Estuaries VoL 22, No. 2A, p. 206-214 June 1999

Benthic Biogeochemistry Beneath the Mississippi River Plume

JOHN W. MORSE 1

GILBERT T. ROWE

Department of Oceanography Texas A & M University College Station, Texas 77843-3146

ABSTRACT: Biogeochemicai processes occurring near the sedlment-water interface of shallow ( -20 m) water sedi- ments lying beneath the Mississippi River p lume on the Louisiana shel f were studied using benthic chambers and sedi- ment cores. Three sites were chosen with distinctly different characteristics. One was overlain by oxic water where aerobic respiration dominated organic matter remineralization. T h e second site was overlain by oxic water but organic matter reminerafization was dominated by sulfate reduction. The third site was overlain by hypoxic water a n d aerobic remin- eralizatlon was of minor significance. Major di f ferences were observed in the fluxes o f CO2(17-56 mmol m e d q), O~(2- 56 mmol m 2 d-i) and nutrients (e.g., NH4 +, 2.6-4.2 mmol m -~ d "a) across the sediment-water interface, and the relative importance of different e lectron acceptors, even though the sites were in close proximity and at nearly the same water depth. Large variations in the eff iciency of organic-C burial (3%-51%) were also calculated based on a simplified mode l o f the relationships between the fraction of organic matter remineralized by sulfate reduction and the fraction of sulfide produced that is buried as pyrite. These observations demonstrate the high degree of spatial helerogenei ty o f benthic biogeochemistry in this important near-deltaic environment.

Introduction

The inner Louisiana shelf is strongly influenced by the westward flowing Mississippi River plume and has been a region of numerous investigations and a focus of a dedicated issue of this journal (Atwood et al. 1994). Our previous work in this area (Lin and Morse 1991; Miller-Way et al. 1994; Morse and Berner 1995) led us to make a more detailed investigation of the benthic biogeochem- istry of this dynamic region. This was done in or- der to better understand the relative importance different benthic processes that play a major role in contributing to summer hypoxic bottom water conditions in this region (Dortch et al. 1994; Mill- er-Way et al. 1994; Justi~ et al. 1996)

The hypoxic conditions can also have a major influence on benthic biogeochemical processes by eliminating aerobic organisms (e.g., Blackwelder et al. 1996; Gupta et al. 1996) and producing an en- vironment in which heterotrophic activity is almost entirely associated with sulfate reducing bacteria. Under hypoxic conditions, the dominant mode of t ranspor t of dissolved components may also change from bioirrigation to molecular diffusion, significantly (perhaps by a factor of about 10) re- ducing benthic fluxes.

In order to investigate the potential influences of seasonal benthic hypoxia, sediments were stud- ied in the area immediately to the west of the Mis- sissippi River delta. Fluxes between the bottom wa-

I Corresponding author; tele: 409/845-9630; fax: 409/84.6- 9631; e-mail: morse@astra.tamu.edu.

ter and the sediments were measured with benthic chambers and sediment geochemical properties were determined under both hypoxic and non- hypoxic conditions. The results are modeled at the end of this paper using the approach of Morse and Berner (1995). Their relatively simple method makes it possible to approximately establish the fraction of organic matter arriving at the sediment- water interface that is oxidized, the relative impor- tance oxic respiration versus sulfate reduction, and the fraction of reduced sulfur that escapes oxida- tion and is buried in the sediment.

Study Area

The area chosen for study is on the Louisiana shelf in an area immediately to the west of the Mis- sissippi River Delta (Fig. 1). Sampling site infor- mation is summarized in Table 1. This area was chosen because a wide range of sediment burial rates was desired where changes in other environ- mental parameters, such as water depth, were rel- atively small. Lin and Morse (1991) worked on how the carbon to sulfur relationship in sediments is affected by sulfate reduction rates in this area. This provided a good database for site selection and sampling. It should be noted that this is the major depositional area for Mississippi River sediments. About half of all reduced sulfur accumulation in Gulf of Mexico sediments takes place within or ad- jacent to our sampling area (Lin and Morse 1991).

Sediment sampling and benthic chamber de- ployments took place on two cruises using the R/ V Gyre. The first was under early spring conditions

�9 1999 Estuarine Research Federation 206

90~ 89~ 30ON 30~

Benthic Biogeochemistn/Beneath a River Plume 207

TABLE 1. Study sites. S is salinity on the practical salinity scale. S and T are near-bot tom water values.

Depth Cruise Site Lat i tude Lxmgdtude (m) S T (~

April 92 l a 28~ ' 89~ ' 13 34.0 19.6 2a 29007.2 ' 89~ ' 18 36.0 20.0

August 94 l b 29000.0 ' 89~ ' 13 30.3 28.1 2b 29~ ' 89044.4 ' 19 32.1 27.2 3 29006.5 ' 8!)~ ' 12 27.9 28.7

29ON 29~

Fig. 1. er delta.

90oW 89~

Location of sampl ing sites west of the Mississippi Riv-

in April, 1992, and the second in m i d - s u m m e r o f August, 1994. Dur ing the second cruise, the two sites f rom the first cruise were re-occupied. A site where hypoxic b o t t o m water condi t ions were pres- ent, as is typical du r ing s u m m e r in this area of the Louis iana shelf (Atwood et al. 1994; Gupta et al. 1996), was also studied. O u r observat ions were m a d e in con junc t ion with the Nat ional Oceanic a n d A t m o s p h e r i c A d m i n i s t r a t i o n N u t r i e n t En- h a n c e d Coastal Ocean Processes (NECOP) Pro- gram, the results o f which provided addi t ional in- f o r m a t i o n on this region dur ing the study.

Methods

BENTHIC CHAMBER FI.UX MF.ASUREMENTS

Fluxes across the sediment-water in terface were measu red in situ by incubat ing b o t t o m water over small areas of the sea f loor in a pair o f diver-de- ployed ben th ic chambers . Changes in the concen- t ra t ion of chemical variables in the chamber s over t ime were used to calculate exchanges of dissolved c o m p o u n d s be tween the water and the sediments , assuming that rates in the b o t t o m water are e i ther negligible or are accoun ted for in controls. The mos t r ecen t descr ipt ions of ou r c h a m b e r system also include their use as pa r t o f a r e m o t e ben th ic l ander for deep-water studies (Rowe et al. 1994). A compar i son of in situ rates m a d e with these c h a m b e r s and a l abora to ry core chemos ta t incu- ba t ion system (Miller-Way et al. 1994), and a com- par ison o f ben th ic respirat ion with that in the wa- ter co lumn in the same area o f study (Dor tch et al. 1994) have also been publ ished.

T h e c h a m b e r s were lowered to the sea f loor in an a l u m i n u m frame, which was a t t ached to a mark-

er buoy on the surface. Tile f r ame holds the elec- t ronic c o m p o n e n t s in a water t ight glass sphere and a pair o f clear and opaque chamber s is a t tached to the glass sphere on the f r ame by an electrical wire umbilicus. In this p a p e r we r epo r t only the results f r o m the opaque chamber s which exclude benth ic photosynthet ic processes.

On the seafloor, the pair o f chamber s (produc- ing 2 m e a s u r e m e n t s per site) is re leased and lifted out o f the f r ame by a diver and then gendy placed on an und i s tu rbed area of sediment . One-way valves on the top of each c h a m b e r allow water to escape when the diver pushes the s h a r p e n e d edge o f the cylindrical chamber s into the sediment . Pen- etrat ion cont inues up to a f lange that s tandardizes pene t ra t ion-cont ro l l ing c h a m b e r volume. Cham- ber area is 900 cm ~ with a vo lume of 7 1. Each is equ ipped with a small st irring motor , which moves water past the e lect rodes and prevents stratifica- tion. The effective diffusive b o u n d a r y layer thick- ness in these chambers is on the o rde r of 0.5 to 1.5 m m (Rowe et al. 1994), which approx ima tes b o u n d a r y layer condi t ions dur ing earl ier ben th ic l ander studies using a d e p o c e n t e r on the cont inen- tal slope. T h e d e p l o y m e n t of the chamber s for pe- riods of 3 to 12 hours was fbund to be appropr ia te , at the flux rates found in our study area.

Syringe samples were rout inely taken f rom each c h a m b e r at intervals o f 1 to 4 hours for measure- m e n t of several dissolved componen t s , including and total dissolved inorganic ca rbon (DIC or ~CO2). The syringes were no t poisoned, bu t the water was fi l tered th rough 0.2 Ixm filters to p reven t en t rance of all bu t the smallest bacter ia into the syringes.

Nutr ients were d e t e r m i n e d at sea soon af ter sam- ple collect ion on a Techn icon autoanalyzer by stan- da rd methods . Oxygen was d e t e r m i n e d by Winkler t i trat ion and also m o n i t o r e d within the chambe r s using a YSI oxygen e lect rode, with the data s tored in a data logger. A cou lomet r i c DIC ti tration system was used on samples taken f r o m our chambers . T h e m e a s u r e m e n t s of CO., fluxes using the cham- bers had analytic e r ro r o f -+0.2% on 20 ml cham- ber samples.

2 0 8 J. Morse and G. Rowe

SEDIMENT GEOCHEMISTRY

Sediment samples were collected by box coring. This was typically done down to about the 30 cm depth. Deeper sed iment samples were obta ined by gravity coring, without a core catcher to avoid dis- turbing the sediments, for depths usually down to about 1 m. At the time of collection cores were closely examined to ensure that the sediment-water interface had no t been substantially disturbed. Subcores were taken immediately f rom the box cores. Cores for shore- laboratory analyses o f solids were sectioned, immediately frozen and kept fro- zen until t ime of analyses. Plexiglas incubat ion chambers , 5 cm long with a volume of about 25 ml, were directly inserted into cores to obtain sam- ples for de te rmina t ion of sulfate reduc t ion rates. They were immediately capped and sealed with electrical tape, and then placed in a water bath at the t empera tu re of the bo t tom water and kept in the dark.

Cores f rom which pore water was extracted were e i ther sect ioned and water extracted by centrifu- gation for non-O 2 sensitive componen t s (e.g., CI-, SO42-, when n o H 2 S present) or placed in a glove bag unde r a high purity Nz a tmosphere where they were sect ioned and then the mud was loaded into Reeburgh (1967)-type sediment squeezers. The pore water f rom the sediment squeezers was passed th rough a 0.45 p~m Nuclepore ~ filter and then into a syringe where it was s tored in the dark and refr igerated until the time of its analysis. Subsam- ples were taken and acidified for later Fe and Mn analysis.

Most solid sediment samples were analyzed for porosity by drying for 24 h at l l0~ grain size weight fraction less than 63 Izm by wet sieving, and weight pe rcen t organic-C and inorganic-C using a LECO carbon analyzer. Acid volatile sulfides (AVS) were de t e rmined by the HCI + SnCI,, m e t h o d of Cornwell and Morse (1987), and TRS was mea- sured using the boiling Cr (II) + acid m e t h o d of Canfield et al. (1986). Pore water nutr ients and ~CO2 were de t e rmined as previously descr ibed for benthic chamber water analyses and H,,S was mea- sured using the Cline (1969) method , but was al- ways below our de tec t ion limit of ~1 ~M.

Sulfate reduc t ion rates were de t e rmined as de- scribed by Lin and Morse (1991). Briefly, ~1 p~Cu of 35S042- was injected into a 25 ml incubat ion chamber conta in ing the sediment sample. Th e chamber was then incubated at in situ bo t tom wa- ter t empera tu re for 24 h and then fast f rozen for later shore laboratory analysis. The 35S concentra- tion in the AVS and TRS fractions was de t e rmined and compared with the ~5SO4" and 8042- concen-

TABLE 2. Fluxes measu red in ben th ic chambers . Fluxes are in mmol m 2 d-L Positive fluxes are out o f sed iments and negative fluxes are into sediments . Note that at Site 3 amb i en t O 2 was very low at only 4 ~M whereas at "all o the r sites it was > 110 o~M.

Site Oz CO s NH 4 PO 4

l a --23 35 3.7 0.26 l b - 1 9 53 3.9 0.26 2a - 2 4 17 3.0 0.024 2b - 5 6 56 4.2 -0 .41 3 - 1 . 9 37 2.6 1.2

trations in the sample to obtain the sulfate reduc- tion rate.

In cores f rom the second cruise, 137Cs was deter- mined by Dr. Peter Santschi (Texas A&M Univer- sity at Galveston) in o rde r to estimate max imum possible sediment accumulat ion rates, organic mat- ter and C / N ratios were d e t e rm in ed by Dr. Luis Cifuentes (Texas A&M University).

Results and Discussion

BENTHIC FLUXES

Benthic flux measurements are summarized in Table 2, where negative values indicate a flux f rom the benthic chamber into the sediment. The flux of O,~ into the sediments (sediment oxygen de- m an d = SOD) varied by about a factor of 30 ( - 1 . 9 to - 5 6 mmol m -2 d- l ) . Th e largest 02 flux oc- cu r red in August at Site 2 and was close to double that observed in the spring at this site. The water temperature was about 7~ warmer in August than in April and this di f ference in SOD may be due to the increase in tempera ture , a l though spatial vari- ability could also contr ibute to observed differenc- es. SOD at Site 1 was about 17% less in August than in April. Since the t empera tu re was 8.5~ h igher in August it was probable that the decrease in SOD was largely the result of a decrease in readily me- tabolizable sedimentary organic matter. This was evidenced by the observat ion that the average wt.% organic-C in the u p p e r 10 cm of sediment de- creased by about a factor of 4, f rom 1.04 to 0.24 wt.%. Site 3 had hypoxic bo t tom water and, be- cause of the very low 02 concent ra t ion in the bot- tom water, SOD was also low co m p ared to the oth- er sites.

ECO2 fluxes were also measured. They are plot- ted against SOD in Fig 2. A flux of CO 2 flux out of and 02 into the sediment should be about equal, if oxic respirat ion was the dominan t fo rm of he te ro t roph ic activity (i.e., CH20 + 02 ---> H 2 0 + CO2) or reoxidat ion of r educed componen t s was comple te with little o ther biologic activity. Carbon- ate mineral dissolution and precipi tat ion reactions and biologic fixation of organic carbon could po-

Benthic Biogeochemistry Beneath a River Plume 2 0 9

6 0 . . . . I . . . . I . . . . I ' * r r [ t r r I . . . . ,

l b S 0

"O o l

'E 4 0 3

, 0

x

�9 2 o" 0

I

0 - 1 0 - 2 0 - 3 0 - 4 0 - 5 0 - 6 0

0 2 Flux (mmol m" = d" I)

Fig. 2. Rela t ionship be tween dissolved O~ f luxes into sedi- m e n t s and CO2 fluxes ou t o f sediments . N u m b e r s are for sam- pl ing sites. T h e heavy line is for an ideal 1:1 re la t ionship and the light l ines are for a ---20% uncertainty.

tentially also inf luence CO 2 to 09 ratios. The CO2/ O,, flux at Site 2b, which had by far the highest SOD, fell close to - 1 . At Site 2a, the CO2/O2 flux ratio was slightly less ( -0 .71 ) than that o f ou r es- t imated range of uncer ta in ty of abou t -+ 14%. Sites la ,b and 3 had substantially larger CO2/O 2 flux ratios o f - 1.5, - 2 . 8 , and - 19.5, respectively, which probably indicate a p r edominance of anaerobic processes at these sites relative to Site 2 where oxic processes dominate , along with a significant burial of r educed componen t s (e.g., pyrite) relative to bi- ologic remineral izat ion.

If remineral izat ion o f sed imentary organic mat- ter were to p roceed with a s toichiometry similar to that of the Redfield C : N : P ratio, it would be ex- pec ted that the absolute value of NH4 (in the ab- sence of processes such as nitrification) and P O 4

flux ratios to 02 and CO2 would be about 0.15 and 0.0094 respectively, and the NH4 to PO4 flux ratio would be close to 16. Calculated values of these ratios are given in Table 3. Only about a quar te r (6 of 25) of the ratios fall within about -+20% of the Redfield ratio. This observat ion reflects the complexi ty of the processes control l ing the fluxes o f dissolved componen t s associated with organic mat ter remineral izat ion. It also suggests nitrifica- t i o n - d e n i t r i f i c a t i o n o c c u r wi th in surf ic ia l sedi- ments. This observat ion implies too that ne t fluxes of N and P are difficult to in te rpre t unambiguous- ly. The use o f net N and P fluxes to discern rela- tionships between O,~ and CO2 fluxes is limited.

TABLE 3. Values o f N H 4 and PO4 flux ratios to the O~ and CO 2 flux ratios, a n d NH 4 to PO4 flux ratios. T h e row RR con- tains values tha t would be expec ted f rom the Redfield Ratio of organic mat te r componen t s . Values in italics are within abou t 20% of the expec ted value based on the Redfield Ratio.

S i t e N H 4 / O z N H 4 / C O z P O J O 2 P O 4 / C O 2 N H J P O ~

RR - 0 . 1 5 0.15 - 0 . 0 0 9 4 0.0094 16

l a -0 ,16 0,11 - 0 , 0 l l 0,007 14 l b - 0 . 21 0.74 - 0 , 0 1 3 0.005 15 2a -0 .13 0.18 -0 .001 0.001 125 2b - 0 . 0 8 0.08 0.007 - 0 . 0 0 7 - 1 0 3 - 1 . 3 7 0.07 - 0 . 6 3 7 0.032 3

SEDIMENT Ct lEMISTRY

Analytical results for solid phase sediment com- ponents , within approximately the top mete r of sediment , are summarized in Table 4. Site 1 was near the mou th of Southwest Pass and sediment accumulat ion in this area is rapid. Values of sedi- menta t ion rates should be taken with caution in the study area, as they are likely to vary bo th sea- sonally and annually in association with changes of input f rom the Mississippi River. However, esti- mates of sedimentat ion rates do have utility when trying to de te rmine possible influences of sedi- menta t ion rate when o rde r of magni tude changes occur. The ~37Cs distr ibution indicates the sedi- menta t ion rate at Site 1 is grea ter than 2 cm y-~. This is in good ag reemen t with z~~ sedi- menta t ion rates at nearby sites of 2.59 and 2.21 cm y-1 (Dr. Bren t McKee, personal communica t ion , 1995). T h e sediments at Site l a were character ized by a decrease in porosity with dep th f rom 0.61 to 0.46, being dominant ly f ine-grained (> 0.93, < 63 txm), and having a low carbonate-C con ten t (0.23 to 0.54 wt.%). Organic-C decreased f rom slightly over 1 wt.% near the sediment-water interface to about 0.6 wt.% at close to 1 m depth. It should be no ted that even though the study area is close to the Mississippi River delta that earlier work here (Gearing et al. 1977) has demons t ra ted that the organic mat ter is p redominant ly o f mar ine origin as evidenced by ~13C values relative to the PDB standard, o f about - 2 1 to 22%0. TRS increases by about a factor of 3 f rom 76 to 229 ~mol g-i and AVS showed a typical max imum near the sediment- water interface where it comprised 28% of TRS and then decl ined with depth. Sediment charac- teristics de t e rmined on a core obta ined dur ing a reoccupa t ion o f Site 1 were generally very similar to those obta ined on the first core, except porosity was slightly h igher and carbonate-C which was slightly lower, perhaps reflect ing enhanced biolog- ic activity at this site dur ing the summer or small scale spatial variability (site l a and l b are about 0.1 nautical miles apart) .

210 J. Morse and G. Rowe

TABLE 4. Data on s ed i men t solids. Porosity is vo lume fraction and fraction < 63 p.m is weight fraction. Carb-C = carbonate carbon; org-c = organic carbon; TRS = total r educed inorganic sulfide; and AVS = acid volatile sulfides.

Depth Fraction Carb-C O~rg-4-C TRS AVS Site (cm) Porosity < 63 o.m (wt%) (wt%) (p.mol/g) (p.mol/g)

l a 0-1 0.61 0.97 0.40 1.04 76 12 1-2 0.61 0.95 0.50 1.08 79 18 2 -4 0.61 0.96 0.54 1.06 97 27 4 -6 0.58 0.97 0.49 1,00 96 9 8-10 0.57 0.96 0.34 1.00 115 4

12-14 0.56 0.95 0.53 0.80 134 4 16--18 0.57 0.96 0.39 0.89 129 4 20-22 0.55 0.94 0.31 0.93 125 3 26--28 0.55 0.95 0.23 1.05 128 4 32-34 0.51 0.95 0.39 0.77 118 2 37-39 0.48 0.95 0.34 0.70 136 2 47-49 0.50 0.95 0.38 0.65 163 0 57-59 0.50 0,96 0.34 0.67 177 1 67-69 0.50 0.93 0.37 0.67 172 1 77-79 0.49 0.96 0.33 0,69 273 2 87-89 0.46 0.93 0.41 0.58 229 1

lb 0-1 0.85 0.96 0.31 1.15 79 3 1-2 0.81 0.92 0.28 1.06 79 7 2--4 0.79 0.91 0.23 1.03 80 12 4 -6 0.78 0.93 0.20 1.06 104 18 8-- 10 0.78 0.98 0.18 1.07 108 6

12-14 0.79 0.99 0.23 1.08 132 10 16-18 0.85 0.97 0.23 1.16 99 12 20-22 0.75 0.98 0.25 0.91 136 5 26-28 0.74 0.98 0.27 0.86 138 1 32-34 0.72 0.96 0.25 0.80 124 4 35-37 0.73 0.98 0.26 0.83 138 3 45--47 0.69 0.96 0.30 0.64 125 2 55-57 0.68 0.97 0.29 0.63 117 2 65-67 0.71 0.97 0.26 0.65 115 1 75-75 0.73 0.96 0.23 0.62 100 1 85-87 0.68 0.77 0.39 0.60 175 0

2a 0-1 0.64 0.95 0.43 1.12 119 6 1-2 0.67 0.96 0.36 1.06 146 6 2 -4 0.57 0.95 0.34 0,99 114 5 4 -6 0.56 0.96 0.26 1.02 119 2 8-10 0.52 0.96 0.36 0.86 158 1

12-14 0.49 0.96 0.44 0.68 124 1 16-18 0.45 0.95 0.26 0.59 106 1 20-22 0.52 0.94 0.54 0.65 82 1 26-28 0.52 0.94 0.33 0.68 102 2 32-34 0.54 0.93 0.35 0.65 90 3 47-49 0.51 0.95 0.28 0.65 96 1 57-59 0.54 0.96 0.35 0.69 91 1 67-69 0.54 0.95 0.27 0,74 83 3 77-79 0.55 0.95 0.33 0.72 106 1 87-89 0.54 0.95 0.33 0.69 199 1 97-99 0.55 0.94 0.27 0.70 122 1

107-109 0.55 0.95 0.28 0.68 138 2

2b 0-1 0.79 0.87 0.60 0.82 158 3 1-2 0.77 0.87 0.46 0,76 136 3 2-4 0.71 0.90 0.40 0,55 82 0 4 -6 0.72 0.96 0.35 0,55 80 0 8-10 0.73 0.99 0.40 0.54 69 0

12-14 0.75 0.99 0.34 0.53 79 0 16-18 0.74 0.99 0.35 0.49 83 0 21}-22 0.73 0.99 0.26 0.54 74 0 26--28 0.71 0.99 0.44 0.48 77 0 32-34 0.75 0,99 0.36 0.55 92 0 35-37 0.74 0,99 0.38 0.54 79 0 4.5-47 0.74 1.00 0.30 0.60 73 0 55-57 0.61 0.99 0.40 0.34 92 0 65--67 0.71 1.00 0.38 0.59 79 0 75-75 [).77 1.00 0.40 0.59 102 0 85-87 0.79 1.00 0.36 0.53 102 0 95-97 0.77 1.00 0.38 0.56 94 0

110-112 0.79 1.00 0.38 0.57 103 0

Benthic Bioge~hemistnJ Beneath a River Plume 21"~

TABLE 4. Cont inued .

Depth Fraction Carb-C Org-C "I'RS AVS Site (cm) Porosity < 63 p,m (wt%) (wt%) (~rnol/g) (p, mol/g)

3 0--1 0.73 0_49 0.39 0.51 88 2 1--2 0.62 0,45 0.37 0.44 100 1 2--4 0.60 0_55 0.58 0.44 94 0 4--6 0.57 0_60 0.49 0.44 95 0 8-10 0.56 0.74 0.42 0.51 144 0

12-14 0.57 0.78 0.59 0,52 182 0 16-18 0.59 0.78 0.38 0,51 169 0 20-22 0.56 0_81 0.41 0.40 81 0 26-28 0.59 0_92 0,40 0.43 122 0 32-34 0.61 0_97 0.42 0.52 68 1 35-37 0.57 0_95 0.45 0.72 84 1 45-47 0.63 0_96 0.35 0A4 57 1 55-57 0.51 0.98 0.43 0,28 100 0 65-67 0.67 0_99 0.26 0.58 295 0 75--75 0.67 1.00 0.53 0.62 83 3 85-87 0.72 1_00 0.43 0.61 99 4 95-97 0.68 1,00 0.45 0.52 50 1

T h e sed imenta t ion rate at Site 2, based on ~'~TCs, was approx imate ly 0.25 cm y-1 210pb accumula t ion rates at nea rby locations exhibi ted a wide range of accumula t ion rates (0.02 to 0.54 cm y z, B. McKee personal commun ica t i on ) that b racke t the rate de- t e rmined at ou r sampl ing site. T h e porosity, grain size, and ca rbona te and organic-C at Site 2a were remarkab ly similar to those for Site 1 a and b. How- ever, TRS did no t exhibi t any clear increase with dep th and AVS t ended to be low t h r o u g h o u t the

A o I 2b

=.

4~ / .'.'1

o t 'it

0 10 20 30 40

Integrated SO 4 Reduction Rate (mmol m -= d "t)

Fig. 3. Rela t ionship be tween the in tegra ted sulfate reduc- tion rate and dissolved 02 fluxes (solid circles) into sed iments and COe fluxes (open triangles) out o f sediments . N u m b e r s are for sampl ing sites T h e solid line is where da ta would fall ac- cord ing to the react ions discussed in the text and the do t ted lines connec t the O~ and CO,, f luxes for the same station (note that they are the same value for station 2b).

core, always compr is ing less than 5% of TRS. As at Site 1, the core ob ta ined in August (2b) had a slightly h igher porosity. TRS and AVS exhib i ted similar t rends in April and August, bu t were slight- ly lower at Site 2b. A significant d i f ference between 2b and 2a was the lower organic-C con ten t in 2b (2b average for top 10 cm = 0.64 wt.% whereas 2a had average of 1.01 wt.% over same dep th inter- val).

The aWCs es t imated accumula t ion rate at Site 3 is < 0.2 cm y- a. Unfortunately, the closest location at which a zl~ accumula t ion rate is avail- able is abou t 8 km away and is the re fo re no t useful for compar i son . The porosity and carbonate-C con- tent o f sediments at Site 3 were similar to those of Site 2b, which was visited at the same t ime. Grain size exhibi ted a ma jo r coarsen ing in the u p p e r 20 cm, d r o p p i n g to only 49% in the < 63 p,m fract ion at the top of the core, indicat ing that sed iment deposi t ion at this site was no t in steady-state. Or- ganic-C exhibi ted no major t rends with dep th (e.g., organic-C was 0.51 wt.% at the core top and 0.52 wt.% at the b o t t o m o f the core) . TRS had a very i r regular distr ibution with depth; AVS also showed no clear variat ion and was always less than 5% of TRS. The average C / S ratio for sediments d e e p e r than 50 cm was 1.30.

In t eg ra t ed sulfate reduc t ion rates for approxi- mately the top m e t e r of sediment , in m m o l m -2 d 1, were: l a = 15.1; l b = 38.6; 2a = 13.7; 2b = 14.3 and 3 = 16.0. O.~ and COt flux rates have been plo t ted against in tegra ted sulfate reduc t ion rates in Fig. 3. I f the genera l react ions for organic mat te r oxidat ion via sulfate reduct ion

2 C H 2 0 + S O 4 2 - ----> 2 C O 2 + S 2- + 2 H 2 0 ( 1 )

and subsequen t oxidat ion of sulfide by oxygen

212 J. Morse and G. Rowe

2Oz + S ~- --) SO4"- (2 )

were to p redomina te in these sediments, then both the 0 2 and CO 2 fluxes would be expec ted to be close to twice the integrated sulfate reduc t ion rate. Clearly this was no t the case for Site 2b, which ap- peared to be domina ted by non-sulfate respiration, and Site 3 where there was little 02 in the overlying water to react with sulfide.

MODEL INTERPRETATION OF OBSERVATIONS

The results demons t ra te that major variations in time and space occur in benthic biogeochemistry, even though this investigation was restricted to a relatively small geographic area (< 20 n.mi.) and narrow dep th range (12 to 19 m). These variations reflect the complex na ture of in ter re la ted process- es over this area of the Louisiana shelf immediately west of the Mississippi River delta (e.g., see previ- ously re fe renced issue o f Estuaries [Atwood et al. 1994]). Clearly data f rom three study sites, two of which were sampled twice, are no t sufficient to re- solve the complex behavior of this system. Because the data f rom the three sites are substantially dif- ferent , they can be used, with considerable cau- tion, in an inverse analysis to get a bet ter under- s tanding o f the dominan t metabolic processes in the sediments in such a spatially patchy area. Our analysis will also fu r the r ou r unders tanding of why the carbon-sulfur relationships in these sediments differ substantially f rom general t rends in more normal environments .

Morse and Be rne r (1995) p resen ted a mode l of processes that control sedimentary organic-C to re- duced-S burial (C/S) ratios in normal mar ine sed- iments in which Fe does no t limit pyrite format ion. The major componen t s of their model have utility for unders tand ing how the relative impor tance of major b iogeochemical processes changes at ou r study sites. C /S ratios in the model d e p e n d on three fundamenta l factors control l ing the burial of organic-C and reduced-S. They are:

1) the fract ion of organic-C reaching the sedi- me n t that is metabol ized (fM);

2) the fract ion of organic-C that is oxidized via sulfate reduc t ion (fcs); and

3) the fract ion of hydrogen sulfide p ro d u ced f rom sulfate reduct ion that is bur ied as re- duced-S primarily in pyrite (fsv).

These factors are re la ted as given in equat ion 1 where R is the C /S molar burial ratio. It is ar- ranged to calculate the fraction of organic mat ter (f,~) that undergoes he te ro t roph ic metabolism.

2 fM = (~)

RfcsfsP + 2

It is impor tan t to no te at this poin t that later we shall be discussing large ext remes in fM values. Be- fore discussing these model results we first present estimates of the degree to which modera te uncer- tainties in o ther model parameters inf luence our calculated values.

Sites 1 and 2 would fall into the category of nor- mal mar ine sediments (see Morse and Be rne r 1995, for discussion), bu t the seasonally hypoxic condi t ions at Site 3 raise a quest ion as to the ap- plicability of the model , at least dur ing par t o f the summer. Canfield (1994), however, has suggested that for sedimenta t ion rates > 0.03 cm y-~, bo t tom water 02 con ten t has little effect on organic-C pres- ervation. Th e lack of detectable (< 1 ~M) H~S and modera te dissolved Fe concentra t ions (typically 50 to 100 ~zM) in pore waters t h r o u g h o u t all cores (Morse, unpubl i shed data) indicate that Fe is un- likely to be limiting for i ron sulfide mineral for- marion. Consequently, there is no a priori reason to believe that the C /S control model is no t appli- cable to these sites. The relationships discussed in this paper should, therefore , be of more than local utility in unders tanding the role of the seafloor in the mar ine carbon cycle.

Based on the work o f Mackin and Swider (1989) and Canfield (1994), Morse and Berne r (1995) as- sumed that anaerobic processes domina te organic mat ter oxidat ion on the Louisiana shelf and that fcs could be taken, as a reasonable approximat ion, at its max imum value of 1 (he te ro t rophic metab- olism is strongly domina ted by sulfate reducing bacteria). T h e relationships shown in Fig. 3 indi- cate that this approximat ion appears to be reason- able for all sites except 2b, because integrated sul- fate rates are similar to or greater than e i ther O2 and CO,, fluxes.

At Site 2b the integrated sulfate reduc t ion rate was close to those observed at sites la, 2a, and 3. However, the Oz and CO 2 fluxes were about twice what would be expec ted if sulfate reduc t ion were the dominan t m o d e of organic mat te r oxidation. Thus fc.s will be taken as - 0 . 5 for Site 2b. Al though it could be simply fortuitous, the excel lent agree- m e n t between 0 2 and CO:, fluxes is probably in- dicative that o the r processes such as carbonate dis- solution and precipi tat ion are of minor impor- tance, at least at this site. An interest ing observa- t ion at the time of sampling was the presence of small goby-like fish with an abundance of about 2- 3 / m "~. These may have been stimulating sediment- water exchanges by burrowing into surficial layers o f sediment and increasing oxic respiration.

Th e fract ion of r educed sulfur that ends up be- ing bur ied (fsv) can be cak:ulated f rom the differ- ence between integrated rate o f sulfate reduct ion and the rate of TRS burial. Th e rate of TRS burial

Benthic Biogeochemistry Beneath a River Plume 213

"J . . . . I . . . . I . . . . I 'l'a'

I~ -2 / o_

-3

i i i b I i i i i ' i , i i I i i r I -4 - 3 - 2 -1 0 1

log Sediment Burlal Rate (g cm "2 y.1)

Fig. 4. Burial o f pyrite-S as a func t ion o f sed imenta t ion rate. U p p e r line is for coastal Connec t i cu t sed imen t s and lower line is for Louis iana she l f sed iments based on the da ta o f Lin and Morse (1991).

was obtained from the sediment burial rate (SBR), which was calculated from porosity (average values of bottom three samples were used for TRS and porosity), assuming a mean solid density of -2 .5 g cm -~. With the exception of Site la, the relation- ships between log fsP and log SBR are remarkably similar (Fig. 4) to those based on the data of Lin and Morse (1991). Consequently, our results re- confirm that fsP is about an order of magnitude less on the Louisiana shelf than in coastal Con- necticut for sediment burial rates of less than about I g cm -2 y-]. The exceptionally high value of fsp (buried S) at Site la may be the result of enhanced preservation because this site had the highest SBR, which was probably due to its close proximity of the mouth of the Mississippi River.

fsP and the C/S ratio can be used to calculate approximate values for f,a, assuming most hetero- trophic metabolism is from sulfate reduct ion (fcs~l) for all sites except 2b where a value of fcs = 0.5 will be used (see above). The percentage of organic-C buried can be calculated from f,~ and also be determined independently using organic- C burial rates and either O~ or CO~ fluxes. How- ever, the 02 derived values must be corrected to account for burial of reduced sulfur (e.g., Berner and Westrich 1985; Chanton and Martens 1987; Sampou and Oviatt 1991). (Note that burial of TRS as FeS2, rather than as S 2-, makes less than a 10% change in the coefficient 2 in equations 1, to a value of about 1.9, depending on which iron min- eral is considered the source of Fe2-.)

TABLE 5. Percen t organic-C bur ied calculated via the Morse and Berner (1995) mode l and f rom correc ted ( ~ e text) 02 and CO2 fluxes.

Si~ Model 02 CO~

la 56 41 51 l b 24 41 27 2a 7 14 22 2b 2 2.6 2.6 3 2 41 5.1

Values of percent organic-C buried obtained by these three methods are given in Table 5. At sites la, lb and 2b, there is good agreement between model results and those obtained from CO 2 fluxes. There is a difference between model and CO 2 flux results of about 3% at the seasonally hypoxic Site 3. This relatively good agreement between these different approachcs at 4 out of 5 sites implies that these approaches are giving fairly reasonable esti- mates of the fraction of organic-C that was being buried.

The only site of major disagreement between the model results and those derived from CO2 fluxes was at site 2a, which is interestingly the only site where 02 derived values are in better agreement with model results than the CO2 derived values. The only really glaring disagreement in results was the percent organic-C burial derived from 02 flux- es at hypoxic Site 3. Since the extent of organic matter oxidation is the inverse of the 02 flux, the 02 fluxes are less than predicted from the model. The explanation for Site 3 being abnormal is sim- ply that there was very little 02 in the overlying water, leading to a low flux of 02 into the sediment and apparently high carbon preservation. If there were no 02 in the overlying water, then there would be no 02 flux into the sediment. If carbon preservation were calculated solely via 02 flux, no oxidation of organic matter would be apparent. During times of the year when this site is overlain by oxic water, a much lower" degree organic-C buri- al would be calculated from 02 fluxes.

Conclusions The Louisiana shelf immediately to the west of

the Mississippi River delta is an area strongly im- pacted by the outflow of the Mississippi River. This leads to complex temporal and spatial variability in the biogeochemistry of sediments in this region. Even at close to the same water depth (--20 m) over a relatively small geographic area (10-30 km), major differences were observed in fluxes across the sediment-water interface. This included the rel- ative importance of different electron acceptors and on what appear to be large differences in the efficiency of organic-C burial. The temporal and

214 J. Morse and G. Rowe

spatial heterogeneity of this important region clearly points to the need more detailed experi- ments and for the establishment of a large data base before we have a complete understanding of regional processes. This study provides a wide va- riety of examples of benthic biogeochemical pro- cesses over the probable range of oxygen-dominat- ed to hypoxic bottom conditions The results from the model, although approximate, demonstrate the large differences in the relative importance of different diagenetic pathways at the study sites.

ACKNOWI,EDGMENTS

This research was supported by grants to J. W. Morse from the NSF Chemical Oceanography Program, and G. T. Rowe and J. W. Morse from National Oceanic and Atmospheric Adminis- tration as part of NECOP studies. The State of Texas also con- tributed support to this study. We thank the crew of the R/V Gyre, and Texas A & M marine technicians for their efforts, and gratefully acknowledge the help of Dr. Don Harper and his stu- dent divers, and Dr. Brian Eadie for running DIC on the April cruise. Greg Boland was responsible for the benthic chambers and several of colleagues helped in various aspects of the sample collection and analysis including Philip Cheeseman, D. Craig Cooper, Jennifer Holmes and Miguel Huerta-Diaz and Qiwei Wang.

LITERATURE CITED

ATWOOD, D. K., A. BRATKOVICH, M. GALIaXGIIER, AND G. L. HITCHCOCK. 1994. Introduction to the dedicated issue. Estu- aries 17:729-731.

BERNER, R. A. AND J. T. WESTRICH. 1985. Bioturbation and the early diagenesis of carbon and sulfur. American Journal of Sci- ence 285:193-206.

BLACKWFI.DER, P., T. HOOD, C. AVAREZ-ZARIKIAN, T. A. NELSON, AND B. McKEE. 1996. Benthic foraminifera from the Nutrient Enhanced Coastal Ocean Processes study area impacted by the Mississippi River plume and seasonal hypoxia. Quaternary International 31:19-36.

CANFIELD, D. E. 1994. Factors influencing organic carbon pre~ ervation in marine sediments. Chemical Ge0/ogy 114:315-329.

CANFIELD, D. E., R. RAISWELL, J. T. WI/.STRICII, C. M. REAVES, AND R. A. BERNER. 1986. The use of chromium reduction in the analysis of reduced inorganic sulfur in sediments and shales. Chemical C, eolog2y 54:149-155.

CHANTON,J. E AND C. S. MARTENS. 1987. Biogeochemical cycling in an organic-rich coastal marine basin. 7. Sulfur mass bal-

ance, oxygen uptake and sulfide retention. Geochimica et Cos- mochimica A cta 51:1187-1199.

CLINE, J. D. 1969. Spectrophotometric determination of hydro- gen sulfide in natural waters. Limnology and Oceanography 14: 454---458.

CORNWELL, J. C. ANDJ. W. MORSE. 1987. The characterization of iron sulfide minerals in marine sediments. Marine Chemistry 22:193-206.

DORTCH, Q., N. RABALMS, R. E. TURNER, AND G. ROWE. 1994. Respiration rates and hypoxia on the Louisiana shelf. 'Estuaries 17:862-872.

GEARING, E, E E. PLUCKER, AND E L. PARKER. 1977. Organic carbon stable isotope ratios of continental margin sediments. Marine Chemistry 5:251-266.

GUPTA, B. K. S., R. E. TURNER, AND N. N. RABALAIS. 1996. Sea- sonal oxygen depletion in continental-shelf waters of Louisi- ana: Historical record of benthic foraminifers. Ge0/ogy 24:227- 230.

JUSTI(;, D., N. N. RABALAIS, AND R. E. TURNER. 1996. Effects of climate change on hypoxia in coastal waters: A doubled CO 2 scenario for the northern Gulf of Mexico. Limnology and Oceanography 41:992-1003.

LIN, S. ANI)J. W. MORSE. 1991. Sulfate reduction and iron sulfide mineral formation in Gulf of Mexico anoxic sediments. Amer- ican Journal of Science 291:55--89.

MACKIN, J. E. AND K. T. SWIDER. 1989. Organic matter decom- position pathways and oxygen consumption in coastal marine sediments. Journal of Marine Research 47:681-716.

MIIA.ER-WAY, T., G. BOLAND, G. ROWE, AND R. TWILLEY. 1994. Sediment oxygen consumption and benthic nutrient fluxes on the Louisiana continental shelf: A methodological com- parison. Estuaries 17:809-815.

MORSE, J. W. AND R. A. BERNFR. 1995. What controls sedimen- tary C/S ratios? Geochimica et Cosmochimica Acta 59:1073-1077.

REEBUrtGli, W. S. 1967. An improved interstitial water sampler. Limnology and Oceanography 14:454-458.

Row~;, G., G. BOI.ANI), W. PHOEL, R. ANDERSON, AND P. BI~;AYE. 1994. Deep-sea floor respiration as an indication of lateral input of biogenic detritus from continental margins. Deep-Sea Research H 41:657-668.

SAMI'OtJ, E AND C. A. OVIATT. 1991. A carbon budget for a eu- trophic marine ecosystem and the role of sulfur metabolism in sedimentary carbon, oxygen and energy dynamics. Journal of Marine Research 49:825-844.

UNPUBLISI tED 1M_ATERIAI.

MCKEE, B. 1995. Personal communication.

Received for consideration, July 21, 1997 Accepted J~ publication, August 19, 1998