Jensen et al. in press

Limnology and Oceanography (06/17/2009)

Title: Phosphorus release with carbonate dissolution coupled to sulfide oxidation in Florida

Bay seagrass sediments

Henning S.Jensen,1,* Ole I. Nielsen,2 Marguerite S. Koch,2 Inmaculada de Vicente,1, 3

1 Institute of Biology, University of Southern Denmark, Odense, Denmark.

2 Aquatic Plant Ecology Lab, Biological Sciences Department, Florida Atlantic University, Boca

Raton, Florida.

3 Present address: Water Research Institute, University of Granada, Spain.

* Corresponding author: hsj@biology.sdu.dk.

Running head: Phosphorus release in seagrass sediment

2

ACKNOWLEDGMENTS

We thank Stephanie Schopmeyer and students in the Florida Atlantic University, Aquatic

Plant Ecology Lab for their assistance in sampling. The manuscript was greatly improved by

qualified reviews from K. J. McGlathery and one anonymous reviewer. The research was funded by

the Critical Ecosystem Studies Initiative (CESI) program of the U.S. Department of Interior (DOI)

through Everglades National Park (Cooperative Agreement H5000 02 0409, J5284 05 0047) and

supplemented by the Danish Natural Science Research Council grants 21030013, 21040034, and

21020463 and by the project OAPN (Organismo Autónomo Parques Nacionales) 129 A/2003.

Everglades National Park’s Florida Bay Interagency Science Center in Key Largo, Florida, and the

South Florida Water Management District, West Palm Beach, Florida, provided field logistical

support.

3

Abstract

We hypothesized that CaCO3 dissolution, coupled to sulfide oxidation, is an important

mechanism by which solid-phase inorganic P (iP) becomes available to seagrass in tropical

carbonate sediments. To examine this supposition, we measured field sulfate reduction rates and

simulated the acidity (10-50 µmol H2SO4 cm-3 sediment) generated by subsequent sulfide oxidation

from high (western) and low (eastern) total P (TP) sediments in Florida Bay. Dissolution

experiments were conducted using sediment slurries at field pH (porewater pH ~6.5 to 7.5). While

CaCO3 dissolution (max 1.3% of sediment dry weight) was evidenced at all sites by leaching of

Ca2+ into slurries, at 3 of the 4 sites, PO43- was primarily recovered in the sediment exchangeable

pools, and at a lower P:Ca ratio than observed in source sediments. In contrast, no PO43- was

recovered from an eastern bay site with the lowest TP and finest-grained sediments, suggesting that

PO43- was either tightly adsorbed or incorporated into the carbonate matrix post-acidification. The

potential for tight PO43- re-sorption by sediment from the low-TP site was also supported by the

rapid rate of isotopic exchange of 33PO43- into the sediment matrix. These adsorptive and

incorporation processes may explain the low PO43- in porewaters in the eastern versus western

regions of Florida Bay, even upon dissolution of carbonates. Carbonate dissolution coupled to

sulfide oxidation could potentially provide 1.8 and 23.5 µmol iP m-2 d-1 at eastern and western bay

sites, respectively, meeting 5% of eastern and 29% of western bay seagrass (Thalassia testudinum)

P requirements.

4

Introduction

Seagrass primary production in carbonate environments is frequently limited by phosphorus

(P) availability (Short et al. 1990; Fourqurean et al. 1992a,b) resulting from low terrestrial P loads

and high sediment affinity for phosphate (PO43-) in these systems. Phosphorus in carbonate

sediments can be strongly bound to the solid phase carbonate matrix, e.g., as carbonate fluoroapatite

(CFA) and other P- and fluoride (F-)-bearing carbonate phases (Kitano et al. 1978; Rude and Aller

1991; Millero et al. 2001), and also adsorbed onto carbonate surfaces (Jensen et al. 1998; Koch et

al. 2001; Zhang et al. 2004). Jensen et al. (1998) observed that both inorganic P (iP) and organic P

were being mobilized from the solid-phase pools by dissolution of the carbonate matrix in the

rooting depth of the tropical seagrass Thalassia testudinum. The same constituents, F-, PO43- and

dissolved organic P, were observed to be elevated in seagrass porewater and released when

carbonate sediments were gradually dissolved using sequential chemical extraction. The study by

Jensen et al. (1998) did not identify any mechanism behind the dissolution process. Ku et al. (1999)

and Burdige and Zimmerman (2002) proposed that O2 released from T. testudinum roots enhances

sulfide oxidation, as well as aerobic mineralization of organic matter in seagrass sediment. As both

of these metabolic oxidation processes generate acidity, they have the potential to dissolve

carbonate sediments, and release sediment-bound P in close proximity to the roots (Burdige and

Zimmerman 2002). Acid generation in seagrass carbonate sediments has been implied by low pH

measured in sediments with high (mmol L-1) porewater sulfide levels in Florida Bay (Ku et al.

1999; Koch et al. 2007) and lower pH in seagrass beds than in bare sediments in the Bahamas

(Morse et al. 1987) and Florida Bay (Long et al. 2008). Another possible process that can release P

from carbonates is the exchangeability of iP with organic acids exuded from growing roots as

shown by Long et al. (2008). The release of organic acids may also stimulate aerobic microbial

metabolism or enhance the coupled sulfate reduction (SR) – sulfide oxidation.

5

Sulfate reduction to H2S and reoxidation by O2 to SO42- is an important biogeochemical

redox process in Florida Bay that has a high potential for acid generation (2H+ generated per H2S

oxidized) and thus for CaCO3 dissolution. Ku et al. (1999) showed that the porewater SO42-:Cl- ratio

was nearly uniform in the upper 30 cm of vegetated sediment in Florida Bay, while the enriched

isotopic signature of 18O and 34S in SO42- implied a rapid recycling of sulfide generated by SR. At

the same time the Ca2+ concentration in the porewater was elevated and model calculations

suggested that carbonate dissolution was stoichiometrically coupled to sulfide oxidation by a 1:2

relationship between H2S oxidized and Ca2+ mobilized to the porewater, assuming formation of

HCO3- from CaCO3 dissolution. Thus, when O2 is transported to the root zone, SR and CaCO3

dissolution may be coupled and, in principle, all sulfide produced in iron-poor carbonate sediments

has the potential to be re-oxidized given sufficient oxygen availability (Ku et al. 1999). Using the

stoichiometry of 1:2, Burdige and Zimmerman (2002) calculated that published rates of net

carbonate dissolution of 4-11 mmol m-2 d-1 in sediments from Florida Bay (Rude and Aller 1991)

could be accounted for by sulfide oxidation, requiring only ~15% of the daily net O2 production of

T. testudinum. Lee and Dunton (2000) also showed that diurnal porewater sulfide varies in seagrass

(T. testudinum) sediments indicating that sufficient O2 can be provided during the day to re-oxidize

most of the sulfide generated at night via SR. Based on these studies, there is indirect evidence that

SR in seagrass beds, and subsequent sulfide oxidation and acid generation, may promote carbonate

dissolution (Walter et al. 2007). Hu and Burdige (2007) suggested that organic matter oxidation

(CO2 production), as enhanced by O2 release from roots, may be more important than sulfide

oxidation for generating the acidity that drives carbonate dissolution in sediments from seagrass

beds in the Bahamas. This CO2 from oxidative respiration dissolves and reacts with water to

produce carbonic acid (H2CO3) which readily dissociates to bicarbonate (HCO3-) and a hydrogen

ion (H+) producing acidity. In a recent modeling study, Burdige et al. (2008) simulated both

6

advective O2 transport and seagrass mediated O2 transport, and concluded that carbonate dissolution

is elevated in vegetated as compared to bare sediments due to enhanced metabolism mediated by

seagrass O2 flux. The different importance of the two biogeochemical reactions for acid formation

in the Bahamas and in Florida Bay can be partly due to the larger grain size of sediment in the

Bahamas site (200-800 µm, Burdige et al. 2008) as compared with the small grain size (<63 µm) of

the Florida Bay site and lower organic matter in highly oligotrophic sediments in the Bahamas

Islands. The larger grain size and lower O2 demand may facilitate porewater advection and O2

transport to an extent where SR is being suppressed.

Since sulfide accumulation is evident in porewater from Florida Bay sites (Walter et al

2007, Koch et al. 2007), and there is isotopic evidence that there is close coupling between SR and

sulfide oxidation (Ku et al. 1999), we focused the present study on the acidity generated from

sulfide oxidation and its effects on P availability. We hypothesized that H2SO4 generated from

sulfide oxidation is neutralized by carbonate dissolution with concomitant release of solid phase iP

and F- with modest changes in porewater pH. By adding H2SO4 to sediment slurries we examined

carbonate sediment dissolution at four sites in Florida Bay at pH levels comparable to those

observed in the field (6.6 – 6.9; Koch et al. 2007). We quantified PO43- mobilization to slurry water,

as well as resorption to surface-bound P pools following CaCO3 dissolution. Fluoride release was

also tracked post-acidification, as its release is considered an indicator of the breakdown of the

tightly bound P associated with the carbonate matrices (e.g., CFA; Jensen et al. 1998). Phosphate

re-adsorption was further elucidated by adding 33PO43- to sediment slurries and following the rates

at which they became homogeneously labeled in the exchangeable and matrix-bound P pools. By

measuring site-specific sulfate reduction rates (SRR) and assuming complete recycling of sulfide to

H2SO4 we calculated a potential iP mobilization rate for eastern and western sites from SR-

7

oxidation coupling and compare this P flux to P release rates based on literature estimates of

carbonate dissolution.

Methods

Sampling sites

Florida Bay is a shallow (< 2 m) semi-enclosed estuarine lagoon at the southern tip of the

Florida peninsula (Fig. 1). The sediment is predominantly aragonite (50-60%) and high-Mg calcite

(30-40%) of biogenic origin, generated internally by primarily calcareous algae and in situ

precipitation and composed of a very low silicate fraction (Rude and Aller 1991; Lyons et al. 2004;

Walter et al. 2007) and low Fe (Chambers et al. 2001; Zhang et al. 2004). Extensive seagrass

meadows of the tropical seagrass T. testudinum dominate primary production in the bay decreasing

in biomass and primary production from southwest to northeast as a consequence of P availability

(Fourqurean and Zieman 2002). This trophic gradient is well established by sediment P

concentrations in the fine fraction of the sediment (Zhang et al. 2004) and in the bulk sediment

(Koch et al. 2001, this study), sediment P adsorption isotherms (Nielsen et al. 2006), and leaf C:P

ratios (Fourqurean et al. 1992a). Sites chosen for this study (Fig. 1) include three sites in the

western bay, Buchanan Bank (BB; 24o54’55” N; 80o45’24” W), Green Mangrove Key (GMK;

24o54’55” N; 80o45’24” W) and Rabbit Key (RK; 24o58’50” N; 80o50’11” W), high sediment P

sites, and two sites in the northeastern bay, Eagle Key (EK; 25o09’32” N; 80o34’58” W) and Black

Betsy Key (BLBK; 25o08’23” N; 80o38’51” W), low sediment P sites, all dominated by T.

testudinum, although, with varying shoot biomass of 97, 41, 226, and 153 g dry wt-1 m-2 at EK,

BLBK, BB, and RK, respectively (Koch et al. 2007; Rosch and Koch 2009). The more enriched

western sites are also characterized as having relatively higher porewater sulfide concentrations (4-6

8

mmol L-1) relative to eastern sites (≤ 2 mmol L-1) (Barber and Carlson 1993; Carlson et al. 1994;

Koch et al. 2007).

P mobilization experiment

Sediment (2-15 cm) was collected in seagrasses at BB, RK, BLBK, and EK sites via coring

(March 2005; 16 cm diameter cores) and subsequently sieved (2 mm) to remove large rhizomes and

roots.The wet sediment (20 g) was added to ambient seawater (80 mL) from the sites. The resulting

sediment slurries were placed on a shaker table and pre-incubated for 16 h prior to the start of the

experiment in order to obtain equilibrium between PO43- in slurry water and PO4

3- adsorbed on

particle surfaces. To examine sediment carbonate dissolution and subsequent release of iP as pH

was lowered through acidification, four different amounts of sulfuric acid: 0, 150, 400, and 800

µmol H2SO4 (~0-50 µmol cm-3 sediment; n=3) were added to sediment from each site. The

maximum H2SO4 addition was chosen so that the observed slurry water pH would represent the

lower values observed in the field. The required amount was found from pre-runs of the experiment.

Since the experimental setup could cause other changes as a result of oxidation of porewater

sulfide or stimulation of aerobic mineralization, the ‘zero addition’ treatment was applied as a

control in the acid addition experiments. Sulfuric acid additions to the carbonate sediments in slurry

generated CO2, as observed by gas bubble formation, suggesting that some CO32- was titrated to

CO2 rather than HCO3-. Under these conditions, 2H+ would be consumed per mole of CaCO3,

probably accounting for a 1:1 relationship between acid treatment and Ca2+ released (see results).

Seawater alkalinity was measured by titration and the acid consumed by changing alkalinity

in the 80 mL seawater was taken into account when calculating the effective acid addition per g dry

wt sediment. Release of PO43-, F-, and Ca2+, as well as pH changes in the slurries, were monitored

throughout the experiment at time 0, 1, and 6 h. The volumes removed by sampling were taken into

9

account when calculating fluxes. The pH of the slurry was measured directly, while PO43-, F-, and

Ca2+ were measured after filtering (GF/F) a subsample (15 mL) of the sediment-water slurry.

Phosphate was determined using standard colorimetric methods (Koroleff 1983) with a detection

limit of 20 nmol L-1. Inorganic fluoride was determined using a combined F- electrode (Orion 96-

09) with 1:1 addition of low-level total ionic strength adjusting buffer (LL TISAB). LL TISAB was

prepared by adding 57 g NaCl and 57 g glacial acetic acid to 550 ml DI water, adjusting pH to 5.65-

6.0 with strong NaOH, and making up volume to 1 L with DI water. The resolution at ambient

seawater F- concentration was ±1.8 μmol L-1. Dissolved Ca2+ was determined by flame atomic

absorbance spectroscopy (Perkin Elmer) also using standard methods. Due to constraints in

volumes available for analyses, we omitted the measurement of organic P (oP) leaching from the

sediment matrix, although, increments in dissolved oP may be an additional indicator of matrix

dissolution (Jensen et al. 1998).

After acid treatments, slurry water was decanted and sequential extractions were conducted

with the wet sediment to quantify P and F that had adsorbed onto carbonates and not recovered in

the dissolved phase. The extraction procedure was a modification of the first two steps of the

extraction scheme used for carbonate sediment by Jensen et al. (1998). In the first step, the sediment

was transferred into 60 mL centrifuge tubes to which 30 mL of 1 mol L-1 MgCl2 was added and put

on a shaker table for 1 h. After centrifugation and decanting, the sediment was washed in 30 mL 0.5

mol L-1 NaCl. The NaCl wash supernatant was added to the MgCl2 supernatant and brought up to

volume with H2O (100 mL). The second sequential step was a 1 h sediment extraction with a

Bicarbonate-Dithionite (BD) reagent (0.11 mol L-1 Na-dithionite and 0.11 mol L-1 NaHCO3)

followed by two BD washes and one NaCl wash. The wash supernatants were again brought to

volume with H2O (200 mL). The first extraction step extracts loosely adsorbed iP while the second

step extracts both Fe-bound and more tightly-bound iP on the surface of carbonates (Jensen et al.

10

1998). In siliclastic temperate sediments MgCl2 extracts loosely adsorbed surface bound iP with BD

extraction, considered highly specific for Fe-bound P (Ruttenberg 1992; Jensen and Thamdrup

1993). However, Jensen et al. (1998) showed that in carbonate sediment BD extracts both Fe and

surface-bound iP without concurrent release of Ca2+. Zhang et al. (2004) also found no correlation

between reactive Fe oxides and BD-extractable iP in the fine-grained fraction of Florida Bay

sediments. PO43- and F- concentrations in the extracts were measured as stated above using standard

curves prepared in the specific matrix extract solutions. While the sediment treatments were similar

on a sediment volume basis, differences in the bulk densities across sites resulted in slightly

variable amounts of MgCl2 and BD amendments per g dry wt sediment. Since eastern sites has

higher bulk densities (see results) this may cause a small, relative, underestimation of the size of

surface-bound P-pools at eastern bay sites.

Sediment characteristics

Porosity, dry:wet wt ratios and % of small grain size (<63 μm) were determined on sediment

from all sites. All parameters except porosity were determined on sediment that was passed through

a 2 mm mesh after removal of larger seagrass debris and seagrass tissue. Total Ca, F, and P

concentrations were quantified by combusting the dried sediment (550°C, 3 h) and subsequently

dissolving the pellet in 1 mol L-1 HCl (120°C, 1 hour). Dissolved Ca2+, F-, and PO43- were analyzed

as above with the exception of F- measurements which required neutralization (NaOH) prior to

analyses. The solid inorganic P fraction in the bulk sediment was determined by dissolving the

sediment in cold 0.5 mol L-1 HCl and analyzing the extract for PO43-.

11

Sediment P-pool labeling experiment

We used a unique radiotracer technique to examine sediment P exchange in eastern fine-

grained P-poor sediment from EK and western coarser P-rich sediment from GMK. This technique

is based on the principle that 33PO43-, when added to sediment slurries in trace amounts under

conditions of equilibrium between surface adsorbed and slurry water 31PO43-, will undergo isotopic

exchange with any surface-bound or solid fraction 31PO43- reservoir. The exchange rates will depend

on the exchangeability potential of a particular iP pool (Jensen et al. 2005). After wet sediment (20

g) was kept in suspension on a shaker table for 16 h in P-free artificial seawater (100 mL), carrier-

free 33PO43- (~42,000 Bq) was added to the slurry. The activity of 33PO4

3- in the slurry water was

followed for 4 h, after which the distribution of the tracer in the exchangeable and matrix-bound

fractions (MgCl2-extractable, BD-extractable, and HCl-extractable P, respectively) was determined

by sequential extraction at time 4.66, 7.66, 33, and 55 h. 33PO43- in solution was measured on a

GF/F filtrate while solid pool tracer distribution was measured on the particles retained on the filter.

The experiments were run in triplicates. Radioactivity was measured on a Beckman LS 6500 Liquid

Scintillation Counter using standard techniques and correcting for background, decay and

quenching with known activity.

Sulfate reduction rates (SRR)

SRR were determined on seagrass sediment cores (diameter 3.5 cm; n=4) from EK, BLBK,

BB, and RK in October 2005 and July 2006 to relate dissolution of carbonates to potential SR and

coupled sulfide oxidation. Intact cores from each site were injected with 35SO42- (37,000 Bq in 2

µL) through silicon stopper holes at 2 cm intervals down to 14-18 cm depth (after Jørgensen 1978).

Cores were incubated under anoxic conditions for 2 h at in situ temperature (26oC October and 31oC

July). Post-incubations, cores were carefully sliced into 2 cm intervals and immediately transferred

12

into zinc acetate (10 mL of 1 mol L-1) for sulfide fixation. Subsequently, the sediment was

centrifuged, decanted and an additional 10 mL of zinc acetate (1 mol L-1) added in order to bind

remaining sulfides and allow removal of excess 35SO42-. Reduced 35S-pools were retrieved using a

two-step distillation method described by Fossing and Jørgensen (1989) and radioactivity was

immediately counted on a scintillation counter. SRR were calculated as the percent recovery of

radiolabelled sulfur in the acid volatile sulfide (AVS) and chromium reducible sulfide (CRS) pools.

Porewater sulfate was measured using a separate non-radioactive core at 2 cm intervals. Sediment

was transferred into centrifuge tubes and centrifuged to extract porewater that was stored frozen

until analysis of sulfate using ion-chromatography.

Statistics

Differences between PO43-, F-, and Ca2+ fluxes with acid treatments relative to control

treatments and release of adsorbed F- and PO43- in chemical extracts were compared using paired t-

tests. Linear regression analyses were used to determine slopes that defined the relationship

between acid addition and ion flux (Sigma Stat version 3.1). Significant differences are reported at

the p < 0.05 level unless otherwise stated.

Results

Sediment characteristics - solid phase

Total sediment phosphorus (TP) and inorganic solid phase phosphorus (iP) (Table 1) closely

followed the established trophic gradient in the bay, depicting greater P enrichment in the western

(TP = 4.40 ± 0.03 and 9.48 ± 0.25) than eastern (TP = 1.77 ± 0.06 and 2.22 ± 0.05) bay sediments.

The importance of the inorganic Ca-bound P pool is shown by this fraction (iP) representing

approximately 50±2% of the total P pool at all four sites (Table 1). Organic matter content,

13

measured as loss on ignition (LOI) also showed an east-west gradient with LOI = 7.8% and 7.6% at

eastern sites and 10.3% and 22.2% at western sites (Table 1). Total Ca concentration (Table 1)

reflect the LOI numbers, as Ca levels can be multiplied by the molar weight of 100 to give the

weight of CaCO3 in the dry sediment. The fact that the two values make up ~100% of dry weight at

all sites, confirms that insignificant amounts of other mineral phases than CaCO3 were present.

Total F- levels were similar across sites with the exception of BB where it was nearly two

times higher. The molar ratios of fluoride to calcium (Table 1) are indicative of sediment dominated

by easily dissolvable aragonite and high-Mg calcite in support of the supposition that dissolution

could be an important mechanism of P recycling in Florida Bay sediments. Carbonate dissolution at

western bay sites has the potential to release greater amounts of sediment P than eastern bay sites

based on a 4-fold higher iP:Ca ratio and an almost 3-fold higher iP concentration in western

compared to eastern bay sediments. The grain size distribution showed that EK had the largest

fraction (83%) of small particles (<63 µm), while this fraction made up 69%, 63%, and 59% at

BLBK, BB, and RK, respectively (Table 1).

Sulfuric acid addition experiment

After 6 h post-acid treatment, pH levels stabilized at all sites to approximately 6.6 to 7.1. As

pH dropped in response to sediment acidification, Ca2+, F-, and PO43- were released. A portion of

the F- and PO43- recovered was in the seawater slurry water (Fig. 2), while another portion was

recovered from the adsorbed pool with MgCl2 and BD extractions (Table 2). The proportionality of

dissolved versus re-adsorbed F- and PO43- was very site specific.

The CaCO3 dissolution, calculated from the increase of Ca2+ in slurry water, amounted to

1.3% (maximum) of the sediment solid phase pool in response to the acid treatments simulating

field pH conditions (Table 2, Fig. 2). Despite Ca2+ mobilization with acid treatments were evident at

14

all sites, PO43- dissolved into the slurry water was only significantly different from controls at

Rabbit Key (Fig. 2, Table 2), the site with the highest sediment TP and iP in the bulk sediment

(Table 1). Sequential extraction of the surface bound iP pools of acidified sediment was necessary

to recover a significant amount of the dissolved iP at several of the sites. These data indicate that a

portion of the mobilized iP became surface bound, as either the loosely adsorbed MgCl2 fraction or

the more strongly adsorbed BD fraction (Table 2). A significant increase in extractable iP in one or

both of the surface-bound pools was found at RK, BB, and BLBK (Table 2). Only at EK were the

changes in surface bound iP insignificant across all acidification treatments and were not recovered

in the slurry water or through sequential extractions with MgCl2 and BD (Table 2).

If we examine the total amount of iP extracted as dissolved and recovered as surface

adsorbed relative to controls, there was significantly more P recovered in western bay sediments

(Table 2) with higher total sediment P and iP in the bulk sediment (Table 1). Fluoride mobilization

to the slurry water with acidification followed the same trend as observed for PO43- with the greatest

F- release at the western sites, RK and BB, and with a significant flux also at BLBK, relative to

controls (Table 2). F- was also mobilized to both the slurry water and recaptured as a surface-bound

pool on the carbonate sediment. While no significant release of F- to the slurry water was found at

EK, all four stations had significant F- recovery in the adsorbed exchangeable pools (Table 2).

The amount of iP dissolved versus the sulfuric acid added correlated positively for RK (p <

0.05; R2 = 0.74) and BLBK (p < 0.05; R2 = 0.44) while there was only a tendency for a positive

relationship at BB (p = 0.08; R2 = 0.33). No iP mobilization was observed for sediments collected

from EK (Fig. 3). For the three sites with observed iP mobilization, PO43- flux as a linear function

of H2SO4 amended was 0.058, 0.105, and 0.359 nmol iP per µmol H2SO4 at BLBK, BB, and RK,

respectively, the same ranking of sites as for iP concentration in the bulk sediment (Table 1). In

general, the amount of Ca2+ and F- dissolved and the sulfuric acid concentrations added showed a

15

strong correlation among sites (Fig. 3). The response of F- to acid treatment was similar at EK,

BLBK and RK sites with slopes of 1.23, 1.37, and 1.25 nmol F- per µmol H2SO4, respectively. In

contrast, BB had a higher slope of 1.61, in correspondence with the higher concentration of F in the

bulk sediment at this site (Table 1). Leaching of Ca2+ with acid treatment did not differ significantly

among sites but there was a tendency for less Ca2+ to be mobilized from eastern (slopes of 0.77 and

0.84 µmol Ca2+ per µmol H2SO4 at EK, and BLBK, respectively) than western sites (slopes of 0.93

and 0.97 µmol Ca2+ per µmol H2SO4 at BB, and RK, respectively). This observation may indicate

that western bay sediments are more easily dissolved than eastern bay sediments. The near 1:1

relationship between Ca2+ leached and H2SO4 added points toward an approximate equimolar

reaction between sulfuric acid and carbonate sediment mineralization (see methods).

The molar ratio between F- and Ca2+ dissolved upon acidification (Table 3) was significantly

lower than F:Ca ratios of the bulk sediment (Table 1), indicating preferential dissolution of a F--

poor carbonate fraction applying modest acid treatments that achieved field pH conditions.

Likewise, less iP was mobilized and subsequently recovered relative to Ca2+ mobilization (Table 3)

than predicted from the bulk sediment iP and Ca distribution among sites (Table 1). This either

indicates dissolution of a relatively P-poor carbonate fraction, or more likely, that a large re-

adsorption and incorporation into solid phase P pools occurred that was not recovered by the

sequential extraction scheme applied. Re-adsorption was less pronounced for F- than for PO43-.

Nevertheless, P:F ratios in the leachates (including slurry water and adsorbed pools) were 65% and

115% higher than in the bulk sediment at the two western sites (BB and RK, respectively), while

similar to the ratio in bulk sediment at BLBK.

16

Tracer experiment

33PO43- amended to sediment slurries rapidly adsorbed to the carbonate particles with 0.35%

and 1.1% left in the water after 60 min at EK and GMK, respectively (Fig. 4). At the eastern bay

site (EK), 33PO43- was distributed in proportion with the cold phosphate after 4 h with 25% found in

the surface-adsorbed P pools, 2.6% in the MgCl2 and 22.4% in the BD extractions, and 75% within

the carbonate matrix (Table 4). In the western bay at GMK, 33PO43- was primarily found in the

surface-bound pools from where it only slowly re-distributed into the HCl-extractable carbonate

matrix pool. Even after 55 h, 53% of the 33PO43- was recovered in the surface adsorbed pools which

only contributed 25% of the solid fraction iP. Despite the difference in rate of incorporation and

despite the clear gradient in sediment iP concentration (Table 1) the proportion between different iP

pools in the sediment, as evaluated by sequential chemical extraction, appeared quite similar in the

eastern EK and the western GMK sediment (Table 4). The results suggest that both 31PO43-

adsorption and the subsequent incorporation into the solid matrix will proceed considerably faster at

EK than at GMK, indicating a very different rate of solid phase iP turnover at eastern versus

western bay sediments.

Sulfate reduction rates

The highest SRR (up to 800 nmol cm-3 d-1) were measured in the western bay sediments,

and in the upper 0-2 cm of sediment (Fig. 5). Depth integrated SRR in the top 14 cm were 2-fold

higher at RK (45.4 ± 18.1 in July and 85.4 ± 19.4 mmol m-2 d-1 in October) compared to BLBK and

EK, while BB had moderate SRR rates, falling between rates measured at RK and the eastern sites

(Table 5). The different SRR among sites resembled the pattern of LOI (Tables 5 and 1). Only at

Rabbit Key did we observe different SRR during the two sampling occasions with rates in the fall

almost twice rates in July, although high variance was also observed (Fig. 5; Table 5). The higher

17

autumn rates could be due to higher DOC availability from senescence of T. testudinum leaves and

a reduction in shoot density (Koch et al. 2007, Rosch and Koch 2009).

Discussion

Our results show that carbonate dissolution caused by sulfuric acid production can mobilize

significant amounts of PO43- at near neutral pH, one mechanism by which seagrass growing in

carbonate sediments can access P bound in the solid matrix. Recycling of sulfide to sulfuric acid

through O2 mediated oxidation, and its coupling to carbonate dissolution, has been established in

seagrass sediments of Florida Bay (Ku et al. 1999; Walter et al. 2007). When sulfide oxidation is

coupled to O2 released from seagrass roots, the iP released is in close proximity of roots. Other

processes like bioturbation and porewater advection (Burdige et al. 2008) can also transport O2

down into the sediment and potentially cause sulfide oxidation. However, the oxidation-reduction

biogeochemical processing around the root zone of seagrass may be the most active because of

concomitant root exudation of organic acids that can stimulate SR, or potentially compete with

PO43- for sorption sites on sediment surfaces (Long et al. 2008). Our study also indicates that

mobilization of carbonate-bound P through dissolution, and its subsequent availability to porewater

and seagrass, is not solely a function of the P:Ca ratio in the bulk sediment. Rather, the capacity of

the sediment to re-sorb PO43- can modify the net mobilization of P during dissolution and control iP

availability in the root zone. We evaluate below the potential magnitude of iP mobilization coupled

with SR and subsequent sulfide oxidation, and provide estimates of its significance for seagrass

nutrition in Florida Bay.

18

Potential rate of P mobilization by sulfuric acid formation

Under our H2SO4 treatments, pH values dropped in the sediment slurry to 6.6-6.8, on par

with those found in porewaters of Florida Bay (6.6-6.9; Koch et al. 2007; Walter and Burton 1990).

These results indicate that the acidification levels generated in this study by H2SO4 represent those

that can occur over time in the bay generated in part by SR and reoxidation of sulfides. As a first

approximation, we estimate potential iP dissolution rates in Florida Bay carbonate sediments based

on sulfuric acid formation applying our field SRR and assuming unity between SR and sulfide

oxidation. Our SRR determined on cores collected from seagrass sites in the field give us values of

187-610 nmol cm-3 d-1. These data compare to those determined by Walter et al. (2007) for

sediments of mudbank and island fringe sites in central Florida Bay (70-579 nmol cm-3 d-1) and

reported by Holmer et al. (2006) for tropical seagrass sediments in East Asia. Using our SRR, with

the assumption that all sulfides generated from SR are re-oxidized in a 1:1 ratio, sulfide oxidation

would contribute the equivalent of 1.4±0.66 and 0.27±0.09 µmol H2SO4 g dry wt-1 d-1 at western

and eastern bay sites, respectively.

Applying the potential daily acid generation with experimental results on PO43- release with

acidification in this study, a maximum of 23.5 and 1.8 µmol P m-2 d-1 would be regenerated at the

western RK and eastern BLBK sites, respectively (Table 6a). These iP mobilization rates are within

the range calculated from literature values for Ca2+ and F- mobilization in Florida Bay (Table 6).

Since we mobilized Ca2+ with a 1:1 ratio between H2SO4 addition and Ca2+ flux, where the ratio

theoretically should be closer to 1:2, we suggest that the range given in Table 6a is a conservative

estimate of the potential effect that sulfide oxidation might have on iP mobilization. We reach

similar rates of iP mobilization using the gross carbonate dissolution rates estimated by Rude and

Aller (1991) in Florida Bay (Bob Allen Key Bank) applying our iP:Ca efflux ratios (Table 6b).

These estimates increase 2-fold when a 2:1 molar ratio between carbonate dissolution and sulfuric

19

acid formation (Ku et al. 1999) is assumed, or if we apply the calculated F- efflux from Rude and

Aller (1991) with our observed site specific iP:F efflux ratios (Table 6c and 6d). A slightly more

conservative estimate for iP mobilization rate is calculated when using the lower carbonate

dissolution rate reported by Walter and Burton (1990) and our site specific iP:Ca efflux ratios

(Table 6e). Averaging (±SE) the estimates presented in Table 6 give us an iP turnover rate of

3.5±1.0, 6.8±1.8, and 28.9±8.4 µmol P m-2 d-1 from BLBK, BB, and RK, respectively. These iP

turnover rates are site specific with an increase at sites with high sediment TP and seagrass

productivity in Florida Bay.

Seagrass P nutrition from potential iP dissolution

We calculated potential seagrass P nutrition attainable from carbonate dissolution by

applying the estimates of iP mobilization in Table 6 and site specific T. testudinum total P

production in leaves, the tissue accounting for the greatest allocation and turnover of nutrients in

this seagrass species. Annual average (±SE) T. testudinum leaf P requirements at BLBK and RK are

34±8 and 80±18 µmol P m-2 d-1, (Rosch and Koch 2009) comparable to estimates for the bay as a

whole (70 µmol P m-2 d-1; Fourqurean 1992a), but clearly showing the higher biomass production at

RK (3.3 g dry wt m-2 d-1) than BLBK (1.0 g dry wt m-2 d-1) during summer months when leaf

elongation rates are maximal (Koch et al. 2007). Based on the leaf TP requirements and average

estimates for iP mobilization from carbonate dissolution (Table 6), we estimate approximately 38%

of the P requirements can be met by iP from carbonate sediment dissolution at RK, while only 10%

at BLBK, clearly indicating that rapid recycling of this solid-phase pool probably assists in

promoting greater seagrass biomass development at western Florida Bay sites.

Our measured SRR also provide an estimate of P mobilization from organic matter

mineralization using the C:P molar ratio of belowground tissue at RK and BLBK (~2290 and 2960,

20

respectively; Rosch and Koch 2009). Applying a respiration quotient of 2 for SR (1 mol SO42-

reduced per 2 mol C oxidized), P recycled from organic matter mineralization would be

approximately 18 µmol P m-2 d-1 at BLBK and 57 µmol P m-2 d-1 at RK. Thus, based on our results,

solid phase inorganic P cycling is 19% and 51% of the organic P recycling generated from SR at

BLBK, and RK. Combining this inorganic and organic P turnover, we estimate that only 64% of the

P requirements for T. testudinum is being met by iP dissolution and SR at the eastern bay BLBK

site, while in contrast, 100 percent (108%) of the P needs are calculated to be satisfied in the

western bay at RK. However, this dissolution effect and iP release can be significantly affected by

the sediment potential for PO43- sorption, which is also site specific.

Phosphate resorption and incorporation by carbonates

Independent of the mechanism that may cause dissolution of the solid carbonate matrix, the

PO43- flux to the porewater is modified by resorption processes. Eagle Key sediments in the

northeastern bay, the site with the lowest sediment iP and the largest fraction of fine particles, was

found to strongly bind iP, even to the point where no PO43- recovery was made from the dissolved

or adsorbed P pools after treatment with 53 µmol H2SO4 g dry wt-1. This result can be explained in

two ways. One, P was not released with acid dissolution, which appears unlikely based on the fact

that CaCO3 was dissolved, or two, P was rapidly (< 1 h) re-adsorbed and incorporated into a non-

exchangeable P pool.

In a previous experiment by the authors on PO43- adsorption-desorption equilibrium (Nielsen

et al. 2006), the crossover concentration (Ce0) where porewater PO43- is in equilibrium with

adsorbed pools and can no longer be taken up, was extremely low (0.017 µmol L-1) at EK. These

results indicate a high capacity for the eastern bay sediments to resorb PO43-, even at low

concentrations. On the contrary, at a western bay site (GMK), Ce0 was 10-times higher (0.292 µmol

21

L-1), suggesting a higher porewater PO43- equilibrium is maintained at this site and higher inorganic

P availability to seagrass roots embedded in the sediment compared to the EK site. Small carbonate

grain sizes (<10 µm) have been shown to be important in P adsorption (Suess 1973) and Florida

Bay sediments have been defined as having high surface energies resulting from small particle size

(Walter et al. 2007). One mechanism behind the high PO43- affinity at EK could be the large

proportion of fine particles, but it is also possible that the degree of iP saturation on and in the

carbonate grains modified PO43- affinity. McGlathery et al. (1994) observed a slower rate of PO4

3-

adsorption (6 h experiment) in more P-enriched carbonate sediments from Bermuda without

detectable differences in grain-size distribution. For Florida Bay such a mechanism would imply

that eastern bay sediment, low in TP, would have higher affinity for PO43- than western bay

sediment, given the same grain size. In the McGlathery et al. (1994) study, however, PO43- was

amended at high concentrations and the median grain size was > 200 µm in the Bermuda sediments.

In contrast, the Nielsen et al. (2006) study in Florida Bay examined affinity at low field porewater

PO43- concentrations with median grain sizes < 63 µm. Thus, the role of grain size cannot be ruled

out as a potential influence on sediment P adsorption contrasting finer sediments in the eastern

compared to coarse grains in the western regions of the bay, although these site differences are

confounded by a P enrichment gradient making it difficult to distinguish between these two effects

on PO43- adsorption.

The potential for tight resorption of iP released from EK sites with fine grains is also

supported by the rapid rate of isotopic exchange of 33PO43- into the HCl-extractable iP pool at this

site. Thus, while the apportionment of iP in the various pools in the bay are more similar based on

31P (Zhang et al. 2004, this study), as revealed through traditional low pH extractions (0.5 mol L-1

HCl), the rates of incorporation and binding efficiency (Nielsen et al. 2006) is quite variable, with

stronger binding of iP in the finer, low TP eastern bay sediments. The fluxes of PO43- to the

22

porewater from carbonate dissolution and re-mineralization may be counteracted by re-adsorption

onto carbonates and/or incorporation into the sediment matrix. Incorporation would be driven either

by diffusion (Froelich 1988) or by re-crystallization (Rude and Aller 1991; Hu and Burdige 2007,

Walter et al. 2007). These adsorptive and incorporation processes may explain the low PO43- in

porewaters in the eastern versus western regions of the bay (Fourqurean et al. 1992b), and the

inability to recover PO43- from EK sediments in this study, even with a clear dissolution of

carbonates.

Because of the importance of resorption of PO43- by carbonates, porewater PO4

3-

concentrations, rather than TP pools, are probably a good proxy for P availability for seagrass

nutrition. This idea is partially supported by the positive correlation between porewater PO43- and

tissue P concentration (Fourqurean et al. 1992a, b), and by PO43- adsorption experiments (Nielsen et

al. 2006) that imply P-limitation of seagrasses in eastern Florida Bay, resulting from low porewater

PO43- concentrations. Also, T. testudinum root PO4

3- uptake kinetics responds linearly as a function

of porewater PO43- concentration in the low ranges found in the bay; and although this uptake can

sequester phosphate down to ~0.030 µmol L-1, at these low concentrations, the plants cannot meet

their P requirements (Nielsen et al. 2006). As porewater PO43- in tropical carbonate seagrass

sediments has a rapid turnover rate of ~2 days (Jensen et al. 1998), it can be reasoned that the

loosely adsorbed and solid phase iP pools, with several orders of magnitude higher P levels than

porewater, are critical in supporting highly productive tropical seagrasses. We contend that this

turnover is driven primarily by dissolution processes associated with O2 enhanced acid formation.

Phosphate availability may be further enhanced by root exudation of organic acids.

In laboratory studies Long et al. (2008) observed that moderate addition of di- and trivalent

organic acids, shown to be released by the roots of T. testudinum, caused significant PO43- leaching

without concomitant F- leaching. No effect was found with HCl at the same acid strength of the

23

organic acid exudates measured in the field. Since this acidity from organic acids added by Long et

al. (2008) to Florida Bay sediments was approximately an order of magnitude lower than in our

experiment herein applying H2SO4, the results suggest that the organic acids leached at the root tips

affected the equilibrium between porewater PO43- and the surface-bound iP pool. However, these

organic acids appeared to generate inadequate acidity to dissolve the solid-phase iP carbonate pool,

again evidenced by a lack of F- mobilization. Indirectly, organic acid exudates can stimulate

carbonate dissolution by enhancing acid generation from oxidative mineralization processes. This is

supported by field observations of increased P and F- along with high concentrations of organic

acids in the root zone of T. testudinum (Long et al. 2008).

Earlier studies (Jensen et al. 1998, Nielsen et al. 2006) provided strong indications that the

loosely adsorbed and solid phase iP pools are potentially critical in supporting highly productive

tropical seagrass in carbonate environments where P availability is low. Here we forward the

supposition that this solid phase iP pool is turned over during the dissolution of carbonate minerals

that occur when sulfuric acid from sulfide oxidation is buffered at near neutral pH. We also show

that subsequent to dissolution, P availability to seagrass roots may be highly dependent on sediment

adsorptive capacities, which are controlled by sediment P loads and characteristics at a particular

site (e.g., grain size). Future research is needed to more clearly define how site differences in

sediment characteristics, nutrient loads, and seagrass plant productivity influence O2 flux into the

sediments (Burdige et al. 2008), and sulfide oxidation, as well as di- or trivalent organic acids,

potential to drive biogeochemical processes and/or assist tropical seagrass in accessing the surface

adsorbed PO43-.

24

References Barber, T. R., and P. R. Carlson. 1993. Effects of seagrass die-off on benthic fluxes and porewater

concentrations of ΣCO2, ΣH2S, and CH4 in Florida Bay sediments, p 530-550. In R. S.

Oremland [ed.], Biogeochemistry of global change: Radiatively active trace gases. Chapman

and Hall.

Burdige, D. J., and R. C. Zimmerman. 2002. Impact of seagrass density on carbonate dissolution in

Bahamian sediments. Limnol. Oceanogr. 47: 1751-1763

Burdige, D. J., R. C. Zimmerman, and Xinping Hu. 2008. Rates of carbonate dissolution in

permeable sediments estimated from pore-water profiles: The role of seagrasses. Limnol.

Oceanogr. 53: 549-565.

Carlson, P. R., L. A. Yarbro, and T. R. Barber. 1994. Relationship of sediment sulfide to mortality

of Thalassia testudinum in Florida Bay. Bull Mar Sci 54: 733-746.

Chambers, R. M., J. W. Fourqurean, S. A. Macko, and R. Hoppenot. 2001. Biogeochemical effects

of iron availability on primary producers in a shallow marine carbonate environment. Limnol

Oceanogr 46: 1278-1286.

Fossing, H., and B. B. Jørgensen. 1989. Measurement of bacterial sulfate reduction in sediments:

Evaluation of a single-step chromium reduction method. Biogeochemistry 8: 205-222.

Fourqurean, J. W., J. C. Zieman, and G. V. N. Powell. 1992a. Phosphorus limitation of primary

production in Florida Bay: Evidence from C:N:P ratios of the dominant seagrass Thalassia

testudinum. Limnol Oceanogr 37: 162-171.

Fourqurean, J. W., J. C. Zieman, and G. V. N. Powell. 1992b. Relationships between porewater

nutrients and seagrasses in a subtropical carbonate environment. Mar Biol 114: 57-65.

25

Fourqurean, J. W., and J. C. Zieman. 2002. Nutrient content of the seagrass Thalassia testudinum

reveals regional patterns of availability of nitrogen and phosphorus in the Florida Keys USA.

Biogeochemistry 61: 229-245.

Froelich, P. N. 1988. Kinetic control of dissolved phosphate in natural rivers and estuaries – a

primer on the phosphate buffer mechanism. Limnol Oceangr 33: 649-668.

Holmer, M., O. Pedersen, and K. Ikejima. 2006. Sulfur cycling and sulfide intrusion in mixed

southeast Asian tropical seagrass meadows. Bot Mar 49: 91-102.

Hu, X., and D. J. Burdige. 2007. Enriched stable carbon isotopes in the porewaters in carbonate

sediments dominated by seagrasses: Evidence for coupled carbonate dissolution and

reprecipitation. Geochim Cosmochim Acta 71: 129-144.

Jensen, H. S., and B. Thamdrup. 1993. Iron-bound phosphorus in marine sediments as measured by

bicarbonate-dithionite extraction. Hydrobiologia 253: 47-59.

Jensen, H. S., K. J. McGlathery, R. Marino, and R. W. Howarth. 1998. Forms and availability of

sediment phosphorus in carbonate sand of Bermuda seagrass beds. Limnol Oceanogr 43:

799-810.

Jensen, H. S., N. Caraco, J. Hansen, and K. K. Christensen. 2005. Humic-bound phosphorus in soil

and sediment, p. 99-107 In H. L. Golterman and L. Serrano [eds.], Phosphate in sediments.

Backhuys Publishers.

Jørgensen, B. B. 1978. A comparison of methods for the quantification of bacterial sulfate reduction

in coastal marine sediments. Geomicrobiology Journal 1: 11-27.

Kitano, Y., M. Okumura, and M. Idogaki, 1978. Uptake of phosphate ions by calcium carbonate.

Geochemical Journal 12: 29-37.

Koch, M. S., E. Benz, and D. T. Rudnick. 2001. Solid-phase phosphorus pools in highly organic

carbonate sediments of northeastern Florida Bay. Estuar Coast Shelf Sci 52: 279-291.

26

Koch, M. S., S. A. Schopmeyer, O. I. Nielsen, C. Kyhn-Hansen, and C. J. Madden. 2007.

Conceptual model of seagrass die-off in Florida Bay: Links to biogeochemical processes. J Exp

Mar Biol Ecol 350: 73-88.

Koroleff, F. 1983. Determination of phosphorus, p. 125-139 In K. Grasshoff, K. M. Ehrhardt and K.

Kremling [eds.], Methods of seawater analysis. Verlag Chemie.

Ku, T. C. W., L. M. Walter, M. L. Coleman, R. E. Blake, and A. M. Martini. 1999. Coupling

between sulfur recycling and syndepostitonal carbonate dissolution: Evidence from oxygen

and sulfur isotope composition of porewater sulfate, South Florida Platform, U.S.A. Geochim

Cosmochim Acta 63: 2529-2546.

Lee, K. S., and K. H. Dunton. 2000. Diurnal changes in pore water sulfide concentration in the

seagrass Thalassia testudinum beds: the effect of seagrasses on sulfide dynamics. J Exp Mar

Biol Ecol 255: 201-214.

Long, M. H., K. J. McGlathery, J. C. Zieman, and P. Berg. 2008. The role of organic acid exudates

in liberating phosphorus from seagrass-vegetated carbonate sediments. Limnol. Oceanogr. 53:

2616-2626.

Lyons, T. W., L. M. Walter, A. M. Gellaty, A. M. Martini, and R. E. Blake. 2004. Sites of

anomalous organic remineralization in the carbonate sediments of South Florida, USA: The

sulfur cycle and carbonate-associated sulfate, p. 161-176 In J. P. Amend, K. J. Edwards and

T. W. Lyons [eds.], Sulfur biogeochemistry- past and present. Geological Society of America

Special Paper 379.

McGlathery, K. J., R. Marino, and R. W. Howarth. 1994. Variable rates of phosphate-uptake by

shallow marine carbonate sediments – mechanisms and ecological significance.

Biogeochemistry 25: 127-146.

27

Millero, F., F. Huang, W. R. Zhu, X. W. Liu, and J. Z. Zhang. 2001. Adsorption and desorption of

phosphate on calcite and aragonite in seawater. Aquat Geochem 7: 33-56.

Morse, J. W., J. J. Zullig, R. L. Iverson, G. R. Choppin, A. Mucci, and F. J. Millero. 1987. The

influence of seagrass beds on carbonate sediments in the Bahamas. Marine Chemistry 22: 71-

83.

Nielsen, O. I., M. S. Koch, H. S. Jensen, and C. J. Madden. 2006. Thalassia testudinum phosphate

uptake kinetics at low in situ concentrations using a 33P radioisotope technique. Limnol

Oceanogr 50: 208-217.

Rosch, K. L., and M. S. Koch. 2009. Nitrogen and phosphorus recycling by a dominant tropical

seagrass (Thalassia testudinum) across a nutrient gradient in Florida Bay. Bull Mar Sci. 84: 1-

24.

Rude, P. D., and R. C. Aller. 1991. Fluorine mobility during early diagenesis of carbonate sediment:

An indicator of mineral transformations. Geochimica Acta 55: 2491-2509.

Ruttenberg, K. C. 1992. Development of a sequential extraction method for different forms of

phosphorus in marine sediments. Limnol Oceanogr 37: 1460-1482.

Short, F. T., W. C. Dennison, and D. G. Capone. 1990. Phosphorus-limited growth of the tropical

seagrass Syringodium filiforme in carbonate sediments. Mar Ecol Prog Ser 62:169-174.

Suess, E. 1973. Interaction of organic compounds with calcium carbonate-II. Organo-carbonate

association in recent sediments.Geochimica et Cosmochimica Acta 37: 2435–2447.

Walter, L. M., and E. A. Burton. 1990. Dissolution of recent platform carbonate sediments in

marine pore fluids. Am J Sci 290: 601-643.

Walter, L. M., T. C. W. Ku, K. Muehlenbachs, W. P. Patterson, and L. Bonnell. 2007. Controls on

the δ13C of dissolved inorganic carbon in marine porewaters: An instructive example from

28

biogenic carbonate sediments (south Florida Platform). Deep Sea Research, Part 2: Studies in

oceanography 54: 1163-1200.

Zhang, J. Z., C. J. Fisher, and P. B. Ortner. 2004. Potential availability of sedimentary phosphorus

to sediment resuspension in Florida Bay. Global Biogeochem. Cy. 18: GB4008 doi:

10.1029/2004GB002255.

29

Table 1. Sediment total calcium (Ca), fluoride (F), and inorganic (iP), and total P (TP), molar ratios of F to Ca and iP to Ca and F, and bulk

density (bd), porosity, and % sediment < 60 µm at Eagle Key (EK), Black Betsy Key (BLBK), Buchanan Bank (BB), Rabbit Key (RK), and

Green Mangrove Key (GMK) in Florida Bay. Elemental and ratio values represent means ± SD (n=3). Sediment bd and porosity are depth

averages (0-14 cm). Not all parameters were measured for GMK.

(mmol g dry wt-1) (μmol g dry wt-1) Molar ratios (x 10-3) g cm-3 %

__________________________________ ___________________ _______________

Site Ca F iP TP F:Ca iP:Ca iP:F bd porosity <63 µm LOI

EK 9.23 28.8 ± 1.3 0.81 ± 0.02 1.77 ± 0.06 3.11 0.09 28 0.70 0.73 83 7.8 ± 0.03

BLBK 9.29 31.5 ± 3.1 1.12 ± 0.02 2.22 ± 0.05 3.39 0.12 36 0.91 0.66 69 7.6 ± 0.07

BB 8.90 58.4 ± 6.3 2.35 ± 0.03 4.40 ± 0.03 6.57 0.26 40 0.34 0.84 63 10.3 ± 0.15

RK 7.80 35.6 ± 3.9 4.66 ± 0.16 9.48 ± 0.25 4.56 0.60 131 0.25 0.89 59 22.2 ± 0.07

GMK 8.87 1.69 ± 0.13 3.44 ± 0.17 0.19 0.42 0.82 46 12.2 ± 0.18

30

Table 2. Inorganic phosphorus (ΔPO43-), fluoride (ΔF-), and calcium (ΔCa2+) mobilized in seawater (SW) after acidification treatments (150,

400, and 800 µmol H2SO4) and ΔiP and ΔF- re-adsorbed but recovered in specific sequential extracts with MgCl2 (loosely adsorbed) and BD

(strongly adsorbed). Values are represented as changes in concentration compared to control treatments with no acid dissolution; (*) represent

significant changes (p < 0.05). Σ, in bold, represents the sum of increments in iP or F- in the three sequential fractions (SW, MgCl2, and BD).

Calcium was only measured in slurry water.

150 µmol H2SO4 400 µmol H2SO4 800 µmol H2SO4

_____________________________________________ ________________________________________________ ___________________________________________________

Site SW MgCl BD Σ SW MgCl BD Σ SW MgCl BD Σ

ΔiP (nmol g dry wt-1)

EK 0.04 0.08 10.47 10.60 0.43 0.07 -16.50 -16.00 0.08 0.09 -9.96 -9.79

BLBK -0.08 -0.60 1.66* 0.98* -0.16 -0.07 2.26 2.03 -0.16 -0.01 4.46* 4.29*

BB -0.62 0.28 3.43 3.10 -0.43 1.33* 8.78 9.67 0.65 2.59* 8.78 12.13

RK 0.07 1.94 34.66* 36.66* 2.29* 5.26 48.62* 56.17* 12.10* 22.11 48.67* 82.87*

31

ΔF- (nmol g dry wt-1)

EK -0.07 7.48 12.72* 20.13 10.84 6.54 28.23 45.61* 13.66 9.06 58.52* 81.23*

BLBK 2.91 2.63 -17.01 -11.47 9.16 7.95 4.33 21.44 21.81* 12.26* 19.26 53.34

BB 17.15 0.64 18.31 36.10 43.86* 2.93 39.54 86.33* 93.83* 8.16* 63.40* 165.39

RK 34.19 0.32 -22.52 11.98 104.28* 16.61* -38.19 82.69* 213.63* 27.17* -72.89 167.91

ΔCa2+ (μmol g dry wt-1)

EK 7.6* 31.0* 50.6*

BLBK 16.9* 28.9* 54.9*

BB 19.4 48.6 94.1

RK 29.6 70.9 152.8

32

Table 3. Ratios between iP, F-, and Ca2+ mobilized

after sulfuric acid additions to carbonate sediments

from Eagle Key (EK), Black Betsy key (BLBK),

Buchanan Bank (BB), and Rabbit Key (RK). The

ratios are calculated from the slopes generated from

Fig. 3.

Atomic ratios (x 10-3)

_______________________

F:Ca iP:Ca P:F

EK 1.45 ns ns

BLBK 1.78 0.07 41

BB 1.71 0.11 66

RK 1.29 0.37 282

No significant (ns) iP mobilization was found at Eagle Key.

33

Table 4. Distribution of 33P (%) in various P-pools as a function of time following addition of trace

amounts of 33PO43- to sediment slurries from the western Florida Bay (Green Mangrove Key;

GMK) and the northeastern Florida Bay (Eagle Key; EK) sites. The percentage distribution of

31PO43- in the sediments is also shown. Values represent mean ± SD (n=3).

Relative 33P distribution (%) 31P dist (%)

______________________________________________ __________

4.66 h 7.66 h 33 h 55 h

GMK MgCl2 26.8 ± 2.9 22.8 ± 3.7 16.1 ± 1.7 16.1 ± 0.4 5.4 ± 1.2

BD 42.7 ± 1.3 44.4 ± 2.2 39.5 ± 0.9 36.9 ± 0.2 20.1 ± 0.9

HCl 30.5 ± 1.6 32.8 ± 1.5 44.4 ± 2.6 47.0 ± 0.2 74.5 ± 2.0

EK MgCl2 3.8 ± 0.0 3.5 ± 0.1 2.3 ± 0.1 2.6 ± 0.2 5.7 ± 0.4

BD 22.8 ± 0.3 24.2 ± 0.2 20.6 ± 1.3 22.9 ± 1.9 17.1 ± 1.8

HCl 73.5 ± 0.3 72.1 ± 0.3 77.1 ± 1.2 74.5 ± 1.8 77.2 ± 1.7

34

Table 5. Seasonal variations in sulfate reduction rates (SRR) integrated to

14 cm depth and mean SRR normalized to sediment volume in seagrass

sediments from Eagle Key (EK), Black Betsy Key (BLBK), Buchanan

Bank (BB), and Rabbit Key (RK) in Florida Bay. Values are mean ± SD

(n=3 cores).

ΣSRR0-14 cm SRR

Site Date (mmol m-2 d-1) (nmol cm-3 d-1)

EK Oct 2005 34.9 ± 3.1 249.3 ± 21.9

Jul 2006 30.3 ± 4.3 216.1 ± 23.2

BLBK Oct 2005 26.2 ± 4.6 187.3 ± 33.1

Jul 2006 28.0 ± 5.5 199.6 ± 26.4

BB Oct 2005 42.7 ± 3.6 304.8 ± 95.2

Jul 2006 50.7 ± 14.6 362.5 ± 104.5

RK Oct 2005 85.4 ± 19.4 610.3 ± 138.8

Jul 2006 45.4 ± 18.1 324.2 ± 129.2

35

Table 6. Potential inorganic solid-phase P (iP) mobilization

rates in Florida Bay (FB) carbonate sediments estimated by

sulfate reduction rates and equivalent acid generation

assuming a (a) 1:1 molar ratio (Fig. 3a) and (b) a 1:2 molar

ratio (Ku at al. 1999) of H2SO4 produced to Ca2+ dissolved,

literature estimates of CaCO3 dissolution in FB sediments

applying our ratios of iP:Ca and iP:F mobilization (Table

3), based on model predicted (Rude and Aller 1991) gross

(c) Ca2+ mobilization of 53 mmol m-2 d-1, (d) F- efflux of

170 µmol m-2 d-1, and (e) based on CaCO3 dissolution rates

of 13.7 mmol m-2 d-1 (Walter and Burton 1990).

iP mobilization estimates*

(µmol P m-2 d-1)

Black Betsy Key Buchanan Bank Rabbit Key

1.8a 5.0a 23.5a

3.8b 10.3b 48.4b

3.7c 5.8c 19.6c

7.0d 11.2d 47.9d

1.0e 1.5e 5.1e

*No estimates were made for Eagle Key because of a lack

of P recovery (dissolved and exchanged) after sediment

acidification.

36

Figure Legends

Figure 1. Study sites in Florida Bay, a semi-enclosed lagoonal estuary at the southern terminus of

the Florida Peninsula, bounded to the north by the Everglades, to the south by the Florida Keys and

Gulf of Mexico to the west. Sediment for SRR and dissolution experiments were collected from

eastern bay sites: Eagle Key (EK) and Black Betsy Key (BLBK) and western bay sites: Rabbit Key

(RK) and Buchanan Bank (BB). Radiotracer experiment with 33PO43- was conducted with sediment

from EK and Green Mangrove Key (GMK). In the bay numerous mangrove islands (dark grey) and

mudbanks (light grey) are noted.

Figure 2. pH changes and efflux of phosphate (PO43-), fluoride (F-), and Ca2+ into the seawater

sediment slurry (dissolved phase) at three different acidification treatments (see legend) and control

over time in the 6 h incubation (n=3).

Figure 3. Relationship between sulfuric acid additions and release of (a) inorganic P (PO43-), (b)

fluoride (F-), and (c) calcium (Ca2+), including both the seawater sediment slurry (dissolved ions)

and the exchangeable (MgCl2 and BD extractable) pools. All fluxes and acid additions are

calculated per g dry wt-1 sediment and the acid treatment is corrected for the acid consumption by

the alkalinity of the added 80 mL seawater.

Figure 4. Changes in the 33P radioactivity over time in the aqueous phase of the sediment slurry,

using sediment from Eagle Key and Green Mangrove Key, eastern and western bay sediments,

respectively.

37

Figure 5. Depth profiles of microbial sulfate reduction rates incubated at in situ temperatures in

intact sediment cores October 2005 (27oC) and July 2006 (31oC) from western (Rabbit Key and

Buchanan Bank) and eastern (Black Betsy Key and Eagle Key) Florida Bay.

38

Figure. 1

Florida Bay

BB

EK

Everglades National Park

25o 0

7′00″N

24o 5

5′20″N

10 km

GMK

North

BLBK

RK

Gulf of Mexico

81o0′0″W 80o30′0″W

39

Figure 2.

6,6

6,8

7,0

7,2

7,4

7,6

7,8

Eagle KeyBuchanan BankRabbit Key

pH

Control 150 μmol H2SO4 400 μmol H2SO4 800 μmol H2SO4

Black Betsy Key

0,0

0,5

1,0

1,5

2,0

iP (μ

mol

L-1)

30

40

50

60

70

F- (μm

ol L

-1)

0 1 2 3 4 5 6 71,0x104

1,5x104

2,0x104

2,5x104

3,0x104

3,5x104

Ca

(μm

ol L

)

Time (h)

0 1 2 3 4 5 6 7

Time (h)

0 1 2 3 4 5 6 7

Time (h)0 1 2 3 4 5 6

Time (h)

PO43-

(μm

olL-1

)

7.87.6

7.27.4

7.06.86.6

2.0

1.5

1.0

0.5

0.0

15

20

25

3035

Ca2+

(mm

olL-1

)

30

40

50

60

70

F- (μm

olL-1

)

control150 μmol H2SO4400 μmol H2SO4800 μmol H2SO4

pH

Rabbit Key Buchanan Bank Black Betsy Key Eagle Key

Time (h)1 2 3 4 5 6

101 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6

40

Figure 3.

0 20 40 60 80 100 120 140

0.00

0.02

0.04

0.06

0.08

0.10

0.12

[iP] (

μmol

g d

w)

[H2SO4] (μmol g dw-1)

a

PO43-

(μm

olg

dry

wt-1

)

0 20 40 60 80 100 120 140

0.00

0.05

0.10

0.15

0.20

0.25

[F] (

μmol

g d

w-1)

[H2SO4] (μmol g dw-1)

b

F- (μm

ol g

dry

wt-1

)

0 20 40 60 80 100 120 1400

25

50

75

100

125

150

175 Rabbit Key Buchanan Bank Black Betsy Key Eagle Key

[Ca]

(μm

ol g

dw

-1)

[H2SO4] (μmol g dw-1)

Ca2+

(μm

ol g

dry

wt-1

)

c

(μmol g dry wt-1)

41

Figure 4.

0 50 100 150 200 2500.0

1.0x103

2.0x103

3.0x103

4.1x104

4.2x104

4.3x104

Slur

ry w

ater

33P

(Bq)

Time (min)

Eagle Key Green Mangrove Key

42

Figure 5.

BB

October 2005 July 2006

RK

BLBK EK

0

2

4

6

8

10

12

14

16

18

0

2

4

6

8

10

12

14

16

18

20

Dep

th (c

m) Oct 05

Jul 06

250 750 1250 250 1250750SRR (nmol cm-3 d-1)