Changes in stoichiometric Si, N and P ratios of Mississippi River water diverted through coastal wetlands to the Gulf of Mexico

Estuarine, Coastal and Shelf Science 60 (2004) 1e10

www.elsevier.com/locate/ECSS

Changes in stoichiometric Si, N and P ratios of Mississippi Riverwater diverted through coastal wetlands to the Gulf of Mexico

Robert R. Lanea,b,), John W. Daya,b, Dubravko Justica,b, Enrique Reyesa,b,Brian Marxc, Jason N. Daya, Emily Hyfielda,b

aCoastal Ecology Institute, Louisiana State University, Baton Rouge, LA 70803, USAbDepartment of Oceanography and Coastal Sciences, Louisiana State University, Baton Rouge, LA 70803, USA

cDepartment of Experimental Statistics, Louisiana State University, Baton Rouge, LA 70803, USA

Received 26 December 2002; accepted 21 November 2003

Abstract

During the spring of 2001, we monitored nutrient concentrations and stoichiometry of diverted Mississippi River water as itflowed through the Breton Sound estuary, Louisiana, USA. River water was discharged through a diversion structure at Caernarvonas a two-week pulse that peaked at 220 m3 s�1. There were reductions in observed concentrations of TN, TP, DIN, DIP and DSi, of

up to 44%, 62%, 57%, 23%, and 38%, respectively, as water flowed through the estuary. TN, TP, DIN, DIP and DSiconcentrations in the river were 137e140, 5.0e5.1, 104e153, 1.1e1.3 and 114e121 mM, respectively, and 36e122, 1.8e3.6, 13e119,0.3e1.8 and 29e110 mM, respectively, at the Gulfward end member stations. The DSi:DIN ratio rose from 0.9 at the Caernarvon

diversion to 2.6 at the Gulf end member station, while the DIN:DIP ratio fell from 107 to 26. This study shows that freshwaterdiversions can significantly alter riverine nutrient concentrations and ratios and reduce the overall amount of exported nitrogen.� 2004 Elsevier Ltd. All rights reserved.

Keywords: river diversion; nutrient ratios; nutrient reduction

1. Introduction

Historically, many rivers had dissolved inorganic sili-con (DSi) concentrations well in excess of dissolvedinorganic nitrogen (DIN) or dissolved inorganic phos-phorus (DIP), thus contributing to N or P deficiency inriver plumes and adjacent coastal waters (Turner andRabalais, 1991; Conley et al., 1993; Humborg et al.,1997; Humborg et al., 2000). However, due to decreas-ing DSi concentrations and increasing DIN and DIPlevels during the 20th century, the proportions of nu-trients have changed in many of the world’s large rivers(Justic et al., 1995a). Using the atomic Si:N:P ratio of16:16:1 (Redfield, 1958) as a criterion for balancednutrient composition, one can distinguish betweenphosphorus deficient rivers (e.g., Changjiang, Huanghe,

) Corresponding author. Coastal Ecology Institute, Louisiana State

University, Baton Rouge, LA 70803, USA.

E-mail address: [email protected] (R.R. Lane).

0272-7714/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ecss.2003.11.015

Mackenzie, Yukon), nitrogen deficient rivers (e.g.,Amazon and Zaire), silicon deficient rivers (e.g., Rhineand Seine), and those where the dissolved Si:N:P ratioapproaches the Redfield ratio (e.g., Po and Mississippi)(Justic et al., 1995b).

In the Mississippi River, there have been significantincreases in DIN and DIP and decreases in DSi con-centrations during the last 50 years, with DIN and DIPconcentrations more than doubling and DSi concen-trations decreasing by half (Rabalais et al., 1996). Asa result, the DSi:DIN ratios in the lower MississippiRiver have changed from about 4:1 to 1:1 (Rabalaiset al., 1996). There has been a concurrent increase in theincidence of phytoplankton blooms and bottom waterhypoxia in the coastal waters of the northern Gulf ofMexico, presumably in response to increased riverineDIN and DIP inputs, and more balanced nutrient ratios(Turner and Rabalais, 1991; Justic et al., 1995b;Rabalais et al., 1996). The northern Gulf of Mexico ispresently the site of the largest (O20,000 km2) and the

2 R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

most severe coastal hypoxic zone in the western AtlanticOcean region (Rabalais and Turner, 2001). It is widelybelieved that decreasing DSi:DIN ratios in riversexacerbate the eutrophication process in the adjacentcoastal waters, primarily by reducing the potential fordiatom growth in favor of noxious non-siliceous formssuch as dinoflagellates (Officer and Ryther, 1980;Cloern, 2001). Diatoms are generally heavily grazedupon and are rarely a nuisance. Heavily silicifieddiatoms, however, may be responsible for large verticalfluxes of carbon to bottom sediments in the Gulf ofMexico during spring, resulting in the formation of thehypoxic zone during the summer (Dortch et al., 2001;Turner, 2001).

The high DIN concentrations in the Mississippi Riverare a result of two factors. First, there was a dramaticincrease in agricultural fertilizer use in the second half ofthe 20th century (Turner and Rabalais, 1991; Rabalaiset al., 1996). Second, wetland and riparian ecosystems inthe Mississippi basin have been greatly reduced anddrainage efficiency of agricultural lands has been greatlyincreased (Mitsch et al., 2001). The lower Mississippiand its major tributaries are largely isolated from floodplain ecosystems by levees. This is especially so in theMississippi delta where the river is confined by leveesalmost to the mouth. This isolation of the river from thedelta is one of the most important factors contributingto the high rate of wetland loss in the Mississippi delta(Day et al., 2000). To counter coastal wetland losscaused by lack of river input, the State of Louisiana andthe Federal Government have developed a comprehen-sive management plan that includes diversions of riverwater back into coastal wetlands (Coast 2050, 1998).One of the largest diversions is located on the east bankof the Mississippi River about 20 km south of NewOrleans at Caernarvon, Louisiana.

The use of coastal wetlands and shallow water bodiesto process Mississippi River water before entering theGulf of Mexico has been proposed as a partial solutionto help reduce the size of the hypoxic zone, as well asrestore and maintain rapidly eroding coastal wetlands(Coast 2050, 1998; Mitsch et al., 2001). Studies haveshown that when water flows through wetland-domi-nated watersheds, there is a reduction in nutrientconcentrations, especially in the concentration of nitrate(Richardson and Nichols, 1985; Fisher et al., 1988;Kadlec and Knight, 1996; Nixon et al., 1996; Boustanyet al., 1997; Lane et al., 1999, 2002; Spieles and Mitsch,2000; Mitsch et al., 2001). Lane et al. (2002), forexample, estimated that 41e47% of Atchafalaya RiverNO3

� was non-conservatively lost from the watercolumn in the Atchafalaya Delta estuarine complexbefore reaching stratified Gulf waters. Similar reduc-tions in nitrogen have been reported to occur in manyestuaries (Smith et al., 1983; Jenkins and Kemp, 1984;Lane et al., 1999) and wetland wastewater treatment

systems (Richardson and Nichols, 1985; Boustany et al.,1997).

The work presented in this paper is based on a largeinterdisciplinary project called PULSES studying theeffects of the Caernarvon diversion on the Breton Soundestuary. We report on changes in nutrient concentra-tions and ratios in water flowing from the Caernarvondiversion through the Breton Sound estuary. Wehypothesized that the diversion would lead to an overalldecrease in DIN concentrations, and an increase inDSi:DIN ratios.

2. Study area

The Caernarvon freshwater diversion structure islocated on the east bank of the Mississippi River southof New Orleans at river mile 81.5 (Fig. 1), and consistsof five 4.6-m-wide box culverts with vertical lift gates,with a capability of passing 226 m3 s�1 of river water.The structure was completed in 1991 and freshwaterdischarge began in August of that year. The BretonSound estuary lies between the Caernarvon diversionstructure and Breton Sound Bay, and comprises ap-proximately 1100 km2 of fresh and brackish wetlandsand shallow water ponds, lakes and bays. Lane et al.(1999) found that salinity was significantly lowered inthe Breton Sound estuary due to operation of theCaernarvon structure, with salinity in the upper half ofthe estuary generally lower than 5 psu throughout theyear. Diverted water must travel about 30e40 kmthrough two major routes dominated by wetlands beforereaching the open waters of Breton Sound, and anadditional 50 km before reaching open Gulf waters. Thetwo major routes are via (1) Lake Leary and BayouTerra aux Boeufs (BTAB) to the east, which carriesapproximately two-third of the flow and (2) Manuel’sCanal and River aux Chene (RAC) to the west of BigMar, which carries about one-third of the flow (Laneet al., 1999). Levees prevent Mississippi River waterfrom entering into the upper Breton Sound estuary.However, river water can flow directly into the baybelow the point where levees end on the eastern bank ofthe river (Fig. 1), and may impact water quality in thatarea.

3. Methods

During spring of 2001, an experimental large pulse ofriver water with a peak flow of 226 m3 s�1 was releasedthrough the Caernarvon structure for 16 days (Fig. 2).Sampling was carried out during weekly transects fromMarch 9 to March 30, 2001. Discrete water sampleswere taken at 19 locations in the Breton Sound estuary(Fig. 1). They were collected in 1 L acid-washed

3R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

Fig. 1. Map of the study area showing the location of the Caernarvon diversion structure and water quality monitoring sites in the Breton Sound

estuary. Arrows along the lower Mississippi River show where the flood control levee ends and water can flow from the river to the estuary during

high river stages. The stippled area indicates the upper 441 km2 of Breton Sound estuary; shaded areas are uplands protected by flood control levees.

polyethylene bottles, were immediately stored on ice,and taken to the laboratory for processing. Within 24 h,60 ml from each water sample was filtered through pre-rinsed 25-mm 0.45-mm Whatman GF/F glass fiber filtersinto acid-washed bottles and frozen until nutrientanalysis. Salinity was determined using an Atago�S-10 hand-held refractometer (accuracy: G 2 psu).Nitrate plus nitrite (NO3CNO2) was determined usingthe automated cadmium reduction method with an

Alpkem� autoanalyzer (Greenberg et al., 1985). Am-monium (NH4-N) was determined by the automatedphenate method, phosphate (PO4-P) by the automatedascorbic acid reduction method, and silicon (SiO4-Si) bythe automated molybdate reagent/oxalic acid method, allwith an Alpkem� autoanalyzer (Greenberg et al., 1985).Total nitrogen (TN) and total phosphorus (TP) weredetermined by the methods described by Valderrama(1981). The accuracy of the nutrient analysis was

4 R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

checked every 20 samples with a known standard, andthe samples were redone if the accuracy was off by 5%.

3.1. Water budget analysis

A water budget analysis was carried out to illustratethe importance of freshwater input from the Caernarvondiversion. Preliminary analysis of the results indicatedthat most chemical reactions occurred in the upper halfof the estuary, making a water budget for this regionessential. As part of another study, the volume of alllakes, bayous, canals and marsh ponds in the upper441 km2 of the Breton Sound estuary was calculated(Erick Swenson, personal communication). The regionwas delineated by Bayou Terra aux Boeufs and theMississippi River levee (see stippled area in Fig. 1). Thewater volume of this region, determined from aerialphotos and field checked depth estimates, was calculatedto be 212 million cubic meters at mean tide (ErickSwenson, personal communication).

3.2. Analysis of nutrient reduction efficiency

Changes in observed nutrient concentrations for TN,TP, DIN, DIP and DSi were calculated using Eq. (1),

Observed nutrient reduction ð%Þ¼ ððMassin �MassoutÞ=MassinÞ100 ð1Þ

where Massin is the concentration of the constituent inincoming river water at the Caernarvon diversion, andMassout is the concentration at the end memberstation(s). Since water passes through two major routes,Bayou Terra aux Boeufs (BTAB) and River aux Chene(RAC), observed nutrient reduction was calculatedseparately for each route and weighted, depending onthe percentage of flow in each route, then summed (Eq.(2)). Five pairs of stations along the two main flowroutes were chosen in order to characterize nutrientreduction: stations 5 and 13; 6 and 14; 8 and 15; 10 and17; and 11 and 18. The stations on each route wereselected because of direct hydrologic connection to each

Fig. 2. Discharge of Mississippi River water from the Caernarvon

diversion structure during spring 2001. Arrows indicate water quality

sampling dates on March 9, 15, 22 and 30, 2001.

other, and paired because of their similarity in distancefrom the diversion structure.

Eq. (1) was modified to account for the different flowin the two pathways as shown in Eq. (2),

Total observed nutrient reduction ð%Þ¼ ð0:66!NRstBTABC0:34!NRstRACÞ100 ð2Þ

where the percentage of diverted Mississippi River waterthrough each route (66% and 34% for the BTAB andRAC routes, respectively) was multiplied by the nutrientreduction of the respective station. The sum of the totalpercent reduction for each route was used to estimatethe percent reduction from station 1 to the respectiveend member.

Analysis of the salinity results discussed below impliesthat water from the diversion had a residence time ofabout two weeks before reaching the outermost stations.For this reason, the loading rate analysis used averagesof March 9 and 15 concentrations at station 1 as initialconcentrations (Massin), and the concentrations at thevarious pairs of end member stations on March 22 asending concentrations (Massout). To account for di-lution with Gulf waters, as evident from salinity at thelast two pairs of stations (10e17 and 11e18, Fig. 4),Eqs. (1) and (2) were modified to Eqs. (3) and (4). Theseequations compensate for the effects of dilution bydividing salinity on March 22 by salinity on March 9,when the outermost stations were not affected by thediversion, and subtracting the resulting fraction from 1.Similarly, Eq. (5) was used to compensate for dilution inthe end member concentration by adding 1 to thefraction obtained above and then multiplying the sumby the respective station concentration. Eqs. (3)e(5)were used to calculate nutrient reduction efficiencies andstoichiometric ratios for the last two pairs of stations.

Observed nutrient reduction ð%Þ¼ fððMassin Mar9 �Massout Mar22Þ=Massin Mar9Þ

�ð1� salMar22=salMar9Þg100 ð3Þ

Total observed nutrient reduction ð%Þ¼ fð0:66!NRstBTABð1� salMar22=salMar9ÞÞ

Cð0:34!NRstRACð1� salMar22=salMar9ÞÞg100 ð4Þ

Undiluted conc: ¼ Concendmembð1CsalMar22=salMar9Þ ð5Þ

3.3. Stoichiometric nutrient ratio analysis

Stoichiometric nutrient ratios were determined usingmolar concentrations of DSi, DIN and DIP. Stoichio-metric ratios for the last two pairs of stations weredetermined from concentration ratios after accountingfor dilution by using Eq. (5).

5R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

3.4. Statistical analyses

A simple linear model was used to test for linearity ofthe various constituents with distance. The two routes,BTAB and RAC, were tested separately in full andreduced models. The full model tested separate slopesand intercepts for each route, while the reduced modelused a common slope and intercept for both routes. Thefull model was used only when the reduced model wasfound to be non-significant using Sum of Squares andMean Square Error terms (a! 0:05).

Statistical analyses of nutrient concentrations andratios were carried out to determine changes in waterquality at the diversion structure compared to the sta-tions located within the estuary. The mean of incomingriver water for the first two dates (March 9 and 15) wascompared to the mean of each paired set of samplestaken on March 22 using Dunnett’s procedure, withincoming river water designated as the control. Themean of all estuarine stations combined was also com-pared to the mean of incoming river water usingStudent’s t-test. Nutrient concentrations used for boththe Dunnett’s procedure and the Student’s t-test weremodified to account for dilution by using Eq. (5).

4. Results

The spring 2001 pulse from the Caernarvon diversionbegan on March 6 with discharge rapidly rising fromnear zero to 212 m3 s�1 within 48 h. Discharge on thewater quality sampling dates was 211 m3 s�1 on March9, 212 m3 s�1 on March 15, 56 m3 s�1 on March 22, and12 m3 s�1 on March 30 (Fig. 2). A total of 269 millioncubic meters of river water was introduced during thepulse, compared to the 212 million cubic meters cal-culated as the volume of the upper 441 km2 of theestuary at mean tide (Erick Swenson, personal commu-nication), indicating a water replacement time of 12.6days. If the 269 million cubic meters were spreaduniformly, it would cover the upper 441 km2 of theestuary with 61 cm of water, or the entire 1100 km2 ofthe estuary with 25 cm. This is much greater than the6 cm of rainfall that occurred during the same period(measured at the Class A weather station at the NewOrleans International Airport).

Winds tended to circulate clockwise in a series of coldfronts with an approximate period of one week duringJanuary, February and March, and decreasing thereaf-ter (Fig. 3). These shifts in wind, combined with thediversion, led to water level variations that oftenoverwhelmed tidal influences and flooded the surround-ing marsh surface for most of the two-week pulse. Up to40 cm of water was measured on the marsh surface inthe region between Big Mar and Grand Lake (GreggSnedden, personal communication). Perez et al. (2000)found extremely high TSS concentrations in FourleagueBay, Louisiana, associated with cold front passages, andhypothesized that erratic water level fluctuations asso-ciated with winter storms are a major mechanism for thetransport of sediments to surrounding marshes. Therewere two frontal passages during the 2001 spring pulse,passing through the system on March 7 and 21 (Fig. 3).

Salinity ranged from 0 psu at the diversion to 17 psuat the end member stations (Fig. 4). The most pro-nounced decrease in salinity occurred on March 22, twoweeks after the onset of the large pulse (March 9). Forexample, salinities at stations 10 and 17, over 25 kmaway from the diversion, decreased from 11 and 4 psuon March 15 to 2 and 1 psu on March 22, respectively.This rapid decrease in salinity at the two end memberstations, followed by just as rapid an increase insalinities after the pulse ended (Fig. 4, March 30),indicates that the Caernarvon diversion had an over-whelming influence on salinity, and that diverted waterscompletely displaced ambient water in the estuary. Thiscorroborates the results of the water volume calculationsthat indicated a total inflow from the diversion of 269million cubic meters compared to the total volume of theupper 441 km2 of the estuary of 212 million cubicmeters.

There were rapid, non-conservative reductions in TN,TP, DIN, and DSi in the first 20 km of the study areawith little change further into the estuary (Figs. 5 and 6).However, DIP was released in the first 20 km and thendecreased (Fig. 6). Concentrations at station 1 (riverwater) of TN, TP, DIN, DIP and DSi were 137e140,5.0e5.1, 104e153, 1.1e1.3 and 114e121 mM, respec-tively, and 36e122, 1.8e3.6, 13e119, 0.3e1.8, and29e110, respectively, at the end member stations (Table 1).There were observed decreases in the concentration

-6-4-20246

January AprilMarchFebruary May

Win

d S

pee

d(m

s-1

)

Cold Front

Cold Front

Fig. 3. Wind speed data taken from New Orleans International Airport. Arrows indicate water quality sampling dates.

6 R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

of up to 44% inTN, 62% inTP, 57% inDIN, 23% inDIPand 38% in DSi (Fig. 6).

Simple linear regression analysis failed to reject thereduced model, and found a linear response withdistance, with a common slope and intercept for bothroutes, for DIN and TN. DIN and TN concentrationswere found to be decreasing with distance from thediversion structure at a rate of 3.1 (p! 0:0001) and 2.0

1 2 123419

135 614

16715

8

17

9

10

18

11

0

5

10

15

0 10 20 30 4

1 2 12341913

5 61416715

8

17

9

10

18

11

0

5

10

15

0 10 20 31 2 123 419135 61416715 8

179

10 18

11

0

5

10

15

0 10 20 30 4

1 2 123

4

1913

5 61416

715

817

9

10

18

11

0

5

10

15

0 10 20 3

March 9’ 01

March 30’ 01March 22’ 01

March 15’ 01

Sal

inity

(ps

u)

Sal

inity

(ps

u)S

alin

ity (

psu)

Sal

inity

(ps

u)

Distance (km) Distance (km)

Distance (km)Distance (km)

Fig. 4. Salinity versus distance from the Caernarvon diversion

structure. Numbers refer to the water quality monitoring stations

shown in Fig. 1. Distance was determined as a straight line from the

structure to the respective sampling stations.

Fig. 5. Changes in observed concentrations versus distance from the

Caernarvon diversion of total nitrogen (TN) and total phosphorus

(TP). Distance was determined as a straight line from the structure to

the respective sampling stations. The symbol (*) indicates statistically

significant differences between the paired stations and Mississippi

River water (a ¼ 0:05).

(p! 0:003) mML�1 km�1, respectively. There was alsoa linear response with distance, with a common slopeand intercept for both routes, for the DIN:DIP ratio.Incoming river water and the last paired stations had anaverage DIN:DIP ratio of 107 and 26, respectively(Fig. 7). The DIN:DIP ratio was found to be decreasingwith distance at a rate of 2.1 mML�1 km�1 (p! 0:0002).The reduced model was rejected for the DSi:DIN ratio,but there was a significant linear response with distanceusing the full model (p ¼ 0:007). The different slopes forthe DSi:DIN ratio, which can be interpreted as rates ofchange for the two routes, were found to be0.08 mML�1 km�1 and 0.03 mML�1 km�1 along theBTAB and RAC routes, respectively (p! 0:0028).

Dunnett’s procedure indicated significant differencebetween river water quality and the last three pairs ofstations for TN and DIN, all pairs for TP, and the lastpair of stations for DSi (indicated by asterisks in Figs.5e7; a ¼ 0:05). Dunnett’s procedure did not discerna significant difference for DSi:DIN, but a significantdifference was detected in DIN:DIP ratios of incomingriver water and the means of the last two pairs ofstations.

Student’s t-test indicated significant differences innutrient concentrations of incoming river water and themean of the estuarine stations for TN (p! 0:0005), TP(p! 0:005), DSi (p! 0:048), and DIN (p! 0:043).Student’s t-test did not discern a significant differencefor DSi:DIN, but did find a significant difference inDIN:DIP ratios of incoming river water and the meanof the estuarine stations.

5. Discussion

The results of this study show significant reductionsin nutrient concentrations and changes in stoichiometricnutrient ratios occurred as diverted river water passedthrough the Breton Sound estuary (Figs. 5 and 6). Thisnutrient uptake may not be a permanent loss pathway,if regeneration of nutrients occurs during the summerseason as water temperature rises (Kemp and Boynton,1984). It is, however, likely that a major mechanism forthe loss of nitrate is through denitrification, which is apermanent sink for nitrogen. Other studies in Louisianaand elsewhere have shown high rates of denitrificationwhen river water with high nitrate concentrations flowsinto estuaries (Jenkins and Kemp, 1984; Smith et al.,1985; Boynton et al., 1995; Nixon et al., 1996; Bachandand Horne, 2000; Brock, 2001).

There were high fluxes of nitrate into the sedimentsand high rates of denitrification during the same periodthat we measured decreases in DIN (R. Twilley,personal communication). The subsurface anaerobiclayer in the sediment supports the reduction of NO3

�

to N2 and N2O, which are released into the atmosphere

7R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

Fig. 6. Changes in observed concentrations versus distance from the Caernarvon diversion of dissolved inorganic silicon (DSi), nitrogen (DIN) and

phosphorus (DIP). Distance was determined as a straight line from the structure to the respective sampling stations. The symbol (*) indicates

statistically significant differences between the paired stations and Mississippi River water (a ¼ 0:05).

(Reddy and Patrick, 1984). This is a very importantpathway for the loss of nitrogen from this systembecause approximately 95% of the DIN in diverted riverwater is in the form of nitrate. The alternate floodingand draining of the marsh soil surface resulting fromfluctuating water levels due to tides and frontal passagesensure that diverted water has substantial contact withthe marsh. Rates of denitrification are greater underconditions of fluctuating redox potential (flooding anddraining cycles) than those where the redox is contin-uously high or continuously low, and is an importantmechanism for the oxidation of ammonia to nitrate andsubsequent denitrification (Smith et al., 1983).

Other potential pathways for nitrogen loss includevascular plant and phytoplankton uptake, burial, andreduction to ammonia (Reddy and Patrick, 1984).

Burial is an important loss pathway of material in theLouisiana coastal zone due to subsidence. DeLauneet al. (1981), for example, studied wetlands in BaratariaBay, Louisiana, and found nitrogen buried in the in-terior marsh at a rate of 13.4 g-Nm�2 yr�1, and Smithet al. (1985) found nitrogen burial of up to 23.0 g-Nm�2 yr�1 in wetlands surrounding the AtchafalayaRiver delta.

In contrast to DIN, DIP is usually buffered in estu-arine systems, being taken up when concentrations arehigh and released when they are low (Patrick andKhalid, 1974). There is also a rapid reduction of phos-phorus during flocculation of Fe, Mn, Al, organic car-bon and humic substances as salinity increases from 0 to15e20 psu (Sholkovitz, 1976). Phosphorus is alsoreadily adsorbed onto the surface of sediments and lost

Table 1

Concentrations of dissolved inorganic silicon (DSi), total nitrogen (TN), dissolved inorganic nitrogen (DIN), total phosphorus (TP), dissolved

inorganic phosphorus (DIP) and salinity of incoming river water (station 1, averaged data from March 9 and 15) and estuarine water quality stations

along Bayou Terra aux Boeufs (BTAB) and River aux Chene (RAC; see Fig. 1)

Station Distance (km) DSi (mM) TN (mM) DIN (mM) TP (mM) DIP (mM) Salinity (psu)

(Avg. March 9 and 15)

1 0 117.8 138.9 128.3 5.1 1.2 0

BTAB (March 22)

5 11.9 106.4 122.4 119.3 3.2 1.8 0

6 15.8 95.7 98.8 71.9 3.2 1.5 0

8 23.1 86.6 86.9 65.6 2.1 1.3 0

10 30.4 59.2 67.3 49.8 1.8 1.1 2

11 38.8 29.2 36.1 12.9 1.9 0.3 7

RAC (March 22)

13 11.6 110.2 95.2 106.5 3.6 1.6 0

14 17.0 97.4 99.8 101.7 2.0 1.3 0

15 19.2 66.9 58.8 35.5 1.8 0.7 0

17 26.2 64.9 53.3 21.2 2.5 0.7 1

18 33.1 66.6 63.7 20.9 2.1 1.3 2

8 R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

Fig. 7. Changes in observed molar ratios with distance from the Caernarvon diversion of DSi:DIN and DIN:DIP ratios. Horizontal dashed lines

indicate the Redfield ratio. Distance was determined as a straight line from the structure to the respective sampling stations. The symbol (*) indicates

statistically significant differences between the paired stations and Mississippi River water (a ¼ 0:05).

from the water column as sediments are deposited onbay bottoms and surrounding marshes (Jitts, 1959).Other major mechanisms for the removal of phosphorusfrom the water column are plant uptake, microbialincorporation and soil fixation (Patrick, 1992). Thefixation of phosphorus is more extensive and lessreversible under alternating floodingedraining thanunder either continuously flooded or continuously moistsoil conditions (Patrick, 1992). Alternate flooding anddrying increases the amount of phosphorus in the ferricphosphate and reluctant-soluble occluded fractions atthe expense of the soluble and aluminum phosphatefractions (Patrick, 1992).

Like phosphorus, silica is readily adsorbed onto thesurface of sediments and lost from the water column assediments are deposited (Mayer and Gloss, 1980).Another major mechanism of loss of dissolved siliconis assimilation into diatom tests and eventual sinking(Day et al., 1989). Regeneration of Si is slow relative toboth N and P. Regeneration of biogenic silica isprimarily a chemical phenomenon whereas the regener-ation of N and P is biologically mediated by grazers andbacteria (Conley et al., 1993). During this study, therewas most likely some DSi decreases due to diatomuptake, with little or no benthic regeneration due to lowspring temperatures, and most of the uptake wasprobably from adsorption onto sediments that weredeposited in the estuary.

Most scientists agree that significant decreases in theMississippi River nutrient concentrations would di-minish the severity of the eutrophication in the northernGulf of Mexico (Rabalais et al., 1996). Consistent withthe hypothesis proposed by Officer and Ryther (1980),an increase in the DSi:DIN ratio would likely reduce thepotential for harmful algal blooms (HAB). Nevertheless,while it is probable that the increasing DSi:DIN ratioswould decrease the potential for non-diatom algalblooms, it is also possible that the increasing stoichio-metric Si availability would increase the proportion ofheavily silicified diatoms in the coastal phytoplankton

assemblages (Dortch et al., 2001). It is generally believedthat heavily silicified diatoms are largely responsible forvertical carbon flux during spring and, consequently, forthe formation of hypoxia on the Louisiana shelf (Dortchet al., 2001; Turner, 2001). Most of the harmful algalspecies are non-diatoms. Out of 24 potential HABspecies that were identified in the Louisiana coastalwaters (Dortch et al., 1999), four are diatoms (Pseudo-nitzschia sp.). Pseudo-nitzschia is considered to bea lightly silicified diatom, whose increasing abundanceover time is thought to reflect not only the increasingoverall nutrient concentrations, but also the decreasingDSi:DIN ratios (Parsons and Dortch, 2002). Assumingthat the overall nutrient loading would remain un-changed, an increase in the riverine DSi:DIN ratioswould probably cause a decrease in the abundance ofPseudo-nitzschia, but also an increase in the abundanceof heavily silicified diatoms. This result would likelyincrease vertical carbon flux and exacerbate hypoxia inriver-dominated coastal waters of the northern Gulf ofMexico. Thus, the most favorable management outcomefor any freshwater diversion, as far as the coastaleutrophication problems are concerned, would be toreduce riverine N and P concentrations, while keepingthe nutrient DSi:DIN ratio at or around 1:1. While thishad not been entirely achieved in the described PULSESexperiment, the study strongly suggests that freshwaterdiversions have a potential to significantly reduce theoverall nitrogen loading and alter the stoichiometricratios at which dissolved nutrients are delivered to thecoastal ocean.

Boesch (1996) suggested that diversions of river waterwithin the Mississippi delta may decrease the nitrogenload reaching hypoxic zone offshore, but may lead tohypoxic zones in inshore areas where none presentlyexists. Nutrient enrichment of coastal waters is knownto cause enhanced production of organic material,leading to subsequent decomposition and oxygen de-pletion near the bottom (Cloern, 2001). During the twoyears of intensive study in the Breton Sound estuary,

9R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

however, hypoxia was not observed. Because watercolumn stratification is an important factor affecting thedevelopment of bottom water hypoxia, it is unlikely thathypoxia will develop in a shallow and well-mixed BretonSound estuary. This is consistent with the findings ofMadden et al. (1988) and Day et al. (1992) forFourleague Bay that is affected by the discharge of theAtchafalaya River. Chlorophyll levels during the di-version experiments were elevated only in the regionsurrounding Grand Lake (Fig. 1), and preliminaryanalysis of the phytoplankton community structure,both during high diversion events and later in thesummer, showed a diverse community dominated bydiatoms, without the strong occurrence of potentiallyharmful algal species (I. Ciugulea, personal communi-cation).

6. Conclusions

This study indicates that the Breton Sound estuaryacted as a sink for dissolved silicon and nitrogen, but atdifferent rates, changing the stoichiometric nutrientratio of water passing through the estuary. TheDSi:DIN ratio rose from 0.9 at the Caernarvondiversion to 2.6 at the marsh end member stations,while the DIN:DIP ratio fell from 107 to 26 at the samestations. This study strongly suggests that freshwaterdiversions have a potential to significantly reduce theoverall nitrogen loading and alter the stoichiometricratios at which dissolved nutrients are delivered to thecoastal ocean.

References

Bachand, P.A.M., Horne, A.J., 2000. Denitrification in constructed

free-water surface wetlands: I. Very high nitrate removal rates in

a macrocosm study. Ecological Engineering 14, 9e15.

Boesch, D.F., 1996. Science and management in four U.S. coastal

ecosystems dominated by landeocean interactions. Journal of

Coastal Conservation 2, 103e114.

Boustany, R.G., Croizer, C.R., Rybczyk, J.M., Twilley, R.R., 1997.

Denitrification in a south Louisiana wetland forest receiving

treated sewage effluent. Wetlands Ecology and Management 4,

273e283.

Boynton, W.R., Garber, J.H., Summers, R., Kemp, W.M., 1995.

Inputs, transformations, and transport of nitrogen and phosphorus

in Chesapeake bay and selected tributaries. Estuaries 18, 285e314.

Brock, D.A., 2001. Nitrogen budget for low and high freshwater

inflows, Nueces Estuary, Texas. Estuaries 24, 509e521.Coast 2050: Toward a Sustainable Coastal Louisiana, 1998. Louisiana

Coastal Wetlands Conservation and Restoration Task Force and

the Wetlands Conservation and Restoration Authority. Louisiana

Department of Natural Resources, Baton Rouge, LA, 161 pp.

Cloern, J.E., 2001. Our evolving conceptual model of the coastal eutro-

phication problem. Marine Ecology Progress Series 210, 223e253.

Conley, D.J., Schelske, C.L., Stoermer, E.F., 1993. Modification of the

biogeochemical cycle of silica with eutrophication. Marine Ecology

Progress Series 101, 179e192.

Day, John W., Hall, Charles A.S., Michael Kemp, W., Yanez-

Arancibia, Alejanro, 1989. Estuarine Ecology. Wiley-Interscience,

New York.

Day, John W., Madden, Christopher J., Twilley, Robert R., Shaw,

Richard F., McKee, Brent A., Dagg, Michael J., Childers, Daniel

L., Raynie, Richard C., Rouse, Lawrence J., 1992. Estuarine

Research Federation Conference Proceedings. The Influence of

Atchafalaya River Discharge on Fourleague Bay, Louisiana

(USA). pp. 151e160.Day, J.W., Britsch, L.D., Hawes, S., Shaffer, G., Reed, D.J., Cahoon,

D., 2000. Pattern and process of land loss in the Mississippi Delta:

a spatial and temporal analysis of wetland habitat change.

Estuaries 23, 425e438.DeLaune, R.D., Reddy, C.N., Patrick, W.H., 1981. Accumulation of

plant nutrients and heavy metals through sedimentation processes

and accretion in a Louisiana salt marsh. Estuaries 4, 328e334.Dortch, Q., Parsons, M.L., Rabalais, N.N., Turner, R.E., 1999. What

is the threat of harmful algal blooms in Louisiana coastal waters?

Recent Research in Coastal Louisiana Conference Proceedings,

Lafayette, Louisiana, pp. 134e144.Dortch, Q., Rabalais, N.N., Turner, R.E., Qureshi, N.A., 2001.

Impacts of changing Si/N ratios and phytoplankton species

composition. In: Rabalais, N.N., Turner, R.E. (Eds.), Coastal

Hypoxia: Consequences for Living Resources and Ecosystems.

American Geophysical Union, Washington, DC, pp. 37e48.

Fisher, T.R., Harding, L.W., Stanley, D.S., Ward, L.G., 1988.

Phytoplankton, nutrients, and turbidity in the Chesapeake,

Delaware, and Hudson estuaries. Estuarine, Coastal and Shelf

Science 27, 61e93.

Greenberg,A.E., Trussell, R.R., Clesceri, L.S., Franson,M.A.H. (Eds.),

1985. Standard Methods for the Examination of Water and

Wastewater.AmericanPublicHealthAssociation,Washington,DC.

Humborg, C., Conley, D.J., Rahm, L., Wulff, F., Cociasu, A.,

Ittekkot, V., 2000. Silicon retention in river basins: far-reaching

effects on biogeochemistry and aquatic food webs in coastal marine

environments. Ambio 29, 45e50.

Humborg, C., Ittekot, V., Cociasu, A., von Bodungen, B., 1997. Effect

of Danube River dam on Black Sea biochemistry and ecosystem

structure. Nature 386, 385e388.Jenkins, M.C., Kemp, W.M., 1984. The coupling of nitrification and

denitrification in two estuarine sediments. Limnology and Ocean-

ography 29, 609e619.Jitts, H.R., 1959. The adsorption of phosphate by estuarine bottom

deposits. Australian Journal of Marine and Freshwater Research

10, 7e21.

Justic, D., Rabalais, N.N., Turner, R.E., Dortch, Q., 1995a. Changes

in the nutrient structure of river-dominated coastal waters:

stoichiometric nutrient balance and its consequences. Estuarine,

Coastal and Shelf Science 40, 339e356.

Justic, D., Rabalais, N.N., Turner, R.E., 1995b. Stoichiometric

nutrient balance and origin of coastal eutrophication. Marine

Pollution Bulletin 30, 41e46.

Kadlec, R.H., Knight, R.L., 1996. Treatment Wetlands, Lewis

Publishers, New York.

Kemp, W.M., Boynton, W.R., 1984. Spatial and temporal coupling of

nutrient inputs to estuarine primary production: the role of

particulate transport and decomposition. Bulletin of Marine

Science 35, 522e535.

Lane, R.R., Day, J.W., Thibodeaux, B., 1999. Water quality analysis

of a freshwater diversion at Caernarvon, Louisiana. Estuaries 22,

327e336.

Lane, R.R., Day, J.W., Marx, B., Reyes, E., Kemp, G.P., 2002.

Seasonal and spatial water quality changes in the outfall plume of

the Atchafalaya River, Louisiana, USA. Estuaries 25, 30e42.

10 R.R. Lane et al. / Estuarine, Coastal and Shelf Science 60 (2004) 1e10

Madden, Christopher J., Day, John W., Randall, John M., 1988.

Freshwater and marine coupling in estuaries of the Mississippi

River deltaic plain. Limnology and Oceanography 33, 982e1004.

Mayer, L.M., Gloss, S.P., 1980. Buffering of silica and phosphate in

a turbid river. Limnology and Oceanography 25, 12e22.

Mitsch, W.J., Day, J.W., Gilliam, J.W., Groffman, P.M., Hey, D.L.,

Randall, G.W., Wang, N., 2001. Reducing nitrogen loading to the

Gulf of Mexico from the Mississippi River basin: strategies to

counter a persistent ecological problem. BioScience 51, 373e388.

Nixon, S.W., Ammerman, J.W., Atkinson, L.P., Berounsky, V.M.,

Billen, G., Boicourt, W.C., Boynton, W.R., Church, T.M., Ditoro,

D.M., Elmgren, R., Garber, J.H., Giblin, A.E., Jahnke, R.A.,

Owens, N.J.P., Pilson, M.E.Q., Seitzinger, S.P., 1996. The fate of

nitrogen and phosphorus at the landesea margin of the North

Atlantic Ocean. Biogeochemistry 35, 141e180.Officer, C.B., Ryther, J.H., 1980. The possible importance of silicon in

marine eutrophication. Marine Ecology Progress Series 3, 83e91.

Parsons, M.L., Dortch, Q., 2002. Sedimentological evidence of an

increase in Pseudo-nitzschia (Bacillariophyceae) abundance in

response to coastal eutrophication. Limnology and Oceanography

47, 551e558.

Patrick, W.H., Khalid, R.A., 1974. Phosphate release and sorption by

soils and sediments: effect of aerobic and anaerobic conditions.

Science 186, 53e55.

Patrick, W.H., 1992. Phosphorus biogeochemistry of wetlands.

Phosphorus, Life and Environment, 199e205.

Perez, B.C., Day, J.W., Rouse, L.J., Shaw, R.F., Wang, M., 2000.

Influence of Atchafalaya River discharge and winter frontal

passage and flux in Four League Bay, Louisiana. Estuarine,

Coastal and Shelf Science 50, 271e290.Rabalais, N.N., Turner, R.E., Justic, D., Dortch, Q., Wiseman, W.J.,

Gupta, B.K.S., 1996. Nutrient changes in the Mississippi River and

system responses on the adjacent continental shelf. Estuaries 19,

386e407.Rabalais, N.N., Turner, R.E., 2001. Hypoxia in the northern Gulf of

Mexico: description, causes and change. In: Rabalais, N.N.,

Turner, R.E. (Eds.), Coastal Hypoxia: Consequences for Living

Resources and Ecosystems. American Geophysical Union,

Washington, DC, pp. 1e36.

Reddy, K.R., Patrick, W.H., 1984. Nitrogen transformation and loss

in flooded soils and sediments. Critical Reviews in Environmental

Control 13, 273e309.

Redfield, A.C., 1958. The biological control of chemical factors in the

environment. American Scientist 46, 205e222.

Richardson, C.J., Nichols, D.S., 1985. Ecological analysis of

wastewater management criteria in wetland ecosystems. In:

Kaynor, E.R., Godfrey, P.J., Pelczarski, S. (Eds.), Ecological

Considerations in Wetlands Treatment of Municipal Wastewaters.

Van Nostrand Reinhold Company, New York, pp. 351e391.

Sholkovitz, E.R., 1976. Flocculation of dissolved organic and in-

organic matter during the mixing of river water and seawater.

Geochimica et Cosmochimica Acta 40, 831e845.

Smith, C.J., DeLaune, R.D., Patrick, W.H., 1983. Nitrous oxide

emission from Gulf Coast Wetlands. Geochimica et Cosmochimica

Acta 47, 1805e1814.

Smith, C.J., DeLaune, R.D., Patrick, W.H., 1985. Fate of riverine

nitrate entering an estuary: I. Denitrification and nitrogen burial.

Estuaries 8, 15e21.

Spieles, D.J., Mitsch, W.J., 2000. The effects of season and hydrologic

and chemical loading on nitrate retention in constructed wetlands:

a comparison of low- and high-nutrient riverine systems. Ecolog-

ical Engineering 14, 77e91.

Turner, R.E., Rabalais, N.N., 1991. Changes in Mississippi River

water quality this century. BioScience 41, 140e147.

Turner, R.E., 2001. Some effects of eutrophication on pelagic and

demersal marine food webs. In: Rabalais, N.N., Turner, R.E.

(Eds.), Coastal Hypoxia: Consequences for Living Resources and

Ecosystems. American Geophysical Union, Washington, DC, pp.

1e36.

Valderrama, J.C., 1981. The simultaneous analysis of total nitrogen

and total phosphorus in natural waters. Marine Chemistry 10,

109e122.

Sources of unpublished materials

Ciugulea, I. Department of Botany and Microbiology, University of

Bucharest, Romania, personal communication.

Gregg, S. Coastal Ecology Institute School of the Coast and the

Environment, Louisiana State University, Baton Rouge, LA 70803,

USA, personal communication.

Twilley, R. University of Louisiana at Lafayette, Lafayette, LA 70504,

USA, personal communication.