Marie-Helene Radenac Æ Yves Dandonneau

Bruno Blanke

Displacements and transformations of nitrate-rich and nitrate-poorwater masses in the tropical Pacific during the 1997 El Nino

Received: 31 August 2004 / Accepted: 24 January 2005 / Published online: 24 February 2005� Springer-Verlag 2005

Abstract A Lagrangian analysis was applied to theoutputs of a coupled physical-biogeochemical model todescribe the redistribution of nitrate-rich and nitrate-poor surface water masses in the tropical Pacificthroughout the major 1997 El Nino. The same tool wasused to analyze the causes of nitrate changes along tra-jectories and to investigate the consequences of the slownitrate uptake in the high nutrient low chlorophyll(HNLC) region during the growth phase of the event.Three patterns were identified during the drift of watermasses. The first mechanism is well known along theequator: oligotrophic waters from the western Pacific areadvected eastward and retain their oligotrophic prop-erties along their drift. The second concerns the persis-tent upwelling in the eastern basin. Water parcels havecomplex trajectories within this retention zone and re-main mesotrophic. This study draws attention to thethird process which is very specific to the HNLC regionand to the El Nino period. During the 1997 El Nino,horizontal and vertical inputs of nitrate decreased sodramatically that nitrate uptake by phytoplankton be-came the only mechanism driving nitrate changes alongpathways. The study shows that because of the slownitrate uptake characteristic of the tropical Pacific

HNLC system, nitrate in the pre-El Nino photic layercan support biological production for a period of severalmonths. As a consequence, the slow nitrate uptake de-lays the gradual onset of oligotrophic conditions overnearly all the area usually occupied by upwelled waters.Owing to this process, mesotrophic conditions persist inthe tropical Pacific during El Nino events.

Keywords Nitrate uptake Æ Tropical Pacific ÆEl Nino Æ Lagrangian analysis

1 Introduction

Nitrate-rich waters of the equatorial divergence in theeastern tropical Pacific are surrounded by oligotrophicregions: the warm pool to the west and the subtropicalgyres north and south. In the photic layer of the equa-torial divergence, the phytoplankton biomass is lowrelative to available nitrate concentrations. High nutri-ent low chlorophyll (HNLC) conditions persist becausethe ecosystem is iron-limited and grazing-balanced(Landry et al. 1997). Biologically available iron in thephotic layer is mainly upwelled from the equatorialundercurrent (EUC) (Coale et al. 1996). The overallconsequences of such an ecosystem are that nitrate up-take is kept at a very low rate (Price et al. 1994) and thata large amount of surface nitrate remains unused insurface water (Coale et al. 1996; Landry et al. 1997). Thenorthern limit of the enriched waters of the equatorialdivergence is sharply defined around 5�N (Wyrtki andKilonsky 1984; Bender and McPhaden 1990). The sur-face waters of the tropical part of the north Pacific gyreare nitrate-depleted with low phytoplankton biomass(Wyrtki and Kilonsky 1984; Hardy et al. 1996). Thedeep chlorophyll maximum is close to the nitracline atabout 100 m depth (Venrick et al. 1973; Levitus et al.1993; Karl and Lukas 1996). South of the HNLC watersof the equatorial divergence, the tropical gyre is proba-bly among the most undersampled regions. Cruise

Communicated by Andreas Oschlies

M.-H. Radenac (&)Laboratoire d’Etudes en Geophysique et OceanographieSpatiale CNRS-IRD-UPS-CNES, 14 avenue Edouard Belin,31401 Toulouse cedex 9, FranceE-mail: marie-helene.radenac@cnes.frTel.: +33-561-333000Fax: +33-561-253205

Y. DandonneauIPSL, Laboratoire d’Oceanographie Dynamiqueet de Climatologie CNRS-IRD-UPMC, case 100,4 place Jussieu, 75252 Paris cedex 05, France

B. BlankeLaboratoire de Physique des Oceans, UFR Scienceset Technique CNRS-IFREMER-UBO,6 avenue Le Gorgeu, 809, 29285 Brest, France

Ocean Dynamics (2005) 55: 34–46DOI 10.1007/s10236-005-0111-5

surveys have shown a deep nitrate-depleted surface layerwith low chlorophyll concentration (Dandonneau 1979;Wyrtki and Kilonsky 1984; Hardy et al. 1996; Rai-mbault et al. 1999). Nutrient-rich surface water as farsouth as 12�S is a frequent feature (Wyrtki and Kilonsky1984; Dandonneau and Eldin 1987; Bender andMcPhaden 1990; Raimbault et al. 1999) and the transi-tion between the mesotrophic and oligotrophic waters ismore gradual than at the northern limit.

West of the equatorial divergence, the transition to-ward the warm pool is marked by a sharp salinity frontwhile the sea surface temperature increases smoothly(Kuroda and McPhaden 1993; Eldin et al. 1997). This isalso the site of transition between the HNLC and oli-gotrophic equatorial ecosystems producing a disconti-nuity of chemical and biological properties such asnutrients, chlorophyll, zooplankton biomass, pCO2, andphytoplankton community structure (Inoue et al. 1996;Eldin et al. 1997; Boutin et al. 1999; Le Borgne et al.2002; Kobayashi and Takahashi 2002). Mean zonalcurrents approach zero in this region of convergence(Picaut et al. 1996). Warm pool waters, defined bytemperature greater than 29�C (McPhaden and Picaut1990), are nitrate-depleted with very low chlorophyll andnew production. The nitracline, the chlorophyll maxi-mum and a weak new production maximum are closelyassociated with the thermocline at a depth of around100 m (Mackey et al. 1995; Radenac and Rodier 1996;Navarette 1998).

Biogeochemical conditions in the tropical Pacificchange drastically during El Nino events. These changesare generally understood to be a significant reduction inprimary and export production in the entire equatorialcold tongue (Barber and Chavez 1983; Strutton andChavez 2000). Measurements relying on oceanographiccruises, merchant ships, moorings, as well as modelstudies have revealed changes in the east-west equatorialasymmetry of nutrients and biology. Nevertheless, it wasduring the 1997–1998 event that this impact was firstvisualized on the tropical Pacific basin scale owing to thetimely launches of new ocean color sensors (Fig. 1). ThePolarization and Directionality of the Earth Reflec-tances (POLDER) (Deschamps et al. 1994) and OceanColor and Temperature Scanner (OCTS) sensors aboardthe ADEOS satellite provided data (November 1996–June 1997) before and during the early phase of theevent. The Sea-viewing Wide Field-of-view Sensor(SeaWiFS) (McClain et al. 1998) mission started inSeptember 1997 and captured the warm phase. Figure 1summarizes the evolution of the 1997 El Nino event asobserved by ocean color sensors. Before the onset of theevent, weak La Nina conditions prevailed in the tropicalbasin in late 1996 (Fig. 1a) and chlorophyll rich waters(Muramaki et al. 2000; Radenac et al. 2001; Ryan et al.2002) extended across most of the equatorial basin. InMay 1997, the satellite chlorophyll (Fig. 1b) had fallenby about 50% since November 1996 and reached about0.1 mg m�3 in the western part of the equatorial coldtongue. East of 160�W however, the surface chlorophyll

remained close to the November 1996 values (about0.2 mg m�3). During the peak of the event, fromNovember 1997 to January 1998 (Fig. 1c), more than6 months after the El Nino event started, the chloro-phyll-rich region was reduced to its narrowest zonalextent (Radenac et al. 2001) and high chlorophyll watersstretch from the Central American coast to the equato-rial western basin (Murtugudde et al. 1999). Unfortu-nately, there was a three-month gap (July 1997–September 1997) in the time series of ocean color datathat corresponded to an essential transition phase of theonset of El Nino.

What were the processes that led to the December1997 biological situation as revealed by ocean colorsensors in the tropical Pacific (Fig. 1c)? Most previousstudies focused on the equatorial zone. Observationsand simulations have shown that the variability of ni-trate, chlorophyll or new production is mainly con-trolled by the fast equatorial dynamics (Chavez et al.1998; Stoens et al. 1999; Friedrichs and Hofmann 2001;

Fig. 1 Modeled surface velocity (vectors, m s�1) superimposed onsatellite chlorophyll (color scale, mg m�3). The November 1996 (a)and May 1997 (b) scenes are POLDER derived chlorophyll, theDecember 1997 panel (c) is SeaWiFS chlorophyll. The white solidline corresponds to 0.1 mg m�3 . The white dashed line in Decemberis the leading edge of eastward spreading ’pre El Nino’oligotrophicwaters

35

Radenac et al. 2001) and their analyses stressed theabrupt decrease of primary production. They did notclosely examine the off-equatorial situation for which wehypothesize here that the low nitrate uptake rate char-acteristic of the HNLC ecosystem should delay the set-ting up of oligotrophic conditions that characterize ElNino. In this study, a Lagrangian analysis (Blanke andRaynaud 1997) was applied to the outputs of a physical-biogeochemical model (Radenac et al. 2001) to addressthe following two issues. First, the goal was to describethe redistribution of surface water masses in the tropicalbasin during the 1997 El Nino and secondly, to inves-tigate the role of the slow nitrate uptake in the HNLCregion during the growth phase of the event.

2 Numerical tools and data

2.1 The physical-biological model

The primitive equation general circulation model(GCM) (Maes et al. 1997; Vialard et al. 2001) covers thetropical Pacific between 120�E and 75�W and between30�N and 30�S. The zonal resolution is 1� and themeridional resolution, close to 0.5� between 5�N and5�S, increases to 2� at the northern and southernboundaries. The vertical resolution is nearly 10 m overthe first 120 m in depth and increases downward. Windstress data derived from the ERS1–2 scatterometer areused to force the model during the 1993–1998 period.Estimates of the heat and freshwater fluxes are com-puted from the 1979–1993 seasonal cycle of the EC-MWF reanalysis. The modeled heat flux isparameterized following Vialard et al. (2001). The longterm SST drift is counterbalanced by a �40 W m�2 K�1

relaxation (with a relaxation timescale of 50 days for a40-m layer) toward the observed Reynolds and Smith(1994) sea surface temperature. A correction term de-rived as by Vialard et al. (2002) is also applied to thefreshwater flux to avoid the sea surface salinity driftcaused by the forcing freshwater flux.

Five-day outputs of the ocean circulation model(zonal velocity u, meridional velocity v, vertical velocityw, vertical diffusion coefficient Kz) and the same spatiallyuniform horizontal eddy coefficient Kh as in the GCMforce the nitrate advection-diffusion equation:

@tNO3 ¼� u@xNO3 � v@yNO3 � w@zNO3 þ KhDh NO3ð Þþ @z Kz@zNO3ð Þ þ S ð1Þ

in which the left-hand side represents the local nitratechange. The first three terms on the right-hand side arethe zonal, meridional and vertical advection, respec-tively. The fourth and fifth terms are the horizontaldiffusion (parameterized by the horizontal Laplacianoperator D h) and the vertical diffusion.

In the euphotic layer, the biological model consists ofa nitrate uptake S (new production) that is biomass-dependant (Price et al. 1994; Landry et al. 1997):

S ¼ �VmaxNO3

NO3 þ KNO3

PUR

PURþ KEChl½ � ð2Þ

in which both the nitrate and light limitations terms areexpressed in Michaelis-Menten form. The photosyn-thetic usable radiation (PUR) is deduced from the shortwave downward radiation as explained by Stoens et al.(1999) and KE (70·10�6 mol photon m�2 s�1) is the halfsaturation constant for PUR. The maximum nitrateuptake rate Vmax (3·10�3 lmol N mg Chl�1 s�1) andthe half saturation concentration KNO3 (0.01 lM) havebeen adjusted in such a way that modeled nitrate fieldsagree the best with concurrently observed vertical sec-tions of nitrate content. Because this model does notexplicitly resolve variations in biomass, a calculationspecific to the tropical Pacific is applied to chlorophyllprofiles empirically derived from modeled nitrate at thesurface (Stoens et al. 1999). New production is locallyexported (exponential shape as by Honjo 1978, with a120-m length scale) below the photic layer and isinstantaneously remineralized into nitrate. Details onthe choice of constants and rationale of the biologicalmodel are given in Stoens et al. (1999) and Radenacet al. (2001).

2.2 The Lagrangian tool

The Lagrangian approach described in detail by Blankeand Raynaud (1997) is applied to the physical-biologicalmodel outputs. Streamlines are computed from the ar-chived, three-dimensional velocity field (5-day runningmeans). Assuming that the velocity is stationary overeach successive sampling period, consecutive portions ofstreamlines represent the trajectories of particles. Prop-erties along trajectories are interpolated from modeledthree-dimensional fields. Both direct and reverse (findingthe origin of water mass) experiments can easily beperformed because it is an off-line technique.

In a Lagrangian form, the equation of conservationof nitrate is:

dtNO3 ¼ KhDh NO3ð Þ þ @z Kz@zNO3ð Þ þ S ð3Þ

in which the nitrate change along the trajectory (dtNO3)corresponds to the effect of the horizontal and verticaldiffusions and of the nitrate sink S (Eq. 2). For off-linediagnostics, this relation is still valid if we assume thatunsampled tracer and velocity fluctuations are smallrelative to their mean values on the time scale of themodel outputs. Previous studies using a similar model(Maes et al. 1997; Lehodey et al. 1998; Stoens et al. 1999;Vialard et al. 2001; Radenac et al. 2001) have shown thatfive-day outputs are relevant for capturing the large-scale variability in the tropical Pacific.

The December 1997 (Fig. 1c) situation was chosen torepresent the peak of the 1997 El Nino and the May1997 situation (Fig. 1b) as the reference for the earlystage of the event. At that time, the surface circulation

36

characteristic of El Nino had clearly emerged while thedisruption of the east-west equatorial trophic contrastwas just starting. Our Lagrangian analyses refer to thesesituations. The aim is to investigate the redistribution oftropical surface water masses and to explain nitratevariations along specific trajectories.

2.3 Identification of water masses

The domain was partitioned into six surface watermasses (Table 1) and one subsurface water mass essen-tially based on the nitrate distribution in May 1997. A1 lM nitrate threshold discriminates oligotrophic watersfrom nitrate-rich waters. Among the oligotrophic sur-face waters, an additional temperature criterion(T‡29�C) was used to identify warm pool water andgeographical limits bounded the northwest subtropicalwater, the northeast subtropical water and the southsubtropical water. Surface nitrate-rich waters haveconcentration between 1 lM and 12 lM. Geographicalcriteria separated the equatorial upwelled water and thecoastal waters of Central and South America. Althoughtwo nitrate criteria are used to define subsurface waters,one single class represents them. Subsurface watersunderlying nitrate-rich surface waters have nitrate con-centration higher than 12 lM and they have nitrateconcentration higher than 1 lM beneath oligotrophicsurface waters. These water masses will be denominatedbelow according to their status in May 1997, even if theirnitrate content (oligotrophic/mesotrophic) changesduring the course of the El Nino.

2.4 Validation data

Three existing data sets were used to validate the mod-eled surface circulation in 1997. The first one consistedof currents data gathered twice a year during cruisesundertaken to maintain the tropical atmosphere ocean/triangle trans ocean buoy network (TAO/TRITON)moorings between 165�E and 95�W (Hayes et al. 1991;McPhaden et al. 1998). During these cruises, ADCPswere operated continuously along meridional sections.

The processing and gridding of these data have beendescribed in detail by Johnson et al. (2000). In the sec-ond data set (Ocean Surface Current Analysis—Realtime, OSCAR), near-surface currents were derived fromsatellite altimeter, scatterometer and sea surface tem-perature (Lagerloef et al. 1999; Bonjean and Lagerloef2002). In this product, the 30-m surface layer current isthe sum of geostrophic and Ekman currents and of abuoyancy term. It encompasses the period since October1992 on a 1�·1� grid with a 10-day temporal resolution.The last set of data used for of the 6-h interpolatedtrajectories was collected from the satellite-trackeddrifting buoys by the Atlantic oceanographic andmeteorological laboratory (AOML) (Pazan and Niiler2004). These data are available on-line (http://www.aoml.noaa.gov/phod/dac/dacdata.html) and theirprocessing has been described by Hansen and Poulain(1996).

3 Results

3.1 Surface circulation

The ‘climatology’ (not shown) that we derived from themodeled 15 m currents averaged over the 1993–1996period (the 1997–1998 ENSO years were excluded be-cause of the strong anomalies present at that time) isvery similar to the one presented by Vialard et al. (2001;their Fig. 5). Consequently, the reader is referred toVialard et al. (2001) and Radenac et al. (2001) for avalidation of the mean modeled surface currents. Thesalient contrast between modeled surface circulationbefore and during the 1997 El Ninois illustrated inFig. 1. The time evolution of the zonal current along155�W as derived from satellites and from the model isshown in Fig. 2. Snapshots of zonal velocity measuredduring the pre-El Nino (December 1996) and El Nino(May 1997 and November 1997) TAO/TRITON cruisesare also part of the validation data set (Fig. 2c).

As was expected, the modeled south equatorial cur-rent (SEC) during cold conditions in late 1996 wasstronger than average and spread well past the date line.The north equatorial countercurrent (NECC) was strong

Table 1 Criteria used to define surface water masses in May 1997: four classes of oligotrophic water with NO3 £ 1 lM and three classes ofnitrate-rich waters

Surface water mass NO3 (lM) T (�C) Latitude Longitude(deg. E)

Warm pool WPW £ 1 ‡29 £ 240North-west subtropical NWSW £ 1 <29 ‡0 £ 180North-east subtropical NESW £ 1 <29 ‡0 >180South subtropical SSW £ 1 <29 <0Equatorial upwelling UPW 1<NO3 £ 12 �12 £ lat £ 10Coastal upwelling CW 1<NO3 £ 12 lat>10 and lat<�12 >180

Note that subsurface waters underlying nitrate-rich surface waters have nitrate concentration higher than 12 lM and that they havenitrate concentration higher than 1 lM beneath oligotrophic surface waters

37

and extended across the basin. These features are con-sistent with surface currents measured during cruises ordeduced from satellite data (Lagerloef et al. 1999;Johnson et al. 2000). Between the boreal fall of 1996 andthe spring of 1997, the NECC shifted southward andgrew in strength (Fig. 2). Its zonal velocity peaked in thefall of 1997. At the equator, the current was mostlyeastward from March to the end of the year, except inJuly and August. These eastward currents merged into alarge eastward flow between 6�N and 2�S (Johnson et al.2000). The model often failed to reproduce the bimodalstructure of the eastward flow during the mature phase(see the November 1997 meridional profile). Actually,

the eastward current at the equator was accuratelysimulated but the NECC was too weak. In November,the maximum zonal current measured during the cruiseor derived from satellite data was close to 1 m s�1,which is about twice the modeled velocity. Besides, acomparison with the TAO zonal current (not shown)shows that the model tended to overestimate the dura-tion and westward extent of the brief-living SEC in July-August, while the satellite derived currents seem to havemissed part of this event.

Trajectories of satellite-tracked drifters integrate thedifferent circulation events that occurred during the 1997El Nino. They resolve circulation features at a highspatial resolution of which the GCM derived trajectoriesare not capable. Nevertheless, their comparison revealsinteresting and consistent large-scale patterns. Drifterspresent between 15�S and 15�N in May 1997 were se-lected and their displacements followed until December1997 (Fig. 3a). The coverage is rather sparse except inthe near equatorial region and in the southern hemi-sphere west of 140�W. To compare these observed tra-jectories with the modeled circulation, virtual drifters

Fig. 2 Zonal velocity component (m s�1) along 155�W betweenNovember 1996 and December 1997. a Time evolution of thesatellite-derived surface zonal current (OSCAR). b Time evolutionof the modeled zonal current at 15 m. The contour interval is0.25 m s�1 . Positive values are shaded. The vertical lines representthe December 1996, May 1997 and November 1997 cruises. cMeridional profiles of ADCP zonal current measured at 15 mduring TAO/TRITON cruises along 155�W in December 1996(small-dashed line), May 1997 (solid line) and November 1997 (long-dashed line)

Fig. 3 Comparison of a in situdrifters and b modeledtrajectories between May andDecember 1997. Colors are1-month intervals between thebeginning (May) and the end(December) of trajectories

38

were released in the modeled velocity field at the locationof the in situ drifting buoys (Fig. 3b): in the northequatorial current (NEC) region, in the near equatorialzone and in the south and south west regions. All sim-ulated trajectories begin in early May 1997 and areintegrated until December 1997 at a constant depth(15 m) in order to better represent the 15 m deep cir-culation followed by in situ drifters (Niiler 2001).

The observed trajectories that start between 10�N and15�N had roughly eastward drifts and remained withinthis latitude band. In the model, the zonal component ofthe NEC was too weak and, as a result the virtualdrifters had too strong a poleward drift. A comparisonwith the Reverdin et al. (1994) climatology confirms thepoleward trend of modeled currents in this region. Ninein situ drifters released in the near equatorial zone (3�S–1�N) were caught in the strong eastward surface flowthat dominates over most of the equatorial basin in Mayand June. They separated into two groups when the SECresumed in July. In the northern group, two of thesedrifters rapidly travelled eastward following the strongNECC. This pathway was well represented by themodel. Virtual drifters released at 150�W in May werearound 3.6�N, 117�W in October, which is close to theposition reached by in situ drifters (3�N, 120�W). Theremaining seven drifters formed the southern group thatwas swept southward in July. In this group, the modelreproduced the timing of the direction changes. While insitu drifters travelled 4–6 degrees further south in thevicinity of 160�W than virtual drifters, in situ and virtualdrifters had about the same southward drift along140�W. These results suggest that the modeled south-ward velocity has been underestimated in the centralwestern Pacific. However, measured and modeledmeridional velocities during the fall 1997 are in agree-ment. Whereas the measured meridional component isgreater than 0.1 m s�1 along 155�W south of 4�S andsmaller further east, the modeled meridional componentalways ranges between 0 m s�1 and 0.05 m s�1. Further

east, the three in situ buoys that started in the equatorialzone near 100�W remained in the eastern basin. In themodel, some of the drifters released at 100�W mean-dered in the complex circulation patterns in the easternpart of the upwelling. The remainder travelled south. Insitu drifters located in the southeast region (115�W–3�Sand 90�W–9�S) show a westward component which isstronger than in the model, especially at the end of theperiod. In situ buoys released in May in the 12�S–13�Sregion had a strong westward drift that was not repro-duced in the central basin by the model. The agreementwas better in the western zone.

3.2 Nitrate and new production

In late 1996, before the onset of the event, the surfacenitrate concentrations in the central Pacific were repre-sentative of cold conditions both in the observations(Strutton and Chavez 2000) and in the model (Fig. 4).However, the measured nitrate was underestimated bythe model by 1.5–2 lM. The nitrate concentration fellabruptly near 2–4�N (convergence and downwellingzone, advection of low nitrate water from the west by theNECC) while the decrease at the southern edge of theupwelling was more gradual. Modeled new productionwas within the range of values found in the literature(McCarthy et al. 1996; Navarette 1998; Raimbault et al.1999; Turk et al. 2001; Aufdenkampe et al. 2002). It waslower than 1 mmol N m�2 day�1 in the oligotrophicregion whereas in the Wyrtki (1981) box (5�S–5�N;90�W–180�) it was 1.9 mmol N m�2 day�1 which isslightly lower than values inferred from measurementsor from other physical-biological coupled models. Themodel reproduced the relationship between measurednitrate concentrations integrated over the euphotic layerand new production as established by Raimbault et al.(1999) for the tropical Pacific. In particular, the rate ofincrease of new production in the HNLC region was

Fig. 4 Meridional distributionsof measured (top panels;adapted from Strutton andChavez 2000) and modeled(bottom panels) surface nitratebefore (late 1996; left panels)and during (May 1997; rightpanels) El Nino along 155�W(triangle) and 170�W (circle)

39

very slow despite high nitrate concentration. Immedi-ately following the strong westerly wind burst of March1997, the warm pool started to move eastward. In themodel, nitrate-poor waters reached 170�W in May and155�W about 1 month later which was consistent withthe satellite-observed progression of low chlorophyllwaters (Radenac et al. 2001; Ryan et al. 2002). As aresult, nitrate disappeared along 170�W and the nitratedecrease along 155�W between 4�S and 4�N afterDecember 1996 was about the same (40%) for theobservations (Strutton and Chavez 2000) and for themodel (Fig. 4). The small nitrate decrease simulatedaround 2�N in May 1997 is related to the weakness ofthe modeled NECC during that period. During the peakof the event, more than 6 months after the El Nino eventstarted, nitrate and new production measured in thecentral equatorial Pacific were representative of warmpool conditions which were reproduced in the model.Conversely, the uplift of the nitracline in the westernequatorial basin caused a relative increase of the chlo-rophyll concentration at the deep chlorophyll maximumand of new production (Turk et al. 2001; Radenac et al.2001; Christian et al. 2002).

3.3 The Lagrangian analysis

3.3.1 Redistribution of water masses between May 1997and December 1997

The distribution of surface water masses in May 1997 isshown in Fig. 5a. The nitrate-depleted layer of the warmpool was about 100 m deep between 160�E and 180�Eclose to the equator (Fig. 5b). It became thinner by 20–30 m poleward, in agreement with measurements(Mackey et al. 1995; Radenac and Rodier 1996; Turket al. 2001). Farther north, the base of the nitrate-de-pleted layer of the northwest subtropical water wasaround 70 m and deepened north of 12�N as reportedduring cruises (Kaneko et al. 1998). For the centralPacific, the model reproduced the asymmetry in nitratedistribution of the north and south subtropical waters(Wyrtki and Kilonsky 1984; Raimbault et al. 1999;Dugdale et al. 2002). The nitrate-depleted layer of thesouth subtropical water (that reaches 150 m south of10�S) was deeper than that of the northeast subtropicalwater (about 80 m). The nitracline of the upwellingequatorial waters shoaled eastward in agreement withthe thermocline slope (Fig. 5b).

In order to find the origin of water masses for a givenmonth, virtual water particles were released at the sur-face at each grid point of the model (5,849 particles).They were tracked backward in the three-dimensionalvelocity field until May 1997, when water masses weredefined. In this way, the release position for the selectedmonth was associated with the water mass of origin. Theredistribution of water masses since May was followedusing such mapping performed for each month betweenJune and December. Two intermediate phases of thisprogression in July (Fig. 5c) and September (Fig. 5d) areshown as well as the peak phase in December (Fig. 5e).Less than 1% of drifters that reach the surface in July,

Fig. 5 Redistribution of surface water masses between May andDecember 1997. a, b Surface distribution of water masses of originand their vertical section along the equator in May 1997. c–e Originof water masses in July, September and December 1997. The colorkey represents water masses: warm pool water (WPW), north-westsubtropical water (NWSW), north-east subtropical water (NESW),south subtropical water (SSW), equatorial upwelling water (UPW),subsurface water (SUBS). Remaining cases, essentially coastalupwelling waters, are shown in black. The thick black line is the1 lM surface nitrate concentration at the time of the release. Thedashed line in December is the leading edge of eastward spreading‘pre El Nino’oligotrophic waters

40

September, or December come from subsurface waters.It should be noted, however, that drifters at the surfacein July, September, or December may come from rela-tively deep within the body of surface water masses.

The main displacements in the surface layer betweenMay and December concerned the oligotrophic warmpool (WPW) and northwest subtropical waters (NWSW)and the nitrate-rich upwelled water (UPW) (Fig. 5).During a first phase (May–July), oligotrophic waterfrom the northwest tropical Pacific expanded south-wards in the western basin while water from the warmpool spread eastward (30� along the equator), breakingthrough the upwelling water (Fig. 5c). In July-August,waters from the northwest had reached the equator andbegan an eastward displacement. Meanwhile, followinga brief episode of westward flow, the upwelling watersshifted westward and separated oligotrophic waters intotwo branches, leading to the September situation(Fig. 5d). During the following months, oligotrophicwaters again spread eastward, with the northern branchprogressing a little bit faster than the southern branch.The south-eastward intrusion of northwest tropical wa-ter toward the central basin in fall 1997 was consistentwith the freshening observed in late 1997 in this region(Johnson et al. 2000). In December (Fig. 5e), a north-east-southwest oriented line (between 8�N–120�W and10�S–140�W) marked the easternmost expansion ofwarm pool and northwest subtropical waters. In thesouth, the eastern limit of south subtropical waters was

oriented northwest-southeast. Waters that originated inthe upwelling were located to the east of these bound-aries.

The leading edge of eastward spreading ‘pre El Nino’oligotrophic waters did not match the limit of the nitraterich surface waters (NO3>1 lM) of the Decemberdwindling upwelling (Fig. 5e). Actually, this simplediagram suggests that the main surface water massesredistributed according to three schemes during the 1997El Nino. First, advection without transformation pre-vailed west of the oligotrophic invasion boundary: oli-gotrophic waters from the western and northwesterntropical Pacific spread toward the central basin withminor nitrate changes. Second, the region inside the ni-trate-rich December 1997 limit was a retention zonewhere drifters have complex trajectories within theupwelled water mass as observed with in situ drifters(Fig. 3). Third, east of the leading edge and outside theremaining upwelling (delimited by the 1 lM isoline),nitrate had been exhausted during the drift of watermasses that were originally nitrate-rich: advection to-gether with transformation predominated. In that case,water masses followed two main paths. (1) North of theDecember residual upwelling, drifters came essentiallyfrom the western part of the May upwelling waters.They were caught in the strong NECC flow and movedeastward toward the Gulf of Panama. (2) South of theDecember upwelling, drifters originally in the upwellingregion followed the southward surface circulation thatdevelops during the fall in the central basin (Fig. 3).Both pathways have been observed with in situ drifters(Fig. 3).

3.3.2 Lagrangian nitrate changes

Figure 6 is an example of nitrate transformations thattake place along trajectories according to the advection/transformation scheme during the May 1997–December

Fig. 6 Results of the direct experiment: drifters are released at150�W in May 1997 and reach their end point in December 1997. aEvolution of the nitrate concentration (lM; color key) along thetrajectories. b Time-series of the mean depth (m; solid line) and NO3

concentration (lM; dashed line) along the northern pathway. cTime-series of the mean depth and NO3 concentration along thesouthern pathway. d Time-series of the horizontal diffusion (blueline), vertical diffusion (red line) and biological uptake (green line)in lmol NO3 m

�3 day�1 along the northern pathway. e Time-series of the horizontal diffusion, vertical diffusion and biologicaluptake along the southern pathway

41

1997 period. These nitrate changes are the consequenceof mixing with the neighboring water masses and bio-logical effects (Eq. 3). In order to investigate thosetransformations, properties such as depth, nitrate con-

centration, horizontal and vertical diffusion and bio-logical effects are sampled along the trajectories ofdrifters released at the surface in the western part of theupwelling in early May (along 150�W between 2�S and2�N).

Within 1 month, drifters split into the north andsouth groups as previously mentioned (Fig. 6a). Theirbehavior was quite similar. Drifters remained in theupper 20 m, except for drifters of the northern groupthat deepened to about 40 m at the end of the period(Fig. 6b, c). It takes about 120 days for drifters releasedin waters with nitrate concentrations of about 2.5 lM tobecome depleted (Fig. 6b, c). In May 1997 and duringthe depletion period, vertical nitrate supply in the centraland eastern equatorial Pacific had greatly decreased(Chavez et al. 1999; Radenac et al. 2001; Christian et al.2002). As expected, the lateral and vertical nitrate dif-fusion remained small along the drift of particles andnitrate changes along trajectories were essentially drivenby the biological uptake between May and September(Fig. 6d, e). Along the northern pathway (Fig. 6d), thenitrate-impoverished drifters reached the NECC flow inSeptember and accumulated in the direction of the Gulfof Panama. During their eastward drift north of theupwelling, the horizontal diffusion balanced the nitrateuptake and the nitrate concentration stayed close tozero. Along the southern pathway (Fig. 6e), nitrate wasassimilated by phytoplankton while the particles driftedslowly. After September, nitrate was exhausted, givingthese waters the same properties as oligotrophic watersof the subtropical south Pacific. For both drifter groups,biology drove the nitrate depletion during this periodwith dynamical characteristics specific of a mature ElNino phase.

Figure 7 shows a generalization of this experimentover the tropical Pacific basin. Drifters were released inDecember in the surface layer at each grid point of themodel and then tracked backward until May. For eachtrajectory, the nitrate consumption between May andDecember (D NO3) was computed and the biological(bio) and diffusion (phy) effects integrated. In order tocompare the order of magnitude of those quantities inDecember, –phy is presented in Fig. 7c (i.e. nitrate gainis negative). As expected in the region west of the limit ofeastward spreading oligotrophic waters, advection ofoligotrophic water prevailed and no major nitratechanges occurred (Fig. 7a). East of this limit, nitratediminished. Drifters that had reached the eastern part ofthe NECC or that were situated southward of theupwelling in December had lost 2–5 lM nitrate. Thenitrate fall could be even higher within the upwellingregion. Some sparse small spots of nitrate increase werefound in the upwelling and off the Central Americancoast. The physical inputs integrated over the May-December period were negligible west of the leadingedge of eastward spreading oligotrophic waters. East ofthe limit, the surface circulation patterns that haddrastically changed at the onset of the event only re-sulted in small physical inputs. The biological response

Fig. 7 Maps of the a NO3 consumption and impacts of the bbiological and c physical effects (lM; color key) computed alongLagrangian backward-trajectories between May and December1997. Positions are the positions at the release of virtual waterparcels in December. The 1 lM surface nitrate isoline (thick line) issuperimposed. The dashed line is the leading edge of eastwardspreading ’pre El Nino’oligotrophic waters

42

was slower and during that period, nitrate assimilationremained rather stable. As a result, the biological con-sumption was greater than the physical inputs and drovethe nitrate impoverishment.

4 Discussion

The reliability of the GCM is an essential point for thisstudy. There was general agreement between modeledand observed surface circulation for 1997, in spite of aslight local underestimation of the meridional velocity.The model accurately reproduced the timing of thereversals of current at the equator. In particular, thestrong eastward flow in May–June and the brief SECreturn in July-August were well captured. Imperfectionsin the mixing scheme caused excessive vertical diffusionin the thermocline and therefore in the nitracline (Via-lard et al. 2001; Radenac et al. 2001; Lengaigne et al.2003). Consequently, the nitrate flux into the modeledsurface mixed layer should be too high. This bias causeda very strong new production during the first years of thespin-up phase that was initially performed for the nitratefield to reach equilibrium with the large-scale dynamics(Radenac et al. 2001). The overall effect of excessivelyhigh diffusion during the spin-up (8 years) was totransfer too much nitrate from just below the pycnoclineinto the mixed layer and then to immediately export it todepth. Finally, this led to abnormally low nitrate con-centrations in the lower part of the nitracline whereasthe modeled nitrate values remained consistent withobservations in the mixed layer.

The constants Vmax and KNO3 in the biogeochemicalmodel had been estimated in order to reproduce asfaithfully as possible the nitrate concentrations observedin the photic layer (Stoens et al. 1999). The model isbased on the common assumption that nitrate fixationequals new and exported production. Vmax was in therange of values found after field experiments (Stoenset al. 1999). On the contrary, KNO3=0.01 lM was muchlower than values generally accepted. This biologicallyunrealistic value was adopted to balance the excessivenitrate diffusion into the mixed layer in the oligotrophicwarm pool waters. Higher KNO3 values indeed led toslow nitrate assimilation when nitrate concentration waslow, so that nitrate was always present in the warm poolphotic layer instead of being depleted. This choice hadno consequences in upwelled waters with significantamounts of nitrate, where nitrate assimilation is drivenby Vmax. These should not apply to rarely reported in-tense phytoplankton blooms where high nutrients con-centration and probably iron availability favored amuch faster nitrate uptake (Ryan et al. 2002). However,such events are too scarce to significantly impact theaverage patterns of nitrate utilization in the equatorialPacific. Since these constants finally reproduced the ni-trate variability in the main oligotrophic and meso-trophic regions and the slow nitrate uptake of theHNLC ecosystem, it was considered that the modeled

nitrate uptake does not largely differ from the actualnitrate uptake. These considerations only concern theoligotrophic and HNLC water masses prior to andduring the 1997–1998 El Nino. They should not apply tothe recovery phase immediately after this event, whenupwelling resumed and brought nitrate-rich waters andiron from below into an entirely oligotrophic ocean(Murtugudde et al. 1999; Ryan et al. 2002).

Chlorophyll concentration estimated from oceancolor data is approximately the mean chlorophyll con-centration between the surface and the first attenuationlength which varies as the inverse of the attenuationcoefficient Kd, at the 490 nm wavelength used in theSeaWiFS algorithm (Morel and Maritorena 2001). It isabout 40 m in oligotrophic waters where chlorophyll islower than 0.1 mg m�3 and about 30 m in waters withchlorophyll ranging between 0.1 mg m�3 and0.3 mg m�3. There are two reasons that support theanalyses of the trajectories of drifters simply released inthe first layer of the model in order to understand thebiological situation as seen by ocean color during thepeak of the 1997 El Nino event. The first one is that theupper two layers of the model (0–20 m) would accountfor more than 60% of the signal in oligotrophic watersand that the first layer alone would account for about50% in upwelling waters. The second one is that duringthe singular period studied, drifters released in the sur-face layer remained within the mixed layer. The reverseexperiment confirmed this result: virtual drifters thatfollowed the strong NECC flow originated from thesurface HNLC waters of the western part of the equa-torial divergence and virtual drifters that reached southof the upwelling in December drifted southward near thesurface. Consequently, the conclusion of the study as tothe origin of water masses of the first layer of the modelis reasonable.

During the 1997–1998 El Nino event, lateral andvertical nitrate inputs are greatly reduced in the HNLCregion because of the deepening of the nutricline(Chavez et al. 1998; Stoens et al. 1999; Strutton andChavez 2000; Radenac et al. 2001; Christian et al.2002) and the decrease in activity of tropical instabilitywaves (Philander et al. 1985). Biological uptake thenbecomes the main process that drives nitrate depletionalong trajectories. Because of the slow nitrate uptakecharacteristic of the tropical Pacific HNLC system,nitrate that was present in the euphotic layer at theearly stage of El Nino was able to sustain biologicalproduction during several months. So, this Lagrangianstudy shows that the slow nitrate uptake characteristicof HNLC waters prevents the tropical Pacific from arapid onset of oligotrophic conditions. Other studieshave shown that the chlorophyll and biological activitycould increase in several regions of the tropical Pacificduring El Nino events. In the western Pacific, the basintilt of the thermocline and nutricline resulted in highernitrate content in the euphotic layer and increasedbiological activity as shown by Radenac et al. (2001),Turk et al. (2001), or Christian et al. (2002) for the

43

1997 event. Similar vertical processes lead to the northequatorial enrichment that has been reported duringthe 1982 El Nino event: abnormally high chlorophyllconcentration measured in a wide area centered in thewarm pool at 170�E, 10�N (Dandonneau 1992) andabnormally high biogenic particle fluxes at 140�W,11�N (Dymond and Collier 1988). In 1997, the SeaW-iFS imagery revealed a large zone of surface chloro-phyll higher than 0.1 mg m�3 that extended from about10�N off the central American coast southwestwardtoward the equator (Murtugudde et al. 1999). Numer-ous equatorial studies show a decrease of the biologicalactivity along the equator during El Nino events.Nevertheless, the above-mentioned observations to-gether with this investigation suggest that the tropicalPacific is not a biological desert throughout the event.Without the ‘‘memory’’ of the HNLC ecosystem, thedecrease of biological production would be more pro-nounced in waters of the Pacific equatorial divergenceand increased biological activity occurs in several off-equatorial locations.

5 Conclusion

This study contributes to the general issue of the relativecontributions of physical and biological processes to thesetting of El Nino oligotrophic conditions. Becausechlorophyll is often used as a proxy for phytoplanktonbiomass, ocean color imagery is a key instrument formonitoring the relative extension of phytoplankton richand poor waters at different timescales. They may behints we can grasp about the underlying mechanisms byobserving surface chlorophyll patterns and their evolu-tion. However, there was a gap in ocean color coveragebetween June and September 1997, i. e. when most of thepre-El Nino surface layer nitrate was being consumed.An insight of these processes is possible using numericalsimulations, although they do not perfectly reproducethe real ocean. A Lagrangian analysis applied to theoutput of a coupled physical-biogeochemical modelshows that surface nitrate-rich and nitrate-poor watermasses were redistributed by three mechanisms duringthe 1997 El Nino. First, oligotrophic waters from thewestern Pacific remained oligotrophic during the courseof their eastward transport. This mechanism is the mostdocumented one along the equator and highlights theimpact that variability of the general circulation mayhave on the biological variability at a first order. Second,an upwelling system persists in the eastern part of thebasin. This study emphasized a third scheme that rep-resents an alternative to generalized nitrate-poor waterconditions. Nitrate that was present in the euphotic layerat the early stage of El Nino was able to sustain bio-logical production during several months. Thus, theslow nitrate uptake characteristic of HNLC watersdelayed the onset of oligotrophic conditions allowingmesotrophic conditions to persist in the tropical Pacificduring this El Nino event.

Acknowledgements We would like to thank G. C. Johnson, F.Bonjean and G. Lagerloef for providing the ADCP cruise data andthe satellite-derived currents. Our thanks as well to the DrifterData Assembly Center, GOOS Center, NOAA/AOML, Miami,Florida, for their quality control of the drifter data and for makingit available through the web. J. Sudre extracted the ocean colordata. Special thanks to C. Maes and F. Melin and to two anony-mous reviewers for constructive comments.

References

Aufdenkampe AK, McCarthy JJ, Navarette C, Rodier M, DunneJ, Murray JW (2002) Biogeochemical and physical controls onnew production in the tropical Pacific. Deep Sea Res II 49(13-14):2619–2648

Barber RT, Chavez FP (1983) Biological consequences of El Nino.Science 222:1203–1210

Bender ML, McPhaden MJ (1990) Anomalous nutrient distribu-tion in the equatorial Pacific in April 1988: evidence for rapidbiological uptake. Deep Sea Res 37:1075–1084

Blanke B, Raynaud S (1997) Kinematics of the Pacific equatorialundercurrent: an eulerian and Lagrangian approach fromGCM results. J Phy Oceanogr 27:1038–1053

Bonjean F, Lagerloef GSE (2002) Diagnostic model and analysis ofthe surface currents in the tropical Pacific Ocean. J Phy Ocea-nogr 32(10):2938–2954

Boutin J, Etcheto J, Dandonneau Y, Bakker DCE, Feely RA, In-oue HY, Ishii M, Ling RD, Nightingale PD, Metzl N, Wann-inkhof R (1999) Satellite sea surface temperature: a powerfultool for interpreting in situ pCO2 measurements in the equa-torial Pacific ocean. Tellus 51(B):490–508

Chavez FP, Strutton PG, McPhaden MJ (1998) Biological-physicalcoupling in the central Pacific during the onset of the 1997–98El Nino. Geophys Res Lett 25(19):3543–3546

Chavez FP, Strutton PG, Friederich GE, Feely RA, Feldman GC,Foley DG, McPhaden MJ (1999) Biological and chemical re-sponse of the equatorial Pacific Ocean to the 1997–1998 ElNino. Science 286:2126–2131

Christian JR, Verschell MA, Murtugudde R, Busalacchi AJ,McClain CR (2002) Biogeochemical modelling of the tropicalPacific ocean. I. Seasonal and interannual variability. Deep SeaRes II 49:509–543

Coale KH, Fitzwater SE, Gordon RM, Johnson KS, Barber RT(1996) Control of community growth and export production byupwelled iron in the equatorial Pacific Ocean. Nature 379:621–624

Dandonneau Y (1979) Concentrations en chlorophylle dans lePacifique tropical sud-ouest: comparaison avec d’autres airesoceaniques tropicales. Oceanol Acta 2:133–142

Dandonneau Y (1992) Surface chlorophyll concentration in theTropical Pacific Ocean: an analysis of data collected bymerchant ships from 1978 to 1989. J Geophys Res 97:3581–3591

Dandonneau Y, Eldin G (1987) Southwestward extent of chloro-phyll-enriched waters from the Peruvian and equatorial upw-ellings between Tahiti and Panama. Mar Ecol Progr Ser 38:283–294

Deschamps PY, Breon FM, Leroy M, Podaire A, Bricaud A, Bu-riez JC, Seze G (1994) The POLDER mission: instrumentcharacteristics and scientific objectives. IEEE Trans GeosciRemote Sensing 32:598–615

Dugdale RC, Barber RT, Chai F, Peng TH, Wilkerson FP(2002) One-dimensional ecosystem model of the equatorialPacific upwelling system. Part II: sensitivity analysis andcomparison with JGOFS EqPac data. Deep Sea Res II49:2747–2768

Dymond J, Collier R (1988) Biogenic particle fluxes in the equa-torial Pacific: evidence for both high and low productivityduring the 1982–1983 El Nino. Global Biogeoch Cycles2(2):129–137

44

Eldin G, Rodier M, Radenac MH (1997) Physical and nutrientvariability in the upper equatorial Pacific associated with wes-terly wind forcing and wave activity in October 1994. Deep SeaRes II 44:1783–1800

Friedrichs MAM, Hofmann EE (2001) Physical control of bio-logical processes in the central equatorial Pacific Ocean. DeepSea Res I 48:1023–1069

Hansen DV, Poulain PM (1996) Quality control and interpolationsof WOCE/TOGA drifter data. J Atmosph Oceanic Technol13:900–909

Hardy J, Hanneman A, Behrenfeld M, Horner R (1996) Environ-mental biogeography of near-surface phytoplankton in thesoutheast Pacific Ocean. Deep Sea Res I 43(10):1647–1659

Hayes SP, Chang P, McPhaden MJ (1991) Variability of the seasurface temperature in the eastern equatorial Pacific during1986–88. J Geophys Res 96:10553–10566

Honjo S (1978) Sedimentation of material in the Sargasso Sea at5,367 m. J Mar Res 36:469–492

Inoue HY, Ishii M, Matsueda H, Ahoyama M (1996) Changes inlongitudinal distribution of the partial pressure of CO2 (pCO2)in the central and western equatorial Pacific, west of 160�W.Geophys Res Lett 14:1781–1784

Johnson GC, McPhaden MJ, Rowe GD, McTaggart KE (2000)Upper equatorial Pacific Ocean current and salinity variabilityduring the 1996–1998 El Nino-LaNinacycle. J Geophys Res105:1037–1053

Kaneko I, Takatsuki Y, Kamiya H, Kawae S (1998) Water prop-erty and current distributions along the WHP-P9 section (137�–142�E) in the western North Pacific. J Geophys Res103(C6):12959–12984

Karl DM, Lukas R (1996) The Hawaii Ocean Time-series (HOT)program: background, rationale and field implementation.Deep Sea Res II 43(2-3):129–156

Kobayashi F, Takahashi K (2002) Distribution of diatoms alongthe equatorial transect in the western and central Pacific duringthe 1999 La Ninaconditions. Deep Sea Res II 49:2801–2821

Kuroda Y, McPhaden MJ (1993) Variability in the western equa-torial Pacific Ocean during Japanese Pacific Climate StudyCruises in 1989 and 1990. J Geophys Res 98:4747–4759

Lagerloef GSE, Mitchum GT, Lukas RB, Niiler PP (1999) TropicalPacific near-surface currents estimated from altimeter, wind,and drifter data. J Geophys Res 104:23313–23326

Landry MR, Barber RT, Bidigare RR, Chai F, Coale KH, DamHG, Lewis MR, Lindley ST, McCarthy JJ, Roman MR,Stoecker DK, Verity PG, White JR (1997) Iron and grazingconstraints on primary production in the central equatorialPacific: an EqPac synthesis. Limnol Oceanogr 42:405–418

Le Borgne R, Barber RT, Delcroix T, Inoue HY, Mackey DJ,Rodier M (2002) Pacific warm pool and divergence: temporaland zonal variations on the equator and their effects on thebiological pump. Deep Sea Res II 49:2471–2512

Lehodey P, Andre JM, Bertignac M, Hampton J, Stoens A,Menkes C, Memery L, Grima N (1998) Predicting skipjack tunaforage distributions in the equatorial Pacific using a coupleddynamical bio-geochemical model. Fish Oceanogr 7:317–325

Lengaigne M, Madec G, Menkes C, Alory G (2003) Impact ofisopycnal mixing on the tropical ocean circulation. J GeophysRes 108(C11). DOI: 10.1029/2002JC001704

Levitus S, Conkright ME, Reid JL, Najjar RG, Mantyla AW(1993) Distribution of nitrate, phosphate, and silicate in theworld oceans. Progr Oceanogr 31:245–273

Mackey DJ, Parslow J, Higgins HW, Griffiths FB, O’Sullivan JE(1995) Plankton productivity and biomass in the westernequatorial Pacific: biological and physical controls. Deep SeaRes II 42:499–533

Maes C, Madec G, Delecluse P (1997) Sensitivity of an equatorialPacific OGCM to the lateral diffusion. Mon Wea Rev 125:958–971

McCarthy JJ, Garside C, Nevins JL, Barber RT (1996) Newproduction along 140�W in the equatorial Pacific during andfollowing the 1992 El Ninoevent. Deep Sea Res II 43:1065–1093

McClain CR, Cleave ML, Feldman GC, Gregg WW, Hooker SB,Kuring N (1998) Science quality SeaWiFS data for globalbiosphere research. Sea Technol 39:10–16

McPhaden MJ, Picaut J (1990) El Nino-SouthenOscillation dis-placements of the western equatorial Pacific warm pool. Science250:1385–1388

McPhaden MJ, Busalacchi AJ, Cheney R, Donguy JR, Gage KS,Halpern D, Ji M, Julian P, Meyers G, Mitchum GT, Niiler PP,Picaut J, Reynolds RW, Smith N, Takeuchi K (1998) TheTropical Ocean-Global Atmosphere observing system: a decadeof progress. J Geophys Res 103:14169–14240

Morel A, Maritorena S (2001) Bio-optical properties of oceanicwaters: a reappraisal. J Geophys Res 106:7163–7180

Murakami H, Ishizaka J, Kawamura H (2000) ADEOS observa-tions of chlorophyll a concentration, sea surface temperature,and wind stress change in the equatorial Pacific during the 1997El Ninoonset. J Geophys Res 105:19551–19559

Murtuggude RG, Signorini SR, Christian JR, Busalacchi AJ,McClain CR, Picaut J (1999) Ocean color variability of thetropical Indo-Pacific basin observed by SeaWiFS during 1997–98. J Geophys Res 104:18351–18365

Navarette C (1998) Dynamique du phytoplancton en ocean equa-torial: mesures cytometriques et mesures isotopiques durant lacampagne FLUPAC, en octobre 1994 dans la partie ouest duPacifique. These de Doctorat de l’Universite Paris VI

Niiler P (2001) The world ocean surface circulation. In: Church J,Siedler G, Gould J (eds) Ocean circulation and climate-observing and modeling the global ocean. Academic, NewYork, pp 193–204

Pazan SE, Niiler P (2004) New global drifter data set available.EOS Trans Am Geophys Union 85(2)

Philander SGH, Halpern D, Hansen D, Legeckis R, Miller L, PaulC, Watts R, Weisberg R, Winbush M (1985) Long waves in theequatorial Pacific Ocean. EOS Trans Am Geophys Union66:154

Picaut J, Ioualalen M, Menkes C, Delcroix T, McPhaden MJ(1996) Mechanisms of the zonal displacements of the Pacificwarm pool: implications for ENSO. Science 274:1486–1489

Price NM, Ahner BA, Morel FMM (1994) The equatorial PacificOcean: grazer-controlled phytoplankton populations in an iron-limited ecosystem. Limnol Oceanogr 39(3):520–534

Radenac MH, Rodier M (1996) Nitrate and chlorophyll distribu-tions in relation to thermohaline and current structures in thewestern tropical Pacific during 1985–1989. Deep Sea Res II43:725–752

Radenac MH, Menkes C, Vialard J, Moulin C, Dandonneau Y,Delcroix T, Dupouy C, Stoens A, Deschamps PY (2001)Modeled and observed impacts of the 1997–1998 El Ninoonnitrate and new production in the equatorial Pacific. J GeophysRes 106:26879–26898

Raimbault P, Slawyk G, Boudjellal B, Coatanoan C, Conan P,Coste B, Garcia N, Moutin T, Pujo-Pay M (1999) Carbon andnitrogen uptake and export in the equatorial Pacific at 150�W:evidence of an efficient regenerated production cycle. J GeophysRes 104:3341–3356

Reverdin G, Frankignoul C, Kestenare E, McPhaden MJ (1994)Seasonal variability in the surface currents of the equatorialPacific. J Geophys Res 99:20323–20344

Reynolds RW, Smith TM (1994) Improved global sea surfacetemperature analyses using optimum interpolation. J Clim7:929–948

Ryan JP, Polito PS, Strutton PG, Chavez FP (2002) Unusual large-scale phytoplankton blooms in the equatorial Pacific. ProgrOceanogr 55(3):263–285

Stoens A, Menkes C, Radenac MH, Grima N, Dandonneau Y,Eldin G, Memery L, Navarette C, Andre JM, Moutin T, Rai-mbault P (1999) The coupled physical-new production systemin the equatorial Pacific during the 1992–1995 El Nino.J Geophys Res 104:3323–3339

Strutton PG, Chavez FP (2000) Primary productivity in theequatorial Pacific during the 1997–98 El Nino. J Geophys Res105:26089–26101

45

Turk D, Lewis MR, Harrison GW, Kawano T, Asanuma I (2001)Geographical distribution of new production in the western/central equatorial Pacific during El Ninoand non-El Ninocon-ditions. J Geophys Res 106:4501–4515

Venrick EL, MacGowan JA, Mantyla AW (1973) Deep maxima ofphotosynthetic chlorophyll in the Pacific Ocean. US Fish Bull71:41–52

Vialard J, Menkes C, Boulanger JP, Delecluse P, Guilyardi E,McPhaden MJ, Madec G (2001) A model study of oceanicmechanisms affecting equatorial Pacific sea surface temperatureduring the 1997–98 El Nino. J Phy Oceanogr 31:1649–1675

Vialard J, Delecluse P, Menkes C (2002) A modeling study ofsalinity variability and its effects in the tropical Pacific oceanduring the 1993–99 period. J Geophys Res 107(C12).DOI:10.1029/2000JC000758

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

Wyrtki K, Kilonsky B (1984) Mean water and current structureduring the Hawaii to Tahiti shuttle experiment. J Phy Oceanogr14:242–254

46