Old Dominion University
ODU Digital Commons
CCPO Publications
Center for Coastal Physical Oceanography
1971
Nutrient-Density Relationships in the Western
North Atlantic Between Cape Lookout and
Bermuda
Unnsteinn Stefánsson
Larry P. Atkinson
Old Dominion University, [email protected]
Follow this and additional works at: http://digitalcommons.odu.edu/ccpo_pubs
Part of the Oceanography Commons
Repository Citation
Stefánsson, Unnsteinn and Atkinson, Larry P., "Nutrient-Density Relationships in the Western North Atlantic Between Cape Lookout
and Bermuda" (1971). CCPO Publications. Paper 178.
http://digitalcommons.odu.edu/ccpo_pubs/178
Original Publication Citation
Stánsson, U., & Atkinson, L. P. (1971). Nutrient-Density Relationships In The Western North Atlantic Between Cape Lookout And
Bermuda. Limnology and Oceanography, 16(1), 51-59. doi: 10.4319/lo.1971.16.1.0051
This Article is brought to you for free and open access by the Center for Coastal Physical Oceanography at ODU Digital Commons. It has been
accepted for inclusion in CCPO Publications by an authorized administrator of ODU Digital Commons. For more information, please contact
[email protected].
NUTRIENT-DENSITY RELATIONSHIPS IN THE WESTERN
NORTH ATLANTIC BETWEEN CAPE LOOKOUT
AND BERMUDA1
Unnsteinn Stefansson2 and Larry P. Atkinson 3
Duke University Marine Laboratory, Beaufort, North Carolina
ABSTRACT
A study of the nutrient-density relationships of Sargasso Sea water reveals a close
correlation; Gulf Stream waters arc anomalously higher and more variable in nutrients for
a given density. Nutrient anomalies can be used in the same way as oxygen anomalies to
map the ~istr~bution of ~aribbean water in sections across the Gulf Stream. They show
that contributions of Caribbean water to the Gulf Stream vary greatly from tinle to time.
Oxidative ratios of AO : dN : dP
-233) : 16.3 : 1.0 were derived from the correlation
between oxygen anomalies and nutrients anomalies of Gulf Stream water in the sigma-t
range 25.0-26.8.
=(
INTRODUCTION
In a variety of ocean areas, water masses
can easily be identified by their characteristic temperature-salinity relationships.
While this familiar technique works well
for distinguishing between Gulf Stream
and slope waters, it fails to differentiate
the waters of the Gulf Stream system from
those of the adjacent Sargasso Sea.
The only chemical methods so far used
with success to trace the distribution of
Gulf Stream water involve the correlation
of oxygen with temperature, salinity, or
density (Dietrich 1937; Rossby 1936; Richards and Redfield 1955). Of these methods
the oxygen-density relationship has proved
most advantageous. Richards and Redfield
( 1955) found a well-defined correlation
between the oxygen content and density
in the Sargasso Sea, while the waters of
the Caribbean Sea, Yucatan Channel, and
Strait of Florida were anomalously lower
in oxygen for a given density. The oxygen
deficiencies could be used to trace the
presence of these waters of southern origin
along the course of the Gulf Stream and
to delineate its right-hand boundary. The
oxygen anomalies have also been useful for
studying seasonal changes in the distribution of these waters along the continental
slope of North Carolina ( Stcfansson et al.
1971).
Since the principal nutrients, phosphates
and nitrates, are normally inversely related
to oxygen, the subsurface Gulf Stream waters would be expected to be anomalously
high in nutrients as compared to Sargasso
Sea water of the same density. To test this
supposition and to investigate the possibility of using nutrient anomalies for tracing
the distribution of Caribbean water in sections across the Gulf Stream, we have examined the relationships between nutiients
and density for Gulf Stream and Sargasso
Sea water. This communication describes
the results and some of their applications.
We wish to express our thanks to Drs.
P. K. Park, A. C. Redfield, and F. A.
Richards for kindly reading the manuscript
and offering helpful suggestions.
MATERIALS AND ANALYTICAL METHODS
1
This research was supported by National Science Foundation Grant GA-4278. Ship time on
the RV Eastward was provided by the Cooperative Program in Biological Oceanography.
2
Marine Research Institute, Reykjavik, Iceland,
and Duke University Marine Laboratory, Beaufort,
North Carolina.
8
Present address: Institute of Oceanography,
Dalhousie University, Halifax, Nova Scotia.
LIMNOLOGY AND OCEANOGRAPHY
Samples for dissolved constituents were
collected in the years 1966-1969 during 12
RV Eastward cruises from the region between the North Carolina continental shelf
and 69 ° 50' W ( Fig. 1).
Salinity was measured conductomctrically
with an inductive salinometer, temperature
51
JANUARY 1971, V. 16(1)
52
UNNSTEINN STEFANSSON AND "LAnRY P . ATKINSON
eo•
,
.
_..,,..,-=, -
/
i~/
35°
~=---6-4-~
I
30°
......
,,
r
,
·"?!
,,
I
35°
2 .J
r
30°
I
I
.t-
25°
25°
eo•
75°
70°
FIG. 1. Location of stations in study area. Sections I , 2, 5, 6, and 7 were occupied six times in
the period February 1966 to April 1967, section 3
twelve times in the period February 1966 to August 1969 and parts of section 4 were occupied in
August 1967, 1968, and 1969. The broken line
indicates the mean axis of the Gulf Stream as given
by von Arx et al. ( 1955).
density surfaces no longer sloped downward from west to east.
Within the Sargasso Sea both nitrntes
and phosphates are almost constant in the
density range a-1 = 24.0-26.0 ( Figs. 2 and
3). At higher densities the values increase
progressively, reaching a maximum [(N0 3-N] = 25 µ,g-atom/ liter, [P043--P] = 1.60 µgatom/liter) at <rt 27.30. In the range CTt
= 27.30- 27.90 the values d ecrease gradually with density. For densities between
CTt = 24.0 and 26.30 nutrient concentrations
remain within narrow limits, between <rt =
26.30 and 26.80 the concentration range for
a given density is somewhat greater, and
at higher densities there is a wide scatter
in the values, especially of samples from
the deepest water strata.
In contrast to the water of the Sargasso
Sea, concentrations of samples taken within
the Gulf Stream were found to b e highly
variable for a given density and anomalously high in the O'"t range 25.0-26.6 (Figs.
2 and 3). At densities exceeding CTt 26.80
the nutrient-density relationships do not
distinguish Gulf Stream water from .that
of the Sargasso Sea.
=
NUTIUENT ANOMALIES
with high-precision reversing thermometers, and oxygen was determined by
Winkler titration. The nutrient samples
were drawn into polyethylene bottles,
quick-frozen, and stored in a deep-freezer
until thawed and analyzed spcctrophotometrically ashore. Nitrate was determined
hy the method of Mullin and Riley ( 1955)
and phosphate according to Murphy and
Riley ( 1962) .
NUTRIBNT-DENSITY RELATIONSIJlPS OF
SARGASSO SEA AND GULF
STREAM WATERS
The phosphate-density and nitrate-density correlations for the Sargasso Sea were
based on 156 and 80 samples respectively,
collected between the right-hand boundary
of the Gulf Stream and 69° 50' W ( section
4, Fig. 1). The Sargasso Sea side of the
Gulf Stream was assumed to be where the
Phosphate and nitrate anomalies were
estimated by subtracting the nutrient concentration of water from the Sargasso Sea
for a given density, as determined from
the curves shown in Figs. 2 and 3, from
the .observed concentration of the sample.
H aving computed the anomalies for each
sample, we plotted their distributions on
sections across the Gulf Stream. Examples
of such distributions are given in Figs. 4
and 5. Oxygen anomalies determined as
described by Richards and Redfield ( 1955)
are shown for comparison.
It is evident from Figs. 4 and 5 that
a) the phosphate, nitrate, and oxygen anomalies show ahnost identiC'al features, oxygen
deficiencies corresponding to nuh·ient excesses as referred to Sargasso Sea water,
b ) numerically the anomalies reach a maximum at the core of the Gulf Stream, where
they arc centered around cr1 = 26.0, disappearing gradually as the right-hand bound-
53
NUT1UENT-DENS1TY IlELAT10NS1IU'S
25
• SARGASSO SEA WATER
1, GULF STREAM WATER
20
"E
E
0
I
en
15
.3z
I
f()
0
2
10
5
24.5
25.0
~-~ _J _ 27.0
26.5
-- ~ ~ ~ - ~ 27.5
26.0
FIG. 2. Correlation between nitrate concentrations and density (er,). Circles denote Sargasso Sea
observations made in section 4 in August 1967, triangles denote Gulf Stream observations made in section 3, February 1966-August 1967.
ary of the stream is approached, and c)
marked variations with time exist in the
magnitude and distribution of Caribbean
water in the section.
A comparison of the anomalies found
during all 12 cruises suggests seasonal variations in the distribution of Caribbean
water: Summer observations showed an
intrusion of this water to the bottom layers
over the shelf, while winter observations
indicated its absence in the shelf area and
near the edge of the continental shelf.
The apparently small contribution of Caribbean water observed in March 1967 represents a unique situation, probably arising
from intense vertical mixing of the surface
layers, as was indicated by almost constant
temperature, salinity, and density from the
surface down to 130 m in the region off
the shelf break ( Stcfansson et al. 1971).
The largest oxygen, nitrate, and phosphate anomalies observed during these
investigations were: -1.27 ml/liter (-113
µ,g-atom/liter), 7.9 µ,g-atom/liter, and 0.53
µg-atom/ liter respectively.
DISCUSSION
The application of oxygen anomalies to
study seasonal variations in the distribution of Caribbean water along the continental slope off North Carolina has been
described ( Stefa:nsson ct al. 1971). Nutrient anomalies could also be used for this
purpose and provide a check on the method
of oxygen anomalies. Although nutrients
arc less easily determined than oxygen,
their concentrations are less affected by
seasonal warming and cooling processes
than is that of oxygen. Consequently, the
nutrient anomaly distribution may provide
54
UNNSTEINN STEFANSSON AND LARRY P. ATKINSON
2.00----------
1.75
•
SARGASSO SEA .~ATER
t..
GULF STREAM WATER
I
150
1.25
~
;
......
E
~
a
, 1.00 ...
"'
:J..
c..
I
0¢
0..
0.75
t..
0.50
0.25
24.5
25.5
26.0
o,
26.5
27.0
27. 5
28.0
Frc. 3. Correlation between phosphate concentrations and density (e1,). Circles denote Sargasso Sea
observations made in section 4 in August 1967, 1968, and 1969, ·triangles denote Gulf Stream observations made in section 3, February 1966-August 1967 and August 1969.
a more suitable water mass tracer, especially in areas where the Gulf Stream is
near the continental slope and seasonal
temperature variations are greater than in
the open sea. With modern autoanalyzer
techniques such surveys could easily be carried out routinely on multistation cruises.
Assuming that the oxygen deficiencies
of Gulf Stream water result from oxidation of organic matter with accompanying
nutrient release, it follows that the correlation between nutrient anomalies and
oxygen anomalies provides a means to
evaluate the oxidative ratios. If it can also
be postulated that the concentrations of
preformed nutrients1 are the same in Sar<1. Following the definition by Redfield ( 1942),
preformed nutrient refers to that fraction of dissolved inorganic nutrient which was present in
the water at the time it sank from the surface.
Since preformed nutrients are entirely independent in origin of processes of organic decomposition and oxidation taking place after the water
left the surface, they are by definition consctvative properties of seawater.
55
NUTRIENT-DENSITY UELATIONSHIPS
ro
s
·- · ···- • + :
···-
1 -·
e
·- .
~---- ~~-.-~- ---,
17
18
19
20
21
22
0
0
~-7tj-;1175·
77•
200
24
······· 1
+
100
100
23
··-r ····· r
200
m -0.50 to - 0 .75 ml /Iller
300
R
4 00 ·
4 00
. . ·___ __]__ ~L--
·· ··- - --- ·
5 00
34
33
0
10
-.
;
0
•
< - 1.00 ml / liter
····-
500
9
- 0.75 to - 1.0 0 ml/liter
17
- ---- - ··
21
19
18
. ·1
22
24
23
100
I 00 -
N
N
200
200
300 -
300
4 00
400
D
••
'.
500
50 0
34
33
32
0
II
10
·-:-· .
100
9
! ...
8
-
2- 4 }lg-atom / liter
4-6 ~g-atom/li!er
17
.,18
/
I
- -- ----
0
/
> 6 pg-atom/liter
19
.. !
20
•
.. - ···--· .. _____ .!__+
·---- -------21
22
23
24
+
100
p
p
200
200
300
300
[3 0. 15 - 0.25)J<l·atom/ liter
~
~ 0 25-0.35 )J<l· otom/l ller
400
0
10
•
20km
> 0.35 )Jg -ol om/ l iter
L
50 0
.. -
-·-··
23-26 FEBRUARY 1966
500
26 - 27 NOVEMBER 1966
FIG. 4. Vertical distribution of oxygen anomalies (0), nitrate anomalies (N), and phosphate anomalies (P) in s<..>etion 3 in February and November 1966. The location of the section is shown on inset map.
Vertical scale shows depth ( m) . H orizontal scale shows station numbe1·.
gasso Sea and Gu1f Stream waters for a content varies seasonally, and reoxygenagiven density, this approach should be tion from the atmosphere causes a disapmore reliable than estimates based on the pearance of negative oxygen anomalies.
apparent oxygen utilization, which involve Furthermore, phytoplankton uptake leads
various errors and uncertainties ( e.g., Red- to reduction or even depletion of nutrients
in the photic zone. To eliminate these
field ct al. 1963; Stefansson 1968).
To investigate this application, a statis- surface effects, near-surface samples were
tical analysis was made of the correlation omitted and only values with negative
between the anomaly values of oxygen oxygen anomalies included in the regresand nutrients in section 3 for the O-t range sion analysis. The results ( Figs. 6 and 7)
24.0-26.8. In the surface layers the oxygen indicate a close correlation between nutri-
56
UNNSTElNN STEJi'ANSSON AND LABllY P. ATKINSON
5
0
6
7
8
m
9
10 11
12
..
·-:-· !""
• + ••
100
5
0
.
6
8
m
.
9
+
100
0
0
200
::: '.': ~
~~j~
~ 175°.
.
__
CAPE
LOO/COi.iT
•
400 -
04'
_ __ _ [ _..
I
2
3
4
a~--:-: -
5
!
- --
6
7
1
500
8
10 11
9
12
I: :- :_
:
~
100-
N
200
~
300
-
< - I.DO ml/liter
4
0
5
.
7
8
7
8
9
100 -
N
· 1
'
.
. )
~
', ,,11·
-----· - ----- - - --
5 6
! !
-
200 -
D
300
,';--- ..............
2-4 µg -olom/llter
I
~
'
,
1 • ......... - - .............
• t I
I I
\•I
10 11
9
40 0
· • > 6 µ g-olom/liter
. .
500 - -
'
, I
I
4-6 µg-otom / liter
/.
:::
+
-0 75 to -I.OD ml / liter
40 0 500
Em- 0.50 to -0.75 ml/liter
300
- - -- -· -·--·
1_2
0
!
--
I 00 -
10 0 -
p
.',{'
200
.
300
~-:
:
'
0
10
.
~ \;~·
r
40 0
.
2 0 ~m
L____J_____J
I
/
.
,
.
p
200 0.15-0 .25 µg-otom t liter
...--- -
1.
l
0.25- 0 35 µg-at om/ liter
400
•
-
·500
500
LJ
3 00
> 0.35 )lg - at om/liter
-----
15-18 MARCH 1967
-- ---- --
·
.....
',
' ',
.\
\
-
.
'-~
' '.
/
_,,/
- ----- ·--- - - - - ·
1-2 AUGUST 1967
FIG. 5. Vertical distribution of oxygen anomalies (0), nitrate anomalies (N), and phosphate anomalies (P) in ·section 3 in March and August 1967. The location of the section is shown on inset map.
Scales as in Fig. 4.
cnt and oxygen anomalies. The calculated
regression lines have the equations
N03 --N anomaly (µg-atom / litcr)
:-:: -0.46-6.23 0 2 anomaly (ml/liter), (1)
POl--P anomaly (µg-atom / liter)
= -0.03- 0.383 0 2 anomaly (ml/liter). (2)
The statistical data ( Table 1) show that
the nitrate to phosphate change derived
from the anomalies is practically the same
as that obtained from the regression of
nitrate concentrations to phosphate concentrations, using all observations from se·ctions 1-7 collected between February 1966
and March 1968. The oxygen to nutrients
ratios of change are in good agreement
with the findings of Richards and Vaccaro
( 1956) for the upper layers of the Cariaco
Trench, but significantly lower than expected from the stoichiometric model of
Richards ( 1965).
It remains to consider the validity of
57
NUTRIENT-DENSITY HELATIONSHIPS
a.or~~---··---,-~·~--~/ ~--
- · , · - - - - - - , - , - - ·-I
0.50-
/
p Om, c8
01
I
IO
7.0 -
/ o
0
I
I
I
IO
I
1;] 4.0 ..
::e:
0
I
~ 2.0
O
.f
I
z
O@
O
0
o
1.0.
80
O
0/
o.
s /
O I
01
I
O;
h, 0.30
p
0
I
I
0
o&o
I
0
- - -· I
I
o/
I
/
I
p
0/
I
01
I
0
I
na:nCW/
I
I
-1.0
0 0
8
0..
1
q./
cSJ 00;
/
<t
0
I
0
ooo
I
I
0
I
0
0..
I
I
I
I
cl
00
I
I
0
0
0
I
z
<(
0 O
0 O
o0
/
w
::::i
<(
::;; 0.20 -
I
@
0
0
I
I
10
I
(I)
I
I
0
~
I
01
0
I
Z
8
I
I
C
I
oO
Oo
I
I
0 0
0
/
3.0
<(
(
0
oo
I
<(
;::::
E
.g
I
00
I
...J
~
0
I
/0
I
~
I
0
80
CY
I
I
0
CO)
0 I
/6
O
;O
-;:- 0.40 ·
o
I
I
ioo
-
lo
I
0
··--- -L - ~- · - L · - · · - - - -
0.50
1.00
-b
I
0 2 -ANOMALIES (-ml/liter)
-· · · ~ - L · · - - - - ~
0.50
0
1.50
1.00
1.50
0 2 -ANOMALI ES (-ml/liter)
Fm. 6. Correlation between nitrate anomalies
and oxygen anomalies based on 93 observations
from section 3, 1966-1967. The broken lines indicate the 95% confidence limits.
Fie. 7. Correlation between phosphate anomalies and oxygen anomalies based on 109 observations from section 3, 1966-1967 and 1969. The
broken lines indicate the 95% confidence limits.
the postulate that preformed nutrients for
a given CTt value are the same in Sargasso
Sea and Gulf Stream waters. Preformed
nitrates estimated from the observed concentrations and tho apparent oxygen utilization [based on Carpenter's ( 1966) solu-
bility data], using a 6.N : ~O ratio of
-16 : 276, increase with decreasing temperatures, as shown in Fig. 8. A similar relationship applies to preformed phosphates.
This increase with decreasing temperatures
can be explained by the fact that the deep
T AnLE 1.
Data
Phosphate anomalies vs.
oxygen anomalies
Nitrate anomalies vs.
oxygen anomalies
Nitrates vs. phosphates, all
obRervations sections 1-7
No.
observ.
Oxygen-nutrient relationships
ll.O
109
-233 ± 13*
92
-233 ± 12•
Ll.N
t,.p
1.0
16.3
16.22 ± 0.10*
1.0
Cariaco Trench, upper layers
-235
15
1.0
Stoichiometric model
-276
16
1.0
1,279
Correlation
coeff.
Reference
+0.86
This work
+0.89
This work
+0.98
This work
" Standard error of estimate of the slopes t.O: t.P, Ll.O: l'IN X 16.3, and Ll.N: l'IP rcspccti.vely.
Richards and
Vaccaro ( 1956)
Richards (1965)
58
UNNSTEINN STEl:<'.ANSSON AND LARRY P. ATKINSON
~o
•c
25 ·
20
15 •
10
~····~- · - JULF STAEAM WATER
..,.,w "' .....
1.0 -
5 .
O·
0 ~·-·
-3.0 1 0
____ J_____ ,--,··· ---1 ------ ---··
5
10
15
T'i::MPERATURE (aC)
F1c. 8. Preformed nitrate-temperatnre relationship base<l on data from sections 1-7.
water is formed in winter in high latitudes
where light conditions limit growth. Consequently on sinking, the water contains
the nutrients in the inorganic form in relatively large concentrations ( cf. Redfield ct
al. 1963). It should be remarked, however,
that the preformed nutrient values found
for temperatures above 20C must be unrealistic, since most of them arc negative.
Even though the oxidative ratio were half
of that used, "negative" values of preformed nutrients would still be obtained.
This suggests that the estimates of the
apparent oxygen utilization cannot be valid.
The discrepancy, which cannot be explained by analytical e1Tors, might be
partly attributable to undersaturation of
the water at the time it sank from the surface, and partly to the atmospheric pressure being then lower than normal. In
spite of these errors, it seems almost certain
that the assumption holds that the "true"
preformed nutrients increase with decreasing temperatures as indicated by Fig. 8.
Plots of in situ temperature agaimt salinity for all observations outside the 200-
34.5
s•/u
·-·
i
J
J
I __ _
35.0
35.5
3 6.0
~6. 5
37,0
Fm. 9. Temperature-salinity relationships of
Gulf Stream (open circles) for stations with depths
>200 m and Sargasso Sea (triangles) waters based
on data from sections 1-7.
m isobath in sections 1-7 show a wide
scatter in the values in the C7't range 2.5.526.5 (Fig. 9). While a considerable overlapping exists, a great number of the Gulf
Stream values tend to have a lower temperature ( and a lower salinity) for a given
<Tt than Sargasso Sea water. This would
lead to slightly higher preformed nutrients
of Gulf Stream water in this <Tt range and
could possibly explain why the oxidative
ratios found were somewhat lower than
expected from the stoichiometric model.
Further studies of the nutrient-density
correlations in the western North Atlantic
are needed, not only to reexamine and
improve the relationships here presented,
but also to investigate to what extent the
distribution of nutrient anomalies can be
applied to trace waters of southern origin
in various parts of the Gulf Stream system.
REFERENCES
J. II. 1966. New m easurements of
oxygen solubility in pure and natural water.
Limnol. Occanogr. 11: 264-277.
l)rnTmCH, C. 1937. -Ober Bewcgung und H er-
CA11PENnm,
NUTRIENT-DENSITY RELATIONSIIIPS
kunft des Golf-stromwassers. Veroeff. Inst.
Meeresk. Univ. Berlin Ser. A 33(2): 53-91.
MuLLIN, J. B., AND J. P. RILEY. 1955. The
colorimetric determination of nitrate in natural water with particular reference to sea
water. Anal. Chim. Acta 12: 464-480.
Mull'HY, J., AND J. P. RILEY. 1962. Modified
single solution method for the determination
of phosphate in natural waters. Anal. Chim.
Acta 27: 31-36.
REDFIELD, A. C. 1942. The processes determining the concentration of oxygen, phosphate
and other organic derivatives within the
depths of the Atlantic Ocean. Pap. Phys.
Oceanogr. Meteorol. 9(2): 1-22.
- - - , B. H. KETCHUM, AND F. A. RICHARDS.
1963. The influence of organisms on the
composition of sea water, p. 26-77. In M.
N. Hill [ed.J, The sea, v. 2. Intersciencc.
fucnAnns, F. A. 1965. Anoxic basins and
fjords, p. 611-645. In J. P. Riley and G.
Skirrow [eds.], Chemical oceanography, v. 1.
Academic.
59
- - - , AND A. C. REDFIELD. 1955. Oxygendensity relationships in the western North
Atlantic. Deep-Sea Res. 2: 182-199.
- - - , AND R. F. VACCARO. 1956. The Cariaco
Trench, an anaerobic basin in the Caribbean
Sea. Deep-Sea Res. 3: 214-228.
RossnY, C. G. 1936. Dynamics of steady ocean
currents in the light of experimental fluid
mechanics. Pap. Phys. Oceanogr. Meteorol.
5(1): 1-43.
STEFANSSON, U. 1968. Dissolved nutrients, oxygen and water masses in the Northern lrminger Sea. Deep-Sea Res. 15: 541-575.
- - - , L. P. ATKINSON, AND D. F. BtrMl'US.
1971. Seasonal studies of hydrographie properties and circulation of the North Carolina
shelf and slope waters. Deep-Sea Res. 18
(2): In press.
voN Anx, W. S., D. F. BuMPus, AND W. S. fucnAnosoN. 1955. On the fine-structure of the
Gulf Stream front. Deep-Sea Res. 3: 46-65.