1.Fate of effluent organic matter and DBP precursors-WR-2009

Fate of effluent organic matter and DBP precursors in aneffluent-dominated river:A case study of wastewater impact on downstream water qualityBaiyang Chen a ,*,Seong-Nam Nam b ,Paul K.Westerhoff c ,Stuart W.Krasner d ,Gary Amy e ,faChinese Environmental Scholars and Professionals Network,3004S 101st Drive,Phoenix,AZ 85353,United StatesbUniversity of Colorado,Dept.of Civil,Environmental &Architectural Engineering,ECOT 441,1111Engineering Dr.,Boulder,CO 80309,United States cArizona State University,Department of Civil &Environmental Engineering,Engineering Center,Box 5306,Tempe,AZ 85287-5306,United States dMetropolitan Water District of Southern California,Water Quality Laboratory,700Moreno Avenue,La Verne,CA 91750,United States eUNESCO-IHE,Institute for Water Education,P.O.Box 3015,2601DA Delft,The Netherlands fTechnical University of Delft,2600AA Delft,The Netherlandsa r t i c l e i n f oArticle history:Received 2October 2008Received in revised form 25December 2008Accepted 14January 2009Published online 3February 2009Keywords:Disinfection by-product Effluent organic matter Dissolved organic nitrogen Wastewater Fate and transporta b s t r a c tThe impact of treated wastewater discharges on downstream water quality was evaluated in an effluent-dominated stream in the Southwest USA.The fate and transport of effluent organic matter (EfOM)and disinfection by-product (DBP)precursors was studied.Nitrifi-cation and biodegradation were important mechanisms.Changes in DBP formation potential along the river appeared to correlate with dissolved organic carbon (DOC)and organic nitrogen concentrations and specific ultraviolet absorbance.The mean oxidation state of carbon (MOC)decreased in value along the river.MOC decreases paralleled decreases in the biodegradability of residual DOC (i.e.,lower biodegradable DOC/DOC ratio).The EfOM was biodegradable by up to 40percent,both in the stream and in a laboratory reactor,and many DBP precursors (e.g.,haloacetonitriles,certain nitrosamines)decreased in concentration.Alternatively,the DBP yields for trihalomethanes or haloacetic acids either remained the same or increased slightly,suggesting that these precursors were part of the recalcitrant organic matter (OM).ª2009Elesevier Ltd.All rights reserved.1.IntroductionIn addition to planned water recycling and reclamation programs,unintentional indirect potable reuse of wastewater has been recognized over the past few decades (Bunch et al.,1961),which will likely increase in the future as upstreamwastewater treatment plants (WWTPs)discharge water into rivers or lakes that serve as downstream drinking-water supplies.Drought and competing in-stream demands may result in substantial (e.g.,37–77%in one effluent-impacted river)contribution of treated wastewater towards the stream flow (Krasner et al.,2008).Attention has recently focused on*Corresponding author .Tel.:þ14808404647;fax:þ16029448605.E-mail addresses:poplar_chen@ (B.Chen),nams@ (S.-N.Nam),p.westerhoff@ (P.K.Westerhoff),skrasner@ (S.W.Krasner),g.amy@ (G.Amy).A v a i l a b l e a t w w w.s c i e n c e d i r e c t.c o mj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /w a t r e s0043-1354/$–see front matter ª2009Elesevier Ltd.All rights reserved.doi:10.1016/j.watres.2009.01.020w a t e r r e s e a r c h 43(2009)1755–1765wastewater-derived pharmaceuticals and endocrine disruptors (Kolpin et al.,2002).However,WWTP discharges are also a source of other contaminants,including disinfection by-products(DBPs)(when chlorine disinfection is practiced)and DBP precursors.Certain DBPs may pose risk to aquatic organ-isms,however such information is yet to be widely published. In addition,certain DBPs are of health and regulatory concern for downstream drinking water treatment plants.Chlorination(oxidation)of amino acids can result in the formation of aldehydes and nitriles,with subsequent or concomitant chlorine substitution to form chloral hydrate (trichloroacetaldehyde)and dichloroacetonitrile,respectively (Trehy et al.,1986).Treated wastewater effluent organic matter (EfOM),which is rich in dissolved organic nitrogen(DON)(Dot-son et al.,2008),has been shown to be a source of precursors for a wide range of DBPs(trihalomethanes[THMs],haloacetic acids [HAAs],haloacetonitriles[HANs],haloacetaldehydes[HAs],and nitrosamines)(Krasner et al.,2008).DBPs in EfOM may be attenuated in receiving waters due to various biogeochemical mechanisms(biodegradation, photolysis,hydrolysis,volatilization,and adsorption)(Chen et al.,2008).A major loss mechanism for certain DBP precur-sors is biodegradation(Krasner et al.,2008).The objective of this study was to evaluate the fate and transport of EfOM and DBP precursors in an effluent-dominated stream,so as to better understand the impact of treated wastewater discharges on downstream drinking-water supplies.2.Materials and methods2.1.Sampling programThe Santa Cruz River(Arizona)was sampled in June2004and February2005.During both sampling periods,there was zero flow in the river above the point of discharge from the Nogales International(NI)WWTP.Thus,this research served as an end-member case study on a100percent effluent-dominated stream.Studying end-member systems facilitates under-standing of more complex hydrologic systems that may have significant dilution,but it is the premise of this study that such extrapolations could be taken by linearly accounting for dilution.This WWTP,a10million gallon per day(MGD, equivalent to37,850cubic meters per day)facility,used aerated lagoon treatment with chlorination/dechlorination prior to discharge.Fig.1shows the sampling locations for this study.Sites0and1(S0and S1)corresponded to the treated wastewater before and after chlorination,respectively. Sampling locations extended14.3miles(22.9km)downstream (to sampling site6[S6])during the June2004sampling campaign and23.8miles(38.1km)in February2005(to sampling site10[S10]).Flow had decreased significantly at the last downstream site for each event.United States Geological Survey(USGS)gauging stations were available on this section of the Santa Cruz River:upstream of the discharge(Station#: 09480500);downstream of the discharge at Tubac(Station#: 09481740)(site6)and at Amado(Station#:09481770)(site10). Theflow velocity and distance were used to estimate travel times.In February2005,the estimated travel time over the entire reach(to Amado)was4.2days(2.5days to Tubac).Diurnal variations inflow rate at the WWTP discharge point were mirrored at downstream USGS gauging stations.A total of seven sites(i.e.,sites0–6)were sampled for eight times from June1to3,2004in the morning(8am–11am), midday(11am–4pm),and at the end of the day(4pm–8pm; except for June3).Not all of the analytical parameters were measured at all sites or at all of the times.A total of nine sites (i.e.,sites0–6,8,and10)were sampled once on February2005 at midday(11am–4pm),except for the pharmaceutical pri-midone(a chemical that can be used as a wastewater tracer or indicator),which was sampled at the WWTP at the beginning and end of the day.Field measurements were made for dis-solved oxygen,pH,and temperature.Grab samples were collected from mid-river at mid-depth.Samples for laboratory analyses were stored in glass or plastic bottles,placed on ice, andfiltered within24h where required.2.2.Analytical parametersThe samples were analyzed for a combination of conventional wastewater and drinking water parameters,including5-day carbonaceous biochemical oxygen demand(CBOD5),chemical oxygen demand(COD),total Kjeldahl nitrogen(TKN),turbidity, electrical conductivity,ammonia(NH4þ),nitrite(NO2À),and nitrate(NO3À).The mean oxidation number of carbon(MOC)was determined(Vogel et al.,2000)by MOC¼4À1.5ÂCOD org/DOC, where COD org¼organic chemical oxygen demand¼total CODÀnitrite-consumed COD in the COD test.In order to determine if any additional water sources impacted the river, some refractory chemicals were also monitored,including chloride,sulfate,phosphate,and primidone(Krasner et al., 2008).Primidone is an anticonvulsant,which has been found to be a conservative tracer(i.e.,relatively recalcitrant)of waste-water impact(Krasner et al.,2006a).General natural organic matter(NOM)measurements included total and dissolved organic carbon(TOC,DOC),dis-solved organic nitrogen(DON),and specific ultraviolet absor-bance(SUVA).The DON method employed dialysis pretreatment to remove dissolved inorganic nitrogen(DIN)to reduce analytical error(Lee and Westerhoff,2005).SUVA was determined by the ratio of UVA at254nm(in mÀ1)to DOC on samples that have been0.45-m mfiltered,and application of the SUVA parameter allowed classification according to humic(>4L/mg m)versus non-humic(<2L/mg m)NOM (Edzwald and Van Benschoten,1990).NOM characterization methods included XAD resin frac-tionation(Aiken et al.,1992),5-day biodegradable dissolved organic carbon(BDOC5)measurements(Allgeier et al.,1996), size-exclusion chromatography(SEC)with DOC detection (SEC-DOC)(Her et al.,2002a),andfluorescence excitation–emission matrix(EEM)measurements(McKnight et al.,2001). Amberlite XAD-8and XAD-4resins were used to separate the NOM into three operationally defined groups:(1)hydrophobic (HPO),(2)transphilic(TPI),and(3)hydrophilic(HPI)organic matter.Although NOM is a heterodisperse mixture of compounds,SEC-DOC provided an estimation of apparent molecular weight(MW)distributions of the NOM,which separated the organic matter into three general components by size:(1)polysaccharides,proteins,and colloids,comprising biopolymers(BP),which elute at early retention times,(2)w a t e r r e s e a r c h43(2009)1755–1765 1756humic substances (HS),and (3)low-MW organic acids (LMWA).Polyethylene glycol (PEG)standards ranging from 200to 10,000Daltons (Da)were used to calibrate molecular weight range assignment during SEC-DOC analysis (Her et al.,2002a ).An EEM was represented by a 3-dimensional spectrum showing fluorescence intensity as a function of excitation wavelength and emission wavelength,with ‘‘peak’’location(s)indicative of NOM composition.Excitation at 280nm and emission at 350nm is an example of a protein-like fluorophore,as a constituent of soluble microbial products (SMPs),whereas excitation at 320nm and emission at 440nm represents a humic-like peak (Nam and Amy,2008).The fluorescence index corresponds to the fluorescence intensity ratio of the emission at 450nm to the emission at 500nm,both at an excitation of 370nm.McKnight and colleagues (2001)found that for terrestrially derived fluorophores,the fluorescence index ranged from 1.3to 1.4,whereas for microbially derived organic matter the index was 1.7–2.0.In addition to the regulated DBPs (i.e.,THMs and HAAs),other known or emerging DBPs that have been or are likely associated with treated wastewater were measured.These included HANs,HAs,N-nitrosodimethylamine (NDMA)(Mitch et al.,2003),and other nitrosamines.In addition to chloral hydrate,dihalogenated and/or brominated analogues of chloral hydrate were also studied (Krasner et al.,2006b ).Both instantaneous DBPs and DBP precursors were measured (Krasner et al.,2008).Disinfectant residuals for THM,HAN,and HAA samples were quenched with ammonium chloride,whereas ascorbic acid was used to quench disinfectant residuals in HA samples.Chinn and colleagues(2007)Fig.1–Location of the Santa Cruz River and sampling sites (see Table 1for distances between sampling points;direction of flow is northward towards Tubac).Source:Reproduced from Krasner et al.(2008).Copyright 2008Awwa Research Foundation.w a t e r r e s e a r c h 43(2009)1755–17651757demonstrated that HAs were stable in water samples when a stoichiometric amount of ascorbic acid(with a slight excess as a safety factor)was added and pH adjustment to3–4(to prevent base-catalyzed hydrolysis)was conducted.Disinfec-tant residuals for nitrosamine samples were quenched with sodium sulfite.The DBP precursors were measured using formation potential(FP)tests(Krasner et al.,2004).Precursors for halogenated DBPs were determined in DBPFP tests con-ducted in the presence of chlorine,whereas NDMA precursors were measured in FP tests done with chloramines.3.Results and discussion3.1.Water quality parameters and wastewater tracerTable1summarizes part of the collected data.Electrical conductivity(800–900m mhos/cm)decreased by less than10 percent along the length of the Santa Cruz River during both sampling events.Chloride(w50mg/L)and sulfate(20mg/L) increased by w10percent probably due to evapotranspiration (with riparian vegetation).Phosphate concentrations(5mg/L) decreased by w10percent.Primidone concentrations did not significantly change(relative to analytical and temporal vari-ability)from the WWTP discharge point(95–149ng/L)to site6 (67–109ng/L)in June2004or from the WWTP discharge point (66–77ng/L)to site10(58ng/L)in February2005.Overall,the results for conductivity,chloride,sulfate,phosphate,and primidone were consistent and reinforced the assumption that no dilution from other unknown surface waters was occurring along the reach;given the depth of the groundwater table,groundwater contributions can be discounted.As expected,ammonia,nitrite,and nitrate concentrations significantly changed along the length of the river(Fig.2).In-stream ammonia was oxidized to nitrite and ultimately to nitrate.Overall,total dissolved nitrogen(TDN)(which included DON)decreased from w28to18mg/L as N in the summer event.Less ammonia transformation was observed in February2005(12–16 C)than in June2004(28–30 C), presumably due to the lower water temperature and less microbial activity.Dissolved oxygen concentrations over the reach decreased from4to2–3mg/L in the summer event.The total N delinquency(w10mg/L)between upstream and downstream sites was presumably due to ammonia volatili-zation from the gaseous form(the water was at pH¼7.8),N fixation by plants,adsorption by sediment,and/or other biogeochemical processes.3.2.Instantaneous DBPsAnalyses for instantaneous DBPs indicated that very low concentrations were produced at the WWTP.The chlorine dose(2.7–5.2mg/L in June2004)was insufficient to break out the ammonia,so combined chlorine(chloramines)was formed.For the midday sampling on June1,2004,the WWTP effluent THM and HAN were all<1m g/L,whereas dihalo-genated HAA(DXAA)and trihalogenated HAA(TXAA) concentrations were6.6and5.1m g/L,respectively.At site2, DXAAs and TXAAs had decreased to1.4and4.1m g/L,respec-tively,and were both<1m g/L by site6.The HAAs were likely removed by biodegradation within the Santa Cruz River;other research has shown DXAAs to be more biodegradable than TXAAs(Baribeau et al.,2005).At other sampling times,the HAAs were well removed before site6too.Nitrosamine concentrations were low(<1–5ng/L of NDMA,<1–2ng/L of N-nitrosomorpholine)in the WWTP effluent and throughout the Santa Cruz River.Overall,instantaneous DBP concentra-tions were very low in the WWTP effluent and stream.Thus, a greater emphasis was placed on DBP precursors.3.3.EfOM parametersanic carbonIn February2005,the TOC and DOC concentrations in the WWTP effluent were21and13mg/L,respectively,indicating a particulate organic carbon(POC)level of w8mg/L.POC concentrations decreased(w50percent)to w4mg/L by site6. This is consistent with a55percent decrease in turbidity in the river,from45NTU in the WWTP effluent to20NTU down-stream.This suggested that sedimentation,filtration during parafluvialflow through gravel banks,and/or interception by aquatic plants of particulate matter occurred in the river channel.The DOC content in the WWTP effluent was slightly higher in February2005(13mg/L)than in June2004(9mg/L).DOC concentration decreased by w40percent along the length of the river(Fig.3).Because in-stream dilution was negligible,the losses of DOC were primarily attributed to microbial metabolic activity in the stream,which utilize DOC as a carbon source and ammonia as the electron donor(Fig.2).3.3.2.DONDON concentrations in the Nogales WWTP effluent ranged from0.70to1.47mg/L as N between the different sampling times and dates.DON concentrations decreased along the length of the Santa Cruz River by17percent on average in June 2004and by41percent in February2005(Fig.3).DON removal was similar to or less than the change of DOC concentration (w40percent).On average,the DOC/DON ratio along the stream length decreased from11:1to8:1by the furthest downstream point in June2004.This trend may reflect DON production by algae,which were visibly present within the stream channel,and/or preferential removal of the DOC. Because algae will add DON and DOC,preferential loss of DOC seems more likely.Even though there were modest changes in the DOC/DON ratio,there was a good linear correlation between DOC and DON for the overall data set (DON¼0.0957ÂDOCþ0.0915;r2¼0.90).3.3.3.UVA at254nmUVA decreased by w10–20percent along the length of the Santa Cruz River,which was a lower percentage removal than either DOC or DON(Fig.3).This suggests that the biodegrad-able fraction of EfOM had less UVA than the refractory frac-tion.In addition,part of the UVA removal may have been due to adsorption onto stream sediment and/or photolysis. Because UVA had a lower percentage change than DOC,SUVA values increased by38and21percent in June2004and February2005,respectively,along the length of the riverw a t e r r e s e a r c h43(2009)1755–1765 1758(Fig.4)(higher SUVA values reflect greater hydrophobicity and less biodegradability).3.3.4.Oxygen demand by organic matterCBOD 5,BDOC,and COD were measured on select samples.COD decreased by w 26percent along the length of the Santa Cruz River in June 2004,but was unchanged in February 2005.MOC values of the Nogales WWTP effluent were between À1and À2in June 2004,and between 1.71and 1.87in February 2005.MOC can theoretically range from À4(most reduced form of carbon,CH 4)to þ4(most oxidized form of carbon,CO 2).Water collected downstream had lower MOC values (i.e.,À2to À3in June 2004and 0.29–0.37in February 2005).Low MOCs (e.g.,high COD/DOC ratios)indicate the presence of more refractory carbon mole-cules,suggesting that more oxygenated organic matter was preferentially biodegraded in the river.BDOC 5was measured at multiple points in the stream (Fig.5).The WWTP effluent had 5.6and 10.8mg/L of BDOC 5in June 2004and in February 2005,respectively.BDOC 5represented 40and 66percent of the initial DOC in these two sample events,respec-tively.Downstream of the WWTP,BDOC 5decreased to 2.8and 4.0mg/L,respectively,which corresponded to 30and 37percent of the DOC at the downstream sites.The difference in BDOC 5between the WWTP effluent and the most downstream site sampled was 2.8and 6.8mg/L in June 2004and in February 2005,respectively.The actual amount of DOC removed in the river in these two sample events was 3.4and 5.2mg/L,respectively,which was consistent with the BDOC 5data.Given the estimated travel time of the entire reach,the 5-day time frame of the BDOC test was appropriate.3.3.5.NOM characterizationXAD fractionation,SEC-DOC,and EEM analyses were con-ducted on several sets of samples from the Santa Cruz River.The samples were characterized by 31–44percent HPO,12–25percent TPI,and 34–55percent HPI.The HPO fraction decreased somewhat in the river (from 4.9–5.8to 3.4–4.2mg/L as C),whereas the HPI fraction was removed to a larger extent (from 5.0–5.9to 3.1–3.4mg/L as C in June 2004and from 9.1to 5.8mg/L as C in February 2005).The HPI fractions of waste-water often contain sugars,free amino acids,and other easily biodegradable compounds (Thurman,1985).Based on SEC-DOC,the samples were 34–54percent HS (similar to the amount of HPO fraction amount),22–35percent BP,and 13–44percent LMWA.Fig.6shows SEC-DOC chromatograms before and after a 5-day BDOC test for the Nogales WWTP effluent (June 2004).As shown,EfOM from this WWTP effluent had diverse MW components.The polydispersity value (r ¼12),which was determined by the weight-average MW (i.e.,5200Da)divided by the number-average MW (i.e.,410Da)for the entire chro-matogram,was much higher than single-sourced NOM such as Suwannee River humic acid (r ¼1.5–2)(Her et al.,2002b;Nam and Amy,2008).In particular,the EfOM contained a large portion of high-MW substances (retention time w 30min,with weight-[M w ]and number-average [M n ]MWs of 14,590and 13,640Da,respectively),due to BP macromolecules.A signifi-cant decrease in the high-MW peak in the BDOC 5test indicates that these BP materials greatly contributed to the biodegrad-ability of the EfOM,whereas HS (retention time w 50min)contributed a minor portion and little was apparently due to LMWA.A slight increase in the LMWA at w 70–80min was observed for the day 5sample,possibly due to the shifting of the HS peak to smaller MW substances by the breakdown of the organic matter.For the entire chromatogram of the EfOM,biodegradation resulted in a reduction in both M w and M n (M w was reduced from 5200to 3790Da and M n was reduced from 410to 310Da).In addition,there was a slight decrease in the polydispersity.These observations confirm the preferential and/or prevalent biodegradation of high-MW substances,which resulted in a larger contribution (and a shifting to)smaller M w and M n .The change in polydispersity,although small,indicates that biodegradation resulted in the remaining organic matter becoming relatively more homogeneous.Based upon the EEMs,a protein-like peak was present in all samples,which diminished in intensity along the river reach.During the June 2004sample event,the fluorescence indices of samples decreased somewhat on average (from 1.47[standard deviation ¼0.07]to 1.41[standard deviation ¼0.02])along the river flow,but were not different within the variability of the results.During the February 2005sample event,the fluores-cence indices ranged from 1.50to 1.71,but the results at sites 1and 10(i.e.,1.64and 1.59,respectively)were essentially the same.These values fall between allochthonous and autoch-thonous NOM,likely reflecting a mixture of SMPs (from bio-logical treatment)and NOM from the corresponding drinking water source.Fig.7shows the EEMs of the aerobic BDOC testing of the Nogales WWTP effluent.After a 5-day BDOC test,the protein-like peak was substantially less intense,whereas the humic-like peaks were relatively unchanged.A differen-tial EEM of the BDOC test shows that the protein-like peak was almost removed,whereas the humic-like peaks still remained.A parallel BDOC test was conducted under anoxic conditions;more humic substances tended to be degraded in the anoxic BDOC test (results not shown).5101520253035Distance (miles)N H 4+ (m g /L a s N )012345678NO 2-, NO 3- (mg/L as N)Fig.2–Variations in the inorganic nitrogen species along the Santa Cruz River (June 2004)from the WWTP discharge point (mile 0)and downstream 14.3miles (22.9km).Error bars represent the standard deviation of samples collected at different times.Source:Reproduced from Krasner et al.(2008).Copyright 2008Awwa Research Foundation.w a t e r r e s e a r c h 43(2009)1755–176517603.4.DBP precursorsPrecursors for all types of DBPs were present in the WWTP effluent,which underwent changes along the length of the Santa Cruz River.Fig.3summarizes changes in EfOM char-acteristics and selected DBP precursor levels during both seasons at multiple sites along the Santa Cruz River.In FP tests,because of the high chlorine dosages and long contact times,chlorine can outcompete (in terms of halogen substi-tution)the bromine produced by oxidation of ambient bromide by chlorine.For these FP samples,chloroform wasthe dominant THM species formed (84–91%of the total THMs on a molar basis).Thus,the observed changes in DBPFP values were attributed to changes in organic matter amount and composition instead of any variations of bromide concentration.THMFP was relatively constant over the length of the Santa Cruz River in June 2004(Fig.3),whereas it decreased some-what (i.e.,26percent,from 267to 197m g/L)in February 2005.Thus,the THM precursors were primarily associated with the non-biodegradable organic matter.Site 0(WWTP effluent)sampled midday on June 2,2004had higher precursor levels,20406080100120140DOCUVA254DON THMFP DXAAFPTXAAFPR e m a i ni n g O r g a n i c M a t t e r (%)0mile2.5mile 5.407.5mile 10.5014.3mileDOCUVA254DONTHMFPDXAAFPTXAAFPHAFPNDMAFPR e m a i n i n g O r g a n i c M a t t e r (%)0mile2.5mile7.5mile14.3mile23.8mileFig.3–Changes in bulk organic matter and selected DBP precursors along the river reach relative to the most upstream location (mile 0)on a percentage basis (Note:seven HA species were detected in June 2004,but only three dihalogenated HA species in February 2005;the error bars represent the standard deviation of 4sampling times on June 2004).Source:Reproduced from Krasner et al.(2008).Copyright 2008Awwa Research Foundation.w a t e r r e s e a r c h 43(2009)1755–17651761which was attributed to a slightly higher DOC concentration (9.8versus 8.1–9.4mg/L at other sample times in June 2004).The THM yield varied between 0.2and 0.3m mol/mg DOC and exhibited a dependence upon SUVA in pooled samples from different sites and different collection times (June 2004:THMFP ¼0.13ÂSUVA À0.02;r 2¼0.69).Similar to the THM precursors,HAAFP was relatively constant over the length of the Santa Cruz River in June 2004(Fig.3),whereas DXAAFP decreased by w 40percent (from 218to 130m g/L)and TXAAFP by 26percent (from 140to 92m g/L)in February 2005.Similar to the THMs,the HAA precursors d especially those associated with the TXAAs d were primarily associated with the non-biodegradable organic matter.The HAA yields exhibited a strong dependence on SUVA and ranged from 0.1to 0.3m mol/mg DOC for DXAAFP (June 2004:DXAAFP ¼0.21ÂSUVA À0.24;r 2¼0.82)and from 0.05to 0.2m mol/mg DOC for TXAAFP (June 2004:TXAAFP ¼0.10ÂSUVA À0.10;r 2¼0.86).As discussed above,MOC decreased in value along the Santa Cruz River.On the one hand,MOC decreases were paralleled with decreases of biodegradability of residual DOC (i.e.,lower biodegradable DOC/DOC ratio).On the other hand,the DBP yields per unit DOC for THMs or HAAs either remained the same or increased somewhat along the river.Again,this indicated that THM and HAA precursors were probably more likely to be part of the recalcitrant NOM.HA precursors decreased along the length of the Santa Cruz River (Fig.3).The sum of the concentration of seven species at sites 0,2,4,and 6sampled midday on June 1,2004were 97,90,73,and 48m g/L,respectively.Chloral hydrate precursors accounted for w 80percent of this DBP class sum.Overall,the change in HA precursors across the reach was 50percent.Changes in nitrogenous DBP precursors were also observed.The removal of HAN precursors in the Santa Cruz River tracked that of the DOC and DON.As for chloropicrin,nitrite was reported to be a potential source of nitrogen in the nitro group of chloropicrin (Choi and Richardson,2004).In this study,there was a peak in chloropicrin FP at site 4in June 2004,then the FP decreased at downstream locations.Similarly,in February 2005,there was a peak in chloropicrin FP at site 3(Table 1).Nitrite is unstable,and its concentration may have differed between the time of nitrite analysis and the time of FP testing.However,these results may confirm early literature that nitrite plays an important role in chloropicrin formation.NDMA was the major nitrosamine whose precursors were detected in all samples.N-nitrosopyrrolidine (NPYR)precur-sors were detected at lower concentrations and were found in more of the samples collected in February 2005.NDMA precursor concentrations exhibited an exponential decrease (from 682–804to 280–347ng/L in terms of FP)along the length of the Santa Cruz River in June 2004(Fig.3).Although most of the data were for the midday sampling on June 1,2004,data from other sampling times corresponded well.During the February 2005sampling,NDMAFP decreased from 1396to 721ng/L over the length of the river.During the latter event,NPYRFP decreased from 35to 19–23ng/L.Fig.8presents correlations between NDMA precursors in the Santa Cruz River and EfOM parameters.Linear0.12.57.514.323.8Distance (miles)B D OC 5 (m g /L )0.000.100.200.300.400.500.60BDOC 5/DOC (mg/mg)FlowFig.5–BDOC 5of samples from the Nogales WWTP and Santa Cruz River (average values for June 2004).Source:Reproduced from Krasner et al.(2008).Copyright 2008Awwa Research Foundation.Fig.6–SEC-DOC chromatograms before and after a 5-day BDOC test for the Nogales WWTP effluent (June,2004)(M w and Mn denote weight-and number-average MW,respectively).Distance (miles)S U V A (L /m g -m )Fig.4–Changes in SUVA along the Santa Cruz River (error bars represent the standard deviation).Source:Reproduced from Krasner et al.(2008).Copyright 2008Awwa Research Foundation.w a t e r r e s e a r c h 43(2009)1755–17651762。