FISHERIES OCEANOGRAPHY Fish. Oceunom. 7:3/4, 311-316, 1998

New production in the equatorial Pacific: a coupled dynamical-biogeochemical model

Lf A. STOENS," C. MENDS, YjDANDONNEAU AND L. MEMERY I

LODyC, CNRS, ORSTOM, Uniuersité Puris VI, Tour.14, 2ème étage, 4 , Place Jussieu, 75252 Paris cedex 05, France

ABSTRACT

A simple 3-D biogeochemical model is coupled to a dynamical model forced by weekly winds deduced from ERS1 scatterometer data, to simulate new production in the equation Pacific from April 1992 to June 1995. The biogeochemistry is modelled as a nitrate sink modulated by chlorophyll, using nitrate/chlorophyll regressions derived from field data. The first simulation was carried out assuming that remineralization below the euphotic layer is totally controlled by sinking particles. In the second simulation, it is shown that the simulation of nitrate and new production by the bio- geochemical model is improved, in comparison with field data, by adding an explicit dissolved organic ni- trogen compartment. In the equatorial band, the model simulates a nitrate-poor region (low new pro- duction) in the fresh warm pool separated from richer waters of the upwelling region by a salinity front. The zonal displacement of this salinity/nitrate front is as- sociated with the El Niño-Southern Oscillation (ENSO). The modelled new production and physics, both in good agreement with the field data, represent useful tools for the study of skipjack tuna (Katsuwonus pelamis) forage distribution in the Pacific.

Key words: dissolved export production, ENSO, equatorial Pacific, model, new production, salinity front

* *Cowes@ondence. Fax: +33 1 44 27 71 59; 1 e-mail: ast@lodyc.jussieu.fr

6'

1

Received for publication 4 April 1998 Accepted for püblication 24 July 1998 -

O 1998 Blackwell Science Ltd. I I

INTRODUCTION

The equatorial Pacific is one of the main productive areas of the world. The upwelling of nutrients main- tains intense biological activity that could account for 50% of global new production (Chavez and Barber, 1987). The inorganic carbon and nitrate contents of surface seawater are controlled by new production, defined by Dugdale and Goering (1967) as the sum of inorganic nitrogen inputs to the photic zone. New production and related transport of organic carbon and associated nutrients are key elements in' oceanic car- bon cycle models (Bacastow and Maier-Reimer, 1991). The equatorial Pacific is also subject to the El Niño- Southern Oscillation (ENSO) that affects the world climate on an interannual time scale and the equa- torial Pacific therefore constitutes a key region for studying the carbon cycle. To understand new pro- duction and carbon fluxes on basin and ENSO t i m 7 scales, a 3-D ocean general circulation model . (OGCM) of the Pacific Ocean has been coupled to a simple biological model. The results from this model are described in this paper.

Until recently, it was thought that the flux of new production was carried mostly by the vertical flux of particulate organic nitrogen (PON). In this respect, we made a first simulation, in which the equivalent of new production in the photic layer is instantaneously remineralized below the photic layer as a vertical flux of sinking particles. This produces, in the upwelling region, an excess accumulation of nitrate at depth, referred to as 'nutrient trapping' (Najjar et al., 1992). However, recent results from JGOFS (Joint Global Ocean Flux Study) process studies show that new production results partly from dissolved organic ni- trogen (DON) and accounts for as much of the new production as does the downward export of particles (Buesseler et al., 1995; Lu0 et al., 1995; Libby and Wheeler, 1997). A second simulation was then made in which new production is fractionated into sedimenting particles, and into DON (50%). Indeed, previous modelling approaches (Bacastow and Maier- Reimer, 1991; Najjar et al., 1992) show that including DON produces better agreement between predicted and observed nutrient concentrations.

311

i - 1 1i-i d S LA OCM M 8 ntai E d RZd'PB M

312 A. Stoens et al.

In the present study, we compare modelled nitrate distributions, both with and without an explicit dis- solved organic nitrogen (DON) compartment] with field data. Then the interannual variability of new production associated with the ENS0 is analysed from April 1992 to June 1995. Implications of this vari- ability are presented in this work.

MODEL DESCRIPTION

We use the 3-D LODyC OGCM (Blanke and Delecluse, 1993) forced with observed weekly winds deduced from the ERS1 scatterometer (Grima et al., 1998) from mid-1992 to mid-1995. The outputs of this model were coupled off line with a simple biological model. The model domain is limited to the equatorial Pacific region and from 20"s to 20"N. At each model grid point and time step, we solved the advection-diffusion equation for nitrate on the OGCM grid:

The left-hand term represents the Eulerian variation of the tracer (nitrate) and is calculated as the sum of the advection terms (first term on the right-hand side), of the horizontal diffusion terms (this one is quasi- negligible here), of vertical diffusion terms and of the biological sink. Currents and diffusion coefficients are taken from the dynamical simulation (Grima et al.,

The last term of equation 1 represents the bioloe ical nitrate sink (biological model). This term is modelled as a simple nitrate sink modulated by chlo- rophyll. To keep the model as simple as' possible, no phytoplankton was explicitly represented in the model. Instead, we estimated the chlorophyll con- centration using a nitrate/chlorophyll linear regression derived from in situ data. This choice is based on the decrease of the chlorophyll concentration at the sur- face from the eastern equatorial upwelling to the western warm pool (Dandonneau, 1992; Yoder et al., 1993), and on a similar decrease in nitrate concen- tration that mirrors the westward temperature increase (Chavez et al., 1996). Finally, the vertical profile of chlorophyll is estimated from the concentration at the surface and statistical equations by Morel and Berthon (1989). The vertical profile of chlorophyll obtained in this way is used to calculate the irradiance at depth

O 1998 Blackwell Science Ltd., Fish. Oceunogr., 7:3/4, 311-316.

~ 1998).

'

'

according to Morel (1991). Photosynthetic nitrate fixation is then given by:

where V,, is the maximum nitrate assimilation rate and &O3 is the half-saturation concentration. PUR is the photosynthetic 'usable' radiation (Morel, 1991) and KE is the half-saturation constant for PUR. The V,,, and KNO, values that give the best fit with the FLUPAC (FLUX dans l'ouest du PACifique équatori- al) and OLIPAC (OLIgotrophie en PACifique) ob- servations of nitrate concentration are 6 x pmol NO3 mg chl-' s-l, and 0.01 pmol NO3 kg-', respec- tively. The daily cycle of irradiance is not represented in this model, and KE is taken to be equal to 70 pEinstein m-'s-', in order to reproduce a limitation by light in agreement with the results of light/photosyn- thesis experiments made by Karima Allali and Marcel Babin during FLUPAC and OLIPAC (Moutin and Coste, 1996). The photic layer is defined as the 1% incident light penetration depth. The particles re- sulting from new production are instantaneously re- mineralized below the photic layer, with an exponential decrease with depth (Honjo et al., 1982). The e-folding length is fixed to 130 m as found by fitting the Honjo et al. (1982) equation to the EqPac sediment trap data (Murray et al., 1996). In the first scheme (referred to as the PON simulation in the following), the export production is only made by sinking particles and is equivalent to new production, as adopted by Eppley and Peterson (1979). In the second scheme (referred to as the DON/PON simu- lation) , the production of dissolved organic matter is also taken into account, by considering that only a fraction (a) of new production is remineralized below the photic layer as sinking particles and that the re- mainder (1 - a) is transformed into dissolved organic matter in the photic layer. Alpha (a) was taken to be equal to 50%, according to Yamanaka and Tajika (1997). We assume that the consumption of the DON pool can be modelled as a .decay process with an e-folding time of 180 days in agreement with Archer et al. (1997).

Initial conditions for the nitrate field are taken from the Levitus climatological nitrate field (Levitus et al., 1983). This initial nitrate field, and the physical fields given by the OGCM, are not in perfect agree- ment. To remedy this, we first ran the coupled model

Modelled new production in the equatorial Pacific 3 13

I'

for 5 years, using the OGCM outputs form April 1992 to April 1993. After this 'spin-up' phase, the model recovers from the perturbation caused by the incone sistency between the initial NO3 conditions and the model dynamics. However, equilibrium is not reached completely. The evolution of integrated nitrate over the entire basin and in the upper lOOm still shows a trend of the order of a 3% increase after 5 years of spin-up, as is the case in most similar models (Sare miento et al., 1993; Chai et al., 1996). Given this small increase, the model is considered to be in quasi-equi- librium.

RESULTS

Comparison of model results and observations The FLUPAC cruise that took place in September- October 1994 in the westem Pacific at 165"E (France-JGOFS), and the CGC94 WOCE cruise in March-April 1994 in the eastern Pacific at 1OO"W, collected samples in contrasting regions: in the warm pool region with warm, low-salinity and nutrient- depleted waters, and in the upwelling region char- acterized by cool, high-salinity and nutrient-rich water, respectively.

In the eastern meridional section at 11O"W (Fig. ld-f), nitrate is depleted north and south of the upwelling region and the nutricline is prominent. At the equator, the spreading of nitrate isopleths is asso- ciated with the Equatorial Under-Current (EUC). The nitrate distribution given by the two versions of the model is in good agreement with the data, in terms of both spatial structure and of values, and the nutricline is particularly well reproduced by the model. However the PON simulation produces an accumulation of ni- trate below the nutricline at 3"N and 3"s ('nutrient trapping': Najjar et al., 1992). The DON/PON simu- lation results in a decrease of the subsurface nitrate concentration at these latitudes, which is in better agreement with observations than in the PON simu- lation.

In the warm pool, the surface mixed layer was de- pleted of nitrate all along the transect including the region close to the equator (Fig. la). Modelled nitrate concentration in the DON/PON simulation (Fig. IC) is in closer agreement with observed distributions (Fig. la) than in the PON simulation (Fig. lb), which produces a deeper nutricline and lower nitrate con- centrations than observed.

In the DON/PON simulation, most of the DON is produced in the upwelling region of the eastern Pacific where advection of nutrient-rich waters into the sur- face layers is important. The DON is then advected

westward in the surface layers via the westward-flow- ing South Equatorial Current (SEC) while part of this DON is exported poleward through the equatorial divergence in less than the e-folding time (180 days) at which DON is remineralized (Yamanaka and Taj- ika, 1997). Hence, some of the production (DON) is exported zonally rather than being entirely reminer- alized below the photic layer as in the PON-only simulation. This decreases the nutrient trapping in the eastern Pacific (Bacastow and Maier-Reimer, 1991; Najjar et al., 1992). The decay time for DON is such that a part of the westward flux of DON by the SEC is remineralized in the warm pool region, thus increasing nitrate at depth in the western Pacific. Consequently, the simulation is improved both in the west, where nitrate concentrations below the nutricline are en- hanced, and in the east, where nitrate in excess below the nutricline in the upwelling region (i.e. the nutri- ent trapping) is removed.

Interannual variability of new production We estimated new production using our model, by integrating equation 2 from 5"N to 5"S, 80"W to 180", and from April 1992 to March 1995. The two esti- mates, with the PON, and the DON/PON simulations, do not differ greatly. They are compared with those from other authors in Table 1. Our new production estimates are similar to the two estimates based on field measurements (Chavez and Barber, 1987; Ku et al., 1995) as well as to two estimations by models (Toggweiler and Carson, 1995; Chai et al., 1996). Another estimate by Chavez and Barber (1987; 6 mmol N m-2 day-') using a mean nitrate concentration and the upwelling index derived by Wyrtki (1981), and assuming that the upwelled nitrate was completely taken up by the phytoplankton, seems to be too high compared with our estimates and those of others (Table 1). This is partly because the Chavez and Barber (1987) estimate ignores the losses of nitrate by physical processes, including tropical instability waves that generate intense downwelling, which export ni- trate prior to its assimilation.

In the model, a well-marked salinity front at the eastern edge of the warm pool separates the fresh and warm waters (28°C) of the warm pool from the higher salinity and colder waters of the upwelling region (Fig. 2b). The longitude of the salinity front varies with time. Picaut et al. (1996) have shown that these displacements were controlled by zonal advection in relation to ENS0 wind-driven surface current varia- tions. In Fig. 2(a), the 34.8 isohaline representative of the salinity front is superimposed on the simulated equatorial zonal current. Globally, an eastward shift of

O 1998 Blackwell Science Ltd., Fish. Oceanogr., 7:3/4, 311-316.

314 A. Stoens et al.

Figure 1. Comparison of meridional vertical sections a t 165"E and 110"W between nitrate observations and modelled nitrate (pmol kg-'). (a) Observed nitrate at 165"E during FLUPAC, (b) modelled nitrate at 165"E from PON simulation, (c) modelled nitrate at 165"E from DON/PON simulation, (d) observed nitrate at ll0"W during CGC94, (e) modelled nitrate at 11O"W from PON simulation, (f) modelled nitrate at 110"W from DON/PON simulation. The thick and dashed lines indicate the bottom of the photic layer (1% incident light) and the bottom of the mixed layer, respectively.

(a) NITRATE FLUPAC 165E (b) PON SIMULATION

40

60

5 80 s

- Q ; 100

120

140

160

40

60

80

1 O0

120

140

160

-10 -5 O 5 -10 -5 O 5 Latitude Latitude

(d) NITRATE CGC 11 OW (e) PON SIMULATION

-20 -10 o 10 20 -20 -10 o 10 20 Latitude Latitude

(c) DON/PON SIMULATION

20

401 I . I \ Ø.- l 60

E - 80 -f:

h

Q 8 100

120

140

160

-10 -5 O 5 Latitude

(f) DON/PON SIMULATION

50

1 O0

E - 150 5

- Q

200

250

300

-20 -10 o 10 20 Latitude

the salinity front is linked to an eastwards-flowing current and a westward shift to a westwards-flowing current. The simulation shows that the 'new produc- tion front' closely follows the salinity front (Fig. 2c). The salinity front thus separates the biologically poor, warm pool region to the west (nitrate-depleted surface layer) from the equatorial cold tongue of upwelled water further to the east. Although the period under investigation is limited to 1992-1995, it is clear from this simulation that the line that delineates the nu- trient-poor pool from the rich pool can be zonally displaced over thousands of kilometres in phase with

O 1998 Blackwell Science Ltd., Fish. Oceanogr., 7:3/4, 311-316.

ENSO. Hence, in the western Pacific, the width of the productive area varies in time according to interan- nua1 ENS0 variability, whereas in the eastern Pacific, the seasonal cycle strongly controls new production (Fig. 2c). Lehodey et al. (1997) have shown that the migrations of the skipjack tuna (Katsuwonus pelamis) population are also correlated to this ENSO-related zonal displacement of the warm pool. In Lehodey et al. (1998, this issue), our modelled new production is used to force numerical simulations of skipjack forage abundance, and to study the coupling between tuna forage biomass and the skipjack fisheries.

Modelled neu broduction in the equatorial Pacific 315

Table 1. Estimates of new production from previous studies and the present study in the 9OoW-18O"W, 5"N-5"S box.

Value I Source (mmol N m-' day-')

,? Chavez and Barber (1987) using Wyrtki's (1981) upwelling rate of 50 Svedrup Chavez and Barber (1987) using primary production and f ratio of 0.44 Chai et al. (1996) using ecosystem model in the upper 120 m Ku et al. (1995) using the '''Ra budget in the euphotic zone (EqPac)

Present study: DON simulation (1992-1995 average) Present study: DON/PON (1992-1995 average)

6.0 2.7 2.3 2.5 3.3 2.75 2.96

Toggweiler and Carson (1995) using ecosystem model

CONCLUSION The dynamical model forced by stresses derived from

The main biogeochemical features observed in the eastern and western Pacific are well reproduced by the model and improved by the addition of an explicit compartment of DON. The model shows that dis- solved organic nitrogen is advected from the eastern Pacific equatorial regions toward the western Pacific where it is injected below the photic layer. The av- erage modelled new production is estimated to be 2.96 mmolN m-2 day-' in the 9O"W-18O0W, 5ON-5"S region and is in agreement with previous estimates from both observations and models.

L

944 Figure 2. Equatorial section of the coupled model surface outputs during April 1992 and June 1995. (a) Zonal velocity component (contour interval: 0.4 m s-l, shaded areas correspond to eastward-flowing current) with superimposed 34.8 isohaline, (b) sea surface salinity (contour interval: 0.4), (c) new production (contour interval: 1 mmol N m-2 day-') with superimposed 34.8 isohaline.

the ERS1 scatterometer is especially successful in re- producing the position, and ENSO-related zonal dis- placements, of the salinity front. These displacements are associated with a nitrate and new production front that separates the nutrient-exhausted warm pool waters from the richer waters in the equatorial cold tongue of upwelled water further east. This ENSO-related zonal displacement of the front has large biogeochemical implications for the global COZ cycle because the front separates a warm pool region, thought to be in global equilibrium with the atmosphere in terms of COZ par- tial pressure, from a region that is the largest marine

(al ZONAL CURRENT \ -

Lo cn cn

d- cn cn 7

M 0, cn T-

a cn cn 7

n

150"E 170"E 170"W 1 5OoW 130"W 1 1 O'W

(b) SURFACE SALINITY (cl NEW PRODUCTION m cn cn 7

d- cn cn

c\l cn cn

150"E 170"E 170"W 150"W 130"W 1 10"W

Q 1998 Blackwell Scien

150"E 170'E 170"W 150'W 130"W 1 10"W

ce Ltd., Fish. Oceanogr., 7:3/4, 311-316.

Lo cn m l--

d- cn cn 7

M 0 2 m 7

(\I m

316 A. Stoens e t al.

source of COZ to the atmosphere (Tans et al., 1990). This ENSO-related zonal displacement of the warm and nutrient-poor pool has a second important implication for the marine ecosystem, because the displacements of tuna forage and of tuna concentrations, are shown to be closely linked to zonal shifts of the front (Lehodey et al., 1998). This simple model that simulates new produc- tion, the distribution of nitrate, and DON (i.e. the main substrate for the growth of bacteria) is thus a powerful tool for enhancing the understanding of other aspects of the ecosystem.

REFERENCES Archer, D., Peltzer, E.T. and Kirchman, D.L. (1997) A timescale

for dissolved organic *carbon production in the equatorial Pacific surface waters. Global Biogeochem. Cycl. 3:435-452.

Bacastow, R. and Maier-Reimer, E. (1991) Dissolved organic carbon in modeling of oceanic new production. Global Biogeochem. Cycl. 5:71-85.

Blanke, B. and Delecluse, P. (1993) Variability of the tropical Atlantic Ocean simulated by a general circulation model with two different mixed-layer physics. J. Phys. Oceanogr. 23:1363-1388.

Buesseler, K.O., Andrew, J.A., Hartman, M.C., Belastock, R. and Chai, F. (1995) Regional estimates of export flux of particulate organic carbon derived from thorium-234 during the JGOFS EqPac program. DeepSea Res. II 42:777-804.

Chai, F., Lindley, S.T. and Barber, R.T. (1996) Origin and maintenance of a high nitrate condition in the equatorial Pacific. Deep-sea Res. II 43:1031-1064.

Chavez, F.P. and Barber, R.T. (1987) An estimate of new pro- duction in the equatorial Pacific. Deepsea Res. 34: 1229-1243.

Chavez, F.P., Service, S.K. and Buttrey, S.E. (1996) Tempera- ture-nitrate relationships in the central and eastern tropical Pacific. J. Geophys. Res. 101:20553-20563.

Dandonneau, Y. (1992) Surface chlorophyll concentration in the tropical Pacific Ocean: an analysis of data collected by merchant ships from 1978 to 1989. J. Geophys. Res. 97:3581-3591.

Dugdale, R.C. and Goering, J.J. (1967) Uptake of new and re- generated forms of nitrogen in primary productivity. Limnol. Oceanogr. 12: 196-206.

Eppley, R.W. and Peterson, B.J. (1979) Particulate organic matter flux and planktonic new production in the deep ocean. Nature 282:677-680.

Grima, N., Bentamy, A., Delecluse, P., Katsaros, K., Levy, C. and Quilfen, Y. (1998). Sensitivity study of an Oceanic General Circulation Model forced by satellite wind-stress fields. J. Geophys. Res. (in press).

Honjo, S., Manganini, S.J. and Cole, J.1, (1982) Sedimentation of biogenic matter in the deep ocean. DeepSea Res. 29: 609-625.

Ku, T.L., Luo, S., Kusakabe, M. and Bishop, J.K.B. (1995) zz8Ra- derived nutrient budgets in the upper equatorial Pacific and the role of "new" silicate in limiting productivity. DeepSen

Lehodey, P., Bertignac, M., Hampton, J., Lewis, A. and Picaut, J. (1997) El Niño Southem Oscillation and tuna in the western Pacific. Nature 389:715-717.

Res. Ir 42:479-497.

Lehodey, P., Andre, J.-M., Bertignac, M., Hampton, J., Stoens, A., Menkes, C., Mémery, L. and Grima, N. (1998) Pre- dicting skipjack tuna forage distributions in the equatorial Pacific using a coupled dynamical bio-geochemical model.

Levitus, S., Reid, J.L., Conkright, M.E. and Najjar, R. (1983) Distribution of nitrate, phosphate and silicate in the world ocean. Progr. Oceanogr. 31:245-273.

Libby, P.S. and Wheeler, P.A. (1997) Particulate and dissolved organic nitrogen in the central and eastem equatorial Pa- cific. Deep-sea Res. I 44345-361.

LUO, S., Ku, T.-L., Kusakabe, M., Bishop, J.K.B. and Yang, Y.L. (1995) Tracing particle cycling in the upper ocean with 23?h and 228Th: an investigation in the equatorial Pacific along 140"W. DeepSea Res. II 42:805-829.

Morel, A. (1991) Light and marine photosynthesis: a spectral model with geochemical and climatological implications. Prog. Oceanogr. 26:263-306.

Morel, A. and Berthon, J.F. (1989) Surface pigments, algal biomass profiles, and potential production of the euphotic layer: relationships reinvestigated in view of remote-sensing applications. Limnol. Oceanogr. 34:1545-1562.

Moutin, T., and Coste, B. (1996) Campagne océanografique OLIPAC à bord du N.O. l'Atalante, 3 Novembre-1 Dé- cembre 1994, Recueil de données (Data report). Centre d'océanologie de Marseille, 266 pp.

Murray, J.W., Young, J., Newton, J., Dunne, J., Chapin, T., Paul, B. and Mc Carthy, J.J. (1996) Export flux of particulate organic carbon from the central equatorial Pacific deter- mined using a combined drifting approach. Deep-sea Res. II 43:1095-1132. I

Najlar, R.G., Sarmiento, J.L. and Toggweiler, J.R. (1992) Downward transport and fate of organic matter in the ocean: ( simulations with a general circulation model. Global Biogeochem. Cycl. 645-76.

Picaut, J., Ioualalen, M., Menkes, C., Delcroix, T. and McPha- den, M.J. (1996) Mechanism of the zonal displacements of the Pacific warm pool. Science 274:1486-1489.

Sarmiento, J.L., Slater, R.D., Fasham, M.J.R., Ducklow, H.W., Toggweiler, J.R. and Evans, G.T. (1993) A seasonal three- dimensional ecosystem model of nitrogen cycling in the North Atlantic euphotic zone. Global Biogeochem. Cycl. 7:417-450.

Tans, P.P., Fung, I.Y. and Takahashi, T. (1990) Observational constraints on the global atmospheric CO2 budget. Science 247:143 1-1438.

Toggweiler, J.R. and Carson, S. (1995) What are upwelling systems contributing to the ocean's carbon and nutrient budgets. In: Upwelling in the Ocean: Modern Processes and Ancient Records. C.P. Summerhayes, M.V. Angel, R.L. Smith and B. Zeitzschel (eds). New York: John Wiley.and Sons Ltd, pp. 337-360.

Wyrtki, K. (1981) An estimate of equatorial upwelling in the Pacific. J.Phys. Oceanogr. 11:1205-1214.

Yamanaka, Y. and Tajika, E. (1997) Role of dissolved organic matter in the marine biogeochemical cycle: studies using an ocean biogeochemical general circulation model. Global Biogeochem. Cycl. 11:599-612.

Yoder, J.A., McClain, C.R., Feldman, G.C. and Esaias, W.E. (1993) Annual cycles of phytoplankton chlorophyll con- centrations in the global ocean: a satellite view. Global

Fish. Oceunogr. 7:317-325. 1

i

8

I

-.

k

Biogeochem. Cycl. 7:181-193. L.

O 1998 Blackwell Science Ltd., Fish. Oceunogr., 7:3/4, 311-316.