Annual Cycle of Primary Production in the Cariaco Basin: Response to Upwelling and Implications for Vertical Export

Home , Cariaco Basin

University of South Florida Scholar Commons

Marine Science Faculty Publications College of Marine Science

3-15-2001

Frank E. Muller-Karger University of South Florida, [email protected]

Ramon Varela Estacion de Investigaciones Marinas de Margarita

Robert Thunell University of South Carolina

Mary Scranton State University of New York

Richard Bohrer University of South Florida

See next page for additional authors

Follow this and additional works at: https://scholarcommons.usf.edu/msc_facpub

Part of the Marine Biology Commons

Scholar Commons Citation Muller-Karger, Frank E.; Varela, Ramon; Thunell, Robert; Scranton, Mary; Bohrer, Richard; Taylor, Gordon; Capelo, Juan; Astor, Yrene; Tappa, Eric; Ho, Tung-Yuan; and Walsh, John J., "Annual Cycle of Primary Production in the Cariaco Basin: Response to Upwelling and Implications for Vertical Export" (2001). Marine Science Faculty Publications. 56. https://scholarcommons.usf.edu/msc_facpub/56

This Article is brought to you for free and open access by the College of Marine Science at Scholar Commons. It has been accepted for inclusion in Marine Science Faculty Publications by an authorized administrator of Scholar Commons. For more information, please contact [email protected]. Authors Frank E. Muller-Karger, Ramon Varela, Robert Thunell, Mary Scranton, Richard Bohrer, Gordon Taylor, Juan Capelo, Yrene Astor, Eric Tappa, Tung-Yuan Ho, and John J. Walsh

This article is available at Scholar Commons: https://scholarcommons.usf.edu/msc_facpub/56 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. C3, PAGES 4527-4542, MARCH 15, 2001

Annual cycle of primary production in the Cariaco Basin: Responseto upwelling and implications for vertical export Frank Muller-Karger,• Ramon Varela,2 Robert Thunell,3 Mary Scranton,4 Richard Bohrer,• Gordon Taylor,4 Juan Capelo,2 Yrene Astor,2 Eric Tappa,3 Tung-Yuan Ho, 4 and John J. Walsh•

Abstract. Monthly hydrographic,primary production,bacterial production,and settling particulatecarbon flux observationswere collectedbetween November 1995 and December 1997 at 10.5øN,64.67øW within the Cariaco Basin, off Venezuela. Upwelling of SubtropicalUnderwater (SUW) startedaround October and lastedthrough approximately Mayof thefollowing year. Wind speeds >7 m s-• wereobserved between January and June,with weaker winds (<5 m s-•) betweenJuly and December. The upwelling cycle was thereforeout of phasewith that of the trade windsby 2-3 months.A seasonalcycle punctuatedby transientextremes associated with subsurfaceventilation events was observedin primary production.High bacterial activityand organiccarbon recycling rates were observednear the oxic-anoxicinterface. Integrated primary productionwas 690 gC m-2 yr-• in 1996and 540 gC m -2 yr-• in 1997.Settling carbon flux measured with sedimenttraps was about 5.6% of integratedprimary productionat 275 m and about 1.7% at 1225m, with no seasonalityin theproportion of verticalflux to primaryproduction. In total,between 10 and11 gC m-2 yr-*were deliveredto thebottom sediment of Cariaco, whichsuggests that between 4 x l0s and1 x 106 t of C yr-• weredelivered to sediments within the upwellingarea of the CariacoBasin. This representspermanent sequestration of carbonpreviously entrained in the North Atlantic gyre in the area of formation of SUW. Resultssuggests that upwelledinorganic nitrogen, rather than nitrogen fixation, is responsiblefor the large productivityand particulatecarbon settling flux in the Cariaco Basin.

1. Introduction Many studieshave focusedon the presentbalance between inorganicand organicpools in the deep ocean [see,e.g., Han- Over geologicaltimescales, geochemical imbalances at the selland Carlson,1998, and referencestherein], but the contri- Earth's surface are closely tied to carbon sequestrationby bution toward this balanceby watersnear continentalmargins marine organismsand to the sinkingflux of particulateorganic remainsunclear. All attemptsat global assessmentsof marine material [Berner,1992]. This variable flux servesas a key to primaryproduction indicate that oceanmargins support on the interpretingpast climatewhen preservedat depth. However, average2-5 timesthe annual productionof open oceanwaters the relationshipbetween fossil flux and oceanographiccondi- [Koblentz-Mishkeet al., 1970; Field et al., 1998]. Indeed, the tions near the ocean'ssurface must be analyzedto understand mean particleflux at 2300 m alongcontinental margins is of the how these records can be used to make inferences about cli- orderof about7.0 gC m-2 yr-•, comparedto ---0.8gC m-2 mate change.There are few placeswhere flux observations yr-• at 2300m in thedeep sea [Deuser et al., 1990;Walsh, 1991; have been linked to climatic and oceanographicforcing con- Pilskalnet al., 1996;Thunell, 1998b]. Globally, over 65% of the ditions[Deuser et al., 1990;Karl et al., 1996; Thunell,1998a]. particulatecarbon that falls below 1000 m is derivedfrom the Indeed, only in a few locationsdo the underlyingsediments slopeand rise of continentalmargins [Jahnke, 1996], and this preservevariability spanning a wide range of timescales.This estimatedoes not include areas shallowerthan 1000 m [see hasmade it difficult to establishthe relationshipbetween pro- alsoHonjo et al., 1982;Spencer, 1984; Walsh et al., 1992].These ductivityin surfacewaters and biogeochemicalfeedbacks act- simple statisticssuggest that the biologicalpump [Volk and ing over decadaland centurytimescales [see Falkowski et al., Liu, 1988] is muchmore efficientin the vicinityof continental 19981. marginsthan in the ocean'sinterior. To try to understandthe connectionbetween surfacepro- 1Collegeof MarineScience, University of SouthFlorida, St. Peters- duction and vertical carbon flux in a productive setting,we burg, Florida. starteda seriesof monthlyhydrographic and productivityob- 2FundacionLa Sallede CienciasNaturales, Estacion de Investiga- cionesMarinas de Margarita, Isla de Margarita, Venezuela. servationsand deployeda set of sedimenttraps in the Cariaco 3Departmentof GeologicalSciences, University of SouthCarolina, Basin, off the coast of Venezuela. These observationsform the Columbia, South Carolina. basisof the Carbon Retention in a Colored Ocean (CARI- 4MarineSciences Research Center, State University of New York, ACO) study.Here we discussthe first 2 yearsof observations Stony Brook, New York. at the CARIACO site. Copyright2001 by the American GeophysicalUnion. The CariacoBasin (Figure 1) offers a unique geographical Paper number 1999JC000291. settingthat permits studiesthat are difficult or impossibleto 0148-0227/01/1999JC000291509.00 conductat other sites.This is a large (---160 km long, 70 km

4527 4528 MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE

20' 10' 66 ø 50' 40' 30' 20' 10' 65 ø 50' 40' 30' 20' 10' 64 ø 40' 40'

30' 30'

20' 20'

10' DE MARGARITA 10' DELA t 'RTUGA.... ISLA DE MARGARITA

11 ø 11 ø

50' 50' SILLA DE CUBAGUA-.

40' ILLA CENTRAL 40'

30' 30'

20' LATAFORMA DE UNA 20'

10' 10'

10 ø 10 ø 20' 10' 66 ø 50' 40' 30' 20' 10' 65 ø 50' 40' 30' 20' 10' 64 ø

Figure 1. The CariacoBasin: bathymetry and locationof the CARIACO time seriesstation [modified from Febres-Ortegaand Herrera, 1975].

wide) and deep (--•1400 m) basinwithin a continentalshelf. to decadal-scalechange [Petersonet al., 1991; Hughen et al., It is bound to the north by a sill connectingMargarita Island 1996, 1998;Haug et al., 1998;Black et al., 1999]. The evidence to Cabo Codera at a mean depth of about 100 m. There are suggeststhat evenweak disturbanceswith a muted expression two channelsbreaching this sill (Figure 1): one in the north- in the deep ocean are recordedtemporarily within the water east of 135 m depth (La Tortuga) and a narrower one in the columnin Cariaco and permanentlyin the sedimentsat the northwestof 146 m depth (Centinela) [Richards,1975; Lidz bottom of the basin. et al., 1969]. In additionto the recentimportance of the CariacoBasin as Current conceptualmodels attribute a marked cyclein up- the site of an important paleo-oceanographictime series,the wellingand associatedsea surface temperature (SST) changes CariacoBasin has servedas a natural laboratoryfor biogeo- in this region primarily to the seasonalintensification of the chemistsfor over 40 years. This basin has been key in con- tradesalong an east-westcoastline [Richards, 1960, 1975;Her- structingstoichiometric models of organicmatter remineral- reraand Febres-Ortega, 1975; Muller-Karger and Aparicio, 1994]. ization [Redfieldet al., 1963; Richards, 1975], developing Similar and concurrentvariations in upwellingoccur along the residencetime and box models,studying metallic sulfides [Ba- entire south/central Caribbean Sea, but we show here that con et al., 1980], and numerousother studies.However, very winds lag behind the onset of upwellingby up to 3 months. little is known about the variabilityin primary productionor Within the CariacoBasin the seasonalupwelling process pro- aboutthe actualprocesses that controlthe carbonand nutrient videsa sourceof nutrientsthat leadsto vigorousphytoplank- fluxes within the basin. ton growth near the surface. Previous studies suggestthat muchof this remainsungrazed and sinks,allowing pigments to 2. Methods reach the sediment [Richards,1975; Richardsand Vaccaro, 1956]. Twenty-sixCARIACO cruiseswere conductedbetween No- In contrastto the situationat other continentalmargins, the vember1995 and December1997 to examinethe hydrography, CariacoBasin has a muted advectiveregime at depth.The sill primary productivity,and vertical flux of material at 10.5øN, restricts water motion and the lateral flux of material, thus 64.67øW.Figure 2 is a schematicof the CARIACO logistics. forming a natural sedimenttrap within a continentalshelf. As The baseof operationsis the Estacionde InvestigacionesMa- the turnoverof basinwaters is slow[Deuser, 1973], the decom- rinasIsla Margarita (EDIMAR) of the FundacionLa Sallede position of the sinking material leads to permanent anoxia CienciasNaturales (FLASA), located on Punta de Piedras, below about275 m depth. The CariacoBasin is well knownto Margarita Island, Venezuela, and cruisesuse the R/V Her- be the largestanoxic basin of truly oceaniccharacter. Indeed, mario Gines(FLASA). the hydrography,chemistry, and sediment deposition pro- Wind data were collected at Santiago Marlrio Airport cessesin the CariacoBasin are sensitiveto changesin climate (10.9øN63.96øW). Hourly recordswere averagedto dailyval- [e.g.,Holmen and Rooth, 1990;Scranton, 1988; Scranton et al., ues,and a five-pointrunning mean was appliedto generatethe 1987;Deuser, 1973]. Varved sedimentsthat accumulatewithin figuresshown in thispaper. Historicalwind recordsfrom Punta the bottom anoxicwaters provide a detailed record of annual- de Piedras,Margarita Island (10.9øN, 64.2øW) were alsoexam- MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE 4529

Infrared & Ocean Color Satellites: NOAA 12/14 l WHOIL USFSt. Pete. (FL) DOC II General Coordination OCTS, SeaWiFS Data Q/C: Nutrients, CTD, Inorg. Carbon BBSR/Bermuda•Image Processingand Analysis/ Synthesis NumericalModeling

SUNY Stony Brook Bacterial metabolism Bacterialproductivity Methane,Acetate cycl.

USC, Columbia (SC) Sedimenttraps ISatellite data products I , Particulate Flux , ', Isotopes ß ': ". Organicgeochemistry IR/V...•ø' Gmeal,, ...... ,'•-3'•,•, •',•, • FLAS•D•AR NumentsCu• I I&e• Co•or, I ! 5 I•4'c •m•y• •ntadebetas Phys.Oc.,I IReflecmce 'l ] •l •oduction[ Logistics BactehaI •...... / • • S•inity,pH,Alk., / [-00m

I I Oce• Color I •uorom.Rosette II•lI• I Radi•ceI•a•cel I i A•c. P. I Satellitedata •. I mr••. I• 1,3• e•chI [ • -200m I D.Oxygen I ECA•O ...... -27m Oxygen ANOXIC Nutbents • -440m Phyto.T•on. [ SedimentTraps -840m [Cariaco TrenchI

Figure 2. Schematicof the CARIACO program showinginstitutions involved and a summaryof observations collected.

ined to assessthe quality of observationsat SantiagoMarlrio (8 L) teflon-coatedNiskin bottleswith teflon-coatedsprings. Airport. The rosettehoused the CTD, YSI oxygenprobe, and Chelsea SSTwas estimated in situ at 1 m depthwith our conductivity- Instrumentsprofiling fluorometer outfitted for chlorophylla temperature-depth(CTD) sensorsand more frequentlyfrom estimates.A SeaTecc beam transmissometer(660 nm) was imagery collectedby the advancedvery high resolutionradi- added to the systemon the eleventhmonth of the time series. ometer (AVHRR) sensorson the NOAA 11, 12, and 14 sat- For the first 10 CARIACO cruises we used a series of Sea- ellites.The satelliteSST were derivedusing the split window Bird SBE-19 CTDs, but we had some difficultywith inconsis- techniques[Walton, 1988; Strong and McClain, 1984;McClain tent and obviouslyerroneous salinity data below the oxic- et al., 1983]. While nonlinear algorithms developed for anoxic interface with these devices. Through repetitive AVHRR data could be used, there is no conclusive evidence attempts,good profiles were obtained for each cruise. How- that thesealgorithms are any better than the multichannelsea ever, the SBE-19 was permanentlyreplaced with an SBE-25 in surface temperature algorithms. The nominal accuracy of September1996. In additionto annual calibrationby the man- AVHRR SST retrievalsis in the range of _+0.3-_+1.0K [see ufacturer, each salinityprofile was correctedwith discretesa- alsoBrown et al., 1985;Minnett, 1991]. In Figure 3 we showa linity observationscollected at 20 depths.The salinitymeasure- smoothedrepresentation of the satellite-derivedSST (100- ments were conducted onshore on a GuildlineTM Portasal 8410 point runningmean) and the monthly1 m CTD temperatures salinometercalibrated with International Associationfor Phys- at CARIACO. ical Sciencesof the Oceans(IAPSO) StandardOcean Water. A minimumof five hydrocastswere performed during each Phytoplanktonbiomass was estimatedwith chlorophylla monthly cruise to collect a suite of core observations.Addi- extractedin methanol and read on a Turner Designsfluorom- tional hydrocastswere performed for specificprocess studies. eter using standard methods [Holm-Hansen et al., 1965; Water wascollected with a SeaBirdTM rosette equipped with 12 Falkowskiand Kiefer, 1985]. During the highproduction season 4530 MULLER-KARGER ET AL.' CARIACO BASIN PRIMARY PRODUCTION CYCLE

(a) 3.0 2.5 2.0 1.5 1.0

0.5

0.0 -0.5

J FMAMJ J ASONDJ FMAMJ J ASONDJ FMAMJ J ASONDJ 1995 1996 1997 1998

(b) -3

J FMAMJ J ASOND J FMAMJ J ASONDJ FMAMJ J ASOND J 1995 1996 1997 1998 (c) 1o

-o 7

J FMAMJ J ASONDJ FMAMJ J ASONDJ FMAMJ J ASONDJ 1995 1996 1997 1998

(d) 29 28-

27 26

25 24-

23-

22-

21 J FMAMJ J ASOND J FMAMJ J ASONDJ FMAMJ J ASONDJ 1995 1996 1997 1998 Figure3. Wind((a) v, (b) u, and(c) scalarspeed (m s-h))at SantiagoMarifio Airport (Margarita Island). Broken line laid over the scalarwind representsthe monthlymean scalarwind. Negativeu and v values representwestward and southward winds, respectively. (d) AVHRR-SSTin a 12 x 12km 2 areacentered at the Cariacosite. Heavy line (squares)overlay represents CARIACO (in situ) temperatureat 1-7 m depth.

(approximatelyJanuary-May) 250 mL seawaterwere filtered alyzer. Cystine standardwas used for calibration,and blank onto 25 mm glassfiber filters(GF/F) filters,while 500 mL were filters were used to determinebackground counts. filteredduring the restof the year. Three replicateswere taken Primary productivityobservations were made usinga mod- per depth exceptwhere biomasswas clearlyat its minimum, ifiedSteeman Nielsen (1952) NaH14CO3 uptake assay. One when only two were collected.Methanol extractionspermitted hour before sunrise, water from 1, 7, 15, 25, 35, 55, 75, and directcomparison of concentrationsestimated separately using 100 m was obtained.Water was poured directlyfrom the Ni- the Kishino et al. [1985] method for measuringthe light ab- skinbottle under low-lightconditions into 250 mL clear poly- sorptioncoefficient by particulatematter. carbonate bottles, which had been acid-washed, rinsed, and Suspendedparticulate organic carbon (POC) wasestimated soaked in deionized water for over 48 hours. Bottles were by filtering water samplesof known volume (generally2 L) rinsed three times before filling using a near-total fill. Four onto precombusted GF/F. The filters were folded inside clear polycarbonatebottles were filled from each depth. One cleaned tin disks and combusted in a PE2400 Elemental An- inoculatedbottle from each depth was wrappedin aluminum MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE 4531

foil to obtainthe dark14C uptake rates. An extrabottle for 1, from Mehrbachet al. [1973]. Values for the other equilibrium 15, 35, and 75 m was filled but not inoculatedto provide time constantswere taken from the review of Millero [1995]. zero (to) filter and seawaterblanks. The t o sampleswere kept On 11 cruises,total bacterialproduction was measured using in the dark in the laboratoryand were filtered after deploying a modification of the method of Kirchman [1993]. Discrete the floating incubationbuoy. samplesfrom 12 to 18 depthswere transferred(in triplicate), Each samplewas inoculatedunder low-lightconditions with under a nitrogen atmosphere,to 40 mL vials with Teflon-lined 1.0mL (4/aCi) of the •4Csodium bicarbonate working solu- butyl rubber septa. Each vial was allowed to overflow two to tion.A 200/aLaliquot for countingtotal added •4C activity was three vial volumesand was sealedwithout headspaceto main- removed from one of the three bottles from each depth and tain ambient redox conditions.Hydrogen 3-1eucine(10 nM placed in a 20 mL glassscintillation vial containing250 /aL final concentration)was introducedinto each vial via syringe. ethanolamineand 10 mL of liquid scintillationcocktail (Cyto- Sampleswere incubated in the dark for 8-10 hours in water ScintTM).The mixturewas held at 5øCuntil subsequentliquid baths maintained at in situ temperatures.Incubations were scintillationanalysis onshore. A 50/aLaliquot of the14C work- terminated by transferring the contents of an entire vial to ing solutionin a vial with ethanolamine(250/aL) and scintil- tubescontaining trichloroacetic acid (5% final concentration), lation cocktail was counted for reference. which lysedcells and precipitatedmacromolecules. Acidified A Licor photosyntheticallyactive radiation (PAR) integrator and chilled sampleswere processedas describedby Kirchman was usedto estimatephotoperiod length. Between December [1993] upon return to EDIMAR. Bacterial production was 1995 and November 1996, sampleswere incubatedfrom 0600 estimatedby measuring3H-leucine incorporation into precip- to 1000LT. StartingDecember 1996, the protocolwas changed itated protein after subtractingactivity of control samplessac- to incubatebetween 0700 and 1100 LT, which more accurately rificed immediatelyafter isotopeaddition to correctfor abiotic representsone third of the daily photoperiodand one third of processes.Carbon productionwas derived by assumingcon- the total energyreceived in 1 day at 10ø30'N,as verified with a stant proportionsof leucine and protein in bacterial biomass Licor PAR light sensor.Incubation times were relativelyshort [Kirchman,1993]. Bacterial production on an arealbasis (mg C becauseof the potentiallyhigh productivity (>1000 mg m-2 m-2 d-1) wascalculated from volumetric rates using trapezoi- d-•) of thiscontinental margin. Upon completion of theincu- dal interpolationsbetween discrete depth intervals. Acetate uptake rate constantswere measuredusing the ra- bations,sample bottles were removedfrom the mooring and diotracer method describedby Wrightand Hobble [1966] and placedin labeleddark plasticbags. A 50 mL aliquotwas with- Lee [1992]. Water sampleswere collectedas describedabove drawnfrom eachproductivity bottle and was filtered onto a 25 for the bacterialproduction assay. One hundredmicroliters of mm Whatman GF/F, which was rinsed with 0.25 mL 0.5 N HC1, a nitrogen-purgedsolution of 14C-acetate(4.4 x 106 dpm placedin a 20 mL glassscintillation vial, sealed,and held at 5øC mL-•) wereadded to eachvial using a gastight syringe. Sam- until subsequentprocessing onshore. Immediately upon return pleswere incubatedin the dark at in situ temperatures.At four to the shore lab and within 15 hours of sample collection,10 time intervals(between 6 and 12 hours),vials were sacrificed mL of liquid scintillationcocktail were added to vials with by filtering duplicate5 mL subsamplesthrough 0.22/am poly- filters. The vials were refrigerateduntil analysisat Empresas carbonateNuclepore filters to determineincorporation. Filters Polar in Caracas, Venezuela. were placed in 20 mL scintillationvials containing5 mL Op- Carbon uptake calculationsfollowed the standardformula- tiflor scintillation liquid. The remaining 30 mL sample was tion outlinedin the Joint Global Ocean Flux Study(JGOFS) killed by addition of 0.5 mL 10 N KOH. The amount of acetate manual [United Nations Educational,Scientific, and Cultural respiredto carbon dioxidewas measuredlater in the labora- Organization(UNESCO), 1994], taking into considerationa tory by acidifyingthe samplein a closedflask with filter soaked (very low) quenchingcurve. To obtain the daily productivity in 2 N KOH [Wrightand Hobbie, 1966;Lee, 1992]. rate, a photoperiodscaling factor that varied between1.89 and Settling carbon flux was measuredwith sedimenttraps. A 3.90, accordingto the fraction of the energy received during mooringwith four Mark-VII automatedsediment traps [Honjo the incubationperiod relativeto the total energyreceived in a and Doherty,1988] hasbeen deployedat the CARIACO site in day, was used. This was determined by the shipboardPAR 1400 m of water sinceNovember 1995. The mooring has been photometer. Depth-integrated and annual production esti- retrieved and redeployedevery 6 months,in May and Novem- mateswere derived by trapezoidal integration over the upper ber. Trapswere positionedat approximately275, 455, 930, and 100 m and then across months. 1225 m within the quiescentflow regime below sill depth. The Total dissolvedinorganic carbon (TCO2, hereinafter here cone-shapedtraps had an apertureof ---0.5m 2 crosssection referredto asDIC) wascalculated from measurementsof total and had a baffle designedto minimize advectivedisturbances. alkalinity and of pH on the total hydrogenion concentration Descendingparticles were funneled into a samplejar at its scaleat 25øC.These estimatesare performedusing the precise base. Each trap had 13 samplingjars filled with high-salinity spectrophotometricdye methods developed by Robert-Baldoet water, 12 of which were poisonedwith formalin. The jars were al. [1985] and Byrne and Breland [1989] and modified from sequentiallyrotated beneath the trap by a microprocessor- Claytonand Byrne [1993] and Breland and Byrne [1993]. A controlledstepping motor, which allowsfor samplecollection single-beamspectrophotometer (Ocean Optics) was used. The over specifiedtime intervals.We synchronizedthe computers pH wascalculated using the originalequations plus an additive to collectsamples for concurrent2 week periodsat eachdepth. factor of 0.0047, which results from a correction to the buffers Upon retrieval,samples were split, and subsampleswere sub- usedin the originalpaper [DelVallsand Dickson,1998]. These jected to a numberof routinegeochemical analyses to estimate methodscircumvent the problem that ariseswhen potentio- carbonate,organic carbon, nitrogen, and biogenicsilica fluxes. metric electrodes are transferred from dilute buffers to seawa- To estimateorganic carbon, samples were dried,homogenized, ter samplesbecause of the sample'shigh ionic strength.Total weighed, and placed in silver boats. To eliminate inorganic DIC was calculatedusing the K1 and K2 equilibriumconstants carbon,boats were placedin a dessicatoralongside a beaker of 4532 MULLER-KARGER ET AL.' CARIACO BASIN PRIMARY PRODUCTION CYCLE

A) Another subsamplewas used for radioisotopeand stableiso- tope determinationsand organicgeochemistry. Approximately half of each sampleremained availablefor additionalstudy.

3. Results 3.1. Trade Wind Figure 3 showsvariations in the wind experiencedat San- 200 tiago Madrio Airport on Margarita Island, about 80 km to the northeastof the CARIACO station.The zonal componentof 250 the wind (u) wasalmost identical to that observedabout 15 km to the west at Punta de Piedras. At both locations the zonal a00 •d FM•Md d•SO•d FMAMdd•SO• componentdominated variations in the wind, showingmarked seasonalitywith values ranging from 4 m s-• betweenabout Augustand January to 10m s-• betweenabout February and ' o •••369 • June.Muller-Karger and Aparicio [1994] compared the windsat 50• •• 36.85• • Margarita Island with thoseobserved at Orchila Island 200 km to the northwest.They concludedthat variationsin the zonal componentof wind at Margarita are representativeof those observedover the region. The meridionalcomponent of the wind (v) observedat the airport in Margarita wasweaker, varyingbetween about 0 and 1 m s-•. Thiscomponent showed cyclic shifts with a periodof about 3 months(Figure 3). Muller-Kargerand Aparicio [1994] found that at both Margarita and Orchila Islandsthe meridi-

300•, • • }, • •,,, , , , , • , v, onal componentof the wind stresswas usually <10% of the NDJ FMAMJ JASONDJ FMAMJJASOND zonal componentof stress,particularly if averagedover peri- C) ods >1 month. ø'11• • .5.•4• •4.4•.••• • 4.•, The tradewinds were relatively weak (order of 4 m s-•) [2 3'6• 3. • 3.6 when we initiated the CARIACO programin November 1995 (Figure3). Windsincreased in Decemberand January,peaked at 9-10 m s-• in May 1996,but decreasedrapidly thereafter, droppingto <5 m s-• againin June1996. Zonal winds re- mainedrelatively weak (u • 5 m s-•) in February-April1997, but the meridional componentshowed substantial northward intensification(v > 2 m s-•) at thistime. Zonalwinds in- creasedto about9 m s-• in May-June,with a nil meridional •o •o.• component,but then u decreasedand v increasedafter July. Thetrade winds remained weak (<5 m s-•) throughthe rest of R• FM•M•d•SO• FM•Mdd•SOR• the year, but they exhibitedan anomalousnorthward compo- nentover this extended period (•1-2 m s-•). D) 0 I 1 ? _ ,•x'"'•2•o,,,70.,,l/w-, )•'•b • •o9• •/] (-" 3.2. Hydrography /'"(• ,,/ ) o 50 t 2130 2196 / BetweenNovember 1995 and March 1996, surfacetemper- 2•3.o 11b'••5 (.A-' 211')• atures decreasedfrom 27.5ø to 23.0øC(Figure 3). The 21øC isothermmigrated from a depth of •130 m in November 1995 to •30 m by April 1996 (Figure 4a). In May 1996 the 22øC isothermreached the surface.A peak in SST of •29.0øC was • ,2•302210••230•221• observedin September 1996. In 1997 an SST minimum of 200• 21.5øC was observed in March as a result of the 21øC isotherm shoalingas early as January-February(Figure 4a). In both yearsan SST decreasewas observedin July and August (Fig- ures 3 and 4a). •d FM•Mdd•SO•d FM•Mdd•SO• Features noted in the temperature data are evident in the other hydrographicparameters. From November1995 through Figure 4. Seasonalvariation in hydrographicproperties at January1996 the salinitymaximum (36.9) rosefrom •100 m to the CARIACO site' (a) temperature (øC), (b) salini•, (c) the surface,and high surfacesalinities were observeduntil May o•gen (mL L-•), and(d) totalDIC (•mol kg-•). 1996 (Figure 4b). A relativelyfresh surfacelayer (•36.3) appeared during August-October, while the salinity maximum concentratedHC1 and put under a vacuum overnight. The retreated to 50-100 m. Salinityisopleths turned upward again sampleswere then folded and combustedin a PE2400 Elemen- after October 1996, and the salinity maximum (>36.75) tal Analyzer. Cystine standard was used for calibration, and reachedthe surfaceagain in January1997. This situationper- blank boats were used for determining backgroundcounts. sistedthrough March. Surface salinitiesbetween August and MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE 4533

Octoberwere fresher(<36.5), and temperatureswere warmer, increasedtenfold during Januarythrough June of both years, coincidentwith the onset of the rainy season. withvalues reaching 15 mg m-3 h-1 andon occasion20-40 Intrusions of Caribbean water, identified on the basis of mgm -3 h-1 in theupper 7 m. The changein incubationtime, oxygendata from the CTD profilerand the discrete(bottle) 02 from 0600-1000 to 0700-1100 LT, starting in January 1996 and DIC data (Figures4c and 4d), were observedin January causeda small decreaseonly in the surface (1 m) primary 1997 at 200-220 m, in June 1997 at 300 m, in August 1997 at productivityestimates. 275 m, and in December 1997 at 220-240 m. The January 1997 Depth-integrated(0-100 m) primary productionchanged eventwas not seenin the Januarybottle data but did appear in substantiallyover the course of our 1996-1997 field effort the February bottle data; it is possiblethat in January the (Figure6), in part becausethe annualcycle was punctuated by bottle at 200 m missedthis feature. In the monthsfollowing an strong events.The average depth-integratedproduction be- event the anomalous water mass could still be detected. tweenJanuary and May 1996was about 2800 mg C m-2 d-1, When an intrusion of water from outside the basin was whichwas heavily influenced by the May 1996upwelling event. detected, waters from depths above the intrusion were ob- Duringthe sameperiod in 1997it wasabout 2600 mg C m-2 servedcloser to the surface.This canbe seenin Figures4c and d-1. Productivityduring June-December was lower (---1000 4d, whereisolines of oxygenand DIC are relativelyflat in early mg C m-2 d-• in 1996and ---860mg C m-2 d-1 in 1997) 1997 below depthsof 200 m, while above those depths,water (Figure 6). Integrated production in 1996 was ---500 (not containinglow oxygenvalues and high DIC are raised toward countingMay) to 690 gC m-2 yr-1 (if May is included).In the surface,tracing the effectsof an intrusionin January.The comparison,production in 1997was between 480 (not counting patternsof intrusion and the effectson hydrographyand geo- January)and 540 gC m-2 yr-1 (countingJanuary). chemistryare presentedin more detailby Scrantonet al. [2001] The assimilationnumber pB (amountof carbonfixed per and Y. Astor et al. (manuscriptin preparation,2001). unitchlorophyll) is shownin Figure7. PB at 7 m wastypically A strongseasonal cycle was observedboth for DIC (Figure the highestthroughout the seriesand was likely the closestto 4d) and near-surfaceca 2 fugacity.The near-surfacefugacity Pm•-B AveragepB at 7 m during January-Junewas about 7 mg values during July-October of both 1996 and 1997 were typi- C mg Ch1-1h -1 (range4.6-9.0 mg C mg Ch1-1h-i), and cally in the 410-415 /.ratmrange. During January-May 1996 duringJuly-November it was12 mg C mg Ch1-1h-1 (range we observedsurface values in the 385-390/.ratm range. During 7.5-14mg C mg Ch1-1h-i). Surface(1 m) PU wasgenerally January-May 1997,values were smaller,typically between 370 6-14 mg C mg Ch1-1h -1 prior to December1996 and de- and 390/.ratm, with one observationof---335/.ratm in March. creasedto 4-9 mg C mg Ch1-1h -1 afterward(not shown). 3.3. Chlorophyll and Primary Production This is likely due to increasedphotoinhibition in the surface A subsurfacechlorophyll a maximum(---0.4 mg m-3) was incubation bottles causedby the later incubation times. This located near 55 m in November 1995, but it rose to 25-30 m in effect was not apparent in deeper samples. December(Figure 5a). BetweenJanuary and March 1996, 3.4. Vertical Organic Carbon Flux valuesof 1.5-3.0mg m -3 chlorophylla were observed in the The first 6 monthsof sedimenttrap sampleswere retrieved upper 15 m. In April and May, values in the upper 25 m exceeded6 mg m-3 andreached 8 mg m-3 at 15 m. Phyto- successfullyin May 1996. Unfortunately, samplesfor the fol- lowing6 monthswere lost.The four trapsclogged immediately plankton concentrationsthen decreased,reaching <0.2 mg m-3 by Juneat the surface.A smallchlorophyll maximum after redeploymentof the mooring,resulting in a data gap in (---0.4-0.8mg m -3) remainedat about35-55 m throughOc- the sediment flux series from May-November 1996. We at- tober. tribute this to unusuallyhigh particlefluxes caused by the large This cyclewas repeatedin 1996-1997, but the bloom started planktonbloom of May 1996(see Figures 5 and 6). Subsequent earlierthan in 1995-1996,with biomass values of ---4mg m -3 redeployments,in November 1996 and May 1997, were suc- cessful. seen at 25 m in November 1996 (Figure 5a). After a brief declinein chlorophylla to <1 mgm -3 in the upper20 m in The vertical flux of organic carbon followed a regular pat- December1996 a bloomreformed in January1997 with 3-5 mg tern, with minima between Septemberand January,maxima m-3 in the upper25 m andlasted through May. betweenFebruary and May, and similartemporal variability at Surface particulate organic matter also increasedtenfold all four depths (Figure 8). In 1996, carbon flux showeda betweenNovember 1995 (40-50 mgC m-3; particulate organic minimumof ---0.01gC m-2 d-1 in Januaryat all four trap nitrogen(PAN) ---2.5 mgNm -3) andMarch 1996 (500-800 depths(Figure 8). Flux increasedin the subsequent2 months mgCm -3, PaN > 60 mgNm-3). Pac andPaN reacheda to peakin March1996 at 0.17gC m-2 d-1 at 275 m and at peak during May 1996, but unlike chlorophyll,which declined about0.06 gC m-2 d-1 at 1225m. The largestorganic carbon in June 1996, Pac (and dissolvedorganic carbon (DaC)) flux recordedin this 2 year period was seenin the 1225 m trap remainedhigh until July 1996. We also commonlyobserved a samplefor early July 1997 and was associatedwith a turbidite minimumin Pac (10-20mgC m -3) justabove the oxic-anoxic generated by an earthquakethat occurredalong the Venezu- interface throughoutthe year. Concentrationsaround the in- elan coast [Thunellet al., 1999]. Otherwise,the highestflux terfaceat 275m werehigher but variable (20-70 mgCm-3). (0.18gC m -2 d-1) wasobserved in the shallowesttrap in May Lowerbut also variable concentrations (25-50 mgC m -3) were 1997. In general,organic carbon flux decreasedfrom the shal- observedin deeperwaters (>275 m). lowest to the deepesttrap. The 275 m trap usuallyexceeded Primary productivityshowed marked seasonality(Figure the flux capturedat 455 m by 20-200% during January-May 5b). Betweenabout July and December,rates were <2.5 mg and frequentlyby 200-400% during July-Novemberperiods. m-3 h-1 in theupper 30 m, withlowest values (<1.0 mgm -3 However, in 27% of collectionsthe flux to the 455 m trap h-•) in Augustand September of both1996 and 1997. Deeper matchedor slightlyexceeded the flux observedin the shallow- waterstypically showed values of < 0.3 mg m-3 h-1. Rates est (275 m) trap. 4534 MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE

6.0 3O

40 4.0

5O 2.0

6O 1.o

0.3

o.1 9O

lOO

4O 3o 2O

..-- 4o 10

e 50

7O 0.5

8O 0.2

lOO

N D J F M A M J J A S O N D J F M A M J J A S O N D I I I I 1995 • 1996 • 1997 I I

Figure5. Seasonalvariation in (a) chlorophyllconcentration (mg m -3) and(b) primaryproduction (mg m-3 h-•) overthe upper100 m at the CARIACO sitefrom December 1995 through December 1997.

3.5. Bacterial Production activity(leucine incorporation and acetateuptake) near the Bacterialabundance, bacterial productivity, acetate concen- oxic-anoxicinterface often are comparableto rates in the sur- tration and acetate uptake rate constants,and dark carbon face mixed layer. fixation rates were measured on six cruises between November 1995 and December 1997, and bacterial productionwas measured on an additional five cruises. The distributions seen for 4. Discussion May 1997(Figures 9 and 10) were similarto thoseobserved on The CariacoBasin serves as a unique laboratoryto examine other dates.The most noticeablefeature in the profilesis the the sinkingflux of material becauseit forms a natural sediment low bacterial activity throughout the water column except at trap. Outside the Cariaco Basin, materials that sink from sur- the surface and at the oxic-anoxic interface. Rates of bacterial face waters and reach the deep parts of the CaribbeanBasin MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE 4535

8OOO •' 7000

• 6000 E o• 5000 .._>,,4000 I

•'000, 1000I iiJ I

i l

Month Figure6. CariacoBasin daily primary production integrated over the upper100 m (mgC m-2 d-•).

may be trapped there for periods of at least 50-800 years substantialand, indeed, redefines earlier estimatesas being IRedfieldet al., 1963; Kinder et al., 1985]. However, sinking minimumvalues. Our annualproduction estimates (500-690 material is quicklydispersed because of the strongCaribbean gC m-2 yr-•) arehigher than those reported previously from Current that washesthe South American continentalmargin. the vicinityof the CariacoBasin (200-400 g C m-2 yr-•) Becauseof the strongcurrents along this margin, this flux is [Ballesterand Margalef, 1965; Curl, 1960; Richards, 1960]. hard to measure.Waters in the CariacoBasin have comparable Thesevalues are comparableto thoseestimated for Monterey residencetimes [Deuser, 1973], but materialsinks near its place Bay(7 yearaverage of -460 gC m-2 yr-•) [seeOlivieri and of origin.In either casethe importantquantity is the amountof Chavez, 2000; Chavez, 1996] and considerablyhigher than material that is incorporatedinto the sedimentsbecause this those estimated for Georges Bank, the New York Shelf, and material may be sequesteredfor periods spanningmultiple theOregon Shelf (380, 300, and 190 gC m-2yr -•, respectively) centuriesor millenia [Haug et al., 1998; Hughenet al., 1998; [Walsh,1988]. Previous estimates of productionin the Cariaco Black et al., 1999]. Basin may be lower either becausethey were obtained by The results from the CARIACO series site demonstrate that indirectmethods [Richards, 1960; Curl, 1960] or derivedfrom productionalong continentalmargins in the tropics can be sparsetemporal data that missedproduction peaks [Ballester

16.00 - 3O

- 29 14.00

- 28

,- 12.00 - 27

r,.) lO.OO- 1 E I -26• i i

• 8.00 I 25 •

z I E I o 24 I- ..•co 6.00 -- ._ E ,_ 23 < 4.00 22

2.00 21

0.00 i i i i i i i i i i i i i i i i I i i i i i i 2O

Figure 7. Time seriesof water temperatureat 7 m measuredwith a CTD (solidline, circles(øC)) and of the assimilationnumber pB at 7 m (brokenline, squares) at the CARIACO stationbetween December 1995 and December 1997. 4536 MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE

200 ample,rapid coolingoccurred between October 1995 and Jan- uary 1996 and againbetween October 1996 and January1997, 150 prior to any significantincrease in zonalwind speedabove 5-6 100 m s-1. Similarly,every year, SST started increasing prior to, or coincidentwith, the onset of the seasonalweakening of the 50 trades,while zonal wind speeds were still in excessof 7 m s-1

o i i i i (Figure3). Even thoughthere appearedto be smalldecreases in SST associatedwith the cyclic increasein the northward componentof the trade wind (Figure 3), there was no clear evidencethat the larger seasonalchanges seen in the hydrog- Figure 8. Time seriesof biweeklyintegrations of organiccar- raphy and SST were driven solelyby the local wind. bonflux observed with sediment traps (mgC m -2 d-1) at 275 At this time, there is no conclusiveexplanation for the out- (open circles,solid line), 455 (squares,broken line), 930 (tri- of-phasechanges in hydrographyrelative to the wind. For the angles,stippled line), and 1225m (crosses)at the CARIACO site. CaribbeanSea a hypothesismay be that the vertical displace- ment of the SUW in the southernCaribbean is controlledby seasonalchanges in the geostrophicflow through the basin [see, e.g., Morrison and Nowlin, 1982; Morrison and Smith, and Margalef,1965]. Clearly, previousworkers have been un- 1990]. Such changesin the meridionalslope of the isopleths able to evaluatevariability in carbonfixation rates in the Cari- are likely connectedto remote processes.We may speculate aco Basin because few observations were available. that the circulation through the Lesser Antilles is strongly affectedby the seasonalswitch between strongcombined in- 4.1. Water Mass Exchange and Upwelling flow from the North Equatorial Current and the Guiana Cur- Why is the southernCaribbean S•ea • so productive?Up- rent betweenJanuary and June and weaker flow betweenJune welling along the southernCaribbean, and specificallyits im- and about December,when the North Equatorial Countercur- pact within the perimeter of Cariaco, has been discussedby rent and the North Brazil Current Retroflectionform [see Richards[1960], Margalef et al. [ 1960],Margalef [1965], Ballester Muller-Kargeret al., 1995, and references therein]. These and Margalef [1965], Morris et al. [1981], Febres-Ortegaand changeswould provide the correctphase relative to the wind in Herrera [1975],Herrera and Febres-Ortega[1975], and Walshet the southernCaribbean. In addition,the seasonalcycle in the al. [1999] among others. Studies with Coastal Zone Color Cariaco Basin may be superimposedon larger interannual Scanner (CZCS) imagery show that the most pronounced variationin the flow throughthe variouspassages in the Ca- blooms take place over the eastern portion of the Cariaco ribbeanSea [seeMurphy et al., 1999]. Basin [Muller-Kargeret al., 1989; Muller-Kargerand Aparicio, At shorter timescalesof 2-3 months, small variations in the 1994]. Both ocean color (CZCS and Sea-viewing Wide hydrographywere apparent,particularly in 1996.These varia- Field-of-viewSensor (SeaWiFS)) and AVHRR satellitedata tions includedrelatively small oscillationsin isotherms,isoha- showthat the upwellingplume extends over 4 x 104 km2 and lines, and oxygenconcentrations to a depth of about 150 m sometimes reaches >9 x 104 km 2. (Figure4). Theywere alsoreflected as roughlyiøC changesin Historicalsections of salinityacross the CaribbeanSea [Mor- the smoothedsatellite-derived SST series.Starting in 1997,the risonand Smith, 1990, and referencestherein] showa north- depth amplitudeof suchvariations was more pronounced,and south trend in salinity at the level of the salinity maximum markedchanges in oxygenand DIC were observedrepeatedly (---36.85) associatedwith SubtropicalUnderwater (SUW). between200 m and the surface.These variations are likely due, SUW lies between 100 and 200 m depth north of 14.5øNand shallowerthan 150 m southof 14.5øN.This upwardtilt in the SUW layer is emphasizedduring the first half of the year DarkC fixation,•vl C d• relative to the secondhalf of the year. Figure 4b showsthe 0.0 0.5 1.0 1.5 2.0 vertical excursion of the 36.85 isohaline in the Cariaco Basin, Beam attenuation coefficient BacterialProduction, pgL -• d'• which is at the southernextreme of this tilting layer.Within the SUW, nitrate is 5-10/aM [Morrisonand Nowlin, 1982], and as 00 1 2./,•0 0.50 0.55 0.60 0.65 thiswater is broughtto the surfacein the CariacoBasin during 0 • • the first half of the year, it leads to elevated nutrients in the 200 - E upper water column,thus stimulatingphytoplankton growth. __ The current conceptualmodel of upwellingin the southern 4•, • +H•S • Caribbean Sea directly associatessurface Ekman transport • • + Bact Prod. drivenby the trade windsto an upwardmigration of the ther- • o •-o OarkC moclinealong the (roughly)east-west oriented coast that de- finesthis continentalmargin. This conceptualmodel hasbeen 0 4 8 12 6 NO•, • 0 1 2 3 4 5 usedwithout rigoroustesting in studiesof modern processes I .... I .... I .... I .... I [e.g., Muller-Kargerand Aparicio, 1994] and paleo-oceano- 0 20 40 60 80 [Bacteria],x10 8 cells L-• graphic studies[e.g., Haug et al., 1998; Hughen et al., 1998; H•S, • Black et al., 1999]. Figure 9. (left) Vertical profilesof bacterialnet production, Upon closerexamination, however, some years show a lag dark carbonfixation, and nitrate and hydrogensulfide concen- betweenthe onsetof upwellingas tracedby a decreasein SST tration measuredin the CariacoBasin in May 1997. (right) in Figure 3 (or the rise in the SUW in Figure 4) and the Bacterialcell counts(circles) and continuousprofile from a seasonalstrengthening of the trade winds(Figure 3). For ex- SeaTec transmissometerfor May 1997. MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE 4537

0.00 0.04 0.08 O. 12 O. 16

100 _

E 200

400 / 5oo•// 60 • --o- C1 (Nov.1995) '(• • O13 (Nov.1996) 5 • C19(May 1997) 200900 i --0--C2• /'t rch1996) Figure 10. Vertical profilesof the total acetateuptake rate constant(respiration plus incorporation) in the CariacoBasin. This rate is higherthroughout the water columnduring upwelling (March and May) relative to nonupwellingseasons (November). The depthof the uptakerate maximumis usuallyclosely related to the depth of the oxic-anoxicinterface.

at least in part, to intrusionof Caribbeanwater into the Cari- observedas early as December-January,1-2 monthsprior to aco Basin at or near sill depth. substantialincreases in the westwardwind speed.Productivity The intrusionof deeperwater (below SUW) from the Ca- tracked temperature changes,which is a proxy for nutrient ribbeanSea into the CariacoBasin is dependenton thiswater availability.In 1997,production was sustainedat a higher level being presentat sill depth, a situationthat occursduring the and for a longerperiod relativeto 1996because of waterswith first half of the year [seeMorrison and Smith, 1990].When the higher nutrientsbeing pushedtoward the surfacefrom below influx is large, it may be classifiedas a ventilation event be- by a strongventilation event [see Scranton et al., 2001;Y. Astor causeit bringssubstantial oxygen into the deeper,oxygen-poor et al., manuscriptin preparation,2001]. watersin the basin.Intrusions are probablymore likely during The biomassspecific carbon uptake rates pB wereclose to times when the SUW layer tilts upward toward the southern the physiologicalmaximum, which suggeststhat phytoplankton Caribbeanmargin [e.g., Morrisonand Smith, 1990], i.e., be- were growingat near-maximumspecific growth rates through- tween about December and the followingJune. out the year and that there was no nutrient limitation in the There is at this time no definitive explanationfor the 2-3 CariacoBasin during the upwellingseason. Nevertheless, there month signalobserved in the 1996 hydrographynor for the wasa distinctseasonal variation in PB (Figure7). WhileBe- repeatedventilations of 1997.There is a 2-3 monthvariation in hrenfeldand Falkowski [1997] suggest that P• shoulddecrease the meridionalcomponent of the wind (Figure 3), whicheven aboveabout 20øC, P• at CARIACO increasedwith tempera- thoughis of a smallamplitude (-1-2 m s-Z),may play a role ture over the seasonaltemperature range of 21ø-29øC.This is in thesefluctuations (Figure 4). Another possiblescenario is likely to be due in part to the temperatureeffect describedby that the hydrographyof the shelfwas affectedby eddiesmov- Eppley [1972] and not to a nutrient effect. At CARIACO, ing west in the Caribbean Current. A synopticview of these nutrientsfollow an inverserelationship to SST,and P• should eddies and their relation to the wider circulation in the Carib- decrease with a decrease in nutrient concentrations. There is bean Sea has emergedfrom recent numerical models of the alsoan inverserelationship between P• andbiomass, suggest- CaribbeanBasin and from satellitealtimeter data [Murphyet ing that successionfavored organismswith higher specific al., 1999; Canon and Chao, 1999]. Once an intrusionor venti- growthrates during periods of lowernutrient availability. PB in lation event occurs,the local wind assistsin further elevating 1997 showedincreased variability relative to the 1996 obser- the SUW layer and in enhancingmixing, thereby lowering the vationspossibly because of an interactionbetween increased temperaturefurther alongthe coastline.A detaileddiscussion photoinhibitioncaused by the changein incubationtime and of the deep geochemicaleffects of intrusionsis providedby nutrient supplydue to increased(but variable)upwelling. Scrantonet al. [2001], and a more detailed descriptionof the hydrographiceffects is presentedby Y. Astor et al. (manuscript 4.3. Sinking Organic Carbon Flux in preparation,2001). Because of their high productivity, surface waters are a sourceof organiccarbon to deeper water. While we have not 4.2. Primary Production measuredvertical velocities of the water beingupwelled at the Similar to the pattern observedin SST, seasonalchanges in CARIACO station,a numericalmodel developedby Walshet primary productivityprecede changesin the wind. Figure 5b al. [1999]suggested upward vertical speeds of 1-2 m d- z under showsthat higher surface productivity (>2 mgm -3 h-z) was westwardwinds of 6-10 m s-•, comparedto 5-10 m d-z about 4538 MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE

160 - 225 rn

140- 455 m •120- • 100- E 80- 93O m 0

E 1,225 m LL X 20-

0 I I I I I 0 500 1000 1500 2000 2500 3000

PP [rag C m-2 d-l] Figure 11. Relationshipbetween monthly values of dailydepth-integrated primary production (mgC m -2 d-•) andmonthly means of the biweeklyintegrations of organiccarbon flux (mgCm -2 d-1) at 275 (dia- monds),455 (squares),930 (triangles),and 1225m (crosses)at the CARIACO site.Lines are leastsquares regressions.January 1996 and 1997 were not included in the regressionsas these were consideredto be possibleoutliers in production.

20 km closer to shore along the Venezuelan coast. With a Acetate uptake rate constants(respiration plus incorpora- prescribedthermocline depth representingconditions for the tion; Figure 10) alsoprovide a measureof organicremineral- first half of the year, i.e., when the SUW layer is closerto the ization under low-oxygenconditions [Henrichs, 1993]. Rate surfacein the southernCaribbean, results suggested that the constantprofiles consistently showed a strongmaximum at the vertical upward displacementvelocities decreased to the north oxic-anoxicinterface. In addition,rate constantsfor upwelling (and west) of the CARIACO station.Nil displacementswere seasons(March and May) were higher than rate constants simulated near the sill, which is located within 100 km of the measuredduring November and July (data for Julynot shown). station.Therefore, even if the actual sinkingrates for partic- The abrupt deepening of the acetate uptake rate constant ulate matter at CARIACO were much less than 100 m d -• maximum in November 1997 as comparedto other dateswas [Deuseret al., 1990],it is likely that the sedimenttraps, located associatedwith a similardeepening of the oxic-anoxicinterface below the 140 m sill depth, are within the statisticalcone of caused by the repeated intrusions of 1997 describedabove sinkingflux [Deuseret al., 1990] that originatesat the station. [Scrantonet al., 2001]. Thesedata supportthe conclusionsfrom We assumethat vertical settlingof particlesdominates hori- the sedimenttraps that more organicmatter is respiredduring zontal losses within this statistical funnel. upwellingperiods than during quiescentperiods at all depths. As mentioned above, the settlingPOC flux in the Cariaco On the basisof the rate of decreaseof fluxeswith depth it Basin followed an annual cyclesimilar to that of surfacepro- appearsthat almost95% of the particulatecarbon produced by duction (see Figures6 and 8). The flux of opaline silicawas phytoplanktonnear the surfaceis consumedor regenerated tightly coupled with organic carbon flux, and the C/N and within the upper 275 m of the water column. This can be 15•3Corgvalues for the sedimenttrap samples mostly varied comparedto about 86-98% in Monterey Bay [Pilskalnet al., from 5.8 to 7.9 vol/vol and from - 18.0 to -22.5%o, respec- 1996] and 84-93% in the Santa Barbara Basin [Thunell, tively [Thunellet al., 2000]. This is consistentwith the idea that 1998b].Our observedcarbon fluxes are in excellentagreement the material reachingthe bottom of Cariacowas diatomaceous (r 2 -- 0.87) withpredicted fluxes determined using the Pace in origin and that the flux of terrigenousmaterial into the et al. [1987]model developed for the easternPacific Ocean and middle of the Cariaco Basin was small at all times of year suggestthat organicmatter degradationoccurring in an anoxic [Thunellet al., 2000]. water column is as efficient as that occurring in well- Vertical carbonflux was directlyproportional to integrated oxygenatedwaters [see Thunellet al., 2000]. production(Figure 11), with the magnitudeof the flux usually SuspendedPOC alsodecreases sharply with deptheven dur- decreasingwith depth. Carbon flux at 275 m was on the aver- ing the high productionseason, from valuesin excessof 200 mg age 5.6% of integratedprimary production. This decreasedto m-3 in the upper25 m to <50 mg m-3 below50 m. The 5.1% at 455 m, althoughin February,March, and April 1996it verticaldistribution of bacteriaand bacterial activity (Figure 9) was essentiallythe sameat both 275 and 455 m at 6-7%. The suggeststhat an activebiological population is located at the averageproportion decreased to 2.8% at 930 m and to 1.7% at oxic-anoxicinterface. While not associatedwith any apparent 1225 m. One of the effects of the decrease in the ratio of increasein bacterial biomass,a secondarymaximum in bacte- settling carbon flux to primary productivitywith increasing ria productivitybetween 100 and 160 m suggeststhat intensive depth is a marked changein the differenceof the flux between carboncycling also takes place in this depth range(Figure 9). 225 and 1225 m with time. SpecificallyOctober, November, Peaks in bacterial activity are commonly found at the oxic- and December(low flux months)feature the smallestvertical anoxictransition zone in many environments. differencesin the flux, while March-May (high flux months) While most of the bacterial productionoccurs in the upper showlarge differences(see Figure 8). This impliesthat during 275 m, a significantfraction occursbelow this depth. Our data the springbloom, remineralizationof the sinkingorganic mat- showthat 4-37% of integratedbacterial production (mean = ter is more efficient than during the nonupwellingseason. 17%; standarddeviation = 11) can occurbelow the depth of MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE 4539

the top sedimenttrap (275 m). This is in contrastwith the Cariaco Basin and adjacent areas,which can be estimatedto sedimenttrap data, which imply that on average,95% of the cover >4 x 104 km 2 and sometimes >9 x 104 km 2 from labile exportproduction is consumedor regeneratedat depths infraredsatellite data, between 0.4 x 10•2 and1.0 x 10TM gC shallower than 275 m and therefore that <5% sinks below this yr-• (4 x 105and 1 x 106 t of C yr-•) maybe deliveredto depth. However, bulk carbon deliveryto the 455 m trap ex- sediments of the southeastern Caribbean Sea. ceededthat deliveredto the 275 m trap in 27% of our obser- The final questionis What is the sourceof the CO2 that vations. Furthermore, on averagewe observe as much, and providesthe carbonthat is depositedin sedimentsof the Cari- frequentlymore, bacterialproduction between 275 and 450 m, aco Basinsystem? In open oceansurface water, carboncycling belowthe oxic-anoxicinterface (17% integratedbacterial pro- tendsto depletenear-surface CO2, resultingin a net flux from duction),as in the overlying175 m (14% integratedbacterial the atmosphereinto the ocean.However, in a systemlike the production). Cariaco Basin the upwellingprocess brings water enriched in Clearly, the interface is a region of vigorouscarbon cycling DIC and with a high fugacityof CO2 into the euphoticzone. fueled by labile carbon. We proposethat this carbon is not Even thoughCO2 fugacityin surfacewaters decreasedduring necessarilyprovided by surface-derivedexport production. the highlyproductive upwelling season, aquatic CO2 remained While bulk carbon delivery to the deepesttrap appeared to near or abovethe atmosphericpartial pressureof ---360 gatm conform to open water predictions[Thunell et al., 2000], the essentiallyat all times. Thus, counterintuitively,the southeast- compositionand sourceterms for this material are not well ern Caribbeanis a sourceof CO2 on a year-roundbasis in spite defined.Rapid heterotrophicactivity at the oxic-anoxicinter- of the substantialprimary productionoccurring there. This is face suggestsintroduction of fresh labile organic matter at importantsince intensification of the "biologicalpump" is of- depth either throughvertical migratorsor possiblythrough in ten consideredto be a key mechanismfor drawing down at- situ productionby chemoautotrophs.Biologically enhanced mosphericCO2. layers establishedby either accumulationof particles along This poses an interesting question regarding the role of isopycnalsor lateral advectionof more productivewaters are continental margins in the global carbon cycle. Viewed as a less likely explanationsin the Cariaco Basin becausewaters purelyvertical system,the Cariacoupwelling system is a source below200 m havenearly constant density and watersacross the of CO2 to the atmosphereon a year-roundbasis. In this respect interface consistentlyhave a uniform temperature-salinity the CARIACO basin, as an analog for continentalmargins, (T-S) signature.Furthermore, these biologicallyenhanced doesnot appear to provide a ready mechanismfor sequestra- layersare persistentfeatures, and flux anomaliesdo not appear tion of anthropogenicatmospheric CO2 in spite of the high to correspondto the intrusionsreported above. If in situ pro- primary productivity and organic carbon sedimentationrates ductionat the interfacecontributes significantly to the vertical seenthere. However, the significanceof the downwardflux of flux, then total fluxes to the seafloor and the chemical and particulatecarbon at this continentalmargin lies in its role as stableisotopic composition of this material may be decoupled a sink for CO2 capturedwithin SUW as it is formed in the from surfaceprocesses. North Atlantic. Because SUW is advected into the Cariaco Eppleyand Peterson[1979] found that in the deep oceanthe system,it is a sourceof new nutrientsas well as of new CO2. fraction of new to total productiongrows as total production Some of this carbonsinks when incorporatedinto the partic- increases.One of our early goalsin the CARIACO program ulate flux, and someis vented to the atmosphereas CO2. The was to test this hypothesisat our continentalmargin site. For upwellingprocess along the southernCaribbean margin there- this purposewe may assumethat the verticalorganic carbon fore representsa remote sink for CO2 previouslysequestered flux observedat the CARIACO station representsnew pro- from the surfacewithin the North Atlantic gyre in the area of duction(export production equals new production)and that formation of the SUW. We may extend this analogyto other our productivityestimates are an approximationof total pro- continentalmargins and view upwellingareas as both nutrient duction. While there was seasonalvariation in the settling traps and areasof CO2 outgassing. organiccarbon flux, there was no noticeableseasonality in the proportionof vertical flux to primary production.Indeed, the 5. Conclusions e ratio was relativelyconstant to within a few percentat any of the trap depths(Figure 11). In Monterey Bay [Pilskalnet al., The CARIACO results show that primary productivityof 1996] and in the Santa Barbara Basin [Thunell, 1998b] the e this tropical ocean margin is extremelyhigh, with the assimi- ratio decreasedas primary productivityincreased. At these lation ratio at its temperaturemaximum all year round. Previ- sitesoff California this result was interpreted as an effect of ous estimatesof annualproduction for the CariacoBasin were lateral advection.In any event, this suggeststhat Eppleyand low by factors of 2-3 becauseof undersamplingof the high Peterson's[1979] relationship applies primarily to areas of productionevents driven by transient physicalstructures. In- lower productivityand not to continentalmargins. deed, upwelling of SUW and ventilation of the Cariaco Basin The trap and primaryproductivity observations suggest then lead to increasedprimary productionin surfacewaters and to that between10 and 11 gC m-2 are deliveredto the bottom increasedsettling flux of particulatecarbon. SurfaceCO2 fu- sedimentof Cariaco everyyear. This agreeswell with the 7-8 gacityis lower during the upwellingseason relative to nonup- gC m-2 seenat 2300m at othercontinental margin locations welling periods, reflecting the lower temperature, complete [cf. Walsh, 1991; Pilskaln et al., 1996; Etcheberet al., 1996; nutrient utilization, and upwelling of water that is nutrient- Thunell,1998b]. Preservation of organicmatter within the an- enriched relative to carbon. This is further evidence that the oxic environment of the Cariaco Basin therefore does not seem CariacoBasin serves as regionalcarbon sink for CO2 entrained to be either enhanced or diminished relative to that of other at remote locations,such as in midlatitude gyreswhere the continentalmargins [Thunell et al., 2000]. SUW is formed. If theseannual rates of carbondelivery to the bottom(10-11 Estimates of rates of carbon remineralization in the water gC m-2 yr-•) applyover the upwelling plume that covers the columnof the CariacoBasin from sedimenttraps suggest remi- 4540 MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE neralization is comparableto that seen in the open ocean in sidad de Oriente, assistedin the bacterial productivityassays. The spite of the suboxic-anoxicnature of much of the water col- BiotechnologyLaboratory of EmpresasPolar in Caracas,Venezuela, haskindly allowed access to their scintillationcounter for our primary umn. Data from direct microbial measurementssuggest that productivityassessments. State University of New York Marine Sci- remineralizationis largelylocalized in the surfacemixed layer ences Research Center contribution number 1196. and near the suboxic-anoxic interface. These data have also shownthat remineralization rates vary somewhatwith season References (and primaryproduction rate). The exportrates (normalized Bacon, M.P., P. G. Brewer, D. W. Spencer,J. W. Murray, and J. to total primaryproduction) seem to varylittle in the transition Goddard, Lead-210, polonium-2210,manganese and iron in the from upwellingto nonupwellingseason. This is different from Cariaco Trench, Deep Sea Res.,Part A, 27, 119-135, 1980. the Eppleyand Peterson[1979] finding that the fraction of new Ballester,A., and R. Margalef, Produccionprimaria, Memoria, XXV, to total production increasesas total productionfor the deep 209-221, 1965. Behrenfeld,M. J., and P. G. Falkowski,Photosynthetic rates derived ocean increases but is similar to what others have found more from satellite-basedchlorophyll concentration, Limnol. Oceanogr., recentlyin assessingexport production.An additionalfinding 42, 1-20, 1997. is that the oxic-anoxicboundary is a zone of both high micro- Berner, R. A., Comments on the role of marine sediment burial as a bial heterotrophicand autotrophicprocesses. repositoryfor anthropogenicCO2, Global Biogeochem.Cycles, 6, 1-2, 1992. Results suggestthat upwelling along the southeasternCa- Black,D. E., L. C. Peterson,J. T. Overpeck,A. Kaplan,M. N. Evans, ribbean Sea margin is not only responsiveto seasonaltrade and M. Kashgarian,Eight centuriesof North Atlantic Ocean atmo- wind intensification.It is possiblethat there is a more direct spherevariability, Science, 286, 1709-1713, 1999. associationwith the meridional tilting of the thermoclinein the Breland,J. A. II, andR. H. Byrne,Spectrophotometric procedures for CaribbeanSea, which reflectsthe remotely forced changesin determinationof seawater alkalinityusing bromocresol green, Deep Sea Res., Part I, 40, 629-641, 1993. geostrophicflow through the basin. Transient eddies in the Brown, O. B., J. W. Brown, and R. H. Evans, Calibration of advanced Caribbean and fluctuationsin the meridional componentof very high resolutionradiometer infrared observations,J. Geophys. the local wind are likely to play a role in ventilation events. Res., 90, 11,667-11,678, 1985. This suggeststhat this site can be used to assessthe impact of Byrne,R. H., and J. A. Breland,High precisionmultiwavelength pH determinationsin seawater usingcresol red, Deep SeaRes., Part A, larger-scalephenomena, which then would alsohelp to under- 36, 803-810, 1989. stand better the paleo-oceanographicrecord. Carton, J. A., and Y. Chao, Caribbean Sea eddies inferred from The CARIACO site is located above sediments that have TOPEX/Poseidon altimetry and a 1/6ø Atlantic Ocean model simu- been shownto retain a high-qualityrecord of sedimentdepo- lation, J. Geophys.Res., 104, 7743-7752, 1999. sitionfor at least the past 12,000years [Hughenet al., 1996].A Chavez,F. P., Forcing and biologicalimpact of onsetof the 1992 E1 Nifio in central California, Geophys.Res. Lett., 23, 265-268, 1996. recent studyof these sediments[Haug et al., 1998] concluded Clayton, T. D., and R. H. Byrne, Spectrophotometricseawater pH that nitrogenfixation plays an importantrole in controllingthe measurements:Total hydrogen ion concentrationscale calibration settlingcarbon flux in the modernCariaco Basin and that this of m-cresolpurple and at-sea results,Deep Sea Res., Part I, 40, is a consequenceof high denitrificationrates. The CARIACO 2115-2129, 1993. Curl, H., Primaryproduction measurements in the north coastalwaters study found the oxic-anoxicinterface at about 250 m, and of South America, Deep Sea Res. Oceanogr.Abstr., 7, 183-189, 1960. nitrate data show an indication of denitrification as shallow as DelValls, T. A., and A. G. Dickson, The pH of buffers based on about 130 m during July-November,the nonupwellingperiod. 2-amino-2-hydroxymethyl-l,3-propanediol('tris') in syntheticsea However, during upwellingthe sourceof nutrientsfor surface water, Deep Sea Res.,Part I, 45, 1541-1554, 1998. production is water associatedwith the SUW, which has an Deuser, W. G., Cariaco Trench: Oxidation of organic matter and residencetime of anoxic water, Nature, 242, 601-603, 1973. elevatednitrate content.Therefore it is likely that mechanisms Deuser, W. G., F. E. Muller-Karger, R. H. Evans,O. B. Brown, W. E. otherthan local nitrogen fixation may control the/5•5N signa- Esaias, and G. C. Feldman, Surface-oceancolor and deep-ocean ture detectedby Haug et al. [1998] in the Cariaco Basin sedi- carbon flux: How close a connection?,Deep Sea Res., Part A, 37, ments. 1331-1343, 1990. Eppley, R. W., Temperature and phytoplanktongrowth in the sea, The results suggestthat the Cariaco Basin servesas a re- Fish. Bull., 70, 1063-1085, 1972. corder of both ancient and modern changesoccurring over Eppley,R. W., and B. J. Peterson,Particulate organic matter flux and regional and possiblybasin scales.The questionremains of planktonicnew productionin the deepocean, Nature, 282, 677-680, howwidely someof thesefindings can be extrapolatedto other 1979. margins. Etcheber, H., S. Heussner, O. Weber, A. Dinet, X. Durrieu de Mad- ron, A. Monaco, R. Buscail,and J. C. Miquel, Organiccarbon fluxes and sediment biogeochemistryon the French Mediterranean and Atlantic margins, in Particle Flux in the Ocean, SCOPE, vol. 57, Acknowledgments.This work was supportedby the National Sci- editedby V. Ittekkot et al., chap. 12, pp. 223-241, JohnWiley, New ence Foundation (NSF grants OCE-9216626, OCE-9729284, OCE- York, 1996. 9401537,OCE-9729697, OCE-9415790, and OCE-9711318),the Na- Falkowski,P., and D. A. Kiefer, Chlorophyllfluorescence in phyto- tional Aeronauticsand SpaceAdministration (NASA grantsNAG5- plankton:Relationship to photosynthesisand biomass,J. Plankton 6448 and NAS5-97128),and the ConsejoNacional de Investigaciones Res., 7, 715-731, 1985. Cientificasy Tecnologicas(CONICIT, Venezuela,grant 96280221). Falkowski, P. G., R. T. Barber, and V. Smetacek,Biogeochemical We are indebtedto the personnelof the FundacionLa Sallede Cien- controlsand feedbackson oceanprimary production, Science, 281, cias Naturales, Estacion de InvestigacionesMarinas Isla Margarita 200-206, 1998. (FLASA/EDIMAR) for its enthusiasmand professionalsupport. In Febres-Ortega,G., and L. E. Herrera, Interpretaciondinamica de particular, we thank Pablo Mandazen (Hermano Gines, Director, algunasde las caracteristicasoceanograficas de la Fosa de Cariaco, FLASA) for his confidencein our activitiesand the crew of the R/V Mar Caribe, Bol. Inst. Oceanogr.Univ. Oriente,14, 3-29, 1975. HermanoGines (FLASA) for its able supportat sea.Jonnathan Gar- Field, C. B., M. J. Behrenfeld, J. T. Randerson, and P. Falkowski, cia, Javier Gutierrez, AnadiuskaRondon (all at FLASA/EDIMAR), Primary production of the biosphere:Integrating terrestrial and JohnAkl, and Ana Lucia Odriozola(at USF) providedessential field oceaniccomponents, Science, 281,237-240, 1998. and laboratorysupport. Luis Troccoli,Wilfredo Patifio,Luis Sanchez, Froelich, P. N., D. K. Atwood, and G. S. Giese, The influence of and William Senior, from the Universidad de Oriente, Cumana, Ven- Amazon River water on surfacesalinity and dissolvedsilicate con- ezuela,provided nutrient data. Maria Iabichela,also from the Univer- centration in the Caribbean Sea, Deep Sea Res.,25, 735-744, 1978. MULLER-KARGERET AL.: CARIACOBASIN PRIMARY PRODUCTION CYCLE 4541

Hansell, D. A., and C. A. Carlson, Net community production of Morris, I., A. E. Smith, and H. E. Glover, Productsof photosynthesis dissolvedorganic carbon, Global Biogeochem.Cycles, 12, 443-453, in phytoplanktonoff the Orinoco River and in the CaribbeanSea, 1998. Limnol. Oceanogr.,26, 1034-1044, 1981. Haug, G., T. Pedersen,D. Sigman, S. Calvert, B. Nielsen, and L. Morrison, J. M., and W. D. Nowlin Jr., General distributions of water Peterson,Glacial/interglacial: Variations in productionand nitrogen masseswithin the easternCaribbean Sea duringwinter of 1972 and fixation in the Cariaco Basin during the last 580 kyr, Paleoceanog- fall of 1983, J. Geophys.Res., 87, 4207-4229, 1982. raphy,13, 427-432, 1998. Morrison, J. M., and O. P. Smith, Geostrophictransport variability Henrichs,S. M., Earlydiagenesis of organicmatter: The dynamics alongthe aves ridge in theeastern Caribbean Sea during 1985-1986, (rates) of cyclingof organiccompounds, in OrganicGeochemistry: J. Geophys.Res., 95, 699-710, 1990. Principlesand Applications,edited by M. H. Engel and S. A. Macko, Muller-Karger, F. E., and R. Aparicio, Mesoscaleprocesses affecting pp. 101-114,Plenum, New York, i993. phytoplankton abundancein the southern Caribbean Sea, Cont. Herrera, L. E., and G. Febres-Ortega, Procesosde surgenciay de Shelf Res., 14, 199-221, 1994. renovacionde Aguas en la Fosa de Cariaco, Mar Caribe, Bol. Inst. Muller-Karger, F. E., C. R. McClain, T. R. Fisher, W. E. Esaias, and Oceanogr.Univ. Oriente,14, 31-44, 1975. R. Varela, Pigmentdistribution in the CaribbeanSea: Observations Holmen, K. J., and C. G. Rooth, Ventilation of the Cariaco Trench, a from Space,Prog. Oceanogr.,23, 23-69, 1989. case of multiple sourcecompetition?, Deep Sea Res., Part A, 37, Muller-Karger, F. E., P. L. Richardson,and D. McGillicuddy, On the 203-225, 1990. offshore dispersalof the Amazon's plume in the North Atlantic, Holm-Hansen, O., C. J. Lorenzen, R. W. Holmes, and J. D. H. Strick- Deep Sea Res.,Part I, 42, 2127-2137, 1995. land, Fluorometricdetermination of chlorophyll,J. Cons.Int. Explor. Murphy, S. J., H. E. Hurlburt, and J. J. O'Brien, The connectivityof Mer, 30, 3-15, 1965. eddy variability in the Caribbean Sea, the Gulf of Mexico, and Honjo, S., and K. Doherty, Large aperture time seriesoceanic sedi- Atlantic Ocean, J. Geophys.Res., 104, 1431-1453, 1999. ment traps:Design objectives,construction and application,Deep Olivieri, R. A., and F. P. Chavez,A model of plankton dynamicsfor Sea Res., Part A, 35, 133-149, 1988. the coastalupwelling system of Monterey Bay, California,Deep Sea Honjo, S., D. W. Spencer, and J. W. Farrington, Deep advective Res., Part H, 47, 1077-1106, 2000. transportof lithogenicparticles in PanamaBasin, Science, 216, 516- Pace, M., G. Knauer, D. Karl, and J. Martin, Primary production,new 518, 1982. production and vertical flux in the eastern Pacific Ocean, Nature, Hughen,K., J. Overpeck,L. Peterson,and S. Trumbore, Rapid climate 325, 803-804, 1987. changesin the tropical Atlantic region during the last glaciation, Peterson,L. C., J. T. Overpeck,N. G. Kipp, and J. Imbrie, A high- Nature, 380, 51-54, 1996. resolutionlate Quaternary upwellingrecord from the anoxicCari- Hughen, K. A., J. T. Overpeck,S. J. Lehman,M. Kashgarian,J. Sou- aCOBasin, Venezuela, Paleoceanography,6, 99-119, 1991. thon, L. C. Peterson,R. Alley, and D. M. Sigman,Deglacial changes Pilskaln, C. H., J. B. Paduan, F. P. Chavez, R. Y. Anderson, and W. M. in ocean circulation from an extended radiocarbon calibration, Na- Berelson.Carbon export and regenerationin the coastalupwelling •ture,391, 65-68, 1998. systemof Monterey Bay, central California,J. Mar. Res.,54, 1149- Jahnke, R. A., The global ocean flux of particulate organic carbon: 1178, 1996. Areal distributionand magnitude,Global Biogeochem.Cycles, 10, Redfield, A. C., B. H. Ketchum, and F. A. Richards, The influence of 71-88, 1996. organismsin the compositionof sea water, in The Sea, edited by Karl, D., J. Christian,J. Dore, D. Hebel, R. Letelier, L. Tupas, and C. M. N. Hill, chap. 2, pp. 26-77, Interscience,New York, 1963. Winn, Seasonaland interannualvariability in primary production Richards, F. A., Some chemical and hydrographicobservations along and particle flux at Station ALOHA, Deep Sea Res., Part H, 43, the north coast of South America, I, Cabo Tres Puntas to Curacao, 539-568, 1996. including the Cariaco Trench and the Gulf of Cariaco, Deep Sea Kinder, T. H., G. W. Heburn, and A. W. Green, Some aspectsof Res., 7, 163-182, 1960. Caribbean circulation, Mar. Geol., 68, 25-52, 1985. Richards,F. A., The Cariaco Basin (Trench), Oceanogr.Mar. Biol. Kirchman, D. L., Leucine incorporation as a measure of biomass Annu. Rev., 13, 11-67, 1975. productionby heterotrophicbacteria, in Handbook of Methodsin Richards, F. A., and R. F. Vaccaro, The Cariaco Trench, an anaerobic AquaticMicrobial Ecology, edited by P. F. Kemp et al., pp. 509-512, basin in the Caribbean Sea, Deep Sea Res., 3, 214-228, 1956. Lewis, Boca Raton, Fla., 1993. Robert-Baldo,G., M. J. Morris, and R. H. Byrne,Spectrophotometric Kishino, M., M. Takahashi, N. Okami, and S. Ichimura, Estimation of determinationof seawaterpH usingphenol red, Anal. Chem., 57, the spectralabsorption coefficients of phytoplanktonin the sea,Bull. 2564-2567, 1985. Mar. Sci., 37, 634-642, 1985. Scranton,M. I., Temporal variationsin the methane content of the Koblentz-Mishke, O. J., V. V. Volkovinsky, and J. G. Kabanova, Cariaco Trench, Deep Sea Res., 35, 1511-1523, 1988. Planktonprimary production of the world ocean,in ScientificExplo- Scranton,M. I., F. L. Sayles,M.P. Bacon,and P. G. Brewer,Temporal ration of theSouthern Pacific, edited by W. S. Wooster,pp. 183-193, changesin the hydrographyand chemistryof the Cariaco Trench, Natl. Acad. of Sci., Washington,D.C., 1970. Deep Sea Res.,Part A, 34, 945-963, 1987. Lee, C., Controlson organiccarbon preservation: The use of stratified Scranton, M. I., Y. Astor, R. Bohrer, T.-Y. Ho, and F. E. MUller - water bodiesto compareintrinsic rates of decompositionin oxicand Karger,The effect of recentsubsurface water massintrusions on the anoxicsystems, Geochim. Cosmochim. Acta, 56, 3323-3335, 1992. geochemistryof the Cariaco Basin,Deep Sea Res., in press,2001. Lidz, L., W. B. Charm, M. M. Ball, and S. Valdes, Marine basinsoff the Spencer,D. W., Aluminumconcentrations and fluxesin the ocean,in coast of Venezuela, Bull. Mar. Sci., 19, 1-17, 1969. Global OceanFlux Study:Proceedings of a Workshop,pp. 206-220, Margalef, R., Composiciony distribuciondel fitoplancton,Memoria, Natl. Acad., Washington,D.C., 1984. XXV, 141-205, 1965. SteemanNielsen, E., The useof radioactivecarbon (14C) for measur- Margalef, R., F. Cervigon,and G. T. Yepez, Exploracionpreliminar de ing organicproduction in the sea, J. Cons.Int. Explor. Mer, 18, las caracteristicashidrograficas y de la distribuciondel fitoplancton 117-140, 1952. en el areade la Isla Margarita(Venezuela),Memorias, XX, 211-221, Strong, A. E., and E. P. McClain, Improved ocean surfacetempera- 1960. tures from space:Comparisons with drifting buoys,Bull. Am. Me- McClain, E. P., W. G. Pichel, C. C. Walton, Z. Ahmad, and J. Sutton, teorol. Soc., 65, 138-142, 1984. Multi-channelimprovements to satellite-derivedglobal sea-surface Thunell, R., Seasonaland annual variability in particle fluxesin the temperatures,Adv. SpaceRes., 2(6), 43-47, 1983. Gulf of California:A responseto climate forcing,Deep SeaRes., 45, Merhbach, C., C. H. Culberson,J. E. Hawley, and R. M. Pytkowicz, 2059-2083, 1998a. Measurement of the apparent dissociationconstants of carbonic Thunell, R., Particlefluxes in a coastalupwelling zone: Sedimenttrap acid in seawaterat atmosphericpressure, Limnol. Oceanogr.,18, results from Santa Barbara Basin, California, Deep Sea Res., 45, 897-907, 1973. 1863-1884, 1998b. Millero, F. J., Thermodynamicsof the carbon dioxide systemin the Thunell, R., E. Tappa, R. Varela, M. Llano, Y. Astor, F. Muller- oceans, Geochim. Cosmochim.Acta, 59, 661-677, 1995. Karger, and R. Bohrer, Increasedmarine sedimentsuspension and Minnett, P. J., Consequencesof seasurface temperature variability on fluxesfollowing an earthquake,Nature, 398, 233-236, 1999. the validation and applicationsof satellite measurements,J. Geo- Thunell, R., R. Varela, M. Llano, J. Collister, F. Muller-Karger, and R. phys.Res., 96, 18,475-18,489, 1991. Bohrer, Organic carbon flux in an anoxicwater column: Sediment 4542 MULLER-KARGER ET AL.: CARIACO BASIN PRIMARY PRODUCTION CYCLE

trap resultsfrom the Cariaco Basin,Limnol. Oceanogr.,45, 300-308, surfacetemperature with AVHRR satellitedata, J. Appl. Meteorol., 2000. 27, 115-124, 1988. United Nations Educational, Scientific, and Cultural Organization Wright, R. T., and J. E. Hobbie, Use of glucoseand acetateby bacteria (UNESCO), Protocolsfor the Joint Global Ocean Flux Study and algae in aquaticecosystems, Ecology, 47, 447-463, 1966. (JGOFS) core measurements,in IOC Manualsand Guides,vol. 29, pp. 128-140, Paris, 1994. Y. Astor, J. Capelo, and R. Varela, FundacionLa Salle de Ciencias Volk, T., and Z. Liu, Controls of CO2 sourcesand sinksin the earth Naturales,Estacion de InvestigacionesMarinas de Margarita, Apar- scalesurface ocean: Temperature and nutrients,Global Biogeochem. tado 144 Porlamar,Isla de Margarita, Venezuela. Cycles,2, 73-89, 1988. R. Bohrer, F. Muller-Karger, and J. J. Walsh, College of Marine Walsh, J. J., On the Nature of ContinentalShelves, 520 pp., Academic, Science, University of South Florida, St. Petersburg, FL 33701. San Diego, Calif., 1988. ([email protected]) Walsh, J. J., Importanceof continentalmargins in the marine biogeo- T.-Y. Ho, M. Scranton,and G. Taylor, Marine SciencesResearch chemicalcycling of carbonand nitrogen,Nature, 350, 53-55, 1991. Center, SUNY, StonyBrook, NY 11794. Walsh, J. J., K. L. Carder, and F. M. Muller-Karger, Meridional fluxes E. Tappa and R. Thunell, Department of GeologicalSciences, Uni- of dissolvedorganic matter in the North Atlantic Ocean,J. Geophys. versityof South Carolina, Columbia,SC 29208. Res., 97, 15,625-15,637, 1992. Walsh,J. J., et al., Simulationof carbon/nitrogencycling during spring upwelling in the Cariaco Basin, J. Geophys.Res., 104, 7807-7825, 1999. (ReceivedMay 7, 1999;revised July 19, 2000; Walton, C. C., Nonlinear multichannelalgorithms for estimatingsea acceptedJuly 27, 2000.)