Grochimica
tv Cosmochimica
Ada Vol.$7,pp.991-1007
Copyright
0 1993 Pergamon
PressLtd.Printedin U.S.A.
M)l6-7037/93/$6.00
+ .OO
Authigenic apatite formation and burial in sediments from non-upwelling,
continental margin environments
KATHLEENC. RUTTENBERG*and ROBERTA. BERNER
Department of Geology and Geophysics, Yale University, New Haven, CT 06511,USA
(Received October9, 1991; accepted in revisedform August 4, 1992)
Abstract-Evidence
for precipitation of authigenic carbonate fiuorapatite (CFA) in Long Island Sound
and Mi~i~ippi Delta sediments suggests that fo~ation
of CFA is not restricted to en~ronmen~
of
active coastal upwelling. We present porewater data suggestive of CFA formation in both these areas.
Application of a sequential leaching procedure, designed specifically to separate authigenic carbonate
fluorapatite from other phosphorus-containing phases, including detrital apatite of igneous or metamorphic
origin, provides strong supporting evidence for authigenic apatite formation in these sediments. The size
of the authigenic apatite reservoir increases with depth, indicating continued formation of CFA during
early diagenesis. This depth increase is mirrored by a decrease in solid-phase organic P at both sites,
suggesting that CFA is forming at the expense of organic P. Mass balance considerations, application of
diagenetic models to interstitial water nutrient data, and the saturation state of the interstitial water are
consistent with this interpretation. Diagenetic redist~bution of phosphorus among the different solidphase reservoirs is observed at both sites, and results in near perfect retention of P by these sediments
over the depth intervals sampled. Formation of CFA in continental margins which do not conform to
the classically defined regions of phosphorite formation renders CFA a quantitatively more important
sink than has previously been recognized. Including this reservoir as a newly identified sink for reactive
P in the ocean, the residence time of P in the modem ocean must be revised downward. The implication
for ancient oceans of CFA formation in continental margin sediments other than phosphorites is that
phosphorite formation may be less a representation of episodicity in removal of reactive P from the
oceans than of localized concentration of CFA in phosphatic sediments by secondary physical processes.
Both of these objectives, but especially the second, are generally pursued for the ultimate purpose of extra~lating to a
global marine P budget which can then be used to formulate
quantitative predictive models of the marine P cycle, the related carbon cycle, and atmospheric chemistry (HOLLAND,
1978; BROECKER,1982a,b; SARMIENTOand TOGGWEILER,
1984; BOYLE, 1990).
Organic P is quantitatively one of the most important reservoirs of P in marine sediments, and organic matter is often
the principle carrier of P to sediments (FROELICH et al., 1982;
RUTTENBERG, 1990). Most of the organic matter (and associated P} which ultimately reaches the ocean bottom is
deposited in continental margin sediments ( BERNER,1982).
Whereas primary production in deep sea regions is largely
regenerated in the water column by the action of bacteria,
with only about 1% buried in the sediments (SUES& 1980;
BROECKERand PENG, 1982 ) , up to 50% reaches the bottom
in continental margin sediments ( MOLLERand SUES, 1979;
SUESS, 1980; JORGENSEN,1983). Continental margin sediments are therefore key sites for remineralization processes
and the primary sinks for P and other bioactive elements
following early diagenesis f HOLLAND, 1978; BERNER, 1982;
JORGENSEN,1983; MARTENS, 1987; BATURIN, 1988; RUTTENBERG,1990) 1
The burial flux of P to sediments is governed by the input
flux to sediments and the fraction of this input flux which is
retained by the sediments. Factors influencing the ability of
a sediment to retain P include: ( I ) the nature of the input
source material, e.g., is it labile or refractory? (2) sedimentation rate, e.g., how rapidly is the input material buried below
PHOSPHORUS(P) IS AN essential nutrient and is believed to
limit marine productivity, especially on geologic time scales
(HOLLAND, 1978; SMITH, 1984; HOWARTH, 1988). Burial
with sediments is the mechanism for ultimate removal of P
from the oceans, and as such is a key factor controlling the
inventory of oceanic P over geologically significant time periods and, hence, the amount of biological production which
can be sustained in the oceans. Attempts to quantify P burial
fluxes have generally been motivated by two objectives. The
first, to assess the fraction of P reaching the sediment-water
interface which is subsequently recycled back into the water
column for renewed participation in biological cycling, is
geared toward understanding the phosphorus cycle in the
modem ocean and often forms the subject of regional or
ecosystem studies (for example, KLUMP and MARTENS,198 1,
1987). The second, to quantify the burial flux of P to sediments in order to draw inferences about paleo-ocean carbon
cycling, has its impetus in understanding the effects on atmospheric COn and O2 of a changing oceanic P inventory,
which would #en modulate photosynthetic productivity (for
example, SANDSTROM,1982; MACH et al., 1987; MOODY et
al., 1988). This latter process has been termed the nutrientCOZ connection ( BROECKER,1982a,b) and has been considered as a possible mechanism for inducing climate change.
* Present address: Department of Marine Chemistry and Geochemistry, Woods Hole OceanographicInstitution,Woods Hole, MA
02543, USA
991
992
K. C. Ruttenherg and R. A. Bemer
the depth at which diffusional escape of reminerahzed P is
possible? (3) the occurrence and intensity of porewater irrigation and of sediment reworking, whether biogenic or abiogenie, and (4 ) formation of secondary authigenic phases due
to diagenetic immobili~tion
reactions such as precipitation
and adsorption. Because these factors vary as a function of
depositions environment, it is clear that phosphors retention by sediments is also a function of depositional environment. For example, the lability of organic matter reaching
the sediment-water interface and the presence and reactivity
of inorganic phosphatic phases, coupled with the sediment
burial rate, will directly control the rate and extent of dissolved
P build-up in porewaters. Particle transport processes and
biogenic irrigation can affect the rate and extent of porewater
phosphate build-up by altering the depth distribution of metabolizable organic matter and the concentration of oxidants
in the sediments (e.g., ALLER, 1980). The nature of the sedimentary matrix (for example, grain size and chemist ) with
which dissolved P interacts can influence the types of secondary P phases which form under elevated porewater phosphate conditions. Iron cycling, a process in which ferric iron
scavenges dissolved P from porewater within oxic regions of
the sediment column, and subsequently releases it when
transported into reducing regions, provides an example of
redox state of bottom water and sediment as controlling parameters. This process can redistribute solid-phase P, in some
cases promoting enhanced retention of P by creating an effective barrier to diffusive loss from sediments (KROM and
BERNER, 1981; FROELICHet al., 1988; HEGGIE et al., 1990;
O’BRIEN et al., 1990 ) . Surfaces of minerals other than ferric
iron phases are known to sorb dissolved P as well, such as
clays and Al oxides ( CHEN et al., 1973 ) and possibly Mn
oxides (BONATTI et al., 197 1) . All sorption reactions will be
enhanced as a function ofdecreasing grain size of the sediment
matrix due to surface area effects. In the case of formation
ofauthigenic minerals, given that saturation is exceeded, some
surfaces may provide nucleation sites which promote precipitation.
In order to construct quantitative, predictive models of
the marine P (and reIated carbon) cycle, and to interpret
paleo-ocean chemistry of phosphorus through study of sediments or the rock record, it is necessary to undemtand the
depositional environmental controls on both the recycling
efficiency of P (e.g., the fraction of initi~ly deposited P which
m-enters the water column) and the burial efficiency of P
(e.g., the amount of initially deposited P which is uhimately
buried). The work presented here focuses on two continental
margin environments, the Mississippi Delta and the FOAM
site in Long Island Sound. We have focused on continental
margin sediments to redress the bias of existing marine P
budgets which emphasize P sinks in pelagic sediments. Detailed solid-phase P speciation profiles presented in this study
indicate that diagenetic reactions within the sediment result
in burial of P in association with phases different from those
initially deposited. This diagenetic redist~bution of P results
in near perfect retention of P by the sediments studied, within
the depth resolution provided by the sampling intervals.
Our specific focus in this paper is on one of the secondarily
precipitated mineral phases, authigenic carbonate fluorapatite
(hereafter abbreviated CFA). CFA is the dominant phosphatic mineral phase of phosphorite deposits. We present
evidence that CFA is forming in sediments from the two
near-shore environments studied, and that its formation enhances retention of P liberated by organic matter degradation
and reductive dissolution of ferric iron phases buried below
the redox boundary. The observation that CFA is forming
authigeni~ally in the FOAM and Mi~issippi Delta sites is
significant, because these sites do not conform to the class&I
phosphorite depositional environment, which is characterized
by eastern boundary current upwelling regimes and low net
accumulation rates for terrigenous material relative to biogenie material (GULBRANDSEN, 1969; MANHEIMet al., 1975;
KOLODNY, 1981; SHELDON, 1981; BURNETT et al., 1983;
RFSMERSand SUESS,1983; RIGGS, I984; COOK, 1984). There
is a clear distinction between authigenic formation of CFA
and phosphorite formation. The former connotes nothing
about the concentration of authigenic CFA within the sediment, whereas phosphorite deposits are defined as sedimentary units containing greater than 5 wt% and up to 40 wt%
P20s ( G~LB~~DSEN, 1966; MCKELVEY, i 967; RIGGS, 1979;
COOK, 1984); this is in contrast to most sedimentary rocks
and sea-floor sediments, which contain less than 0.3 wt%
PZ05 (0.13 wt% P) (RIGGS, 1979). Phosphor&s are generally
believed to be the result of concentration of CFA from a
primary sediment deposit, which contained more widely disseminated CFA, by secondary physical processes such as
winnowing and reworking. CFA is not generally recognized
as an important authigenic phase in marine sediments unless
it has been secondarily concentrated, as it is in modem phosphor&es on the Peru margin, for example (BURNETT, 1977;
GLENN and ARTHUR, 1988).
The occurrence of CFA formation in depositions environments such as FOAM and the Mi~i~ippi Delta raises the
possibility that CFA is an important sink for reactive P from
the oceans not only in environments which conform to the
classical upwelling-associated phosphor&e formation environments, but in “normal”, terrigenous-dominated sediments
in non-upwelling environments as well. Authigenic precipitation of CFA is known to occur in other environments which
differ from the classical upwelling-associated phosphor&e
formation environment,
such as seamounts (BURNETT et
al., 1983, and references therein) and the non-upwelling environment of the East Australian continental margin, which
is characterized by low sediment accumulation rates and active winno~ng and erosion (O’BRIENet al., 198 1,1986,1990;
WEGGIE et al., 1990). The occurrence of CFA formation in
environments such as FOAM and the Mississippi Delta,
which are characterized by relatively rapidly accumulating
sediments dominated by terrigenous debris, is a distinct phenomenon which has not been reported previously. The concentration of CFA in sediments from the FOAM and Mississippi Delta sites is considerably diluted to well below levels
which would constitute a phosphorite deposit, due to high
net accumulation rates of mostly terrigenous debris. For this
reason CFA has not previously been recognized as a distinct
phase in these sediments. Despite low concent~tions,
the
high net sediment a~um~ation
rates of these sites result in
a significant P burial flux in association with CFA.
If the formation of CFA in environments such as FOAM
and the Mississippi Delta is a widespread phenomenon, the
implications for both the modem ocean and for ancient
oceans, as represented by the sedimentary rock record, are
Authigenic apatite from continental margins
significant. For the modem ocean, formation of authigenic
CFA in non-upwelling continental margin sediments represents an important sink, previously unrecognized, for reactive
P. This is one of two authigenic phases (the other is deltaic
ferric iron bound-P) which are potentially important sinks
for reactive P from the oceans which have not been included
in previously constructed marine P budgets. Taken together
with the larger burial flux of organic P observed for continental margin sediments relative to the deep sea (RUBENBERG, 1990), we conclude that continental
margins are
quantitatively more important sites for P burial than the deep
sea. Inclusion of the continental margin burial flux of P in a
revised budget for the marine P cycle will undoubtedly necessitate new estimates of oceanic residence time for P in the
modern ocean. These will be lower than currently accepted
estimates.
The implications for paleoceanographic interpretations of
the marine P cycle of the observation that CFA forms in
continental margin environments of the type discussed here
are of two kinds. First, although there is a definite episodicity
in the occurrence of phosphorite deposits (COOK and McELHINNY, 1979), this may not translate directly to a difference
in the quantity of reactive P being removed from the oceans
in times of phosphor&e deposition. The possibility that authigenic carbonate fluorapatite forms ubiquitously in normal,
tenigenous-dominated
sediments in non-upwelling environments, sediments which are precursors for black or gray shales
in the sedimentary record, raises the question of whether removal of reactive P from the oceans as CFA during nonphosphogenic periods may be of comparable magnitude to
that during episodes of phosphorite formation. The second
implication is not specific to authigenic CFA formation, but
applies to any secondarily precipitated phase which derives
its phosphorus from the remineralization of organic matter,
and is illustrated clearly by the data on CFA formation presented here. This is that models which extrapolate from bulk
sedimentary organic C:P ratios to fluxes of P removal to sediments will underestimate the removal term if remineralized
P is retained in sediments by secondary immobilization reactions. The extent to which the removal term may be underestimated can be quantified, as illustrated in this work,
by mapping out the diagenetic redistribution occurring in
sediments as reflected by depth changes in porewater and
solid phase chemistry.
METHODS
Core Sampling and Sediment Processing
In order to obtain deeply buried sediment samples, the sediment
core from the FOAM site in Long Island Sound was collected using
a vibrocorer. The corer, which was fitted with a core cutter to support
acetate-butyrate liners of 6.67 cm i.d., was lowered to the sediment
surface on a boom line and positioned with the aid of a diver; deep
penetration into the sediment was facilitated by vibration. Cores three
times the length of ordinary gravity cores were collected ( 3 m) . Porewater profiles measured for vibrocore samples compare well with
those from gravity cores (K. C. Ruttenberg, unpubl. data), suggesting
that porewater disturbance and compaction resulting from vibration
is not significant over the sampling intervals used ( 10 cm). The sediment surface appeared undisturbed, suggesting that using a diver to
insert the core prior to initiating the vibration/penetration stage of
coring minimized disturbance of the interface. Bottom water was
993
diver-collected in acid-washed polypropylene bottles at the time of
coring. Processing of cores was initiated within 6 h of sample collection. Sediments were processed at room temperature, which, at the
time of core collection, was within 5 “C of bottom water temperatures.
Visual descriptions of cores were recorded and then desired sampling
depth intervals below the sediment-water interface were ruled on the
outside of the core liner. Core liners were then cut into lo-20 cm
lengths, immediately capped, and placed in a N-filled, collapsible
glove bag. Inside the glove bag, sediment was collected from core
liners by extruding with a plunger fitted to the core liner, briefly
homogenized, and packed into screw-capped, sterile polypropylene
centrifuge tubes. Sediment immediately adjacent to the cut ends of
the core liner was discarded. Tubes were either frozen for future solid
phase analysis or centrifuged for 5-10 min at 3900 X g in a refrigerated
centrifuge, maintained at 15°C and returned to the glove bag. Inside
the glove bag, porewater was extracted into sterile plastic syringes
and filtered through 0.45 pm Millipore filters into acid-washed scintillation vials. Two vials were collected for each interval: one acidified
to pH I and the other left unacidified. Porewater vials were wrapped
in parafilm and stored refrigerated in Ziplock plastic bags until analysis.
The sediment core from the Mississippi Delta site was collected
in an acetate-butyrate liner (4.5 cm id.) fitted within a stainless steel
gravity corer topped with lead weights. The gravity corer was lowered
on a boom line to within a few feet of the sediment surface, and then
allowed to free fall and sink under its own weight into the sediment.
The interface appeared undisturbed. The core was taken immediately
to a refrigerated lab on board ship (the RV Gyre) and stored until
processed. Bottom water was collected from water overlying the sediment in the core tubes and filtered into acid-washed polypropylene
bottles. Core processing was initiated within 4 days after collection
and was conducted in a cold lab kept at ca. 3°C. The bottom core
cap was removed and replaced with a plunger fitted to the core liner.
The top of the core was sealed inside a N-filled, collapsible glove bag,
and sediment was extruded in 2 cm intervals and packed into polypropylene screw-cap centrifuge tubes. Tubes were either frozen for
future solid phase analysis or centrifuged for 2-3 min at 15,000 rpm
in a gimballed Sorvall SS-3 centrifuge located in the cold lab. Porewater was collected as described for the FOAM core, except that it
was performed outside the glove bag. The acidified aliquot (final pH
of 1) was rapidly dispensed directly into a vial containing 12 M HCI
to minimize loss of phosphate due to adsorption.
Porewater Analyses
Dissolved fluoride concentrations were determined with a precision
of +3S using a fluoride-specific electrode. The potentiometric and
spike methods for concentration determination were employed
(WARNER,197 1). Precision was estimated on the basis of replicate
electrode analyses, by comparing results from the previously mentioned electrode methods with the slope method used by FR&LICH
et al. ( 1983) and with results obtained usina the calorimetric techniaue
of GREENHALCHand RILEY( 196 1). Dissolved phosphate and ammonia were determined using the molybdate blue method and the
indophenol method of KOROLEFF( 1976), with precisions of &3W
and ?5%, respectively, based on replicate analyses. Chloride was
measured using the Mohr titration (STRICKLAND
and PARSONS,
1972), with a precision of +3% based on replicate analyses. Fluoride,
chloride, and ammonia analyses were performed on unacidified
samples, and phosphate on acidified samples. Sediment pH for FOAM
sediments was measured during core processing under nitrogen by
the punch-in method. In this method, the pH electrode is inserted
into a section of sediment prior to homogenization and transfer to
centrifuge tubes in order to minimize degassing. The electrode is
standardized before the first measurement and after the final measurement to monitor electrode drift; based on the observed drift, an
accuracy of kO.1 pH unit is estimated.
Solid Phase Analyses
Detailed solid-phase P speciation of the sediments was obtained
using a selective, sequential chemical extraction technique (hereafter
referred to as the SEDEX method) for P. The SEDEX method is
illustrated schematically in Fig. 1. The five sedimentary P reservoirs
994
K. C. Ruttenberg and R. A. Bernet
I-B
Exchangeable M
loosely sorbed P
Residue
CDB (pH 7.6)
Bh, 25” C
I
I
I
Hz0 wash
2h, 25’ C
\
Fe bound-P
II-B
lHz0 wash
2h,WC
Acetate Buffer (pH 4.0)
6h, 25” C
III-A
i,
G
I
III-B
I
I
Kesidue
MgCl~(pHg)
2h.
25’ C
AuthigenicApatite
plus CaCO, - bound P
plus Biogenic Apatite
i
III-C
III-D
Residue
I
MgC$(pH8)
2h. 2P C
Hz0 wash
2h. 25” C
t
IV
Residue
tMHC1
16h,
Detrital Apatite
plus other inorganic
-P
V
V
Residue
Ash5WC
+
1 M HCI, 16h, 25’ C
Organic - P
I
RG. f . The wet chemical sequential extraction scheme (SEDEX) for quantification of five sedimentary P reservoirs: ( I ) exchangeable or loosely sorbed P; (2) ferric iron bound P, (3) authigenic apatite + biogenic apatite + CaCO3
bound P; (4) detrital apatite; and (5) organic P. Step IIA extractant CDB is citrate-dithionite-bi~rbonate.
Details of
the development and standardization of this method can be found in RUTTENBERG ( 1990, 1992 ) .
separated by this technique are: ( I) authigenic carbonate fluorapatite
f biogenic apatite + CaCO,-bound P; (2) detrital apatite of igneous
or metamorphic origin; (3) P bound to ferric oxides and oxyhydroxides; (4) organic P; and (5) exchangeable or loosely sorbed P.
The SEDEX method is capable of measuring concentrations as low
as 0.005 wt% P with a reproducibility, based on analysis of replicates,
of rt4%. A detailed description of the SEDEX method is given elsewhere ( RU~ENB~RG,
1990, 1992 ) . Three salient features which distinguish the SEDEX method from other selective extraction methods
are briefly summarized here, First, a means for the chemical separation
of authigenic apatite from detrital apatite of igneous or metamorphic
origin has been devised. This is an important distinction since authigenic apatite represents an oceanic sink for reactive P, whereas
detrital apatite does not. Second, it has been demonstmt~ that the
matrix effect, or the redistribution of P onto residual solid surfaces
during a given extraction step, can be extreme and result in large
errors in the identification and quantification of sedimentary P reservoirs. The use ofa MgClr wash following those extraction steps for
which secondary re-adsorption onto the residual sedimentary matrix
is a problem reverses most of it. Finally, each extraction step has
been standardized indi~du~iy to ensure the selectivity and specificity
of the extractants, using phases and materials which are anaiogues
for naturally occurring sedimentary P phases. The SEDEX method
was applied to freeze-dried sediment which had been ground to < 125
pm. It is also important to note that, due to the very low phosphorus
content of most marine sediments, indirect operationally defined
methods such as the SEDEX method are the only available means
for determining detailed sohd-phase speciation information on sedimentary phosphorus.
SITE DE!3CRIPTiONS
The FOAM site in Long Island Sound is located in 9 m water
depth in the vicinity of the Thimble Istands off the coast of Connecticut, 41”14.5’ N lat., 72”44.8’W long. (Fig. 2a). The sediment
at this site is intermittently oxygenated via irrigation and bioturbation
to depths of lo- 15cm, an oxic cap of sediment is observed ( 1-3 cm)
at most times of the year, and shallow porewater and solid phase
Authigenic apatite from ~ntinental
995
margins
Mobile
FIG. 2. Location maps showina core locations at (a) the FOAM site in Long Island Sound, and (b) site CT1 5 on
the Mississippi Delta in the Gulf of Mexico.
.
chemistry are perturbed as a result of infaunal activity f ALLER,1977;
WESTRICH,1983; CANFIELD,1988; RUTTENBERG,1990). As uoted
by CANFIELD( 1988)) belowthe zone of bioturbation porewater and
solid phase profiles approximate a system at steady-state. Porewater
~dients appear to be controlled by molecular diffusion in this depth
region, despite evidence from x-radiographs that sedimentation at
FOAM may be interrupted by periodic storm events which resuspend
sediments and leave shell lag deposits and imperfect sediment laminations. These laminations and lag deposits are preserved because
of the small size and intermittent activity of the benthic fauna at
FOAM (ALLER,
1977). Porewater sulfate concentrations drop to
zero at depths of 100-l 30 cm, and sulfate reduction rates reach 80100 mM/yr in the top 5 cm during late summer and early fall
( WESTRICH,1983). Porewater nutrient concentrations are high, reflecting remineralization of relatively labile organic matter, reaching
levels of 2-3.5 mM ammonium and 200-300 pM phosphate at depths
of 1-3 m (GOLDHABERet al., 1971; ALLER, 1977; RU’MXNBERG,
1990). Sedimentation rates are estimated at between 0. I and 0.3 cm/
y on the basis of radiometric and geochemical analyses ( BERNER,
1978; GOLDHABERet al., 1977: KRISHNASWAMI
et al., 1982).
Mississippi Delta core CT15-GC4 was collected in i 10 m water
depth on the lower Mississippi Delta-upper Mississippi Cone east of
the Mississippi Trough and the extension ofSouthwest Pass, 28*47.X’
N lat., 89” 19.11’ W long. (Fig. 2b). Burrowswere observed and core
color was a mottled brown and gray to a depth of 10 cm. Below 10
cm, the uniform gray color of the sediment indicates little or no
burrowing. Integrated sulfate reduction rates of 7.7 mmol/m’/day
were measured for the top 26 cm, and porewater sulfate con~ntration
is constant at between 3 i mM and 32 mM to a depth of 19 cm ( LIN,
1990). Porewater ammonium and phosphate reach values of 700
pM and 86 PM, respectively, at a depth of 1 m. The shape of the
nutrient profiles and the vertical porewater sulfate profile suggest that
the sediment has been disturbed, either by bioturbation or slumping,
to a depth of 45 cm. Below this depth,
nutrient profiles
diffusionai. Sedimentation
is estimated
be 0.8
from the
depth profile
1990).
RESIJLTS
profiles
ammonia are
sites. The
of
chloride,
in Fig.
for both
denote bottom
The fluoride
concentration maximum
the sediment-water
from the
that of
and a
phosphate,
FOAM and
fluoride consite displays
bottom water,
drop to
bottom water
in the
region below
cm. The
in dissolved
near the
water interface
the presence
a source
fluoride
to
within this
range, and
sub~quent
drop
removal of
from porewater
the solid
deeper in
core. The
responsible for
features are
in the
text. The
profile
from
Mississippi Delta
is negatively
from botwater fluoride
by roughly
FM. Fluoride
is relatively
at a
of 45-46
in the
65 cm
the first
depth intervals).
70 cm,
concentrations drop
this level
a minimum
37 FM
89 cm.
salinity profiles
rived from
concentration measurements)
both
the
and Delta
are invariant
depth, suggesting
the variations
fluoride concentration
at these
result from
other than
which would
perturb chloride,
as freshwater
Several other
taken at
FOAM site
the Mississippi
display
features
to those
here (
1990).
The
nutrient profiles
both sites
steep positive
above bottom
concentrations
near
sediment-water interface.
this, the
profile at
FOAM site
fairly monotonically
the depth
the core.
phosphate also
creases, althou~
is a
~oncentmtion gradient
the lo-40
depth interval.
the ammonia
phosphate profiles the Mississippi
core can
divided
into
distinct depth
which result
different
physico-chemical
The region
50 cm
undisturbed build-up
porewater metabolites,
in
the
cm region,
absence of monotonic~ly increasing
profile, normally
by diffusional
of the
gradient, indicates
physical disturbance
the sediment due either to sediment
slumping, deep sediment mixing, or irrigation by benthos.
The solid-phase phosphorus data obtained using the
SEDEX method are given in Table 1 for the FOAM site and
996
K. C. Ruttenberg and R. A. Renter
IF- ] uM & Salinity (s)
INHfl PM
1HPOTJ uM
100
200
I .
0
1000 2000 3NtO 4
2
5
ioa
eb
Depth
(cm)
i
200
%
300
0
,
200
400
600
I
0
IYJ
200
40Depth
(cm)
60
0”
0
0
n
0
60-
0
n
FIG. 3. Porewater profiles for salinity (derived from chloride measurements), guoride, phosphate, and ammonia at
the FOAM and Mississippi Delta sites. Arrows indicate bottom water fluoride concentration. Bottom water phosphate
and ammonia at both sites are below 2 and 5 PM, respectively. The sediment-water interface of both cores FOAMVC-D and CT1 5-CC4 appeared undisturbed. Similar box core and gravity core porewater profiles for the Delta site
confirm an intact gravity core sediment-water interface (RUTTENBERG,
1990).
the Mississippi Delta site, respectively. The quantity of P in
each of the five reservoirs separated by the SEDEX scheme
is given as the fraction of the total sedimentary P in percent.
The total P concentration in gg P/g dry wt is also given. The
solid-phase profiles of organic P and of authigenic P are plotted together for both sites in Fig. 4 to illustrate the near mirrorimage relationship between these two P reservoirs. The significance of this relationship is discussed next.
DISCUSSION
The Case for Authi~enjc Carbonate ~luo~patite
at the FOAM and Mississippi Delta Sites
Formation
Diagnostic indicators of authigenic carbonatejluorapatite
formation
It is necessary to use indirect diagnostic indicators of CFA
when the quantity of solid-phase P present as CFA in sediments is below the detection limit of direct identification
methods, such as X-ray diffraction or petrographic analysis.
The majority of marine sediments fall in this category, as
they contain ~0.1 wt% total P, already an order of magnitude
lower than the detection limit of X-ray diffmction. Three of
these diagnostic indicators use porewater chemistry to infer
the presence of CFA in sediments. These are porewater fluoride profiles, porewater phosphate profiles (which are decoupled from porewater ammonia profiles), and saturation
state of the porewater with respect to the authigenic mineral
phase, CFA. These indicators are described in the following
text. The fourth diagnostic indicator is indirect identification
of CFA in the solid phase using the SEDEX method.
Downward decreasing porewater fluoride concentration
indicates removal of fluoride to the solid phase. The most
common explanation for removal of fluoride to the solid
phase is incorporation into CFA (e.g., FROELICHet al., 1983;
JAHNKEet al., 1983; BURNETTet al., 1987; FROELICH et al.,
1988; SCHUFFERT, 1988). Results of recent work on early
diagenetic fluoride chemistry suggest that phases other than
CFA, such as nonphosphatic carbonate minerals (RUDE and
ALLER, 1991) and ferric iron phases ( RU~ENB~RG and
997
Authigenic apatite from continental margins
Table la. Solid-Phase PhosphorusSpeciationfor FOAM-W-D.*
&Ph
(cm)
O-10
50-60
100-110
150-160
200-210
250-260
290-300
STEP I
STEP II
Exch. & Loosely
SorbedP
(% of Total-P)
Iron ox./oxyhydrox.
bound P
(46of Total-P)
8
:
:
z
< d.1.)
< d.l.
< d.1.
< d.1.
< d.1.
< d.1.
< d.1.
STEPN
STEP III
Authigettic-P
(S of Total-P)
DettitaI
Inorganic-P
(% of Total-P)
29
;:
z;
40
39
41
45
44
47
41
41
36
STEPV
STEPS I-N
STEPS I-V
Organic-P
(W of Total-P)
Inorganic-Pt
(% of Total-P)
Total-P*
WgP/g)
22
12
13
78
:;
89
86
::
;:
11
19
436
450
448
456
449
429
417
* All values are the average over duplicate analyses; reproducibility errors are discussed in the text.
t Inorganic-P is the sum of Steps I-IV.
* Total-P is the sum of Steps I-V, repotted in units of concenmation.
5 < d.1indicates concenuation is below the detection limit.
1988; RUTTENBERG, 1990), may play an important role in controlling interstitial water fluoride profiles, as
well. This additional
aspect of early diagenetic
fluoride
chemistry and its bearing on interpreting fluoride profiles as
indicators of CFA precipitation are discussed in the following
text.
Porewater phosphate concentrations
are the result of the
balance between input from microbially mediated organic
matter decomposition,
anoxic reduction of ferric iron phases
and liberation of associated phosphate, and fish-bone dissolution; and removal by diffusional loss or by secondary immobilization
reactions, of which formation
of authigenic
carbonate fluorapatite is one. When the rate of removal to
an authigenic phase exceeds the input rate, the dissolved
phosphate profile will show a reversal in gradient from positive
to negative. This has been observed in classical upwelling
sites where CFA has been directly identified in sediments,
such as the Peru and Baja California shelf regions ( FROELICH
et al., 1983, 1988; JAHNKE et al., 1983; SCHUFFERT, 1988).
When a gradient reversal is not apparent, it is possible to
detect removal of phosphate from porewater using ammonia
as a conservative
tracer of organic matter decomposition.
This is a variant of the stoichiometric
nutrient regeneration
CANFIELD,
model for anoxic sediments described by BERNER ( 1977).
These models have been employed to ascertain the C:N:P
ratio of organic matter undergoing decomposition
from the
ratio of metabolites in the interstitial water, or to predict the
ratio of porewater metabolites
from measured solid-phase
C:N:P ( HARTMANN et al., 1973, 1976; BERNER, 1974, 1977;
MARTENS et al., 1978; KROM and BERNER, 198 1; ROSENFELD, 198 1 ), and also to estimate fluxes of nutrients out of
the sediment ( KLUMP and MARTENS, 1987 ) . Deviations from
stoichiometric
release of phosphate are explained as being
due to removal via adsorption or precipitation of authigenic
minerals (MARTENS et al., 1978; KLUMP and MARTENS,
198 1, 1987). In the current treatment, we assume stoichiometric release of phosphate and ammonia from organic matter decomposition,
and steady state deposition and burial.
We also take into account differential diffusion and adsorption, and use an empirical fit of the ammonia profile to predict
the shape of the phosphate profile. If porewater phosphate
and ammonia profiles are decoupled such that the prediction
based on the fit to the ammonia profile overestimates
the
measured phosphate profile, the deficit is interpreted to be
the result of removal to an authigenic solid phase.
The third porewater indicator is the calculated saturation
Table lb. Solid-Phase Phosphorus Speciation for Mississippi Delta core CTl5-GC4.
Depth
(cm)
o-2
4-6
8-10
12-14
16-18
20-22
24-26
28-30
32-34
36-38
40-42
48-50
52-54
56-58
60-62
6466
68-70
72-74
STEP I
Exch. &Loosely
sorbed P
(W of Total-P)
STEP 11
Iron ox./oxyhydrox.
bound P
(% of Total-P)
Authigenic-P
(% of Total-P)
:
z
::
8
27
26
27
28
24
23
26
24
;z
25
26
27
z
;
4
4
5
2
:
2
25
:
2’:
z?l
2
STEP III
::
::
::
;:
34
::,
31
* Inorganic-P is the sum of Steps I-N, reported in units of concentration.
t Total-P is the sum of Steps I-V.
STEP IV
Denital
Inorganic-P
(% of Total-P)
11
14
14
16
16
16
17
16
14
14
15
14
15
::
11
13
13
STEP v
STEPS I-IV
STEPS I-V
Organic-P
(% of Total-P)
Inorganic-P’
(% of Total-p)
Total-Pt
(pgP/g)
::
72
74
74
605
546
587
549
561
578
586
553
555
546
548
553
565
587
548
547
622
598
26
27
24
:;
23
23
24
22
:::
22
;;
20
20
:z
78
Yl
77
::
76
77
78
76
::,
80
998
K. C. Ruttenberg and R. A. Berner
Long Island Sound:
FOAM-W-D
Mississippi Delta:
wt% P
O.CNl
0.01
CT15GC4
wt% P
i
0
0.010
0.015
0.020
0
ia
FIG. 4. Solid-phase profiles of authigenic P and organic P vs. depth at the FOAM and Mississippi Delta sites. Data
are from STEP III and STEP V, respectively, of the SEDEX method (Table 1). The curves drawn through the data
points are empirical best-fit curves which are given explicitly in the notes accompanying Table 2 and are discussed in
the text.
state with respect to the authigenic mineral. Where the ion
activity product (IAP) is equal to or exceeds the solubility
product (Ksp) for CFA, the porewaters are saturated or supersaturated with respect to CFA, respectively, and precipitation is thermodynamically
favored. This indirect approach
to inferring the presence of authigenic minerals in sediments
has been widely used ( BERNER, 1974; EMERSON, 1976;
EMERSON and WIDMER, 1978; MARTENS et al., 1978; MURRAY et al., 1978; MULLER and SUESS, 1979; ALLER, 1980;
JAHNKE et al., 1983; DE LANGE, 1986; CANFIELD, 1989).
Because the solubility of CFA is a sensitive function of the
activity of dissolved carbonate ( JAHNKE, 1984; GLENN and
ARTHUR, 1988 ), a single KS,,does not adequately reflect the
saturation conditions at all depths. To account for changes
in porewater chemistry with depth, separate values for KsP
must be calculated at each depth.
These indirect diagnostic indicators provide the most convincing argument for the presence of CFA in sediments when
they are used in concert. For example, we have quantified
the removal flux of fluoride over a given depth interval and
compared it to the quantity of CFA in that depth interval as
measured by the SEDEX method. A deficit in porewater
phosphate has been identified using the measured ammonia
profile and a stoichiometric
nutrient regeneration model, indicating removal of dissolved phosphate to the solid phase.
The increase in CFA has been compared to the decrease in
organic P measured by the SEDEX method. When these various indicators of CFA formation yield consistent results, a
convincing
sediment.
argument
is made for formation
of CFA in a
Formation of authigenic CFA at the expense of organic
phosphorus: Solid-phase diagnostic indicators
A large fraction of the total P at both the FOAM and the
Mississippi Delta sites is in the authigenic CFA + biogenic
apatite + CaC03 reservoir (Table 1) . All three of these phases,
extractable in STEP III of the SEDEX scheme, are authigenic
because they represent removal of reactive P from either the
water column or from porewater. For the purposes of this
discussion, the term authigenic P will be used to denote all
three. Authigenic P, present in appreciable quantities at all
depths at both the FOAM and the Mississippi Delta sites,
displays a depth trend that indicates an increase in the size
of this reservoir with depth (Fig. 4). Evidence suggests that
the increase with depth in the authigenic P reservoir is the
result of formation of authigenic CFA, and that the authigenic
phosphate present in the top sediment interval (0- 10 cm at
the FOAM site and O-2 cm at the Delta) is actually a nonCFA background
phase. Candidates
for this background
phase are suggested on the basis of results obtained during
standardization
of the SEDEX method. Sedimentary P phases
identified as being soluble in pH 4 acetate buffer (STEP III)
during specificity standardization
experiments
are CaC03
bound P, biogenic apatite (fish bones, teeth, and scales), and
P associated with smectite ( RUTTENBERG, 1990, 1992). Of
999
Authigenic apatite from continental margins
these possible non-CFA background phases, finely divided
biogenic apatite and/or P associated with smectite are considered the most likely. Phosphorus associated with CaC03
is ruled out at the FOAM site because, although some sediment layers contain significant shell material, this material
was found to contain only minor P. This is consistent with
the findings of SHERWOODet al. ( 1987) who also found that
shell material contains very little P. CaC03 is ruled out as a
non-CFA background phase at the Delta site because these
sediments contain negligible amounts of CaC03 (<OS wtW).
Biogenic apatite debris, which was not readily visible upon
microscopic analysis, may be present in very finely divided
form. The presence of P associated with a smectite-like clay
in these sediments is possible. If the background phase is
biogenic apatite, the entire authigenic P reservoir is truly authigenic, in the sense that it represents removal of reactive P
from seawater. If smectite is the background phase, whether
it is truly authigenic or not depends on whether the P became
associated with the clay before or after the clay entered the
marine system.
An alternative explanation for the presence of appreciable
authigenic P at the sediment-water interface is that it reflects
virtually instantaneous formation of authigenic CFA within
the first 2-10 cm of sediment below the sediment-water interface. This is believed to be the case for CFA forming in
sediments on the Mexican ( JAHNKE et al., 1983) and Peru
margins ( BURNETTet al., 1988; FROELICHet al., 1988; GLENN
and ARTHUR, 1988 ). Although rapid, near-interface precipitation of CFA at FOAM and the Mississippi Delta is not
supported by the quantitative arguments based on porewater
and other solid-phase data presented in the following text, it
cannot be ruled out, since near-interface processes cannot be
resolved by the current data sets, which provide only cmscale depth resolution. Collection of porewater and detailed
solid-phase phosphorus data over finer depth scales (e.g., mm)
may reveal that CFA formation is occurring at accelerated
rates very near the sediment-water interface.
Assuming that the background phase at FOAM and the
Delta is not CFA, the increase above background levels reflects authigenic formation of CFA within these sediments.
The depth trends of authigenic P and organic P clearly mirror
each other (Fig. 4), and we suggest that authigenic carbonate
fluorapatite is forming from the organic P which has been
liberated to porewater as a result of microbial remineralization
of organic matter. Comparison of the integrated amount of
authigenic P represented by the increase above background
to the integrated amount of organic P decrease gives a measure
of the degree to which the transfer of phosphorus from one
reservoir to another (sink-switching) can explain the mirrorimage depth profiles of these two reservoirs. Best-fit curves
were drawn through the data for authigenic and organic P at
both sites, and the area under these curves was obtained by
integration. Logarithmic functions yielded the best fit for both
authigenic and organic P at both sites. At the FOAM site,
these curves were integrated from the interface to 110 cm
depth (the region of the largest gradient). At the Delta these
curves were integrated over the full data set as the gradients
appear to persist throughout the depth of the core. The integrated values for both sites are listed in Table 2. The organic
P deficit matches the authigenic P increase at the FOAM site
Table 2. Depth Integrated CFAP Increase versus Depth integrated Organic-P Decrease.
StZItiOtl
Depth Interval of
Integration (cm)
lntegnted Organic-P
Deficit
Standing Crop
(pg P/cm2)
Integrated CFA
Increase
Standing Crop
(Kg P/cm?
Long Island Sound
FOAM-WI-D
O-110
28W+ IW
2900 f 200
0-103
3200 f 200
3700 f 200
Mississippi Delta
CT15.cc4
Notes:
a) ?ix equations for the curves integrated for the FOAM site are:
Organic-P:
y = 0.011516
[0.003082. LOG(x
R* = 0.929.
Authigenic-P:
y = 0.010338 + IO.003450 * LoG(x
Rz = 0.992.
b) The equations for the curves tntegnted for the Delta site xe’
Organic-P:
y = 0.016187
[O.oOlY92 . LOG(x)];
Authigenic-P:
y =0.0121511 + [O.W3849 * LOG(x
R2 = 0.674.
R2 = 0.668.
c) The value for the organic-P decreasewas obtained by subtracting the area under the
curve from the area of the rectangle which has dimensions described by the origin, the
wt% organic-P value of the first depth interval, and the lower depth limit of integration.
d) The value for the CFA increase was obtained by subtractingthe area of the rectangle
which has dimensions described by the origin, the wt% authigenic-P value of the fist
depth interval, and the lower depth limit of integration. from the area under the curve.
e) llx integrated areas under the curves have teen converted from units of (pg P/g x cm)
to units of standing crop (pg Plcmz) using the following equation:
kgP/cm* = (I-$)[p (gkm’)llArea (g P/g x cm)l(lpg P/l~lO-~g P)l,
where o = 0.65 and p = 2.5 g/cm’.
f) Tbe etmr reported for the standing crop was obtamed by calculating the area under
cwcs described by maximum and minimum values for organic-P and authigenic-P,
nspcctively. The maximum and minimum values for each organic-P and authigenic-P
data point were assumedto be +4’% of the measured value, where +_4% is the average
relative cmx in reproducibility determined for the SEDEX pmcedurr (Ruttenkrg,
IWO). The reported deviation from the best-tit value is the maxmxmt difference
between the standingcrop derived from the area under the maxmwn or minimum
curves and the best fit curve.
to within the calculated error, and the match for the Mississippi Delta core falls just outside the calculated error. The
fact that the profiles do not perfectly mirror each other indicates that the transfer of phosphorus from the organic to
the authigenic P reservoir is not the only process operating,
or that the two processes operate at different rates. In addition,
sources of P other than organic matter, such as ferric iron
associated P or biogenic apatite (fish debris), could be important, especially near the interface.
Near-quantitative transferral of dissolved P from remineralized organic matter to CFA implies that loss of dissolved
phosphate by diffusion out of the sediment prior to precipitation as CFA is minimal. Although there is significant diffusive flux out of the upper 10-l 5 cm of sediment at the
FOAM site during most of the year, as documented by previous workers ( ALLER, 1980; KROM and BERNER,198 1) and
indicated by the upper 20 cm ofthe FOAM phosphate profile
discussed here (Fig. 3), the essentially quantitative transfer
of solid-phase P from the organic to the authigenic reservoir
suggests that beneath this horizon diffusive loss is insignificant.
Near-coinciding minima in the dissolved phosphate and fluoride profiles at 40-60 cm depth provide supporting evidence
for this interpretation, implying diffusion of phosphate and
fluoride into the sediments and subsequent removal to the
solid phase in this depth interval. We note also that it is not
uncommon for minima in phosphate profiles to be observed
in sediment cores from the FOAM site (see, for example,
GOLDHABER
et al., 1977; ALLER, 1980). The largest increase
in the authigenic P reservoir, and therefore the most CFA
precipitation, occurs in the top 60 cm. The coarsely spaced
1000
K. C. Ruttenberg and R. A. Bemer
depth intervals of the solid-phase data prevent us from ascertaining whether precipitation occurs throughout this interval or is localized in a discrete depth horizon; however,
the minima in the porewater phosphate and fluoride profiles
suggest that formation of CFA occurs predominantly within
the 40-60 cm depth interval. (We restrict our discussion of
porewater to the FOAM site, since the Mississippi Delta
porewater profiles show evidence of recent disturbance.)
It is the observed mirror image of the organic P and authigenic P profiles, and the match between integrated loss of
organic P and gain in authigenic P, which compels us to
argue that authigenic P is forming at the expense of organic
P. Diffusive loss of phosphate from surficial sediments does
not invalidate this conclusion. Calculated diffusive loss of
phosphate based on near-interface porewater data will have
limited bearing on diagenetic transfer of phosphorus occurring
at depth. In addition, surficial porewater profiles at the FOAM
site are known to display seasonal variability. A diffusive flux
calculated from a single porewater profile represents an instantaneous flux and will not necessarily reflect past input
and perturbation of porewater which may have been recorded
in the solid phase.
In general, the fluoride profile at the FOAM site displays
a negative concentration gradient (Fig. 3)) consistent with
removal to the solid phase. In the region of steepest negative
slope, between 0 and 60 cm depth, mass balance calculations
suggest that this removal is the result of authigenic CFA formation.
The amount of fluoride removed from porewater, approximated by the negative dissolved fluoride concentration gradient over the O-60 cm depth interval and its change to a
roughly constant fluoride concentration below 60 cm, is calculated using Fick’s first law of diffusion. The removal flux
of fluoride to the solid phase within a box defined by this
depth interval is assumed to be equal to the amount of CFA
buried at 60 cm depth, and the system is assumed to be at
steady state. In other words, essentially all dissolved fluoride
entering the top of the box by diffusion is removed at the
bottom of the box as CFA buried with sediment. This calculated removal flux of fluoride is then compared to the increase in the size of this reservoir, as determined by the
SEDEX method. The amount of fluoride removed from
porewater to the solid phase is calculated using Eqn. 1:
Formation of authigenic CFA: Porewater diagnostic
indicators
where
The use of porewater chemical profiles to infer processes
occurring in the solid phase is widespread (GOLDHABERet
al., 1977; MARTENSet al., 1978; MURRAYet al., 1978; FROELICH et al., 1982, 1983; JAHNKE et al., 1989). The main
advantage offered is that changes in porewater chemistry are
sensitive indicators of changes in the solid phase, due to much
lower concentrations in porewaters. Most models which use
porewater profiles to infer solid-phase processes require an
assumption of steady state. This requirement presents the
most serious limitation to solid phase arguments based on
porewater chemistry. Especially in regions of dynamic physical processes, which characterize most continental margin
areas, the steady state assumption often cannot be made.
This is so in the Mississippi Delta site discussed in this study.
The FOAM site appears to be in steady state, as evidenced
by the shape of the porewater nutrient profiles presented here
(Fig. 3 ) , as well as other porewater chemical profiles for this
same core (SO,; H2S; CANFIELD, 1988), and by the success
of steady state models presented in this and in previous studies
at describing both solid-phase and porewater data ( BERNER,
1977; ALLER, 1977; GOLDHABERet al., 1977; MARTENSet
al., 1978; KROM and BERNER, 1981; BOUDREAUand CANFIELD, 1988). In contrast, porewater profiles from the Mississippi Delta site show clear evidence of disturbance, as discussed in the Results section. The following discussion of
CFA formation based on porewater data will be made, therefore, exclusively for the FOAM site. We note, however, that
porewater profiles from the Delta are qualitatively consistent
with the solid-phase evidence for CFA formation at this site:
fluoride profiles indicate removal to the solid phase, and
phosphate profiles increase with depth up to nearly 100 FM.
The similar porefluid chemistries of the two sites suggest that
interstitial water should be at saturation or greater with respect
to CFA in the Delta core.
-DFe =JF,
4 = 0.7 ( KROM and BERNER, 1980),
DF = 200 cm’/yr
ACF
ax=
(VAN CAPPELLENand BERNER, 1988),
and
-14 pmol/L
60cm
.
The mass of CFA-P per unit area leaving the bottom of the
box is obtained by multiplying the observed flux of fluoride,
calculated using Eqn. 1, by a P/F ratio of 2.5, which is typical
for authigenic marine CFA ( FROELICHet al., 1988). Division
by a sediment accumulation rate appropriate for the FOAM
site (0.075 g/cm’ yr) yields a value of 0.0034 wt% authigenic
P. The measured increase of authigenic P above background
in the O-60 cm depth interval is 0.0032 wt%.
There is additional structure in the porewater fluoride profile, however, which merits discussion both to ensure that
the processes responsible for these other features are not at
odds with the interpretation rendered thus far (e.g., that fluoride removal from porewater in the upper 60 cm of the core
is due to incorporation into CFA), and to address the question
of general fluoride reactivity during early diagenesis. The subsurface maximum observed in the FOAM core has been observed in other cores taken at this and other sites (Sachem,
NWC) in Long Island Sound ( RUTTENBERG,1990). We have
observed that this concentration maximum is coincident with
the redox boundary, and that it coincides with a maximum
in porewater [Fe’+] and with the most rapid decrease in solidphase reactive ferric iron ( RUTTENBERGand CANFIELD,1988;
RUTTENBERG, 1990). These observations, coupled with results from laboratory adsorption experiments ( RUTTENBERG,
1990) which indicate that fluoride adsorbs readily onto ferric
oxyhydroxides at porewater pH (7.4 to 7.6), suggest that the
observed sub-surface maximum in porewater fluoride is related to redox chemistry in the sediment. Fluoride is adsorbed
onto reactive ferric iron phases in the surficial oxidized layer
Authigenic apatite from continental margins
of the sediments, and when these phases are buried below
the redox boundary, the fluoride is released to porewater upon
reduction of the ferric iron. Significant levels of fluoride adsorption onto other phases was also observed in laboratory
adsorption experiments, especially clays (smectite and kaolinite, but not illite) and humic acid ( RUTTENBERG,1990).
Because the sub-surface maximum in porewater fluoride
concentration implies a source of fluoride to porewater, however, it is most readily explained by association with redoxsensitive reactive iron phases. The structure of the fluoride
profile below the top 60 cm may, however, be the result of
adsorption or desorption reactions involving non-iron phases,
or indeed may be due to continued CFA formation. A gradual
increase in fluoride concentration occurs immediately below
the O-60 cm zone of CFA formation, and below 100 cm
concentrations drop once again, with a negative slope much
shallower than that observed in the upper 60 cm of the core.
The increase observed between 50 and 100 cm implies input
of fluoride from the solid phase. The fact that the size of the
CFA reservoir continues to increase within this depth interval
(Table la) argues against dissolution of CFA as the process
responsible for the increase in fluoride concentration. Other
phases which are possible candidates as the source of this
fluoride input, based on the adsorption experiment results
previously discussed, are clays or organic matter, although
more work on the early diagenetic chemistry of fluoride is
required before a more definite interpretation can be made.
The return to a negative concentration gradient below 100
cm indicates removal from porewater to the solid phase is
occurring below 100 cm depth. As there is no systematic
change in the CFA reservoir in this interval, the solid-phase
chemical data do not provide an unequivocal indication that
this is the result of continued CFA formation. The possibility
exists that another phase is responsible for removing fluoride
in this region of the core, or that CFA formed in the upper
regions of the sediment column is recrystallizing with a more
fluorine-rich stoichiometry.
The negative offset of porewater from bottom water fluoride concentration observed in the Mississippi Delta core
(Fig. 3) is believed to be the result of fluoride adsorption
onto reactive ferric oxyhydroxides which are present in sediments throughout this 100 cm core ( RUTTENBERG, 1990).
There is a trend of more gently decreasing porewater fluoride
concentrations superimposed on the more dramatic negative
offset of porewater from bottom water, indicative of removal
of dissolved fluoride from porewater by a separate mechanism. Solid-phase sedimentary phosphorus data from this
site support the argument that this removal is the result of
incorporation into carbonate fluorapatite.
There is an interesting link between the processes of sorption of fluoride onto ferric-oxyhydroxides and of authigenic
carbonate fluorapatite formation. The results of the FOAM
site will be used to illustrate this link, but the argument can
be generalized. Ferric-oxyhydroxides are capable of adsorbing
both fluoride and phosphate from seawater. There is a subsurface maximum in porewater fluoride as a result of reduction of these ferric phases, as previously discussed, and there
is a similar spike in porewater phosphate near the sedimentwater interface as a result of reduction of ferric-oxyhydroxides
and release of associated phosphate to porewater (KROM and
BERNER, 198 1). The spike in concentration of fluoride and
1001
phosphate which results from release of these two solutes at
or below the redox boundary in sediments may provide the
necessary conditions of supersaturation required for the precipitation of authigenic carbonate fluorapatite ( RU~ENBERG,
1990) _This has been proposed for the phosphorites forming
on the East Australian continental margin (HEGGIE et al.,
1990) and may be important at FOAM and the Mississippi
Delta, as well. A more detailed treatment of the early diagenetic chemistry of fluoride in terrigenous sediments will be
presented elsewhere (K. C. Ruttenberg, unpubl. data).
To assess whether the porewater phosphate profile is decoupled from the ammonia profile at the FOAM site, an
empirically derived best-fit curve was drawn through the ammonia data. The equation for this curve was used to predict
the porewater phosphate profile using a stoichiometric nutrient regeneration model (Fig. 5 ). The details of the model
are given in the Appendix. There is clearly a deficit of porewater phosphate relative to the model-predicted concentrations. This result illustrates that the process of P removal
from porewaters to form secondary phases can be identified
in a depth zone where porewater phosphate concentrations
are increasing as a result of release of remineralized organic
P. A quantitative comparison of the porewater phosphate
deficit to the solid-phase CFA increase is complicated by the
fact that the upper limit of integration for the porewater curves
was taken at 25 cm to avoid bioturbation effects, and that
due to the coarse sampling intervals used in this core, diffusional loss of phosphate cannot be adequately estimated
from the profile. The much coarser sample intervals in the
solid-phase profile also make a detailed comparison difficult.
Given the limitations of the data set and of the model, this
should be viewed as a qualitative comparison, the results of
which are consistent with removal of porewater phosphate
to form authigenic CFA.
The saturation state of porewaters at the FOAM site with
respect to CFA was calculated using the ion activity product
(IAP) derived from porewater data for apatite of stoichiometry Ca&%.X,M&.l~ (~0~)~,8(~~3)~.2(~)2,~8,
mmthd
to the solubility product ( Ksp) for the authigenic phase (IAP/
K,,). The solubility product was calculated in the same manner described by JAHNKEet al. ( 1983), where Ksp is evaluated
for each depth interval using the experimental relationship
between carbonate activity and Ksp shown in JAHNKE et al.
( 1983, Fig. 5 ) and the corresponding carbonate activity obtained from the measured alkalinity and pH at each depth.
Porewater data are tabulated in RUTTENBERG( 1990). We
chose the stoichiometry for carbonate fluorapatite used by
JAHNKE in his experimental and field work characterizing
the solubility of carbonate-rich apatites in marine systems
( JAHNKEet al., 1983; JAHNKE, 1984). The interstitial water
is supersaturated by a factor of lo4 to lo6 throughout the
length of the core, indicating that precipitation of the authigenie phase is thermodynamically
favored. The extensive
and complex substitution chemistry displayed by apatite (MCCLELLAN, 1980; JAHNKE, 1984; LEGEROS, 1965;
LEGEROS et al., 1967, 1980; MCARTHUR, 1985), however,
results in a large degree of uncertainty in the parameter Ksp.
Also, any effects of recrystallization of authigenic CFA with
depth, or any continued uptake of fluoride through exchange
processes (e.g., without additional growth or formation of
new material), are excluded from this simple interpretation.
K. C. Ruttenberg and R. A. Bemer
1002
[HPOi]
[NH,+1 PM
100
0
0
200
pM
300
!
400
-0
0
100
0
100
0
0
0
0
Depth
(cm)
g
0
0
2ot
00
2ot
0
0
0
0
0
B
30
<0
A
\
30
FK. 5. The empirical exponential fit to the porewater ammonia data is shown in (A); the equation is given in the
Appendix. The porewater phosphate profile predicted from the ammonia curve using the stoichiometric nutrient
regeneration model described in the Appendix is compared to the porewater phosphate data in (B). The actual data
are represented by open circles; the model fit and prediction by solid lines. The discrepancy between the predicted and
the actual phosphate profiles indicates that some portion of the phosphate regenerated to porewater via microbial
deeomnosition of organic matter has been removed. We interpret this removal as being due to formation of an authigenic
phosphate phase. Such processes could result in a situation where the porewater
derived (IAP / Ksp) relation does not necessarily reflect the
solubility of the solid phase. In spite of these uncertainties,
the saturation state calculation results are consistent with the
conclusion that CFA is forming in these sediments.
In sum, all four diagnostic indicators show positive, internally consistent evidence of CFA formation at the FOAM
site. Mass balance arguments coupling porewater fluoride to
solid-phase CFA and the match between SEDEX-determined
solid-phase CFA and solid-phase organic P provide a quantitatively coherent argument for authigenic CFA formation
in these sediments. Evidence for a deficit in porewater phosphate as predicted by the ammonia profile and the saturation
state of porewaters with respect to CFA are consistent with
this interpretation.
At the Mississippi Delta site, the match
between SEDEX-determined
solid-phase
CFA and solidphase organic P is strongly suggestive of authigenic CFA formation, and the fact that the porewater fluoride profile indicates removal of fluoride to the solid phase is consistent
with this argument. However, because the physical disturbance of this core makes the assumption of steady state untenable, porewater arguments cannot be used in a straightforward way to supplement the solid-phase evidence for CFA
formation. This particular case illustrates the point that solidphase chemistry can be a more accurate, integrated measure
of early diagenetic processes than is porewater. Porewater
can be useful as a very sensitive indicator of solid-phase pro-
cesses, but it is also much more easily perturbed than the
solid phase. For this reason, it is crucial that porewater chemistry be viewed in conjunction
with solid-phase chemistry.
Using this combined
porewater/solid-phase
approach to
characterizing
early diagenetic chemistry, the risk of incorrectly inferring solid-phase processes from porewater profiles
alone is greatly reduced.
That the comparison of integrated areas of the decrease in
organic P and the increase in CFA for both the FOAM and
Delta should yield such good agreement is surprising, given
that non-steady state effects of any significance should promote irregular solid-phase profiles. At the FOAM site, we
conclude that either sediment deposition has occurred in a
steady state manner, or that bioturbational
smoothing has
effectively overprinted
any non-steady state signal. At the
Delta, the smoothly varying solid-phase profiles provide evidence that the porewater disturbance reflected in the nutrient
profiles is due either to physical pumping or to biological
irrigation and not to sediment slumping. At this site the solidphase is a more accurate integrated measure of early diagenetic processes than is the porewater.
Diagenetic Redistribution of Phosphorus
The formation of CFA at the expense of organic P at the
FOAM and Mississippi Delta sites is a clear example of diagenetic redistribution
of P in sediments (Fig. 4). The five
Authigenic apatite from
vals,
the sediments are nearly 100% retentive of P. The evidence for growth of CFA at the expense of organic P indicates
that the mechanisms of retention include diagenetic redistribution among reservoirs. Direct measurement of phosphate
diffusive flux and estimates based on porewater phosphate
profiles document a tlux of phosphate from surficial sediments
to bottom water during most times of year (ALLER, 1980;
KROM and BERNER, 198 1). Finer sampling intervals near
the sediment-water interface should therefore reveal net loss
of solid-phase P near this interface at the FOAM site. Diffusive
loss in surficial sediments, however, need have little bearing
on subsequent diagenetic transfer that occurs at greater depth,
as argued previously for the FOAM site. Reconciling diffusive
loss of phosphate from FOAM surficial sediments with retention via diagenetic redistribution at depth, we place the
horizon below which maximum retention occurs at 20 cm.
The porewater phosphate gradient below this depth is in fact
supplying phosphate to the zone of P removal and does not
ultimately result in loss to overlying water. The impact of
enhanced retention on P recycling back into the water column, and more precise determination of the depth below
which diffusive loss of phosphate becomes insignifi~nt, will
require finer depth-scale sampling near the sediment-water
interface. Since short-term perturbation of porewater profiles
will not necessarily disrupt the time-integrated record provided by the solid phase, detailed P speciation of the solid
phase sampled on a fine depth-scale will be ~~~1~1~
useful
in cases, like the Mississippi Delta core discussed here, in
which porewater has been disturbed. Despite the limitations
imposed by the coarse depth intervals employed in the cores
studied, the impact of enhanced sediment retention of P via
diagenetic r~is~bution
on the ultimate burial of P can be
addressed with the current data set.
An important consequence of the diagenetic redistribution
of P is the ultimate burial of P in association with phases
different from its initial association, or sink-switching. The
case discussed here, that of fo~ation of CFA at the expense
of organic P, has important ramifications for the use of sedimentary organic C:P ratios to calculate phosphorus burial
fluxes. Generally, one value is taken as the best estimate of
an average sedimentary organic C:P ratio, and this is coupled
to a global estimate for carbon burial to obtain a value for P
burial f FR~ELICH et al., 1982; MACH et al., 1987; DELANEY
and BOYLE, 1988). It is probably unreasonable to assume
one C:P ratio as representative of average sedimentary organic
matter ( RUTTENBERG, 1990; INGALLand VAN CAPPELLEN,
1990), but particularly if this parameter is taken at a depth
below the sediment-water interface, or averaged over an entire
core, it may not reflect the original burial term for organic
P. If a large portion of the P initially deposited in association
with organic matter is remineralized within the sediment, it
may ultimately be buried in association with a secondary
inorganic phase, such as authigenic CFA ( FROELICHet al.,
1982; RUTTENBERG, 1990). At the FOAM and Delta sites,
roughly 50% of the organic P is lost to CFA and is retained
in the sediment by this mechanism. Thus organic C:P ratios
based on con~ntration me~urements of organic P at depth
in sediments may seriously underestimate the true quantity
of organic P that was originally deposited. This will also have
implications for paleoceanographic and paleoclimate models,
which are based on the premise that oceanic biological cycles
affect atmosphe~c COZ and 02 as a result of changes in the
balance between biological production, respiration, and burial
of reduced carbon ( BROECKER,1982a,b; MCELROY, 1983;
HOLLAND,1984; KNOX and MCELROY, 1984; BOYLE,1990).
Use of coupled water column-sediment
models which employ sedimentary organic C:P ratios to link marine productivity to atmospheric CO2 and 0, will be useful only if this
parameter is carefully evaluated. Quantification of secondary
minerals (e.g., CFA) formed from remineralized P may provide a means for estimating the C:P ratio of organic matter
initially buried, unless sources other than organic P are significant.
Implications of Authigenic CFA Formation in
Non-upwelling Environments
The observation that authigenic carbonate fluompatite
forms in non-upwelling environments such as those represented by the FOAM and Mississippi Delta sites indicates
that conditions necessary for CFA formation are not restricted
to upwelling areas. The key characteristic of deposits formed
beneath upwelling currents may be that CFA con~ntmtions
are less diluted by continentally derived terrigenous sediment
flux ( MANHEIMet al., 1975; CAI.VERTand PRICE, 1983; BATURIN, 1983; RIGGS, 1984). We suggest that authigenic CFA
is forming at FOAM and the Mississippi Delta, but that the
concentration of P as CFA in these sediments is well below
that which defines a phosphorite deposit. The distinguishing
features of CFA formation in environments such as these
from environments of phosphorite formation may not imply
a difference in the mechanism of formation of the authigenic
mineral, but of the general ~imenta~
en~ronment. The
feature of minimal dilution of CFA can be enhanced if the
sediment is exposed to the winnowing action of currents, a
generally accepted phase in the genesis of phosphor&e deposits
( KOLODNY,198 1, and references therein; although reworking
is not a necessary factor in concentration of CFA, e.g., BURNETT et al., 1988). The possibility that CFA forms by a different mechanism in upwelling environments cannot be ruled
out, however. The data presented here indicate that CFA
formation occurs dispersed in the sediments of the non-upwelling environments studied, whereas CFA occurs in discrete
layers in many upwelling environments ( JAHNKEet al., t 983;
VAN CAPPELLENand BERNER, 1988; SCHUFYERT,1988). In
addition, remineralized organic P appears nearly sufficient
as a source for authigenic CFA at FOAM and the Mississippi
Delta, whereas other sources are required to explain CFA
formation in Peru margin sediments (~OELICH et al., 1988).
There have been other observations of CFA formation in
non-upwelling environments, but these have usually been
attributed to special, possibly site-specific formation mechanisms (e.g., MC&LVEY, 1967; GAUDETTE and LYONS,
1980; O’BRIEN et al., 198 1; HEGCIE et al., 1990). Unlike
these specialized environments, the formation of dispersed
authigenic apatite in non-upwelling environments such as
those represented by the FOAM and Mississippi Delta sites
1004
K. C. Rutten~~ and R. A. Berner
represents a potentially large burial flux for phosphorus which
has previously gone unrecognized. These moderately organicrich ( - 1 wt% C,,) sediments, dominated by terrigenous debris and characterized by high sediment accumulation rates,
are representative of an important fraction of continental
margin sediments. If CFA formation is a general occurrence
in sediments of this type, the burial flux of reactive P as authigenic CFA will be signi~~n~y larger than estimated in
previous budgets (l?ROELICH et al., 1982) which considered
burial of P as CFA in upwelling regions only.
The deltaic ferric iron P reservoir is another potential sink
for reactive P which has only been identified through use of
the SEDEX method (Table 1b). Considering this reservoir
in conjunction with the CFA reservoir in non-upwelling continental margin environments, it is clear that the burial flux
of P to continents margins has been underestimate
in previous studies. It is necessary to obtain a larger, more representative solid-phase P data set before a well-constrained
budget can be constructed. Preliminary estimates of the effect
of continental margin burial fluxes of P on the oceanic residence time of P are that it may drop by as much as 50% of
the currently accepted value ( RUTTENBERG, 1990).
If formation of CFA in sediments similar to those represented by the FOAM and Mississippi Delta is a ubiquitous
phenomenon in the modem ocean, it is likely that this was
also the case in the geologic past. Due to the small quantity
and dispersed nature of CFA in these sediments, this phase
is not readily apparent through normal mineral identification
techniques. It seems likely, therefore, that many if not most
black or gray marine shales host dispersed CFA in quantities
similar to those found in the FOAM and Mississippi Delta
sites. In view of this possibility, the notion that the episodic
formation of phosphor&es throu~out geologic time (COOK
and ~~ELHINNY, 1979) is of impo~ance to the marine P
cycle may require reassessment. The phenomenon of episodicity may hold true for phosphorite deposits, but not for
authigenic CFA formation which occurs dispersed in organicrich sediments. Although the average P content of shales is
less than that of phosphorites (0.07 wt% P in shales, TUREKIAN, 1972; vs. 2-18 wt% P in phosphor&es, e.g., COOK,
1984), the greater abundance of shales over phosphor&es
throughout geologic time ( RONO~, 1976) indicates that far
more P is present in shales. If the fraction of the total P as
authigenic CFA which is present in average shale is similar
to that found in sediments from FOAM and the Delta (e.g.,
13% to 9% of the total P, respectively), the quantity of P
buried as CFA in shales is equally as important or more important than P buried as CFA in phosphorites as a marine
sedimentary sink for reactive phosphorus. The significance
of this distinction is that times of phosphorite deposition may
reflect less a change in marine sedimentary conditions favorable to pr~ipitation of authigenic CFA than a change in
depositional environmental conditions to a regime conducive
to segregation and concentration of authigenic CFA into discrete horizons.
CONCLUSIONS
Porewater fluoride and nutrient data, solid-phase P data,
and mass balance models indicate that authigenic carbonate
fluorapatite is currently forming in sediments from the FOAM
site in Long Island Sound and the shallow Mississippi Delta.
The size of the authigenic CFA reservoir increases with depth
at both sites and is mirrored by a decrease in size of the
organic P reservoir, implying that authigenic CFA is being
formed at the expense of organic P. This scenario of sinkswitching, from organic P to authigenic CFA, is supported
by the fact that the calculated integrated amount of the CFA
increase matches the integrated amount of the organic P decrease at both the FOAM site and the Mississippi Delta site.
Furthermore, the amount of CFA formation predicted on
the basis of the porewater fluoride gradient observed at the
FOAM site is quite close to the measured value of CFA at
this site. One consequence of diagenetic redistribution of P
from organic P to CFA is that organic P burial efficiency will
not be accurately quantified by sedimentary organic C:P ratios
in cases where P initially deposited as organic P is ultimately
buried in association with CFA.
In both study sites, the amount of total P is virtually constant with depth, whereas the proportion of P contained in
the different solid-phase P reservoirs is quite variable. This
indicates that the sediments are retentive of P over the depth
intervals sampled, and that one of the mechanisms which
enhances retention is transfo~ation
of P from one reservoir
to another.
The identilication of authigenic CFA formed in non-upwelling environments raises the question: What is the difference between precipitation of CFA in terrigenous dominated
sediments in non-upwelling environments vs. biogenic sediments in classical phosphorite deposit environments? Authigenic CFA precipitation may proceed by different chemical
mechanisms in the two depositional environments. It is
equally possible, however, that phospho~tc deposits are distinctive simply due to the fact that the classical en~ronmen~
in which they form are impoverished in terrigenous detritus
and are subject to winnowing events, resulting in both primary
and secondary concentration of the authigenic phase, whereas
CFA in non-upwelling environments is diluted by terrigenous
debris to the point that it is only identifiable with difficulty.
In the latter case the mechanisms of precipitation need not
be different. The question of the relative quantitative importance of the two environments as sites for removal of
reactive P from the oceans can be addressed despite the fact
that the mechanism of formation in the two environments
remains unresolved. The well-known episodicity curve for
phosphorites through geologic time (COOK and MCELHINNY,
1979) is based on the number and size of economically important deposits. Inclusion of P as CFA in normal sedimentary rocks for which terrigenous dominated sediments in nonupwelling environments are the precursors, e.g., black or gray
shales, in a curve of phosphors burial through time, may
reveal that although episodicity in formation of phospho~te
deposits occurs, this does not translate directly to a difference
in the quantity of P being removed from the oceans.
The identification of authigenic CFA formed in non-upwelling environments as a previously unrecognized sink for
reactive P requires that a revised budget for marine P which
includes this continental margin sink be constructed. Preliminary burial flux calculations indicate that the currently accepted residence time for P may be as much as 50% too high
Authigenic apatite from Continental margins
if continental margins are not included as sites of P burial
( RUTTENBERG, 1990). This would place the residence time
of P on the time scale of facial-interracial
transitions.
Acknowledgments-We thank Don Canfield, Mike Pimer, and Toby
Baldwin for assistance in collecting the FOAM vibrocores; Tim and
Karen Kurkjy Lyons for obtaining the vibrocorer, and R. N. Ginsburg
at the University of Miami for offering us the use of the vibrocorer;
Art Goodhue f&r assisting in modification of the vibrocorer to accommodate core liners; John Morse for inviting KCR on the Mississippi Delta cruise; Saulwood Lin, Miguel Huerta-Diaz, Martha
Scott, and the crew of the R/V Gyre for assistance in core collection
in the Gulf of Mexico; Toby Baldwin and Karen Lyons for assistance
in the lab, Don Canfield and Philippe Van Cappellen for helpful
discussions on modeling. The manuscript profited from the careful
and insightful reviews of Bill Burnett, Jeff Cornwall, Rick Jahnke,
Bill Martin, and Dan McCorkle. This work was supported by NSF
grants EAR-8420334, EAR-86 17600, and OCE-8903442 to RAB,
and OCE-9101495 to KCR.
Editorial handling: R. C. Aller
REFERENCES
ALLER
R. C. ( 1977) The influence of macrobenthos on chemical
diagenesis of marine sediments. Ph.D. dissertation, Yale Univ.
ALLERR. C. ( 1980) Diagenetic processes near the sediment-water
interface of Long island Sound, 1. Decomposition and nutrient
element ge~hemist~ (S,N,P). in Estaur~~e Physics and Chemistry: Studies in Long Island Sound (ed. 8. SALTZMAN
); Advances
in Geophysics, Vol. 22, pp. 235-348.
BATURING. N. (1983) Some unique sedimentological and geochemical features of deposits in coastal upwelling regions. In
Coastal Uowefffn~-its Sediment Record ted. J. THIEDEand E.
SUESS),P&t B: Sedimentary Records 0f~4ncie~tCoastal ~pweffjng,
pp. 1l-27. Plenum Press.
BATURING. N. ( 1988) Disseminated phosphorus in oceanic sediments-a review. hfar. Geol. 84, 95-104.
BERNER R. A. ( 1974) Kinetic models for the early diagenesis of
nitrogen, sulfur, phosphorus, and silicon in anoxic marine sediments. In The Sea fed. E. GOLDBERG). Vol. 5. DU.
. . 247-450. J.
Wiley & Sons.
BERNERR. A. ( 1977) Stoichiometric models for nutrient regeneration
in anoxic sediments. Limnol. Oceanogr. 22,781-786.
BERNERR. A. ( 1978) Sulfate reduction and the rate ofdecomposition
of marine sediments. Earth Planet. Sci. Lett. 37,492-498.
BERNERR. A. ( 1980) Early Diagenesis: A Theoretical Approach.
Princeton Univ. Press.
BERNERR. A. ( 1982) Burial of organic carbon and pyrite sulfur in
the modem oceans: Its geochemieal and environmental significance. Amer. J. Sci. 282, 451-473.
BONATTIE., FISHERD. E., JOENSUUO., and RYDELLH. S. ( 1971)
Post-depositional mobility of some transition elements, phosphoNS, uranium and thorium in deep sea sediments. Geochim. Cosmochim. Acta 35, 189-201.
BOUDREAUB. P. and CANFIELDD. E. ( 1988) A provisional diagenetic
model for pH in anoxic porewaters: Application to the FOAM
site. J. Mar. Rex 46, 429-455.
BOYLEE. A. ( 1990) Quatemary deepwater paleoceanography. Science
249,863-870.
BROECKERW. S. ( 1982a) Glacial to interglacial changes in ocean
chemistry. Progr. Oceanogr. 11, 15 l-197.
BROECKERW. S. ( 1982b) Ocean chemistry during glacial time. Geuchim. Cosmochim. Acta 46, 1689- 1705.
BROECKERW. S. and FENGT.-H. ( 1982) Tracers in the Sea. Eldigio
Press.
BURNETTW. D. ( 1977) Geochemistry and origin of phosphorite
deposits from off Peru and Chile. GSA Bull. 88, 8 13-823.
BURNETTW. C., ROE K. K., and PIPER D. 2. ( 1983) Upwellingand
phosphorite formation in the ocean. In Coastal Upweffing-Its
1005
Sediment Record (ed. J. THIEDEand E. SUESS), Part A: Responses
ofthesedimentary Regime to Present Coastal Upwe[fing, pp. 377391. Plenum.
BURNED W. C., LANDINGB., CHIN P., LYONSW. B., and OREM
B. ( 1987) Jellyfish Lake, Palau: A modern analogue for island
phosphorite formation? Eos 68, 168 1.
BURNETT W. C., BAKER K. B., CHIN P. A., MCCABE W., and
DITCHBURNR. ( 1988) Uranium-series and AMS “‘C studies of
modern phosphatic pellets from Peru shelf muds. Mar. Geof. SO,
215-230.
CALVERTS. E. and PRICEN. B. ( 1983) Geochemistry of Namibian
Shelf sediments. In Coastal Upwefling-Its Sediment Record (ed.
J. THIEDEand E. SUE%), Part A: Re~sponsesof the Sedimentary
Regime to Present Coastal Upweffing, pp. 337-375. Plenum.
CANFIELDD. E. ( 1988) Sulfate reduction and the diagenesis of iron
in anoxic marine sediments. Ph.D. dissertation, Yale Univ.
CANFIELDD. E. ( 1989) Reactive iron in marine sediments. Geochim.
Cosmochim. Acta 53,619-632.
CHEN Y. S. R., BUTLERJ. N., and STUMMW. ( 1973) Kinetic study
of phosphate reaction with aluminum oxide and kaolinite. Environ.
Sci. Tech. 7, 327-332.
COOK P. J. ( 1984) Spatial and temporal controls on the formation
of phosphate deposits: A review. In Phosphate Minerals (ed. J. 0.
NRIAGUand P. B. MOORE), pp. 242-273. Springer-Verlag.
COOK P. J. and MCEL.HINNYM. W. ( 1979) A reevaluation of the
spatial and temporal dist~but~on of ~dimenta~ phospho~te deposits in the light of plate tectonics. &an. Geol. 74, 3 15-330.
DELANEYM. L. and BOYLEE. A. ( 1988) Tertiary paieoceanic chemical variability: Unintended consequences of simple geochemical
models. Paleoceanography 3, 137-l 56.
DE LANGE G. J. ( 1986) Early diagenetic reactions in interbedded
pelagic and turbiditic sediments in the Nares Abyssal Plain (western
North Atlantic): Consequences for the composition of sediment
and interstitial water. Geochim. Cosmochim. Acta 50,2543-256 1.
EMERSONS. E. ( 1976) Early diagenesis in anaerobic lake sediments:
Chemical equilibria in interstitial waters. Geochim. Cosmochim.
Acta 40,925-934.
EMERSONS. and WIDMERG. ( 1978) Early diagenesis in anaerobic
lake sediments. II. Thermodynamic and kinetic factors controlling
formation of iron phosphate. Geochim. Cosmochim. Acta 42, 13071316.
FROELICHP. N., BENDERM. L., LUEDTKEN. A., HEATH G. R.,
and DEVRIEST. ( 1982) The marine phosphorus cycle. Amer. J.
Sci. 282,474-5 11.
FROELICHP. N., KIM K. H., JAHNKER., BURNETTW. C., SOUTAR
A., and DEAKINM. ( 1983) Pore-water fluoride in Peru continental
margin sediments. Uptake from seawater. Geochim. Cosmochim.
Acia 47, 1605-1612.
FROELICHP. N., ARTHUR M. A., BURNETTW. C., DEAKIN M.,
HENSIXYV., JAHNKER., KAUL L., KIM K.-H., ROE K., SOUTAR
A., and VATHAKANON
C. ( 1988) Early diagenesis of organic matter
in Peru continental margin sediments: Phosphorite precipitation.
Mar. Geof. 80, 309-343.
GAUDETTEH. E. and LYONSW. B. ( 1980) Phosphate ge~hemist~
in nearshore carbonate sediments: A suggestion of apatite formation, In Marine Phosphorites (ed. Y. K. BENTOR);SEPM Spec
Pub. No. 29, pp. 2 15-225.
GLENNC. R. and ARTHURM. A. ( 1988) Carbon isotopic composition
and lattice-bound carbonate of Peru-Chile margin phosphor&es.
Mar. Geol. SO, 287-307.
GOLDHABERM. B., ALLER R. C., COCHRANJ. K., ROSENFELD
J. K., MARTENSC. S., and BERNERR. A. ( 1977) Sulfate reduction,
diffusion, and bioturbation in Long Island Sound sediments: Report
from the FOAM group. Amer. J. Sci. 277, 193-237.
GREENHALGHR. and RILEYJ. P. ( 196 1) The determination of fluorides in natural waters with particular reference to sea water. Anal.
Chim. Acta 25, 179- 188.
GULBRANDSEN
R. A. ( 1966) Chemical composition of phosphorites
of the Phosphoria Formation. Geochim. Cosmochim. rlcta 30,769778.
GL’LBRANDSEN
R. A. ( 1969) Physical and chemical factors in the
formation of marine apatite. Econ. Geof. 64, 365-382.
HARTMANNM., MOL.LERP., SUESS E., and VAN VER WEUDEN
1006
K. C. Ruttenberg and R. A. Berner
C. H. ( 1973) Oxidation of organic matter in recent marine sediments. Meteor Forsch-Ergebnisse C, 74-86.
HARTMANNM., MOLLER P., SUESSE., and VAN DER WEIJDEN
C. H. ( 1976) Chemistry of late quaternary sediments and their
interstitial waters from the NW African continental margin. Meteor
Forsch-Ergebnisse C, 1-67.
HECGIED. T., SKYRINCG. W., O'BRIENG. W., REIMERSC., HERCZEG A., MORIARTY D. J. W., BURNE’I-~W. C., and MILNES
A. R. ( 1990) Organic carbon cycling and modem phosphorite
formation on the East Australian continental margin: An overview.
In Phosphorite Research and Development (ed. A. G. J. NOTHOLT
and 1. JARVIS); Geol. Sot. Spec. Publ. No. 52, pp. 87-117. The
Geol. Sot. London.
HOLLAND H. D. (1978) The Chemistry of the Atmosphere and
Oceans. Wiley Interscience.
HOLLANDH. D. ( 1984) The Chemical Evolution of the Atmosphere
and Oceans. Princeton Univ. Press.
HOWARTHR. W. (1988) Nutrient limitation of net primary production in marine ecosystems. Ann. Rev. Ecoi. 19, 89-I IO.
INGALLE. I. and VAN CAPPELLENP. ( 1990) Relation between sedimentation rate and burial of organic phosphorus and organic carbon in marine sediments. Geochim. Cosmochim. Acta 54, 373386.
JAHNKER. A. ( 1984) The synthesis and solubility of carbonate fluorapatite. Amer. J. Sci. 284, 58-78.
JAHNKE R. J., EMERSONS. E., ROE K. K., and BURNETTW. C.
( 1983) The present-day formation of apatite in Mexican continental margin sediments. Geochim. Cosmochim. Acta 47, 259266.
JAHNKER. A., EMERSONS. R., REIMERSC. E., SCHUFFERTJ., RUTTENBERGK., and ARCHERD. ( 1989) Benthic recycling of biogenic
debris in the eastern tropical Atlantic ocean. Geochim. Cosmochim.
Acta 53, 2947-2960.
JORGENSENB. B. ( 1983) Processes at the sediment-water interface.
In The Major Biogeochemical Cycles and Their Interactions (ed.
B. BOLINand R. B. COOK); Scope 21, pp. 477-507. Wiley.
KLUMP J. V. and MARTENSC. S. ( 1981) Biogenic recycling in an
organic-rich coastal marine basin. II. Nutrient sediment-water exchange processes. Geochim. Cosmochim. Acta 45, 10 1- 12 1.
KLUMP J. V. and MARTENSC. S. ( 1987) Biogeochemical recycling
in an organic-rich coastal marine basin. 5. Sedimentary nitrogen
and phosphorus budgets based upon kinetic models, mass balances,
and the stoichiometry of nutrient regeneration. Geochim. Cosmochim. Acta 51, 1161-l 173.
KNOX F. and MCELROYM. B. ( 1984) Changes in atmospheric COr:
Influence of the marine biota at high latitude. J. Geophys. Res.
89,4629-4637.
KOLODNYY. ( 198 1) Phosphorites. In The Sea (ed. C. EMILIANI),
Vol. 7, pp. 981-1023. Wiley.
KOROLEFFF. ( 1976) Determination of nutrients. In Methods ofSeawater Analysis (ed. K. GRASSHOFF),pp. 125-157. VerIag-Chimie.
KRISHNASWAMI
S., WESTRICHJ., BENNETTJ. T., MONAGHANM.,
and TUREKIANK. K. ( 1982) Radionuclide and elemental study
of sediment cores from the FOAM site. Eos 63, 114.
KROM M. D. and BERNERR. A. ( 1980) The diffusion coefficients
of sulfate, ammonium and phosphate in anoxic marine sediments.
Limnol. Oceanogr. 25, 327-337.
KFCOMM. D. and BERNERR. A. ( 1981) The diagenesis of phosphorus
in a nearshore marine sediment. Geochim. Cosmochim. Acta 45,
207-216.
LEGEROSR. Z. ( 1965 ) Effect of carbonate on the lattice parameters
of apatite. Nature 206, 403-404.
LEGEROSR. Z., LEGEROSJ. P., TRAUTZ0. R., and SHIRRAW. P.
( 1967) Apatite crystallites: Effects of carbonate ion on morphology.
Science 155, 1409-1411.
LEGEROSR. Z., Go P., VANDEMAELEK. H., QUIROLGICOG. B.,
and LEGEROSJ. P. ( 1980) Transformation of eaalciumcarbonates
and calcium phosphates to carbonate apatites: Possible mechanism
for phosphorite formation. Zndlnt. Congress on Phosphorus Compounds, 89- 103.
LIN S. ( 1990) Environmental controls on sulfate reduction and iron
sulfide mineral formation. Ph.D. dissertation, Texas A&M Univ.
MACH D. M., RAMIREZA., and HOLLANDH. D. ( 1987) Organic
phosphorus and carbon in marine sediments. Amer. J. Sci. 278,
429-441.
MANHEIMF. T., ROWEG. T., and JIPAD. (1975) Marine phosphorite
formation off Peru. J. Sediment Petrol. 45, 243-25 1.
MARTENSC. S. ( 1987) Labile organic matter input and benthic biogeochemical cycling processes in continental slope sediments. Report of the Workshop on the Impact of Ocean Boundaries on the
Interior Ocean, 27-38.
MARTENSC. S., BERNERR. A., and ROSENFELDJ. K. ( 1978) Interstitial water chemistry of anoxic Long Island Sound sediments.
2. Nutrient regeneration and phosphate removal. Limnoi. Oceanogr. 23,605-617.
MCARTHUR J. M. ( 1985) Francolite geochemistry: Compositional
controls during formation, diagenesis, metamorphism and weathering. Geochim. Cosmochim. Acta 49, 23-35.
MCCLELLAN,G. H. ( 1980) The mineralogy of carbonate fluorapatite.
J. Geol. Sot. London l31,675-68 1.
MCELROYM. B. ( 1983) Marine biological controls on atmospheric
CO2 and climate. Nature 302, 328-329.
MCKELVEYV. E. ( 1967) Phosphate deposits. U.S. Geol. Survey Bull.
1252-D, 1-22.
MOODYJ. B., CHABOUDYL. R., and WORSLEYT. R. ( 1988) Pacific
pelagic phosphorus accumulation during the last 10 M.Y. Paieoceanography 3, 113- 136.
MULLERP. J. and SUESSE. ( 1979) Productivity, sedimentation rate,
and sedimentary organic matter in the oceans. I. Organic carbon
preservation. Deep-Sea Res. 26A, 1347- 1362.
MURRAYJ. W., GRUNDMANISV., and SMETHIEW. M. J. ( 1978)
Interstitial water chemistry in the sediments of Saanich Inlet. Geochim. Cosmochim. Acta 42, 101 I-1026.
O’BRIENG. W., HARRISJ. R., MILNESA. R., and VEEHH. H. ( 1981)
Bacterial origin of East Australian continental margin phosphor&.
Nature 294,442-444.
O’BRIEN G. W., VEEH H. H., CULLEN D. J., and MILNES A. R.
( 1986) Uranium-series isotopic studies of marine phosphorites and
associated sediments from the East Australian continental margin.
Earth Planet. Sci. Lett. 80, 19-35.
O’BRIEN,MILNESA. R., VEEH H. H., HEGGIED. T., RIGGS S. R.,
CULLEND. J., MARSHALLJ. F., and COOKP. J. ( 1990) Sedimentation dynamics and redox iron-cycling: Controlling factors for
the apatite-glauconite association on the East Australian continental
margin. In Phosphorite Research and Development (ed. A. G. J.
NOTHOLTand I. JARVIS); Geol. Sot. Spec. Pubi. No. 52, pp. 6186. The Geol. Sot. London.
REIMERSC. E. and SUESSE. ( 1983) Spatial and temporal patterns
of organic matter accumulation on the Peru continental margin.
In Coastal Upwelling-Its Sediment Record (ed. J. THIEDEand
E. SUESS), Part B: Sedimentary Records of Ancient Coastal Upwelling, pp. 377-397. Plenum.
RIGGS S. R. ( 1979) Petrology of the Tertiary phosphorite system of
Florida. Econ. Geol. 74, 195-220.
RIGGSS. ( 1984) Paleoceanographic model of Neogene phosphorite
deposition, U.S. Atlantic continental margin. Science 223, 123131.
RONOVA. B. ( 1976) Global carbon geochemistry, vulcanism, carbonate accumulation, and life. Geochemistry International 13,172195 (English translation of Geokhimiya).
ROSENFELD
J. K. ( 1979) Ammonium adsorption in nearshore anoxic
sediments. Limnoi. Oceanogr. 24, 356-364.
ROSENFELDJ. K. ( I98 1) Nitrogen diagenesis in Long Island Sound
sediments. Amer. J. Sci. 281, 436-462.
RUDE P. D. and ALLERR. C. ( 199 1) Fluorine mobility during early
diagenesis of carbonate sediment: An indicator of mineral transformations. Geochim. Cosmochim. Acta 55,249 1-25 10.
RUTTENBERGK. C. ( 1990) Diagenesis and burial of phosphorus in
marine sediments: Implications for the marine phosphorus budget.
Ph.D. dissertation, Yale Univ.
RUTTENBERGK. C. ( 1992) Development of a sequential extraction
method for different forms of phosphorus in marine sediments.
Limnol. Oceanogr. 37, 1460-1482.
RUTTENBERG
K. C. and CANF’IELD
D. E. ( 1988) Fluoride adsorption
onto ferric oxyhydroxides in sediments. Eos 69, 1235.
SANDSTROM
M. W. ( 1982 ) Diagenesis of organic phosphorus in ma-
1007
Authigenic apatite from continental margins
tine sediments: Implications for the global carbon and phosphorus
cycles. In Cycling of Carbon, Nitrogen, Sulfur and Phosphorusin
Terrestrialand Aquatic Ecosystems (ed. J. R. FRENEY and I. E.
GALBALLY), pp. I33- 14 I. Springer-Verlag.
SARMIENTO
J. L. and TOGGWEILERJ. R. ( 1984) A new model for
the role of the oceans in determining atmospheric Pee. Nature
308,620-624.
SCHUFFERTJ. D. ( 1988) Multi-layered authigenic apatite formation
off southern Baja California. Eos 69, 1234.
SHELIXINR. P. ( 1981) Ancient marine phosphorites. Ann. Rev. Earth
Planet. Sci. 9, 25 l-284.
SHERWOODB. A., SAGERS. L., and HOLLANDH. D. ( 1987) Phosphorus in foraminiferal sediments from North Atlantic Ridge cores
and in pure limestones. Geochim. Cosmochim. Acta 51, 18611866.
SMITH S. V. ( 1984) Phosphorus versus nitrogen limitation in the
marine environment. Limnol. Gceanogr. 29, 1149- 1160.
STRICKLAND
J. D. H. and PARSONST. R. ( 1972) A PracticalHandbook of SeawaterAnalysis. Fish. Res. Bd. Canada Bull.
SUESSE. ( 1980) Particulate organic carbon flux in the oceans: Surface
productivity and oxygen utilization. Nature 288, 260-263.
TUREKIANK. K. ( 1972) Chemistry of the Earth. Holt-ReinhartWinston.
VANCAPPELLEN
P. and BERNERR. A. ( 1988) A mathematical model
for the early diagenesis of phosphorus and fluorine in marine sediments: Apatite precipitation. Amer. J. Sci. 288, 289-333.
WARNER T. B. ( 1971) Electrode determination of fluoride in illcharacterized natural waters. WaterRes. 5,459-465.
WESTRICHJ. T. ( 1983) The consequences and controls of bacterial
sulfate reduction in marine sediments. Ph.D. diss., Yale Univ.
APPENDIX
The principal assumption required when use is made of a stoichiometric nutrient regeneration model is that regeneration rates of
N and P bear a constant ratio to one another. Other required assumptions are: steady state deposition and burial; first order decomposition kinetics; rapid, reversible adsorption of phosphate and ammonia which follows simple linear isotherms; negligible compaction
and porewater flow; simple molecular diffusion below the mixed zone.
The ammonia profile was fit to the equation ( BERNER,1980)
where
C = Co + A[1 - exp(-Bx)],
(Al)
A = FN,,w*/Dku + (1 + K)w*, and
(A2)
B = (ku/w).
(A3)
For the FOAM site, the sedimentation rate (0) = 0.1 cm/yr; F
= [( 1 - g)/&~], = 1.35 g/cm’; the molecular diffusion coefficient
(II) for ammonia = 309 cm*/yr (KROM and BERNER, 1980); the
reversible adsorption constant for ammonia (K) = 1.3( ROSENFELD,
1979). From the fit the constants were evaluated as A = 3200 rmol/
L, B = 0.0065 cm-‘. Evaluation of the first order rate constant ku
from B and w yields a value of 0.00065 yr-’ . Evaluation of No from
A yields a value of 53.1 rmoles N/g. This compares favorably to
measured organic nitrogen values obtained for the FOAM site (RoSENFELD,1980) of 62.3 rmoles N/g.
The rate constant obtained above for ammonia regeneration (ku)
should be equal to the rate constant for phosphate regeneration according to the assumption of stoichiometric nutrient regeneration,
e.g., ku = kr (BERNER, 1977). The pre-exponential constant A was
evaluated for phosphate using Eqn. A2, the value for kr calculated
from the empirical fit to the ammonia data, and the appropriate
constants for phosphate: the molecular diffusion coefficient (D) for
phosphate = 100 cm*/yr (KROM and BERNER,198 I ); the reversible
adsorption constant for phosphate (K) = 2.0 ( KROM and BERNER,
198 I ); the initial concentration of solid-phase organic P (PO) = 3.1
rmoles P/g (value is from the O-10 cm interval SEDEX organicP). The pre-exponential constant is evaluated as A = 435. The resulting predicted profile is shown in Fig. 5.