JOURNAL OF GEOPHYSICAL
RESEARCH, VOL. 99, NO. C7, PAGES 14,255-14,268, JULY 15, 1994
New production in the warm waters of the tropical
Pacific
Ocean
M. AngelicaPefia,MarlonR. Lewis,and JohnJ. Cullen1
Department of Oceanography,Dalhousie University, Halifax, Nova Scotia, Canada
Abstract.
Pacific
The averagedepth-integratedrate of new production in the tropical
Ocean was estimated
from a calculation
of horizontal
and vertical
nitrate
balance over the region enclosedby the climatological26øC isotherm. The net
turbulent flux of nitrate into the regionwas computedin terms of the climatological
net surfaceheat flux and the nitrate-temperature relationship at the base of the
26øC isotherm. The net advectivetransport of nitrate into the region was estimated
using the mean nitrate distribution obtained from the analysisof historical data
and previousresults of a generalcirculationmodel of the tropical Pacific. The rate
of new production resulting from vertical turbulent fluxes of nitrate was found to
be similarin magnitudeto that due to advectivetransport. Most (about 75%) of
the advective input of nitrate was due to the horizontal transport of nutrient-rich
water from the easternequatorialregionrather than from equatoria•upwelling. An
averagerate of new productionof 14.5-16g C rn-2 yr-• wasfoundfor the warm
waters of the tropica• Pacific region. These values are in good agreement with
previousestimatesfor this region and are almost five times lessthan is estimated
for the eastern equatorial Pacific, where most of the nutrient upwellingoccurs.
1. Introduction
eastern equatorial Pacific, little is known about the contribution to global new production by the warm pools
The equatorial Pacific region is among the most pro- of the westernequatorial Pacific. Sincein this regionniductiveoceanicregimes[Betzeret al., 1984;Chavezand trate upwelling is lessintense, turbulent vertical transBarber, 1987] and, becauseof the upwellingof deep port of nitrate should be a relatively more important
nutrient-rich water to the euphotic •.onethroughout the physical input.
New productionis definedas the productionof phytoyear, may contribute a significantfraction of the global
new production. In the eastern and central equatorial plankton resultingfrom nitrogensuppliedfrom outside
Pacific, the effect of upwelling is evident as a tongue the euphotic•.one[Dugdaleand Goering,1967],whereas
of cold nutrient-richsurfacewater [W•trtki, 1981]. In regeneratedproduction is that productionassociated
contrast, the thermocline and nutricline are deeper in with nitrogenregeneratedwithin the euphotic•.one.In
the western equatorial Pacific than in the eastern and the open ocean,the main external sourceof nutrientsto
centralregions[Craig et al., 1982],so upwellingbrings the euphotic •.oneis the upwellingand turbulent transmostly warm, nutrient-poor water from above the ther- port of nutrientsfrom deepwater. New productioncan
mocline(nutricline)into the uppersurfacelayer. It has be measuredby a number of methods suchas sediment
beenestimated[Chavezand Barber, 1987]that the ni- traps at the bottom of the euphotic•.one,the uptake of
nitrate, and the net flux of nitrate into the
trate upwelled eastward of the dateline in the equato- 15N-labeled
rim Pacificcouldsupporta highproportion(18-56%)of euphotic •.one. New production also givesa measureof
the global new production. However,this representsan the primary productionwhichcan be removedfrom the
upper bound estimate for that region, sinceonly a frac- surfacelayer of the ocean without destroyingthe long[Epple•landPetion of the nitrate upwelled is used by phytoplankton term integrityof the pelagicecosystem
terson,
1979;
Platt
et
al.,
1989].
As
such,
ratesof new
locally. In fact, a significantamount of the nitrate is
production,
when
averaged
over
timescales
longerthan
transportedawayfrom the region[Care, 1991;Fiedler
thebiologically
mediated
transport
of
et al., 1991]. In contrastto what is known about the a year,represent
carbon from the ocean surface layer to the ocean interior. This is not net transport, however. Severalstudies
have shown considerablespatial and temporal variabilAlso at Bigelow Laboratory for Ocean Sciences,McKown ity in the rates of new production even in open ocean
Point• West Boothbay Harbor, Maine.
regions. Therefore the way measurementsare averaged
Copyright 1994 by the American GeophysicalUnion.
hasa profoundeffecton the results[Platt andHarrison,
1985].
The abilities to evaluate the importance of the tropi-
Paper number 94JC00603.
0148-0227/94/94JC-00603505.00
cal Pacific region to global new productionand to pre14,255
14,256
PE•A ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
dict anomalousconditionshave been constrainedby a
lack of synoptic observationsin this large and remote
area of the ocean. Attempts have been made to estimate new productionin coastalupwelling[Duõdale
et al., 1989]and frontal regions[$ath•tendranath
et al.,
1991]from satelliteobservations
of seasurfacetemper-
bicmeter),and$ is theirradiance
vector(J m-2 s-1).
ature, the observed correlation between nitrate concentration and temperature, and empirical relationships
between nitrate concentrationand new production. In
the North Atlantic, a lower limit to new productionhas
a fixed volume of water which is bounded by a threedimensionalsurfaceof a constantannum mean temper-
Overbarsindicate mean values,and primesindicate departures from mean conditions.
A simple model of the heat flux in the warm-water
pool of tropical regions has been obtained by Nillet
and Stevenson
[1982]by integrating(1) and (2) over
ature (AT) at depth,the oceanseasurface(As), and
the insulating continental land masses. By virtue of
been estimatedfrom changesin phytoplanktonbiomass the constant-temperatureboundary condition, the addetermined from the coastal zone color scanner data and
vective term vanishes, because water that enters the
variationsin the depth of the nutricline[Campbelland volume is on averageof the same temperature as water
Aarup, 1992]. In oceanicregions,asin the westerntrop- which leavesit, and hence there is no net flux of heat
ical Pacific, where surface nitrate concentrationsare by mean advection. Thereforethe steadystate solution
generally undetectableby conventionaltechniquesand is a balance between the net surface heat flux into the
whereseasonalvariationsin phytoplanktonbiomassare volume and the turbulent heat transport out of the vol-
small[Dandonneau,1992],the aforementioned
methods ume.The turbulenttransport
•:•at removes
heatfrom
the pool can be causedby a varietyof processes
operating on different time and spacescalesfrom seasonal
sentedby Lewis[1992]with resultsfrom a generalcir- to microstructure. For instance,this includeseddy heat
culation model of the tropical Pacific[Philanderet al., flux due to time-dependent ocean motions on scalesof
1987] and historical temperature and nitrate data to tens to hundreds of kilometers and turbulent diffusion
for estimating new production are not applicable.
Here, we combine the new production model pre-
estimate the annum mean rate of new production in
the warm waters of the tropical Pacific region. The
analysis is constrained by the horizontal and vertical
nitrate balance and employs the surface heat flux, a
quantity amenableto remote sensing[Liu, 1988;
and Gautier, 1990], to estimatenitrate fluxes. While
other estimations of new production based on nitrate
balance[Chavezand Barber, 1987;Fledlet et al., 1991]
in this region consideronly the input of nitrate due to
upwelling, we evaluate the transport of nitrate due to
turbulent fluxes, which are difficult to estimateat large
scales as well.
operating on scalesof a few meters to a few millimeters. In the tropical Pacificwarm-waterpool, Nillet and
Stevenson
[1982]foundthat mostof the heattransport
is due to vertical turbulent transport, since hori•.ontal
turbulent fluxes are relatively very small. Therefore
/ /As
QO
dA
- cp
p/ /AT
W'T'
ds (3)
whereQ0 is the net surfaceheatflux (wattspersquared
meter), whichis integratedover the surfacearea As;
WtT • is the mean annum vertical turbulent flux of heat,
whichis integrated overthe subsurfaceisothermof area
AT, and dA and ds refer to the surface and subsur-
2. Approach
face elements,respectively.Equation(3) statesthat
the net surface heat flux into a volume of ocean which
In the tropics, large warm poolsof near-surfacewater is bounded by coast and at depth by a constantmean
are enclosedat the marginsby continentalland masses annum temperature surface is transported out of the
and at the surfaceand depth by constantannummean volume by turbulent vertical mixing only. A term to
temperatureisopleths[Niiler and Stevenson,
1982]. In accountfor the penetratingirradiancethroughthe suba steady state ocean, heat is conservedand new prosurface[Lewiset al., 1990]is small,giventhe depthof
duction is balanced by the net input of essentiMnuthe chosenisotherm(26øC).
trients into the euphotic •.one. To estimate the annum
meannew productionoverthe warmwaters(temperature, •26øC) of the tropicalPacificOcean,webaseour
2.2 Fluxes
of Nitrate
Similarly, if the mean concentrationof nitrate in the
upper layer of the ocean is constant, the net physical
input of nitrate into this layer must be equivalent to
2.1 Fluxes
of Heat
the nitrate consumption,which basicallyrepresentsnew
In the upper layer of the ocean, the heat and mass production. Then, the nitrate conservationequationin
the upper layer of the oceanis
conservationequation may be written as
arguments on the closerelationshipbetweenthe fluxes
of heat and nutrients in this region.
+ v.
V. UN q- V. U'N' - P.
- v. $
(4)
whereN is the nitrateconcentration
(retoolN m-3),
andP, is the localnewproduction
rate (retoolN m-3
whereT is temperature(Celsius),U is thevelocity(me- s-X). We canproceed
by integrating
(4) and (2) over
ters per second),cpis the specificheat of seawater(J the samesurface(As)and subsurface
(A•) boundaries
kg-x øC-X), p is the waterdensity(kilograms
per cu- as above. In this case, since the nitrate concentrationis
v.v-o
(2)
PE•IA ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
14,257
not constantalong the isothermalsubsurfaceboundary
AT, a net advective input of nitrate can occur if water
enteringthe volumehason averagea higherconcentration of nitrate than waterleavingthe volume.Therefore
new productionover the wholevolumeenclosedby the
were reformatted, checkedfor grosserrors, and merged.
Duplicate data, detected by having identical date, location, depth, temperature, and nitrate concentration,
surface(As) and subsurface
(AT) boundariesand the
used here were collected mainly between 1967 and 1972
continental landm•ses is equivalent to the net advective and turbulent input of nitrate,
to increase the probability that similar analytical meth-
wereeliminated.Most of the stations(about80%)were
obtained
from
the NODC
data
set.
The
NODC
data
(Figure2). We haveexcludeddata collectedbefore1952
ods were used in the determination
of nitrate
concen-
tration. Data were available for all months of the year.
On the basis of the distribution of temperature and
nitrate concentration, the 26øC isotherm was chosenas
whereP•v(=j•T p, dz, where
z• is thedepthof the the subsurfaceboundary AT becauseit enclosesmost of
subsurface
isothermalboundary)represents
the depth- the tropical Pacific region and becausenitrate is genintegrated
newproduction
(retoolN m-2 s-i), andUN erally detectable along the bottom of it. About 3300
and U•N • are the annual mean net fluxes of nitrate due
stations were located inside the region enclosedat the
to advective and turbulent transports, respectively.
surfaceby the climatological 260C isotherm. At each of
these
stations the depth of the 26øC isotherm and the
2.3 Nitrate-Temperature
Relationship
nitrate concentrationat that depth weredetermined. In
In the ocean, temperature and nitrate are well cor- the region between 50 and 12øSlatitude and 1250 and
related[e.g.,Kamykowskiand Zentara,1986].We can 140øE longitude, a few stations had surface temperaexploit this empiricalobservationto combine(3) and tures cooler than 26øC, and they were excluded from
(5), neglectingfor now-horizontalturbulentfluxesof the analysis. Standard deviation statistics for the obnitrate, and the spatial covariancebetween heat flux served temperature, isotherm depth, and nitrate data
across the isothermal boundary and the slope of the were calculated for 50 by 50 squaresof latitude and lonnitrate-temperature relationshipin the region to yield
gitude. Since discretevertical profileswere usedin this
analysis, the slope of the nitrate-temperature relationship at the base of the 260C isotherm was operationally
defined as that obtained using temperatures between
$
$
ds+dN
d-•cp-•
//.aPNdA--//.aT
UN
I //.aQødA
(6)
200 and 27øC. Given the vertical
Here, the annual mean rate of new production is
due to advection plus the turbulent input of nitrate
expressedas the product of the net surface heat flux
and the average slope of the nitrate-temperature relationship at the base of the subsurfaceboundary AT.
Note that as before, the turbulent transport refers to a
resolution
of the data
set, it was not possibleto use a narrower range of temperature. To estimate the mean slope of the region,
the surface area enclosedby the 26øC isotherm was divided into 178 boxesof 50 of latitude and longitude. In
each of them, the slope was obtained from a linear re-
gressionof temperature(independentvariable)versus
variety of spaceand time scalesof motions(seasonal
nitrate concentrationof the pooled data. Only those 50
to microstructure)and thus is not equivalentto the squarescontainingthree or more observationswereused
vertical turbulent
diffusion of nitrate alone.
The net
in the analysis. The averageslopefor the entire region
surface heat flux can be evaluated from climatology
or from satellite observations;the nitrate-temperature
relationship can be taken from historical data. Similarly, the advective transport of nitrate can be obtained
from the mean water transport normal to the subsurface
boundary AT and the annum mean nitrate concentration along this boundary.
and its 95% confidence intervals were calculated from
the slope values of all the boxes for which data were
available(164 boxes).Second-and third-orderpolynomial regressionsbetween nitrate and temperature were
tested and rejected on the basis of a small increment in
the correlation
volume of water
3.
Data
Approximately 4870 vertical profileswith simultaneous observationsof nitrate concentrationand temperature from the tropical Pacific between 20øN and 20øS
of latitude and 120øEand 80øW of longitudehave been
usedin this analysis(Figure 1). This data set wasobtained from three sources:(1) the National OceanographicData Center(NODC), (2) the oceanatlas
borneet al., 1992],and (3) work by Fiedlerand colleaguesin the easternequatorialPacificbetweenAugust
and December1986to 1988[Fietileret al., 1991;P.C.
Fiedler, personalcommunication,
1993]. Thesedata
coefficient
over the linear model.
The advective transport of nitrate into and out of the
which is warmer
than 260 C was calcu-
lated as the product of the mean nitrate concentration
at the subsurfaceboundary and the mean water transport normal to the boundary. To estimate horizontal
nitrate transport at the lateral boundaries,we consider
that the warm-water regionwas boundedon the sideby
an isothermal well-mixed layer of 50-m depth. The concentration of nitrate at the 260C boundary wasobtained
from the contouredmap of nitrate which was basedon
the pooled data. The mean water transport acrossthe
part of the boundary where nitrate was detectable was
calculated by spatial integration of the mean circulation valuesreportedby Philanderet al. [1987]usingthe
trapezoidal rule. They obtained these values from the
14,258
PEI•IAET AL.: NEW PRODUCTIONIN THE TROPICALPACIFICOCEAN
120 ø E
140 ø
20 ø N ' -
'
b' •'-' ,;-
i••-:•.....
i
; ';
ß
ß
ß
v
ß ......
L•-:..
'
ß
ß.
. .:
5o ß ß...............
0'
&......
ooß .,
5o •'j•
-e•
180 ø
' '
i
ß
,
'
-.
. .
..••i•.-.
•
ß•
vv v
i
i
i
•
ß
•ß .
"
• =
e•
.
•
. ''., '
v
' ....
•
v
ß
.....
. . •• .........
:..?
i
120 ø
•
•
,
j_
'
-•t•
oo, i
100 ø
•
...... •.....
• :- o ....•:•:•:•::
80 ø W
•
•;••-'-••••J•
- - '":::
:::..
••
.
i
'
v
v
..•,,,•--
,
i
o
ø.
6 _
•o • •o
•
v
ß
! _
•••]
/
,•o• • -•o•o•
..'.i
I
.
ß
ß
•
ß .
•• •t
ß
.
I
-•ß . •
. . .;'
ß•
•,.-'?
ß .•v •.•
I ee ß......ß
v
t
•
,.
-'le
.: ....:: ß ß '
i
i
i
i
!
...,
,... ,.., :...
•5o J ••:<'•:•,.•
•••••'•,.•'•..:
ß , II•
ß ß,
ß
1 •"'-'-'"•••-••••• . :•':
.',ß I ' "0, .....
ß
ß
V :'. v•:'
ß ßßßßß•
i
,
ß . 0..
-o. oo .•
ß v
.-.:• •.•.,.• .•.........
I.
n--
•
' ß' 0- .
..... , ,....[
....
..........
ß•
140 ø
•, _
;;
'
-•:•:.•••••,:•::•.'.•.'' ..,I
• • •
ß•
ß
I
•
.
••/'•':'•:••:•.••
•0o••--:•,•
160 ø
i
ß ve
:..:
, •.
I
.- " " --.-',•.
.., ;
........
,-.,•- - .-- .....
•
10ø
160 ø
- ,
•o
• - •P-'-• -:••:•:•L
• ? T • L ' '-'-•••'••F
ß
ev v ß
•...,ß •
..-•
I
t •oL' •ß f•o'•
-• ••-•
., ßß . •.{.• ': ?'
.'• .•ß •:
•
'
.
, <
i
I ß ß _ I
L
I
ß I
I
Figure 1. Locationsof the stationsusedin this analysis.Solidcircles,National Oceanographic
Data Center(NODC; 3845stations);triangles,oceanatlas(454 stations);opencircles,Fledlet
and co]leagues
(574 stations).
third year of simulationof a generalcirculationmodel of
the tropical Pacific which was forced with monthly averagedclimatologicalwinds. The surfacearea enclosed
by the 26øC isotherm and the mean annual net surface heat flux integrated over this area were calculated
usingthree differentclimatologicaldata sets: Weareclimatology,which has a 50 by 50 latitude-longitudegrid
[ Weare et al., 1981]; Esbensenand Kushni•climatol-
6OO-
400-
between
200-
Distribution
10 øN and 15ø S. The standard
deviations
of sur-
the coast. A distribution similar to that of temperature
wasfoundfor the surfacenitrate concentrations
(Plate
2a), wherethe highestnitrate concentrations
wereas-
Years
sociated with the coldest surfacewaters. However, the
nitrate-enriched zone was almost symmetricalabout the
equator,whereasthe cold-watertonguewaslesssharply
definedto the south. Along the equator, surfacenitrate
600
o
concentrations were >4 /•M eastward of 160øW. Waters with > 1/•M nitrate extended to about 170øElongitudinally and beyond5øN and 5øSlatitudinally. Outsidethis region,nitrate concentrationwasgenerallylow
5oo
03 400
o 300
(<0.5/•M) or undetectable.The standarddeviations
of nitrate concentration
(Plate 2b) werehigherin the
..• 200
eastern side between the equator and 10øS than else-
100
where.
z
In the tropical Pacific, water warmer than 26øC was
found beyond 20øN and 20øS westward of 150øW and
o
Months
Figure 3.
NODC
and Nitrate
facetemperature(Plate lb) showedhighervaluesnear
o
•
4.1 Temperature
main featurespresentin the annualmean climatologies.
In the eastern equatorial Pacific, surfacetemperatures
were relatively cool and the gradientswere large, indicating upwellingat the equatorialdivergenceas well
as advection from the Peru upwelling region. Farther
west, a large pool of water of more than 28øC wasfound
-
m 800-
E
Results
data in the tropicalPacific(Plate la) reproduces
the
o
:•
ß
4.
A contour plot of all the sea surface temperature
•1200
:.•1000
ogyon a 40 latitude by 50 longitudegrid [Esber•ser•
and.
Kushr•ir,1981];and Oberhuberclimatologyon a 2ø by
20 latitude-longitudegrid [Oberhuber,
1988].
data.
Yearly and monthly distribution of the
between30 and 15øN toward the east, coincidingwith
the North Equatorial Countercurrent.In this region,nitrate was detectableonly in the southeasternboundary.
Vertically,the 26øCisothermsubsurface
boundarywas
locatedmainlyat depthsof 80 to 120m (Figure3a)
PEi•TAET AL ß NEW PRODUCTION IN TIlE TROPICAL PACIFIC OCEAN
• 20 ø E
140':'
I
1 •;,0 o
I
I
180 o
i
!
160 ':'
i
I
14 0':'
i
I
12 0 o
I
i
14,259
80 o W
10 0 ø
!
!
i
15 ø
10 ø
•
5 ø
o
10 <,
15 ø
20 ø S
"21
2'2
120 ø E
20ON
25
140':'
I
,
I
24
160':'
I
!
2õ
2(;
1RO':'
!
I
27
1 •;;0ø
I
I
28
14 0':'
I
I
'-29
120 ø
I
I
30
100':'
I
80':' W
I
•
15 ø
10
ø •
5 ø
15'o
20 ø S
0 0
0 ._5
1 0
1 õ
2.0
Plate 1. (a) Surfacedistributionof temperature(degrees
Celsius)in the tropicalPacificregion
and (b) its 5ø-square
standarddeviation.The locationof the 26øCisothermis shownby a white
line o
in the western tropical Pacific. Toward the east and
away from the equator, the depth of the 26øC isotherm
of the equator, between 145ø and 150øE, a marked gra-
shoaled.
isotherm(Figure4a) wasnot constantbut followeda
reflected in the depth of the isotherm. An examination of the data available for this region showedthat
pattern similar to that of the isotherm depth. Higher
the nutricline
The
concentration
of nitrate
at
the
26øC
dient
of nitrate
concentrations
was sometimes
was found
shallower
that
than
was not
the 26øC
nitrate concentrations
(>5 #M) were found near the isotherm.The variabilities
in the depth(Figure3b)and
equator, where the isotherm was deeper. Beyond 15øN in the nitrate concentration(Figure 4b) at the 26øC
and 15øS, nitrate concentrations were usually unde-
tectablein waters of 26øC. This distribution agreeswith
previousobservationsalong the west coast of North and
isotherm were higher on the eastern and western sides
than centrally.
SouthAmerica[Ze•tara andKam•tkowski,
1977],where 4.2 Nitrate-Temperature
it has beenfound that a givennutrient concentrationoccurs at colder temperatures as latitude increases.North
Relationship
A scatter diagram of temperature versusnitrate concentration in the upper 250 m of the surface area en-
14,260
PEi•TAET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
12,:)ø E
140 ø
20øN,
160 ø
I
I
18(I ø
I
I
160 ø
I
I
140 ø
I
I
120 ø
I
I
80 ø W
100 ø
I
I
I
15 ø
10 ø
5 o
o
o
10 ø
15 ø
20 ø S
;';
120 ø E
20'--' N
1
140 ø
3
160"'
I
'
o
I
4
180 ø
I
I
r--,
160 o
I
I
6
14 0 ø
I
I
7
12 0 ø
I
I
80':' W
10 0 ø
I
I
I
I
I
15 ø
10 ø
o
o
5 o
10 ø
(b)
15 ø
20 o $.
(, ,')
,-'_,,
5
1 0
1 5
"2 0
'.--'6
3 0
3 5
Plate 2. (a) Surfacedistributionof nitrate concentration
(micromolar)in the tropicalPacific
regionand (b) its 5ø-squarestandarddeviation.The locationof the 26øCisothermis shownby
a white
lineo
nitrate were poorly correlated, indicating that surface
temperature is not a good predictor of surface nitrate
(1) a regionwheretemperatureislow(generally
<15øC) concentration in this region. Studies of the relationship between temperature and nitrate concentrationin
but nitrate concentration
hasa highrange(15-36
closed by the 26øC boundary is shown in Figure 5.
Three different regionscan be distinguishedin this plot:
of values,(2) an intermediatezone wherethe gradient of temperatureis high (15-26øC)and changesin
temperatureand nitrate are well correlated,and (3)
a warm (>26øC) upper layer wherenitrate concentra-
the upperlayerof the ocean[Karn•tkowski,
1987;Minas
and Minas, 1992]haveshownthat seasonal
andspatial
variations in the slope are related to the rate of nitrate
uptake by phytoplankton. Below this upper layer, the
tion is undetectable or, when it is present, the slope value depends on the balance between bacterial denitriof the temperature-nitrate relationship is lower than fication and nitrate regenerationfrom decomposing
orthat of the intermediate region. When only surface ganic matter as well as on physicaltransport processes
data wereconsidered
(Figure5 inset),temperatureand [Zentaraand Karngtkowski,
1977].
PE•IA ET AL.' NEW PRODUCTIONIN THE TROPICALPACIFIC OCEAN
120 ø E
140 ø
160 ø
180 ø
160 ø
140 ø
20 ø N
•
120 ø
14,261
80 ø W
1O0 ø
'
15 ø
10 ø
o
Oo
o
10 ø
,so
(a)
20 ø S
•T•3'•
20
120 ø E
40
140 ø
60
160 ø
80
180 ø
100
160 ø
20 ø N
120
140 ø
•'
140
120 ø
1O0 ø
80 ø W
•
15 ø
10 ø
o
Oø
o
10 ø
,so
(b)
20 ø S
•'I'•T•
0
5
10
15
20
25
30
35
Figure a. (a) Thedepthofthe260C isotherm
subsurface
boundary
(meters)and(b) its 5ø-square
standard
deviation.
sistent with the values estimated here. The average
the baseof the 26øCisotherm(Figure6a) showed
lon- slope of the nitrate-temperature relationship was-0.89
intervals,
-1.00to-0.78
gitudinal as well as latitudinal variations. Along the /•M øC-1 (95%confidence
The slope of the nitrate-temperature relationshipat
tropicalPacific,the slopesin the easternsector((-1.4
0C- 1) in the regionenclosed
by the 260C isotherm.
/•M 0C-Z) weresteeperthanthosein the westernsector (>-0.8 /•M øC-Z). Latitudinally,
the valueswere
4.3
lessvariable, with the least steepslopesoccurringin the
westernside poleward of 10øN and 15øSwhere the nutricline was deeper than the 26øC boundary. The variabil-
cal 26øC isotherm
Nitrate
Fluxes
The total surface area enclosedby the climatologiand continental
land masses and the
annual mean net surface heat flux into this region are
ity in the slope(Figure6b) washighernearthe equator shown in Table 1. Slight variations were obtained in
between
1600 and 180øW
than elsewhere.
These differ-
the total surface area between the data sets, mostly be-
encesin slopesimply that a higher turbulent exchange cause of differences in the northern location of the 26øC
of nitrate occurs in the eastern side, where the slopes isotherm. The net surfaceheat flux into the region obare steeper, than in the western side for an equivalent
tainedfromclimatologies
rangedbetween
8.3x 10TMand
turbulent flux of heat. Previousstudiesin this region 11x 10TMW. Usingthe meanslopefor the regiongiven
havefoundslopesof-l.12 and-2.04/•M øC-Z between above, the new production resultingfrom turbulent in180o to 93øW and 1.1øSand 1.1øN [Halpernand Feld- put of nitrate into the Pacific warm water pool ranges
man, 1993] and eastwardof 130øW between5øS and from 5.47x 1012to 7.26x 1012toolN yr-1 or an aver20øN [Fietiler et al., 1991],respectively,
whichare con- ageof 0.25-0.31mmolN m-2 d-• overthe entirearea.
14,262
PE•IA ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
120 ø E
140 ø
160 ø
180 ø
160 ø
140 ø
20 ø N
120 ø
•
100 ø
80 ø W
'•
15 ø
10 ø
o
o
o
10 ø
(a)
20 ø S
0
120 ø E
1
140 ø
2
160 ø
3
180 ø
160 ø
20 ø N
4
140 ø
'
15o
5
120 ø
6
1O0 ø
80 ø W
'
•iiiiiii!•ii!!••:•=•=•;•=•
5o
........
0.0
0.5
1.0
:.m_•:.:.•:.:.:.:..
::::::::::::::::::::::
1.5
2.0
2.5
3o0
3.5
4.0
Figure 4. (a) Nitrateconcentration
(micromolar)
at thedepthof the 26øCisothermand (b) its
5ø-square standard deviation.
For comparison,this can be expressedin stoichiometri- shows that upwelling is constrained to the upper 100
cally equivalent carbon units using a Redfield ratio of m. The longitudinal pattern in upwelling velocitieshas
6.6; the new productionrate is then 4.37x 10TMto 5.74 beensupportedby other studiesin this region[Halpern
et al., 1989; Halpern and Freitag, 1987]. Also, calculax 1014g C yr-1 or 19.9-24.9mg C m-2 d-1.
The net advective input of nitrate into the regionen-
tions of upwellingas a function of depth basedon direct
of divergence[Halpernand Freitag,1987]
closedby the 26øC isotherm(Table 2) is due to up- observations
welling (26%) as well as horizontaltransport (74%) and on a model of equatorialcirculation[Br!ldenand
of nitrate into the region. At the 26øC surface, most Brad!h 1985] show that vertical velocitiesfall sharply
(•60%) of the horizontaladvectionof nitrate occurs above 50 m and below 100 m. Although vertical nitrate
meridionally from the south eastward of 110øW rather
than zonally from the east between 2.5øN and 10øS.Ac-
input is constrained to within a degree of the equator
[Halper•a•d Freitag,1987],divergence
ofupwelled
wa-
ter producesa nutrient-rich area larger in meridional
and
zonal extent than the zone of upwelling.
centrationof nitrate at the boundaryis small(about-•-1
cordingto the data available,the variability in the con-
/•M). Most of the verticaltransportof nitrateis dueto
upwellingwithin 1o of the equator, whereasdownwelling 5. Discussion
occursnorth of the equatorand represents•30% of the
We have estimated
the mean annual rate of new
nitrate upwelled. Along the equator, the model of Philawderet al. [1987]indicatedmaximumupwellingnear production in the warm water of the tropical Pacific
1500to 180øWwith diminishedupwellingeastwardand from a considerationof the large-scalenutrient balance
westward of this region. In the vertical, their model based on the net surface heat flux, mean ocean circu-
PE•IA ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
from othersources[e.g.,Epple•landPeterson,1979].In
somelocations,however,nitrogenfixation [Caponeand
Carpenter,1982; Carpenterand Romans,1991]can be
50
ß
-
4-5-
15
ß
.?..-..:..
ß
•,•.;,•...
ß 10
ß
40-
ß
significant. Although we are not awareof nitrogenfixation measurementsin the tropical Pacific, this input is
;-
...--.1•.•
,..• ß-.-. ß
ß ,5
?:'.'•, ';.'...
ß
ß
,';: i-;2.•
. ..,•..--
"0
35-
.
./,.•,,•..,..
.,,•,........
.
' '
likely to be small,sincethis process
requiresiron [Morel
et al., 1991], an elementscarcein this region[Martin
et al., 1991]. Another considerationin the definition
.
'1"¾•I
I"
..
2o 25
-
20
is that we assumethat depth-integratednew production over the depth of the 26øC isotherm is equal to
that of the euphotic zone. In the eastern tropical Pacific, the depth of the euphotic zone is around 60 to
-.
...:%-
u25
3o
ß
.•..
14,263
.
_
70 m [Pega et al., 1990;Fledlet et al., 1991],whichis
closeto that of the 26øCisotherm(Figure3a). In the
-
western equatorial Pacific, the euphotic zone depth as
estimated from the average surfacechlorophyllconcen-
15-
trationwithin100latitudeoftheequator(0.11mgm-3
10-
[D,ndonne,u, 1992])and from Morel's[1988]modelis
about
95 m which
is almost
25 m shallower
than
the
i
26øC isotherm depth. However,severalstudies[Murray et al., 1989; Barber and Chavez,1991]of the tropi-
i
cal Pacific have shownphytoplankton production at the
0.1% light level, which is severalmeters below the euphotic zone, arbitrarily defined as the depth at which
0
,5
10
1,5 20
25
Temperalure
30
35
40
(oC)
Figure 5. Scatterplot of temperature(degreesCelsius)versusnitrate concentration
(micromolar)
in the
upper 250 m of the regionenclosedby the 26øCisotherm
light is reducedto 1% of surfaceirradiance. Therefore
our assumptionregarding enclosureof the most significant productive layer is satisfiedfor the entire area consideredo
To estimate
turbulent
fluxes of nitrate
into the re-
(n=48,100). The samerelationship
is shownfor the sur- gion, we have assumedthat there is no significantspatial covariance between the net heat flux across the 26øC
facevaluesin the insert (n=4,870).
isotherm and the slope of the nitrate-temperature 'rela-
lation, and nitrate and temperaturecorrelation.Previ- tionship. Although there are no data availablefor estiously,new productionhas been inferredfrom nitrate- matingits value,the generaltrend in heat flux differs
temperaturecorrelationand experimentalrelationships noticeably from the general trends in the slope of the
betweennitrate concentrationand the uptake of nitrate
temperature-nitrate relationship. Using the clhnatolog-
[Dugdaleet al., 1989]or the f ratio,i.e., the ratioof new ical net surfaceheat flux data of Weareet al. [1981]
productionto total production[Sath•lendranath
et al., as a rough indicator of what occurs to the subsurface
'1991]. The approachusedin this study is potentially heat flux and our slope of the nitrate temperature relamore robust than previous approachesbecauseit does
not requireestimatesof total productionand it is independentof the relationshipbetweenf ratio and nitrate
concentration,whichhavelarge uncertainties.However,
our parameterizationof new productionprovidesonly a
large-scaleaverageand cannotprovidedetail on spatial
or temporal variability in the rate of new production.
Also,it doesnot allowa directcomparisonto estimates
tionship,wefoundno significant
covariance
(r2=-0.16)
between these variables. On an annual scale, net surface heat fluxes in the tropical Pacific show mostly latitudinal variations, being higher near the equator and
decreasingaway from it [Esbensenand Kushnir, 1981;
Weare et al., 1981; Oberhuber,1988]. Similarly,turbulent vertical fluxes are generally three to five times
higher within 20 of the equator than they are poleward
_
of newproduction
basedonnitrateuptake(•5Nexperi- [Peters et al., 1989]. In comparison,We have found
ments)[Dugdale
et al., 1992;Pegaet al., 1992],which, mainly longitudinal variationsin the 5ø-squareslopeof
though more sensitive,are discreteand restrictedto
short timescales.
5.1
Potential
Sources
of Error
Model assumptions. In this study, new production has been definedas the productionresultingfrom
the net physical transport of nitrate into the volume
bordered by the 26øC isotherm. This operational definition neglectsother inputs of nitrogento the region
suchas inputs from rain and from biologicalnitrogen
fLxation. In most of the ocean,fluxesof nitrate from below the nutricline are much more important than those
the temperature-nitrate relationship,with highervalues
in the western tropical Pacific, whereasheat fluxes vary
latitudinally rather than longitudinally. Also, the slope
of the nitrate-temperature relationshipwithin 1ø of the
equatorwasnot significantly
different(p <<0.1) from
that poleward.
Finally, we have assumedthat horizontal turbulent
fluxes of nitrate into the region were small compared
to turbulent vertical fluxes. At the 26øC isotherm,
the concentration of nitrate in the upper layer was detectable only in the southeastern boundary that surrounded
the
cold nutrient-rich
waters
of the
eastern
14,264
PE•IA ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
120 ø E
140 ø
20 ø N '
'
'
160 ø
'
•
180 ø
•
160 ø
•
'
'
140 ø
'
'
120 ø
100 ø
80 ø W
'
!ii•ii::i!!•ii::•iiiiii::i::i::::::::::::::::::::::::::i::::iiii:•i::iii:•i::::•::::•
10
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
....................
:..................
::::::::::::::::::
.......
::::::::::::::::::::::::::::::::::::::::::::::::::::
.....-....
o
5
..•:•:•:•:•:•:•:•:•:::::•:::::::::•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•::•
•-'---•.
----•.
-___
_•--:---_-•:---- i-9-•
•:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::i:i:i:!:!:•:•:i:i:•:i:i:i:i:i:i:i:i:i:i:!:i:i:i:i:!:i:•:!:!:•:!:i:i:i:!:•:!:•:•:•:•:•:•:•:•:•:•+•i•;;•
"'-•
............ -•-•--•
•x•'-'-'-'-'-'-'-'-'-'-,-'-'--'-'-'-'-'-'-'-'-'-'-'-'-'
'•'•'•'•••••:•:•:•:•:
.......................................................................................................................................................
...............
-•..-_'-•':•_:•_::•:.:•:•:., .........
•.•-•j•
.........
:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:+:•:•:•:•:•:•:•:•:•:•:•:•:•:•:+:•:•:•:•:•:•:•:•:•:•:•:•:•:.•:•:•:;•.•:•:•:•:•:!:?•.•:•:•:•:•:•:•:•:•:
.:.:.:.:.:.:.:;:+:.:.:.:+:.:+:.:.:?:.:.:.:•;•.:•:.:_•:.:.:+::•:_
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
_•,.-•
_::;::::•:::
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::•..•
"::::::::::::::::::::::::::::::::::
•:_•2_•:_•:._.'•.;_•:_•:_•:_•:•;"•;:?•_:.:;:?:
ß
:,:.:.:+;•.
',:•._'•.....'.'.'.'.','.;•2,;•i'
.ii•':•:m':':'-'-'-"'-"'""•;••':'•.:.
..,.,,.,.?•
•;•,!•
•.'•:.'.'.'•-.;.'•
0.............
•i•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•i•
•-':
.•'
....
......
...........
•..;.,,,
_.•.•...•......•;•....•..•;.•.,.,•.....•.
ß . •;•,,,,.•.:.;.:+:.:,:o,,,,.',•;,
ß,,•;•,
10
,;• ,_•,'_•,%•:_•,',
ß
.,,,,,,,..
'='='='•-•='='='='=',
•=:=:=•=••;•;•;•=•`•;•=•=•:•e• •='='='='='.•:':':':':':':':':':':':':':':':':':•':-:':':
(a)
20øS
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
0.0
120 ø E
-0.2
-0.4
140 ø
-0.6
160 ø
-0.8
180 ø
-1.0
-1.2
160 ø
-1.4
140 ø
20 ø N
'
-1.6
120 ø
-1.8
-2.0
1O0ø
80 ø W
'
•=•;i:i:i:i:i:i:i:!:•:i:!:i:i:i:i:i:i:i:•:i:i:i:i:i:i:i:i:i:i:i:i:i`•i•=i
"=•.'i.'i;•.'•i•..
,•,,,.•,,,,,,.;•,..:.:.:.:..,.,•
;.•_:.:.:.:.•.:.:.:.:.:,:.:.:.:.:.:.:.:.:,:,,,
•_,,,,
10 o
..•.......-.,...•...=•=...,,=,=......,,•-....:,.,,........•.=.....,.,:...
_•,...,,...
,,:,,.,,.,:,:,.,,.,,:.:,.,,:::,•_.,,,,:,:•:,:_•:,•:.:,,.
ß:,:•.,,,,,.•,=•:.,,=g_'•:;,,,
tii
.•..=,..:,•........,_=:....
'"":-'"':'"'";"':'"'•'"'"•';'"'•'
'"'""%%'"";'=
'='=•=
'"?'%'"'""•"
";:'•"•';";"
..•..ii•;•i•i!i•i:
:[:::
!;•:.=•.•;%.•;•;•_.•:%%
ß
0o
;•:•.
:.:.:.:.:•-.:.:.:•..:.:.:.:,:.:•.•-•.:::.:.l•:::.;•.:.:.:
:•:.•:-:•;.•-"•:>.'•::::::::::
:::::.'•-.'•.;i.'•'-.•.';•;•:::";:!i•
:.:•:•:.:
ß:;::i•[•:•_:;';'i.'.':':':':':':':':
;-•&;•:•:•:':';•':+•:
':':'_.'•':'
,,,,,•,,?•...:...:.:.'.'.:...:.•_•,
,.,;?,,;•,,•,,,,
•:_;
ß
:,,'_•,'?•_•,,%•,'?-..,...:.....:.:...:...
:_•.'_;:•?,,'g_•,•
gg,:,.¾,',,',
5o
,,,,,5,,,,
•1'
.':':-•:•':•:-•;'•::':'::
:':':':':::':':;:•-'•_::•?•i::::•:•:_:i:
15ø
:::::::::::::::::::::::::::::::::::::::::::
•,-.-i
•.=•=•:•;•=•;•=..•;.:.•.=•.:;;•;i•.:•:!•.;•.;•;•.•;•;•=•;•.•i?•;•;•=.;•;•i•=:>.•:.:•.:....::•:...•:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.a:.:.:.:.:...:...:.:..:.:.:.:.:.:.:.:.:.:.:.:
20 øS
..............................................................................................................................
•
T
0.0
0.1
0.2
0.3
0.4
•_L
0.5
Figure 6. (a) The slopeof the temperature-nitrate
relationship
(micromolar
perdegreeCelsius)
at the baseof the 26øCisothermand (b) the standarddeviationof the slopeat every5ø-square
in the region enclosedby the 26øC isotherm.
equatorial Pacific. In this region the maximum hori-
they reproducethe main featurespresentin the clima-
zontalnitrate gradientwas10-5 mmolm-4. An upper tologicaltemperature field.
layer of 50 m and a lateral turbulent exchangecoefficient
With the data available,we haveobtaineda 95% con-
of 103m2 s-• [Br!tde•a•d Brad[I,1989]will resultin fidence interval of about -4-12% in the estimation of the
the transportof 3.3 x 10TMmmolN d-• intothe region, averageslopeof the nitrate-temperaturerelationshipfor
which, though it may be significantlocally, is 2 orders the region. In comparison,the individualheat flux comof magnitude less than the estimated total turbulent ponentshavelargeuncertainties
[Weareet al., 1981],so
vertical transport.
that the total heat flux has very large error bounds. In
Model parameter estimates. Climatologicaltem- the tropical Pacific, the different estimatesof net annual
perature fields [Ezbe•ze• a•d Kuzh•ir, 1981; Weareet heat flux agreewithin a factor of 2 [Niiler a•d Ste•e•al., 1981; Oberhuber,
1988]in the tropicalPacificshow zoo,1982].Thereforemostof the uncertaintyin thenew
that
water
warmer
than
26øC
reaches near 21ø-24øN
production estimate due to turbulent input of nitrate is
and 20ø-21øSin the western region, whereasour data associatedwith errors in the magnitudeof the net surset coverageis limited to 20o latitude from the equator. face heat flux. The error in the estimation of the net
This discrepancycan be ignoredwith little consequence advective transport of nitrate associatedwith changes
in our analysis, where the horizontal grid boxesare as of 1/•M in the nitrate concentrationof the sourcewater
wide as 5ø latitude and longitude. Also, despite the is about 26% of the net advectivetransport. Although
uneven temporal and spatial distributions of the data, there are no data available for evaluating the errors in
PE•A ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
14,265
Table 1. Heat Flux, Nitrate Transport, SurfaceArea Enclosedby 26øC Isotherm
in the Tropical Pacific,and CorrespondingNew Production
Net
Mean
Tm'bulent
Heat Flux,
Area,
Nitrate Transport,
New Production,
W x 10•4
m2 x 10•a
toolN yr-• x10•2
mgC m-2 d-•*
Source
11.0
10.7
6.31
6.44
7.26
7.04
24.9
23.7
Weare et al., 1981
Oberhuber, 1988
8.3
5.98
5.47
19.9
Esbensen
and
Kushnir, 1981
*DeterminedusingRedfieldC/N ratio of 6.6 (atom/atom)
the mean water transport, the mean ocean circulation
valuesgivenby Philanderet al. [1987]agreewell with
fieldobservations
[HalpernandFreitag,1987]and other
modeloutputs[Br•ldenandBrad161985;Halpernet al.,
1989]. We concludethat the main quantitiablesource
of error in our estimation of new production is related
to uncertainty in the value of the net surface heat flux
used in the calculation.
For the western equatorial Pacific, annual mean surface heat fluxes reported in the literature range from
25 to 100W m-2 [Esbensen
andKushnir,1981;Reed,
tions of net surface heat flux in the tropical Pacific derived from metereological reports from merchant ships
and fishing vesselshave uncertainties related to poor
spatial and temporal coverage. Heat flux estimations
basedon satellite data, which can providebasin-wide
coveragefor days to years, may considerablyimprove
the estimation of this parameter and thus reduce the
error in our new production estimate.
5.2 Comparison
With
Other
New Production
Estimates
1985; Gordon, 1989]. In this regionthe oceancirculation is too slow and the temperaturegradientis too
weak for either advectionor mixing to carry any significant heat awayfrom the region[Niiler and Stevenson,
1982]. Thus in orderto maintaina steadyannualmean
temperature, the net heat flux through the sea surface
Measurementsto calculatethe long-termnew production of the tropical Pacific are unavailable and likely to
remain so for a considerabletime. Thus the only values
available for a comparison of new producticn estimates
at these scalesis that obtained by Badaztoivand Maier-
must balance the vertical
carbon cycle in the ocean. They have estimated a new
turbulent
fluxes of heat.
How-
ever, because the near-isothermal salt-stratified layer
tteimer [1990]from a three-dimensional
model of the
production
rateof 18-20g C m-2 yr-1 which'is
only
[LukasandLindstrom,1991]foundin the westernequa- slightlyhigherthan the onewehaveobtained(14.5-16g
torial Pacific seemsto insulate the deep oceanfrom the
C m-2 yr-1) for approximately
thesame!:eg!on.
Given
surfacelayer muchof the time [Godfrei!and Lindstrom,
the uncertainties
1989],eitherthenet surface
heatfluxes(18-22W m-2)
the approximations involved in theseparameteri•.ations,
the agreementbetween the two estimates'isremarkably
good.
In the eastern equatorial Pacific region, a mean inte-
usedby Nitlet and Stevenson[1982]are too large, or
the vertical fluxes of heat occur only episodically, for
example, in association with wind bursts. For exam-
in both
our and their
estimates
and
ple, Godfrelland Lindstrom[1989]estimatea turbulent gratedannualnewproduction
of 1.3x 1015to 1.9x 1015
verticalheat flux of only 10-16W m- 2 in the top 100 g C yr-1 (or 320-470mgC m-2 d-1) hasbeenreported
m of the western equatorial Pacific from data obtained
in calm conditions, which is substantially smaller than
within 50 of the equator and between 900 and 180øW
[Chavezand Barber, 1987]. Thesevalueswerebasedon
that estimatedby Niiler and Stevenson[1982]. How- estimation of nitrate upwelling and are approximately
ever, Godfrey and Lindstrom suggestedthat the strong 30 to 90% higherthan the total integratednewproduc-
westerly winds that occur intermittently in this region tion estimated in this study. The area consideredin the
might producestrongmixing events.Field observations presentanalysisis approximatelysix times larger than
in the western Pacific between November 1989 and Janthat considered by Chavez. and Barber and includes
uary 1990 have shownstrongvertical sheardown to 150 large unproductive areas in the western Pacific not in-
m in response
to westerlywindevents[McPhadenet al.,
1992]. Althoughthe meanwind speedsare weakin the
cludedby them. Severalstudies[e.g., Thomas,1979]
have shown the persistence of unused nitrate in surface
waters of the eastern and central equatorial Pacific despite irradiance and stratification conditions favorable
cific [Keen,1988]showsthat this area experiences
such to complete nutrient uptake. This persistence results
wind events more often than regions to the east and in an important fraction of the nitrate upwelled being
that these events are far more frequent in the northern transportedawayfrom the upwellingzone[Cart, 1991;
western equatorial Pacific, their variability is large. A
climatology of westerly wind bursts in the western Pa-
winter
season.
Fledlet et al., 1991]. BecauseChavez.and Barber did
Considerableprogresshas beenmade in the computa- not consideredthis loss'of nitrate, their estimation reption of the most important terms of surface heat fluxes resentsthe upper limit of local new production. How(i.e., surfacesolarirradianceandlatentheatflux) from ever, since downwelling of nitrate out of the euphotic
satelliteobservations
[Liu and Gautier, 1990]. Estima- zone(whichrepresents
a nutrientsink)wassmallin the
14,266
PE•IA ET AL.: NEW PRODUCTIONIN THE TROPICALPACIFICOCEAN
Table 2. Mean Advective Nitrate Balanceof Region
Enclosed by 26øC Isotherm in the Tropical Pacific and
CorrespondingMean New Production
Mechanism
Nitrate •ransport,
New Production,
mol N yr-1
mg C m-9' d-1 *
2.0 x 1012
-5.5 x 10•x
1.5 x 1012
?.0
-1.9
5.1
Meridional
2.6 x !019'
9.1
Zonal
1.6 x 1012
Net input
4.2 x 1019'
14.8
Net transport
5.7 x 10•2
19.9
Vertical Advection
Upwelling
Døwnwelling
Net input'
Horizontal
.,
Advection
.
5.7
*Determined
usingRedfieldC/N ratioof 6.6 (atom/atom)
region consideredin this study, most of the upwelled
nitrate is eventually consumedby phytoplankton, but
consumptionmay occur well away from the equator.
A net advective balance of nitrate into the euphotic
This implies a decreasein the vertical transport of heat
through the thermoclineand thereforea decreasein the
input of nitrate into the surfacelayer. Model analysis
has shownthat changesin the tropical Pacific sea surzoneof the easterntropicalPacific[Fietileret al., 1991] face temperature are associatedmostly with anomalous
zonal advection of warm water eastward and with varigavean estimateof new productionof 200 mg C m
d-1 eastwardof 130øWand within 50 of the equator, ationsin the depth of the thermocline[W•lrtki, 1975;
which was about 30% of total production. A similar Seaõer,1989;Halpern,1987]. As a resultof variations
valuefornewproduction
(196mgC m-2 d-1) wasesti- in the westerly winds, the thermoclineis raised in the
mated
from
turbulent
and advective
fluxes of nitrate
westand depressed
in the east [Halpern,1987]. Since
within 50 of the equator along a transect at 150øW the thermoclineis depressedbelow the depth of entrain[Cart, 1991]. In this region,turbulentdiffusionof ni- ment in the eastern Pacific, the water upwelled toward
trate was about 30% of the advectivetransportof ni- the surfaceis warm and depleted of nutrients.
trate, and the resulting new productionwas about 40%
of the total production. TheSe values are almost five
timeshigherthan our meannewproduction(44 mg C 5.8 New Production and the Relationship to Net
m-2 d-X). Considering
a meantotal production
of 312 Air-Sea Exchange of Carbon
mgC m-2 d-1, asfoundbetween160øWand 140øEand
within 50 of the equator[Barberand Chavez,1991],the
new production estimated here is about 14% of the total production. This representsa lower-boundestimate,
sincewe have consideredregionsawayfrom the equator
Most of the new production in the oceanresultsfrom
the supply of inorganic nitrogen from below the eu-
wheretotal productionis likely to be lower.
The productivity of the tropical Pacific region is
vertical transport of dissolvedinorganiccarbon and nitrate is in approximately the Redfield ratio and if the
C and nitrate are taken up in the euphotic zone and
exported out of it in the same ratio, there would be
no net biologicallymediated transport of carbonto the
very strongly affected by interannual events like E1
photic zone [e.g., Epple•land Peterson,1979]. If the
Nifio-Southern Oscillation(ENSO). Studiesof both
surfaceproductivity and fluxesto the sediments
mond and Collier, 1988]suggestthat the interannual deepocean[Epple•landPeterson,1979].Unlikenitrate,
variability associatedwith these eventsfar exceedsthe
however, atmospheric CO2 enters the ocean by sea-air
seasonalvariability. It has been estimated[Chavez gasexchangeand is redistributedthroughthe combined
and Barber, 1985]that the total primary productivity effectsof advection,mixing, formationof particulate
within 5ø•of the equatorand between820and 172øW matter in the surface layer, and subsequentremineralwas 2-3 times lower during E1 Nifio than in normal
oceanographicconditions. ENSO conditionsalso suppressed new production at a site close to the equator
but enhancedit away from the equator [D•lmondand
Collier, 1988]. An important manifestationof E1Nifio
is the increase in sea surface temperatures above their
long-term mean valuesby as much as 4øC in the central
and eastern Pacific.
This increase in the heat content of
the upperoceangreatlyexceeded
(about10 times)the
estimatedchange in the surfaceheat flux [Reed,1986].
ization in deep water. If there were sufficientdata available for the tropical Pacific to allow us to determinethe
inorganic carbon concentrationversustemperature relationship at the base of the subsurfaceboundary, the
mean annual vertical ttux of inorganiccarbon could be
estimated using the same computationas that for the
turbulent nitrate fluxes. Then the rate of new production could be compared with the rate of carbon ttuxes,
and the net air-sea exchangeof carbondioxidecouldbe
evaluated.
PEI•IA ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
6. Conclusion
14,267
Capone, D. G., and E. J. Carpenter, Nitrogen fixation in
the marine environment, Science, 1•7, 1140-1142, 1982.
Carpenter, E. J., and K. Romans, Major role of the cyanobacterium Trichodesmiumin nutrient cycling in the North
Our analysisgivesan estimateof new productionin
the warm waters of the tropical Pacific from a considAtlantic Ocean, Science, l•5j, 1356-1358, 1991.
eration of the large-scalenutrient balance. In this re- Cart, M. E., Interactions between physical and biological
processesin the equatorial Pacific, Ph.D. thesis,Dalhousie
gion, the new productionresultingfrom vertical turbuUniversity, Halifax, Nova Scotia, Canada, 1991.
lent fluxesof nitrate was of the same magnitude as that
due to advectivenitrate transport. Most (about 75%)
of the advective transport of nitrate was due to hori-
zontaltransport(zonaland meridional)from the eastern equatorial Pacific rather than upwelling. Previous
estimations of new production from nutrient budgets
[Chavezand Barber, 1987; Fiediet et al., 1991]in the
tropical Pacificregionhaveignoredthe input of nitrate
due to vertical turbulent transport. The method employed here, which considersthe input of nitrate due
to turbulent diffusion as well as to advection, yields
more representativeestimatesof the large-scaleaverage
new production than those calculated from advection
alone. We have estimated the new production resulting from vertical turbulent fluxes of nitrate based on
fluxes of heat at the sea surface. This approach can
be applied to other regionsof the world'soceans. The
possibilitythereforeexiststhat there is a meansto estimate new productionfrom remotelysensedobservations
which doesnot requirethe determinationof physiological parametersof nitrate uptake from ship observations
that are notoriouslyvariable. Therefore although our
estimation of new production is based on several assumptionsand is lesssensitivethan direct estimations,
it providesan estimate over oceanographicallysignificant scalesof time and spacewhich are inaccessibleto
ship observations.
Chavez, F. P., and R. T. Barber, Plankton production during E1 Nifio, WCRP Publ. Set., 4, 23-31, 1985.
Chavez, F. P., and R. T. Barber, An estimate of new production in the equatorial Pacific, Deep Sea Res., $j, 12291243, 1987.
Craig, H., W. S. Broecker, and D. W. Spencer, GEOSECS
Pacific Ezpedition: Sections and Profiles, vol. 4, 251
pp., U.S. GovernmentPrinting Office, Washington,D.C.,
1982.
Dandonneau, Y., Surface chlorophyll concentration in the
tropical Pacific Ocean: An analysis of data collectedby
merchant ships from 1978 to 1989, J. Geophys. Res., 97,
3581-3591, 1992.
Dugdale, R. C., and J. J. Goering,Uptake of new and regenerated forms of nitrogen in primary production, Limnol.
Oceanogr., •t•, 196-206, 1967.
Dugdale, R. C., A. Morel, A. Bricaud, and F. P. Wilkerson,
Modeling new production in upwellingcenters: A case
study of modeling new production from remotely sensed
temperature and color, J. Geophys. Res., 95, 18,11918,132, 1989.
Dugdale, R. C., F. P. Wilkerson, R. T. Barber, and F. P.
Chavez, Estimating new productionin the equatorialPacific Ocean at 150øW, J. Geophys. Res., 97, 681-686,
1992.
Dymond, J., and R. Collier, Biogenic particle ttuxes in
the equatorial Pacific: Evidence for both high and low
productivity during the 1982-1983E1 Nifio, GlobalBiogeochem. Cycles, 1•, 129-137, 1988.
Eppley, R. W., and B. J. Peterson,Particulate organicmatter ttu.xand planktonic new production in the deep ocean,
Nature, 1•8•, 677-680, 1979.
Acknowledgments.
We thank W. G. Harrison and D.
Esbensen,S. K., and Y. Kushnir, The heat budget of the
Kelley for helpful discussionand commentson the manuscript.
global ocean: An atlas based on estimatesfrom marine
Our thanks also to P. C. Fledlet and V. Philbrick for proobservations, Rep. 29, Clim. Res. Inst., Oreg. State
riding temperature and nitrate data and to the National
Univ., Corvallis, 1981.
Center for Atmospheric Researchfor supplying us with the
Fledlet, P. C., V. Philbrick, and F. P. Chavez, Oceanic upheat flux data. This work was supported by the Natural
Sciencesand EngineeringResearchCouncil(Canada), the
National Aeronautics and Space Administration, and the
Naval Oceanographic Laboratory.
References
welling and productivity in the eastern tropical Pacific,
Limnol. Oceanogr., 36, 1834-1850, 1991.
Godfrey, g. S., and E. J. Lindstrom, The heat budget of the
equatorial Western Pacific surfacemixed layer, J. Geophys. Res., 9j, 8007-8017, 1989.
Gordon, C., Comparisonof the surfacefluxesin the tropical
Pacific Ocean derived from an atmosphericgeneral circulation model and from climatology, Trop. Ocean Atmos.
Bacastow, R., and E. Maier-Reimer, Ocean-circulationmodel
Ne•vsl., 5•, pp. 1-4, Coop. Inst. Mar. Atmos. Sci.,
of the carbon cycle, Clim. Dyn., j, 95-125, 1990.
RSMAS, Univ. Miami, Fla., 1989.
Barber, R. T., and F. P. Chavez, Regulation of primary productivity rate in the equatorial Pacific, Limnol. Oceanogr., Halpern, D., Observationsof annual and E1 Nifio thermal
and flow variations at 0ø 110øW and 0ø 95øW during
36, 1803-1815, 1991.
1980-1985, J. Geophys. Res., 91•,8197-8212, 1987.
Betzer, P. R., W. J. Showers,E. A. Laws, C. D. Winn, G.
R. DiTullio, and P.M. Kroopnick, Primary productivity Halpern, D., and G. C. Feldman, Annual and interannual
variations of phytoplankton pigment concentration and
and particle fluxeson a transectof the equatorat 153øW
upwelling along the Pacific equator, J. Geophys. Res.,
in the Pacific Ocean, Deep Sea Res., 31, 1-11, 1984.
99, 7347-, 7354, 1994.
Bryden, H. L., and E. C. Brady, Diagnostic model of the
three-dimensional circulation in the upper equatorial Pa- Halpern, D., and P. H. Freitag, Vertical motion in the upper
ocean of the equatorial eastern Pacific, Oceanol. Acta, 6,
cific Ocean, J. Phys. Oceanogr., •5, 1255-1273, 1985.
19-26, 1987.
Bryden, H. L., and E. C. Brady, Eddy momentum and heat
fluxes and their effects on the circulation of the equatorial Halpern, D., R. A. Knox, D. S. Luthier, and S. G. H. Philander, Estimates of equatorial upwellingbetween 140ø and
Pacific, J. Mar. Res., j7, 55-79, 1989.
110øW during 1984, J. Geophys. Res., 9j, 8018-8020,
Campbell, J. W., and T. Aarup, New production in the
North Atlantic derived from seasonalpatterns of surface
chlorophyll, Deep Sea Res., 39, 1669-1694, 1992.
1989.
Kamykowski, D., A preliminary biophysical model of the
14,268
PEfIA ET AL.: NEW PRODUCTION IN THE TROPICAL PACIFIC OCEAN
relationship between temperatare and plant nutrients in
the upper ocean, Deep Sea Res., 5•t, 1067-1079,1987.
Kamykowski,D., and S.-J. Zentara, Predictingplant nutrient concentrationfrom temperatare and sigma-t in the
world ocean, Deep Sea Res., 55, 89-105, 1986.
Keen, R., Equatorial westerlies and the Southern Oscillation, in Proceedingof the U.S. TOGA Western Pacific
Air-Sea Interaction Workshop,edited by R. Lulls and P.
Webster, pp. 121-140, UCAR Tech. Rep. USTOGA-8,
Univ. Corp. for Atmos. Res., Boulder, Colo., 1988.
Lewis, M. R., Satellite ocean color observationsof global
biogeochemicalcycles, in Primary Productivityand BiogeochemicalCycles in the Sea, edited by P. G. Falkowski
and A.D. Wooclhead, pp.139-156, Plenum, New York,
Osborne, J., J. Swift, and E. P. Flhtchem, Ocean atlas for
Macintosh: A microcomputer application for examining
oceanographicdata, version 2.0, S. L O. Ref. 9•-•9,
Scripps Inst. of Oceanogr., La Jolla, Calif., 1992.
Pefia, M. A., M. R. Lewis, and W. G. Harrison, Phytoplankton productivity and size structure of phytoplankton biomasson a transect of the equator at 135øW in the
Pacific Ocean, Deep Sea Res., 37, 295-315, 1990.
Pefia, M. A., W. G. Harrison, and M. R. Lewis, New production in the central equatorial Pacific, Mar. Ecol. Prog.
Set., 80, 265-274, 1992.
Peters, H., M. C. Gregg, and J. M. Toole, Meridional variability of turbulence through the equatorial undercurrent,
J. Geophys. Res., 94, 18,003-18,010, 1989.
1992.
Philander, S. G. H., W. J. Hurlin, and A.D. Seigel, SimuLewis, M. R., M. E. Cart, G. C. Feldman, W. Esaias,and C.
lation of the seasonalcycle of the tropical Pacific Ocean,
McClain, Influence of penetrating solar radiation on the
J. Phys. Oceanogr., 17, 1986-2002, 1987.
heat budget of the equatorial Pacific Ocean, Nattire, 5g7, Platt, T., and W. G. Harrison, Biogenic fluxes of carbon
543-545, 1990.
and oxygenin the ocean, Nature, 318, 55-58, 1985.
Liu, T. W., Moistare and latent heat flux variabilities in the
Platt, T., W. G. Harrison, M. R. Lewis, W. K. W. Li, S.
tropical Pacific derived from satellite data, J. GeopAys.
Sathyendranath,R. E. Smith, and A. F. Vezina, Biological
Res., 95, 6749-6760, 1988.
production of the oceans: The casefor a consensus,Mar.
Liu, T. W., and C. Gautier, Thermal forcingon the tropical
Ecol. Prog. Set., 5•, 77-88, 1989.
Pacific from satellite data, J. Geophys.Res., 95, 13,209- Reed, R. K., An estimate of the climatologicalfluxes over
13,217, 1990.
the tropical Pacific Ocean, J. Clint. Appl. Meteorol., •J,
Lukas, R., and E. Lindstrom, The mixed layer of the west833-840, 1985.
een equatorial Pacific Ocean, J. Geophys.Res., 96, 3343- Reed, R. K., Effects of surfaceheat flux during the 1972 and
3357, 1991.
1982 E1 Nifio episodes, Nature, 3•, 449-450, 1986.
Martin, J. H., R. M. Gordon, and S. E. Fitzwater, The case Sathyendranath, S., T. Platt, E. P. W. Horne, W. H. Harrifor iron, œintnol. Oceanogr.,56, 1793-1802, 1991.
son, O. Ulloa, R. Outerbridge, and N. Hocpffner,Estimation of new production in the ocean by compoundremote
McPhaden,M. J., F. Bahr, Y. du Penhoat,E. Firing, S. P.
sensing, Nature, 353, 129-133, 1991.
Hayes, P. P. N'filer, P. L. Richardson,and J. M. Toole,
Seaget, R., Modeling tropical Pacific sea surfacetemperaThe responseof the western equatorial Pacific Ocean to
tare: 1970-87, J. Phys. Oceanogr., 19, 419-434, 1989.
westerly wind bursts during November 1989 to January
Thomas, W. H., Anomalous nutrient chlorophyll interrela1990, J. Geophys. Res., 97, 14,289-14,303, 1992.
tionships in the offshore eastern tropical Pacific Ocean.,
Minas, H. J., and M. Minas, Net community production in
J. Mar. Res., 37, 327-335, 1979.
"high nutrient-low chlorophyll"waters of the tropical and
Antarctic oceans: Grazing vs iron hypothesis, Oceanol. Weare, C., P. T. Strub, and M.D. Samuel, Annual mean
surface heat fluxes in the tropical Pacific Ocean, J. Phys.
Acta, 15, 145-162, 1992.
Oceanogr., 11, 705-717, 1981.
Morel, A., Optical modelling of the upper oceanin relation
to its biogenousmatter content(caseI waters),J. Geo- Wyrtki, K., E1 Nifio-The dynamic responseof the equatophys. Res., 93, 10,749-10,768, 1988.
Morel, F. M. M., J. G. Rueter, and N.M. Price, Iron nutrition of phytoplankton and its possible importance in
the ecology of ocean regions with high nutrient and low
biomass, Oceanography,J, 56-61, 1991.
Murray, J. W., J. N. Downs, S. Strom, C.-L. Wei, and H. W.
Jannasch, Nutrient assimilation, export production and
2a4Thscavenging
in the easternequatorialPacific,Deep
Sea Res., 36, 1471-1489, 1989.
N'filer, P., and J. Stevenson, The heat budget of tropical
ocean warm-water pools, J. Mar. Res., gO, suppl., 465480, 1982.
Oberhuber, J. M., An atlas based on the 'COADS' data
set: The budgets of heat, buoyancy and turbulent kinetic energy at the surface of the global ocean, Rep. 15,
Max Planck Institute for Metereology, Hamburg, Germany, 1988.
rial currents in the western Pacific Ocean to atmospheric
forcing, J. Phys. Oceanogr.,5, 572-584, 1975.
Wyrtki, K., An estimate of equatorial upwelling in the Pacific, J. Phys. Oceanogr., 11, 1205-1214, 1981.
Zentara, S. J., and D. Kamykowski, Latitudinal relationships among temperature and selected plant nutrients
along the west coastof North and South America, J. Mar.
Res., 35, 321-337, 1977.
J. J. Cullen, M. R. Lewis, and M. A. Pefia, Department
of Oceanography,Dalhousie University, Halifax, Nova Scotia
B3H 4J1, Canada.
(ReceivedJune 18, 1993;revisedDecember20, 1993;
acceptedFebruary7, 1994.)