LimnoI. Oceanogr., 33(6, part 2), 1988, 1559-1567
0 1988, by the American Society of Llmnology and Oceanography, Inc.
Twin cosmogonic radiotracer studies of phosphorus recycling and
chemical fluxes in the upper ocean
D. Lal
Scripps Institution of Oceanography, University of California, San Diego, La Jolla 92093
Y. Chung
Institute of Marine Geology, National Sun Yat-sen University, Kaohsiung, Taiwan 80424
T. Platt
Marine Ecology Laboratory, Bedford Institute of Oceanography, Dartmouth, Nova Scotia B2Y 4A2
T. Lee
Scripps Institution of Oceanography,
Abstract
We determined the feasibility of using two cosmogonic radionuclides, 32P(half-life, 14.3 d) and
‘SP (half-life, 25.3 d) as tracers for studying nutrient cycling and chemical fluxes in the upper layers
of the ocean. Various chemical procedures were tried to extract dissolved inorganic phosphorus
(DIP) from several thousand liters of seawater as required to measure the natural activities of 32P
and 33P.Wc developed a simple and convenient procedure for extraction of DIP from about 103
liters of seawater h 1, simultaneously from several depths. The activities of 32P and 33P were
measured with a low-level Q-gas flow counter. The 32Pactivities in coastal waters off La Jolla and
San Diego range between 0.5 and 0.9 dpm m-’ of seawater. The 32P-specific activities in coastal
plankton samples range between 20 and 100 [dpm (g P)- ‘], zooplankton samples being consistently
lower than phytoplankton samples. The phytoplankton samples have 32Pactivities (dpm m-’ of
seawater) close to those in surface waters.
The observed concentrations and specific activities of these tracers in dissolved and particulate
phosphorus pools are in excellent agreement with expectations. Our studies demonstrate their
usefulness as suitable tracers for studies of P pathways through the pelagic food chain and P flux
through the thermocline, averaged over a time scale of 1-2 months.
Nutrient cycling and chemical fluxes in
the oceans are recognized as central problems in biological and chemical oceanography. Most particulate organic matter is
recycled in the upper layers by heterotrophic organisms of all sizes. Inventories of
Acknowledgments
We are grateful to H. Craig for allowing us to use
the beta-counting facilities and to S. P. Wallace for a
timely gift of acrilan fibers.
We acknowledge assistance and advice on chemical
procedures from M. Honda, S. Krishnaswami, K. Nishiizumi, and B. L. K. Somayajulu. For discussions
and advice, we are thankful to Farooq Azam and George
Jackson. For valuable and extensive comments on the
manuscript, we are thankful to P. A. Jumars.
This research was conducted in part under ONR
Grant USN NOOO14-87-K-0005 and in part by a gift
from Mr. Myron Eichen. Seed funds for testing the
feasibility of the method were made available by LeRoy
Dorman. Support from S10 throughout the research
work, including ship time, is gratefully acknowledged;
we particularly thank M. Mullin and G. Shor for their
encouragement.
dissolved inorganic and organic C, N, and
P depend on a number of very complex
planktonic food webs. During the last few
years, several important discoveries, e.g.
substantial primary production by picoplankton (Platt and Li 1986) and a dynamic
microbial food web (Azam 1986) have altered our perception of the structure and
function of the marine ecosystem.
Various techniques are currently available for measuring the net flux of carbon
out of the euphotic zone, which equals “new
production” in steady state. The first technical advance in this direction was made by
Dugdale andGoering(1967) who suggested
production due to allochthonous nitrogen
in the euphotic zone as a measure of the
new production. Measurements of new production using ]5N-labeled N03- and NH4-’
were summarized by Eppley and Peterson
(1979). Hayward (1987) estimated nitrate
flux through the nitracline. Particulate flux-
1559
La[ et al.
1560
Table 1. Cosmogonicradionuclidesin the oceanswith half-lives >10 d.
=
Principal target element(s)
—
Half-life
32p
33p
37Ar
‘Be
35s
22Na
3H
32C&
B~Ar
“c
4’Ca*
SIKr
36C1
$
2cA]
10Be
14.3 d
25.3 d
35.0 d
53.3 d
87.4 d
2.6 yr
12.3 yr
-200 yr
269 yr
5,730 yr
1.0x105yr
2.1x105yr
3.Ox105yr
7.2x 105yr
1.6x 10’yr
.—
Atm.
In situ
Ar
Ar
Ar
N, O
Ar
Ar
N, O
Ar
Ar
N, O
—
Kr
Ar
Ar
N, O
Cl, S, K
Cl, S, K
K, Ca
0
Cl, Ca, K
Na
(), 2H
S, Ca
K, Ca
0’
Ca
Sr
c1
S, K, Ca
0
Global avg surface
injection rate*
~
Integrated in situ oceanic
production rate~
(atoms cm- 2 rein-’)
—
5.82 x10-’
6.93 x 10-3
9.1OX1O-’
1.27
2.84 x 10-2
3.75 x 10-3
1.39 XI0’
9.60 x10-3
2.00X 10-1
I.2OX1O2
—
2.30x 10-’
6.60 x 10-2
8.40 x 10-3
2.70
7.6x 1O-4
2.9 x 1o-4
8.1X10”””
6.OX10-3
5.1 x 10-4
3.9 x 10”-4
1.2 X1C)-2
2.5 x 10-~
1.2x 10-””
9.OX1O-3
2.4 X10-” (~)
1.9X 10-’ (~)
1.06x 10-[ (n)
6.8 X10-6
1.8x 10-3
* Ilased on atmospheric production estimates of Lal and Peters 1967. The mean sr.ratosphere-t roposphere exchange time is taken to be 2 yr,
Tropospheric rcsidcncc lime of isotopes that can bc scavenged by wet precipitation is taken to be 40 d. Atmospheric residence times for lW, 39Ar,
and 81Kr are taken to be 10, 270, and 170 yr.
t Present calculations.
$ Flux to oceans from rivers has to bc considered duc to production in rocks and soil by 40Ca(n, ~) 4’Ca reaction no estimates are given here because
of Iargc uncertai rrtics in these calculations.
$ As above; due to 35C1(rz,T) 3’CI reaction.
es at different depths in the oceans have
routinely been made with sediment traps
(Knauer and Martin 198 1).
Techniques for in situ carbon flux measurements have, however, suffered several
problems. Since the rates of recycling and
new production are temporally variable, new
production estimates must take into account net biomass changes in the euphotic
zone (Platt et al. 1984). Marked temporal
variations in nitrate flux across the thermocl ine have been observed (Hayward
1987) whereby principal nitrate influx may
happen during short episodes of high eddy
diffusivity. Measurements of particle fluxes
also present problems. Recent work has
shown that the fluxes of particulate organic
carbcm (PO~) and calcareous and silicate
materials at depths are in large part ( > 90°/0)
due to mechanisms other than transport by
zooplankton
fecal pellets (Pilskaln
and
Honjo 1987). Food web interactions reduce
fecal particle contributions to mass flux in
the deep sea by processes such as degradation of fecal particles by microbial activity,
coprophagy, and the feeding of vertically
migrating zooplankton in the upper 5001,000 m.
Considering the importance
of understanding nutrient recycling in the upper
ocean to marine food webs and of trace element and C032- regulation via sinking flux
of organic materials, and realizing that the
artificial tracer methods are not truly representative of field-scale processes, it would
be a significant step forward if any of the
naturally occurring nuclides could be used
as tracers. In this paper we discuss that indeed there are two naturally occurring radiotracers available, belonging to the element phosphorus, 32P (half-life, 14.3 d) and
33P (half-life, 25.3 d). We also demonstrate
that they can be measured conveniently and
with the required degree of accuracy to
quantifj cycling between inorganic and organic dissolved and particulate phosphorus
pools and fluxes out of the mixed layer averagecl over time scales of 2–4 weeks.
Materials and methods
Background on cosmogonic radio nuclides
in the ocean — Several cosmogonic radionuclides are present in the ocean (Table 1)
as a result of injection from the atmosphere
(Lal and Peters 1967) and by their direct
production in the oceans (Davis and Schaeffer 1955; Lal and Peters 1967). Their atmospheric injection rates are estimated from
production rates and fallout calculation procedures outlined by “Lal and Peters (1967)
1561
Cosmogonic 32P, 33P in oceans
and Lal (1988). The in situ oceanic production of 3GC1due to thermal neutron capture in 35C1was first discussed by Davis and
Schaeffer (1955). We have now made estimates for in situ integrated oceanic production rates of isotopes due to thermal”
neutron capture and nuclear spallation reactions due to nuclei of energy >10 MeV
(Table 1). The principal target elements from
which the isotopes are produced are listed
in columns 3 and 4 for atmospheric and
oceanic production. It should be noted here
that the in situ production of isotopes in the
oceans decreases exponentially with an absorption mean free path of 150 cm. Thus
> 950/0of the in situ production occurs within the upper 5 m of the water column.
Except for 37Ar, the principal source of
radionuclides in the ocean is their injection
from the atmosphere; in situ production
corresponds to 10-4 to <9 x 10-’ times
their injection from the atmosphere. In the
case of the radionuclides 32P and 33P, their
in situ productions amount to 13.1 and 4.20/o
of their atmospheric contributions.
The
generally low in situ production rates arise
from two factors: the cosmic ray beam is
appreciably attenuated at the sea surface,
and the oceanic concentrations of most target elements of interest are low. The oceanic
in situ production of 37Ar becomes important because most of the atmospherically
produced 37Ar decays in the atmosphere
(slow gas-exchange rate); that of 36C1is important because of an appreciable concentration of Cl in seawater and the relatively
high cross section for the 35C1 (n, ~) 3bCl
reaction, 43 barns.
The production rates of the two radioisotopes 32Pand 33P(Lal and Peters 1967), and
their fallout in rainwater (Lal et al. 1960),
are consistent with a mean removal time
from the atmosphere, r~ of 40 d (Lal and
Peters 1967). The tropospheric production
rates are fairly independent of latitude (figure 16 in Lal and Peters 1967). On the land,
their fallout is primarily due to wet precipitation; dry fallout constitutes < 20°/0.
The radionuclide 32P is also produced in
situ in oceans by decay of cosmogonic 32Si,
but this source is not important at depths
<500 m (Somayajulu et al. 1987). The
oceanic source function of 32P and 33P can
Table 2. Oceanic source functions of cosmogonic
32Pand 33P.
Parameier
Mean global stratospheric
production rate (atoms
cm-2
~i~
33p
3.2x 1O-Z
2.8x 10-Z
1.6x 1O-’
1.3x1 o-2
5.8x 1o-J
6.9x 1O-3
7.6x 10-”
2.9x 1O-’
1.9X102
3.8x102
0.78
0.65
‘l)*
Mean global tropospheric
production rate (atoms
f.m-2
12p
rein-l)*
Mean tropospheric fallout
rate (atoms cm-2 min 1)*
Mean in situ oceanic production~ (atoms cm-2
rein-])
Mean inventory in the
oceans (atoms cm-2)
Activity in the mixed layer
(50 m) with (l/B)= 30 d
(dpm m-’)
* Based on L.al and Peters 1967.
T From Table 1.
be represented (Table 2) by an injection of
a total of 6.58 and 7.22 x 10-3 atoms cm-2
rein-’ (within the upper 5 m of the ocean
surface. We have also estimated their mean
concentrations in the mixed layer (taken to
be 50 m), calculated on the assumption that
the mean residence time of phosphorus isotopes in this layer is 30 d (removal constant
B = 1/30 d-’).
Extraction of dissolved inorganic 32Pand
“J3Pfrom marine samples — Several largevolume (1–5 X 103 liters) seawater samples
were collected and processed for 32Pand 33P
activities with radiochemical and counting.
procedures in the literature (Lal et al. 1960;
Somayajulu et al. 1987). In experiments at
the Scripps Institution of Oceanography
(S10) pier, we used a portable swimming
pool of 104 liters capacity as a sample container. Shipboard chemical extractions were
also carried out with Fe(OH)3 precipitation
but in large plastic tanks of -1 -m3 volume.
The sample was acidified with HC1 to a pH
of about 2, and a FeC13 solution was added
(* 10-3 g liter-’ of water). Ammonium hydroxide was added slowly to bring the pH
up to 10 in a period of a few hours. The
supernatant was then discarded, and the
Fe(OH)3 precipitate was collected for laboratory purification of the phosphorus. Dissolved inorganic phosphorus (DIP) was
extracted with > 900/0 efficiency. The concentration of DIP in the seawater was de-
La[ et al.
A
AA
a)
A
o 20-50
●
●
mesh
50-100
mesh
A fiber
●
Oao
.
●
o
,
200
...
c>
..
0
1000
600
Space
I.:
+-
0,”
1800
1400
(h-l)
velocity
(D
100
80
60
40
Fe (OH)3 - fiber
20
0
0
Total
,
1
volume
of
1
1
2
3
seawater
1
filtered
4
(103 liters)
Fig. 1. a. Measured adsorption efficiencies for dissolved inorganic phosphorus with Fe(OH)3-form Dowex 20-50- and 50-100-mesh
resins, and Fe(0H)3-coatcd acrilan fibers as a function of space velocity. The
column volume is 1 liter in each case. b. The measured
decrease in the adsorption efficiencies for two Fe(O”H)3iiber columns as a function of the total amount of
seawater filtered through the column for two experiments.
termined with the spcctrophotometric
method developed by Strickland and Parsons (1968); it may give a slight overestimate because
this method
includes
some labile organophosphorus compounds
(Cerebella et al. 1984).
Whereas the Fe(OH)~ coprecipitation experiments proved very successful for the
quantitative extraction of DIP, this technique could not be adapted for routine analyses of 32P and 33P activities, primarily because of the inconvenience
relate~ to
chemical processing on the deck of the ship.
We therefore aimed at developing quantitative techniques involving seawater passage directly through a suitable matrix in
which the DIP could be efficiently adsorbed.
W.e tried two methods for these experiments: an Fe(OH)3 ion-exchange column,
designated the Honda column (Merrill et al.
1960) and Fe(OH)3-coated acrilan fibers (Lal
et al. 1964; Krishnaswami et al. 1972). The
former technique was developed for extraction of dissolved beryllium from seawater
at extremely low concentrations ( <10-’0 M)
and the latter for dissolved silica from surface and deep-sea water samples. The
Fe(OH)3 fibers towed during the GEOSECS
programs showed high extraction efficiency
for DIP (Somayajulu pers. comm.); this finally was confirmed in the laboratory for
seawater containing 0.1 –O.3 ~g-atoms P liter-’.
The Fe(OH)3-form ion-exchange resin and
the fibers were loaded in the conventional
1-liter cartridges used conumercially for filtering and softening industrial water; our
cartridges were obtained from Culligan Industrial. Systems. Seawater was first passed
through a 5-pm particulate filter cartridge
and then through the cartridge containing
the Honda resin or the Fe(OH)3-coated fibers. The particulate filters very efficiently
removed particles (organic and inorganic
phases) as evidenced from the presence of
particles only in the outermost layers of the
filter cartridge. Practically no particles were
seen in layers 5-7 mm from the outer surface, even after passing about 10 m3 of seawater.
With the S10 pier seawater containing
about 0.3 pg-atoms P liter-’ of DIP, DIP
extraction efficiency for the Honda resin is
a sensitive function of the space velocity.in
the range considered (Fig. 1a). Space velocity (h-’) is defined as the volumetric flow
rate (cm3 h-’) normalized to the column
volume (i.e. cm–3). The efficiency, q for
Honda resin of 50–1 00-mesh size is about
1.5 x higher than that for 25-50 mesh. The
effective amount of water cleared of P (t x
space velocity) for the 1-liter cartridge is
-150 liters h-1 for 50–1 00-mesh resin and
70 liters h--l for 25-50 mesh, over a wide
range of space velocities. The result speaks
strongly against the use of the Honda column for extraction of DIP from 3–5 m3 of
seawater within a reasonable period of time
onboard ship. Even so, we did carry out tests
onboard to check the extraction efficiency
of the Honda column in a full-scale test by
passing several tons of water through it. The
results obtained for seawater from Southern
California Bight Study (SCBS) Sta. 205 (off
Cosmogonic 32P, 33Pin oceans
Table 3. Measured concentrations
Sample location
S10 pier
(La Jolla)
S10 pier
Off San Diego
(32°53.5’N,117’’35.5’W)
S10 pier
Sampling
depth
(m)
7
of ‘2P in seawater.
Code
(date)
Effective
volume of
water
processed
(m’)
Swl .SCP
6.3
Fe(OH)~
coprecipitation
0.54*0.07
1.0
Fe(OH),
coprecipitation
Fe(OH),
coprecipitation
Adsorption onto
Fe(OH)~fibers
O.87*O.1O
(23 Aug 85)
7
30
7
1563
SW2.SCP
(1 Ott 85)
SW1.SD
(9 NOV 85)
SW3.SCP
(4 Sep 87)
Santa Catalina Island) showed that the initial DIP extraction efficiency was as determined from shore-based short duration tests
with the S10 pier water, but it quickly
dropped (within an hour) to half the initial
value and kept dropping further with passage of seawater through the column. We
have not been able to ascertain the reason,
but based on the observed presence of a
colloidal precipitate in the Honda column,
we infer that the resin did pick up some
other anions and that its behavior was considerably modified from that expected for
the Honda resin, namely adsorption of dissolved Si, P, and group III elements; channeling in working with high space velocities
may also be an important contributing factor.
We therefore concentrated on the use of
Fe(OH)~-coated acrilan fibers (20-~m diam),
prepared following the procedure outlined
by Somayajulu et al. ( 1987). The shore-based
extraction efficiencies ranged between 95 and
10OO/oover the full range of space velocities,
100–1 ,000 h-1. The extraction efficiencies
were also reproducible in full-scale tests carried out at SCBS Sta. 205. In Fig. lb we
show the DIP extraction efficiency for the
Fe(OH)g fibers as a function of time for the
duration of the experiment (8 h). The slow,
monotonic decrease in efficiency is attributed to saturation of the column primarily
by dissolved silicon, whose concentration is
about an order of magnitude higher than
that of DIP; it is the major dissolved element expected to be adsorbed
on the
Fe(OH)~-coated fibers from seawater.
Four surface samples, three from the S10
pier and one from off San Diego, were ana-
1.6
1.7
Method used for
DIP extraction
32Pactivity
(dpm m-’)
0.83+0.10
0.68+0.07
lyzed for 32P activities (Table 3). Rainwater
collected in a swimming pool on 25 November 1985 at the S10 pier was analyzed
for 32P activity.
Extraction of particulate organic 32Pand
-?3Pfromplankton samples –Photoplankton
samples were collected with a O.5-m ring
net of 76-~m mesh and zooplankton samples with 0.75- and 1.O-m ring nets of 280and 452-pm mesh; the coarser net was used
to sample chaetognaths. Both kinds of samples (Table 4) were collected by series of
oblique tows between 50 m and the surface.
The Celtic Sea Sample was from 50”39’N,
07”00’W, off the SW coast of England. The
Bedford Basin samples were taken from
44°3 1.3’N, 63”38.3’W (Table 4). Seven samples of plankton (14-45 g dry wt), collected
from the Celtic Sea and Bedford Basin, were
processed to determine their 32P activities
(Table 4). Some of the samples consist primarily of particular phyto- or zooplankton
taxa; others contain mixtures.
Plankton samples were first dried at 70”C
Table 4. Specific activities of 32P [dpm (gP)-’] in
marine plankton samples.
Sample lypc
Phytoplankton
Diatom
Diatom + copepod
nauplii
Dinoflagellates +
copepod nauplii
Zooplankton
Chaetognaths
Unspecified
Bedford Basin
Celtic SCa
16 Jun 86 3 Jul 86 16 Jul 8620 Aug 86
—
59.1
–
–
—
76.6
–
—
—
244
–
–
20.3
40.0
22.4
–
—
46.5
–
1564
Lal et aL
I *
o SW I.SCP
●
“’b
● RW. SO
IO
“%
.,
~
● PHY.
BF3
n zoo.
BF3
■ zoo.
BF3
I
‘1(1
~1°000000
1.0
–
b...
● .e.
●
t
●
omm
●
c)
G
o
0
●*O.
00
0
0.2
0.3
0
%
00
:
.E“
.
a>
\
10
5
15
20
~
I
!
10
L
20
1
1
1
I
J
30
40
50
60
70
w
a
I
b
oZNI.
SC1
AZ N2.
SC1
n
1
O MxN . SC2
o ZON .SC2
0.5
❑
%@a
0.2
L
0
I
I
I
10
20
30
Day
of
.-l
0.2
40
@
o
($
I
L
,
I
,
,
o
10
20
30
40
“u
60
I
J
60
counting
Day
-Fig. 2. Observed gross beta-counting
rates of radiochemically
pure phosphorus extracts from two seawater samples and a rainwater sample as a function of
time of counting; first day of counting is taken to be O
and is’7-l O d after sample collection, The background
of the beta counters is in the range of 0.2-0.3 cpm.
and then combusted at 4.00°-5000C. The reduced mass was then digested in HNO~; the
residue was repeatedly digested with HNO~.
Phosphorus was then extracted in 1:1
HNO~, and a molybdate precipitation was
carried out. Subsequent chemical procedures were the same as followed for seawater (Lal et al. 1960).
No phosphorus carrier was added in the
case of seawater or plankton samples. The
results of 32P and 33P determinations are
obtained directly in terms of their measured
disintegration rates, (g of P)-], “basedon the
amount of phosphorus deposited in the
counting planchet.
%qies
were counted for their beta activity. The maximal energy of 32P /3-radiation is 1.71 MeV while that of 33P is 0.249
MeV. We used rectangular Plexiglas lowIevel Q-gas flow counters with a 1 mg cm-2
window. The cross sections of the counters
used were 4.75 and 10 cm2; the total available areas for deposition of the sample were
7.3 and 17.7 cm2 respectively. Details of the
construction of the counters are given else-
of
counting
Fig. 3. Observed gross counting rates, as in Fig. ‘2,
for phosphorus extracted from ph yto- and zooplankton
samples. The filled, symbols in the upper diagram represent counting of the samples wit”h an absorber 10 mg
cm-2 thick placed between the counter window and the
sample. (Sample details given in Table 4.)
where (La] and Schink 1960). More than six.
counters were used during the work. The
small counters had backgrounds in the range
of 0.2–0.25 cpm and the larger had 0.3-0.4
cpm. The counters displayed high stability
in counting rates of background and cosmic
ray mu.-mesons over periods of >1 yr.
Results
Repeated measurements of seawater and
rainwater extracts showed good stability and
stable backgrounds over periods of 1-2
months (Figs. 2 and 3). The results for the
four surface seawater samples (3-7 m deep
off La Jolla and San Diego) yield 32Pactivity
values in the range 0.5–0.9 dpm m 3 (Table
3).
In this work, our emphasis was on determining 32Pactivities. The best procedure for
an accurate determination of the activity of
‘3P (Lal .et al. 1960) is to measure the activity of the sample deposits with and without an absorber (-10 mg cm-2 thick). In
thick deposits (> 10 mg cm-2) the relative
Cosmogonic
32p,
contributions of 33P are reduced appreciably due to absorption of its weak (3-radiation
(half-thickness of absorption -5 mg cm-’).
The half-thickness for 32P@-radiation is -80
mg cm–2. Our thickest sample deposit was
-30 mg cm-’, in the case of plankton samples. However, we did measure the 33P activities in one seawater (SW 1.SCP; Table 3)
and two plankton samples from Bedford
Basin collected on 20 August 1986. The results yield values in the range of 1.O1.5 ~0.35 for the dpm ratio of 33P and 32P
activities —not inconsistent with expectations (Table 2).
Discussion
The measured 32P activities in surface
waters (Table 3) agree well with the expected
model concentrations (Table 2), indicating
that levels of 32Pactivity in the surface waters
arc close to those expected for removal from
a - 50-m-thick mixed layer with an average
removal time of 30 d. For the case of no
removal, the expected concentration would
be 1.34 dpm m-3. In practice, however, since
mixing would not be complete on the time
scales of the half-lives of these isotopes, one
can expect a considerable gradient in concentrations within the mixed layer and much
higher values at the surface than the average
in Table 2.
A vertical profile of DIP from SCBS Sta.
205, 33”17.3’N, 118”1O.4’W (CalCOFI line
90; in the vicinity of Sta. 32) was studied
by Lal and Lee (1988). 32P activity decreased exponentially with a half-concentration depth of 11.5 m. The surface concentration, *2 dpm 32Pm-3 is 2-3 x higher
than in the surface ocean water samples off
San Diego (Table 3). The concentration gradient is expected to be caused by two factors: radioactive decay during downward
mixing and upward mixing of 32P-poor
waters through the thermocline. The 2–3 x
lower surface concentrations in waters off
San Diego are probably caused by a greater
vertical mixing in these waters and represent the average concentration of the mixed
layer.
The mean observed 32P activities in different plankton samples (Table 4) show an
order-of-magnitude
in range; the phytoplankton values are higher than those for
33p in oceans
1565
zooplankton samples. The observed values
for the Bedford Basin phytoplankton samples collected on 3 and 16 July 1986 translate to DIP concentrations of 0.9 and 1.2
dpm 32Pm-3 of seawater. These values compare well with those observed for surface
waters (Table 3). The plankton sample collected on 20 August 1986 is considerably
more active, however (Table 4); the DIP
value is calculated to be 3.8 dpm m-3 of
seawater. On examining the data on water
discharge from the Sackville River to Bedford Basin, we note that the discharge during 1 week before the date of sampling in
August was about 2.5 x that during the
whole of July. The river discharge on 19
and 20 August was about the same as that
during the 1 week in July; the average discharge during August was about 2 x that in
July (Environ. Can. 1986).
We attribute the high value in the plankton sample of 20 August 1986 to a large
influx of river water. The increase in 32Pspecific activity in the plankton sample is
in the range expected, considering the appreciably higher concentration
of 32P in
rainwater, about three orders of magnitude
higher than in surface waters. The discharge
during the period of interest, 1 week before
sampling, amounted to >0. 50/0of the total
volume of water in the basin (Sameoto
197 3). For comparison, the rain sample collected at La Jolla on 25 November
1985
(Fig. 2) had a concentration of 1.65 iO. 15
dpm 32Pliter-*, and the average concentration of 32Pin rains at Bombay, India, during
1958 was 1.46 dpm 32P liter-~ (Lal et al.
1960).
The chaetognath sample collected on 20
August 1986 from the Bedford Basin also
shows an enhancement in its 32P activity;
the increase is -2x
instead of -3x
as observed for the phytoplankton sample. This
is consistent with the phase lag in transfer
of 32P to carnivorous zooplankton via the
phytoplankton and herbivorous zooplankton. The mean specific activities of 32P in
two phytoplankton and zooplankton samples collected during July are 67.8 and 21.35
dpm 32P (g P)-’. Organisms higher up the
food chain would be expected to have progressively lower 32P-specific activities. On
the basis of a simple model of assimilation
:1566
Lal et al.
Riverine
Disso
Input
Deep Sea Input
>~
Inorganic
+——
Organ,c
Bacteria
~
~
/
Particulate
Organic
(dead)
<
?
‘ Phytoplankton
LJ
——~
—
LJ
‘{
v
-
Zooplankton
Nekton
—
Fecal pellets,
scales, molt
}+
v
.T9 .... ..................................Q9.P ..................... ......... ................9a
Fig. 4. A simplified description of the principal phosphorus pc)ols in the upper ocean and their interactions.
The model incorporates the microbial loop recently outlined by Azam (1986; see also Beers 1986).
of phosphorus by the zooplankton, the 3.2x
lower specific activity in zooplankton corresponds to a mean lag time of about 40 d,
consistent with estimates made by Eppley
et al. (1983).
Results to date show that the 32Pconcentrations in surface waters are 1–2 dpm nl-3.
The column inventory of 32P at SCBS Sta.
205 was found to be 4.1 x 10-3 dpm cm-2,
or 120 32P atoms cm-2 (Lal and Lee 1988).
This value is 640/o of the inventory estimated from atmospheric injection and in
situ production (Table 2). The deficiency in
the inventory of 32P must be explained
largely as due to losses from export of particulate organic matter (POM) from surface
waters. The standing inventory of 32P in
POM is estimated to contribute < 3?40of the
DIP inventory; that of 32P in dissolved organic matter (DOM) may constitute - 10%
of that in the DIP.
In Fig. 4, we show an oversimplified schematic description of the phosphorus pools
and fluxes in the mixed layer. To date we
have primarily studied 32Pand ‘3P activities
in DIP and POP; some attempts have been
made to measure 32P and 33P activities in
DOM (Lal and Lee 1988). With further developments, we should be able to study the
concentrations of the two radiotracers in all
the phosphorus pools. The specific activities
of 32P and 33P and their relative concentrations would allow us to determine the time
scales of exchange between or effux from
the various reservoirs (Fig. 4). These relationships are expected to critically decide
the net C, N, and P fluxes out of the mixed
layer. New production is studied by measuring photosynthesis due to nitrate flux in
the cuphotic zone. In a steady state situation, new production must equal the net
export of C, N, and P from the euphotic
zone. In our studies it would be manifest in
a number of different ways: departures from
expected secular equilibrium inventories of
S*P and 33P in the mixed layer, variations
in the concentrations of 32P and 3SPwithin
the mixed layer, and in the fractional inventories of 32Pand 33Pin the layers below
the mi:xed layer.
The scope of the proposed studies of the
twin radiotracers 32p and 33p in studies Of
Cosmogonic
32p, 33p
in
oceans
1567
KRISHNASWAMI,
S., AND OTHERS. 1972. Silicon, radium, thorium and lead in seawater: In situ extraction by synthetic fiber. Earth Planet. Sci. Lett.
16:84-90.
LAL, D. 1988. In situ-produced cosmogonic isotopes
in terrestrial rocks. Annu. Rev. Earth Planet. Sci.
16:355-388.
—,
J. R. ARNOLD,ANDB. L. K. SOMAYAJULU.1964.
A method for the extraction of trace elements from
seawater. Geochim Cosmochim. Acts 28: 11111117.
—,
ANDS. KRISHNASWAMI.1988. Isotope tracers
in the oceans: Old and new, in press. in Festschrift
for M. G. K. Menon. Tata Press, Bombay.
—,
AND T. LEE. 1988. Cosmogonic 32Pand 33P
used as tracers to study phosphorus recycling in
the upper ocean. Nature 333: 752–754.
—,
AND B. PETERS. 1967. Cosmic ray produced
radioactivity on the Earth, p. 55 1–6 12. In Encyclopedia of physics. V. 46/2. Springer.
—,
RAMA, AND P. K. ZUTSHI. 1960. Radioisotopes P32,Be’, and S35in the atmosphere. J. Geophys. RCS. 65: 669–674.
—,
ANDD. SCHINK. 1960. Low background thinwall flow counters for measuring the beta activity
of solids. Rev. Sci. Instr. 31: 395–398.
MERRILL,J. R., M. HONDA,ANDJ. R. ARNOLD. 1960.
Methods for separation and determination of beReferences
ryllium in sediments and natural waters. Anal.
AZAM,
F. 1986. Nutrient cycling and foodweb dyChem. 32:1420-1428.
PILSKALN,
C. H., ANDS. HONJO. 1987. The fecal pellet
namics in the Southern California Bight: The mifraction of biogeochemical particle fluxes to the
crobial foodweb, p. 274-288. In R. W. Eppley
[cd.], Plankton dynamics of the Southern Califordeep sea. Global Biogeochem. Cycles 1:31-48.
PLATT,T., M. LEWIS,ANDR. E. GEIDER. 1984. Thernia Bight. Springer.
modynamics of the pelagic ecosystem: Elementary
BEERS,J. R. 1986. Organisms and the foodweb, p.
closure conditions for biological production in the
84-175. In R. W. Eppley [cd.], Plankton dynamics
of the Southern California Bight. Springer.
open ocean, p. 49-84. In Flows of energy and materials in marine ecosystems: Theory and practice.
CEMBELLA,A. D., N. J. ANTIA, AND P. J. HARRISON.
1984. The utilization of inorganic and organic
NATO Conf. Ser. 4, Mar. Sci. V. 13. Plenum.
AND W. K. W. Lr. 1986.
Photosynthetic piphosphorus compounds as nutrients by eukaryotic
co~lankton. Can. Bull. Fish. Aquat. Sci. 214.
microalgae: A multidisciplinary perspective: Part
1. CRC Crit. Rev. Microbiol. 10:317-391.
SAMEOTO,D. D. 1973. Annual life cycle and proDAVIS,R., ANDO. A. SCHAEFFER.1955. Chlorine-36
duction of the chactognath Sagitta elegans in Bedin nature. Ann. N.Y. Acad. Sci. 62: 105-122.
ford Basin, Nova Scotia. J. Fish. Res. Bd. Can. 30:
DUGDALE,R. C., AND J. J. GOERING. 1967. Uptake
333-344.
of ncw and regenerated forms of nitrogen in priSILKER,W. B. 1972. Horizontal and vertical distrimary production. Limnol. Oceanogr. 12:196-206.
butions of radionuclides in the North Pacific Ocean.
ENVIRONMENT
CANADA. 1986. Annual meteorologiJ. Gcophys. RCS. 77:1061-1070.
cal summary. Halifax-Dartmouth, N .S. (ShearSOMAYAJULU,B. L. K., R. RANGARAJAN,D. LAL, R.
water “A”).
F. WEISS,AND H. CRAIG. 1987. GEOSECS AtEPPLEY,R. W., ANDB. J. PETERSON.1979. Particulate
lantic 32Siprofiles. Earth Planet. Sci. Lett. 85:329organic matter flux and planktonic new produc342.
tion in the deep ocean. Nature 282:677-680.
STRICKLAND,
J. D. H., AND T. R. PARSONS. 1968. A
—,
E. H. RENGER,ANDP. R. BETZER. 1983. The
practical handbook of seawater analysis. Fish. Res.
residence time of particulate organic carbon in the
Board Can. Bd. 167.
surface layer ofthc ocean. Deep-Sea Rcs. 30:311YOUNG, J. A., AND W. B. SILKER. 1974. The deter323.
mination of air–sea exchange and oceanic mixing
HAYWARD,T. L. 1987. The nutrient distribution and
rates usimz ‘Be during the Bomex Ex~eriment. J.
primary production in the central North Pacific.
Geophys. ‘Res. 79: 44~ 1-4489.
Deep-Sea Res. 34:1593-1627.
KNAUER,G. A., AND J. H. MARTIN. 1981. Primary
Submitted: 15 February 1988
production and carbon-nitrogen fluxes in the upAccepted: 26 May 1988
per 1,500 m of the northeast Pacific. Limnol.
Oceanogr. 26:97-108.
Revised: 18 August 1988
biological pathways of P fluxes would be
appreciably augmented by the inclusion of
studies of concentration of another cosrnogenic radionuclide, 7Be(half-life, 53.3 d),
in the same water. Analogous to 32P and
33P,7Beis produced in the atmosphere and
injected into the oceans by wet precipitation. The usefulness of 7Be for determining
eddy diffusivity has already been demonstrated(Silker 1972; Young and Silker 1974).
Beryllium is biologically inactive; most of
the inventory of 7Be in the oceans is in the
dissolved form (Silker 1972; Young and
Silker 1974). Considering these facts, it is
proposed (Lal and Krishnaswami 1988) that
simultaneous studies of 7Be would provide
information on two valuable parameters: the
tropospheric injection rates of the cosmogonic nuclides 32P and 33P, and the eddy
diffusivity in the upper ocean layers, averaged over a period of the order of 2 months.