journal of Hydrology 409 (2011) 157–171

further refined understanding of various factors, such as soil pH,with respect to phosphorus mobility or plant availability and theimportance of clay minerals, in addition to iron and aluminum oxi-des for controlling the distribution of phosphorus between soil andwater.

It was hypothesized for this study that phosphorus movementin the unsaturated zone and along a groundwater flow path to astream or tile drain would be different with location because ofhydrological and chemical factors. Processes affecting the rate ofwater movement through the unsaturated zone are likely to beimportant for phosphorus transport as well as processes that affectsorption or dissolution. Chemical equilibrium may not be achievedin the unsaturated zone because of short residence times at somelocations, if transport times are fast such as might occur becauseof preferential flow paths, and because of competition for sorptionsites from competing ions. It was further hypothesized thatgroundwater processes affecting phosphorus would more likelybe closer to equilibrium relative to the unsaturated zone as resi-dence time increases and groundwater chemistry stabilizes allow-ing equilibrium conditions between the fluid and the solids.Groundwater/surface water interactions should also differ withlocation relative to the amount and direction of exchange of wateracross the streambed interface. Unsaturated zone, groundwater,and streambed materials and water chemistry are expected to af-fect the transport or immobilization of phosphorus. In a previousstudy at a set of five watersheds across a range of climatic and soiltypes within the United States, Domagalski et al. (2008a) showedthat baseflow contributed more than 20% of the annual ortho Pload in some of the streams of this study. Tesoriero et al. (2009)examined the baseflow index of some of the same streams, as thisstudy, with a range of groundwater contributions to discharge.They found that the baseflow index was negatively correlated within-stream ortho P concentrations at most sites. However, at sitesdominated by baseflow or having favorable geochemical condi-tions (e.g. aquifers having low dissolved oxygen concentrationsor lacking in iron and aluminum oxides, clay minerals, and calciumcarbonate) groundwater contributions of ortho P to streams areimportant (Tesoriero et al., 2009). The purpose of this investigationwas to examine the geochemical conditions and reactions affectingortho P concentrations in the subsurface environmental compart-ments (unsaturated zone, groundwater, groundwater/surfacewater interface) of the five watersheds from the Domagalskiet al. (2008a) and Tesoriero et al. (2009) studies. Geochemicalmodeling was used to determine the level of saturation of orthoP with respect to minerals and sorption onto iron oxides. The re-sults of the modeling were used to interpret the subsurface orthoP concentrations and assess the potential for phosphorus transportalong the groundwater flow paths.

2. Study areas

Five study areas were chosen in agricultural watersheds nestedwithin larger basins that are being investigated by the US Geolog-ical Survey (USGS) NAWQA Program (US Geological Survey, 2010a)(Fig. 1). Specific characteristics of the aquifers, crop types, andphosphorus use for each study area are shown in Table 1.

Study areas were chosen from a previous investigation that in-cluded important agricultural systems and to cover a range ofhydrologic settings using the hydrologic landscape concept (Capelet al., 2008). The focus of the previous investigation was on nitro-gen and pesticide transport. Transport of water, nitrogen, and pes-ticides had previously been examined within and across differenthydrologic compartments including the atmosphere, the unsatu-rated zone, groundwater, and the stream/groundwater interface(Fisher and Healy, 2008; Green et al., 2008; Essaid et al., 2008;

Puckett et al., 2008; Steele et al., 2008; Domagalski et al., 2008a,2008b; Hancock et al., 2008). Watersheds selected were the DR2Drain within the Yakima River Basin, Washington; the lower Mer-ced River, within the San Joaquin-Tulare Basins, California; MapleCreek within the Central Nebraska Basins, Nebraska; the Leary We-ber Ditch within the White River Basin, Indiana; and Morgan Creekwithin the Delmarva Peninsula, Maryland. Detailed site locationshave been described previously (Capel et al., 2008).

The lower Merced River Basin is located in the eastern San Joa-quin Valley of California and occupies 832 km2. The climate issemi-arid with most of the rainfall occurring in the winter. The ma-jor crops are orchards and vineyards with some row crops and pas-ture. Soils are mainly loams and sandy loams. There is a narrowstrip of natural vegetation adjacent to the river. All agriculture re-quires irrigation and most of that water is transported from outsideof the immediate study area. The groundwater flow in the immedi-ate study area is towards the Merced River. The riverbed consists ofcoarse sand and gravel alluvial deposits. Recharge in the study areaoccurs primarily at the land surface with secondary contributionsfrom subsurface inflow from upgradient areas. About two thirdsof areal recharge is attributed to irrigation return flow, and theremainder to precipitation. Irrigation drives a downward gradientof recharge water in the agricultural fields and orchards (Domagal-ski et al., 2008b). About one third of the applied phosphorus ischemical and the remainder is manure (Gronberg and Kratzer,2006).

The DR2 Drain is located in south-central Washington and occu-pies 5.5 km2. The climate is also arid to semi-arid, with very littleor no rainfall occurring during the growing season, and crops in-clude corn, other row crops, orchards, vineyards, and dairies. Mostof the soils are well drained, are sandy to clayey in texture, and arevery shallow to very deep. The upper 3–10 m of aquifer materialare silty sands and clayey silts (Payne et al., 2007). Irrigation wateris supplied from outside the DR2 basin from the Yakima Riverthrough a series of canals. Shallow subsurface flow in this areahas been modified by an extensive system of buried drains (Capelet al., 2008). The shallow subsurface flow discharges to the DR2Drain and contributes greatly to year-round flow (Capel et al.,2008). The bed of the DR2 Drain consists of silty sands and clayeysilts similar to the aquifer material and, locally, has relatively highorganic matter content. Phosphorus use data are not available forthis watershed, but it is known that a combination of chemical fer-tilizer and manure is used, with considerable amounts of manureapplication.

Maple Creek is a tributary to the Elkhorn River in eastern Ne-braska and occupies 956 km2. Land use is predominately agricul-ture consisting mainly of corn, soybeans, and alfalfa. Glacial tilland Quaternary-age loess mantles the hills and forms terraces oversand and gravel deposits that make up the primary aquifer materi-als. Soils are fine-textured aeolian sand, silt, and loess. The climateis humid continental and precipitation is supplemented with irri-gation from groundwater to supply crop water needs. Ground-water discharge supports baseflow during the late growing seasonand winter; this discharge occurs where stream channels intersectthe sand and gravel aquifer (Fredrick et al., 2006; Puckett et al.,2008). Amounts of phosphorus from manure are not available.

The Leary Weber Ditch is a 7.5 km2 watershed in Central Indi-ana, with the agriculture dominated by corn and soybeans. The cli-mate is humid continental and precipitation supplies all crop waterdemand. The watershed is underlain by glacial till deposits, andsoils are loams or silty loams. The till-derived soils have poordrainage, which necessitates the use of tile drains to prevent waterlogging of the soils (Baker et al., 2006). Most of the applied phos-phorus is chemical, and manure use is minimal (Lathrop, 2006).

Morgan Creek is a 31 km2 watershed in eastern Maryland with-in the Chester River Basin. Crops include corn, soybeans and grains.

Table 1Aquifer characteristics, crop types, and phosphorus use for each location.

Washington California Nebraska Indiana MarylandParameterCrop Corn/orchards Almonds/Corn Corn/soybeans Corn/soybeans Corn/soybeans

Irrigation rate (yr, cm) (Fisherand Healy, 2008)

74.4 120 20.3 0 0

Phosphorus use, 2004 (kg/ha) Not available 21 (Gronberg andKratzer, 2006)

45 (Fredrick et al., 2006) 67 (Lathrop, 2006) 48 (Hancock andBrayton, 2006)

Water table depths (m) 4.8 7.2 4.4–21.8 1.0–1.8 4.4–11.1Soil texture Silty clay to fine sand Fine to medium sand Deep, well to poorly drained,

silt and silty claySilty clay loam to siltloam

Fine sandy loam tomedium sand

Average weight (%) sand 52.6 92.4 56 14 62.1Average weight (%) silt 42.5 6.5 36 73.5 31.2Average weight (%) clay 4.9 1.1 8 12.5 6.7Bulk mineralogy Quartz, plagioclase,

feldspar, calciteQuartz, plagioclase,feldspar

Quartz, plagioclase, feldspar Quartz, dolomite,plagioclase, calcite

Quartz, clay, feldstar

Organic matter, % 2 0.4 4.4 3.6 2.2

Explanation

State boundary

Basin Location

1000 km

Washington

California Nebraska

IndianaMaryland

1.

2.

3.

4.5.

Approximate study area location

1.) Central Columbia Plateau and Yakima River Basin2.) San Joaquin-Tulare Basins--Lower Merced River3.) Central Nebaska Basins--Maple Creek4.) White, Great and Little Miami Rivers Basins--Leary Weber Ditch5.) Potamac River Basin and Delmarva Peninsula-- Morgan Creek

Fig. 1. Map showing area of larger river basins within which the study areas are located. Locations of actual study areas are approximate.

J.L. Domagalski, H.M. Johnson / Journal of Hydrology 409 (2011) 157–171 159

The climate is humid subtropical (Hancock and Brayton, 2006) andprecipitation supplies most of the crop water needs except duringdry periods. Soils are mainly well to moderately well drained finesilt loams with some clay. The watershed is underlain by quartzsands and gravels (Owens and Denny, 1979; Owens and Minard,1979). Within the study reach, the entire flood plain consists of a1–2 m thick layer of heavy silt and clay. This impervious clay layerprevents direct movement of groundwater through the streambed;however groundwater discharges from seepage zones at the lateralmargins of the flood plain. Both chemical fertilizer and manure areused (Hancock and Brayton, 2006).

3. Methods

3.1. Field methods

Unsaturated-zone monitoring locations were equipped withinstrumentation to measure surficial soil temperature, soil heat

flux, soil water matric potential and soil moisture at depth in orderto estimate the flux of water out of the soil and the amount ofwater in the soil. Polyvinylchloride lysimeters with ceramic porouscups were installed at various depths to collect unsaturated zonewater for chemical analysis by applied suction. The shallowestlysimeter was located just below the root zone and the deepest justabove the water table. In all cases, lysimeters were always installedat the upper portion of the flow path location. Additional lysime-ters were installed at other locations in order to obtain more infor-mation on recharge characteristics.

Details of the study design have been described previously (Ca-pel et al., 2008). In general, clusters of monitoring wells were in-stalled at sites along a groundwater transect that was oriented inthe general direction of groundwater flow. The direction of flowwas determined by water levels and monitoring of hydraulic head.Wells were drilled at appropriate depths at each study site in orderto capture the chemistry and hydrology of the recently rechargedwater and water of various ages along the flow path to the stream.


Because of the drainage pattern at the Indiana location, the studydesign was modified and tile drains were sampled.

River transects, where appropriate, were established to gatherinformation on the interaction of groundwater and surface water.Transects were established down gradient of the groundwater flowpath. Wells were installed in the riverbed at 0.5 m and 2–3 mdepth and equipped with pressure transducers and temperatureloggers to measure direction of groundwater flow. Five sets ofthese well pairs were installed including a set on each bank andthree pairs within the channel. To collect water samples, 4.5 cmstainless steel drive-point tips with 2-cm screens were co-locatedwith each well set. Nylon tubing (6 mm) was attached to thedrive-point tips and routed to a shelter on the bank for samplingaccess.

Recharge and water travel times from land surface to variousdepths within the unsaturated zone were estimated by means ofa bromide tracer test conducted at each study site. Bromide wasapplied to a level of 12–15 g/m2 to the surface of a 12 m2 sectionof land overlying the lysimeter clusters. Water samples collectedfrom the lysimeters over time were analyzed by ion chromatogra-phy to define bromide breakthrough curves. Analyses of these sam-ples enabled calculation of a range of travel times, velocities, andspecific water fluxes for each site. Travel time and velocity of watermovement in the unsaturated zone was also estimated by a previ-ously described piston-flow model (Fisher and Healy, 2008).

Water samples were collected by various methods dependingon the environmental compartment. Samples were collected fromthe lysimeters, after 3–24 h of vacuum, through nylon tubing. Thegroundwater samples were collected with stainless steel pumpsand Teflon tubing, and the surface water samples were collectedin Teflon bottles and split into equivalent aliquots using a Teflonchurn. Further details about sample collection methods are avail-able from Capel et al. (2008). Methods for collection of surfacewater or overland flow samples are available from Capel et al.(2008), or Domagalski et al. (2008a).

3.2. Analytical methods

Water samples were processed by filtering through 0.45 lm fil-ters for dissolved inorganic constituents (orthophosphate, majorcations and anions, iron, manganese, and alkalinity). Dissolvedorthophosphate (ortho P) was analyzed using US EnvironmentalProtection Agency (US EPA) method 365.1 (US Environmental Pro-tection Agency, 1993). The method detection limit for ortho P was0.008 mg/L. Orthophosphate concentrations are expressed in thispaper as mg/L, or mmol/L as P. Major ions (Na, Ca, K, Mg, Cl, andSO4) were analyzed by the methods of Fishman (1993), Fishmanand Friedman (1989), and the American Public Health Association(1998). The estimated age of groundwater samples was deter-mined using measurements of sulfur hexafluoride (Busenbergand Plummer, 2000) and corrected for recharge temperature usingmeasurements of dissolved nitrogen and argon in the groundwatersamples (US Geological Survey, 2010b). Quality assurance con-sisted of the collection of routine field and equipment blanks andreplicates. Analysis of the blanks indicated that samples were freeof artifacts, cross-contamination, or carry-over, and the analysis ofreplicates indicated that reproducibility was consistent with theanalytical method.

Selected samples of sediment from the unsaturated zone,groundwater flow path, and at the groundwater/surface waterlocations were analyzed for labile phosphorus concentrationsusing two methods. Samples were collected by either soil augeror through a hollow-stem auger from the drilling. The Olsen meth-od (Olsen et al., 1954) utilizes a sodium bicarbonate solution at pH8.5, and is thought to be most appropriate for alkaline or calcare-ous soils. The Mehlich-3 method (Mehlich, 1984) utilizes two acids

(acetic and nitric), two salts (ammonium fluoride and ammoniumnitrate) and a chelating agent (ethylenediaminetetraacetic acid).Neither method recovers total phosphorus, but only that looselybound to the soil or sediment. A more thorough phosphorus diges-tion method of the US Environmental Protection Agency (US EPA3051) was also used (US Environmental Protection Agency,1995a, 1995b). Samples from drill cores were also analyzed forthe apparent concentration of amorphous iron oxide by extractionwith 0.5 M HCl plus hydroxylamine hydrochloride (Lovley andPhillips, 1987).

3.3. Geochemical modeling

Geochemical modeling was completed using the code PHREEQC(Parkhurst and Appelo, 1999). The PHREEQC model calculateschemical equilibria using a thermodynamic database, and per-forms other simulations with solutions and solid phases, includingmixing and adsorption. In this analysis, PHREEQC was used to cal-culate the saturation state of various minerals containing phospho-rus relative to phosphorus concentrations measured in water andalso to calculate the equilibrium surface adsorption of phosphorus(PO3�

4 , HPO2�4 , H2PO�4 ) and competing ions onto iron oxide surfaces.

The default databases contain thermodynamic data for a surfacenamed ‘‘Hfo’’ (Hydrous ferric oxide) that are derived from Dzom-bak and Morel (1990). Two sites are defined in the databases: astrong binding site, Hfo_s, and a weak binding site Hfo_w. Dzom-bak and Morel (1990) used 0.2 mol weak sites and 0.005 molstrong sites per mol Fe, a surface area of 53,300 m2/mol Fe, and agram-formula weight of 89 g Hfo/mol Fe; to be consistent withtheir model, the relative number of strong and weak sites shouldremain constant as the total number of sites varies (Parkhurstand Appelo, 1999). Total amounts of iron oxide present were calcu-lated from the hydroxylamine hydrochloride analysis as previouslydescribed. Water chemistry, as measured from each well, was usedto calculate saturation indices and sorption competition, and to setoxidation–reduction levels. Complete details of the sorption calcu-lation model are given by Parkhurst and Appelo (1999).

4. Results and discussion

4.1. Phosphorus concentrations and water chemistry

Concentrations of ortho P for each hydrological compartment(unsaturated zone, groundwater, groundwater/surface water inter-face, streams, and tile drains (Indiana site only) are shown in Fig. 2,along with the US EPA criteria for total phosphorus in specific eco-regions (US Environmental Protection Agency, 2010). Concentra-tions of ortho P in rainfall were below the laboratory-reportinglimit of 0.008 mg/L. Therefore rain was not considered a sourceof phosphorus to the soil.

Concentrations of ortho P at each location were assessed amongthe hydrological compartments by a post hoc multiple comparisontest (Simes, 1986; Hochberg, 1988) on the results of pairwise two-sample Wilcoxon rank sum tests. Statistical designations are givenon Fig. 2 and show the hydrological compartments within a studyarea that have median ortho P concentrations which are similar ordifferent at a level of significance of 0.05.

The highest concentrations of ortho P were measured in theunsaturated zone of the Washington site (concentration maxi-mum: 4.5 mg/L). However, concentrations drop rapidly from theunsaturated zone to groundwater. Surface water concentrationswere slightly higher than concentrations at the groundwater orgroundwater/surface water sites suggesting an additional sourceof ortho P. Overland flow contributed up to the 19% of thedischarge at the DR2 Drain (Domagalski et al., 2008a), and is a


possible source. However, the median concentration of ortho P inoverland flow water was 0.009 mg/L as compared to the medianconcentration of ortho P in the DR2 Drain of 0.1 mg/L. Concentra-tions of ortho P in tile drains were even less. At the California loca-tion, a similar drop in concentrations from the unsaturated zone togroundwater occurs, but not to the extent as at the Washingtonsite. The median concentration of ortho P in the unsaturated zoneis different from all environmental compartments at the Californiasite, with median concentrations decreasing from the unsaturatedzone to the groundwater, to the groundwater/surface water inter-face, and to the stream.

In contrast to the California and Washington sites, the medianconcentration of ortho P of the unsaturated zone and groundwateris similar at the Nebraska site. Groundwater ortho P is similar tothe stream, the unsaturated zone, and the groundwater/surfacewater interface. Both the groundwater and the groundwater/sur-face water interface are higher than the stream.

Concentrations of ortho P are uniformly low at all of the envi-ronmental compartments of the Indiana location. Tile drains lar-gely control the discharge of the Leary Weber Ditch. In additionto tile drains, overland flow contributes to the Leary Weber Ditchdischarge (Domagalski et al., 2008a). Concentrations of ortho Pfor the Leary Weber Ditch are similar to that for groundwaterand tile drains. Some higher concentrations of ortho P were mea-sured at the Leary Weber Ditch as indicated by the 75th and90th percentile of the data range (Fig. 2). Median concentrationof ortho P in overland flow was 0.29 mg/L and that for total Pwas 0.58 mg/L. Overland flow probably is the source of the higherortho P concentrations in the stream. The unsaturated zone wateris largely removed by tile drains at the Indiana site (Domagalskiet al., 2008a) and therefore only minimally contributes to the fluxof water to the aquifer where the drains are present.

Similar to Indiana, concentrations of ortho P at the Marylandsite are uniformly low. Groundwater concentrations are statisti-cally less than all of the other environmental compartments. Therange of ortho P concentrations of the unsaturated zone is higherthan the other hydrologic compartments. An impervious clay layerprevents direct movement of groundwater through the streambed;however, groundwater discharges from seepage zones at the lat-eral margins of the floodplain and then flows across the floodplainin small channels and as diffuse sheet flow to the creek. Some ofthose seeps may contribute ortho P to the stream if the concentra-tions are from shallow horizons with concentrations similar to theunsaturated zone.

Suggested criteria for total phosphorus in surface water for spe-cific ecoregions, as specified by the US EPA (US Environmental Pro-tection Agency, 2010) are shown in Fig. 2. Concentrations of orthoP in excess of the criterion are exceeded for the majority of mea-surements made at the Washington and Nebraska sites. The crite-rion is also exceeded for the unsaturated zone, groundwater andthe groundwater/surface water interface at the California location,but the stream water is mostly within the established criteria.Most of the measurements taken at the Indiana site indicate thatthat location is within the recommended criterion, at least with re-gard to ortho P. The unsaturated zone concentrations at the Mary-land location are generally higher than the total P criteria, as aresome of the groundwater/surface water interface concentrations.About 25% of the measurements taken at the Maryland streamare above the recommended criterion, but the groundwater onlycontributes minimally to those concentrations. Groundwater maycontribute to the exceedance of the recommended criteria at theWashington, California, and Nebraska sites, but not the Indianaand Maryland.

Domagalski et al. (2008a) calculated basin yields of ortho P andtotal phosphorus of the streams from these locations and providedestimates of the yearly load that was contributed by baseflow. In

addition, Tesoriero et al. (2009) calculated base flow index (ratioof base flow to total flow volume) at three of these locations andthe contribution of ortho P loading from base flow. At the Washing-ton site, baseflow contributes about 41% of the discharge withintroduced canal water contributing about 40% and overland flowthe remainder (Domagalski et al., 2008a). About 25% of the annualload of ortho P was contributed to the stream by baseflow duringthe irrigation season. Tesoriero et al. (2009) calculated a baseflowindex of 0.57 for the Washington site and that 32% of the annualortho P load could be attributed to baseflow. The Washington sitehad the highest yield of ortho P of any of the study sites at 1 kg/ha(Domagalski et al., 2008a). At the Nebraska site (Maple Creek),about 20% of the annual load of ortho P was contributed bygroundwater baseflow (Domagalski et al., 2008a). Tesoriero et al.(2009) calculated a baseflow index of 0.35 for the Nebraska siteand that 30% of the annual ortho P load was from baseflow. The cal-culation method was different between the two studies and there-fore there are differences in the calculated amounts of phosphoruscontributed from base flow.

Baseflow does contribute to the load of ortho P in some cases tothe studied streams, and it was the intent of this study to deter-mine the connection of the different hydrologic compartments tothe transport of ortho P and the processes that result in ortho Ptransport. In other words, does application of phosphorus at theland surface result in some component of downward movementbelow the root zone and once the agriculturally-applied phospho-rus is transported to the aquifer, can it move along a flow path to astream interface. Alternatively, the phosphorus may be taken up byplant tissue, be sorbed onto mineral surfaces, or precipitated as anew phase either in the unsaturated zone or the groundwater. Con-centrations measured in groundwater might then be the result oflocal chemical equilibrium, as opposed to advective movement.

Geochemical conditions, such as redox, pH, or major elementchemistry, might change along the flow path and result in a differ-ent potential for adsorption, precipitation or, alternatively, dissolu-tion. The connection between the hydrological compartments toone another was initially investigated by comparing the major ele-ment chemistry by using piper plots (Fig. 3). Major element chem-istry varies widely at these locations and is controlled by acombination of water and sediment interactions, chemical equilib-rium, types of agricultural chemicals applied, and direction ofwater movement. Mineralogical buffers and soil amendmentsinfluence the amounts of individual ions. At the Indiana location,dolomite and calcite exert a strong influence on the ion distribu-tion resulting in waters dominated by Ca, Mg, and HCO3 in all envi-ronmental compartments. The cations and anions of the waters atthe Nebraska site are also tightly clustered suggesting a similarsource of water that is close to equilibrium with the aquifer sedi-ments. At the other extreme is the Maryland location. The upper3–5 m of the unsaturated zone is fluvial sand and gravel that areunderlain by moderately weathered marine quartz and glauconiticsand and silt. The underlying aquifer material is composed mainlyof marine quartz and glauconitic sand. The minerals at the Mary-land site are very insoluble, and the water chemistry of the unsat-urated zone and groundwater is influenced by agricultural soilamendments. The unsaturated zone water has a wide range of an-ions with sulfate dominant in several samples. In contrast to theunsaturated zone, most of the groundwater samples have a similarchemistry with calcium and chloride as the dominant ions, and lit-tle sulfate. Applications of fertilizers containing sulfate explain theelevated concentrations of sulfate. Ion composition at the ground-water/surface water interface is variable and suggests differentsources of water. Puckett et al. (2008) found no evidence of signif-icant exchange between groundwater and stream due to the claylayer. However, abundant seeps and upward head gradients alongthe margin of the floodplain allow groundwater discharge and con-

0

0.5

1

1.5

2

2.5

3

orth

o-P,

in m

g/L

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Stre

am

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Stre

am

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Stre

am

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Stre

am

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Stre

am

Tile

Dra

in

Washington: 0.022 mg/LCalifornia: 0.047 mg/LNebraska: 0.076 mg/LIndiana 0.076 mg/LMaryland: 0.031 mg/L

90th percentile

75th percentileMedian

25th percentile

10th percentile

Washington

CaliforniaNebraska

Indiana Maryland

A

B BC

A

AB

AC ABC A C

B

B

BC B

A

C

A

AD

Sample Size and Statiscal Designation Letter

A

A

33

55 5765

8

35

52

22

118

49 54 74

26 22 48 39 6061

31

23

70

Fig. 2. Boxplots of dissolved orthophosphate concentrations in hydrological compartments per study area and US Environmental Protection Agency total phosphorusecoregion specific criteria for protection of surface water resources. Statistical groupings (A–C) indicate which groups of samples are statistically similar.


tribution to Morgan Creek. The major ion chemistry of the seeps isnot uniform, however, and it is difficult to assign a specific sourceof water to the seeps.

At the Washington location, stream water is influenced by theunsaturated zone, groundwater, and exchange across the ground-water/surface water interface. The unsaturated zone is slightly en-riched in sulfate from agricultural amendments, while thegroundwater is more enriched in bicarbonate from the dissolutionof carbonate minerals present in small amounts in the aquifer.Water moves from the unsaturated zone to groundwater, and thenalong short flow paths to the stream (Domagalski et al., 2008a).Therefore, the major element chemistry is tightly clustered andthe stream chemistry plots between that of the other hydrologicalcompartments. The water chemistry at the California site is vari-able and subject to wide variations depending on what types ofagricultural chemicals were used, and the source of the water.Water movement can be in either direction across the streambedat the California site and therefore the chemistry might be similarto the adjacent groundwater system, or may be the result of riverwater from upstream sources (Domagalski et al., 2008b).

The pH of the waters in the unsaturated zone, groundwater andgroundwater/surface water interfaces varies across these locations(Fig. 4). The pH of the Maryland location is the lowest for ground-water and the groundwater/surface water interface, because ofprecipitation pH values below 4.5 (National Atmospheric Deposi-tion Program, 2010) and the low buffering capacity of the sedi-ments that make up the aquifer. The pH is close to the neutralrange for the other areas with the highest pH generally found inthe Washington location. The water pH is not necessarily a goodpredictor of phosphorus mobility as low concentrations occur atboth the Maryland site, with acidic groundwater and at the Indianasite with groundwater of alkaline pH. Ortho P concentrations, how-ever, might be related to pH at the Maryland site since adsorptiononto iron oxides is dependent on the pH of the solution. Sorption ofanions increases with decreasing pH (Stumm, 1992). This occurs

because the oxide surfaces become more positively charged andtherefore have a greater capacity to adsorb anions, such as ortho-P.

4.2. Contents of phosphorus and iron in aquifer and unsaturated zonesolids

Only a certain percentage of phosphorus in soil or on aquifersolids can be mobilized into solution or exchanged between solu-tion and a solid. The portion of phosphorus that is reactive in thismanner has sometimes been termed the environmentally availablephosphorus (Devau et al., 2009). Estimates were made of the envi-ronmentally available phosphorus in aquifer or unsaturated zonesolids by measuring the amount extracted by two methods: Olsen(Olsen et al., 1954) and Mehlich-3 (Mehlich, 1984) (Fig. 5). Bothmethods have long been used to determine levels of soil saturationwith respect to phosphorus (Maguire and Sims, 2002; Sims et al.,2000; Thompson and McFarland, 2010; Heckrath et al., 1995).The Mehlich-3 method has been demonstrated to accurately pre-dict the amount of water-soluble phosphorus and the amount thatcould be desorbed from iron oxides in soils of the mid-Atlantic re-gion of the US (Sims et al., 2002). The Olsen test generally is recom-mended only for alkaline soils (Sims, 2009) whereas the Mehlich-3provides reasonable estimates of exchangeable phosphorus inmost soils (Sims, 2009). For the Washington, California, and Ne-braska samples, the Mehlich-3 method results for the aquifer sed-iments are higher than the Olsen, but for the Maryland samples,they are similar. Given the fact that the Mehlich-3 utilizes a mixof extracting agents, it is not surprising that more phosphorus ismobilized relative to the Olsen method in some locations. For theIndiana location, contents of phosphorus in aquifer solids were lessthan the Mehlich-3 detection limit suggesting that little of thephosphorus was labile. The contents measured in the Washingtonand California aquifers were similar as were those measured in Ne-braska, Indiana, and Maryland. In all cases, total phosphorus insediment (Fig. 5) exceeds the amount considered labile. There is

Unsaturated Zone

Groundwater

Groundwater/ Surface WaterSurface Water

Seep 1 (Plot E only)

Seep 2 (Plot E only)

% meq/l

a

b

c d

e

Fig. 3. Piper plots. (a) Washington, (b) California, (c) Nebraska, (d) Indiana and (e) Maryland.


a wide range of total phosphorus in the aquifer material with thehighest at the Washington site and the lowest at the Maryland site.

This is likely attributable to different minerals with varyingamounts of phosphorus. The content of total phosphorus in sedi-


ment has no relation to the concentrations of ortho P in groundwa-ter. The Olsen or Mehlich-3 methods do not dissolve the mineral-ogical forms of that phosphorus.

Contents of labile phosphorus were higher in the unsaturatedzone relative to the aquifers in all cases. The median contents ofphosphorus measured by the two methods are given in Fig. 5. Inall cases, the contents of phosphorus were higher near the surfaceand decreased with depth in the unsaturated zone.

It is generally recognized that iron oxides and hydroxides arethe most important reservoirs of exchangeable phosphorus in mostsoils (Hemwell, 1957; Zhang and Huang, 2007; Hingston et al.,1974; Sigg and Stumm, 1980; Hawke et al., 1989; Parfitt, 1978),although aluminum oxyhydroxides, certain clay minerals, and car-bonate minerals may serve as the primary reservoir of exchange-able phosphorus in some environments (Hemwell, 1957;Shariatmadari and Mermut, 1999; Parfitt, 1978). Iron can be pres-ent in sediments in a variety of forms including crystalline phasessuch as hematite or magnetite, or poorly crystalline phases includ-ing goethite or limonite. It was assumed that the forms of ironmost likely to result in sorption of ortho P would be the poorlycrystalline phases, because they would likely have the highest sur-face area, and that the concentration of those could be estimatedby selective extraction. Amounts of iron recovered by hydroxyl-amine hydrochloride in the aquifers are shown in Fig. 6. The high-est median content (1500 lg/g) was at the Washington location,followed by the California location (765 lg/g). The Nebraska loca-tion had the lowest amount (60 lg/g). Reducing conditions also oc-cur in the Nebraska location, as indicated by the presence ofdissolved ferrous iron, and the lower relative amounts of iron oxi-des might be partially attributable to ferric iron reduction. At theother locations, some reducing conditions occur, but the lack ofelevated concentrations of dissolved ferrous iron suggests thatthe oxides are stable.

4.3. Geochemical modeling

It was assumed that measured concentrations of ortho P in theunsaturated zone and within the aquifer would be limited by equi-librium with solids, either by precipitation or co-precipitation asmineral phases, or by sorption onto iron oxides and other minerals,and that those relationships could be examined by geochemicalmodeling. Phosphorus concentrations might be limited, for exam-ple, by precipitation of hydroxyapatite (Ca5(PO4)3(OH)). However,only a few instances of supersaturation with hydroxyapatite, as indi-cated by the model, were found, such as the unsaturated zone at theWashington, California, and Nebraska sites, because of high ortho Pand calcium concentrations. A few instances of supersaturation withrespect to hydroxyapatite in the groundwater of the California loca-tion were also found. It has been reported previously that supersat-uration with respect to hydroxyapatite has been observed ingroundwater plumes down gradient of septic systems but that pre-cipitation is not an effective control on the transport of phosphoruspossibly because of slow kinetics (Robertson et al., 1998). In general,precipitation of hydroxyapatite cannot be invoked as a primarymechanism controlling the transport of phosphorus through theflow paths of these agricultural systems since most measured con-centrations were below the saturation index.

Since precipitation of hydroxyapatite may not be controllingphosphorus concentrations, sorption onto iron oxides may be animportant factor. The observed concentrations of ortho P at the timeof sampling would then be the result of equilibrium with the sorbingphase. It would be useful to know the degree of saturation of ironoxides of the unsaturated zone or aquifer that result in the observedortho P concentrations. The level of saturation with respect to sorp-tion will dictate whether or not the aquifers might be a source ofphosphorus to the stream. As the oxides become saturated, phos-

phorus transport would become more likely. In order to understandthat relationship, a geochemical model was utilized whereby con-centrations of ortho P were changed to levels above and below theactually observed to determine the relative levels of saturation with-in the hydrological compartments along the flow path.

A plot of the modeled amount of sorbed phosphorus relative toortho P is shown in Fig. 7 for the unsaturated zone, aquifer, andgroundwater/stream interface for the Washington location.

In Fig. 7, it is apparent that dissolved concentrations of phos-phorus are low until the iron oxides are saturated, if sorption ontooxides is the primary factor limiting phosphorus transport in thesubsurface. Saturation, or near saturation of the oxides, is definedin this discussion as that point in the plots where there is a rapidincrease in the concentration of ortho P with only minimal changesin the amount of phosphorus adsorbed. The high concentrations ofortho P measured in the Washington unsaturated zone (Fig. 2) arethe result of saturation or near saturation of the oxides with phos-phorus. Use of manure as a fertilizer probably accounts for the highconcentrations in the unsaturated zone. The groundwater concen-trations of phosphorus are much lower and the median concentra-tion (0.0015 mmol/L) is close to the point where the oxides arestarting to become saturated. The highest concentration measuredin groundwater was 0.006 mmol/L. Phosphorus in the unsaturatedzone is therefore a source to the groundwater and it is likely thatthe levels sorbed onto aquifer solids will increase under the cur-rent management practices. Most of the ortho P concentrationsat the groundwater/surface water interface are slightly greaterthan the recommended US EPA criteria for protection of surfacewater. Therefore, groundwater discharge might contribute to theload of phosphorus in the stream as well as concentrations abovethe recommended criteria.

Competing ions for sorption sites on the iron oxide surfaces forthe Washington location include hydrogen ion, and magnesium.The competing ions limit the potential sites for phosphorus sorp-tion. At the condition of the median phosphorus concentration inthe unsaturated zone, hydrogen ions occupy about 41% of theavailable sorption sites, while phosphorus occupies about 40%.Magnesium occupies about 16% of the available sorption surface.Other cations or anions are not important with respect to sorption.Although sulfate is present in relatively high concentration at55 mg/L, very little sulfate sorbs to iron oxide surfaces under theseconditions. At the highest dissolved concentration of ortho P in theunsaturated zone (0.146 mmol/L), representing a high level ofoxide saturation with respect to phosphorus, about 46% of theavailable oxide surface is occupied by phosphorus while hydrogenions occupy about 35%. In contrast to the unsaturated zone, phos-phorus occupies about 18% of the oxide surfaces in the saturatedzone. This is less than half of that for the unsaturated zone. Aswas the case for the unsaturated zone, the main competing ionsin the saturated zone are hydrogen ions and magnesium.

The average annual rate of water movement through the unsat-urated zone of the Washington location was the slowest of allstudy units as estimated by a piston-flow model at 0.1 cm/day(Fisher and Healy, 2008). This slow rate of water movement resultsin an average unsaturated zone residence time of 8.1 years (Fisherand Healy, 2008). However, during the summer irrigation season,downward movement is much faster. Use of the bromide tracersuggested a downward movement of 2.5–2.7 cm/d (Fisher andHealy, 2008).

The predicted amount of phosphorus sorbed onto aquifer sedi-ment at the Washington location, based on the median ortho Pconcentration in the groundwater, was about 0.275 mmol/kg(8.5 mg/kg). This lies between the amount analyzed by the Meh-lich-3 and Olsen methods. Therefore, the modeled amounts ofsorbed phosphorus are probably close to the potentially environ-mentally available phosphorus.

3

4

5

6

7

8

9

pH

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Washington California

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Nebraska

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Indiana

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

Maryland

Uns

atur

ated

zon

e

Gro

undw

ater

Gro

undw

ater

/Sur

face

Wat

er

90th percentile


25th percentile

10th percentile

Fig. 4. pH by environmental compartment and study location.

0

10

20

30

40

50

60

70

80

90

Phos

phor

us, i

n m

g/kg

Mehlich-3 method detection limit

Meh

lich-

3

Ols

en

Olsen method detection limitWashington

California Nebraska Indiana Maryland

90th percentile


25th percentile

10th percentile

Total P: 750Total P: 206 Total P: 67

Total P: 217 Total P: 36Unsaturated Zone: 20.5 ; 8 Unsaturated

Zone: 26, 9Unsaturated Zone: 12 ; 3.5

Unsaturated Zone: 10 ; 8

Unsaturated Zone: 18 ; 10

Meh

lich-

3

Ols

en

Meh

lich-

3

Ols

en

Meh

lich-

3

Ols

en

Meh

lich-

3

Ols

en

Fig. 5. Boxplots of phosphorus measured in aquifer solids by the Mehlich-3 and Olsen methods; median concentrations of total phosphorus in aquifer solids for each studyarea, in mg/kg; and, median concentrations of phosphorus measured by the Mehlich-3 and Olsen methods in unsaturated zone solids, in mg/kg.


A plot of modeled sorption of phosphorus onto iron oxides forthe unsaturated zone, aquifer, and groundwater/stream waterinterface of the California location is shown in Fig. 8.

The unsaturated zone solids of the California location have thehighest level of saturation with respect to phosphorus and themedian groundwater phosphorus concentration indicates thatthose solids are approaching saturation. The groundwater/surface

water interface has a lower level of oxide saturation. The amountof phosphorus predicted by the model as sorbed to iron oxides inthe aquifer at the California location was 0.9 mmol/kg (28 mg/kg), based on the median ortho P concentration in groundwater.This is within the range of concentrations measured by the Meh-lich-3 method. The model predicted that phosphorus coveredabout 50% of the available sorption sites within the aquifer at the

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

Iron

Ext

ract

ed b

y H

ydro

xyla

min

e

H

ydro

chlo

ride

, mg

/ g

Was

hing

ton

Cal

ifor

nia

Neb

rask

a

Indi

ana

Mar

ylan

d

90th percentile


25th percentile

10th percentile

Fig. 6. Iron extracted from sediments by hydroxylamine hydrochloride.


top of the flow path where ortho P concentrations were the high-est. The major competing ion was hydrogen, which accounted formost of the remaining sorption sites.

A schematic of the flow system at the California location isshown in Fig. 9. At the California location, ortho P has migrateddown into the aquifer, but does not appear to have been trans-ported to the river interface.

There is an order of magnitude difference in ortho P concentra-tions between the wells of the middle portion of the flow path andthose situated in the riparian zone. Since there is no fertilization inthe riparian zone, ortho P concentrations are lower, probably closeto natural levels. It is likely that agricultural use of fertilizers, eitherchemicals or manure have affected the aquifer at the top and mid-dle portion of the flow path, and the downward gradients ofgroundwater flow from irrigation have allowed phosphorus to mi-grate below the root zone. Transport time of water in the unsatu-rated zone of the California location was one of the higher of thesestudy areas. The rate of unsaturated zone water movement wasestimated by a piston-flow model as 1.0 cm/day (Fisher and Healy,2008). This translates to an unsaturated zone residence time of1.9 years (Fisher and Healy, 2008). The bromide tracer moved fromland surface to a depth of 6.1 m in approximately 8 months, a rateof 2.5 cm/d. This translates to an unsaturated zone residence timeof 9 months. As was observed at the Washington site, the estimatefrom the bromide tracer yields a shorter residence time comparedto the annual average piston-flow estimate, and is due to applica-tion of irrigation water during the summer months. Because of thenear saturation of the oxides with respect to phosphorus sorptionand the rapid movement of water driven by the irrigation induceddownward gradient, phosphorus has migrated into the aquifer atelevated concentrations. However, the phosphorus plume doesnot reach the river, as there is an approximate 200-m thick riparianzone, which acts as a buffer. Previously it was shown (Domagalskiet al., 2008b) that agriculturally derived nitrate moves through thisriparian zone with little attenuation, and that the chemistry of theshallow groundwater directly adjacent to the Merced Rivermatches that of the upgradient wells below the agricultural fields.Phosphorus transport is limited across the riparian zone because ofsorption onto the aquifer solids and the lack of an external sourceof phosphorus. Therefore, groundwater transport of phosphorus, atthis specific location, does not currently contribute significantly to

the concentrations of phosphorus in the river, and the ortho P inthe river is largely within the recommendations of the US EPA.

At the Nebraska location, ortho P concentrations measured at allhydrologic compartments along the flow path were generally inexcess of the US EPA recommended criteria for phosphorus instreams most of the time (Fig. 2). A plot of the modeled concentra-tions of phosphorus sorbed onto iron oxides is shown in Fig. 10.The unsaturated zone at the Nebraska location is the thickest ofall of the study areas, with a maximum thickness of about 22 m.Shallow groundwater is about 10–20 years old and deeper is 40–50 years. Therefore, shallow groundwater is being recharged withrecent water and associated chemicals. The rate of water move-ment, as estimated by the bromide tracer test, was about 1.8 cm/day. This translates to an unsaturated zone residence time of3 years.

The median concentration of ortho P in the unsaturated zone,groundwater and groundwater/surface water interface are at satu-ration levels with respect to iron oxides. This is in spite of a rela-tively thick unsaturated zone, between 15 and 22 m of thickness.At the Nebraska location, the level of iron oxides in the sedimentswas the lowest, and therefore the amount sorbed is over an orderof magnitude less than in the other study areas. Reducing condi-tions exist in the aquifer and one reason for the lower oxide con-centrations is reductive dissolution. Because of this lowercapacity of the sediments to sorb phosphorus, the flow path doescontribute to the stream load and to elevated concentrations ofphosphorus in the stream.

The predicted amount of phosphorus sorbed onto iron oxidesfor the Nebraska location, based on the median ortho P concentra-tion in groundwater, was about 0.022 mmol/kg (0.7 mg/kg). Thisconcentration is within the range of the amount of phosphorus ex-tracted by the Olsen and Mehlich-3 methods. The model predictsthat phosphorus sorption covers 37% of the available oxide sur-faces while hydrogen ions cover 59%. Most of the remaining sur-face is occupied by magnesium. Therefore, the low amounts ofiron oxides in the flow path allow for elevated concentrations ofortho P along the flow path and groundwater has become a signif-icant source of phosphorus concentrations to the stream.

The modeled sorption plot for the Maryland location (Fig. 11) isdifferent from the others showing that the sediments of the aquifercan sorb substantially more phosphorus than that of the unsatu-rated zone.

The difference between the sorption in the unsaturated zoneand the aquifer for the Maryland location is the effect of pH, asthe amount of iron oxides does not change. The sorption of phos-phorus onto iron oxides increases with decreasing pH. There is asubstantial decrease in pH from the unsaturated zone to the aqui-fer (Fig. 3). Rain in equilibrium with quartz and other insolubleminerals in the aquifer sediments recharge the aquifer water. Be-cause of the lack of significant reactions with aquifer minerals,the pH reflects equilibrium with atmospheric CO2 and the pH islow relative to the other sites. The unsaturated zone solids areactually close to saturation with respect to phosphorus sorptiononto iron oxides, but the aquifer solids are very unsaturated.Although groundwater can contribute to the load of phosphorusin the stream, the concentrations are always less than the recom-mended criteria of the US EPA (Fig. 2). In fact, the stream concen-trations are less than the criteria for most of the measurementstaken.

The predicted amount of phosphorus sorbed by iron oxides inthe aquifer at the Maryland site, based on the median concentra-tion of ortho P in groundwater was 0.5 mmol/kg (15.5 mg/kg). Thisis similar to the range of concentrations measured by either theMehlich-3 or Olsen methods. Phosphorus occupies about 21% ofthe available sorption sites within the aquifer while hydrogen ionsoccupy 68%. Because of the low pH of the Maryland groundwater,

0.0

0.0

0.2

0.4

0.6

0.8

Unsaturated Zone Groundwater Groundwater / Surface Water

0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

ortho P, mmol/L

Phos

phor

us s

orbe

d, m

mol

/ kg

Unsaturated zone maximum

Unsaturated zone median

Groundwater and Groundwater / Surface Water median

Fig. 7. Plot of dissolved orthophosphate and modeled phosphorus concentrations sorbed onto iron oxides, Washington location.

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4



Unsaturated zone median

Phos

phor

us s

orbe

d, m

mol

/ kg

ortho P, mmol/L

Groundwater median

Groundwater / Surface water median

Fig. 8. Plot of dissolved orthophosphate and modeled phosphorus concentrations sorbed onto iron oxides, California location.


the oxides are undersaturated with respect to phosphorus andgroundwater is not currently a source of high concentrations ofphosphorus to the stream. In contrast to the other sites, there issome sorption of sulfate. Sulfate occupies about 10% of the sorptionsites. The annual average rate of water movement within theunsaturated zone is 0.4 cm/d with a residence time of 7.1 years(Fisher and Healy, 2008). Much of the recharge, however, occursduring the non-growing season. Active recharge occurs duringabout 70% of the year and yields an average rate of 0.57 cm/dand a residence time of 2–4 years. Given the high affinity of theaquifer solids for phosphorus, it is unlikely that groundwater trans-port will contribute significant concentrations of phosphorus to thestream under the given agricultural management.

Movement of water and associated chemicals is different at theIndiana location relative to the others because of the presence oftile drains, which were installed because of poor drainage. Thepresence of tile drains greatly reduces the amount of water re-

charged to the aquifer since water is rapidly transported out ofthe soil to the Ditch. Therefore, the movement of water and chem-icals is primarily through the unsaturated zone to the drains anddirectly to the stream, except for an additional component of over-land flow. Although the soils have poor drainage, preferential flowpaths do exist (Stone and Wilson, 2006; Fisher and Healy, 2008).Evidence for preferential flow paths was obtained from the bro-mide tracer experiment as bromide was detected within a weekafter application at a depth of 1.5 m. Presumably, these preferentialflow paths also allowed for the transport of agricultural chemicals,including phosphorus. However, bromide concentrations mea-sured in lysimeters were steady for some time after applicationin lysimeters at depths of 0.9 and 1.2 m suggesting that water thatwas initially transported by preferential flow paths becametrapped in the soil matrix (Fisher and Healy, 2008). Some soilsare also more slowly drained because of their distance to the tiles.Concentrations of ortho P are always low in both the unsaturated

27

15

3

-9

-21

-34

irrigated

ripariansprinkler

flood

Direction of vertical flow (horizontal = no vertical gradient)Typical horizontal and vertical hydraulic gradients are small,on the order of 0.005

Lysimeter; Concentration in mmol/L

Groundwater well; Concentration in mmol/L

.03

.02.02

0.1

.01

.004

Approximate groundwater table

.02

.03

.02

Merced River

.002

.002

.002

.003

.006Confining layer, dashed where location is uncertain

Age: 9 years

Approximate age contour

Age: 25 years

Ele

vatio

n, m

eter

s

1000 600 0200400800

Distance from Merced River (meters)

Age: 20 years

0.0016

0.0007

Fig. 9. Flow system of the California location showing well depth and location, dissolved orthophosphate concentrations in mmol/L, age of groundwater, and direction ofgroundwater flow.

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035


ortho P, mmol/L

Phos

phor

us s

orbe

d, m

mol

/ kg


Unsaturated zone, groundwater, and groundwater / surfacewater median

Groundwater Maximum

Fig. 10. Plot of dissolved orthophosphate and modeled phosphorus concentrations sorbed onto iron oxides, Nebraska location.


zone and the tile drains. Concentrations of ortho P are slightlyhigher in the stream water, but the unsaturated zone and the tiledrains do not appear to be a major source of phosphorus to thestream. Most measurements of ortho P at the Indiana locationare lower than the USEPA recommended criteria for protection ofsurface water. Higher concentrations in the stream, relative tothe tile drains, are likely the result of overland runoff. Levels of la-bile phosphorus are low at the Indiana location. Most of the anal-yses for the Mehlich-3 method yielded results below the method

detection limit. Modeling of the sorption of phosphorus onto ironoxides suggests undersaturation at the concentrations of ortho Pof the unsaturated zone. The model predicts, based on a medianconcentration of ortho P in the unsaturated zone of 0.01 mg/L,about 0.08 mmol/kg (2.5 mg/kg) of phosphorus will be adsorbedonto iron oxides. This amount is higher than the levels analyzedby either the Mehlich-3 or the Olsen methods.

Because of the low amount of labile phosphorus in the unsatu-rated zone, the oxide surfaces are very undersaturated with respect


to phosphorus relative to the other study locations. Possible rea-sons for the low amount of oxide sorption at Indiana might be low-er use of phosphorus for agriculture or a very low amount ofphosphorus transport below the root zone. Sequestration in theroot zone might be very effective at the Indiana location. Concen-trations of near-surface soil phosphorus as measured by the Meh-lich-3 method at a depth of 7.6–23 cm were between 54 and100 mg/kg. Immediately below this horizon at a soil depth of 28–48 cm, the concentration of Mehlich-3 extractable P was down to11 mg/kg, which then decreases down to 5 mg/kg immediately be-low. The Indiana location has the highest levels of silts and claysrelative to the other four study areas, and movement of phospho-rus past the clays of the root zone may be limited. In addition, thehighest level of sedimentary calcium (total calcium) occurs at Indi-ana. The median level of total calcium at the Indiana site was96,870 mg/kg, and the next highest was at the Washington site(14,870 mg/kg). Calcium in sediment was much lower at the otherthree sites because of the lack of carbonate minerals in the drain-age basin. These higher levels of calcium at the Indiana location arebecause calcite and dolomite are abundant in the unsaturated zoneand aquifer sediments due to of the presence of carbonate out-crops. It has been suggested that carbonate minerals are an effec-tive sink of phosphate in soils (Shariatmadari and Mermut, 1999;Freeman and Rowell, 1981; Avnimelech, 1980). The unsaturatedzone water is slightly supersaturated with respect to calcite andthe mineral probably undergoes both dissolution and precipitationthereby possibly providing a source of new surface for phosphatesorption or co-precipitation. However, other studies have sug-gested that carbonate minerals in soil are not an effective sink ofphosphorus, but that sorption onto oxides and non-carbonate clayis more effective (Zhou and Li, 2001). The actual mechanism ofphosphate sequestration in the root zone was beyond the scopeof this study.

Apparent distribution coefficients, Kd, were estimated for theaquifer and unsaturated zone of these locations using the medianconcentration as measured by the Mehlich-3 method and medianconcentrations of ortho P in either the unsaturated zone or theaquifer. These distribution coefficients were calculated as the log-arithm of the concentration of phosphorus in soil or sediment mea-sured by the Mehlich-3 method divided by the median waterconcentration (units = L/kg). The apparent logKd for the Washing-

0.00 0.02 0.04 0.06 0.

0.0

0.5

1.0

1.5

2.0

2.5

Phos

phor

us s

orbe

d, m

mol

/ kg

ortho P,

Unsaturated zone media

Groundwater median

Fig. 11. Plot of dissolved orthophosphate and modeled phosphoru

ton aquifer was 2.3. The value for the California aquifer was 1.4.The value for the Nebraska aquifer was 1.3, and that for the Mary-land aquifer was 6. Values of Kd for the unsaturated zone weresometimes similar to those for the aquifers, and in some cases dif-ferent. The logKd for the Washington unsaturated zone was 1.1.That can be understood with respect to the high concentrationsof ortho P in the unsaturated zone. The logKd for the unsaturatedzone of California and Nebraska were each 1.5, similar to the aqui-fers. The logKd for the Indiana unsaturated zone was 3.0, whichwas the highest of any of the study sites. This can be attributedto the larger amount of clay and calcium carbonate in the soil.The logKd for the Maryland unsaturated zone was 2.5. These appar-ent Kd values can be understood in the context of the concentra-tions of the sorbent materials, water chemistry, and theagricultural practices. The Maryland aquifer has the highest Kd,which is much higher than any other location, because of the effectof the low pH, which allows for increased sorption on iron oxides.In contrast, the Maryland unsaturated zone Kd is lower because ofthe higher pH. The apparent Kd for the California location was cal-culated for the part of the aquifer in the upper portion of the flowpath where continuous fertilization has resulted in a net down-ward movement of phosphorus into the aquifer. Because of nearsaturation on the oxides, at a neutral pH, the apparent Kd is rela-tively low, and the soil is losing its capacity to adsorb additionalphosphorus. The California unsaturated zone soil is also close tosaturation with phosphorus and therefore, the Kd is close to thatof the aquifer. Similarly, the Nebraska location has a low value ofthe apparent Kd because of low concentrations of iron oxide ineither the aquifer or unsaturated zone that are also at or near sat-uration with respect to phosphorus. The Washington aquifer has ahigher apparent Kd, relative to California and Nebraska, because ofhigher concentrations of iron oxides in the aquifer material, but theunsaturated zone Kd is lower probably because of the high use ofmanure, which leads to high concentrations of ortho P in the unsat-urated zone water. The Kd values obtained in this study were sim-ilar to those measured in other studies. For example, in a study ofNigerian soils (Egwu et al., 2010) measured P partition coefficientsranged from 1.29 to 4.06 (log values). The lowest value in thatstudy was in a soil with the highest sand content (84%).

The modeling of phosphorus sorption onto iron oxides is com-pared among the five locations on Fig. 12. Concentrations of ortho

08 0.10 0.12 0.14 0.16 0.18

GroundwaterUnsaturated Zone

mmol/L

n

s concentrations sorbed onto iron oxides, Maryland location.

0.00001 0.0001 0.001 0.01 0.1 1

0

1

2

3

4

5

Mod

eled

ort

ho P

, mg/

L

Modeled ortho P / Sorbed Phosphorus

Washington California Nebraska Indiana Maryland

Fig. 12. Ratio of modeled orthophosphate to sorbed phosphorus vs dissolvedorthophosphate at five study locations. Best lines are drawn to show ratio ofdissolved orthophosphate to sorbed phosphorus at saturation conditions.


P, at any location, will be low as a result of sorption until the avail-able sites are saturated. The point of saturation is indicated by abreak of slope (Fig. 12). The modeling results for the Marylandlocation plot closest to the y-axis as the oxides have the greatestaffinity for phosphorus because of the effect of pH even thoughthe concentration of oxides is less relative to California and Wash-ington. Therefore, the relative amount of phosphorus on the oxidesurfaces is much greater than that in equilibrium with the dis-solved compared to the other locations. The plots for Washington,California, and Indiana are similar because of either similar levelsof oxides or solution chemistry that results in a similar amountof sorption. A contrasting situation is for the Nebraska locationwhere the break in slope occurs furthest to the left in the plot ata ratio of dissolved to sorbed P of about 0.1. Because of the lowamount of available oxide surfaces due to low iron concentrations,phosphorus is not effectively sorbed and the sub-surface transportof phosphorus can occur to a much greater extent.

5. Summary and conclusion

Phosphorus can be transported below the root zone to theunsaturated zone and through an aquifer in agricultural settings.This study showed that under conditions where phosphorus iseither not entirely taken up by plant tissue or where soil chemistrydoes not favor either precipitation or sorption, that sub-surfacetransport can result in elevated concentrations in groundwater orloadings to receiving streams. In this study, phosphorus transportto a stream was demonstrated most clearly at the Nebraska loca-tion, although groundwater transport did contribute to the streamloads at other sites. A low concentration of iron oxides in the sand,silt, and loess soils at the Nebraska location resulted in low sorp-tive capacity. The median concentration of ortho P at the ground-water/stream interface was 0.2 mg/L at the Nebraska location. Insome cases, agriculturally applied phosphorus will likely not betransported through the aquifer under current management prac-tices. That was demonstrated most clearly at the Maryland loca-tion. Although oxides were close to saturation with respect tophosphorus in the unsaturated zone, the aquifer solids had highsorptive capacity because of the low groundwater pH (<5) withinthe aquifer. Under that condition, ortho P concentrations werelow to undetectable in the aquifer and transport to the streamwas not a concern. A similar situation occurred at Indiana althoughfor different reasons. The apparent situation for phosphorus inIndiana was very low transport out of the root zone to the sub-sur-

face. It was surmised that the clay and carbonate chemistry of theroot zone effectively sequestered phosphorus and as a result oxidesin the subsurface sorbed what little environmentally availablephosphorus that was present.

In contrast to Indiana, transport of phosphorus out of the rootzone to the unsaturated zone occurred to a greater extent at theWashington and California locations. High unsaturated zone con-centrations in Washington were the result of the widespread useof manure. Phosphorus was also applied throughout the growingseason at the California site and strong downward gradients ofwater, driven by irrigation, resulted in transport of phosphorus be-low the sandy root zone and into the unsaturated zone and aquifer.Phosphorus was effectively sorbed within the aquifer at the Wash-ington location because of higher iron oxide contents, relative tothe other study areas. It was demonstrated at the California loca-tion that the current agricultural management practices result inmovement of phosphorus through the aquifer in the zone ofgroundwater that had been dated between 10 and 25 years. Phos-phorus was not transported to the stream interface at the Califor-nia location because of the presence of a riparian zone with naturalvegetation with no external inputs of phosphorus.

This study also showed that concentrations of ortho P in thesub-surface that are close to or above the eco-region specific crite-ria for protection of surface water are probably the result of aquifersolids that are at or above saturation with respect to sorbed phos-phorus. In those instances, agricultural management practicesshould consider methods to prevent plumes of sub-surface phos-phorus from reaching discharge areas at stream interfaces.

References

Allen, B.L., Mallarino, A.P., 2008. Effect of liquid swine manure rate, incorporation,and timing of rainfall on phosphorus loss with surface runoff. J. Environ. Qual.37, 125–137.

American Public Health Association, 1998. Standard Methods for the Examinationof Water and Wastewater, 20th ed. American Public Health Association,American Water Works Association, and Water Environment Federation,Washington, DC.

Anderegg, J.C., Naylor, D.V., 1988. Phosphorus and pH relationships in an acidic soilwith surface and incorporated organic amendments. Plant Soil 107, 273–278.

Avnimelech, Y., 1980. Calcium–carbonate–phosphate surface complex in calcareoussystems. Nature 288, 255–257.

Baker, J.L., Campbell, K.L., Johnson, H.P., Hanway, J.J., 1975. Nitrate, phosphorus, andsulfate in subsurface drainage water. J. Environ. Qual. 4, 406–412.

Baker, N.T., Stone, W.W., Wilson, J.T., Meyer, M.T., 2006. Occurrence and Transportof Agricultural Chemicals in the Leary Weber ditch, Hancock County, Indiana,2003–2004. US Geological Survey Scientific Investigations Report 06-5251.USGS, Washington, DC.

Boorman, G.A., 1999. Drinking water disinfection byproducts: review and approachto toxicity evaluation. Environ. Health Perspect. 107 (Suppl. 1), 207–217.

Burwell, R.E., Schuman, G.E., Heinemann, H.G., Spomer, R.G., 1977. Nitrogen andphosphorus movement from agricultural watersheds. J. Soil Water Conserv. 32,226–230.

Busenberg, E., Plummer, L.N., 2000. Dating young groundwater with sulfurhexafluoride: natural and anthropogenic sources of sulfur hexafluoride. WaterResour. Res. 36, 3011–3030.

Capel, P.D., McCarthy, K.A., Barbash, J.E., 2008. National, holistic, watershed-scaleapproach to understand the sources, transport, and fate of agriculturalchemicals. J. Environ. Qual. 37, 983–993.

Cole, C.V., Olsen, S.R., Scott, C.O., 1953. The nature of phosphate sorption by calciumcarbonate. Soil Sci. Soc. Am. J. 17, 352–356.

Devau, N., Le Cadre, E., Hinsinger, P., Jaillard, B., Gérard, F., 2009. Soil pH controls theenvironmental availability of phosphorus: experimental and mechanisticmodeling approaches. Appl. Geochem. 24, 2163–2174.

Diaz, R.J., Nestlerode, J., Diaz, M.L., 2004. A global perspective on the effects ofeutrophication and hypoxia on aquatic biota. In: Rupp, G.L., White, M.D. (Eds.),Proc. 7th International Symposium Fish Physiology, Toxicology and WaterQuality, Tallinn, Estonia. 600/R-04/049. US Environmental Protection Agency,Ecosystems Research Division, Athens, Georgia.

Domagalski, J.L., Ator, S., Coupe, R., McCarthy, K., Lampe, D., Sandstrom, M., Baker,N., 2008a. Comparative study of transport processes of nitrogen, phosphorus,and herbicides to streams in five agricultural basins, USA. J. Environ. Qual. 37,1158–1169.

Domagalski, J.L., Phillips, S.P., Bayless, E.R., Zamora, C., Kendall, C., Wildman Jr., R.A.,Hering, J.G., 2008b. Influences of the unsaturated, saturated, and riparian zones


on the transport of nitrate near the Merced River, California, USA. Hydrol. J. 16,675–690.

Dzombak, D.A., Morel, F.M.M., 1990. Surface Complexation Modeling—HydrousFerric Oxide. John Wiley, New York, 393p.

Egwu, G.N., Uyovbisere, E.O., Okunola, O.J., 2010. Phosphorus sorption isothermsand parameter estimates of soils of the savanna, and effect of soil properties onthese parameters. Electron. J. Environ., Agric. Food Chem. 9, 866–874.

Essaid, H.I., Zamora, C.M., McCarthy, K.A., Vogel, J.R., Wilson, J.T., 2008. Using heat tocharacterize streambed water flux variability in four stream reaches. J. Environ.Qual. 37, 1010–1023.

Fisher, L.H., Healy, R.W., 2008. Water movement within the unsaturated zone infour agricultural areas of the United States. J. Environ. Qual. 37, 1051–1063.

Fishman, M.J., 1993. Methods of Analysis by the US Geological Survey NationalWater Quality Laboratory—Determination of Inorganic and OrganicConstituents in Water and Fluvial Sediments. US Geological Survey Open-FileRep. 93-125. USGS, Denver, CO.

Fishman, M.J., Friedman, L.C., 1989. Methods for Determination of InorganicSubstances in Water and Fluvial Sediments. Techniques of Water-ResourcesInvestigations. Book 5. US Geological Survey, Reston, VA, 545pp (Chapter A1).

Fredrick, B.S., Linard, J.I., Carpenter, J.L., 2006. Environmental Setting of Maple CreekWatershed, Nebraska. US Geological Survey Scientific Investigations Report2006-5037. USGS, Washington, DC.

Freeman, J.S., Rowell, D.L., 1981. The adsorption and precipitation of phosphate ontocalcite. Soil Sci. 32, 75–84.

Green, C.T., Puckett, L.P., Böhlke, J.K., Bekins, B.A., Phillips, S.P., Kauffman, L.J.,Denver, J.M., Johnson, H.M., 2008. Limited occurrence of denitrification in fourshallow aquifers in agricultural areas of the United States. J. Environ. Qual. 37,994–1009.

Gronberg, J.M., Kratzer, C.R., 2006. Environmental Setting of the Lower Merced RiverBasin, California. US Geological Survey Scientific Investigations Report, 2006-5152. USGS, Washington, DC.

Gunjigake, N., Wada, K., 1981. Effects of phosphorus concentration and pH onphosphate retention by active aluminum and iron of Ando soils. Soil Sci. 132,347–352.

Hancock, T.C., Brayton, M.J., 2006. Environmental Setting of the Morgan Creek Basin,Maryland, 2002–2004. US Geological Survey Open-File Report 2006-1151.USGS, Washington, DC.

Hancock, T.C., Sandstrom, M.W., Vogel, J.R., Webb, R.M.T., Bayless, E.R., Barbash, J.E.,2008. Pesticide fate and transport throughout unsaturated zones in fiveagricultural settings, USA. J. Environ. Qual. 37, 1086–1100.

Hawke, D., Carpenter, P.D., Hunter, K.A., 1989. Competitive adsorption of phosphateon goethite in marine electrolytes. Environ. Sci. Technol. 23, 187–191.

Heckrath, G., Brookes, P.C., Poulton, P.R., Goulding, K.W.T., 1995. Phosphorusleaching from soils containing different phosphorus concentrations in theBroadbalk experiment. J. Environ. Qual. 24, 904–910.

Hemwell, J.B., 1957. The fixation of phosphorus by soils. In: Norman, A.G. (Ed.),Advances in Agronomy, vol. 9. Academic Press, New York, pp. 95–112.

Hergert, G.W., Klausner, S.D., Bouldin, D.R., Zwerman, P.J., 1981. Effects of dairymanure on phosphorus concentrations and losses in tile effluent. J. Environ.Qual. 10, 345–349.

Hingston, F.J., Posner, A.M., Quirk, J.P., 1974. Anion adsorption by goethite andgibbsite 2. Desorption of anions from hydrous oxide surfaces. Soil Science 25,16–26.

Hochberg, Y., 1988. A sharper Bonferroni procedure for multiple tests ofsignificance. Biometrika 75, 800–802.

Klausner, S.D., Zwerman, P.J., Ellis, D.F., 1976. Nitrogen and phosphorus losses fromwinter disposal of dairy manure. J. Environ. Qual. 5, 47–49.

Lathrop, T.R., 2006. Environmental Setting of the Sugar Creek and Leary WeberDitch Basins, Indiana, 2002–2004. US Geological Survey Scientific InvestigationsReport 2006-5170l. USGS, Washington, DC.

Lovley, D.R., Phillips, E.J.P., 1987. Rapid assay for microbially reducible ferric iron inaquatic sediments. Appl. Environ. Microbiol. 53, 1536–1540.

Maguire, R.O., Sims, J.T., 2002. Measuring agronomic and environmental soilphosphorus saturation and predicting phosphorus leaching with Mehlich-3. SoilSci. Soc. Am. J. 66, 2033–2039.

Mehlich, A., 1984. Mehlich 3 soil test extractant: a modification of Mehlich 2extractant. Commun. Soil Sci. Plant Anal. 15, 1409–1416.

National Atmospheric Deposition Program, 2010. <http://nadp.sws.uiuc.edu/sites/ntnmap.asp?> (accessed 03.18.10).

National Research Council (NRC), 2005. Clean Coastal Waters: Understanding andReducing the Effects of Nutrient Pollution. National Academy Press,Washington, DC.

Oliver, B.G., Shindler, D.B., 1980. Trihalomethanes from the chlorination of aquaticalgae. Environ. Sci. Technol. 14, 1502–1505.

Olsen, S.R., Cole, C.V., Watanabe, F.S., Dean, L.A., 1954. Estimation of AvailablePhosphorus in Soils by Extraction with Sodium Bicarbonate. US Department ofAgriculture Circular, p. 939.

Owens, J.P., Denny, C.S., 1979. Upper Cenozoic deposits of the central DelmarvaPeninsula, Maryland and Delaware. US Geological Survey Professional Paper1067-A. USGS, Reston, VA.

Owens, J.P., Minard, J.P., 1979. Upper Cenozoic Sediments of the Lower DelawareValley and the Northern Delmarva Peninsula, New Jersey, Pennsylvania,Delaware, and Maryland. US Geological Survey Professional Paper 1067-D.USGS, Reston, VA.

Parfitt, R.L., 1978. Anion adsorption by soils and soil materials. In: Brady, N.C. (Ed.),Advances in Agronomy, vol. 30. Academic Press, New York, pp. 1–50.

Parkhurst, D.L., Appelo, C.A.J., 1999. User’s Guide To PHREEQC (Version 2)—AComputer Program For Speciation, Batch-Reaction, One-Dimensional Transport,And Inverse Geochemical Calculations. US Geological Survey Water-ResourcesInvestigations Report 99-4259. USGS, Washington, DC.

Payne, K.L., Johnson, H.M., Black, R.W., 2007. Environmental Setting of the GrangerDrain and DR2 Basins, Washington, 2003-04, U.S. Geological Survey ScientificInvestigations Report 2007-5102, Washington, DC.

Puckett, L.J., Zamora, C., Essaid, H., Wilson, J.T., Johnson, H.M., Brayton, M.J., Vogel,J.R., 2008. Transport and fate of nitrate at the ground-water/surface-waterinterface. J. Environ. Qual. 37, 1034–1050.

Robertson, W.D., Schiff, S.L., Ptacek, C.J., 1998. Review of phosphate mobility andpersistence in 10 septic system plumes. Ground Water 36, 1000–1010.

Salvia-Castellvi, M., Iffly, J.F., Borght, P.V., Hoffman, L., 2005. Dissolved andparticulate nutrient export from rural catchments: a case study fromLuxembourg. Sci. Total Environ. 344, 51–65.

Scully, F.E., Howell, G.D., Kravitz, R., Jewell, J.T., Hahn, V., Speed, M., 1988. Proteins innatural waters and their relation to the formation of chlorinated organicsduring water disinfection. Environ. Sci. Technol. 22, 537–542.

Shariatmadari, H., Mermut, A.R., 1999. Magnesium- and silicon-induced phosphatedesorption in smectite-, palygorskite-, and sepiolite-calcite systems. Soil Sci.Soc. Am. J. 63, 1167–1173.

Sharpley, A.N., Syers, J.K., 1979. Phosphorus inputs into a stream draining anagricultural watershed II. Amounts and relative significance of runoff types.Water Air Soil Pollut. 9, 417–428.

Sharpley, A.N., Daniel, T., Sims, T., Lemunyon, J., Stevens, R., Parry, R., 2003.Agricultural Phosphorus and Eutrophication, second ed. US Department ofAgriculture, Agricultural Research Service, ARS-149.

Sigg, L., Stumm, W., 1980. The interaction of anions and weak acids with hydrousgoethite (a-FeOOH) surface. Colloids Surf. 2, 101–117.

Simes, J.R., 1986. An improved Bonferroni procedure for multiple tests ofsignificance. Biometrika 73, 750–754.

Sims, J.T., 2009. Soil test phosphorus: principles and methods. In: Kovar, J.L.,Pierzynski, G.M. (Ed.), Methods for Phosphorus Analysis for Soils, Sediments,Residuals, and Waters, second ed. Southern Cooperative Series Bulletin 408,Virginia Tech University.

Sims, J., Edwards, A.C., Schoumans, O.F., Simard, R.R., 2000. Integrating soilphosphorus testing into environmentally based agricultural managementpractices. J. Environ. Qual. 29, 60–71.

Sims, J.T., Maguire, R.O., Leytem, A.B., Gartley, K.L., Pautler, M.C., 2002. Evaluation ofMehlich 3 as an agri-environmental soil phosphorus test for the mid-AtlanticUnited States of America. Soil Sci. Soc. Am. J. 66, 2016–2032.

Sprague, L.A., Lorenz, D.L., 2009. Regional nutrient trends in streams and rivers ofthe United States, 1993–2003. Environ. Sci. Technol. 43, 3430–3435.

Steele, G.V., Johnson, H.M., Sandstrom, M.W., Capel, P.D., Barbash, J.E., 2008.Occurrence and fate of pesticides in four contrasting agricultural settings in theUnited States. J. Environ. Qual. 37, 1116–1132.

Stone, W.W., Wilson, J.T., 2006. Preferential flow estimates to an agricultural tile drainwith implications for glyphosate transport. J. Environ. Qual. 35, 1825–1835.

Stumm, W., 1992. Chemistry of the Solid-water Interface. Wiley-Interscience, NewYork.

Tesoriero, A.J., Duff, J.H., Wolock, D.M., Spahr, N.E., 2009. Identifying pathways andprocesses affecting nitrate and orthophosphate inputs to streams in agriculturalwatersheds. J. Environ. Qual. 38, 1892–1900.

Thompson, C.A., McFarland, A.M.S., 2010. Effects of surface and groundwaterinteractions on phosphorus transport within streambank sediments. J. Environ.Qual. 39, 548–557.

US Environmental Protection Agency, 1993. Methods for the Determination ofInorganic Substances in Environmental Samples, EPA/600/R-93/100. USEPA,Washington, DC.

US Environmental Protection Agency, 1995a. Test Methods for Evaluating SolidWaste, Physical/Chemical Methods, SW-846, third ed. USEPA, Washington, DC.

US Environmental Protection Agency, 1995b. Federal Register, vol. 60, pp. 37974–37980.

US Environmental Protection Agency, 2010. Ecoregional Nutrient CriteriaDocuments for Rivers & Streams. <http://water.epa.gov/scitech/swguidance/waterquality/standards/criteria/aqlife/pollutants/nutrient/rivers_index.cfm>(accessed 01.10.10).

US Geological Survey, 2010a. About Regional Assessments on Water-Quality Statusand Trends. <http://water.usgs.gov/nawqa/studies/regional_assessments.html>(accessed 01.10.10).

US Geological Survey, 2010b. <http://water.usgs.gov/lab/dissolved-gas> (accessed01.10.10).

Vollenweider, R.A., 1968. Scientific Fundamentals of the Eutrophication of Lakes andFlowing Waters, with Particular Reference to Nitrogen and Phosphorus asFactors of Eutrophication. Report Number GP OE/515. Organization forEconomic and Development, Paris.

Zhang, J.Z., Huang, X.L., 2007. Relative importance of solid-phase phosphorus andiron on the sorption behavior of sediments. Environ. Sci. Technol. 41, 2789–2795.

Zhou, M., Li, Y., 2001. Phosphorus-sorption characteristics of calcareous soils andlimestone from the southern everglades and adjacent farmlands. Soil Sci. Soc.Am. J. 65, 1404–1412.

http://nadp.sws.uiuc.edu/sites/ntnmap.asp?

http://nadp.sws.uiuc.edu/sites/ntnmap.asp?

http://water.epa.gov/scitech/swguidance/waterquality/standards/criteria/aqlife/pollutants/nutrient/rivers_index.cfm

http://water.epa.gov/scitech/swguidance/waterquality/standards/criteria/aqlife/pollutants/nutrient/rivers_index.cfm

http://water.usgs.gov/nawqa/studies/regional_assessments.html

http://water.usgs.gov/lab/dissolved-gas

Journal of Hydrology - California Water Science Center … ·...

Documents

Transcript of Journal of Hydrology - California Water Science Center … ·...