Characteristics of Wind-Generated Rings in the Eastern Tropical Pacific Ocean

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Scholar Commons | University of South Florida Research

University of South Florida Scholar Commons

Marine Science Faculty Publications College of Marine Science

1-15-2000

Characteristics of Wind-Generated Rings in the Eastern Tropical Pacific Ocean

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

Cesar Fuentes-Yaco University of South Florida

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. and Fuentes-Yaco, Cesar, "Characteristics of Wind-Generated Rings in the Eastern Tropical Pacific Ocean" (2000). Marine Science Faculty Publications. 53. https://scholarcommons.usf.edu/msc_facpub/53

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]. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 105, NO. C1, PAGES 1271-1284, JANUARY 15, 2000

Characteristics of wind-generated rings in the eastern tropical Pacific Ocean

Frank E. Mfiller-Karger and C•sar Fuentes-Yaco Department of Marine Science,University of South Florida, St. Petersburg

Abstract. Eddies are generatedin the easterntropical Pacific(3øS-23øN, 75ø-105øW) by winds blowing through Central American mountain passesfrom the Atlantic. We used CoastalZone Color Scanner(CZCS) and advancedvery high resolutionradiometer (AVHRR) satelliteimagery complemented with monthlyin situ sea surfacetemperature and wind seriesfrom the ComprehensiveOcean-Atmosphere Data Set (COADS) to study these eddies and their effect on pigment concentrationsin the region. Pigment valuesin the Gulf of Tehuantepecgenerally reach higher valuesin November-March before those in the Gulf of Papagayo.The eddiesgenerated in the Gulf of Tehuantepecare associated with passagesof cold fronts acrossthe Gulf of Mexico from the north, while the eddiesoff Papagayoand Panamfiare associatedwith increasesin trade wind intensity.CZCS images showedlarger numbersof eddiesper seasonthan have been previouslyreported on the basis of in situ and AVHRR observations or numerical simulations. We counted 13 eddies in 1979-1980, 8 in 1984-1985, and 6 in 1985-1986. The eddiestransfer both energy and biologicalconstituents from the continentalmargin to the offshoretropical Pacific.The eddiesfrequently moved distancesin excessof 1500 km from their point of origin. Both anticyclonicand cycloniceddies are generated,but in general,there are more anticyclones.Anticyclonic eddies generally moved to the southwest.Some cycloniceddies movedto the south and southeastalong the Central American coastand appearedto be trapped by the cyclonicCosta Rica thermal dome. Eddies traveled at speedsvarying between9 and21 cms -• andhad diameters of 100-500km. Phytoplankton concentrationsassociated with the eddiesvaried from -2 to >10 mgm -3 within-70 km of the coastto -1 mgm -3 up to 600km of thecoast. Between late April andOctober, fewereddies were observed, and phytoplankton concentrations were lower (<0.25 mg m -3) and more uniform over the region.

Abajo se escapael mar al., 1999a,b]. Three mountainpasses lead to very distincteddy en la misma luz que se entrega, generationregions in the Pacificoff Central America, namely, y aunque se escapa,no sale de las manos de la tierra. the Gulf of Tehuantepec,the Gulf of Papagayo,and the Gulf of Panamfi.Eddies generatedin these regionsare referred to Underneath the sea escapes as Tehuano,Papagayo, and Panamefio,respectively. The east- in the samelight it gives, and even though it escapes,it does not leave ern tropical PacificOcean alsofeatures a large cyclonicstruc- the hands of the land. ture calledthe CostaRica thermaldome (CRTD), whichhas a Alfonso Reyes, Poet diameter ranging between 100 and 900 km and is typically (Mexico, 1889-1959) locatednear 8-11øN, 87-90øW [Cromwell,1958]. The eddiesare characterizedby low temperaturesrelative to 1. Introduction surroundingwaters and also by high concentrationsof phytoplankton, which cause marked changesin the color of the The easterntropical PacificOcean, between 75 ø and 160øW water. The high concentrationin pigmentsresults at first from and within the Tropic of Cancer, features numerouscyclonic coastalupwelling associated with the wind jets, and the blooms and anticycloniceddies [Stumpf, 1975; Stumpf and Legeckis, are advectedoffshore in large filaments created by the wind 1977;Hoffman et al., 1981;McCreary et al., 1989;Barton et al., jets. The blooms are later promoted or maintained by up- 1993;Fiedler et al., 1991]. These eddiesare generatedas sea- welling within the eddy structure.Because of the temperature sonal winds in the Gulf of M•xico, and the trade winds from and color signalsof the eddies, satellite imagery servesas an the Caribbean are funneled through narrow mountain passes effective tool to visualize and quantify eddy motions in this in the CentralAmerican mountain range (Figure 1), producing area [cf. Stumpf, 1975; Clarke, 1988;Legeckis, 1988; Umatani strongwind jets that blow over the coastalwaters of the Pacific and Yamagata,1991; Lluch-Cota et al., 1997]. In this studywe Ocean [Matsuuraand Yamagata,1982; Clarke, 1988;Alvafez et provide new insight on the dynamicphytoplankton pigment al., 1989;McCreary et al., 1989;Hansen and Maul, 1991;Fiedler et al., 1991; Lavin et al., 1992; Trasvi•a et al., 1995; Chelton et distributionpatterns associated with the eddiesof this region and present new information in an effort to understandthe Copyright2000 by the American GeophysicalUnion. mechanismswhereby carbon and other materials are trans- Paper number 1999JC900257. ported from the marginsto the interior of the easterntropical 0148-0227/00/1999 JC900257509.00 Pacific Ocean.

1271 1272 MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS

-110 -105 -100 -95 -90 -85 -80 -75 -70

..... ß:.:'-' '"'": ii;.:'½•. . ' ..

ß•:...... ::..•.-(::::..' ... .. • Jet "-' •- Tehuano ' O0 JetPapagayo •'•"Z•.•_•• •l' JetPanamefio :•!::: ••

'L_ -110 -105 -100 -95 -90 -85 -80 -75 -70 Figure 1. Major topographicfeatures of Mesoamericaand adjacentareas based on the ETOP05 database (National GeophysicalData Center (NGDC)). Lowlands(solid), midlands(light shading),and highlands (dark shading)of the studyarea are shown.Arrows indicate the name,location, and directionof the windjets. Wind intensityplots shownin Figure 3 were derivedfrom the row of 2ø x 2ø boxesshown centered on liøN.

We showthat the spatialscales of the eddiesare similar to farther west in the CZCS data. We complementedthese data those of eddiesproduced by the Gulf Stream and Kuroshio with monthly SST and wind series from the Comprehensive currents(---250 km) and that they are typicallygenerated over Ocean-AtmosphereData Set (COADS, 1946-1987). 3-10 days [seeBarton et al., 1993]. We also describeanticy- We named eddiesin the image serieswith a numericalprefix clonic eddiesthat propagateover long distances(100-1,000 identifyingthe year of occurrence.We constructededdy names km) and that last for periodsof up to severalmonths. It was by choosingthe first letter in alphabeticalorder and addinga more difficult to discern coherent cycloniceddies because suffix identifyingthe sequentialeddy number. Finally, a letter thesedissipate within one to severalweeks [see McCreary et al., identifies the region of origin of the eddy (i.e., T, Y, or P, 1989]. The frequencyof eddy generationis highestwhen the dependingon whether the eddy was Tehuano, Papagayo,or wind jets are strongest,which is usuallyin boreal winter and Panamefio,respectively), and a sign denotesthe directionof springevery year, but the Tehuanoeddies are generatedinde- rotation (i.e., plus signfor cyclonicand minussign for anticy- pendentlyfrom thosein the more southerlypasses of Papagayo clonic). and Panamfi. We showthat much larger numbersof eddiesform in the 2.1. Ocean Color Imagery easterntropical Pacificthan has been previouslyreported on The CZCS [Hoviset al., 1980] was an experimentalsensor the basisof in situ observationsor than is predictedby numer- operated betweenOctober 1978 and June 1986 by NASA on ical modelsthat do not includerealistic, high-frequency wind the Nimbus7 satellite.Only imagesthat coveredat least some forcing.This studycontributes to documentingbetter the sta- portion of the eastern tropical Pacific and that contained tisticsand characteristicsof anticyclonicversus cyclonic eddies cloud-freeareas >---200 x 200 km2 wereselected. Figure 2 in this region.Finally, we suggestthat the wind-inducededdies summarizesthe temporal coverageachieved over the lifespan in the easterntropical Pacific Ocean transportmass, energy, of the CZCS over the area of interest. Data were screened with nutrients,plants, and animalsfrom the ocean margin to the a quick-look facility developed at NASA's Goddard Space ocean interior over distancesexceeding 1000 km. Flight Center (Greenbelt, Maryland; softwareby G. Feldman and N. Kuring). There are no major riversdischarging water into our area of 2. Methods study, and therefore the patterns observedin CZCS imagery We examinedthe shape and speedof ringswith multiyear representvariations in colordue to phytoplanktongrowth time seriesof satellite images.Specifically, we use imagery stimulatedby nutrients suppliedvia upwelling and vertical derived from NASA's Coastal Zone Color Scanner(CZCS) mixing. Concentrationswere derived from ratios of the blue and from NOAA's advancedvery high resolutionradiometers (443 nm) or blue-green(520 nm) water-leavingradiance to the (AVHRRs). The area of studywas limited to 75ø-105øWand greenradiance (550 nm), accordingto Gordonet al. [1983a;see 3øS-23øN,even though some eddies were observedmoving alsoGordon et al., 1983b,1988]. At low concentrations(0.04- MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS 1273

½ 50 o E

a. 40

E •-. 30 o

• 20

Date (Month/Year) Figure 2. Temporalcoverage (number of imagesper month) achievedwith the CoastalZone Color Scanner (CZCS) over the easterntropical PacificOcean for the 1978-1986 period. Imageswere countedif they containedan arealarger than 200 x 200km 2 withinour studyregion.

0.5 mgm -3) the CZCSpigment product represents the aver- ActiveArchive Center (JPL-PODAAC; MCSST product).The age pigment concentrationwithin a layer of -1-10 m depth. data are sorted by time and groupedinto weekly bins. Subse- Imageswere examinedat a spatialresolution of -4 km or one quently, data points are geographicallybinned into pixelsof a sixteenthof the original CZCS resolution[see Feldman et al., 2048 x 1024 matrix coveringthe globe(cylindrical equidistant 1989].Clouds were maskedusing a simplethreshold test on the projection),with a resolutionof-18 x 18km 2 at theequator. 750 nm band (channel5). The thresholdwas selectedas the A Laplacian interpolationwas used to fill gaps,with the con- valuewhere the CZCS visiblechannels, particularly the 670 nm dition that one valid retrieval exist within nine pixels of the band (channel4), beganto saturate,a point at which atmo- pixel being evaluated.For this work we extracted a window sphericcorrection is no longerpossible. These pixels, as well as coveringour region of interest from each of 235 weekly aver- areaspotentially contaminated by sunglint,were coveredwith agedglobal maps (October 1981 to December1989). a mask.All imageswere mappedto congruentcylindrical equi- distantprojections. 2.3. Eddy Tracking Becauseof the irregularcoverage provided by the CZCS, we We usedsoftware developed with the IDL TM(Research Sys- used daily and weekly compositesto resolvecoherent spatial tems, Inc.) environmentto visualize the daily and weekly patterns.Available imageswere used to derive weekly arith- CZCS and AVHRR images,to identify eddies,and to observe metic averagepigment concentrations.Valid pixelswere de- the spatial and temporal evolution of each eddy. The visual fined asthose having pigment concentrations between 0.04 and analysesof these sequencesallowed us to determine eddy 7.0 mg m-3; that is, we excludedmissing data, clouds, and rotation direction.Eddy translationspeeds were estimatedby extremelyhigh pigment values.The resultingcomposite im- determiningthe displacementdistance of an eddy center over ageshad the samespatial resolution as the input images.Lo- a specific period using the daily and weekly CZCS and cationsaffected by cloudsor missingdata in successiveimages AVHRR scenes.When available,the daily CZCS data were resultedin smallertemporal bins, thus annual and interannual examinedto assessbetter the positionof an eddy and improve variations in the number of images may reflect changesin cloud cover and the number of eddies observed. speedestimates derived from the weeklyimages. Given that we usedweekly meansand data with spatiallydegraded resolution 2.2. Sea Surface Temperature Imagery (particularlyfor the AVHRR), we estimatethat our error in We examinedsea surface temperature (SST) fieldsusing the the estimateof the positionof the center could be as high as National Oceanicand AtmosphericAdministration (NOAA) 20% relativeto the diameterof the eddy (assuming,for exam- multichannelsea surfacetemperature (MCSST) operational ple, a meantranslation speed of 20 km d-l). productsderived from AVHRR data (MCSST techniquesare Because of the seasonalityin cloud cover in this tropical describedby McClain et al. [1983],Strong and McClain [1984], region,it was difficultto find a sequenceof imagesunder clear and Walton[1988]). The algorithmsused are thoseof McClain skiesbetween May and Octoberexcept during 1979 (Figure 2). et al. [1985]. This product was provided by NOAA (global One Tehuano and three Papagayoseddies were observedbe- retrievaltapes), gridded at the Universityof Miami [seeOlson tween August 9 and 30, 1979. There were few imagesavailable et al., 1988], and is now routinely distributedby NASA's Jet for summers1980-1986, but enoughimages were availableto Propulsion Laboratory, Physical OceanographyDistributed infer some of the characteristicsof the eddies present. In 1274 MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS general, there were fewer clear scenesbetween Panama and and when the trade windsare strongestin the Northern Hemi- the equator than to the north. sphere(December-February). We examinedthe positionof the ITCZ with the COADS data in the 78ø-90øW meridional 2.4. Ancillary Data band. The data showthat the exactposition of the ITCZ over We examined multiyear series (1946-1987) of monthly our area of interestis not predictablewith certainty.However, mean SST and surfacewind speedgridded at 2ø x 2ø from the the annual migrationis evident,showing a 10-14 month peri- COADS. To aid our analysis,we accentuatedvariability in odicity. Its northernmostposition (9ø-12øN) is reachedbe- surfacewinds at various locationsin the region by using the tween June and August, and its southernmost(2ø-3øN) is squareof the wind speed.We also screenedthe NOAA Na- reachedbetween January and March. The COADS wind fields tional OceanographicData Center (NODC) databasefor tem- near Papagayo(Figures 1 and 3) and the satellite-derivedda- perature profilesfrom expendablebathythermograph (XBT) tabaseshow that the frequencyof eddy generationis highest and hydrographiccasts for the period 1966-1990. Where con- when the wind jets are strongest.The wind immediatelyto the current satellite and cast data were available,surface temper- west of the Papagayopass (Figure 3c) is higher than at adja- ature values were compared.We used the NGDC ETOP05 cent COADS grid points.Figure 4 showsmonthly averagesof topographicdatabase, with a spatialresolution of --•5 x 5 min, wind speed immediately to the west of all three mountain to examinethe morphologyof the passesin the Mesoamerican passes.The Tehuano jet seemsto increasein intensityduring Range.These data providedinformation on the anglebetween the boreal winter before the Papagayoand Panamefiojets. As wind jets and the coast. mentionedabove, this phasingaffects the period of eddygeneration. Clearly, we were not able to resolve high-frequency 3. Results variabilityin the wind. Cheltonet al. [1999a]however, show the 3.1. Ring Formation advantagesof usinga satellitescatterometer to observespatial and temporal variabilityin the wind forcing over this region. The three passesin the Mesoamerican(Central America) The strong seasonalwinds from the north (commonlyre- Range focusthe force of winds from the Atlantic Ocean onto ferred to as Nortes throughoutthe region) lead to wind jets narrow stretchesof the coastof the PacificOcean. This forcing overthe Gulf of Tehuantepecwith speeds > 20m s- • sustained varies seasonallywith changesin intensityof (1) the trade over severaldays. These strong, focused winds extend 300-600 windsin the Caribbeanand of (2) windsassociated with cold km over the PacificOcean, generatingfilamentous ocean jets fronts that originate over the North American continent and and high wavesalong their trajectory.Upon the onset of the that move southand east crossingthe Gulf of M6xico [Barton strongwind jets it appearsthat an anticycloniceddy is always et al., 1993;Lluch-Cota et al., 1997].In the Gulf of M6xico,high formed first, curlingto the right of the wind jet [Lavin et al., pressurebehind each cold front createsa large sea level pres- 1992;Trasviha et al., 1995].Cyclonic eddies seem to form when sure differenceacross the Isthmusof Tehuantepec,which gen- a shorterburst of strongwinds is followedby a more quiescent erates northerly winds through Chivela Pass and an intense wind jet over the Gulf of Tehuantepec.Chelton et al. [1999a] period of low-intensitywinds, as offshorewaters seemto act use a NASA scatterometer to show how cold fronts in the Gulf more effectivelyas a barrier to an extendingoceanic filament of M6xico can continuesoutheastward to createrelatively high in this case[Alvafez et al., 1989;McCreary et al., 1989;Trasviha, surfacepressure a day or two later in the southwesternCarib- 1991]. Current meandersare observedbefore circulationpat- bean Sea. The pressuredifference between the Caribbeanand terns closeonto themselvesto form an eddy. These observations are consistent with the first detailed in situ observations the Pacificcoastlines occasionally drives a wind jet that blows acrossthe Nicaraguanlake district.This resultsin strongeast- of the responseof the Gulf of Tehuantepecto the intense,but erly surfacewinds that can extendfar into the easterntropical intermittent,winter offshorewind [Bartonet al., 1993]. Pacific west of the Gulf of Papagayo.High surfacepressure 3.2. Eddy Trains over the southwesternCaribbean Sea alsogenerates northerly surface winds across the Isthmus of Panama and over the Gulf The CZCS data were usefulin identifyingthe date of eddy of Panama in the easterntropical Pacific. genesis,in characterizingtheir morphology,and in tracingthe Cheltonet al. [1999b]show that the Papagayoand Panamefio evolution and dissipationof eddy trains over very long dis- jets are, however, most stronglyinfluenced by tropical atmo- tances(--•2000 km). The high pigmentconcentrations in the sphericcirculation features that have little or no influenceon eddies initially result from coastalupwelling associatedwith the Tehuano jet. They concludethat the Papagayoand Pana- the wind jets. These blooms are advectedand subsequently mefio wind jets are primarily associatedwith variationsin the maintainedor stimulatedby upwellingwithin the eddy struc- trade winds and that exceptfor the occasionalcold front that ture. Thus the featuresseen are not only the result of upward propagatessouth, they are generallydecoupled from low-level or downward Ekman pumping within the core of an eddy. atmosphericvariability at midlatitudes.These findingsconfirm Indeed, it seems that the advected blooms from the coastal the observationsof Legeckis[1988], who suggestedthat there is upwellingregion dominatethe patternsobserved in the CZCS no synchronyin the frequencyof generationof eddiesin these data, both in cyclonicand anticycloniceddies. The AVHRR three regionson the basisof features seen in AVHRR imag- imagesare much lessuseful than the CZCS imagesto follow ery.Lluch-Cota et al. [1997],using CZCS images,arrived at the the eddiesas thesemove offshorebecause surface warming of same conclusion.Lluch-Cota et al. [1997, Figure 2] showthat waters rapidly rendersthe eddiesinvisible in the IR imagery. pigment values in the Gulf of Tehuantepecgenerally reach The combined CZCS and AVHRR satellite data show much highervalues in November-March,before thosein the Gulf of larger numbersof eddiesin the easterntropical PacificOcean Papagayo(February-April). than have been previouslyreported on the basis of in situ As a generalrule, all three jets intensifywhen the Intertrop- observations,AVHRR data alone, or predictionsby present ical ConvergenceZone (ITCZ) is at its southernmostposition numericalmodels. Using both setsof images,we identified 13 MOLLER-KARGER AND FUENTES-YACO' EASTERN TROPICAL PACIFIC RINGS 1275

•'-'- 200 = a) 1'1øN, 81øW Cari'bbean Sea _-- % 150_-- • lOO

_ 0 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988

200• b)1'1øN, 83øW Ca•bbean Sea Z 150= = •100

1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988

'200 c) 1'1øN,85øW Pac'ific Ocean 150

._? 100 •- 50 _• o 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988

..--200: d) • 1505

._• lOO • 50--

1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988

200 =-- e) 1'1øN,89øW PaCific Ocean &•150

._? 100 5O

968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988

Figure 3. Surfacewind intensityfor the period 1968-1987 from ComprehensiveOcean-Atmosphere Data Set (COADS) at five locationsbetween 10 ø and 12øNnear the Papagayowind jet: (a) lløN, 81øW,Caribbean Sea; (b) lløN, 83øW, CaribbeanSea; (c) lløN, 85øW,Papagayo region (Pacific Ocean); (d) 11øN, 87øW, westernPapagayo region (PacificOcean); and (e) lløN, 89øW,northern side of the thermal dome region (PacificOcean). Intensitieswere computedas the squareof the wind speedin order to accentuateevents.

eddies in 1979-1980, 8 in 1984-1985, and 6 in 1985-1986. showfour eddies (Plate 2): 84FEDE3Y-, 84ELADIO2Y-, Tables 1, 2, and 3 presentthe characteristicsof these eddies. 85JOSE5Y-, and 85FLORA6T+. The imagerysequence sug- Plate i shows CZCS images recorded in February and geststhat the Papagayoeddy 84FEDE3Y- (ring a) movedto March 1980(weeks 6-12). Thesedemonstrate the largenum- the NW andthen dissipaied. A similarpattern was described ber of uniquestructures in the regionthat can occurat any one by Hansen and Maul [1991, Figure lb]. The track of buoy time. In February 1980, five eddieswere characterizedon im- NOAA/2084 described two lobes near 12øN, 105øW. Table 3 ages 80W06 to 80W09: 80LINDA9T+, 80LOCO3P-, shows the characteristics of eddies identified in 1985-1986. 80FRANK7T-, 80POLLERA4P-, and 79FEDE2Y-. Four Plate 3, correspondingto January 1986, showsthree Te- more eddieswere describedon March 1980 (80W10-80W12): huanoeddies, two anticyclonicones named 85ANDRES3T- and 80MARTA8Y+, 80MANUEL8T-, 80MARIA10T+, and 85ALE1T- and one cycloniceddy named 85CRISTY2T+, 80JORGE3Y-. Their location is given in Table 1. and the Papagayoanticyclone 85FRANK3Y-. Tracesof other- The trajectoryof buoy NOAA/2212 [seeHansen and Maul, swirls are also visible in the images, which suggestsactive 1991, Figure la] agreeswith the satellitepigment patternsof dissipationof eddies(as in Plate 2). Plate 1. The buoy was draggedto the NW by the Costa Rica The eddiesare most frequentbetween November and April Coastal Current and entrained, alongwith buoy NOAA/2218, of everyyear. Outside this period, we only detectededdies in into 79FEDE2Y- (ring e in Plate 1). After completingtwo August1979 (Plate 4). We noted a minor upwellingevent off revolutionsthe buoysdrifted to the west. Estimatesof trans- Papagayoevery June-July.This is known in the region as the lation velocitiesby Hansenand Maul [1991] coincidewith ours, Veranillo de SanJuan ("San Juan's Little Summer")and is the which are in part based on the propagationof 79FEDE2Y- result of a high atmosphericpressure feature moving into the (Table 1). region. It is accompaniedby a brief hiatus in the rainy season The CZCS compositesfor March and April 1985 (Table 2) and a secondaryannual wind speedmaximum [Ramœrez,1983]. 1276 MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS

10 - ii -- TEHUANO JET - " PAPAGAVO JET _ : ', PANAME•O JET

, ,

:',.:: •i i :.• • •',::• • , • :I i ,',

o i i i i i i i i i I i i 1978 1980 1982 1984 1986 TENYEAR PERIOD ßJAN. 1978 - DEC.1987 Figure 4. Tehuano(solid), Pagapayo (dotted), and Panamefio(dashes) wind jet speedsfor the period 1978-1987. Data were derived from COADS.

3.3. Anticyclonicand CyclonicEddies eddiesgenerally exit our area of studytraveling to the west The anticycloniceddies (identified with a negativesign in (Table 4). Tables1, 2, and 3) are particularlyevident in the warmwaters Plates 1, 2, and 3 show17 anticycloniceddies. From Febru- west of 90øW. These are the most conspicuouseddies gener- ary to March 1980 (Plate 1), therewere four Tehuanos,three ated in all three upwellingsystems. The Tehuano and Pana- Papagayos,and two Panamefiosanticyclonic eddies. Between mefiojets turn anticyclonicallyto the westin a mannerthat is March and April 1985 (Plate 2), one Tehuanoand three Pa- consistentwith jets that are inertiallybalanced at the coastand pagayoswere identified.In January1986 (Plate 3), threeTe- becomeprogressively more geostrophicallybalanced with in- huanosand one Papagayowere seen. In total these anticy- creasingdistance from the coast[Chelton et al., 1999b].By clonicrings account for --•60%of the identifiededdies. trackingfeatures along the peripheryof someeddies we in- Cycloniceddies (identified with a positivesign in Tables1, 2, ferred that a revolutionmay be completedwithin --•18-39 days and3) survivea fewweeks at most.They pump water from the and that the diameterranges between --•150 and 267 km. Av- deeperparts of their centralstructure toward the surfaceand eragetranslation speeds are between 14 and16 cms -q, and havelower temperatures and higher pigment concentrations at

Table 1. Characteristics of 13 Eddies Identified in 1979 and 1980

Final Center Pigment Position Velocity Concentration (Center/ Identifi- Latitude, Longitude, Diameter, Speed, Border), cation in Eddy Name Dates øN øW Shape km Direction cm s-] mg m-3 Figure 5 79ALE1T- Sep.25 to Oct. 9 14.0 96.0 Circle 189-216-225 236 14 0.31/2.50 79CRISTY2T+ Oct. 31 to Nov. 5 15.0 95.0 Circle 118-176 250 13 0.50/4.50 79ELADIO3T- Nov. 11 to Jan. 31 12.0 104.0 Circle 150-184-344 257 14 0.30 79EVA4T+ Dec. 8 to Jan. 11 14.0 95.0 E(NW) 95-111-115 0.20/0.50 79FEDE2Y- Dec. 15 to Feb. 26 9.0 93.0 Circle 339-383 266 11 0.10/0.30 80LINDA9T+ Jan.30 to Feb. 7 13.8 95.0 E(NW) 200-500 0.50/3.5 '80POLLERA4P- Jan.31 to Feb. 16 6.0 79.0 E(NE) 311-195/339 247-208 5.0/0.35 80LOCO3P- Feb. 6 to Feb. 7 3.5 83.0 E(NE) 376 209 15 3.0/0.35 80FRANK7T- Feb. 7 to Feb. 15 13.0 97.0 Circle 247-328 242 18 0.18/0.50 80JORGE3Y- Feb. 6 to March 23 11.5 91.5 Circle 100-374 241-202 14-19 0.07/5.0 80MARTA8Y+ March5 to April 7 10.0 88.0 E(NW) 280 0.35/5.0 80MANUEL8T- March 6 to March 26 15.0 96.5 E(NW) 250 258 16 0.25/0.05 80MARIA10T+ March 8 to March 31 15.0 94.5 E(NE) 150 0.10/0.50 Eddiesare identified as Tehuanos (T; 8 eddies),Papagayos (Y; 3 eddies),and Panamefios (P; 2 eddies).Eddy Name is NNAAAANXS, where NN is year,AAAA is eddyname, with first letter chosen in alphabeticalorder, N is sequentialnumber identifying eddy in a familyor sequence of rings,X is T for Tehuanoeddies, Y for Papagayo,and P for Panamefio,and S is sign,plus sign for cyclonicand minus sign for anticyclonic. Datesindicate the period when the eddy was detectable. Under shape, E is ellipse(long axis orientation). Diameter is varying during eddy life. Velocityis the speedand direction of eddy.Pigment concentration is the rangebetween the centerand the border. MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS 1277

• t

-10, ., ß

-100

0.04 0 85 4.5 (m 9 m"-3 Plate 1. A seriesof weeklycomposites of CZC$-dcri¾½dpigment concentration in the easterntropical Pacific.The sequencestarts with week 6 in mid-February1980 (80W06) and endsin late March with week 12 (80W12). The •ollowin• eddiesc•n be identified:a, 80LINDA9T+; b, 80LOCOSP-; c, 80FRANKTT-; d, 80POLLERA4P-; e, 79FEDE2Y-; f, 80MARTA8Y+; g, 80MANUEL8T-; h, 80MARIA10T+; and i, 80JORGE3Y-.Characteristics for theseeddies are presentedin Table1. Concentrations(mg m -3) were color coded,with blue representinglow pigmentconcentrations and yellow and red indicatinghigher concentrations.Land was maskedgray, the coastlinewas maskedwhite, and cloudsand missingdata are black. their centers than waters surroundingthe eddies. Cyclones The longer series of AVHRR data showed a cyclonic typicallypropagate to the left of the wind jet and have diam- nucleus of cold (21ø-22øC) water between March 7 and eters rangingfrom 95 to 500 km (Tables 1-4). Plate 1 shows April 21, 1985, off Panamfi [see Quir6s and Miiller-Karger, two Tehuanosand one Papagayocyclones, while Plate 2 illus- 1997, Figure 6]. This feature moved over 1180 km from the trates only one cyclonicTehuano. We did not observecyclonic Gulf of Panamfi to the Galfipagos (0.66øS, 87.00øW). Its eddiesgenerated by the Panamfijet in the CZCS data. It is speedwas ---30cm s-• with a directionof ---220ø. Onceit possiblethat we missedsuch eddies because of the extensive reached the equatorial region, it followed a 270ø trajectory cloud cover affecting this region. toward the Galfipagos. 1278 MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS

-8, 80

e .o ß 0.3 o 4 Plate 1. (continued)

Table 2. Characteristicsof Eight EddiesIdentified in 1984-1985'Five Tehuanosand Three Papagayos

Final Center Pigment Position Velocity Concentration (Center/ Identifi- Latitude, Longitude, Diameter, Speed, Border), cation in --3 EddyName Dates øN øW km Direction cms-• mg m Figure 6

84ALEIT- Oct. 28 to Nov. 18 15.5 95.5 125-195-234 237 9 0.24/0.44 84ELADIO2Y- Nov. 14 to March 8 10.0 97.0 258-230-227 247 10 0.20/O.40 84FEDE3Y- Dec. 29 to March 30 9.0 91.5 198-185-267-400 258 9 O.4O/4.72 85JORGE5T- Dec. 11 to Jan. 21 13.0 101 228-374 293 21 0.35/O.20 85FRANK2T- Jan. 14 to Jan. 21 15.0 95.0 225 0.35/5.50 85LINDA3T+ Jan. 14 to Jan. 21 15.0 93.0 100-400 0.35/1.0 85JOSE5Y- Feb. 22 to April 19 11.0 91.0 295 0.65/6.13 85FLORA6T+ March 30 to April 3 13.2 95.1 290 0.08/0.27 MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS 1279

C .

0.0 0.3 .5 (rng Plate 2. Seriesof weekly CZCS imagesshowing the evolutionof severaleddies between March and April 1985 (weeks 9, 10, 12, and 13 or 85W09, 85W10, 85W12, and 85W13, respectively):a, 84FEDE3Y-; b, 84ELADIO2Y-; c, 85JOSE5Y-; and d, 85FLORA6T+.

Table 3. Characteristicsof Six Eddies Identified in 1985-1986:Four Tehuanos,One Papagayo,and One Panamefio

Final Center Pigment SST Position Velocity Concentration difference (Center/ Identifi- (Center/ Latitude, Longitude, Diameter, Speed, Border), cation in Border), Eddy Name Dates øN øW Shape km Direction cm s-• mgm -3 Figure7 øC

85CRISTY2T+ Dec. 4 to Jan. 14 14.0 93.0 247 0.08/1.5 c 22.5/25.1 85ALE1T- Dec. 27 to Jan. 14 15.0 96.0 281 272 16 0.50/0.10 b 28.1/27.9 85ANDRES3T- Dec. 27 to Feb. 5 13.9 98.6 Circle 319-127 267 19 0.30/3.0 a 22.1/27.3 85FRANK3Y- Dec. 27 to Feb. 14 11.0 90.3 Circle 204-384-354 262 21 0.20/5.0 d 29.9/29.5 86FUGAZ1P- Feb. 5 3.0 86.3 135 O.25/O.5O 86JOSE4T- Feb. 5 to March 5 14.7 95.6 E(N) 230/186 0.20/3.0 28.1/28.3 1280 MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS

o .•

.=•o<'• •x

.,•

,. c:;

o. MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS 1281

Table 4. Characteristicsof the Eastern Tropical Pacific Rings

Velocity Position, Deepest Swirl Gyre Diameter, Speed, (Latitude), Layer, Speed, Period, Eddy Family and Type km Direction cm s- • øN m cm s- • days

Tehuano anticyclonic 240 258 16 14 Tehuano cyclonic 220 254 14 14 -- Papagayoanticyclonic 285 251 14 10 100a 70 18 Papagayocyclonic 150 15 700b 25 39 Panamefioanticyclonic 267 218 15 4 -- Panamefiocyclonic • •

Averagesof parametersfrom Tables 1, 2, and 3. aFromMcCreary et al. [1989]. bFromWyrtki [1964].

4. Discussion eddiesmove between 11ø and 10øN.The total changein plan- etary vorticity is ---20-10% in each case. 4.1. Eddy Characteristics Panamefio eddies typically travel to the SW, faster than An eddy'stranslation speed, shape, sense of rotation, and Tehuano and Papagayo rings, and maintain an ellipsoidal other dynamicproperties are defined in great measureby the form. Plate 1 (ringsb in 80W06 and d in 80W07) showsthese environmentalconditions under which it is generated.When eddieswith trajectoriesmore toward the south than the tra- an eddy spins up, a baroclinic structure is developed, and jectory of the Tehuano and Papagayoeddies. Panamefio ed- geostrophicmotion ensues.Its shapemay be due to friction. dies reach low latitudes in a rapid traverse and are then en- For example, the distorted shape that seemstypical of Te- trapped in the equatorial waveguide. huanocyclonic eddies (for example Plate 1, eddylabeled a in 80W06) maybe due to two factors:(1) advectionby the Costa 4.2. Costa Rica Thermal Dome Rica CoastalCurrent and (2) interactionwith the continental shelf [Bartonet al., 1993; Trasviha,1991]. If the continental The Costa Rica thermal dome forms by action of the curl of shelf is about as wide as the eddy, the bottom can affect the the wind stress[Hoffman et al., 1981]. It is intensifiedby ad- speeddifferently in different parts of the eddy. Such is prob- vection from the south via the Equatorial Countercurrent ably the case for eddies located off E1 Salvador,where the (ECC), from the east via the Costa Rica Coastal Current slopeis 0.5 x 10-3 m km-• withinthe first 40 km of thecoast, (CRCC), and from the north in part by the North Equatorial and it increases to 100 x 10-3 m km- • at 100 km from the Current (NEC) [Cromwell, 1958; Umatani and Yamagata, coast. These spatial scalesspan the width of the incipient 1991]. The CRTD is visiblein the AVHRR imageryas a zone eddies,and therefore the movementover the shelfbreak likely where SSTs are consistentlylower by as much as 2-5øC than influencestheir shape. adjacent waters. Its diameter is --•120 km in October and The satellitedata suggestthat the phasespeed of both cy- reachesover 960 km in January.Horizontal temperature gra- clonic and anticycloniceddies, and the number of eddies,are dients between the center of the CRTD and adjacentwaters largerthan what is predictedby models[see, e.g., McCreary et are of the order of 5 x 10-3øCkm -• duringits seasonal al., 1989;Umatani and Yamagata,1991]. It is possiblethat this intensificationand dissipationphases. During maturity, gradi- is due to the frequentuse of monthlymean wind speedto drive entsas high as 30 x 10-3øCkm -• wereobserved. A compar- models,a practicethat typicallyunderestimates and smooths ison between the COADS winds and the AVHRR data con- wind stressvariability. Stronger winds would lead to a deeper firms that the intensity of the CRTD is modulated by the jet and therefore deeper eddies.Forcing with realisticwinds meridional migration of the ITCZ, with strongestSST gradi- with high-frequencyvariability would probably also lead to ents during boreal winter. This supportsLegeckis' [1988] ob- more eddies being generated during a modeled season,as servationsand Umatani and Yamagata's[1991] model. indicatedby the scatterometrywind studiesof Cheltonet al. Starting approximatelyin January, Papagayoanticyclonic [1999a,b]. eddieserode the CRTD and may contribute to its disappear- Cycloniceddies of the Gulf Stream [The Ring Group, 1981] ance as the ITCZ beginsits northwardmigration (April). At differ from the ones observedin our region as follows: (1) other times,however, the CRTD may be reinforcedby cyclonic mesoamericaneddies have translation speedsof 16-18 km eddies.Because of suchinteractions, the thermal gradient be- d-•, whileGulf Streamcold core rings have average transla- tween the center and edgesof the CRTD may intensifyby a tionvelocities of 4-9 km d-•; (2) mesoamericaneddies last factor of 3. This intensificationprocess enlarges the area cov- 1-3 months,while Gulf Streamrings last 7-18 months;and (3) ered by the CRTD suchthat for example,in 1984 it covered the angularvelocity of mesoamericancyclones is --•20cm s -•, some300,000 km 2. while that of Gulf Streamrings is --•150cm s-•. Both the Figure 5 representsSST variationsin the region of the Pa- mesoamericanand Gulf Stream cyclonesare --•700-900 m pagayojet and the CRTD as per COADS. Coastalwaters near deep [Wyrtki,1964] and have diametersof 150-500 km. Also, the Papagayojet do not showa clear annual cycleas opposed the changein planetaryvorticity associated with the meridional to adjacentwaters where the CRTD forms.At the locationof excursionof the eddiesis similar in both regions:Gulf Stream the CRTD the series showstwo minimum temperatures per eddiestraverse latitudes ranging from 37ø to 30øN,while Te- winter for 1980 and 1982-1986. During boreal fall the intensi- huano eddies move between 15ø and 12øN, and Papagayos ficationof the NECC [Hansenand Maul, 1991]helps the dom- 1282 Mf.)LLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS

I I I I , I , [ I 32

•' 30

0 28 o

o• 26

24 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 Papagayo Region, Ten Year Series ' 1978 - 1987

liltlilt[ II,lllSll Illlllll, IlllJlJll IllBllJll IlBlllllJ lrlllllll Illt]llll IlJlllJll ,llltllll 32

•' 30

O 28 o

o• 26

24 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 Costa Rica Thermal Dome Region, Ten Year Series ß1978 - 1987 Figure 5. SST variationin (a) the Papagayowind jet area and (b) the CostaRica Thermal Dome for the period January1978 to December1987 from the COADS. The horizontallines in Figure 5a correspondto gapsin the seriessubstituted by the 20 year mean for this period. ing phenomenon,which carrieswaters from severalhundred layer of thiswestern region were higher relativeto the central metersdepth toward the surfaceand stimulateshigh biological (0.48-2.06mg m -3) or theeastern (1.0-2.37 mg m -3) regions. productionin a manner similar to that seen in the thermal In general, their ship-basedpigment concentrationpatterns dome of the Gulf of Guinea [Voituriez,1981]. agreewith the pigmentvalues and eddyfeatures seen in the CZCS images.The high phytoplanktonbiomass promoted by 4.3. Impact on the Regional Ecosystem the wind-drivenupwelling and eddiesformed during the strong Norres seasonis responsiblefor the high fish stocksin the The inhabitantsof coastalregions of the Isthmus of Tehu- region. Using a model of fish productionfor the Gulf of Te- antepecare familiar with the strongNortes that arrive in win- huantepec,Ritter and Guzmc•n[1984] showedthat the winter ter. Similarly,people living in Nicaragua'slake districtnear the Nortes explain86% of the variabilityin fish concentrationout Gulf of Papagayoand in the PanamfiCanal zone are familiar of 11 environmentalvariables. Fiedler et al. [1991] also ob- with the seasonalincrease in winds as the trade windsintensify. served high phytoplanktonbiomass and productivityin the The effectsof thesestrong winds on fisheriesare well known to euphoticzone in surfacewaters around the CostaRica dome, local fishermen,and internationalpelagic fishing fleets heavily further confirmingthe patterns observedin the CZCS data. exploit this rich area. Indeed, the easterntropical Pacificcon- tainssome of the mostproductive waters of the world'soceans 5. Conclusions [Fiedleret al., 1991].In particular,the Gulf of Tehuantepechas great economicimportance to Mfixico [Ritter and Guzmgn, The three passesin the MesoamericanRange focus the 1984;Robles-Jarero and Lara-Lara, 1993],while the CostaRica seasonal north and trade winds onto the coastal waters of the dome area is vital for the internationaltuna fishery[Cromwell, easterntropical Pacific,leading to the periodicformation of 1958; Wyrtki,1965; Hoffman et al., 1981]. oceanicjets off the Gulf of Tehuantepec,the Gulf of Papa- Trasviga[1991] and Layin et al. [1992]proposed that waters gayo,and the Panamficoast. The frequencyof eddygeneration in the upperlayer (150-200 m) of the Gulf of Tehuantepeccan is highestwhen the wind jets are strongest,usually between be classifiedinto three hydrographicregimes: (1) the western November and April of every year. However, the Tehuano region,characterized by the presenceof eddieswith colderand eddies are formed in responseto strongwinter winds in the saltier water than in the eastern and central Gulf, (2) the Gulf of Mfixico and, particularly,the occurrenceof Nortes centralregion, which is directlyaffected by the northernwinds associatedwith cold fronts, which propagatesouthward from andwhere upwellingbrings cold and saltywater to the surface, the U.S. continental land mass.The Papagayoand Panamfi and(3) theeastern region, with thermohaline characteristics close eddiesrespond more directly to the seasonalintensification of to thoseof the offshore,warm easterntropical Pacific Ocean. the trade winds.The eddiesdominate the dynamicsof surface Robles-Jareroand Lara-Lara [1993] studied the effectof the watersin the easterntropical Pacific in borealwinter and spring. winter Nortes on phytoplanktonbiomass and productivityin The CZCS data were useful in identifyingthe date of eddy the Gulf of Tehuantepec.They noted that biomassand pro- genesis,in characterizingtheir morphology,and in tracingthe ductivitywere similar in the eastern and central regionsbut evolutionand dissipationof eddytrains. Eddies were visiblein that the westernregion was statisticallydifferent. They found the CZCS imagesbecause of their higher phytoplanktoncon- that chlorophylla maxima(0.37-11.2 mg m-3) in the photic centration. This high pigment concentrationresults at first MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS 1283 from coastal upwelling associatedwith the wind jets. These productivityin the eastern tropical Pacific,Limnol. Oceanogr.,36, features are advectedoffshore in meanderingfilaments, and 1834-1850, 1991. Gordon, H. R., D. K. Clark, J. W. Brown, O. B. Brown, R. H. Evans, phytoplanktongrowth is later maintainedor promotedby up- and W. W. Broenkow,Phytoplankton pigment concentrations in the welling within the eddy structure.The satelliteimages suggest Middle Atlantic Bight: Comparison of ship determinationsand that the eddiestransfer both energyand biologicalconstituents CZCS estimates,Appl. Opt., 22, 20-36, 1983a. to the offshoretropical Pacific,which would otherwiseremain Gordon, H. R., J. W. Brown, O. B. Brown, R. H. Evans, and D. K. oligotrophic.Our resultsshow that much larger numbersof Clark, Nimbus 7 CZCS: Reductionof its radiometricsensitivity with time, Appl. Opt., 22, 3929-3931, 1983b. eddies form in the eastern tropical Pacific Ocean than have Gordon, H. R., O. B. Brown, R. H. Evans, J. W. Brown, R. C. Smith, been previouslyreported on the basisof the in situ observa- K. S. Baker, and D. K. Clark, A semianalyticradiance model of tions or predictionsby numericalmodels that do not include ocean color, J. Geophys.Res., 93, 10,909-10,924, 1988. realistic,high-frequency wind forcing.This studyimproved the Hansen, D. V., and G. A. Maul, Anticycloniccurrent rings in the eastern tropical Pacific Ocean, J. Geophys.Res., 96, 6965-6979, statisticsand characteristicsof anticyclonicversus cyclonic eddies. 1991. Our analyseslead to a seriesof questions:(1) What is the Hoffman, E. E., A. J. Busalacchi,and J. J. O'Brien, Wind generation fate of Tehuano eddies as they approach the Papagayoup- of the Costa Rican dome, Science,214, 552-554, 1981. welling region?, (2) Does this processof translationlead to Hovis, W. A., et al., Nimbus-7 coastal zone color scanner: System new cyclonesoff Papagayo?,(3) What is the interactionbe- descriptionand initial imagery,Science, 210, 60-63, 1980. tween suchcyclones and anticyclones?,(4) What is the typical Lavin, M. F., J. M. Robles, M. L. Argote, E. D. Barton, R. Smith, J. Brown, M. Kosro, A. Trasvifia, H. S. Vdlez, and J. Garcia, Fisica del depth, transport, and life of such eddies?,(5) What is the Golfo de Tehuantepec,Cien. Desarrollo, 18, 97-108, 1992. contribution of eddies to the CRCC, the NEC, and the Legeckis,R., Upwelling off the gulfsof Panamaand Papagayoin the NECC?, and (6) What is the fate of Panamefioeddies? The tropical Pacific during March 1985, J. Geophys.Res., 93, 15,485- new information generatedby the JapaneseOcean Color and 15,489, 1988. TemperatureScanner (OCTS) (November1996 to June 1997) Lluch-Cota, S. E., S. Alvarez-Borrego,E. M. SantamariaDel Angel, F. E. Mfiller-Karger, and S. Hernandez Vazquez, E1 Golfo de Te- and the U.S. Sea-viewingWide Field-of-view Sensor (Sea- huantepecy areas adyacentes:Variaci6n espacio-temporalde pig- WiFS) launchedin August1997 will help answersome of these mentos fotosintdticosderivados de satdlite, Cien. Mar., 23, 329-340, questions. 1997. Matsuura, T., and T. Yamagata, On the evolutionof nonlinearplan- etary eddieslarger than the radius of deformation,J. Phys.Ocean- Acknowledgments.We appreciatethe input of Mark Luther, Rob- ogr., 12, 440-456, 1982. ert Weisberg, Guillermo Quir6s, and Gilberto Gaxiola Castro and McClain, E. P., W. G. Pichel, C. C. Walton, Z. Ahmad, and J. Sutton, commentsby two anonymousreviewers. Processing of the data was Multi-channel improvementsto satellite-derivedglobal sea-surface largely carried out on the software environment dsp developed at temperatures,Adv. SpaceRes., 2(6), 43-47, 1983. RSMAS, Universityof Miami, and implementedat the Universityof McClain, E. P., W. G. Pichel,and C. C. Walton, Comparativeperfor- SouthFlorida. We thank Gene Feldmanat the GoddardSpace Flight manceof AVHRR-based multichannelsea surfacetemperature, J. Center (NASA, Greenbelt,Maryland) for providingthe data for the Geophys.Res., 90, 11,587-11,601, 1985. CZCS archive and Elizabeth Smith and Rubby Lassanyi(NODS/ McCreary,J.P., Jr., H. S. Lee, and D. B. Enfield,The responseof the NASA/JPL) for providingthe AVHRR data. Initial processingof the coastalocean to strongoffshore winds: With applicationto circula- serieswas carried out by Guillermo Quir6s (UniversidadNacional, tions in the Gulfs of Tehuantepecand Papagayo,J. Mar. Res., 47, Heredia, CostaRica). Postprocessingand final analyseswere carried 81-109, 1989. out usingsoftware developed at the Universityof SouthFlorida. Laure National GeophysicalData Center, ETOP05: Digital relief of the Devine Castonguay helped with editorial comments, and Denis surface of the Earth, Boulder, Colo., 1988. Nadeau helped with someimage processes.This work was supported Olson, D., G. P. Podesta, R. H. Evans, and O. Brown, Temporal by the CostaRican CONICIT and the UniversidadNacional at Here- variationin the separationof Brazil and Malvinascurrents, Deep Sea dia, Costa Rica, by the Universityof South Florida at St. Petersburg, Res., Part A, 35, 1971-1990, 1988. and by the National Aeronauticsand SpaceAdministration (grant Quir6s, G., and F. E. Mfiller-Karger, Dynamics of satellite-viewed NAGW-678). ringsnear the CostaRica thermal dome, Geofisica,46, 67-103, 1997. Ramirez, P., Estudio meteoro16gicode los veranillos en Costa Rica, Nota de Invest. 5, Inst. Meteorol. Nacl., San Josd, Costa Rica, 1983. Ritter, O. W., and Guzmfin,R. S., Modelo generalizadode producci6n References pesqueracon dependenciaambiental: Una aplicaci6nal Golfo de Alvarez, L. G., A. Badan-Dangon,and A. Valle, On coastalcurrents Tehuantepec,Rev. Geofis.,20, 21-29, 1984. off Tehuantepec,Estuarine Coastal Shelf Sci., 29, 89-96, 1989. Robles-Jarero,E.G., and J. R. Lara-Lara, Phytoplanktonbiomass and Barton, E. D., M. L. Argote, J. Brown, P.M. Kosro, M. Lavin, J. M. primary productivityby size classesin the Gulf of Tehuantepec, Robles, R. L. Smith, A. Trasvifia, and H. S. Velez, Supersquirt: Mdxico, J. Plankton Res., 15, 1341-1358, 1993. Dynamicsof the Gulf of Tehuantepec,Mexico, Oceanography,6, Strong,A. E., and E. P. McClain, Improved oceansurface tempera- 23-30, 1993. tures from space:Comparisons with drifting buoys,Bull. Am. Me- Chelton, D. B., M. H. Freilich, and S. K. Esbensen, Satellite observa- teorol. Soc., 65, 138-142, 1984. tions of the wind jets off Central America, I, Case studiesand Stumpf, H. G., Satellite detectionof upwellingin the Gulf of Tehu- statisticalcharacteristics, Mon. WeatherRev., in press,1999a. antepec,Mexico, J. Phys.Oceanogr., 5, 648-658, 1975. Chelton, D. B., M. H. Freilich, and S. K. Esbensen, Satellite observa- Stumpf,H. G., and R. V. Legeckis,Satellite observations of mesoscale tions of Central American coastalwind jets, II, Regional relation- eddydynamics in the easterntropical Pacific Ocean, J. Phys.Ocean- shipsand dynamicalconsiderations, Mon. WeatherRev., in press, ogr., 7, 648-658, 1977. 1999b. The Ring Group, Gulf Streamcold-core rings: Their physics,chemis- Clarke, A. J., Inertial wind path and seasurface temperature patterns try, and biology,Science, 212, 1091-1100, 1981. near the Gulf of Tehuantepecand Gulf of Papagayo,J. Geophys. Trasvifia, A., Offshore wind forcing in a coastalocean: Observations Res., 93, 15,491-15,501, 1988. and modellingof the Gulf of Tehuantepec,Mdxico, Ph.D. thesis,91 Cromwell,T., Thermoclinetopography, horizontal currents and "ridg- pp., Univ. of Wales, Cardiff, Wales, U.K., 1991. ing" in the eastern tropical Pacific,Bull. Inter.-Am. Trop. Tuna Trasvifia, A., E. D. Barton, J. Brown, H. S. Velez, P.M. Kosro, and Comm., 3, 135-164, 1958. R. L. Smith, Offshore wind forcing in the Gulf of Tehuantepec, Feldman, G., et al., Ocean color, availabilityof the global data set,Eos Mexico: The asymmetriccirculation, J. Geophys.Res., 100, 20,649- Trans.AGU, 70(23), pp. 634-635, 640-641, 1989. 20,663, 1995. Fiedler, P. C., V. Philbrick,and F. P. Chavez,Oceanic upwelling and Umatani, S., andT. Yamagata,Response of the easterntropical Pacific 1284 MOLLER-KARGER AND FUENTES-YACO: EASTERN TROPICAL PACIFIC RINGS

to meridionalmigration of the ITCZ: The generationof the Costa Wyrtki, K., Corrientes superficialesdel Oceano Pacifico Oriental Rica dome,J. Phys.Oceanogr., 21, 346-363, 1991. Tropical,Bol. Com.Int. Attin Trop.,9, 271-304, 1965. Voituriez,B., Les sous-courantsdquatoriaux nord et sudet la formation des d6mes thermiquestropicaux, Oceanol. Acta, 4, 497-506, C. Fuentes-Yacoand F. E. Mfiller-Karger,Department of Marine 1981. Science,University of SouthFlorida, 140 SeventhAvenue South,St. Walton, C. C., Nonlinearmultichannel algorithms for estimatingsea Petersburg,FL 33701.([email protected]. edu) surfacetemperature with AVHRR satellitedata, J. Appl. Meteorol., 27, 115-127, 1988. Wyrtki, K., Upwelling in the Costa Rican dome,Fish. Bull., 63, 355- (ReceivedJanuary 25, 1999;revised August 6, 1999; 372, 1964. acceptedSeptember 21, 1999.)