Oceanogr.,46(6). 2001, 1278-1286
Limnitiol.
Inc.
?D2001. by theAmericanSocietyof Limnologyand Oceanography,
Biogeochemicaleffectsof ironavailabilityon primaryproducersin a shallowmarine
carbonateenvironment
RandolphM. Chambers1
Biology Departmentand Virginia Instituteof Marine Science, College of William and Mary, Williamsburg, Virginia
23187
JamesW. Fourqurean
Biology Departmentand Southeast EnvironmentalResearch Center; Florida InternationalUniversity,Miami, Florida 33199
StephenA. Macko
Departmentof EnvironmentalSciences, Universityof Virginia, Charlottesville,Virginia 22903
Regis Hoppenot
Biology Department,Fairfield University,Fairfield,Connecticut 06430
Abstract
We completeda synopticsurveyof iron,phosphorus,and sulfurconcentrations
in shallow marinecarbonate
sedimentsfromsouthFlorida.Totalextractedironconcentrations
typicallywere <50 ,umolg-' dryweight(DW)
mostly
and tendedto decreaseaway fromtheFloridamainland,whereastotalextracted
phosphorusconcentrations
were <10 /ctmol
of reduced
g-I DW and tendedto decreasefromwestto east acrossFloridaBay. Concentrations
sulfurcompounds,up to 40 ,umolg-' DW, tendedto covarywithsedimentironconcentrations,
suggestingthat
sulfidemineralformation
was iron-limited.
An indexof ironavailabilityderivedfromsedimentdatawas negatively
a concentrations
theclose couplingof sediment-water
colrelatedwithchlorophyll
in surfacewaters,demonstrating
columnprocesses.Eightmonthsafterapplyinga surfacelayerof ironoxide granulesto experimental
plots,sediment
and sulfurwere elevatedto a depthof 10 cm relativeto controlplots.Biomass of the seagrass
iron,phosphorus,
Thalassia testuidinuim
was not different
betweencontroland ironadditionplots,but individualshootgrowthrates
were significantly
higherin experimental
plots after8 months.Althoughthe ironcontentof leaf tissueswas sigin phosphorus
contentof T. testutdinuin
leaves was observed.
nificantly
higherfromironadditionplots,no difference
Iron additionalteredplantexposureto freesulfide,documentedby a significantly
higher634S of leaf tissuefrom
but
experimental
plotsrelativeto controls.Ironas a bufferto toxicsulfidesmaypromoteindividualshootgrowth,
in carbonatesediments.
phosphorusavailabilityto plantsstillappearsto limitproduction
to benthicprimaryproducers(Jensenet al. 1998) and retardingthereleaseof P fromthesedimentsto theoverlying
watercolumn(Chamberset al. 1995). Second,ironmayrethemas iron-sulfur
act withsedimentsulfides,precipitating
mineralsand "buffering"the effectsof toxic sulfideson
primary
producers(Carlsonet al. 1994; Terradoset al. 1999;
Erskineand Koch 2000). Because of bothdirectand indirect
effectsof ironon thegrowthof plantsand algae, thedistrimarineenvironbutionand availabilityof ironin different
of
the structure
mentsare important
variablesinfluencing
ecosystems.
and
A marinesystemideal forstudyingthe distribution
geochemicaleffectsof iron is the shallow carbonateenvii Corresponding
author([email protected]).
whererelativeto teiTigenous
systems,theconcenronment,
Acknowledgments
trationsof iron are very low (Duarte et al. 1995). Pyrite
Sedimentsamplesfromthesouthwest
Floridashelfwereobtained formation
in marinecarbonateentypicallyis iron-limited
by Brad Petersonand theSERC seagrassmonitoring
team.Thanks
vironments
(Berner1984), and freesulfidesproducedas a
to Kevin Cunniff,
Lisa Millman,Toru Endo, Susie Escorcia,and
of sulfatereductiontendto accumulateto high
by-product
Dana Madeux forfieldand lab assistance.Principalfundingforthe
concentrations
(Barberand Carlson 1993). In biogenicsedprojectwas providedby NationalPark Service grantCA5280-7without
muchiron,P sorptionto calciumcarbonate
iments
9024; additionalfundingforlaboratoryanalysisand data interprelow P
mineralsis theprincipalmechanismformaintaining
tationwas providedby Florida Coastal EvergladesLTER (NSF/
in solution(Rosenfeld 1979; Morse et al.
OCE-9910514). This is contribution148 from the Southeast concentrations
Environmental
ResearchCenter.
primaryproducersP-limited(Powell et
1985) and rendering
The effect of iron limitation on growth of primary producers in marine systemshas been shown forphytoplankton
(Martin and Fitzwater 1988; Barber and Chavez 1991; Coale
et al. 1996) and suggested for seagrasses growing in carbonate sediments (Duarte et al. 1995). In addition to its observed biochemical effectson photosynthesis(Geider et al.
1993) and suspected control of nitrogenfixation(Wu et al.
2000), iron participates in at least two other geochemical
reactions influencingprimaryproductionin marine systems.
First,reactive iron can serve as a sorptionsite for inorganic
phosphate, therebychanging the availability of phosphorus
1278
Iron effectsin carbonate environmenfts
1279
density,
thenashed at 450?C for5 h. For sedimentscollected
fromthe southwestFlorida shelf,the ashed samples were
resuspendedin 10 ml of 1 N HCl and placed on a shaker
to
of acid was sufficient
table for24 h. This concentration
dissolve the carbonatematrixof the sedimentsand associbut did not dissolve moreresistant
ated iron monosulfides
ironsulfidecompoundssuch as pyrite.The solutionwas filwas analyzedcoloritered(WhatmanGFC) and the filtrate
foriron (Fe) and phosphorus(P) using standard
metrically
on
methods.HCl-Fe and totalextractedP were determined
these samples. For sedimentscollectedfromFlorida Bay,
by summing
HCl-Fe and totalextractedP were determined
the resultsof a four-stepsequentialextractionscheme for
sedimentsmodifiedafterJensenet al. (1998). The fourfracextracted
by
tionsincludedironand phosphorussequentially
acetate
1 N MgCl,,buffered
citrate-dithionite
(BD), buffered
we definedthe "reactive
(Ac), and 1 N HCl. Furthermore,
iron" fractionas the suM of MgCl2-Fe+ BD-Fe + Ac-Fe,
withmorechemicallyresistant
detritalironextractedby the
finalHCl step.
A two-stepextraction
schemewas completedto determine
Materials and methods
totalsedimentsulfideconcentrations
(Chamberset al. 1994).
Site description-Thestudywas completedby sampling Acid-volatilesulfide(AVS) was extractedwith1 N HCl and
carbonatesedimentslocatedin FloridaBay and theadjacent includedfreedissolvedsulfideand ironmonosulfide(FeS);
sulfide(CRS) was extractedwithboilchromium-reducible
southwestFlorida shelfin the Gulf of Mexico (-24.5?N,
HCl and was assumed
81.0?W). FloridaBay, as definedby the shallowmarine/es- ing,reducedchromium/concentrated
of CRS-Fe, therefore,
tuarineenvironment
borderedon thenorthby themainland to be pyrite(FeS2). The concentration
of CRS and was added to HClof Florida,on thesoutheastby theFloridaKeys islandchain, was halfthe concentration
and on the west by the boundaryof EvergladesNational Fe to calculatetotalextractedFe in sediment.
of total
maps of the concentrations
Spatiallyinterpolated
Park,is a veryshallow(<2 m deep) embayment
generally
and sulfidein surfacesediments
extractediron,phosphorus,
lined by fine-grained
carbonatesedimentssupportingseaiso(pointkriging,
A strong wereproducedusinga krigingalgorithm
grass beds, dominatedby Thalassia testudinum.
gradientof decreasingseagrassbiomass and primarypro- tropiclinearvariogrammodel withno driftand no nugget).
ductionfromwest to east in FloridaBay is a consequence
Iron enrichmeent
experiment-Theeffectsof ironaddition
of decreasingP availabilityin pore wateralong the same
to carbonatesedimentswere measuredin a dense T. testugradient(Fourqureanet al. 1992a,b).
The 51 samplingstationson the southwestFloridaShelf dinumseagrassbed in water1 m deep in RabbitKey Basin,
westernFloridaBay (24?58.8'N, 80?50.8'W). In November
wereall locatedin less than20 m of waterand typicallyhad
muddy-sandcarbonate sedimentsand sparse macroalgal 1998, three0.25-iM2plots were establishedforironenrichcommunities
insteadof seagrasses.Twentyof the24 stations ment;threeadjacentplotsservedas controls.To each of the
plots,ironoxide granules(PeerlessC)Ironfiles
sampledin Florida Bay make up a portionof the Florida experimental
ETI 8/50,surfacearea 1.2944 m2 g- 1) were shakenout and
Bay monitoring
programthathas trackedsurfacewaterqual0.5 g cm-. In Novemspreadto a densityof approximately
itysince 1989 (Fourqureanet al. 1993; Boyer et al. 1997,
ber (priorto ironaddition),2 monthsafteraddition(January
a (Chl a)
1999); we obtaineddata on themean chlorophyll
1999) and 8 monthsafteraddition(July 1999), sediment
concentrations
in the watercolumnat the 20 Florida Bay
stationsfortheperiodJune1989-July1999 fromthismon- cores to a depthof 30 cm were collectedfromeach of the
controland experimental
plots.Subcoreswerethentakenat
itoringprogram.
intervalsof 2.5-5 cm down thelengthof each core; 0.5-ml
of bulk
sedimentsampleswerecollectedfordeterminations
Synopticsurveyfor iron,phosphorus,and sulfur-At all
and forHCl-Fe, totalP, and totalsulfide
75 stations,0.5-ml sedimentsampleswere collectedin dusedimentproperties
usinganalyticalproceduresdescribedforsurplicateor triplicateusingan open-end3-ml syringepushed concentrations
1 cm intothesedimentsurface.Sampleswereplacedin tared face sediments.
The responseof the seagrassto ironadditionswas deteranalglass vials,capped,and kepton ice priorto laboratory
mined in Januaryand July1999 by biomass and growth
ysis forbulk sedimentpropertiesand extractableiron and
of T. testudinum
measures.Net abovegroundproductivity
phosphorus.A second set of 0.5-ml sedimentsampleswas
zinc acetatesolutionto sta- was measuredin bothJanuaryand July1999 at each experplaced in 3 ml of 1 M buffered
bilize reducedsulfurcompoundspriorto laboratoryextrac- imentaland controlplotusinga modifiedleafmarkingtechtionand analysis.
nique (Zieman 1974; Ziemanet al. 1999). A single200-cm2
in each plot;withineach quadSamples forbulk sedimentpropertieswere weighedwet, quadratwas placed arbitrarily
oven driedat 80?C for4 d, weighedto determine
drybulk rat, all shortshoots of the seagrass T. testudinumwere
al. 1989; Fourqureanet al. 1992a, 1993; Erftemeijer
1994).
By reducingsulfidetoxicityand contributing
to phosphorus
limitation,
however,thegeochemicalbehaviorof ironcould
createconditionsthatsimultaneously
enhanceand decrease
plantproduction.
To date,therelativeeffectsof sulfidetoxicity and phosphoruslimitationon primaryproductionin
are unknown.
iron-poorenvironments
In thepresentstudy,we determined
the surfacesediment
concentrations
of iron,phosphorus,and sulfurthroughout
of southFlorida,in areas vegetated
carbonateenvironments
by rootedseagrasses(FloridaBay), and by calcareousalgae
(southwestFloridashelf).We also comparedtheavailability
of reactiveiron in sedimentswith watercolumnprimary
production.
Finally,we measuredtheeffectsof ironenrichmenton phosphorusretention,
toxic sulfidebuffering,
and
froma sitein FloridaBay. Overall,our
seagrassproduction
thedual geochemicalinfluence
ofiron
goal was to determine
on sulfurandphosphorus
in carbonatesedimentsandto evaluate thepotentialeffecton primaryproducers.
1280
Chamberset al.
markedby drivingan 18-gaugehypodermic
needle through
the base of the leaves. Care was takennot to disturbother
plant and animal taxa in the quadrats.The markedshort
shootswere allowed to grow for 10-14 d, afterwhichall
abovegroundseagrassmaterialin thequadratswas harvested. Plant materialwas separatedby seagrass species, and
leaves of all species were counted,measured(lengthand
widthto nearestmm),cleanedof epiphytesby gentlescraping, and driedto constantmass at 70?C. We quantifiedT.
testudinum
standingcrop as the dryweightof greenleaves
as dry weightof green
per square meterand productivity
leaves producedper shortshootper day.
Because phosphoruscontentin seagrassleaves is a good
proxyforthe availabilityof P to plants(Fourqureanet al.
1992a), we measuredP contentof leaves collectedfromboth
and controlplots in Januaryand July1999.
experimental
Fromeach plot,fiveintactshortshootsof T. testudinum
were
collectedarbitrarily;
whenHalodule wrightii
was present,
we
collected20 intactshortshoots,whichwerereturned
to the
lab. All attachedgreenleaves werecutfromtheshortshoots
and cleaned of adheringepiphytesby gentlyscrapingwith
a razorblade. All leaves froma sitewerepooled and dried
at 80?C. Dried leaves weregroundto a finepowderusinga
ceramicmortarand pestle.Powderedsampleswereanalyzed
in duplicateusinga dryoxidation,acid hydrolysis
extraction
followedby a colorimetric
analysisforphosphate(Fourqurean et al. 1992a). To determine
whetherironenrichment
of
sedimentswould lead to increasedFe uptakeby plants,the
Fe contentof leaves was determined
usinga similaroxidaand colorimetric
tion/acid
extraction
analysisscheme.
Plant exposureto sulfidestresswas measuredby deterofleaftissue.Sedminingthestablesulfurisotopicsignature
imentsulfidesare generatedby microbialsulfatereduction,
a process that fractionatessulfurisotopes very strongly
(Chambersand Trudinger
1979). The sulfidepool is depleted
in theheavier34S isotoperelativeto thesourcesulfate.This
fractionation
providesa toolforexaminingtherelativeeffect
of sulfideson plants,because thestablesulfurisotopiccompositionof planttissuesis determined
by the sulfursource
(Carlson and Forrest1982; Striblinget al. 1998). We hypothesizedthatwithremovalof isotopicallylightersulfide
via authigenicpyriteformation
in ironadditionplots,an increase in the 834S contentof seagrassleaves would indicate
an alleviationof sulfideexposure.
Samples of dried seagrasseswere convertedto SQ2 for
isotopeanalysisusinga Carlo Erba elementalanalyzer(EA)
coupledto an OPTIMA stableisotoperatiomass spectrometer(Micromass).In theEA, theorganicsulfuris pyrolyzed
at 1,050?Cusinga combination
oxidationand reductionfurnace system.The resultingSO2 gas is chemicallydriedand
injecteddirectlyinto the source of the mass spectrometer.
The stableisotopicratiois reportedas
534S =
-1]
[Rsample/Rstandard
X 103(%o),
whereR is theabundanceratioof theheavyto lightisotopes
of sulfurand is correctedforthemass overlapwith
(34S/32S)
theisotopesof oxygen.The international
standardfor34S iS
theCanyonDiablo Troilite(CDT), whichis assigneda 834S
value of 0.0%o.The reproducibility
of themeasurement
typicallyis betterthan +0.2%o whenusingthecontinuousflow
interface
on theOPTIMA. In thelaboratory,
thesamplesare
commonlymeasuredagainsta sulfurdioxide gas thathas
been calibratedagainstNBS 127.
Results of the iron additionexperiments
were analyzed
usingANOVA withsignificance
level P ' 0.05. We treated
timeand treatment
as maineffects;we did notuse a repeated
measuresdesignbecause we did not samplethe same short
shootseach samplinginterval,nor did we recordthe plot
withina treatment
fromwhicheach sampleoriginated.
Results
Synopticsurveyofiron,phosphorus,and sulfur-Relative
to terrigenous
Fe concentrations
sediments,
totalextracted
in
FloridaBay wereextremely
low and exhibiteda stronggradient of decreasingconcentration
away fromthe Florida
mainland(Fig. 1). The concentrations
in FloridaBay ranged
froma highof 66 ,umolg-' dryweight(DW) in thenortheast cornerof FloridaBay to a low of 5.2 ,umolg-I DW at
the southernbay boundary.Fe concentrations
on thesouthwestFloridashelf,in contrast,
generallywerelowerand exhibitedno clear gradient.Some higherconcentrations
were
measurednearestthewesterncoast of Florida,and a distinct
regionof low concentration
was foundwest of the Florida
mainland.
Sedimentconcentrations
of total P fromthe southwest
Florida shelfwere distributed
in a patternsimilarto iron,
withlow concentrations
foundwestof theFloridamainland
and higherconcentrations
nearthecoast and alongthewesternboundaryof FloridaBay (Fig. 1). A gradientof decreasing sedimentphosphorusconcentrations,
however,was observedfromwestto east acrossFloridaBay.
On average,mostof thereducedsulfurcompoundsin sedimentswere extractedas CRS, with<10% as AVS. Total
sedimentsulfideconcentrations
were higherin FloridaBay
thanfromthe southwestFlorida shelfand generallyhigher
neartheFloridamainland(Fig. 1).
For FloridaBay, we plottedthe distributions
of reactive
Fe and CRS-Fe and thencalculatedan index of ironavailability(Fig. 2). The index measuresthe degreeof sulfidization(Chamberset al. 2000) and computesthefractionof
the ironpool available forreactionwithfreesulfides(i.e.,
reactiveFe/[reactive
Fe + CRS-Fe]). The small amountof
AVS-Fe relativeto CRS-Fe could not be determinedand
includedin the calculation,so the index is a slightoverestimateof iron availability.Lowest values of the index occurredin the north-central
basin of FloridaBay, wherethe
capacityof reactiveironto buffersulfidetoxicityvia ironsulfidemineralformationis diminished.In contrast,the
highestvalues of theindexoccurredalong thewesternmargin and in thenortheastern
portionsof thebay.
We thencomparedthesedimentindexof ironavailability
withthe1997-1998 annualaveragewatercolumnChl a concentrationsobtained fromthe Florida Bay water quality
monitoring
program(Boyer et al. 1997, 1999). The index
was negativelycorrelatedwithwatercolumnChl a concentrations(r2 = 0.54) (Fig. 3). Higherindicesof reactiveiron
availabilityin sedimentsoccurredwherewatercolumnconcentrations
of Chl a were low.
Iron effectsin carbonateenvironments
Total ironin sediment
(g.ml
14
4
30450
I0I020
4
4t:^
|
1____
26
24
22
20
18
16
14
12
10
ReactiveFe (g mol g'-1)
470
~~~~~kilometers
10 20 30 40 50
00
|
g-1)
1281
25ON
4.
.
.i ' 1'"91?W
~
o
X
O
5.
CRS-Fe (mol~~~ g-1)g
~~~~~24
otal phosphorusin sediment
_
2
4
6
8
10
15 20
"O
2
26
22
20
_
18
16
14
12
10
8
6
(pmol g1)
0
4
lckilometers
10 12
4
2
0
_r~~~~Toa
sufri
sdmn
Index ofiron availability
0.75
0.70
0.65
0.60
~~~~~~~~~~~~~~1
I
t +
_
Flori
S _ 4
~~~~It
ef,andc
4_
44
)n
{ Amol
0.55
1
0.40
25 30 35
l 5 10
Pa 15 A fm
a 4s I
ey_ ot |~~~~
m osal sumplfinsrom
se di
0.35
y
4
'' _44.
4
th 7 sapig
sttin shon. TesuhFoid
manndi
ope symbl<r
apestsfo
hotws
shwLnwie
Floid shlf an loesmos
ar sapestsfomFoiaBy
0.30
~~~~~~~0.25
ofreactive
Fe andCRS-Fethroughout
survey
Fig.2. Synoptic
andthecalculated
indexofiron
FloridaBay Q.tmol
g' dryweight)
availability. Isopleths of equal value were generated by spatial in-
stations
arounddata collectedfromthe24 sampling
terpolation
shown.
Iron enrichment experiment-In
January, 2 months after
of HCI-Fe, totalP,and total
ironaddition,theconcentrations
sulfide in the upper 2.5 cm of sediment were higher and
morevariablein theironadditionplotsrelativeto theinitial
and to controls(Fig. 4). In July,
sedimentconcentrations
after 8 months, the effect had extended down-core to a depth
bioturbaof 10 cm, perhapsdue to geochemicalmigration,
tion, or both. Althoughpore-watersulfideconcentrations
were not measured,free sulfidecould not be detectedby
smell in the upper 10 cm of sedimentfromiron addition
1282
Chambers et al.
4
4
-Fe
~J3
Control
additionT
-\3
2
2
1~~~
Chl a =-5.85 (index) + 4.79
r2=
0
0.54
p<O.OOl
0.2
0.3
0
0.4
0.5
0.6
0.7
plots. We also observed differencesin the amounts of Fe, P,
and S below the zone of influence of iron additions. For
example, the down-core concentrationsof sulfurin iron addition and control plots were measured up to 60 ,umol g-
January
0
0
10
10-
-0--
20
~
30
0
100
*~
200
Fe Addition
300
Y--- Fe Addition
0
100
Iron (,umolgdw-1)
200
300
400
Iron (imol gdw-1)
0.
0-
10
10-
20
20-
30
30-
0
50
100
150
0
200
Phosphorus(,umolgdw-')
0-
0'
100
150
200
X10 -
20-
20-
30-
30-
0
50
Phosphorus(,umolgdw-1)
lo0-
20 40 60 80 100 120 140 160
Sulfur(,tmolgdw-')
Jan99
Jul99
50
iita
30
400
.t.
150
of Thalassia testutdiinumi,
Fig. 5. Standingcropand productivity
2 and 8 monthsafteriron additionsin Januaryand July,respectively.
20---
ia
L
0
July
Jul99
200T
Index ofironavailability
Fig. 3. Plot of watercolumnChl a concentration
versusthe
sedimentindexof ironavailabilityin FloridaBay.
Jan99
0
20 40 60 80 100 120 140 160
Sulfur(,tmolgdw-1)
Fig. 4. Resultsof ironadditionto carbonatesedimentson HClFe, totalP and S profiles.Data fromcontroland experimental
plots
2 and 8 monthsafterironadditions(in Januaryand July,respectively)are plottedalong withdata collectedpriorto ironaddition.
DW in Januarybutwereless than20 ,umolg-I DW in July.
and controlplots
Also, sedimentsfromboth experimental
werehigherin phosphorusrelativeto initialconditionsmeasuredin November.
The measuredresponseof the seagrassT. testudinurnto
was
ironenrichment
was variable.Shortshootproductivity
significantly
higherin ironadditionplotsin Julyduringthe
growingseason (t-test,t = -3.6, df 4, P = 0.02) (Fig. 5).
In contrast,seagrass standingcrop was significantly
lower
in ironadditionplotsin January(t = 3.764, P = 0.02), with
no difference
betweenironadditionand controlplotsin July
(t = 0.558, P = 0.61).
The ironcontentof T. testudinumleaves was significantly
higherin the iron additionplots in both Januaryand July
(ANOVA, TREATMENT main effect,F = 7.2, P = 0.02)
(Fig. 6). Leaf ironcontentdecreasedfromJanuaryto July
(ANOVA, TIME main effect,F = 8.0, P = 0.01) . The
and control
colonizedbothexperimental
seagrassH. wrightii
plotsbetweenJanuaryand July.In July,theironcontentof
H. wrightiialso was higherin theironadditionplots(t-test,
t = -2.85, df 4, P = 0.05).
was
The phosphoruscontentof leaves of T. testudinum
higherin JulythanJanuary(ANOVA, TIME maineffect,F
= 9.7, P = 0.01), buttheP contentof leaves fromtheiron
differadditionplotsand controlplotswas not significantly
ent (ANOVA, TREATMENT main effect,F =0.58, P=
Iron effectsin carbonate environments
January
0.8
January
July
20
Control
Fe addition
0.6
IT
m Control
-15
-5
10
T
CZ
CA
no
0.0
data
TlhalassiaHalodule
l
ThlalassiaHalodile
5
0
no
~~~~data
-5
Thalassia Halodule
0.16
July
OMFe addition
0.4
0.2
1283
Thalassia Halodule
Fig. 7. 834S signatureof seagrassleaves fromThcalassiatestuc2 and8 months
dinunm.
andHalodiulewrightii,
afterironadditions
in January
andJuly,
respectively.
0.14
skeletalremainsto thissedimentinorganismscontributing
clude
calcareous
green
algae, seagrassepiphytes(spirorbid
0.12
polychaetes,soritidforaminiferans,
encrusting
corallinealgae), mollusks,and stonycorals(Nelsenand Ginsburg1986;
Bosence 1989; Frankovichand Zieman 1995). Iron distri0.10 T
no
butionin these sediments,howevei; is somewhatvariable
data
and the mechanismsfordepositingiron and generatingits
0.08
observeddistributions
in southFlorida hydroscapeare not
ThlalassiaHalodule
ThlalassiaHalodule
of ironare fairlyhigh
well known.First,theconcentrations
(Duarteet al. 1995;
Fig. 6. Iron and phosphoruscontentof leaf tissuefromThail- relativeto othercarbonateenvironments
2 and 8 monthsafteriron
assia testutdinilisn
and Halodule wrightii,
sedKoch et al. 2001), althoughnotrelativeto terrigenous
additionsin Januaryand July,respectively.
iments(e.g., Chamberset al. 2000). In FloridaBay, theiron
gradientdecreasingaway fromthe Florida mainlandsuggests a continental,clastic source (Eugene Shinn pers.
TIME by TREATMENT
0.46), nor was therea significant
comm.),
possiblyderivedfromtransport
throughthe Ever=
=
interaction
(ANOVA, F
0.004, P
0.95) (Fig. 6). The P
of iron in sediments
contentof H. wrightiileaves collectedin Julywas higher glades. The more patchydistribution
Floridashelfmaybe indicativeof a nonfromironadditionplotsat P = 0.08 (t-test,t = -2.287, df fromthesouthwest
of
iron
uniform
source
or sedimentredistribution
by water
4).
eolian
or
other
currents,
transport,
physical
process
(e.g.,
Both samplingtimeand ironadditionhad a markedeffect
Muhs et al. 1990; Duce and Tindale 1991).
on the sulfurisotopiccompositionof seagrassleaves (Fig.
Based on our results,totalFe and P exhibitsimilarpat7). The 834S signatureof T. testudinumwas higherin July
externs
of abundancein shelfsediments.Using a different
thanJanuaryin controlplots (ANOVA, TIME main effect,
et
traction
Yarbro
al.
found
that
Fe
scheme,
(pers.
comm.)
=
=
F
65.7, P
0.001). The 834S signaturealso was higher
in ironadditionplots(ANOVA, TREATMENT maineffect, and P species werepositivelycorrelatedin FloridaBay sedF = 63.9, P = 0.001). In January,
T. testudinunleaves from iments.We view the covariationin Fe and P fromshelf
sediments(Fig. 1) and frombay sediments(Yarbroet al.
controlplotshad an average834Sof -2.6 ? 1.4%ocompared
is to some extent
to + 6.7 ? 1.2%oin ironadditionplots.In July,T. testudinumn 1997) as evidence thatP concentration
leaves fromironadditionplotsalso wereisotopicallyheavier associatedwithFe availability,even in iron-poorcarbonate
environments
whereP sorptionto calcium carbonateminthanleaves fromcontrolplots(+ 12.3 ? 0.2%ocomparedto
erals is the principalgeochemicalmechanisminfluencing
+ 6.7 ? 0.2%o). Comparedto T testudinumn,
leaves of H.
aqueous phosphorusconcentrations
(Morse et al. 1985). We
weremoreenrichedin 34S (Fig. 7); in July,theisowrightii
of H. wrightii
leaves fromironadditionplots observeda decreasinggradientin sedimentP fromwest to
topicsignature
east acrossFloridaBay, consistentwithpreviousworkdocwas higherthancontrolplots (+ 19.9 ? 0.4%ocomparedto
umentingphosphoruslimitationof seagrass biomass and
+11.3 ? 0.4%o;t-test,t = -14.2, df 2, P = 0.005).
productionalong the same gradient(Fourqureanet al.
1992a).
Discussion
The computedindexof ironavailabilitywas lowestin the
The fine-grained
basin of FloridaBay, a regionwhereextensive
carbonatesedimentsof our studyarea in
north-central
southern
Floridaare of biogenicorigin.The mostimportant seagrassdie-offoccurredovera decade ago (Fourqureanand
1284
Chacmbers
et al.
Robblee 1999). The low indexcould be eithera cause or a
consequenceof seagrassdie-off,butthefactthata majority
of sedimentFe in thisregionhas been used forFe-S mineral
formationsuggestsa stronglinkagebetweenseagrassproand sedimentFe-S geochemistry.
duction/decomposition
Furthermore,
this linkage extends to the water column,
wherewe founda significant
negativecolTelationbetween
sedimentiron availabilityand watercolumnprimaryproof these data is thatwitha reduction.Our inter-pretation
ductionin sedimentironavailabilityvia seagrassdecomposition,mineralizedor iron-boundphosphoruscan migrate
fromthe sedimentto the watercolumn,therebypromoting
basin (e.g., Phlipset al.
in thenorth-central
algal production
1999). Ratherthanphytoplankton
beingiron-limited
(Martin
and Fitzwater1988), we speculatethatlow availabilityof
of
reactiveironin surfacesedimentsoccurswherediffusion
P into the watercolumnstimulatesalgal blooms. In these
of labile organicmattermay
areas,anaerobicdecomposition
releasemoreP thancan be biogeochemically
retainedin the
sulfideto depletethe
sedimentsand maygeneratesufficient
reactiveironpool. Studiesof nutrient
fluxbetweensediment
and watercolumnarerequiredto testtheeffectsofdecreased
ironavailability.
of increasediron availability
What are the ramifications
in sediments?First,the sedimentpool of P was increased
by iron addition,but we are not sure whetherthis P was
available for seagrass growth.We measureda significant
changein foliarP forT. testudinum
betweensamplingdates
but saw no significant
difference
betweencontroland iron
additionplots (Fig. 6). Even thoughmore totalP was reP mayhave been in a forminaccessible
tainedin sediments,
by T. testudinumn
(e.g., sorbedto ironor carbonateminerals).
In contrast,
ironadditionled to significantly
greaterironuptake by seagrass,but tissue Fe was lower in Julythanin
January.The decreasein foliarFe fromironadditionplots
mineralproprobablywas due to theincreasediron-sulfide
ductionduringthegrowingseason (Fig. 4).
accumulationof sulfideminUsing the depth-integrated
eralsin ironadditionplots,we calculateda sulfidedeposition
rate of approximately130 /Cmol cm-2 in 8 months.This
amountof sulfidebuffering
is comparableto an estimated
sulfatereductionrate of 200 fLmolcm-2 yr-I in vegetated
carbonatesedimentsofFloridaBay (Ku et al. 1999). In other
sulfidepoolwords,a significant
portionof thepore-water
locationby Carlsonet
measured>1 mM at our experiment
in
al. (1994)-was sequesteredby sulfidemineralformation
ironadditionplots.Withisotopicallylightersulfideremoved
frompore watei;a heaviercombinedsulfate+ sulfidepool
was availableforuptakeby seagrassesin ironadditionplots
relativeto controls(Fig. 7). All 834S signatureswere less
thansurfaceseawater(+21), howevei;suggestingtheinfluence of activesulfatereduction/oxidation
cycles in theroot
of elementalsulfur(Canzone, bacterialdisproportionation
fieldand Thamdrup1994), or both.Furthermore,
thelighter
seagrass834S values in Januaryvs. Julymay reflecta seasonal shiftin the relativeimportanceof sulfideoxidation;
as pyriteduringthesummergrowlightsulfideprecipitated
ing season may be reoxidizedduringthewinter,
generating
lightersulfateforrootuptake(Striblinget al. 1998).
Whetherby iron-mediated
pyriteformation
or otherpro-
cesses, lowerpore-water
sulfideconcentrations
in vegetated
sedimentswouldprovideobviousbenefitsto seagrassessensitiveto sulfide(Erskineand Koch 2000). Owing to nonhomogeneity
of seagrassdensityat the initiationof theexperiment,
however,T. testudinunstandingcropactuallywas
lower in the ironadditionplots in January.An increasein
standingcropto thelevel of controlplotsmayhave reflected
the observedfastergrowthrate of T. testudinumn
shootsin
theironadditionplotsin July(Fig. 5).
Colonizationof controland iron additionplots by H.
wrightiiallowed for a comparisonof the responseof two
has a less roseagrassspecies to ironadditions.H. wrightii
and its
bust root and rhizomesystemthan T. testudinurn,
rhizomesare notburiedas deeplyin thesedimentas T. testudinum(-3-5 cm, comparedto -20 cm at ourexperimental site). For bothspecies,foliarFe was higherin ironadditionplots,indicatingthatdissolvedFe was moreavailable
for root uptake.In contrastto T. testudinumi,
however;P
contentof H. wrightiiwas higherin theironadditionplots
(at P = 0.08) (Fig. 6), suggestinggreaterP availabilityto
H. wrightii.Also, H. wrightiiwas isotopicallyheavierthan
T. testudinumn,
consistentwiththe presumedspecies differences in sulfide exposure. The more shallow-rootedH.
was exposedto lowersulfideconcentrations;
hence,
wrightii
in H. wrightii
the34S signature
leaves was heavierthandeepIn theone previouscomparisonof
ly rootedT. testudinumn.
the sulfurisotopiccontentof thesetwo species,H. wrightii
was more enrichedwith34S thanT. testudinun(Fry et al.
thatdeep- and shallow1982). These resultsdemonstrate
condirooted seagrasses are respondingto environmental
tionsat different
depthsin thesediment(Williams1990).
SeagrassesfromthesouthFloridahydroscapedo notseem
foliarconcentrations
of ironfromcontrolplots
iron-limited;
in FloridaBay were above thecriticallevel forironlimitationsuggestedby Duarteet al. (1995), and in fact,foliarFe
actuallydecreasedduringthe growingseason. Furthermore,
the absence of an increasein P contentof the dominant
seagrass and only a modest increase in individualshoot
did not enhance
growthratesindicatethatironenrichment
P availabilityfor T. testudinumn.
Even withiron-controlled
exalleviationof sulfidestress,we suspectthatP limitation
ertsa primarycontrolon T. testudinumn
productionin carbonatesediments.More researchis necessaryto determine
whetherthe indirect,geochemicaleffectsof iron additions
of shallow-rooted
influenceprimary
significantly
production
and of benthicalgae.
seagrassspecies like H. wvrightii
in sedimentsof FloridaBay and the
Iron concentrations
southwestFloridashelfare variableand low relativeto terrigenoussedimentsbutare highrelativeto otherstudiedcaris ironbonateenvironments.
Still,sedimentpyriteformation
limited.In FloridaBay,thecalculatedindexof sedimentiron
available to buffertoxic sulfidescorrelatesnegativelywith
a strongbenthic-pelagic
couwatercolumnChl a, indicating
pling in this shallow subtropicalembayment.Additionof
ironto vegetatedsedimentsincreasespyritemineralformation,reducesplantexposureto sulfide,and stimulatesindividual shootgrowthof T. testudinumn.
Iron additionalso increases sediment P but does not appear to increase
of seagrass
to overcomeP limitation
availabilitysufficiently
production.
Iron effectsin carbonate environments
1285
ratiosof thedominantseagrassThalassia testudinuin.Limnol.
Oceanogr.37: 162-171.
, R. D. JONES,
ANDJ.C. ZIEMAN. 1993. ProcessesinfluencBARBER, R. T., AND E P CHAVEZ. 1991. Regulationof primary
in theequatorialPacificOcean. Limnol.Oceanogr.
ing watercolumnnutrient
characteristics
and phosphoruslimproductivity
itationof phytoplankton
biomass in Florida Bay, FL, USA:
36: 1803-1815.
BARBER, T. R., AND P R. CARLSON,JR. 1993. Effectsof seagrass
Inferencesfromspatialdistributions.
Estuar.Coast. ShelfSci.
die-offon benthicfluxes and porewaterconcentrations
of
36: 295-314.
FRANKOVICH,T. A., AND J. C. ZIEMAN. 1995. A comparisonof
p. 530-550.
ICO2, IH2S, and CH, in FloridaBay sediments,
In R. S. Oremland[ed.], Biogeochemistry
of global change:
methodsfortheaccuratemeasurement
of epiphytic
carbonate.
Estuaries18: 279-284.
Radiativelyactivetracegases. Chapmanand Hall.
J. K. WINTERS,
ANDP L. PARKER.1982.
BERNER, R. A. 1984. Sedimentary
pyriteformation:An update. FRY,B., R. S. SCANLON,
Sulfuruptakeby saltgrasses,mangrovesand seagrassesin anGeochim.Cosmochim.Acta 48: 605-615.
sedimentsof Florida Bay.
aerobic sediments.Geochim. Cosmochim.Acta 46: 1121BOSENCE, D. 1989. Surfacesublittoral
1124.
Bull. Mar. Sci. 44: 434-453.
GEIDER, R. J.,J. LA ROCHE, R. M. GREENE, AND M. OLAIZOLA.
BOYER, J. N., J.W. FoURQUREAN,AND R. D. JONES. 1997. Spatial
1993. Responseof thephotosynthetic
apparatusof Phaeodacof waterqualityin Florida Bay and Whitecharacterization
waterBay by multivariate
analyses:Zones of similarinfluence.
tyliwin
tricornzuttrnz
(Bacillariophyceae)to nitrate,
phosphateor
ironstarvation.
J.Phycol.29: 755-766.
Estuaries20: 743-758.
H. S., K. J. McGLATHERY,
R. MARINO,ANDR. W. HoJENSEN,
, J.W. FoURQUREAN,AND R. D. JONES. 1999. Seasonal and
WARTH.1998. Formsand availabilityof sedimentphosphorus
long-term
trendsin waterqualityof FloridaBay (1989-1997).
in carbonatesand of Bermudaseagrassbeds. Limnol.OceanEstuaries22: 417-430.
CANFIELD,D. E., AND B. THAMDRUP.1994. The production
of 34Sogr.43: 799-810.
of elemen- KOCH, M. S., R. E. BENZ,ANDD. T. RUDNICK.2001. Solid-phase
depletedsulfideduringbacterialdisproportionation
tal sulfuirScience 266: 1973-1975.
phosphoruspools in highlyorganiccarbonatesedimentsof
CARLSON, P R., JR., AND J. FORREST. 1982. Uptakeof dissolved
north-eastern
FloridaBay. Estuar.Coast. ShelfSci. 52: 279sulfide by Spartincaaltermifora: evidence from natural sulfur
291.
M. L. COLEMAN,
isotopeabundanceratios.Science 216: 633-635.
R. E. BLAKE,AND
Ku, T. C. W., L. M. WALTER,
1994. Relationshipof
A. M. MARTIN. 1999. Couplingbetweensulfurrecyclingand
9 L. A. YARBRO, AND T. R. BARBER.
in Florida
sedimentsulfideto mortality
of Thalassiatestutdinumzn
carbonatedissolution:Evidence fromoxygen
syndepositional
and sulfurisotopecompositionsof pore watersulfate,South
Bay. Bull. Mar. Sci. 54: 733-746.
Florida platform,U.S.A. Geochim. Cosmochim. Acta 63:
CHAMBERS, L. A., AND P A. TRUDINGER. 1979. Microbiological
fractionation
of stablesulfurisotopes:A reviewand critique.
2529-2546.
limits
MARTIN, J. H., AND S. E. FITZWATER.
1988. Irondeficiency
Geomicrobiol.J. 1: 249-293.
Pacificsubarctic.Nagrowthin the north-east
CHAMBERS, R. M., J.T. HOLLIBAUGH, AND S. M. VINK. 1994. Sulphytoplankton
fatereductionand sedimentmetabolismin TomalesBay, CA.
ture331: 341-343.
Biogeochem.25: 1-18.
MORSE, J. W., J. J. ZULLIG, L. D. BERNSTEIN,E J. MILLERO, P.
of
MILNE, A. MUCCI, AND G. R. CHOPPIN. 1985. Chemistry
, J. W. FoURQUREAN,J. T. HOLLIBAUGH,AND S. M. VINK.
calciumcarbonate-rich
shallowwatersedimentsin the Baha1995. Importanceof terrestrially-derived,
particulate
phosphorus to phosphorusdynamicsin a westcoast estuary.Estuaries
mas. Am. J. Sci. 285: 147-185.
18: 518-526.
MUHS, D. R., C. A BUSH, K. C. STEWART, T. R. ROWLAND, AND
R. CRITTENDEN. 1990. Geochemicalevidenceof Saharandust
, J.T. HOLLIBAUGH,C. S. SNIVELY, AND J.N. PLANT. 2000.
limestones
Iron, sulfurand carbon diagenesisin sedimentsof Tomales
parentmaterialforsoils developedon Quaternary
of Caribbeanand westernAtlanticislands.Quat.Res. 33: 157Bay, California.Estuaries23: 1-9.
177.
COALE, K., AND THE IRONEX GROUP. 1996. A massivephytoplankton bloom induced by an ecosystem-scale iron fertilizationexNELSEN, J.E., JR., AND R. N. GINSBURG. 1986. Calciumcarbonate
in theequatorialPacificOcean. Nature383: 495-501.
periment
productionby epibiontson Thzalassiain FloridaBay. J. Sediment.Petrol.56: 622-628.
DUARTE, C. M., M. MERINO, AND M. GALLEGOS. 1995. Evidence
in seagrassesgrowingabove carbonatesedof irondeficiency
PHLIPS, E. J.,S. BADYLAK, AND T. C. LYNCH. 1999. Blooms of the
iments.Limnol.Oceanogr.40: 1153-1158.
picoplanktonic
cyanobacterium
Synechococcusin FloridaBay,
a subtropical
of
inner-shelf
DUCE, R. A., AND N. W. TINDALE. 1991. Atmospheric
transport
lagoon.Limnol.Oceanogr.44: 1166iron and its depositionin the ocean. Limnol. Oceanogr.36:
1175.
1715-1726.
POWELL, G. V. N., W. J. KENWORTHY, AND J. W. FoURQUREAN.
of seagrass
P L. A. 1994. Differences
in nutrient
evidencefornutrient
limitation
concentrations
ERFTEMEIJER,
1989. Experimental
circulation.Bull.
and resourcesbetweenseagrasscommunities
on carbonateand
growthin a tropicalestuarywithrestricted
sedimentsin SouthSulawesi,Indonesia.Bull. Man.
Mar. Sci. 44: 324-340.
terTigenous
waterand sedimentchemistry
Sci. 54: 403-419.
ROSENFIELD, J. K. 1979. Interstitial
on Thalassia
of two cores fromFloridaBay. J. Sediment.Petrol.49: 989ERSKINE,J.M., AND M. S. KOCH.2000. Sulfideeffects
testutdinurn
carbon balance and adenylate energycharge. Aquat.
994.
Bot. 67: 275-285.
STRIBLING,J. M., J. C. CORNWELL,AND C. CURRIN. 1998. VariFoURQUREAN,J. W., AND M. B. ROBBLEE. 1999. Florida Bay: A
abilityof stable sulfurisotoperatiosin Spartinza
altermifora.
Mar. Ecol. Prog.Ser. 166: 73-81.
historyof recent ecological changes. Estuaries 22: 345-357.
, J.C. ZIEMAN,AND G. V. N. POWELL. 1992a. Relationships TERRADOS, J.,AND OTHERS. 1999. Are seagrassgrowthand survival
betweenporewaternutrients
and seagrassesin a subtropical
constrained
by thereducingconditionsof thesediment?
Aquat.
Mar. Biol. 114: 57-65.
carbonateenvironment.
Bot. 65: 175-197.
. 1992b.Phosphorus
studiesof Caribbeanseagrass
limitation
of pri, AND
WILLIAMS, S. L. 1990. Experimental
mary production in Florida Bay: Evidence from the C : N: P
bed development.
Ecol. Monogar60: 449-469.
References
9
1286
Chaibers et al.
WU, J.,W. SUNDA,E. A. BOYLE, AND D. M. KARL. 2000. Phosphate
depletionin the westernNorthAtlanticOcean. Science 289:
759-762.
ZIEMAN,J. C. 1974. Methodsforthestudyof thegrowthand productionof turtlegrass,Thalassia testudinumlswl.
Aquaculture4:
139-143.
, J. W. FOURQUREAN,AND T. A. FRANKOVICH.1999. Sea-
grass die-offin FloridaBay: Long-termtrendsin abundance
and growthof turtlegrass,Tlhalassiatestucdiinluwn.
Estuaries22:
460-470.
Received. 29 Novemnber2000
Amnended:28 March 2001
Accepted: 23 April 2001