Influence of Drought and Municipal Sewage Effluentson the Baseflow Water Chemistry of an UpperPiedmont River

Jin Hur & Mark A. Schlautman & Tanju Karanfil & John Smink &

Hocheol Song & Stephen J. Klaine & John C. Hayes

Received: 2 May 2006 /Accepted: 15 September 2006 / Published online: 14 December 2006# Springer Science + Business Media B.V. 2006

Abstract The Reedy River in South Carolina isaffected by the urban area of Greenville, the third mostpopulous city in the state, and by the effluents from twolarge-scale municipal wastewater treatment plants(WWTPs) located on the river. Riverine water chemistrywas characterized using grab samples collected annuallyunder spring season baseflow conditions. During the 4-year time period associated with this study, climatic

variations included two severe drought spring seasons(2001 and 2002), one above-normal precipitation springseason (2003), and one below-normal precipitationspring season (2004). The influence of drought andhuman activities on the baseflow chemistry of the riverwas evaluated by comparing concentrations of dis-solved anions, total metals, and other important waterchemistry parameters for these different years. Concen-trations of copper and zinc, common non-point sourcecontaminants related to urban activities, were notsubstantially elevated in the river within the urban areaunder baseflow conditions when compared with head-water and tributary samples. In contrast, nitrate concen-trations increased from 1.2–1.6 mg/l up to 2.6–2.9 mg/l through the urban stream reach. Concentrations of othermajor anions (e.g., sulfate, nitrate) also increased alongthe reach, suggesting that the river receives continuousinputs of these species from within the urban area. Thehighest concentrations of major cations and anionstypically were observed immediately downstream fromthe two WWTP effluent discharge locations. Attenuationof nitrate downstream from the WWTPs did not alwaystrack chloride changes, suggesting that nitrate concen-trations were being controlled by biochemical processesin addition to physical processes. The relative trends indecreasing nitrate concentrations with downstream dis-tance appeared to depend on drought versus non-droughtconditions, with biological processes presumably servingas a more important control during non-drought springseasons.

Environ Monit Assess (2007) 132:171–187DOI 10.1007/s10661-006-9513-1

J. HurDepartment of Earth & Environmental Sciences,Sejong University, Seoul, 143-747, South Korea

M. A. Schlautman (*) : T. Karanfil :H. SongDepartment of Environmental Engineering & Science,School of the Environment, Clemson University,Clemson, SC 29634-0919, USAe-mail: mschlau@clemson.edu

M. A. SchlautmanDepartment of Geological Sciences,School of the Environment, Clemson University,Clemson, SC 29634-0919, USA

M. A. Schlautman : J. Smink : S. J. KlaineInstitute of Environmental Toxicology,Clemson University, Pendleton, SC 29670, USA

J. Smink : J. C. HayesDepartment of Agricultural and Biological Engineering,Clemson University, Clemson, SC 29634, USA

S. J. KlaineDepartment of Biological Sciences, Clemson University,Clemson, SC 29634, USA

Keywords Drought . Urban river . Urban watershed .

Wastewater treatment .Water chemistry

1 Introduction

Population growth and urbanization are closelyrelated to changes in water chemistry and quality inmany watersheds (Chen, He, Zhang, & Cui, 2004;Rhodes, 2001; Whitall, Castro, & Driscoll, 2004).Degradation of water quality brought about byanthropogenic inputs directly affects the health ofhumans and aquatic ecosystems. For example, it iswell known that elevated nutrient concentrationscontributes to eutrophication resulting in proliferationof noxious algae, and reduced species abundance anddiversity in lakes and oceans (Whitall et al., 2004).Elevated pollution concentrations in watersheds typ-ically can be attributed to a combination of point andnon-point sources (Chen et al., 2004; Rhodes, 2001).

Rural (i.e., nonurbanized) watersheds are affected bydiverse contaminant sources from an array of agricul-tural activities and land use practices. Nutrient and otherpollutant concentrations in the rivers and streams ofsuch watersheds tend to be elevated under increasedflow conditions (i.e., storm events) because chemicalloadings are typically dominated by non-point sources(Chen et al., 2004). There have been conflictingconclusions drawn from previous investigations ofurban area effects on stream water chemistry (e.g.,nutrients and other solute concentrations) in whichsampling was conducted over long time periodsencompassing various flow conditions. For example,Sonoda, Yeakley, and Wallker (2001) sampled streamsmonthly from many urban and non-urban locations inan urbanizing watershed and found that highernitrogen concentrations were highly correlated withnon-urban land use. Herpe and Troch (2000) alsoobserved no significant relationship between streamwater nitrate concentrations and the degree of urban-ization based on temporal and spatial data from sixcatchments with mixed land uses. In contrast, otherstudies have shown positive relationships betweenpercent urban area and sulfate, nitrate and chlorideconcentrations (Rhodes, 2001; Wayland et al., 2003). Itshould be noted that under baseflow conditions,groundwater sources can be responsible for the elevatedanion concentrations. For example, septic systems,leaky sewer pipes, and lawn fertilizers are known to be

major sources of nitrate contamination in groundwaterbeneath urban residential areas (Clawges & Vowinkel,1996; Ellis & Revitt, 2002). In fact, one studyconcluded that nitrate inputs to streams from groundwa-ter are higher in both urban and agricultural areas than inundeveloped areas (Clawges & Vowinkel, 1996).

Industrial wastewater and municipal sewage efflu-ents can have major impacts on chemical loads insurface waters downstream of developed urban areas(Andersen, Lewis, & Sargent, 2004; Marti, Aumatell,Gode, Poch, & Sabater, 2004; Rhodes, 2001) and theireffluents are likely to have greater effects on rivers andstreams during drought conditions (Andersen et al.,2004; Jordan, Weller, & Correll, 2003). Previousinvestigators have reported high nutrient concentra-tions in the effluents of wastewater treatment plants(WWTPs), demonstrating that WWTPs can be majorpoint sources of pollution in urban-impacted water-sheds (Andersen et al., 2004; Kim et al., 2002; Martiet al., 2004). Marti et al. (2004) surveyed 15 streamsreceiving effluents from WWTPs, and reported thatphosphate loads from the plants were, on average, over10 times greater than the upstream loads and contrib-uted approximately 90% of the total phosphatemeasured downstream of the WWTPs. Rivers andstreams dominated by WWTP effluents are typicallylocated in certain western states in the United States(U.S.) (Berestov, Fernando, & Fox, 1998; Brookset al., 2004; McMahon, Tindall, Collins, Lull, &Nuttle, 1995; Reilly, Horne, & Miller, 2000; Saunders& Lewis, 2003) and in other arid or semi-arid climateregions around the world (Deksissa, Ashton, &Vanrolleghem, 2003; Marti et al., 2004). Because suchrivers and streams are less common in the eastern U.S.,however, fewer studies of WWTP effluent effects ondownstream hydrology and water chemistry have beenconducted in eastern U.S. states (Andersen et al., 2004;Diamond, Hall, Pattie, & Gruber, 1994). Recently,Andersen et al. (2004) described the effects of WWTPeffluent discharges on the water chemistry and nutrientconcentrations in the Bush River, South Carolina (SC).They demonstrated that WWTP effluent effects ondownstream water chemistry could be substantial in awatershed located in temperate climate regions undersevere drought conditions. However, the baseflowsampling of their study was limited to a very short periodof time (i.e., a fewmonths), and thus was not sufficient todiscuss potential temporal variations in the WWTPeffluent impacts (e.g., effects of long-term drought).

172 Environ Monit Assess (2007) 132:171–187

Previous long-term monitoring studies have shownthat the water chemistry and nutrient concentrations inwatersheds are influenced by extreme precipitationconditions (Jordan et al., 2003; Lucey & Goolsby,1993) as well as by human activities and land uses (Ha& Bae, 2001; Robson & Neal, 1997). For example,Jordan et al. (2003) reported that atmospheric sourcesof nitrogen to a river estuary become more importantduring low-precipitation years. Over a 17-year sam-pling period, Lucey andGoolsby (1993) reported nitrateconcentrations above the maximum contaminant levelin two water supply reservoirs shortly after twodrought years. Such drought effects are confoundedby watershed characteristics and contaminant sources.For example, Parr and Mason (2003) reported increas-ing nitrogen contamination in watersheds affected bysewage effluents under drought conditions, whereasCaruso (2002) observed that nutrient concentrations inforested and rural watersheds decreased in extremedrought years due to the lack of rainfall and runoff eventsand subsequent low inputs from non-point sources.

The Reedy River is located in the northwestern,upper Piedmont region of SC (Figure 1) on the eastcoast of the U.S. The flow rate and water chemistry ofthe river is currently being affected by the effluentsfrom two large-scale municipal WWTPs, and its waterchemistry is also likely being influenced by the urbanarea of Greenville, the third most populous city in thestate. In this study, the chemistry of the Reedy Riverwas evaluated annually during the spring seasonunder non-storm event conditions from 2001 through2004. During this time period, there were two severedrought spring seasons (2001 and 2002), one above-normal precipitation spring season (2003), and onebelow-normal precipitation spring season (2004). Thespecific objectives of the study were to (1) quantify theeffects ofWWTP effluents and urban areas on importantwater chemistry parameters under baseflow conditionsover the period 2001–2004, and (2) examine potentialeffects of drought conditions on nitrate attenuation inthe river downstream of the WWTPs.

2 Methods

2.1 Study site

The Reedy River watershed is one of the majorwatersheds comprising the Saluda River Basin in SC.

The river originates near the town of Travelers Rest,SC, in the foothills of the Appalachian Mountainswhere two groundwater-fed streams meet. The slopeof the terrain in the Reedy River watershed rangesfrom 2 to 45%, with an average slope being 15%(Beasley, Criss, Quattro, & Scharf, 2001). Thewatershed drains more than 676 km2 (167,000 acres)of land and contains 523 km (325 mi) of streams.

The Reedy River is generally shallow along its entirechannel length, with typical depths under baseflowconditions ranging from approximately one meter orless in the northern portion of the watershed toapproximately two meters or less in the southern portionof the watershed. In addition to its shallow depth,numerous small falls and shoals provide good mixingand reaeration such that overall concentrations ofdissolved oxygen (DO) in the river typically remainabove 6 mg/l. Because many reaches of the Reedy Riverhave been deliberately channelized in the past and/orchannelized due to stream incision, the river has becomedisconnected from its floodplain and thus water levelsrise and fall quickly from storm water runoff duringprecipitation events (Beasley et al., 2001). This phe-nomenon results in a large variation in mean dailyflows. The river has two major impoundments (LakeConestee and Boyd Mill Pond) above its dischargepoint into Lake Greenwood (Figure 1). Three long-term U.S. Geological Survey (USGS) gauging stationsare located on the main channel of the Reedy River(Figure 1).

The overall land use/cover in the Reedy Riverwatershed is 66% forest, 16% agriculture (croplandand pasture), and 18% urban (Beasley et al., 2001).However, over 75% of the southern portion of thewatershed is forested compared to less than 60% inthe northern portion where the city of Greenville andsurrounding communities are located. Although theofficial population of the city of Greenville is onlyabout 60,000, the total population of GreenvilleCounty is more than 400,000. Thus, urban land userepresents over 28% of the northern portion of theReedy River watershed, but only 7.5% of the southernportion. There is little difference in the percentage ofagricultural area between the northern and southernportions of the watershed (Beasley et al., 2001).

The upper piedmont region of SC has a warmtemperate climate, with an average daily maximumtemperature of 22 °C and an average of more than50 days/year in which the temperature exceeds 32 °C

Environ Monit Assess (2007) 132:171–187 173

(Andersen et al., 2004). The annual rainfall averages127 cm (50.5 in) in Greenville, SC, which is similarto the state’s average annual rainfall of nearly 120 cm(47 in.). Drought conditions occurred throughoutmuch of SC from 1998 through 2002. According tothe National Drought Mitigation Center (http://drought.unl.edu), the Reedy River watershed wasunder severe drought conditions on the samplingdates of this study in 2001 (28 March) and 2002(9 May). By the sampling date of 2003 (12 May),abundant rainfall had returned to SC and thus no

drought condition was specified by the Center for thattime period. In 2004, however, rainfall had again fallenshort of normal amounts and the Center characterizedthe condition as being abnormally dry during that year’ssampling date (10 May).

Hydrologic drought is commonly defined by thenumber of consecutive days or months that a stream-flow is deficient (e.g., within the lowest 50% of along-term average record) over a specific period oftime (Maidment, 1993). Records from the USGSgauging stations on the Reedy River near Greenville,

Figure 1 Map of the ReedyRiver (SC) watershed,showing the locations of thethree USGS gauging sta-tions, two major impound-ments, two wastewatertreatment plants, and grabsample locations for theheadwaters, tributaries, andon the main channel of theriver.

174 Environ Monit Assess (2007) 132:171–187

SC, reflect the changes in drought versus non-droughtconditions during the 4-year time period associatedwith this study (Figure 2). In 2001 and 2002, most ofthe daily mean discharges were lower than the 60-year average value. The flow rate increased in 2003due to abundant precipitation throughout the year, butthen returned again to relatively low flow rates in2004. The drought conditions that existed during thesampling dates of this study also can be illustrated bycalculating average antecedent mean daily streamflows over an appropriate time interval prior to eachsample date as a percentage of the 60-year averagemean daily flows. For example, 30-day antecedentaverage flow rates were 67%, 53%, 133%, and 46%of normal on 28 March 2001, 9 May 2002, 12 May2003, and 10 May 2004, respectively. For these samedates, the 90-day antecedent average stream flowrates were 54%, 48%, 145%, and 45% of normal.

Two major municipal WWTPs are located alongthe Reedy River. The Mauldin Road WWTP (MR-WWTP) discharges directly to the river near thesouthern boundary of Greenville and just upstream ofLake Conestee. The Lower Reedy WWTP (LR-WWTP) also discharges to the main river channel,approximately 12 km downstream of the MR-WWTP(Figure 1). Typically, the discharge flow of the LR-WWTP is about 20% to 30% that of the MR-WWTP(personal communication from Western CarolinaRegion Sewer Authority).

2.2 Sample collection and analyses

Sampling sites were selected along the main channel ofthe Reedy River from the headwater, upper suburbanarea, down through the upstream arm of Lake Green-wood (Figure 1). Annual sampling events were sched-uled in early to late spring (March for 2001, May for2002–2004) on days with at least two antecedent daysof no precipitation. Discharge of the Reedy River oneach sampling date was similar to the level prior to theantecedent rain event, indicating baseflow samples werecollected for this study (Figure 2).

Sample collections were completed within 12 h,starting at the bottom of the watershed and movingupstream. At each site, grab samples were taken atbridge crossings from midstream with a plastic pail,poured into appropriate bottles, placed in ice chests,and transported back to the laboratory. All bottleswere acid-washed prior to use, and the sampling pail

was thoroughly rinsed with river water beforecollecting the final sample for analysis. In 2002, insitu measurements of temperature, pH, turbidity,conductivity, and DO were also recorded at all of thesampling sites with a Hydrolab DataSonde 4 to verifythe subsequent laboratory processing and results.Unfiltered grab samples were used for the determinationof total suspended solids (TSS), turbidity, total metalconcentrations, pH, and alkalinity. Laboratory analysesof the unfiltered samples were undertaken within 48 hafter sample collection. Digestion of unfiltered sampleswas conducted by transferring a volume (9 ml) toconical centrifuge tubes, adding 1 ml of concentratednitric acid (HNO3, 70%, EM Omintrace), and heatingwith a water bath on a hot plate at a minimumtemperature of 100 °C for 15 min. Aliquots of thedigestates were then decanted into clean 15 ml plasticbottles (Wheaton HDPE bottle), and sent to theClemson Agricultural Services Laboratory (Clemson,SC, USA) for selected element concentrations (P, Na,Ca, Mg, K, Cu, Fe, Zn, Al and B) by inductivelycoupled plasma atomic emission spectrometry. Ali-quots of the grab samples were filtered (0.2 μm) priorto determining major anion concentrations (nitrate,nitrite, chloride, sulfate, phosphate and bromide) byion chromatography. Additional aliquots of the grabsamples were filtered (0.45 μm) prior to measuringultraviolet absorbance (UVA) at 254 nm and dissolvedorganic carbon (DOC) concentrations following pro-cedures previously described by Karanfil, Erdogan,and Schlautman (2003). Specific UVA (SUVA254) wascalculated for each filtered sample by dividing theUVA value at 254 nm by DOC concentration. Allfiltered samples were stored at 4 °C until analyzed. Allchemical analyses were completed within a week aftersample collection. Appropriate control and blanksamples were analyzed in parallel with all experimentalsamples. Covariance between the concentrations ofdifferent species was examined by creating scatterplotsand conducting correlation analyses using standardtools available in Microsoft Excel.

3 Results and Discussion

3.1 Water chemistry in Reedy River watershedheadwaters and tributaries

Sampling results from tributaries and headwatersprovide information about the background water

Environ Monit Assess (2007) 132:171–187 175

chemistry of the Reedy River under spring baseflowconditions. Mean concentrations and associated con-centration ranges for major solutes in the headwatersand tributaries indicate that the river has relatively

low background concentrations of dissolved ions(Table I). Except for sulfate, average concentrationsof major cations and anions are two to four timeslower than the values reported for the headwaters and

0

5

10

15

20

0 365 730 1095 1460 1825

Dis

char

ge

(m3/s

)

2000 2001 2002 2003 2004Years

0

5

10

15

20

1 3 5 7 9 11 13 15

March 28, 2001

0

5

10

15

20

1 3 5 7 9 11 13 15

May 9, 2002

0

5

10

15

20

1 3 5 7 9 11 13 15

May 12, 2003

0

5

10

15

20

1 3 5 7 9 11 13 15

May 10, 2004

3/13 3/18 3/23 3/28 4/24 4/29 5/4 5/9 4/27 5/2 5/7 5/12 4/25 4/30 5/5 5/10

Dis

char

ge (

m3 /

s)

0

5

10

15

20

1 3 5 7 9 11 13 15

March 28, 2001

0

5

10

15

20

1 3 5 7 9 11 13 15

May 9, 2002

0

5

10

15

20

1 3 5 7 9 11 13 15

May 12, 2003

0

5

10

15

20

1 3 5 7 9 11 13 15

May 10, 2004

0

5

10

15

20

1 3 5 7 9 11 13 15

March 28, 2001

0

5

10

15

20

1 3 5 7 9 11 13 15

May 9, 2002

0

5

10

15

20

1 3 5 7 9 11 13 15

May 12, 2003

0

5

10

15

20

1 3 5 7 9 11 13 15

May 10, 2004

3/13 3/18 3/23 3/28 4/24 4/29 5/4 5/9 4/27 5/2 5/7 5/12 4/25 4/30 5/5 5/10

Dis

char

ge (

m3 /

s)

Figure 2 Records of daily mean discharge for the USGSgauging station at Greenville from 2000 to 2004 (top) and forthe 15 days prior to each sampling date (bottom; sampling datesindicated by vertical arrows). The dotted line in the top figurerepresents the 60-year average discharge value. Note that the

discharge of the Greenville USGS gauging station is notaffected by the effluent from the two wastewater treatmentfacilities nor from the two impoundments, all of which arelocated downstream of the station.

176 Environ Monit Assess (2007) 132:171–187

tributaries of the Bush River (Andersen et al., 2004).The Bush River watershed is another major watershedin the Saluda River basin that is near our studylocation, although it has substantially more agricul-tural land use/cover (∼56% forest, 36% agriculture,8% urban; Andersen et al., 2004) than the ReedyRiver watershed. The low concentrations of back-ground solutes suggest that if their concentrations arehigher downstream of Greenville, those increasesshould be related to the urban area and/or WWTPeffluents.

Total phosphorous concentrations in the ReedyRiver headwaters and tributaries were less than0.05 mg/l. Ammonia was not detectable. Nitrate wasthe dominant (99%) form of dissolved inorganicnitrogen, averaging about 1.2 mg/l. This concentra-tion is higher than the range reported for undevelopedwatersheds in the U.S. (Clark, Mueller, & Mast,2000), and likely reflects the small but significant

land use contributions of agriculture, acreages withhorses, and household septic systems in the headwaterregion of the watershed. Chloride concentrationsranged from 2.0 to 3.9 mg/l, with an average of about3 mg/l. No temporal trends in water chemistry werereadily apparent over the 4-year period, except forsodium which had relatively lower concentrations inthe severe drought spring season of 2002 versus theother years. DOC concentrations in the headwatersand tributaries ranged from 1.1 to 6.4 mg/l, with anaverage of about 2.7 mg/l, and SUVA254 valuesranged from 1.9 to 4.2 l/mg C-m, with an averageof about 2.8 l/mg C-m.

3.2 Influence of urban areas on water chemistryparameters in the Reedy River

Riverine water chemistry in urban areas can often beaffected by a combination of point and nonpoint

Table I Solute concentration ranges, various water chemistry parameters, and their mean values for all headwater and tributarysamples (2001 to 2004)a

Solute, unit 2001 2002 2003 2004

Range Mean Range Mean Range Mean Range Mean

Na, mg/L NMb – 0.6–1.6 1.0 2.2–3.9 3.5 2.9–5.2 3.5K, mg/L NM – 0.9–1.2 1.1 0.8–4.8 1.6 1.2–1.9 1.5Ca, mg/L NM – 3.1–5.3 3.9 2.5–4.7 3.4 1.7–4.4 3.3Mg, mg/L NM – 0.8–1.4 1.3 0.8–1.3 1.1 0.8–1.5 1.2*Cl, mg/L 2.9–3.5 3.1 2.9–3.8 3.3 2.5–3.3 2.8 2.0–3.9 2.6*NO3, mg/L 1.3–1.9 1.6 0.7–1.6 1.3 0.4–1.8 1.3 0.6–1.6 1.2NO3/Cl, unitless

c 0.44–0.59 0.53 0.26–0.46 0.39 0.58–0.72 0.65 0.41–0.54 0.50*NO2, mg/L NDd – ND – ND – 0.01–0.02 0.01*SO4, mg/L 2.5–6.3 4.8 2.2–5.5 3.6 1.7–5.3 3.4 2.3–5.9 4.1*PO4, mg/L ND – ND – ND – 0.03–0.07 0.04*Br, mg/L 0.0–0.05 0.03 ND – 0.0–0.05 0.04 0.03–0.04 0.03*DOC, mg/L 1.6–3.3 2.3 2.4–3.5 3.1 1.8–6.4 3.2 1.1–2.6 1.9*SUVA254, L/mgC-m NM – 2.1–4.2 3.2 1.9–3.7 2.9 2.3–2.7 2.4P, mg/L NM – ND – ND – 0.03–0.05 0.04Al, mg/L NM – 0.2–0.7 0.3 0.2–0.6 0.3 0.4–1.2 0.8Fe, mg/L NM – 0.8–3.6 2.0 0.7–2.1 1.5 1.0–3.1 2.3Cu, mg/L NM – ND – 0.01–0.03 0.01 0.004–0.006 0.005Zn, mg/L NM – 0.0–0.036 0.02 ND – 0.02–0.05 0.04TSS, mg/L 5.8–23.5 15.7 2.4–10.0 5.9 3.9–15.0 8.7 9.7–27.9 13.4DO, mg/L NM – 6.5–7.8 7.0 NM – NM –pH 6.6–6.9 6.8 7.2–7.7 7.5 6.5–6.8 6.7 6.5–6.7 6.6Alkalinity, mg/Le 16–28 19.5 NM – 16–28 20.8 16–40 24.8

a Sampling dates were 28 March 2001, 9 May 2002, 12 May 2003, and 10 May 2004. The headwater sampling was conducted at aminimum of six sampling sites upstream of the urban area. The tributary samples were collected from three major tributariesdischarging into the main channel of the Reedy River. Solutes indicated by asterisk are for filtered samples. b NM, not measured.cMass ratio. d ND, not detected. e As CaCO3.

Environ Monit Assess (2007) 132:171–187 177

pollutant sources. The major nonpoint sources inurban areas include storm water runoff and over-flowing sewage. Thus, direct impacts of nonpointsources from urban areas on the stream chemistry arelikely to be minimal under baseflow conditions.

Very small increases in the total concentrations ofbase cations (Na, K, Ca and Mg) were observed alongthe urban reach within the city of Greenville (up to∼15 km downstream of RR30; Figure 3). Corre-spondingly, major anions such as chloride, sulfate,and nitrate also increased slightly within the urbanarea (Figure 4). These increases in cation and anionconcentrations suggest input of water containinghigher concentrations of the solutes from the urban

landscape. It is interesting that such source inputs intothe river from the urban watershed were observedunder baseflow conditions.

Elevated baseflow concentrations of total copperand total zinc generally were not observed within theurban Greenville area during this study (Figure 5).Typically, contamination by these two trace metals inurban watersheds is closely associated with particu-late sources such as debris from automobiles andpaved roads which becomes more important duringstorm events. For example, Andrews and Sutherland(2004) noted that copper and zinc in sediments of anurban watershed sampled after a storm event wereenriched two and five times respectively, compared to

0

10

20

30

40

50

0 10 20 30 40 50 60 70 80 90

Na

K

Ca

Mg

2003

0

10

20

30

40

50

0 10 20 30 40 50 60 70 80 90

Na

K

Ca

Mg

2004

0

10

20

30

40

50

0 10 20 30 40 50 60 70 80 90

Na

K

Ca

Mg

2002

Distance downstream of RR30 (km)

Con

cent

ratio

n (m

g/L)

Urban

MR-WWTP

LR-WWTP

0

10

20

30

40

50

0 10 20 30 40 50 60 70 80 90

Na

K

Ca

Mg

2003

0

10

20

30

40

50

0 10 20 30 40 50 60 70 80 90

Na

K

Ca

Mg

2004

0

10

20

30

40

50

0 10 20 30 40 50 60 70 80 90

Na

K

Ca

Mg

2002

Distance downstream of RR30 (km)

Con

cent

ratio

n (m

g/L)

Urban

MR-WWTP

LR-WWTP

Figure 3 Concentrations oftotal sodium, potassium,calcium, and magnesiumdownstream of samplingsite RR30 for 2002 to 2004(cation concentrations werenot measured in 2001). Thevertical lines indicate thewastewater treatment plantlocations.

178 Environ Monit Assess (2007) 132:171–187

samples collected under baseflow conditions. Similarenrichments have also been observed in our studyarea during storm events (unpublished data).

The most visible trend within the urban area ofGreenville was the dramatic decrease (i.e., ∼70% in2002, ∼50% in 2003 and 2004) in total iron concen-trations with downstream distance (Figure 5). To date,we have no explanation for this phenomenon, butshould point out that it does not appear to be related toprocesses affecting TSS or DOC. For example, noconsistent trends for TSS concentrations, DOC con-centrations, or SUVA254 values (Figure 6) were found

that paralleled the total iron concentration trend. It isalso noteworthy that the total aluminum concentrationdid not exhibit a similar trend to iron (Figure 5), asmight be expected for TSS- and/or DOC-controlledprocesses.

3.3 Effect of WWTP effluents on water chemistryparameters in the Reedy River

Effluent flow rate data for the two municipal WWTPssituated on the Reedy River are shown in Table II forthe four sampling dates. The effluent flow from the

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2003

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2001

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2002

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2004

Urban

MR-WWTP LR-WWTP

Distance downstream of RR30 (km)

Con

cent

ratio

n (m

g/L)

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2003

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2001

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2002

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2004

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2001

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2002

0

10

20

30

40

0 10 20 30 40 50 60 70 80 90

Cl

SO4

NO3

2004

Urban

MR-WWTP LR-WWTP

Distance downstream of RR30 (km)

Con

cent

ratio

n (m

g/L)

Figure 4 Concentrations ofdissolved chloride, sulfate,and nitrate downstream ofsampling site RR30 for2001 to 2004. The verticallines indicate the wastewa-ter treatment plant locations.

Environ Monit Assess (2007) 132:171–187 179

MR-WWTP is three to four times higher than thatfrom the LR-WWTP, suggesting that the upstreamWWTP has a larger effect on the overall waterchemistry of the Reedy River. The contribution ofeach WWTP effluent to the stream flow immediatelydownstream from its discharge point was estimatedfrom the values in Table II and data from the USGSgauging station nearest each WWTP. The contributionof the MR-WWTP ranged from 26% to 51%, with thehighest and lowest contribution being observed forthe most severe drought spring season (2002) and theabove-normal precipitation spring season (2003),respectively (Table II). The flow contribution of theLR-WWTP, conversely, never exceeded 10%, andexhibited a much smaller annual variation that wasnot directly related to the drought versus non-droughtconditions.

Except for magnesium, concentrations of basecations in the Reedy River generally increased imme-

diately downstream of the MR-WWTP (Figure 3). Themagnesium concentrations in the MR-WWTP effluents(personal communication from Western Carolina Re-gion Sewer Authority) were similar to those of theheadwater and tributary samples, whereas the otherbase cation effluent concentrations were between 5 and20 times higher than their background concentrations.The latter observation is consistent with that ofVengosh and Keren (1996), who reported that WWTPeffluents are likely the major contributor to enrichmentof salts in rivers under baseflow conditions. Sodiumappears to be the most abundant base cation in both ofthe WWTP effluents on the Reedy River, as suggestedby the observed trends shown in Figure 3. Base cationconcentrations generally decreased with distancedownstream from the MR-WWTP, but did show aslight enhancement again at the LR-WWTP (Figure 3).The decreasing trend with distance is conceptuallysimilar to that predicted by a point source being diluted

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2002

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2003

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2004

Urban

MR-WWTP

LR-WWTP

Cu

or Z

n co

ncen

trat

ion

(mg/

L)

Fe

or A

l con

cent

ratio

n (m

g/L)

Distance downstream of RR30 (km)

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2002

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2003

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2004

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2002

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2003

0

0.02

0.04

0.06

0.08

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4Cu

Zn

Fe

Al

2004

Urban

MR-WWTP

LR-WWTP

Cu

or Z

n co

ncen

trat

ion

(mg/

L)

Fe

or A

l con

cent

ratio

n (m

g/L)

Distance downstream of RR30 (km)

Figure 5 Concentrations oftotal copper, zinc, iron, andaluminum downstream ofsampling site RR30 for2002 to 2004 (metal con-centrations were not mea-sured in 2001). The verticallines indicate the wastewa-ter treatment plant locations.

180 Environ Monit Assess (2007) 132:171–187

as described by Andersen et al. (2004), in whichconservative solutes are diluted by groundwater andtributary inputs.

Chloride is often assumed to be a conservativetracer in many aquatic systems. Therefore, the

decreasing chloride concentrations downstream ofthe WWTPs (Figure 4) likely reflect physical dilutionby tributaries and groundwater inputs. The ReedyRiver experiences the effluent effects of the twoWWTPs as far as 85 km downstream from MR-

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

TSS2001

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4

5

6TSSDOCSUVA

2002

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4

5

6TSSDOCSUVA

2003

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4

5

6TSSDOCSUVA

2004

Urban

MR-WWTPLR-WWTP

Distance downstream of RR30 (km)

TS

S c

once

ntra

tion

(mg/

L)

DO

C c

once

ntra

tion

(mg/

L) o

r S

UV

A25

4(L

/mg-

m)

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

TSS2001

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4

5

6TSSDOCSUVA

2002

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4

5

6TSSDOCSUVA

2003

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90

0

1

2

3

4

5

6TSSDOCSUVA

2004

Urban

MR-WWTPLR-WWTP

Distance downstream of RR30 (km)

TS

S c

once

ntra

tion

(mg/

L)

DO

C c

once

ntra

tion

(mg/

L) o

r S

UV

A25

4(L

/mg-

m)

Figure 6 Concentrations oftotal suspended solids, dis-solved organic carbon, andspecific UV absorbance val-ues for dissolved organicmatter downstream of sam-pling site RR30 for 2001 to2004 (not all parameterswere measured in 2001).The vertical lines indicatethe wastewater treatmentplant locations.

Table II Effluent flowrates (m3/s) for the Mauldin Road and Lower Reedy wastewater treatment plants, and their estimated percentcontributions to the flow in the main channel of the Reedy River immediately downstream of their respective discharge pointsa

Sampling date MR-WWTP Contribution (%) LR-WWTP Contribution (%)

28 March 2001 0.795 39 0.189 4.79 May 2002 0.908 51 0.216 7.212 May 2003 0.810 26 0.295 4.610 May 2004 0.677 44 0.222 7.5

a Data based on personal communication from Western Carolina Region Sewer Authority and USGS gauging station flowrate data.

Environ Monit Assess (2007) 132:171–187 181

WWTP, before the chloride concentrations againbecome similar to the background values observedfor the headwaters and tributaries. In particular,dramatic increases in anion concentrations immedi-ately below the MR-WWTP were observed in 2001and 2002 when its contribution to the flow in theReedy River was the greatest (Table II). Although thegeneral trends of downstream concentration decreaseswere similar for all anions, the relative rate of changefor each anion differed in some instances. Forexample, sulfate concentrations decreased moreabruptly than chloride for the 2003 spring season,although they declined more gradually along thereach for the other years. Very high correlationsbetween sulfate and chloride were observed in thesamples downstream of MR-WWTP for the 2001,2002, and 2004 spring seasons (r=0.98, 0.99, and0.93, respectively), suggesting that sulfate was be-having relatively conservatively in those particularyears. However, the correlation between sulfate andchloride was weaker (r=0.81) in 2003, suggesting thatchemical and/or biological processes may haveenhanced the downstream attenuation of sulfate thatspring season.

The lack of any major changes in the concentrationsof iron and aluminum immediately downstream of theWWTP locations (Figure 5) demonstrates that theireffluents did not greatly affect the levels of thesemetals in the Reedy River. In contrast, the concen-trations of copper and zinc downstream of the WWTPlocations were irregular from year to year, suggestingthat their concentrations were affected by temporalvariations in the compositions of the WWTP effluents.In particular, zinc concentrations appeared to increasedownstream of the WWTPs in 2002 and 2004, but notin 2003. Presumably, the latter observation resultedfrom dilution effects in 2003, although the zinc trendsmay be biased due to the low concentration rangesmeasured. Trends for both iron and aluminum concen-trations downstream of the MR-WWTP were consis-tent with the TSS concentration data (Figure 6). Theratio between aluminum and iron molar concentrations(Al:Fe) increased from 0.1 to 0.6 along the reach, andthis trend was consistent for the 2002 to 2004 springseasons (data not shown; metal concentrations notmeasured in 2001). With increasing distance down-stream, Al:Fe approached the molar concentration ratioof 0.9 as reported for aluminum and iron in the soils ofthe upper Piedmont region of SC (Franklin, Duis,

Smith, Brown, & Toler, 2003). This suggests that thecontribution of soils and sediment to the downstreamwater chemistry in the Reedy River becomes greater,presumably because the water has a longer time tocontact the soils and sediments along the reach.

Effluent discharge from the MR-WWTP into theReedy River resulted in sharp increases (70% to210%) in DOC concentrations (Figure 6). In contrast,SUVA254 values were similar to values measured forthe headwater and tributary samples. These resultssuggest that the SUVA254 characteristics (i.e., aromat-ic content) of the WWTP effluents are not greatlydifferent from those of the natural stream, presumablybecause the physiochemical and biological WWTPprocesses remove the DOC components containinghigh SUVA254 values (Herpe & Troch, 2000).Alternatively, it may be that the SUVA254 measure-ments merely are not sensitive enough to distinguishbetween the carbon composition of the WWTPeffluents and the background carbon composition ofthe main channel.

The overall trends of measured water chemistryparameters along the reach indicate that the ReedyRiver is most affected by the MR-WWTP effluent,which is primarily responsible for the concentrationsof major anions and base cations being higher thanbackground levels. Although the LR-WWTP effluentis partially responsible for changes in the downstreamwater chemistry, its contribution is much smallerrelative to the MR-WWTP effluent. In addition,although two impoundments are located on the mainchannel of the Reedy River, their potential effects onwater chemistry were not readily apparent based onmerely examining solute concentrations, becausenotable and consistent changes in solute concentra-tions were absent immediately downstream from theimpoundments.

3.4 Effects of drought on the attenuation of nitratefrom WWTPs in the Reedy River

In general, changes in solute concentrations downstreamof a point source reflect the net result of physical,chemical, and/or biological uptake and release process-es. For conservative (i.e., non-reactive) solutes, concen-tration changes are controlled by physical processesonly. To investigate the potential reactive behavior ofnitrate due to biochemical processes, nitrate andchloride concentrations downstream from the MR-

182 Environ Monit Assess (2007) 132:171–187

WWTP were normalized by their respective concen-trations at RR19, the nearest sampling site downstreamfrom the effluent discharge location and which wasexpected to have the highest concentrations.

A distinct difference in the relative decreasingconcentration trends downstream of the MR-WWTPwas observed for the two severe drought springseasons of 2001 and 2002 versus the above-normalprecipitation spring season of 2003 (Figure 7). Underdrought conditions, normalized nitrate concentrationswere slightly higher than the chloride values, with thedifference showing a somewhat increasing trend withdownstream distance. The apparently weaker dilutionof nitrate versus chloride is likely due to the relativelyhigher nitrate concentrations that were being added tothe river by groundwater and tributaries. Excellentpositive correlations (r > 0.96) between chloride andnitrate concentrations in 2001 and 2002 (Figure 8)suggest that physical processes were primarily con-trolling downstream nitrate concentrations. Althoughdenitrification has often been suggested as a majorremoval process of nitrate in watersheds (Kellman &Hillaire-Marcel, 1998; Sjodin, William, & Saunders,1997), its role appears to have been minimal in theReedy River during the 2001 and 2002 springseasons. This may have been due to the relativelysufficient oxygen content of the river (e.g., DO levelsalways above 6.0 mg/l, data not shown), but more likelyreflects the fact that biological uptake, an importantnitrate processing function in many watersheds, mayhave been limited under the severe drought conditions.Biological processing of nitrate, in turn, likely waslimited due to reduction of critical algal communities(Dahm, Baker, Moore, & Thibault, 2003; Wall, Phillips,& Riva-Murray, 1998). For example, Wall et al. (1998)reported that uptake by diatoms was responsible fordecreasing nitrate concentrations in Canajoharie Creek,New York, based on the strong correlation betweennitrate and silica in samples collected along the mainchannel. They also suggested that the diatom popula-tion might decline during drought periods due to thelack of necessary elements supplied by groundwater.

In contrast to the drought spring seasons, normal-ized nitrate concentrations declined more rapidly thannormalized chloride concentrations in 2003 (Figure 7).For the below-normal precipitation spring season of2004, the trend was intermediate between the droughtconditions versus the above-normal precipitationcondition. Up to ∼20 km downstream of the MR-

WWTP, normalized nitrate concentrations werehigher than those for chloride but fell below normal-ized chloride concentrations beyond this distance.Much weaker correlations (r=0.71 and 0.80 for 2003and 2004, respectively) between chloride and nitrateconcentrations were observed for above- and below-normal precipitation spring seasons (Figure 8). Theseobservations thus can be interpreted as the reactivebehavior of nitrate downstream from the MR-WWTP.Biological uptake, the process most likely responsiblefor nitrate removal, therefore appears to be moreimportant when sufficient nutrients and elements canbe supplied from groundwater.

To further investigate the apparent conservative-like behavior of nitrate during the two drought springseasons, nitrate concentrations were divided by theircorresponding chloride concentrations at each sam-pling location (Figure 9), which is equivalent to theratio of nitrate to chloride mass loads in the river ateach location. Thus, if nitrate concentrations and/ormass loads were only being affected by physicalprocesses, the nitrate to chloride ratios shown inFigure 9 should have remained reasonably constant inthe river or perhaps approached the ratios of theheadwater and tributary samples. However, the resultsclearly demonstrate a distinct difference in the trendsfor the ratios in drought versus non-drought springseasons downstream of the MR-WWTP. For the twodrought spring seasons, a slight increase in the ratiowas observed downstream from MR-WWTP, thevalues being closer to the ratio for headwater andtributary samples. This observation suggests that thedownstream nitrate is controlled primarily by physicalrather than biochemical processes. Although theeffluent from the LR-WWTP may have affected theratio, its low flow rate contribution to the mainchannel would have made any such effect small.Therefore, no discernable change in the ratio wasobserved near LR-WWTP.

In contrast to the two drought spring seasons, netreductions in the nitrate to chloride ratios wereobserved downstream of the MR-WWTP in 2003 and2004. In fact, the lowest values for the ratio were evenlower than the range observed in headwater andtributary samples those two years (Table I), indicatingthat the decrease in nitrate was attributable to bothbiochemical and physical processes. The decrease wasmore in the 2003 above-normal precipitation springseason than it was for the 2004 below-normal precip-

Environ Monit Assess (2007) 132:171–187 183

Distance downstream from MR-WWTP (km)

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

ClNO3

2001

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

ClNO3

2002

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

Cl

NO3

2003

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

Cl

NO3

2004

Downstream from MR-WWTP (km)

Nor

mal

ized

con

cent

ratio

ns0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

ClNO3

2001

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

ClNO3

2002

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

Cl

NO3

2003

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50 60 70

Cl

NO3

2004

Downstream from MR-WWTP (km)

Nor

mal

ized

con

cent

ratio

ns

Downstream from MR-WWTP (km)

Nor

mal

ized

con

cent

ratio

ns

Figure 7 Concentrations ofchloride and nitrate normal-ized by their respectiveconcentrations at RR19, thenearest sampling site down-stream of the Mauldin Roadwastewater treatment plant.

184 Environ Monit Assess (2007) 132:171–187

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 10 20 30 40 50 60 70 80 90

Distance downstream of RR30 (km)

Nitra

te/c

hlo

rid

e r

atio

2001

2002

2003

2004

MR-WWTP LR-WWTP Boyd Mill Pond

UrbanFigure 9 Nitrate-to-chlorideconcentration ratios down-stream of RR30 for 2001 to2004. The vertical lines indi-cate the locations of thewastewater treatment plantsand one of the impound-ments on the Reedy River(Boyd Mill Pond).

Chloride concentration (mg/L)

Nitr

ate

conc

entr

atio

n (m

g/L)

0

2

4

6

8

0 10 20 30 40

2001

r = 0.961

0

3

6

9

12

15

0 10 20 30 40 50

2002

r = 0.973

0

2

4

6

8

0 3 6 9 12 15

2003

r = 0.713

0

2

4

6

8

10

0 5 10 15 20 25

2004

r = 0.797

Chloride concentration (mg/L)

Nitr

ate

conc

entr

atio

n (m

g/L)

0

2

4

6

8

0 10 20 30 40

2001

r = 0.961

0

3

6

9

12

15

0 10 20 30 40 50

2002

r = 0.973

0

2

4

6

8

0 3 6 9 12 15

2003

r = 0.713

0

2

4

6

8

10

0 5 10 15 20 25

2004

r = 0.797

Figure 8 Scatterplots and correlations between concentrations of nitrate and chloride downstream from the Mauldin Road wastewatertreatment plant.

Environ Monit Assess (2007) 132:171–187 185

itation spring season. The difference in trends between2001 and 2002 versus 2003 and 2004 providesadditional evidence of how drought can alter the reac-tive behavior of nitrate in upper Piedmont watersheds.

It is interesting to note that the nitrate trends weobserved in 2004 were different from those in 2001and 2002 despite the fact that all three spring seasonshad comparable 30- and 90-day antecedent averagestream flow rates. Instead, the major differencebetween 2004 and 2001/2002 was the total length oftime that precipitation shortfalls generally had existedprior to the sampling dates. For example, the springseasons of 2001 and 2002 followed approximately 3and 4 years, respectively, of drought conditionswhereas the spring season of 2004 followed theabove-normal precipitation year of 2003. If ourhypothesis is correct that reduction in algal commu-nities is primarily responsible for the conservative-like behavior of nitrate in the Reedy River in 2001and 2002, then it appears that a drought must be inplace sufficiently long to substantially decrease algaepopulations. It also appears that algal communities areable to recover quickly when precipitation and flowrates increase again, as evidenced by the nitrate trendsin 2003. These observations are qualitatively consis-tent with previously reported effects of long-term,intense droughts on aquatic ecosystems (Boulton,2003; Humphries & Baldwin, 2003; Lake, 2003).For example, Lake (2003) has suggested that droughtsact like ‘ramps’ to bring about concomitant ramp-likedetrimental responses in organisms that increase withtime. Alternatively, Boulton (2003) has argued thatalthough droughts act as ramp disturbances, theresponse of aquatic biota is more appropriatelydescribed as a ‘stepped’ rather than ‘ramped’ processas a drought progresses and various hydrological and/or geomorphological thresholds are crossed.

4 Conclusions

In this study, effects of drought and urban activities,including WWTP effluents, on the water chemistry ofthe Reedy River were evaluated using grab samplescollected annually under non-storm event conditionsfrom 2001 through 2004. Headwater and tributarysamples showed that the Reedy River has relativelylow ion concentrations compared to other nearbywatersheds, and that its spring season background

water chemistry was reasonably stable from year toyear, even under drought conditions. Relatively feweffects of the urbanized land area of Greenville itselfwere found for the concentrations of metals such ascopper and zinc in the river. Concentrations of majoranions, including nitrate, increased along the urban-ized reach, suggesting that the river receives contin-uous inputs of these solutes from the urban area evenunder non-storm event conditions. Effects of WWTPeffluents on the water chemistry of the Reedy Riverwere evident, with the highest concentrations of majorcations and anions typically being observed directlydownstream from the WWTP locations. Our resultssuggest that nitrate concentrations downstream of theWWTPs were influenced by chemical and biologicalprocesses in the river, as well as by physical processessuch as dilution. Biological processing of the nitratedischarged by WWTPs to the Reedy River appears tobe greatly diminished under drought conditions.

Acknowledgments We greatly appreciate the laboratory andfield assistance provided to us throughout this study by IlkeErdogan, Miranda Williams, Saugath Lahiri, Jesse Cannon,Adam Ryan, Eric Van Genderen, Kat Hanson, and Laura Terry.We also wish to acknowledge one anonymous reviewer for theconstructive comments on the organization and content ofthe manuscript. Primary funding for this study was provided bythe Natural Resources Conservation Service of the U.S.Department of Agriculture (NRCS-69-4639-1-0010) throughthe Changing Land Use and Environment (CLUE) Project atClemson University. Additional support was provided by theCooperative State Research, Education, and Extension Serviceof the U.S. Department of Agriculture (CSREES/USDA), underproject number SC-1700278. Any opinions, findings, conclu-sions or recommendations expressed in this publication aresolely those of the authors and do not necessarily reflect theviews of the USDA/NRCS. This is Technical Contribution No.5199 of the Clemson University Experiment Station.

References

Andersen, B. C., Lewis, G. P., & Sargent, K. A. (2004).Influence of wastewater-treatment effluent on concentra-tions and fluxes of solutes in the Bush River, SouthCarolina, during extreme drought conditions. Environmen-tal Geosciences, 11, 28–41.

Andrews, S., & Sutherland, R. A. (2004). Cu, Pb and Zncontamination in Nuuanu watershed, Ohau, Hawaii.Science of the Total Environment, 324, 173–182.

Beasley, B. R., Criss, M., Quattro, L., & Scharf, R. (2001). TheReedy River report: Managing a watershed. Report no. 22(p. 116). Columbia, SC: South Carolina Department ofNatural Resources: Land, Water and Conservation Division.

Berestov, A. L., Fernando, H. J. S., & Fox, P. (1998). Modeling

186 Environ Monit Assess (2007) 132:171–187

of transport and seepage in effluent-dominated streams.Water Resources Research, 34, 3025–3033.

Boulton, A. J. (2003). Parallels and contrasts in the effects ofdrought on stream macroinvertebrate assemblages. Fresh-water Biology, 48, 1173–1185.

Brooks, B. W., Stanley, J. K., White, J. C., Turner, P. K., Wu,K. B., & La Point, T. W. (2004). Laboratory and fieldresponses to cadmium: An experimental study in effluent-dominated stream mesocosms. Environmental Toxicologyand Chemistry, 103, 1057–1064.

Caruso, B. S. (2002). Temporal and spatial patterns of extremelow flows and effects on stream ecosystems in Otago,New Zealand. Journal of Hydrology, 257, 115–133.

Chen J., He, D., Zhang, N., & Cui, S. (2004). Characteristics ofand human influences on nitrogen contamination inYellow River system, China. Environmental Monitoringand Assessment, 93, 125–138.

Clark, G. M., Mueller, D. K., & Mast, M. A. (2000). Nutrientconcentrations and yields in undeveloped stream basins ofthe United States. Journal of the American WaterResources Association, 36, 849–860.

Clawges, R. M., & Vowinkel, E. F. (1996). Variables indicatingnitrate contamination in bedrock aquifers, Newark Basin,New Jersey. Water Resources Bulletin, 32, 1055–1066.

Dahm, C. N., Baker, M. A., Moore, D. I., & Thibault, J. R.(2003). Coupled biogeochemical and hydrologicalresponses of streams and rivers to drought. FreshwaterBiology, 48, 1219–1231.

Deksissa, T., Ashton, P. J., & Vanrolleghem, P. A. (2003).Control options for river water quality improvement: Acase study of TDS and inorganic nitrogen in the CrocodileRiver (South Africa). Water SA, 29, 209–217.

Diamond, J. M., Hall, J. C., Pattie, D. M., & Gruber, D. (1994).Use of an integrated monitoring approach to determinesite-specific effluent metal limits. Water EnvironmentResearch, 66, 733–743.

Ellis, J. B., & Revitt, D. M. (2002). Sewer losses and interactionswith groundwater quality. Water Science and Technology,45, 195–202.

Franklin, R. E., Duis, L., Smith, B. R., Brown, R., & Toler, J. E.(2003). Elemental concentrations in soils of South Carolina.Soil Science, 168, 280–291.

Ha, S.-R., & Bae, M.-S. (2001). Effects of land use and municipalwastewater treatment changes on streamwater quality.Environmental Monitoring and Assessment, 70, 135–151.

Herpe, Y. V., & Troch, P. A. (2000). Spatial and temporalvariations in surface water nitrate concentrations in amixed land use catchment under humid temperate climateconditions. Hydrological Processes, 14, 2439–2455.

Humphries, P., & Baldwin, D. S. (2003). Drought and aquaticecosystems: An introduction. Freshwater Biology, 48,1141–1146.

Jordan, T. E., Weller, D. E., & Correll, D. L. (2003). Sources ofnutrient inputs to the Patuxent river estuary. Estuaries, 26,226–243.

Karanfil, T., Erdogan, I., & Schlautman, M. A. (2003).Selecting filter membranes for measuring DOC andUV254. Journal of the American Water Works Associa-tion, 95, 86–100.

Kellman, L., & Hillaire-Marcel, C. (1998). Nitrate cycling in

streams: Using natural abundances of NO3-Sigma-15N tomeasure in-situ denitrification. Biogeochemistry, 43, 273–292.

Kim, K., Lee, J. S., Oh, C.-W., Hwang, G.-S., Kim, J., Yeo, S.,et al. (2002). Inorganic chemicals in an effluent-dominatedstream as indicators for chemical reactions and streamflows. Journal of Hydrology, 264, 147–156.

Lake, P. S. (2003). Ecological effects of perturbation by drought inflowing waters. Freshwater Biology, 48, 1161–1172.

Lucey, K. J., & Goolsby, D. A. (1993). Effects of climatevariations over 11 years on nitrate–nitrogen concentrationsin the Raccoon river, Iowa. Journal of EnvironmentalQuality, 22, 38–46.

Maidment, D. R. (1993). Handbook of hydrology. New York:McGraw-Hill.

Marti, E., Aumatell, J., Gode, L., Poch, M., & Sabater, F.(2004). Nutrient retention efficiency in stream receivinginputs from wastewater treatment plants. Journal ofEnvironmental Quality, 33, 285–293.

McMahon, P. B., Tindall, J. A., Collins, J. A., Lull, K. J., & Nuttle,J. R. (1995). Hydrologic and geochemical effects on oxygen-uptake in bottom sediments of an effluent-dominated river.Water Resources Research, 31, 2561–2569.

Parr, L. B., & Mason, C. F. (2003). Long-term trends in waterquality and their impact onmacroinvertebrate assemblages ineutrophic Lowland River. Water Research, 37, 2969–2979.

Reilly, J. F., Horne, A. J., & Miller, C. D. (2000). Nitrateremoval from a drinking water supply with large free-surface constructed wetlands prior to groundwater re-charge. Ecological Engineering, 14, 33–47.

Rhodes, A. L. (2001). Influence of land use on water quality ofa diverse New England watershed. Environmental Science& Technology, 35, 3640–3645.

Robson, A. J., &Neal, C. (1997). Regional water quality of the RiverTweed. Science of the Total Environment, 194, 173–192.

Saunders, J. F., & Lewis, W. M. (2003). Implications ofclimatic variability for regulatory low flows in the SouthPlatte river basin, Colorado. Journal of the AmericanWater Resources Association, 39, 33–45.

Sjodin, A. L., William, M. L., & Saunders, J. F. (1997).Denitrification as a component of the nitrogen budget for alarge plains river. Biogeochemistry, 39, 327–342.

Sonoda, K., Yeakley, J. A., & Walker, C. E. (2001). Near-stream land use effects on streamwater nutrient distribu-tion in an urbanizing watershed. Journal of the AmericanWater Resources Association, 37, 1517–1532.

Vengosh, A., & Keren, R. (1996). Chemical modifications ofgroundwater contaminated by recharge of treated sewageeffluent. Journal of Contaminant Hydrology, 23, 347–360.

Wall, G. R., Phillips, P. J., & Riva-Murray, K. (1998). Seasonaland spatial patterns of nitrate and silica concentrations inCanajoharie Creek, New York. Journal of EnvironmentalQuality, 27, 381–389.

Wayland, K. G., Long, D. T., Hyndman, D. W., Pijanowski, B. C.,Woodhams, S. M., & Haack, S. K. (2003). Identifyingrelationships between baseflow geochemistry and land usewith synoptic sampling and R-mode factor analysis. Journalof Environmental Quality, 32, 180–190.

Whitall, D., Castro, M., & Driscoll, C. (2004). Evaluation ofmanagement strategies for reducing nitrogen loadings to fourUS estuaries. Science of the Total Environment, 15, 25–36.

Environ Monit Assess (2007) 132:171–187 187