Importance of solar subsurface heating in ocean ... - Wiley Online Library

15 downloads 222 Views 2MB Size Report
Dec 15, 2001 - examined when solar heating, as given by monthly mean kvA• and PAR fields, ... The solar heating of the upper ocean occurs through the.
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. C12, PAGES 30923-30938,

DECEMBER

15, 2001

Importance of solar subsurface heating in ocean general circulation

models

Peter A. Rochford Naval ResearchLaboratory,StennisSpaceCenter, Mississippi,USA

A. Birol Kara• SverdrupTechnologyInc., StennisSpaceCenter,Mississippi,USA

Alan J. Wallcraft

and Robert A. Arnone

Naval ResearchLaboratory,StennisSpaceCenter, Mississippi,USA

Abstract. The importanceof subsurfaceheatingon surfacemixed layer propertiesin an ocean generalcirculationmodel (OGCM) is examinedusing attenuationof solarirradiancewith depth belowthe oceansurface.The depth-dependent attenuationof subsurface heatingis givenby global monthly mean fields for the attenuationof photosyntheticallyavailableradiation(PAR), kpAR. Theseglobal fields of kpARare derivedfrom Sea-viewingWide Field-of-view Sensor(SeaWiFS) data on the spectraldiffuse attenuationcoefficientat 490 nm (k490),and have been processedto have the smoothlyvarying and continuouscoveragenecessaryfor use in OGCM applications. Thesemonthlyfieldsprovidethe first completeglobaldatasetsof subsurfaceopticalfieldsthat can be usedfor OGCM applicationsof subsurfaceheatingand bio-opticalprocesses.The effect on globalOGCM predictionof seasurfacetemperature(SST) andsurfacemixed layerdepth(MLD) is examinedwhen solarheating,as given by monthlymean kvA• and PAR fields, is includedin the model. It is fbund that subsurfaceheatingyields a marked increasein the SST predictiveskill of the OGCM at low latitudes.No significantimprovementin MLD predictiveskill is obtainedwhen includingsubsurfaceheating.Use of the monthlymean lcp^Rproducesan SST decreaseof up to 0.8øCand a MLD increaseof up to only 4-5 m for climatologicalsurfaceforcing,with this

primarily confined to theequatorial regions. Remarkably, a constant kv^R valueof 0.06m-• , which is indicativeof opticallyclearopenoceanconditions,is foundto servevery well for OGCM predictionof SST and MLD over most of the global ocean. 1.

Introduction

1985]. This is becausethe solar radiationcan penetrateto depths below the surfacemixed layer under suchcircumstances, thereby requiringthat proper accountbe taken of the solarheatingwithin the mixed layer. Studieshave indicatedthe necessityof including penetrating solar irradiancefor SST predictionsin the tropical Pacific [Zaneveld et al., 1981; Woo& et al., 1984, Lewis et al., 1990]. While this hasbeenclearly shownwith one-dimensionalmodels,the majority of oceangeneralcirculationmodels(OGCM) with mixed layersdo not considerspace-and time-varyingsolar attenuationwith depth [Baturin and Niiler, 1997; Carton and Zhou, 1997; Murtuguddeet al., 1995; Schopfand Loughe, 1995; Sterl and Kattenberg,1994; Yuen et al., 1992]. The reason is there has been insufficient information on the regional and seasonal distribution of the irradianceabsorptioncoefficientover basinscalesto use it as part of the heatflux forcingin an OGCM. As a consequence, the extent to which subsurfaceheatingis..•important for global SST prediction

The solar heating of the upper ocean occurs through the absorptionof solar irradiance(400-800 nm) with depth beneath the oceansurface,with this rate of absorptionvarying in spaceand time. The distributionof this solar heatingwith depth within the surfacemixed layer directly affectsthe sea surfacetemperature (SST) [e.g., Kantha and Clayson,1994]. It also indirectlyaffects the mixed layer depth (MLD) by altering the vertical turbulent mixing in the upperwater column[e.g., Lewis et al., 1990]. The positivebuoyancyforcinggeneratedby this subsurfaceheatingis primarily responsible for stabilizing the surface waters [e.g., Martin, 1985], which in turn contributesto the formation of the upperoceanthermocline. Knowledgeof the seasonaland maximum annual depth of the thermoclineis of importancefor heat storagein the ocean[White et al., 1998; Gallimoreand Houghton,1987;Meehl, 1984; Stevenson and Niiler, 1983]. SST predictionin one-dimensional mixed layer has been difficult to determine. modelshas been shownto be sensitiveto the parameterizationof Recently available ocean color remote sensing data (1997the solar extinctionwith depth in the upper water column when 1998) being acquiredwith the Sea-viewingWide Field-of-view thereis a shallowmixed layer [Kanthaand Clayson,1994;Martin, Sensor(SeaWiFS) cannow be usedto estimatesurfaceattenuation rates on global scales.These data collectedfrom the commencement of SeaWiFS operationin September1997 provide, for the •Nowat Centerfor Ocean-Atmospheric Prediction Studies, Florida first time, the global spatial variation of the rate of absorption StateUniversity,Tallahassee,Florida. USA. which can be used to quantify the solar influence in OGCM applications.SeaWiFS ocean color data provide information on Copyright2001 by the AmericanGeophysicalUnion. water leaving radiancesat selectedwavelengths[McClain et al., 1998], from which the wavelength dependenceof the diffuse Papernumber2000JC000355. 0148-0227/01/2000JC000355509.00 attenuationcoefficientcan be constructedfor the photosyntheti30923

30924

ROCHFORD

ET AL.:

SUBSURFACE

cally availableradiation(PAR) portionof the solarspectrum.PAR is definedas the portionof the spectrumthat penetrates beyonda few centimeters of the surfaceandrepresents the light availablefor photosynthesis [e.g., Mobley, 1994]. The diffuseattenuationcoefficient at 490 nm (k49o)is in particularabundancebecausethe strongpigmentabsorptionat 490 nm in phytoplankton is of interest in remotesensing[Gordonand Clark, 1980]. Given that 490 nm lies within the 420-580 nm wavelengthrangehavingthe deepest penetrationinto the ocean[Austinand Petzold,1986;Morel, 1988], knowledgeof k49oallowsthe attenuationof PAR (kpAR)to be easily determined[Austinand PetzoM, 1986]. In the past, suchremote sensing data, which were available from sensors,such as the CoastalZone Color Scanner[e.g., Arnone et al., 1992], were too limited in arealcoverageto be usedaspart of the heatflux forcing in a basin-scale OGCM. However,with SeaWiFSnow in operation sinceSeptember1997, sufficientoceancolor data at high-resolution global coveragehave been acquiredto constructa complete annual cycle of monthly means that can be used for OGCM forcing. In additionto SST, knowledgeof the attenuationof PAR with depth, as given by the kpARmonthly fields, can be used to characterizethe light environmentfor models of ocean biology [Dadou et al., 1996; Flierl and Davis, 1993; Keen et al., 1997; McCreary et al., 1996]. The variation of solar irradiancewith depth defines the euphotic zone of the upper ocean, where photosynthesiscan occur, and has a direct influence upon growth rates for phytoplankton.The vertical PAR distribution directly controlsthe rate of photosynthesis within the phytoplankton subcommunityof the ocean ecosystem[e.g., Platt et al., 1988]. As phytoplanktonare an importantfirst link in the food chain, the variation of PAR with depth has a strongand direct impact on primary productionand thereforethe world's fisheries. The availability of these kpAR fields provides the opportunityto implement satellite-basedoptics in biophysical coupled OGCMs of primary productivity [Keen et al., 1997; McCreary et al., 1996]. In this studywe reportfor the first time on the construction of k49oand kpARdata setsof monthly meansfrom recentlyavailable 1997-1998

SeaWiFS

ocean

color

data

and comment

on the

HEATING

IN AN

OGCM

the quantityof interestfor subsurfacesolarheatingin an OGCM with an embeddedmixed layer. The attenuationof surfacesolarirradiance(QsoL)with depthz over the solar spectrumis typically representedas a simple exponentialdecay,

QsoL(X,z)= Qsou(X,0) exp(-kxz), where kx is definedas the spectraldiffuseattenuationcoefficient at wavelengthX. The kx valuesvary stronglyas a functionof the optical characteristicsof the water, such as the amount of suspended particles.There are methodsfor specifyingkx given the known attenuationat a single wavelengthXo [Austin and PetzoM, 1986;Morel, 1988] or via setsof parametersfor different Jefiov water types [ZaneveM and Spintad, 1980]. Typically, subsurface heatingis calculatedby dividing the solar spectrum into a number of narrow wavelengthbands [e.g., Paulson and Simpson,1977; Morel and Antoine, 1994], with kx calculatedat the centerwavelengthfor each of the bands.However,it is also possibleto derive a simpleexponentialdecayfor the attenuation of PAR with depth, which is given in terms of a single wavelengthXo [Austinand Petzold,1986]. This allowsa simpler expressionto be used

PAR(z): PAR(O)exp(-kp•a•z),

(2)

where PAR(O) is the PAR at the oceansurface.For OGCM with bulk mixedlayers,wherethe upperoceanis typicallyrepresented by a vertically uniform layer 0.1 indicate theincrease in SSTpredictive skillof theNLOMwhenincluding subsurface heatingusingthe monthlykpA R.

30934

ROCHFORD

ET AL.'

SUBSURFACE

,o• / •

1

\

/

-- Arabian Sea (15øN, 60øE) is included. Values of ASS > 0.1 indicate an increase in IndoPacific (5øN, 145øE) - - South America (5øS, 84øW) predictive skill. The SS difference (Plate 6) reveals that the

of subsurfaceheating improvesNLOM SST predictive -•-South Africa (5øS, 5øE• inclusion skill in the equatorialregionsand markedly increasesthis skill



in the Arabian Sea, the Bay of Bengal, and the shallower IndonesianPacific. The regions of ASS < 0 are confined to isolated areas of smaller extent and are smaller in magnitude

0.5

^

/ ', -

\

o

,t

than the areas of ASS > 0. While the annual mean difference in

'*-i

i

i

i

1N AN OGCM

scores (monthly kpAR minus large kpAR SS) indicate an improvementin NLOM SST predictionwhen subsurfaceheating

1.5

/

HEAT1NG

i

i

i

i

i

i

i

0'5jan FebMarAprMayJunJulAugSepOctNovDec

SST may be small, the positive increasesin SS of 0.3-0.5 clearly indicate that NLOM SST prediction is significantly improvedby including subsurfaceheating. The improvementin NLOM SST predictionthat resultsfrom includingsubsurfaceheatingis observedto be mostlyconfinedto regionswhere the NLOM MLD is _2

(B6)

w•--- 40k [max( 0,hk- h•-)/h•-] 2 k> 2

(B7)

/kPk-- max(Pk+l - Pk-- /kp•-,0)

(B8)

Notethat•k+ -w• andf•k denote areaaverages of therespective quantitiesover the oceanmodel domain.

• Bjk •Vki --Gi,

(A2)

k=l

where B.2.

By•-- {Afa(ry)Afa(r•))

(^3)

qi- (AfB(ry)AfB(ri))

(A4)

We list herethe symbolsappearingin the NLOM equationsthat are not definedelsewherein the manuscript. A,coefficient of horizontal eddy viscosity (1500

are the covariancesin the field value errors Afs. The n x n symmetric matrix B is the error covariancematrix among the backgroundfield valuesat the noninterpolated locations(j, k: 1, ..., n), and the n vector C is the correspondingerror covariance vector of the backgroundfield valuesat noninterpolated locations with respectto the interpolatedlocation [Daley, 1991, equation (4.2.10)]. For lack of informationon the root-mean-square errors of the mean k490, the covariances are assigned correlation functions

of

Byk -- exp[-lry- r•./ro]

(A5)

C•,= exp[-rj - ril/ro],

(A6)

with a distancescaleof r0: 1ø. The value of r0 was chosento be consistentwith the grid resolution. The data filling of the SeaWiFS k4o0using SI proceedsby searchingfor data voids along latitudetransectsstartingfrom the equatorand movingpolewardin alternatingnorth-southlatitudes. Searchesalonglongitudeproceedfrom westto eastand thenfrom eastto west.An interpolatinggrid box of 5 x 5 centeredat the data void is used for the SI in all but the last 2 latitude rows at the

polewardboundaries.A 3 x 3 box is usedfor the secondlastrows, and a 3 x

2 box

is used for the boundaries.

Data

voids

Notation

at

noncentrallocationsin the interpolatingbox are excludedwhen performingthe SI.

m2s - 1). Co coefficientof bottom friction (0.002). C:u coefficient of additional interfacial friction on entrainment.

Cpa specific heatcapacity forair(1004J kg-1 øC-i). Cpwspecific heatcapacity forwater (3988Jkg-• øC-i). D(x, y) total depthof the oceanat rest. e• saturatedvapor pressure. f Coriolisparameter.

f, Coriolis parameter at 5ølatitude (2.54x 10-s s-i). g gravitational acceleration (9.81m s-2). hm + minimum mixedlayerdepth(10 m). hJ kthlayerthickness at whichentrainment starts. h• kth layer thicknessat which detrainmentstarts. Ho constantreferencelayer thickness(100 m). H• kth layer thicknessat rest. K,coefficient of horizontaltemperaturediffusivity

(0 m2 s-l). Lv the latentheat of vaporization. m• TKE constants(m• = 6.25, m3 = 7.5, ms = 1.8, m6 = 0.1). n• TKE constant(n• = 1). P• surfaceair pressure(1013 mbar). q• polynomial coefficientsfor qsatas given by Lowe [1977]. qsat saturatedmixing ratio at sea surface.

QN•(x,y) meansurface heatflux climatology. Rgas themolecular gasconstant. Tt, temperature just belowthe mixedlayer(øC). Appendix B: NLOM Notation Td dewpointtemperature (øC). B.1. Definitions ApJ minimumdensity contrast for (p•+ AT• temperature shear(øC). We list here the mathematicaldefinitionsof variousquantities

AT[ minimum temperature shear(0.2øC). c•(T) coefficient ofthermal expansion ofseawater(øC-i).

appearingin the NLOM equations.

Gk• --{g

I_> k

(B1)

ATm temperature changeacrossthe mixedlayer(øC).

f• rotation rateof Earth(7.292x 10-s s-•).

ROCHFORD

ET AL.'

SUBSURFACE

4) longitude(degrees).

p"• model density climatology forlayerk (kgm-3). po reference density in the ocean (1000kgm-3). -1

c% --1

%

e-foldingtime for MLD relaxation(1 day).

densityclimatology relaxation e-foldingtimefor hk=

Ho. q-a surfacewind stress.

•e(x,y) meanclimatological evaporation rate. &k kth interfacereferenceverticalmixing velocity. Q•(x, y) kth interfaceglobalmixing correctionscalefactor.

Acknowledgments. The authorswould like to thank Daniel Fox of the Naval ResearchLaboratoryfor providingthe statisticalinterpolation FORTRAN code that was used in constructingthe kpA}•data sets.Ocean color data used in this study were producedby the SeaWiFS Project at GoddardSpaceFlight Center.The data were obtainedfrom the Goddard DistributedActive Archive Center under the auspicesof the National Aeronauticsand SpaceAdministration.Use of this data is in accordwith the SeaWiFSResearchData Use Termsand ConditionsAgreement.This work was funded by the Office of Naval Research (ONR) and is a contributionof the Basin-ScalePredictionSystemprojectunder program element 602435N and the Dynamics of Coupled Air-Ocean Models Study under program element 61153N. Sverdrup TechnologyInc. is funded under subcontractfrom the Naval ResearchLaboratory. This is contributionNRL/JA/7323-99-0028 and has been approvedfor public release.

References Anderson,S. P., R. A. Weller, and R. B. Lukas, Surfacebuoyancyforcing and the mixed layer of the westernPacificwarm pool: Observations and 1D model results,•/. Clim., 9, 3056-3085, 1996. Amone, R. A., G. Terrie,L. Estep,and R. A. Oriol, OceanOptical Database,NOARL Tech.Note 254, 36 pp., Naval Res. Lab., StennisSpace Center, Miss., 1992. Austin, R. W., and T. J. Petzold, Spectraldependenceof the diffuse attenuationcoefficientof light in oceanwaters,Opt. Eng., 25, 471-479, 1986.

Baturin, N. G., and P. P. Niiler, Effects of instabilitywaves in the mixed layer of the equatorialPacific,•/. Geophys.Res., 102, 27,771-27,793,

HEATING

IN AN

OGCM

30937

Gallimore,R. G., and D. D. Houghton,Approximationof oceanheat storage by ocean-atmosphere energy exchange:Implicationsfor seasonal cycle mixed layer ocean formulations,d. Phys. Oceanogr.,17, 12141231, 1987.

Gordon,H. R., and D. K. Clark, Atmosphericeffectsin the remotesensing of phytoplanktonpigments,BoundaryLayer Meteorol., 18, 300- 314, 1980.

Hellerman, S., and M. Rosenstein,Normal monthly wind stressover the world oceanwith error estimates,d. Phys. Oceanogr.,13, 1093-1104, 1983.

Hurlburt,H. E., andE. J. Metzger,Bifurcationof the KuroshioExtensionat the ShatskyRise, d. Geophys.Res., 103, 7549-7566, 1998. Hurlburt, H. E., A. J. Wallcraft, W. J. Schmitz, P. J. Hogan, and E. J. Metzger, Dynamics of the Kuroshio/Oyashiocurrent system using eddy-resolvingmodelsof the North Pacific Ocean,d. Geophys.Res., 101, 941-976,

1996.

Jerlov,N. G., Marine Optics,ElsevierOceanogr.Set.,vol. 14, ElsevierSci., New York, 1976.

Kantha, H. K., and C. A. Clayson,An improvedmixed layer model for geophysicalapplications,d. Geophys.Res., 99, 25,235- 25,266, 1994. Kara, A. B., P. A. Rochford,and H. E. Hurlburt, Efficient and accuratebulk parameterizations of air-seafluxesfor usein generalcirculationmodels, d. Atmos. Ocean. Technol., 17, 1421-1438, 2000a.

Kara, A. B., P. A. Rochford,and H. E. Hurlburt,An optimaldefinitionfor ocean mixed layer depth, d. Geophys. Res., 105, 16,803-16,821, 2000b.

Keen, T. R., J. C. Kindle, and D. K. Young, The interactionof southwest monsoonupwelling, advectionand primary productionin the northwest Arabian Sea,d. Mar. Syst.,13, 61-82, 1997. Kraus, E. B., and J. S. Turner, A one-dimensionalmodel of the seasonal thermocline,II, The generaltheoryandits consequences, Tellus,19, 98106, 1967.

Large,W. G., J. C. McWilliams, and S.C. Doney, Oceanicverticalmixing: A review and a model with a nonlocalboundarylayer parameterization, Rev. Geophys.,32, 363-403, 1994. Levitus,S., and T. P. Boyer, WormOceanAtlas 1994, vol. 4, Temperature, NOAA Atlas NESDIS 4, 117 pp., Natl. Oceanic and Atmos. Admin., Silver Spring,Md., 1994. Levitus, S., R. Burgett,and T. P. Boyer, Worm OceanAtlas 1994, vol. 3, Salinity,NOAA Atlas NESDIS 3, 99 pp., Natl. Oceanicand Atmos. Admin., Silver Spring,Md., 1994. Lewis, M. R., M. E. Carr, G. Feldman, C. R. McClain, and W. Esaias, Influence of penetratingradiation on the heat budget of the equatorial Pacific Ocean, Nature, 34 7, 543- 545, 1990.

Lorenc,A. C., A globalthree-dimensional multivariatestatisticalinterpolation scheme,Mon. WeatherRev., 701 - 721, 1981. Bishop,J. K. B., and W. B. Rossow,Spatialand temporalvariability of Lowe, P. R., An approximatingpolynomialfor the computationof saturaglobal surfacesolar irradiance,d. Geophys.Res., 96, 16,839-16,858, 1991. tion vaporpressure,d. Appl. Meteorol., 16, 100-103, 1977. Brock, J. C., C. R. McClain, M. E. Luther, and W. W. Hay, The phyto- Martin, P., Simulationof the mixed layerat OWS NovemberandPapawith severalmodels,d. Geophys.Res., 90, 903-916, 1985. planktonbloom in the northwesternArabian Sea during the southwest McCreary, J.P., K. E. Kohler, R. R. Hood, and D. B. Olsen, A fourmonsoonof 1979, d. Geophys.Res., 96, 20,623-20,642, 1991. componentecosystemmodel of biological activity in the Arabian Sea, Carton,J. A., and Z. Zhou, Annual cycle of seasurfacetemperaturein the Prog. OceanoAr.,37, 193-240, 1996. tropicalAtlanticOcean,d. Geophys.Res.,102, 27,813-27,824, 1997. Dadou,I., V. Garqon, V. Anderson,G. R. Flierl, and C. S. Davis, Impact of McClain, C. R., M. L. Cleave, G. C. Feldman,W. W. Gregg,S. B. Hooker, and N. Kuring, Sciencequality SeaWiFS data for global biosphererethe northequatorialcurrentmeanderingon a pelagicecosystem: A modsearch,Sea Technol., 39, 10-16, 1998. eling approach,d. Mar. Res., 54, 311-342, 1996. Daley, R., AtmosphericData Analysis,457 pp., CambridgeUniv. Press, McClain, C. R., E. J. Ainsworth,R. A. Barnes,R. E. Eplee Jr., F. S. Patt, New York, 1991. W. D. Robinson,M. Wang, and S. W. Bailey, SeaWiFSpostlaunchcaliDarzi, M., SeaWiFSsciencealgorithmflow chart,GSFC/CR-1998-206848, brationandvalidationanalyses,part 1, NASATech.Memo.,2000-206892, vol. 9, 82 pp., 2000a. 30 pp., NASA GoddardSpaceFlight Center,Greenbelt,Md., 1998. McClain, C. R., et al., SeaWiFSpostlaunchcalibrationand validationanaDa Silva,A.M., C. C. Young,and S. Levitus,Atlas of SurfaceMarine Data lyses,part 2, NASA Tech.Memo., 2000-206892, vol. 10, 57 pp., 2000b. 1994, vol. 1, AlgorithmsandProcedures,NOAA AtlasNESDIS 6, 83 pp., Meehl, G. A., A calculation of ocean heat storageand effective ocean Natl. Oceanicand Atmos. Admin., Silver Spring,Md., 1994. surfacelayer depthsfor the Northern Hemisphere,d. Phys. OceanoAr., EuropeanCentre for Medium-RangeWeatherForecasts(ECMWF), The 14, 1747-1761, 1984. descriptionof the ECMWF/WCRP Level III, a global atmosphericdata archiveEuropeanCentrefor MediumRangeWeatherForecasts technical Metzger,E. J., and H. E. Hurlburt, Coupleddynamicsof the SouthChina attachment,Reading,England, 1993. Sea,the SuluSea,andthePacificOcean,d. Geophys.Res., 101, 12,33112,352, 1996. EuropeanCentre for Medium-RangeWeatherForecasts(ECMWF), Asnatureof Kuroshio pectsof the re-analysedclimate,ECMWF Re-AnalysisProj. Rep. Ser. Metzger,E. J., andH. E. Hurlburt,The nondeterministic 2, 89 pp., Reading,England, 1997. Penetrationand eddy sheddingin the SouthChina Sea,d. Phys. Oceanoar., 31, 1712-1732, 2001. Firestone,E. R., and S. B. Hooker, SeaWiFSprelaunchtechnicalreport series final cumulative index, NASA Tech. Memo., 1998-104566, 43, 69 Metzger,E. J., H. E. Hurlburt,J. C. Kindle, Z. Sirkes,and J. M. Pringle, Hindcastingof wind-drivenanomaliesusing a reduced-gravityglobal pp., 1998. oceanmodel, Mar. Technol.Soc. d., 26, 23-32, 1992. Flierl, G. R., and C. S. Davis, Biological effectsof Gulf StreammeanMobley,C. D.,Light and Water,592 pp., Academic,SanDiego, Calif., 1994. dering,d. Mar. Res., 51, 529-560, 1993. Fu, G., K. S. Baith, and C. R. McClain, SeaDAS: The SeaWiFS Data AnaMorel, A., Opticalmodelingof the upperoceanin relationto its biogenous lysis System,paperpresentedat the 4th Pacific OceanRemote Sensing matter content(case I waters),d. Geophys.Res., 93, 10,749-10,768, 1988. Conference,NASA GoddardSpaceFlight Center,Qingdao,China, 1998. Gallacher,P. C., and P. A. Rochford,Numerical simulationsof the Arabian Morel, A., and D. Antoine, Heating rate within the upperoceanin relation Seausingtracersasproxiesfor phytoplanktonbiomass,d. Geophys.Res., to its bio-opticalstate,d. Phys. OceanoAr.,24, 1652-1665, 1994. 100, 18,565-18,579, 1995. Murphy, A. H., Skill scoresbased on the mean squareerror and their 1997.

30938

ROCHFORD

ET AL.:

SUBSURFACE

relationshipsto the correlation coefficient, Mon. WeatherRev., 116, 2417-2424,

1995.

Murtugudde,R., M. Cane, and V. Prasad,A reduced-gravity,primitive equation,isopycnalocean GCM: Formulationand simulations,Mort. WeatherRev., 123, 2864-2887,

1995.

Niiler, P. P., and E. B. Kraus,One-dimensional modelsof the upperocean, in Modelingand Predictionof the UpperLayersof the Ocean,pp. 143172, Pergamon,New York, 1977. O'Reilly, J. E., et al., SeaWiFSpostlaunchcalibrationand validationanalyses,part 3, NASA Tech.Memo. 2000-206892, 11, 49 pp., 2000. Paulson,C. A., and J. J. Simpson,Irradiancemeasurements in the upper ocean,d. Phys. Oceanogr.,7, 952- 956, 1977. Platt, T., S. Sathyrendranath, C. M. Caverhill, and M. R. Lewis, Ocean primary productionand available light: Further algorithmsfor remote sensing,Deep Sea Res., 35, 855-879, 1988. Rochford,P. A., J. C. Kindle, P. C. Gallacher,and R. A. Weller, Sensitivity of the ArabianSeamixed layer to 1994-1995operationalwind products, d. Geophys.Res.• 105, 14,141-14,162, 2000. Schopf,P.S., and A. Loughe,A reduced-gravityisopycnaloceanmodel: Hindcastsof E1 Nifio, Mort. WeatherRev., 123, 2839-2863, 1995. Shriver,J. F., and H. E. Hurlburt, The contributionof the globalthermohaline circulationto the Pacificto Indian Oceanthroughflowvia Indonensia,d. Geophys.Res., 102, 5491-5511, 1997. Sterl,A., andA. Kattenberg,Embeddinga mixed layermodelinto an ocean generalcirculationmodeof the Atlantic: The importanceof surfacemixing for heat flux and temperature, d. Geophys.Res.,99, 14,139-14,157, 1994.

Stevenson,J. W., and P. P. Niiler, Upper ocean heat budget during the Hawaii-to-Tahiti shuttle experiment, d. Phys. Oceanogr., 13, 18941907, 1983.

HEAT1NG

1N AN

OGCM

White,W. B., D. R. Cayan,and J. Lean,Globalupperoceanheatstorageto radiativeforcingfromchangingsolarirradiance andincreasing greenhouse gas/aerosol concentrations, d. Geophys.Res., 103, 21,355-21,366, 1998. Woods,J. D., W. Barkman,and A. Horch, Solar heatingof the oceans, diurnal,seasonaland meridionalvariation,Q. d. R. Meteorol.Soc., 110, 633-656,

1984.

Yuen, C. W., J. Y. Chemiawsky,C. A. Lin, and L. A. Mysak, An upper oceangeneralcirculationmodel for climate studies:Global simulation with seasonalcycle, Clim. Dyn., 7, 1-18, 1992. Zaneveld,J. R., and R. W. Spinrad,An arctangentmodel of irradiancein the sea,d. Geophys.Res., 85, 4919- 4922, 1980. Zaneveld,J. R., J. C. Kitchen,and H. Pak, The influenceof opticalwater typeontheheatingrateof a constant depthmixedlayer,d. Geophys. Res., 96, 6426-6428, 1981. Zaneveld, J. R., J. C. Kitchen, and J. L. Mtieller, Vertical structureof

productivityand its verticalintegrationas derivedfrom remotelysensed observations, Limnol. Oceanogr.,38, 1384-1393, 1993.

R. A. Areone, P. A. Rochford,and A. J. Wallcraft, Oceanography Division,Naval ResearchLaboratory,StennisSpaceCenter,Mississippi, MS 39529, USA. (amone•nrlssc.navy. mil; rochford•nrlssc.navy. mil; wallcraft•nrlssc.navy.mil) A. B. Kara, Centerfor Ocean-Atmospheric PredictionStudies,Florida StateUniversity,Tallahassee,FL 32306, USA. (ReceivedApril 6, 2000; revisedJune 11,2001; acceptedJune20, 2001.)