Seasonal patterns of carbonate diagenesis in

0 downloads 0 Views 705KB Size Report
Jun 9, 1998 - Samples were diluted in 0.2% Ultrex- ..... other studies (Carpenter, 1969; Baker et al., 1982; Rude and Aller, 1991; Green et al.,. 1998]. 1109.
Journal of Marine Research, 56, 1097–1123, 1998

Seasonal patterns of carbonate diagenesis in nearshore terrigenous muds: Relation to spring phytoplankton bloom and temperature by Mark A. Green1,2 and Robert C. Aller1 ABSTRACT Pore water saturation state with respect to calcite and aragonite minerals in Long Island Sound sediments  uctuates from saturated and near saturated conditions during late fall, to undersaturated during winter, before slowly changing to supersaturated conditions during late spring. Undersaturation occurs during cold, winter periods when lower rates of S CO2 production (low rates of heterotrophic metabolism) and oxidation of reduced minerals such as FeS lower calcium carbonate saturation states. Direct evidence that dissolution of both calcites and aragonite are occurring during this season comes from the simultaneous increases in excess pore water carbonate dissolution products Ca21 , F2 , and Sr21 during periods of pore water undersaturation.Higher S CO2 production rates during warmer periods cause the CO322 concentrationto become supersaturatedfor both calcite and aragonite. S CO2 production is controlled by both temperature and substrate availability so that benthic deposition of organic matter produced during the spring bloom accelerates the seasonal progression of pore waters to supersaturation. These patterns control carbonate dynamics in temperate, nearshore regions, and result in a regularly observed, yearly cycling of calcium carbonate dominated by alternating periods of net dissolution and net precipitation.

1. Introduction Relatively little research has considered the diagenetic carbonate cycle in nearshore regions of terrigenous sediment deposition. These deposits are seldom more than 1–3% CaCO3 by weight, and are generally not viewed as quantitatively important calcium carbonate sinks compared to platform carbonates or deep-sea sediments. However, studies have shown that nearshore regions are sites of dynamic and quantitatively signiŽ cant carbonate cycling, with seasonal patterns of net dissolution and deposition (Aller, 1982; Reaves, 1986; McNichol et al., 1988; Green et al., 1992, 1993). The time variation of carbonate saturation state and its relation to environmental factors are still not well documented in nearshore deposits where large, rapid changes in temperature and carbon  ux typically occur seasonally. In the present study, saturation states in estuarine surŽ cial 1. Marine Sciences Research Center, State University of New York at Stony Brook, Stony Brook, New York, 11794-5000, U.S.A. 2. Present address: Saint Joseph’s College, Department of Environmental Science, 278 White Bridge Road, Standish, Maine, 04084-5263, U.S.A. 1097

1098

Journal of Marine Research

[56, 5

Figure 1. Map of Long Island Sound showing PULSE sampling station. Contour lines represent 10 meter depth-intervals.

muds were examined in detail during a period of dynamic seasonal change in temperature and C-organic  ux: the spring phytoplankton bloom. More speciŽ cally, this study evaluated pore water composition and thermodynamic saturation states with respect to calcite and aragonite at approximately two week intervals for 11 successive cruises during the winter and spring in Long Island Sound (LIS). Analysis of Ca21 , F2 , and Sr21 solute concentrations provide evidence for dissolution of different carbonate minerals during periods of undersaturation. Stoichiometric relations of F2 vs. Ca21 , Sr21 vs. Ca21 , and F2 vs. Sr21 in pore water were used to infer the likely reacting mineral phases: lo-Mg calcite, hi-Mg calcite, and aragonite, which differ in F2 , Sr21 , Mg21 and Ca21 elemental ratios. In addition, we consider the overall reactions responsible for in uencing pore water saturation state, with particular emphasis on the production of S CO2 before and after benthic deposition of the spring bloom phytoplankton, and on the oxidation of acid-volatile sulŽ des. A one-dimensional transport-reaction model is used to describe the overall contribution of CaCO3-derived S CO2 to pore water during periods of greatest pore-water carbonate undersaturation. 2. Station location and sampling Sampling was conducted at a single station in central Long Island Sound (LIS; Fig. 1) as part of the PULSE study (41°10.03N, 72°57.43W) during winter–spring 1993. This project investigated a wide array of benthic biogeochemical responses to deposition of the spring bloom in LIS sediments. The general location of the study has been extensively researched in the past from a geochemical as well as a biological perspective (e.g., Station NWC: Aller and Cochran, 1976; Aller, 1977, 1982; Benninger et al., 1979; Station P: Green et al., 1992, 1993). Water depth at this station averages , 15 meters. Sediments are muddy, consist of

Green & Aller: Seasonal patterns of carbonate diagenesis

1998]

1099

Table 1. Cruise date, corresponding year day, and surface and bottom water temperatures (°C). Cruise

Cruise date

Year day

Surface/bottom water temp.

Pulse 1 Pulse 2 Pulse 3 Pulse 4 Pulse 5 Pulse 6 Pulse 7 Pulse 8 Pulse 9 Pulse 10 Pulse 11

12/15/92 1/21/93 2/8/93 3/1/93 3/22/93 4/12/93 4/26/93 5/10/93 5/24/93 6/7/93 6/28/93

2 16 21 39 60 81 102 116 130 144 158 179

6.0/6.6 3.1/3.1 0.8/1.5 0.2/0.4 1.0/0.1 3.8/3.6 7.7/4.6 13.1/8.7 11.1/7.8 12.3/11.1 18.6/15.8

, 2% organic carbon and CaCO3 by weight, and have an average sediment porosity (f ; 0–1 cm) of , 0.8. Seasonally averaged macrofaunal abundances (. 0.5 mm sieve) are 949 6 720 m2 2. The numerical order of relative abundances of macrobenthos is bivalves . polychaetes . crustaceans. A Soutar-type box corer (0.1 m2) was used to collect bottom sediment and overlying water during each cruise. Cruises were at approximately 14-day intervals which provided relatively high sampling resolution. Sampling dates, corresponding progressive year days, and surface and bottom water temperatures are given in Table 1. 3. Methods A large subcore (14.5 cm I.D.; CAB tube) was taken from a box core during each cruise and stored in a shipboard incubator at in situ temperature for transport back to the laboratory. Once back in the laboratory (within , 6 hr of core retrieval), pH was determined throughout individual cores by direct insertion of a combinationmicroelectrode (Microelectrodes Inc., Londenderry, NH; calibration with pH 4 and 7 NBS-traceable buffers) into sediment (1 mm intervals). Next, cores were sectioned under nitrogen in a glove bag, and pore water separated at in situ temperature without air contact using centrifugation (10–15 min, 5000 rpm, 20–40 ml pore water). Sectioned intervals varied slightly between cruises 1–3 and 4–11. Depth intervals sectioned were 0–0.5, 0.5–1.0, 1.0–2.0, 2.0–3.0, 3.0–4.0, 4.0–5.0, and 5.0–7.0 cm during cruises 1–3 and 0–0.25, 0.25–0.5, 0.5–0.75, 0.75–1.0, 1.0–1.5, 1.5–2.0, 2.0–3.0, 3.0–4.0, 4.0–5.0, and 5.0–7.0 cm during cruises 4–11. Pore waters were Ž ltered through a 0.45 µm Acrodisc Ž lter, acidiŽ ed with concentrated hydrochloric acid (10 µl 12N HCl per 1 ml sample) to prevent solid phase precipitation, stored in acid-washed and distilled-water-rinsed scintillation vials, and refrigerated for later analysis. Chloride (Cl2 ) concentrations were measured on unacidiŽ ed aliquots using a Radiometer CMT chloride titrator (precision , 1.0%). Total alkalinity was measured on 2 ml samples after Edmond (1970). S CO2 was measured on unacidiŽ ed samples at the time of collection by  ow injection analysis (FIA) and conductivity detection (Hall and Aller;

1100

Journal of Marine Research

[56, 5

1992). Carbonate (CO322 ) ion concentrations were estimated using alkalinity, Cl2 , pH and temperature data using the Ž rst and second dissociation constant of carbonic acid in seawater (Mehrbach et al., 1974) as Ž t to in situ temperature and salinity conditions (Millero, 1979). Portions of the acidiŽ ed sample were diluted in La21 (1 mg ml2 1) and Ca21 analyzed with a precision of 6 3.0% on a Hitachi Z-8100 Polarized Zeeman Atomic Absorption Spectrophotometer using an air-acetylene  ame. Overlying- and pore-water saturation states with respect to calcite and aragonite, V (V 5 IMP/K8sp), were estimated for each of the 11 cruises by calculating the ion molarity products (IMP; mCa21 3 mCO32 2) and comparing them to the apparent equilibrium constants (K8sp) using the salinity- and temperature-dependent equation of Gieskes (1974). Due to the potential problems associated with use of NBS pH buffers in high ionic strength media such as sea water, additional estimates of pore water calcite and aragonite saturation states were made from S CO2 and alkalinity measurements using the relationships of Morse and Mackenzie (1990) to estimate pore water H1 activity (aH1 ). aH1 were used to determine CO322 using pore water alkalinity, Cl2 , temperature, and the Ž rst and second dissociation constant of carbonic acid under the given temperature and salinity conditions (Millero, 1979). A pore water subsample was used for Sr21 analysis on a Perkin Elmer HGA Graphite Furnace Atomic Absorption Spectrophotometer. Samples were diluted in 0.2% Ultrexgrade HNO32 and atomized at 2600°C using a Mg(NO3)2 matrix modiŽ er. F2 was determined using an ion slective electrode (Frant and Ross, 1966) by the method of standard addition. Sediment incubation experiments were used to measure net rates of NH41 , S CO2, Fe21 , and F2 production reactions (Goldhaber et al., 1977; Aller and Mackin, 1989). Two butyrate subcores (I.D. 5 14.5 cm) collected with overlying water were removed from a box core during each cruise and placed in an incubator at in-situ temperature for transport back to the laboratory. Overlying water was continuously aerated to maintain in-situ geochemical gradients, taking care not to disturb the sediment-water interface during bubbling. In the laboratory (within 5–8 h after core retrieval) the core was placed in an N2-purged glove bag, carefully drained of overlying water, and sediments extruded upward until the interface was level with the edge of the core liner. Approximately twenty-four 10 cc syringes, with the tips cut off at the base to resemble small piston-coring devices, were carefully inserted into the sediment while keeping the plunger even with the sediment interface. Syringe subcores were inserted to an exact depth of 4 cm in each core. Sediment from the 2–4 cm interval and the 0–2 cm interval were separately extruded into acidwashed, distilled-water-rinsed scintillation vials. This procedure was repeated to completely Ž ll each vial. Vials were sealed, removed from the glove bag, and incubated without air contact at in situ temperatures by burying them in a bucket Ž lled with anoxic mud taken from the study location.After an initial time zero sample, incubated samples were retrieved every 12–24 h for 3–4 days and pore water collected after centrifugation. S CO2, and F2

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1101

were analyzed as described previously. NH41 was measured using FIA (Hall and Aller, 1992). Total dissolved iron (Fe21 ) was measured using ferrozine (Stookey, 1970). Immediately following pore water extraction, the sediment pellet from the time zero incubation was frozen and saved for acid volatile sulŽ de analysis. Prior to analysis, sediment was thawed under N2 in a glove bag and homogenized by gently stirring. Care was taken to remove any surface sediment that appeared slightly dry or oxidized prior to homogenization.A subcore of the pellet was taken in a pre-weighed syringe, capped prior to removal from the glove bag, and reweighed. This sediment was quickly dropped into a test tube containing N2-purged 6N HCl and the emitted sulŽ de trapped in two inline test tubes by precipitation with 0.05 M Zn-acetate. HS2 was analyzed spectrophotometrically with diamine reagent after Cline (1969), and acid volatile sulŽ des expressed as inventories from 0–4 cm (µmol FeS cm2 2). Sediment oxygen penetration depth was measured on each cruise using a platinumtipped, Clarke combination microelectrode calibrated using air- and nitrogen-bubbled seawater. Measurements were taken on board ship from three butyrate subcores (4.5 cm I.D) each taken from separate box cores immediately after retrieval of the cores. A Brinkman micro-manipulator allowed for 0.1 mm depth resolution. The O2 penetration depth was expressed as an average value of two proŽ les from each of the three cores. 4. Results a. Estimates of pore water calcite/aragonite thermodynamic saturation states Although some small deviations existed between the CO322 concentration calculated using directly measured pH, and pH calculated from S CO2 and alkalinity, the overall time-dependent pattern of calcite and aragonite saturation states remained the same. As little difference existed between the two values, and because previous studies in LIS and other nearshore regions employed the use of NBS pH buffers for calculation of aH1 (Aller, 1982; Reaves, 1986; McNichol et al., 1988; Walter and Burton, 1990; Rude and Aller; 1991), the saturation states based on directly-measured values of pH are used. This provides a consistent data set necessary for comparison with previous studies. The ratio of calculated ion molarity products (IMP) relative to apparent solubilities (K8sp) for both calcite and aragonite in the top 7 cm from cores taken during each cruise are plotted as omega (V 5 IMP/K8sp) values in Figure 2. Differences in the bottom water K8sp values for both calcite and aragonite from cruise to cruise resulted mostly from changing bottom water temperatures. However, small changes in Cl2 were also observed (, 0.2%) and caused a small down-core variability in K8sp during some cruises. There were no measurable temperature gradients (6 0.1°C) within the top 7 cm of retrieved cores during any of the 11 cruises. There are regular changes in bottom water and pore water calcite and aragonite saturation states during the cruise period. With the exception of cruises 4 and 5, bottom waters are either at equilibrium (V 5 1) or supersaturated (V . 1) with respect to calcite.

1102

Journal of Marine Research

[56, 5

Figure 2. Pore water calcite (closed circles) and aragonite (open circles) saturation states expressed as omega (IMP/K8sp) values during PULSE cruises 1–11. Overlying water saturation states are given by solid (calcite) and open (aragonite)triangle symbols. Undersaturationoccurs to the left of the dashed line and supersaturationis to the right.

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1103

For aragonite, bottom waters are undersaturated (V , 1) during cruises 1–9 and supersaturated during cruises 10 and 11 (V . 1). Pore water calcite saturation states vary regularly from mostly supersaturated conditions during cruise 1, to increasing degrees of undersaturation during cruises 2–6. Undersaturation is initially conŽ ned to a zone just below the sediment-water interface during cruise 1, which propagates downward during successive cruises to encompass the entire 0–7 cm interval by middle cruises. Saturation or supersaturation progresses in the same fashion, beginning Ž rst at the sediment-water interface during cruise 7, and subsequently migrating downward during cruises 8 and 9. The entire sampled interval is supersaturated during cruises 10 and 11. For aragonite, pore waters generally follow the same relative pattern, but with greater degrees of undersaturation. Sediments are supersaturated with respect to aragonite by cruise 11. b. Interstitial water Ca21 , F2 , Sr21 In order to alleviate the slight between-cruise and down-core concentration  uctuations due to small changes in Cl2 and emphasize inter-cruise differences in pore water proŽ les of the carbonate mineral constituents Ca21 , F2 , and Sr21 , these solutes are expressed as ‘‘excess’’ concentrations.This excess is relative to values expected from ratios of overlying water solute concentrations to Cl2 , and the measured pore-water solute concentration ratio to Cl2 according to: Excess [x]pw 5

[x]pw 2

([Cl2 ]pw 3

[x]ow/[Cl2 ]ow)

(1)

where [x]pw 5 concentration of solute x in pore water, [Cl2 ]pw 5 concentration of Cl2 in pore water, [x]ow 5 concentration of solute x in overlying water, and [Cl2 ]ow 5 concentration of Cl2 in overlying water. Plots of pore water Ca21 , F2 , and Sr21 in excess of overlying water values are plotted for each cruise in Figures 3–5. While some variability exists between elements, the general trend through the 11-cruise study period is: (1) relatively low excess values of pore water solutes during late fall and early winter; (2) increasing concentrations to maximum excesses during low temperature, winter cruises; and (3) decreasing values approaching no excess following bloom input and warming temperature. c. Sediment incubations A time series of S CO2 and NH41 production rates from 0–2 and 2–4 cm sediment intervals were calculated from incubation experiments during each cruise. Sediment incubations were assumed to provide good estimates of solute production rates as short incubation times (, 5 days with 12–24 hour sampling intervals) provided linear slopes of solute production, suggesting minimal artifacts such as inhibition from reaction products. Production rates for S CO2 and NH41 were converted to whole sediment volume units of µmol S CO2 cm2 3 d2 1 and nmol NH41 cm2 3 d2 1 using sediment porosities determined for each cruise and are shown in Figure 6.

1104

Journal of Marine Research

[56, 5

Figure 3. Pore water Ca21 (mM) in excess of overlying water values during PULSE cruises 1–11.

Production rates for S CO2 (Fig. 6a) show a regular time dependent pattern of change, with rates decreasing from cruise 1 and 2, remaining low for middle cruises (3–8), before increasing rapidly. Production rates from the 0–2 cm interval range in value from a low of 0.06 µmol S CO2 cm2 3 d2 1 during cruise 3 to a high of 0.64 µmol S CO2 cm2 3 d2 1 during

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1105

Figure 4. Pore water F2 (µM) in excess of overlying water values during PULSE cruises 1–11.

cruise 11. Production rates from 2–4 cm vary from a cruise 3 low of 0.023 µmol S CO2 cm2 3 d2 1, to a high during cruise 11 of 0.33 µmol S CO2 cm2 3 d2 1. With the exception of cruise 9, production rates in the 0–2 cm interval are higher than those from 2–4 cm.

Journal of Marine Research

1106

[56, 5

Figure 5. Pore water Sr21 (µM) in excess of overlying water values during PULSE cruises 1–11.

NH41 production was corrected for adsorption using a linear adsorption coefficient, K, (Mackin and Aller, 1984) by the relationship: K5 where: f

5

(1 2

f )/f 3

r

s

3

weighted mean porosity from 0–2 cm, r

K* s

5

(2) dry sediment density (assumed

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1107

Figure 6. S CO2 and NH41 production in incubated sediment samples during PULSE cruises 1–11. Incubated sediment intervals were from 0–2 cm (open circles) and 2–4 cm (closed circles). The period of the spring phytoplanktonbloom and its subsequentdeposition to underlying sediments is noted.

2.65 g cm2 3), and K* 5 ratio of rapidly-exchangeable NH41 concentration in µmol · g2 1 dry weight sediment to pore water NH41 concentration in mM relative to pore water volume (measured K* 5 1.21 g ml2 1). Production rates for NH41 (Fig. 6b) follow a pattern similar to S CO2 production in the 0–2 cm interval, decreasing following cruise 1 (, 40 nmol NH41 cm2 3 d2 3) to lows of , 10 nmol NH41 cm2 3 d2 1 during cruises 2–8, before increasing rapidly during cruises 10 and 11 (cruise 9, 0–2 cm NH41 production rate

1108

Journal of Marine Research

[56, 5

Figure 7. Sediment inventories of acid volatile sulŽ des (µmol FeS cm2 2; 0–4 cm) along with oxygen penetration depth into sediment during PULSE cruises 1–11.

not available; cruise 10 5 75 and cruise 11 5 81 nmol NH41 cm2 3 d2 1). With the exception of cruise 2, NH41 production rates from 2–4 cm are lower than those from 0–2 cm. The main deposition period of spring bloom C-organic to sediments determined from sediment-Chl-a inventories (Gerino et al., 1998) is also shown in Figure 6. The annual deposition of phytoplankton detritus to sediments is associated with increases in S CO2 production and, as discussed later, is largely responsible for pore water supersaturation during the post-bloom cruises. d. Sediment acid volatile sulŽ des Acid volatile sulŽ de inventories (µmol FeS cm2 2; 0–4 cm), are plotted along with depth of oxygen penetration into sediment during each cruise in Figure 7. FeS inventories range from a high of 9.5 µmol FeS cm2 2 during cruise 10, to a low of 2.7 µmol FeS cm2 2 during cruise 3 (year day 39). With the exception of cruise 4 (year day 60), FeS inventories are generally lowest during middle, winter time cruises and highest during earlier and later cruise periods. Oxygen penetration into sediment shows a strong seasonal pattern with increasing penetration through the Ž rst four cruises to a maximum depth of , 5.5 mm, after which shallowing occurs through a cruise 11 minimum of , 1 mm (year day 179). 5. Discussion a. Calcite and aragonite saturation states During this study, thermodynamic predictions of pore water saturation state suggest that calcite dissolution is possible in surface sediments during cruises 2–7 and for aragonite during cruises 2–10 (Fig. 2). These saturation states represent typical winter–early spring conditions in Long Island Sound (LIS). Aller (1982) and Green et al. (1993) calculated

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1109

similar states of pore water calcite/aragonite undersaturation during winter–spring at this station. Reaves (1986) also reported pore water carbonate undersaturation during winter in tidal creek sediments along the north shore of LIS. McNichol et al. (1988) recorded the highest degrees of pore water undersaturation during June and August in Buzzards Bay, MA; however, large increases in pore water Ca21 were seen during March, suggesting dissolution of CaCO3 during this period as well. Data from all studies are therefore consistent with general winter–early spring undersaturation and possible site-speciŽ c patterns during other seasons. The present study demonstrates that in LIS both under- and supersaturation are initiated near the sediment surface and propagate downward during early winter and late spring, respectively. b. Evidence for CaCO3 dissolution Thermodynamic predictions of pore water undersaturation are not conclusive evidence that dissolution occurs, only that chemical conditions exist that could make it possible. Additional factors that can control reactions include inhibition by solutes such as orthophosphate (Morse and Berner, 1972; Morse, 1974; Walter and Burton, 1986), magnesium and sulfate (Sjoberg, 1978); kinetic hindrance due to rate limitation by transport processes at mineral surfaces (Berner and Morse, 1974; Plummer et al., 1978); variable grain size composition (Morse, 1983; Walter and Morse, 1984); uncertainty of the exact composition of the dissolving mineral phases; and differences between the calculated stoichiometric solubility products (K8sp) and actual equilibrium K8sp values of the solution. Direct evidence that carbonate dissolution is occurring comes from the simultaneous increases in excess pore water carbonate dissolution products Ca21 , F2 , and Sr21 during periods of pore water undersaturation (Figs. 3–5). A comparative plot of the depthintegrated calcite and aragonite saturation state from 0–5 cm and the inventory of excess Ca21 and F2 over this same depth interval as a function of time clearly shows radically increasing excess concentrations of these solutes as pore waters become more undersaturated during winter (Fig 8). As explained in detail subsequently, although F2 has multiple mineral sources, dissolution of CaCO3 is the likely explanation for increases in this solute during carbonate undersaturation. An increase in inventories could also result from decreasing transport rates, causing an increase in pore water solute concentrations at a constant dissolution rate by allowing reaction products to build up. However, Br2 tracer experiments to evaluate solute transport processes during each cruise showed that while transport decreased slightly due to temperature over the initial period of increasing inventories and decreasing saturation state (cruises 1–4), transport changes alone were too small to explain the large inventory excesses observed (Green and Aller, in prep.). Actual rates of dissolution appear to be increasing steadily, producing larger reaction product inventories as pore waters become more undersaturated. c. Dissolving mineral phases Pore water values of Ca21 , Sr21 , and F2 have been used to infer dissolution processes in other studies (Carpenter, 1969; Baker et al., 1982; Rude and Aller, 1991; Green et al.,

1110

Journal of Marine Research

[56, 5

Figure 8. Comparative plots of the depth-integratedcalcite (closed diamonds) and aragonite (closed triangles) saturation states from 0–5 cm and excess Ca21 and F2 inventories over this same depth range during PULSE cruises 1–11.

1992). In principle, stoichiometric relations of F2 vs. Ca21 , Sr21 vs. Ca21 , and F2 vs. Sr21 in pore water re ect the relative contribution of different reacting mineral phases. Solute ratios from cruises 2–8 (net dissolution period) were Ž rst corrected for differential diffusion using published values of diffusion at inŽ nite dilution (Li and Gregory, 1974). Average slope values yield ratios of F2 /Ca21 , 6.8 6 8.7 (mmol/mol), Sr21 /Ca21 , 8.6 6 7.5 (mmol/mol), and F2 /Sr21 , 0.8 6 0.9 (mol/mol), respectively, and are consistent with the dissolution of both aragonite (F 2 /Ca21 , 4.0–11.0, Sr21 /Ca21 , 9.0–12.0, and F2 /Sr21 , 0.5–1.0) and high magnesian calcite (F2 /Ca21 , 1–9, Sr21 /Ca21 , 2.5–4.0, and F2 /Sr21 , 0.3–3.0; Morse and Mackenzie, 1990; Rude and Aller, 1991). The ratios expected for low magnesian calcite would be F2 /Ca21 , 0.3, Sr21 /Ca21 , 2, and F2 /Sr21 , 0.2

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1111

The exact ratio of the dissolving mineral phases may be different from the net value observed in interstitial water. Simultaneous reprecipitation can remove solutes with a distribution coefficient different from that of a dissolving mineral. As a result, the net release of any given solute may depart from that produced from a simple congruent dissolution reaction. Thermodynamic stabilities, and the in uences of mineralogy and grain microstructure, indicate that dissolution and precipitation reactions of numerous mineral phases can occur within the same sedimentary zone (Walter and Morse, 1985). For example, the existence of diffusive gradients below the Sr21 and F2 subsurface peaks are consistent with a reprecipitating mineral phase or deep irrigation transport by macrofauna. In addition, Ca21 proŽ les reach near constant concentrations at depths $ 2 cm, despite undersaturation with respect to both calcite and aragonite, also indicating interaction with another mineral phase, or deep irrigation. A calcium  uoro-phosphate phase may be precipitating at or near this region of the deposit (Jahnke et al., 1983; Froelich et al., 1983; McNichol et al., 1988; Ruttenberg and Berner, 1993). Interstitial water separated from vibra-corer sections taken from LIS was shown to be supersaturated with respect to carbonate  uorapatite by a factor of 104 to 106 making precipitation thermodynamically favorable (Ruttenberg and Berner, 1993). The high resolution core sectioning used in this study reveals that carbonate  uorapatite might be precipitating much closer to the sediment-water interface than was described by Ruttenberg and Berner (1993). Another possible source of F2 to pore waters is from release during reduction of Fe-oxyhydroxides (Ruttenberg, 1990; Rude and Aller, 1993). In the present cases F2 peaks are consistently 1–2 cm higher in the deposit than the peak of Fe21 produced during the reduction of ferric iron. In addition, Fe21 and F2 production in pore waters of incubated sediment (0–2 cm) during cruises 3–6 differ in time-dependent release patterns. F2 production increases for 8–14 days before noticeable changes in interstitial Fe21 occur (Fig. 9). Therefore, F2 is likely associated with dissolving calcium carbonate. Increases in Ca21 were also detected during the initial sampling points of sediment incubations (Green and Aller, in prep.). By cruise 11 however, destruction of the ferric iron carrier phase was the likely source of the F2 ion as warmer temperatures and greater reaction rates resulted in rapid production of Fe21 , and a corresponding increase in F2 . d. Controls on saturation state Seasonal variations in saturation state and dissolution/precipitation depend on balances between metabolic acid production, sedimentary transport processes, and temperaturedependent thermodynamic stabilities. The seasonal pattern of diagenetic reactions in surŽ cial LIS sediments consists of warm periods where organic inputs are high and build-up of reaction products dominate, and colder periods of relatively low heterotrophic metabolism and net oxidation of previously stored anaerobic metabolites. These patterns appear to control carbonate dynamics in temperate, nearshore regions, and result in a regularly observed, yearly cycling of calcium carbonate dominated by alternating periods of net dissolution and net precipitation.

1112

Journal of Marine Research

[56, 5

Figure 9. Fe21 (open circles) and F2 (closed circles) production in pore water from incubated sediment (0–2 cm) during PULSE cruises 3, 4, 5, 6, and 11.

Winter-time is the dominant period of carbonate undersaturation, which results predominantly from the production of metabolic acids. The aerobic decomposition of organic carbon is a primary source of metabolic carbonic acid and can be expressed as: (CH2O)x(NH3)y(H3PO4)z1 (x1 2y)O2 ® xCO2 1 (x 1 y)H2O 1 yHNO3 1 zH3PO4

(3)

Assuming RedŽ eld stoichiometry of x 5 106, y 5 16, and z 5 1, 106 carbon dioxide molecules should be produced for every 138 oxygen molecules consumed, further assuming complete oxidation of NH41 ® NO22 ® NO32 . If acid production is dominated by aerobic carbon oxidation and there is no oxidation of earlier precipitated metabolites, or, if suboxic and anoxic metabolic pathways are active but anaerobic products such as Fe21 and S2 are reoxidized, then an average dissolved oxidized carbon to oxygen  ux ratio of , 0.77 should exist. Nonsteady state oxidation of previously stored reduced S will lower this theoretical ratio, and storage of reduced metabolites (e.g., Fe21 , HS2 ) produced during anaerobic organic carbon decomposition will increase it. Calculated oxygen  uxes into sediments were compared with  uxes of S CO2 out of

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1113

Figure 10. S CO2/O2 diffusive  ux ratio during PULSE cruises 1–11. The solid horizontal line represents a ratio based on a RedŽ eld stoichiometric relationship.

sediments during each cruise to evaluate the likely diagenetic reaction balances producing acid and controlling saturation state at any give time. S CO2 concentration gradients over 0–0.5 cm during cruises 1–3, and over 0–0.25 cm during cruises 4–11, and the linear portion of O2 gradients (0.1 mm resolution) were used to calculate  uxes (J) from Fick’s Law (Berner, 1980): J5

2 f

oDs(dC/d )

(4)

where: f o 5 porosity at depth z 5 0, Ds 5 whole sediment diffusion coefficient, and dC/dz 5 S CO2 or O2 concentration gradient. Flux estimates for cruises 9–11 substituted the whole sediment diffusion coefficient (Ds) with apparent diffusion coefficients (DI) determined from the transport of Br2 from overlying water into incubated microcosm cores retrieved during each cruise. Experiments which evaluated solute and particle transport throughout the study period are described in detail in Gerino et al. (1998) and Green and Aller (in prep.). The calculated diffusive  ux ratios of S CO2 and O2 are shown in Figure 10. There was a regular pattern of change throughout the experimental period with ratios of , 1.15 during cruise 1, decreasing through winter to a low during cruise 6 of , 0.25, before increasing again to . 0.77 during later cruises. During middle cruises 2–7, J-S CO2/J-O2 ratios were much less that 0.77, suggesting an additional sink of oxygen which cannot be explained by S CO2 produced during respiration. One possible O2 sink is the oxidation of sedimentary Fe-sulŽ des. In the present case, the acid-volatile sulŽ de portion of the Fe-sulŽ de pool was used as an indicator of net reactions. The oxidation of acid-volatile sulŽ des (FeS) is rapid and capable of taking place abiogenically within the pH range typical of marine sediments (e.g., Aller, 1980). Cold winter temperatures, low metabolic activity, and the relatively deep sedimentary oxygen

Journal of Marine Research

1114

[56, 5

penetration probably result in coupled oxidation-dissolution reactions of the type: FeS 1 FeS2 1

9/4O2 1 15/2O2 1

2CaCO3 1 4CaCO3 1

5/2H2O ® 2Ca21 1 7/2H2O ® 4Ca21 1

2HCO32 1 4HCO32 1

SO422 1

Fe(OH) 3

(5)

2SO422 1

Fe(OH) 3

(6)

Note that these oxidation-dissolution reactions produce apparent S CO2/O2 ratios of , 1.0 (0.89–0.53) due to the contribution of CaCO3-S CO2 and raise the observed J-S CO2/J-O2 ratios by contributing CaCO3-derived S CO2 to pore waters. Hence, actual J-S CO2/J-O2 ratios based solely on reduction-oxidation reactions involving carbon remineralization are less than those calculated here and suggest that an even larger O2 sink actually exists relative to estimates of J-S CO2 and an assumed stoichiometry of the reacting organic carbon. Measured FeS inventories changed seasonally, and dropped from , 8.5 µmol FeS cm2 2 (0–4 cm) during cruises 1 and 2, to mid-winter inventories of , 3 µmol FeS cm2 2 (Fig. 7). Assuming reaction (5) to be signiŽ cant and essentially complete during periods of declining FeS inventories, then the oxidation of , 0.05 µmol FeS cm2 2 d2 1 (from 0–4 cm) between cruises 1–5 should support a production  ux of Ca21 equal to , 2 1.0 mmol Ca21 m2 2 d2 1. This estimate agrees well with values of Ca21  ux from a similar location in LIS calculated by Aller (1982) and Green and Aller (in prep.) of 2 0.55 and 2 1.9 mmol m2 2 d2 1, respectively. Therefore, Fe-sulŽ de oxidation is capable of driving the observed cold-water dissolution of CaCO3 in the absence of direct metabolic CO2 production. If CaCO3 dissolution is not rapid enough to neutralize the local production of acid, CO322 may diffuse from nearby regions (Emerson and Bender, 1981). Such diffusion may promote undersaturation deeper in the deposit where oxidation of Fe-sulŽ des is unlikely. The net oxidation of Fe-sulŽ des during winter represents the loss of anaerobic metabolites stored during the preceding warmer periods of greater respiration. In contrast, supersaturation, prevails during late spring and summer, and results from an increase in anaerobic CO2 production, as well as anaerobic metabolite storage (e.g., FeS, FeS2). In temperate regions such as Long Island Sound, there is a regularly observed seasonality in water column productivity with maximum water column phytoplankton concentrations occurring generally during late winter (spring bloom) and a smaller, secondary bloom occurring usually in November (autumn bloom; Conover, 1956). During the present study, the autumn bloom occurred prior to cruise 1 and the spring bloom occurred between cruises 6 and 7 (for a complete review see Gerino et al., 1998). The freshly deposited reactive carbon from the spring bloom produced high sedimentary Chl-a inventories (Gerino et al., 1998), and was rapidly mixed below the oxic region. Rising temperatures, increased organic loading, and rapid penetration into anoxic zones, promoted anaerobic pathways of decomposition. The associated production of HCO32 (Froleich et al., 1979; Berner, 1980; Emerson et al., 1981), resulted in progressive pore water supersaturation with respect to calcite and aragonite following the bloom input. Benthic foram production and juvenile bivalve settlement also occur in conjunction with the spring

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1115

Figure 11. Comparative plot of the depth-integrated (0–5 cm) pore water saturation state of calcite (closed circles) and aragonite (closed diamonds) saturation states and the S CO2 production (µmol cm2 3 d2 1) during PULSE cruises 1–11.

bloom, presumably extracting a portion of the enhanced interstitial HCO32 into shell material (e.g., Green et al., 1993). S CO2 production during each cruise was estimated from sediment incubation experiments as described earlier. As can be seen in Figure 11, S CO2 production (open circles) from 0–4 cm correlates almost perfectly with calcite and aragonite saturation state. The highest S CO2 production is associated with supersaturated conditions, shallow O2 penetration, and warmest temperatures (cruises 1, and 9–11), J-S CO2/J-O2 ratios greater than 0.77 during cruise 1 and cruises 8–11 result from release of anaerobically produced S CO2. In addition, the precipitation and temporary storage of at least some of the reduced elements produced (i.e., Fe21 , S2 ) will also act to increase J-S CO2/J-O2 ratios. During this study, water temperature increased rapidly along with the  ux of C-organic to sediments (Table 1). This increase is one reason for increased S CO2 production (CO322 ) and the associated increase in pore water supersaturation. If between season increasing production of S CO2 resulted solely from increasing temperature, then a natural log transformation of S CO2 production rate from 0–2 cm and 2–4 cm (Fig. 12a) versus (1/T) should give equal slopes in the two depth intervals (proportional to apparent activation energy of reaction). However, the supply of organic-C substrate was also time dependent and was deposited at the sediment-water interface. Although reworking by marine organisms transported this reactive substrate to depth, the bulk of the freshly deposited

1116

Journal of Marine Research

[56, 5

Figure 12. (a) S CO2 production rates from 0–2, 2–4 and 0–4 cm sediment interval incubations during PULSE cruises 1–11 and (b) Apparent activation energies (kJ mol2 1) calculated from log transformations of the 0–2 cm and 2–4 cm S CO2 production rates during PUSLE cruises 1–3 and 9–11, respectively.

reactive carbon was located in the upper 0–2 cm (Gerino et al., 1998). The effects of the reactive substrate addition and of temperature are indicated by slope differences between the 0–2 and 2–4 cm depth intervals during periods of higher carbon  ux (Fig. 12b). In the absence of unsteady inputs of C, it is more common to observe lower apparent activation energies in zones of relatively higher net remineralization rates (Westrich and Berner, 1988; Middleburg et al., 1996). In addition to metabolic reaction rates, the solubility, and therefore K8sp, of both calcite and aragonite are controlled by temperature. The saturation CO322 concentration increases

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1117

with decreasing temperature and decreases with increasing temperature. K8sp for calcite is plotted along with temperature during each cruise in Figure 11 (inset). Therefore, during winter when S CO2 production is lowest, the concentration of CO322 necessary to reach equilibrium is highest. The opposite is true during summer months. This pattern of oscillating solubility due to the dependence of K8sp on temperature is unique to nearshore regions prone to radical changes in temperature. Hence, change from undersaturation to supersaturation due to increased production of S CO2 (anaerobic) results from both the effect of increasing temperature on mineral solubility and the changes in the deposition of reactive carbon to sediments. Maintenance of periods of under- and supersaturation also imply that dissolution or precipitation rates respectively, are slower than acid or base production. e. Effects of CaCO3 dissolution on S CO2 production pools An independent estimate of the S CO2 contributed to sediment pore water by the dissolution of CaCO3 during periods of pore water undersaturation can be made by applying a single-layer diffusion-reaction model to results from the sediment incubations (Berner, 1980). In this case, NH41 production rate measurements calculated from sediment incubations (0–2 cm) during cruises 4–7 (pore water undersaturation) were converted to theoretical S CO2 production rates, assuming C/N production ratios determined from incubations during cruises 1 and 11. Pore water from cruises 1 and 11 were supersaturated with respect to calcite, and to a large extent, aragonite (Fig. 2). During these cruises S CO2 release is assumed to come solely from organic carbon diagenesis. Hence, NH41 production rates from sediment incubations during those times should yield accurate estimates of S CO2 production rates from C-org remineralization alone. C/N production ratios of 7 and 7.9 were determined from sediment incubations (0–2 cm) during cruises 1 and 11, respectively (corrected for reversible adsorption).An average C/N value of 7.5 was used to calculate theoretical S CO2 production from NH41 production during cruises 4–7. This C/N ratio is consistent with previously reported C/N ratios found in pore waters of nearshore regions with terrigenous organic carbon inputs (Muller, 1977, Anderson et al., 1986), and with values estimated from stoichiometric analysis of pore water proŽ les (Cruise 1 and 11; Green, 1996). With the above assumptions, predicted pore water proŽ les of S CO2 were calculated for cruises 1, 4–7, and 11 using the following equation: ­ C/­ t 5

05 x5

Ds(­ 2C/­ x2) 1 0, C 5

CT

dC/dx ® 0

Roe2

a x

(7a) (7b) (7c)

where: C 5 pore water S CO2 concentration, x 5 space coordinate with origin Ž xed at the sediment surface and positive into sediment, Ds 5 bulk sediment diffusion coefficient, Ro 5 S CO2 production rate at the sediment-water interface, and a 5 the S CO2 production

Journal of Marine Research

1118

[56, 5

rate attenuation coefficient. Physical advection was ignored due to low sedimentation rates in Long Island Sound. Local steady state was assumed because the spatial scale of interest (7 cm) is small, and the processes being measured are relatively fast compared to the time resolution of the study (, 2 week sampling interval). Ro values for cruises 1, 4–7, and 11 were calculated by taking the integral of the measured NH41 production (Rm) rate from 0–2 cm, Rm 5

e 2 1

2 0

Roe2

a x

dx

(8)

and solving for Ro so that, Ro 5

Rm 3

2a /(1 2

e2

a 2

).

(9)

The attenuation coefficient, a , was estimated by taking the ratio of the integral of the NH41 production rate measurement from 0–2 cm and 2–4 cm and adjusting the value of a until, Rm0–2cm/Rm2–4cm 5

(1 2

e2

2a

)/(e2

2a

2

e2

4a

)

(10)

where Rm0–2cm and Rm2–4cm are the measured NH41 production rates from 0–2 and 2–4 cm, corrected for adsorption, in units of µmol NH41 cm2 3 sed2 1 d2 1. Values of Ro for cruises 1, 4–7, and 11 are 0.38, 0.08, 0.19, 0.19, 0.1, and 1.39 respectively, with corresponding a values of 0.95, 0.51, 0.86, 0.74, 0.57, and 0.65. These Ro values were converted to equivalent S CO2 production rates using C/N 5 7.5 (a is assumed to be identical for S CO2 and NH41 ). Measured pore water S CO2 concentrations and Ž ts of pore water S CO2 to the model equations 7a–c are shown in Figure 13. There is good agreement between model and measured S CO2 data during cruises 1 and 11 suggesting that methods used to calculate Ro, a , and diffusive transport are reasonable.Although no correction for biogenic transport was necessary during cruises 4–7, model calculations consistently underestimate measured S CO2 pore water concentrations by 20–50%. Loss of NH 41 through nitriŽ cationdenitriŽ cation would increase the discrepancy. Dissolution of CaCO3 during cruises 4–7 would result in higher pore water S CO2 concentrations than model predicted values. Assuming model-predicted S CO2 proŽ les accurately describe S CO2 produced from organic carbon decomposition, and that dissolution is the only other major source of S CO2, then pore water S CO2 during periods of undersaturation with respect to calcite and aragonite are 20–50% higher due to contribution of CO322 from dissolution (Fig. 13b). These contribution estimates agree well with predictions based on C/N stoichiometric analysis of pore water proŽ les which indicate net C/N ratios nearly a factor of 2 larger during middle cruises when pore waters are undersaturated (Fig. 14), although nitriŽ cation-dentiŽ cation may accentuate these differences. The estimated carbonate dissolution S CO2 contribution, also agrees well with direct measurements of Ca21  uxes, as well as estimates of S CO2 release based on the dissolution

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1119

Figure 13. (a) Measured (closed circles) and modeled (dashed line) pore water S CO2 concentrations during PULSE cruises 1, 4–7, and 11 and (b) the predicted contribution to the pore water S CO2 pool from CaCO3 dissolutionbased on the model and measured S CO2 concentrationdiscrepancies.

of foraminifera during winter (Green et al., 1993). These estimates are also comparable to predictions by McNichol et al. (1988), who determined that roughly 25% of the S CO2  ux came from dissolution of CaCO3 within coastal sediments from Buzzards Bay, MA. 6. Conclusions Pore water saturation states with respect to calcite and aragonite minerals in Long Island Sound sediments follow a regular seasonal pattern, with saturated and near-saturated

1120

Journal of Marine Research

[56, 5

Figure 14. Linear regression analysis of pore water S CO2 vs. NH41 concentrationswhen pore waters were saturated and/or supersaturated (solid circles) and when pore waters were undersaturated (open circles) with respect to calcite and aragonite mineral phases. S CO2 concentrationsare given as D -S CO2 calculated from overlying water S CO2 values. There was no detectable NH41 in overlying water during any cruise.

conditions during late fall, undersaturation during winter, and supersaturation during late spring/early summer. Extensive dissolution of calcium carbonate having variable mineralogies is evident from the large increases of Ca21 , Sr21 , and F2 pore water inventories during periods of undersaturation. Oscillating calcite/aragonite saturation states in temperate regions such as LIS result from the interplay between temperature control on the saturation CO322 concentration, the oxidation of reduced metabolites which acts to lower CO322 concentrations, and S CO2 production from aerobic and anaerobic metabolism which vary both as a function of temperature as well as supply of reactive carbon. Supersaturation occurs during warmer periods when the lower saturation CO322 concentration for both calcite and aragonite is exceeded by higher anaerobic S CO2 (HCO32 ) production rates. Deposition of organic matter produced during the spring bloom accelerates the seasonal progression of pore waters to supersaturation, as S CO2 production is both temperature and substrate controlled. Undersaturation occurs during winter time when lower rates of S CO2 production, and oxidation of previously formed reduced minerals such as FeS and FeS2, lower CO322 below the saturation CO322 concentrations.

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1121

Acknowledgments. We thank members of the PULSE research team: Josephine Aller, Magali Gerino, Ian Stupakoff, David Hirschsberg, and Kristen Sullivan for their assistance in the Ž eld and laboratory. In addition, we are particularly indebted to the captain and crew of the R/V Onrust for getting us to where we needed to go. Margaret Lima, Rebecca Yingst and Steve Lucatuorto provided help in the laboratory. Josephine Aller, Cindy Lee, Bob Cerrato, Kirk Cochran, and Clare Reimers provided reviews of an early version of this manuscript. Lastly, we thank Rick Jahnke and John Morse for their thorough reviews of the completed manuscript. This research was supported by Department of Energy grant DEFG0292ER61464 as part of the initial Ocean Margins Program. REFERENCES Aller, R. C. 1977. The in uence of macrobenthos on chemical diagenesis of marine sediments, Ph.D. dissertation,Yale University, New Haven, CT, 600 pp. ——1980. Diagenetic processes near the sediment-water interface of Long Island Sound. II. Fe and Mn. Adv. Geophy., 22, 351–415. ——1982. Carbonate dissolution in nearshore terrigenous muds: the role of physical and biological reworking. J. Geol., 90, 79–95. Aller, R. C. and J. K. Cochran. 1976. 234Th/238U Disequilibrium in near-shore sediment: Particle reworking and diagenetic time scales. Earth Planet. Sci. Lett., 29, 37–50. Aller, R. C. and J. E. Mackin. 1989. Open-incubation, diffusion methods for measuring solute reaction rates in sediments. J. Mar. Res., 47, 411–440. Anderson, L. G., P. O. J. Hall, A. Iverfeldt, M. M. Rutgers van der Loeff, B. Sundby and S. F. G. Westerlund. 1986. Benthic respiration measured by total carbonate production. Limnol. Oceanogr., 31, 319–329. Baker, P. A., J. M. Gieskes and H. ElderŽ eld. 1982. Diagenesis of carbonates in deep-sea sediments: evidence from Sr/Ca ratios and interstitial dissolved Sr21 data. J. Sed. Petrol., 52, 0071–0082. Benninger, L. K., R. C. Aller, J. K. Cochran and K. K. Turekian. 1979. Effects of biological sediment mixing on the 210Pb chronologyand trace metal distribution in a Long Island Sound sediment core. Earth Planet. Sci. Lett., 43, 241–259. Berner, R. A. 1980. Early Diagenesis-ATheoreticalApproach. Princeton University Press, Princeton, NJ, 241 pp. Berner, R. A. and J. W. Morse. 1974. Dissolution kinetics of calcium carbonate in sea water. IV. Theory of calcite dissolution.Am. J. Sci., 274, 108–134. Carpenter, R. 1969. Factors controlling the marine geochemistry of  uorine. Geochim. Cosmochim. Acta, 33, 1153–1167. Cline, J. D. 1969. Spectrophotometric determination of hydrogen sulŽ de in natural waters. Limnol. Oceanogr., 14, 454–458. Conover, S. A. M. 1956. Oceanography of Long Island Sound, 1952–1954. IV. Phytoplankton. Bull. Bingham Oceanogr. Coll., 15, 62–112. Edmond, J. M. 1970. High precision determination of titration alkalinity and total carbon dioxide content of seawater by potentiometric titration. Deep-Sea Res., 17, 737–750. Emerson, S. and M. Bender. 1981. Carbon  uxes at the sediment-water interface of the deep-sea: calcium carbonate preservation. J Mar. Res., 39, 139–162. Frant, M. S. and J. W. Ross, Jr. 1966. Electrode for sensing  uoride ion activity in solution. Science, 154, 1553–1555. Froelich, P. N., G. P. Klinkhammer, M. L. Bender, N. A. Luedtke, G. R. Heath, D. Cullen, P. Dauphin, D. Hammond, B. Hartman and Val Maynard. 1979. Early oxidation of organic matter in pelagic sediments of the eastern equatorial Atlantic: suboxic diagenesis. Geochim. Cosmochim. Acta, 43, 1075–1090.

1122

Journal of Marine Research

[56, 5

Gerino, M., R. C. Aller, C. Lee, J. K. Cochran, J. Y. Aller, M. A. Green and D. Hirschberg 1998. Comparison of different tracers and methods used to quantify bioturbation during a spring bloom: 234-Thorium, luminophores, and chlorophyll-a. Estuar. Coast. Shelf Sci., 46, 531–547. Gieskes, J. M. 1974. The alkalinity-totalcarbon dioxide system in seawater, in The Sea, Vol. 5, Ch. 3, E. D. Goldberg, ed., J. Wiley & Sons, 123–151. Goldhaber, M. B., R. C. Aller, J. K. Cochran, J. K. Rosenfeld, C. S. Martens and R. A. Berner. 1977. Sulphate reduction, diffusion and bioturbation in Long Island Sound sediments: Report of the FOAM group. Am. J. Sci., 277, 193–237. Green, M. A. 1996. The biogeochemicalmechanisms driving calcite and aragonite saturation states in Long Island Sound sediments: The effects on juvenile bivalves, benthic foraminifera, and carbonate debris preservation, Ph.D. dissertation, State University of New York, Stony Brook, 257 pp. Green, M. A. and R. C. Aller. 1998. Rapid, short term changes in calcite and aragonite saturation states during winter in Long Island Sound and the associated  uxes and dissolved Ca21 , Mg21 , Sr21 , and F2 . (In preparation). Green, M. A., R. C. Aller and J. Y. Aller. 1992. Experimental evaluation of the in uences of biogenic reworking on carbonate preservation in nearshore sediments. Mar. Geol., 107, 175–181. ——1993. Carbonate dissolution and temporal abundances of foraminifera in Long Island Sound sediments. Limnol. Oceanogr., 38, 331–345. ——1998. The in uence of carbonate dissolution on the survival of shell-bearing meiobenthos in nearshore sediments. Limnol. Oceanogr., 43, 18–28. Hall, P. O. J. and R. C. Aller. 1992. Rapid, small-volume,  ow injection analysis for S CO2 and NH41 in marine and freshwaters. Limnol. Oceanogr., 37, 1113–1119. Jahnke, R. A., S. R. Emerson, K. K. Roe and W. C. Burnett. 1983. The present day formation of apatite in the Mexican continental margin sediments. Geochim. Cosmochim. Acta, 47, 259–266. Li, Y. M. and S. Gregory. 1974. Diffusion of ions in seawater and in deep-sea sediments. Geochim. Cosmochim. Acta, 38, 703–614. McNichol, A. P., C. Lee and E. R. M. Druffel. 1988. Carbon cycling in coastal sediments: 1. A quantitative estimate of the remineralization of organic carbon in the sediments of Buzzards Bay, MA. Geochim. Cosmochim. Acta, 52, 1531–1543. Mehrbach, C., C. H. Culberson, J. E. Hawley and R. M. Pytkowicz. 1974. Measurements of the apparent dissociation constants of carbonic acid in seawater at atmospheric pressure. Limnol. Oceanogr., 18, 897–907. Middleburg, J. J., G. Klaver, J. Nieuwenhuize, A. Wielemaker, W. de Haas, T. Vlug, J. F. W. A. Vander Nat. 1996. Organic matter mineralization in intertidal sediments along an estuarine gradient. Mar. Ecol. Prog. Ser., 132, 157–168. Millero, F. J. 1979. The thermodynamicsof the carbonate system in seawater. Geochim. Cosmochim. Acta, 43, 1651–1661. Morse, J. W. 1974. Dissolution kinetics of calcium carbonate in sea water. V: Effects of natural inhibitor and the position of the chemical lysocline.Am. J. Sci. 274, 638–647. ——1983. The kinectics of calcium carbonate dissolution and precipitation. Rev. Mineral., 11, 227–264. Morse, J. W. and R. A. Berner. 1972. Dissolution kinetics of calcium carbonate in sea water: II. A kineic origin for the lysocline.Am. J. Sci., 272, 840–851. Morse, J. W. and F. T. Mackenzie. 1990. Geochemistry of sedimentary carbonates, Elsevier Science Publishers, NY, 707 pp. Muller, P. J. 1977. C/N ratios in PaciŽ c deep-sea sediments: Effects of inorganic ammonium and organic nitrogen compounds sorbed by clays. Geochim. Cosmochim. Acta, 41, 765–776.

1998]

Green & Aller: Seasonal patterns of carbonate diagenesis

1123

Plummer, L. N., T. M. L. Wigley and D. L. Parkhurst. 1978. CO2-water systems at 5 to 60°C and 0.0 to 1.0 atm CO2. Am. J. Sci., 278, 179–216. Reaves, C. M. 1986. Organic matter metabolizability and calcium carbonate dissolution in nearshore marine muds. J. Sed. Petrol., 56, 486–494. Rude, P. D. and R. C. Aller. 1991. Fluorine mobility during early diagenesis of carbonate sediment: An indicator of mineral transformations.Geochim. Cosmochim. Acta, 55, 2491–2509. ——1993. The in uence of Mg21 on the adsorption of  uoride by hydrous oxides in seawater.Am. J. Sci., 293, 1–24. Ruttenberg, K. C. 1990. Diagenesis and burial of phosphorous in marine sediments: Implications for the marine phosphorus budget, Ph.D. dissertation,Yale University, New Haven, CT. Ruttenberg, K. C. and R. A. Berner. 1993. Authigenic apatite formation and burial in sediments from non-upwelling, continential margin sediments. Geochim. Cosmochim. Acta, 57, 991–1007. Sjoberg, E. L. 1978. Kinetics and mechanism of calcite dissolution in ageous solutions at low temperatures.Acta Univ. Stockholm, Contrib. Geol., 332, 1–92. Stookey, L. L. 1970. Ferrozine-a new spectrophotometric reagent for iron. Anal. Chem., 42, 779–781. Walter, L. M. and E. A. Burton. 1986. The effect of orthophosphate on carbonate mineral dissolution rates in seawater. Chem. Geol., 56, 313–323. ——1990. Dissolution of recent platform carbonate sediments in marine pore  uids. Am. J. Sci., 290, 601–643. Walter, L. M. and J. W. Morse. 1984. Magnesian calcite stabilities: A reevaluation. Geochim Cosmochim. Acta, 48, 1059–1069. ——1985. The dissolution kinetics of shallow marine carbonates in seawater: A laboratory study. Geochim. Cosmochim. Acta, 49, 1503–1513. Westrich, J. T. and R. A. Berner. 1988. The effect of temperature on rates of sulfate reduction in marine sediments. Geomicrob. J., 6, 99–117.

Received: 9 December, 1997; revised: 9 June, 1998.