Saharan Dust and Florida Red Tides: the Cyanophyte Connection

Home , Saharan dust

University of South Florida Scholar Commons

Marine Science Faculty Publications College of Marine Science

6-15-2001

Saharan Dust and Florida Red Tides: The Cyanophyte Connection

John J. Walsh University of South Florida, [email protected]

Karen A. Steidinger Florida Fish and Wildlife Conservation Commission

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

Part of the Marine Biology Commons

Scholar Commons Citation Walsh, John J. and Steidinger, Karen A., "Saharan Dust and Florida Red Tides: The Cyanophyte Connection" (2001). Marine Science Faculty Publications. 163. https://scholarcommons.usf.edu/msc_facpub/163

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. 106, NO. C6, PAGES 11,597-11,612, JUNE 15, 2001

Saharan dust and Florida red tides: The cyanophyte connection

John J. Walsh College of Marine Science,University of South Florida, St. Petersburg,Florida

Karen A. Steidinger Florida Marine ResearchInstitute, Florida Fish and Wildlife ConservationCommission, St. Petersburg,Florida

Abstract. Predictionof the consequencesof harmful algal bloomsfor humansand other vertebratesis constrainedby an inadequateunderstanding of the factorsthat promote their initiation. A simpleexponential growth model of net productionis used for analysis of four time seriesat different samplingintervals over ---40years of red tide strandings, associatedfish kills, and concomitantdust loadingson the West Florida shelf. At least large summerblooms of a toxic dinoflagellateGymnodinium breve appear to be primed regularlyby an aeolian supplyof nutrients.Wet depositionof Saharanmineral aerosols may alleviateiron limitation of diazotrophiccyanophytes, which in turn fuel the nitrogen economyof red tides in the easternGulf of Mexico. Vagaries of the wind-induced circulationand of selectivegrazing pressure on phytoplanktoncompetitors within phosphorusreplete coastalwaters then determineeach year the residencetimes for exposureof G. breve-mediatedneurotoxins to fish, manatees,and humansalong the southeastern United States.

1. Introduction lation is still the result of growthfrom nutrient suppliesof the aphoticzone in thesetemperate ecosystems [Longhurst, 1998]. Both the origin of and the nitrogen supplyfor red tides in In contrast,10% of surfacephotosynthetically active radia- the easternGulf of Mexico have remained obscuredespite a tion (PAR) can be found at a depth of ---30m on the oligo- long history of bloom eventsfirst recorded in 1542 by A. N. trophic subtropicalWest Florida shelf. Thus the chlorophyll Cabezade Vaca. Over the past centurythe seasonalduration stockswithin the upper 0.5 cm of sedimentare usuallytwofold of red tides eachyear off west Florida hasvaried from none to to fourfold those of the overlyingwater column (G. Vargo, 12 months [Steidingeret al., 1998], with >80% of the blooms personalcommunication, 2000), as is alsofound on the Geor- apparentlyoccurring in late summer-fall.Their causativeagent gia shelf [Nelsonet al., 1999]. Pelagicbiomass is then <0.5 •g was not identifieduntil 1947 [Davis,1948] as the dinoflagellate chl L -•. Gymnodiniumbreve, which producesbrevetoxins. Evena modestfall bloomof G. breve(Figure 2a), amounting Bloomsof G. breveoriginate ---20-75 km offshore[Steidinger to ---10•g chlL -• withinthe upper5-10 m, or 50-100 mgchl and Haddad, 1981; Testerand Steidinger,1997] of central Flor- m-2, anda particulateorganic nitrogen (PON)/chl (•mol/•g) ida at depthsof ---15-40 m (Figure 1), yieldingsurface stocks ratio of 0.4 for shade-adaptedpopulations [Shanley and Vargo, of asmuch as 30-60 •g chl L -1 [Carderand Steward,1985]. 1993], with a C/chl ratio of 30, suggeststhat at least 20-40 These red tides force closureof shellfishbeds and regularly mmolN m-2 of dissolvednew nitrogen are requiredto yield drive touristsfrom Florida beaches,with occasionalimpact as such dinoflagellatestocks on the West Florida shelf. Here, far away as North Carolina,where an economicloss of $25 however,the nitratestocks above the 40 m isobath(Figure 2c) aregenerally <0.25 •mol NO3 kg -•, i.e.,<10 mmolNO 3 m-2, million occurredin 1987 [Testerand Fowler, 1990]. At popula- exceptduring episodic intrusions of the Loop Current [Haddad tion levelsof >1 x l0 s cellsL -1, or a biomassof >1 •g chl and Carder,1979; Paluskiewicz et al., 1983] denotedby higher L -1, comparedto backgroundabundances of -<1 x 103 cells salinities (Figure 2b). Recent high-sensitivityfluorometric L -1 in the offshoreGulf of Mexico[Geesey and Tester,1993], methods[Masserini and Fanning,2000] suggestthat the back- discolorations,fish and manateekills, neurotoxicshellfish poi- ground nitrate stockson the West Florida shelf are actually soning,and respiratoryirritation of humansensue within near- <0.01 /xmolNO 3 kg-1. shore waters. At suchlow backgroundlevels, local anthropogenicnutrient While physicalaggregation mechanisms at tidal fronts may supplies,i.e., deposition of NOx in Florida coastal waters lead to local bloomsof dinoflagellatesin the English Channel [Paerl, 1997], may now modulatethe presentyield of today's [Holliganet al., 1983; Tett, 1987] and North Sea [Wyattand phytoplanktonpopulations. However, at only ---37% of the Saborido-Rey,1993; Joint et al., 1997], on Georges Bank estuarinesupply, they are not yet a major factor unlessindus- [Franksand Chen,1996;Anderson, 1997], and within the Bay of trial releaseswere also large in 1542 and somehowmissed by Fundy [Balch,1981; White,1987], aswell as in the Seto Inland de Vaca! Indeed, we shall find that there was no difference in Sea [Kishiand Ikeda, 1986;Yanagi et al., 1995], their accumu- the intensityand durationof red tides between1957 and 1992. Furthermore, during 1947 a large Florida red tide was found Copyright2001 by the American GeophysicalUnion. [Guntheret al., 1948], when eutrophicationpresumably had Paper number 1999JC000123. little impact on the Gulf of Mexico [Walshet al., 1981]. 0148-0227/01/1999JC000123509.00 Moreover,the supplyof nitrate from either northernFlorida

11,597 11,598 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

:/ ... x\ • '.-• i•.... \ '"'. "x"-/X • X X

\ -,X..X-,\ " "'--/•

x.. \\"\ /

i \\

27ø Charlotte \

•8ø -}. FloridaMiddle •rounds / 100m• 200m ,'•i0 m 60 4dm [--] 1958-1961USFWS Time Series •\ •t m 20 rn /• 1966-1967FWCC Time Series \\ ß 1976FWCC/Mote Surveys \ " / 25ø ¸ 1980 FWCCdaily red tide logs • / / / / / / d• 1986FWCC diel time series $ 1957-1998USFWS/FWCC logs

i i i I I i i / 87 ø 86 ø 85 ø 84 ø 83 ø 82 ø 81 ø

Figure 1. The location of 1969-1970 zooplanktontime series(plus) on the 30 m isobathof the Florida Middle Groundin relationto 1958-1961nutrient time serieson the 15 and 40 m isobaths(open squares), to 1966-1967 plankton time serieson the --•5, 25, and 40 m isobaths(open triangles),to 1976 planktontime seriesbetween the 5 and 40 m isobaths(solid pentagons), to 1986nutrient and red tide time serieson the 20 m isobath(solid diamond), and to 1957-1998red tide time serieson the 5-10 m isobaths(open and solidcircles) of the West Florida shelf.

riversor the Loop Current insteadyields both pelagicdiatoms nitrate led to the earlyhypothesis that [Laskerand Smith,1954, in the water columnand benthicones within the upper 0.5 cm p. 175] "G. breve... red tide organismsmight be able to utilize of sediment [Khromov,1969; Saundersand Glenn, 1969; G. atmosphericnitrogen as someof the blue-greenalgae are ca- Vargo, personalcommunication, 2000], not dinoflagellates,as pable of doing."We, instead,pose a direct link betweenpop- alsofound on the Louisiana[Dortch and Whitledge,1992; Nel- ulation incrementsof T. erythraeumand G. breve,as previously sonand Dortch, 1996]and Georgia [Nelsonet al., 1999] shelves. suggested[Chew, 1955; Steidinger, 1973; Taylor and Pollingher, This is hardly a surprisesince the former can grow --•10times 1987], but with Saharan dust now identified as the trigger faster than G. breve and most dinoflagellateshave a half- mechanismfor initiation of at least large red tides. saturationkso, of •7.2/xmolNO 3 kg-• comparedto one of At individualstations of our fixed samplinggrids, maximal •1.3 /xmolNO 3 kg-• for diatoms[Smayda, 1997]. Thus the diazotrophpopulations during any one prior cruiseto the West oxidizedforms of inorganicanthropogenic nitrogen deposited Floridashelf, were instead --•5-10 txg chl L -•, suchthat the within Tampa and SarasotaBays [Paerl, 1997] may also be above tenfold higher amount in May 2000 might have been consumedby diatomsrather than by dinoflagellates. inducedby physicalconcentration, for example,Langmuir cells Although consideredan oceanic resident of tropical seas, and internal waves, effecting small-scalepatchiness [Franks, the diazotroph Trichodesmiumerythraeum appears to play an 1997]. These nitrogen fixers are also found in the South At- importantrole in the nitrogeneconomy of toxicred tidesof G. lantic Bight, where Trichodesmiumspp. impact the plankton breve within coastal waters of the southeastern United States. dynamicsof both the Georgia [Dunstanand Hosford,1977] and Off the west coast of Florida, T. erythraeumis the dominant North Carolina [Guo and Tester,1994] shelves.Here seasonal speciesof these colonialnitrogen fixers,while 'Trichodesmium maximaof Trichodesmiumspp. occur in July,when the Georgia thiebautii tends to be the most abundant form off the east shelfpopulations of --•4.0/xgchl L -• are 100-foldthose of the coast. At very high levels of discolorationseen by aircraft Gulf Stream. These cyanophytesare also associatedwith red above the West Florida shelf the surfacestocks of these cya- tides of G. brevein North Carolinawaters [Westet al., 1996]. nophytesamount to >200/xgchl L -•, asfound off St.Peters- Diazotrophs,such as T. erythraeumfound regularlyoff West burg Beach during May 2000 (C. Heil, personalcommunica- Florida beaches[King, 1950], require iron to form the enzyme tion, 2000). nitrogenasefor fixation of Nn and show increasedrates in The 1947 red tide [Guntheret al., 1948] and the paucityof responseto Fe additions[Rueter, 1988; Rueter et al., 1990;Paerl WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES 11,599

_ _

ß ß ß ß ß ß ß ß ß ß ß ß ß ß ß 1 100 A) ...... \0.1• II /-',' .' .0.1 ...... '10.1 0.1,•._...• ...... io! ...... 20 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I ON DJ FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J

: -/ - D E P T 2O- H 30- (m)

ß ß ß ß ß ß ß ß ß ß ß 4O I i i I I I I I I I I I I I I I I I I I I I I I I I I I I I I I ON D d F MAMJ d ASO NOd F MAMJ d A SON D d F MA MJ d

10' 0 5 ß , ß ß ß ß ß ß ß ß 'e 40 I I I I I I I I I I I I I I I I I I I I I I I I I I I I ON D J FMAMJ J A SON D J FMAMJ J A SON D J F MA MJ J 1958 1959 1960 1961 Figure2. Timeseries of (a) redtides of Gymnodiniumbreve (10 s cellsL -•) on the 15 m isobathand of (b) salinityand (c) nitrate(/•mol L -•) on the 40 m isobathof the WestFlorida shelf during 1958-1961. et al., 1994].During summerthe worldwidehistorical reports of prior to iron-stimulatedblooms of Trichodesmium;afterward, the subtropicaloceanic distributionsof Trichodesmiumspp. theseinorganic and organicforms of phosphorusare undetect- [Carpenter,1983] indeedsuggests that their relativelyhigh but able [Leneset al., 2001].When the coastalpopulations of these nonbloom[Capone and Carpenter, 1982] abundances of >102 cyanophytesare neither Fe- nor P-limited [Karl et al., 1992], trichomesL -• ("-104 cellsL -• or 0.01/•g chlL -•) withinthe are not grazedby mostcopepods [Roman, 1978; O'Neil et al., upper5 m parallelthe >100 mgm -2 yr-• isoplethof annual 1996], and remain at their optimal temperaturesof 25ø-30øC Fe depositionin the differentoceans [Duce and Tindale,1991], [Carpenterand Romans, 1991], they would exhibit maximal downstreamof the adjacentdesert source areas [Sokolikand growth,yielding observed biomasses of at least tenfold greater Toon, 1996]. amountson the West Florida and Georgia shelvesthan off Within the lee of Barbados, however, their seasonalmaxima Barbados. amountto only0.3 /•g chl L- • [Stevenand Glombitza,1972; Subsequentexcretion of dissolvedorganic nitrogen (DON) Borstad,1982] when Saharandust is high [Prosperoand Nees, by Trichodesmium,at a rate of ---50% of their nitrogenfixation 1986]but where ---0.05/•mol PO 4 kg-• prevailover the upper [Glibeftand Bronk, 1994], is consistentwith a ---50% respira- 25 m of the water column [Stevenet al., 1970], typical of tion lossof carbonduring gross photosynthesis [Carpenter and P-limited offshoreregions [Codispoti, 1989; Tyrell, 1999]. In Roenneberg,1995]. This DON could then serveas the initial contrast, the molar ratios of dissolved inorganic nitrogen nitrogensource in offshorebreeding waters for the subsequent (NO 3 + NH 4 + NO2) and phosphatein the Peace River, noxious red tides found later at the coast. During 1999 we enteringCharlotte Harbor [Fraserand Wilcox,1981; McPher- found incrementsof DON after populationincreases of Tri- son et al., 1990], and in the Alafia River, enteringTampa Bay, chodesmiumon the West Florida shelf[Lenes et al., 2001] (also are instead usually <2. They both drain the phosphate-rich found off Hawaii [Karl et al., 1992] and on the Great Barrier Hawthorne formation of centralwest Florida [Dragovichet al., Reef (C. Heil, personalcommunication, 2000)). 1968],such that at least---0.20/•mol PO 4 kg -•, aswell as ---0.20 Like otherdinoflagellates [Pomeroy et al., 1956;Carlsson and /•moldissolved organic phosphorus (DOP) kg-• [Leneset al., Graneli, 1993], G. brevecan alsoutilize ammonium,vitamins, 2001], are alwaysfound within 2-4 km of the Florida coast amino acids,and urea at varyingrates [Wilson,1966; Baden [Dragovitchet al., 1961, 1963;Vargo and Shanley,1985]. andMende, 1979; Shimizu and Wrensford,1993; Steidinger et al., At times,surface stocks of ---0.40/•molP kg-• of bothPO 4 1998]. Some of the DON utilized by G. brevemay be of estu- and DOP are also found above the 40 m isobath at midshelf, arine or seabed origin, as well as from pelagicdiazotrophs. 11,600 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

Since G. breveblooms originate offshore away from near- combineddry (25%) and wet (75%) depositionrate [Prospero shoresupplies of both iron and DON [Ingleand Martin, 1971; et al., 1987]of ---1.25g m-2 yr-•. Mostof thisinflux at Miami Kim and Martin, 1974], however,we considerhere the food occurredin summermonths of the rainy seasonduring 1982 chain impact of nitrogen fixation from an aeolian Fe supply, and 1983;note that there were large red tides alongthe West describedby Darwin [1839,pp. 5-6]: "I collecteda little packet Florida coastin both years(Figure 3). We thusexamine rain- of this brown-colored fine dust... From the direction of the fall and wind data from Tampa, Sarasota,Venice, and Fort wind wheneverit hasfallen,... we maybe surethat it all comes Myers recordsduring 1976-1996. from Africa... It has often fallen on ships... even more than With an iron compositionof 3.5% of mineralaerosols [Duce a thousandmiles from the coastof Africa and at pointssixteen et al., 1991;Zhu et al., 1997] an influx of ---80% of the annual hundred miles distant in a north and south direction." loading,i.e., 0.6 mmol Fem -2, mayoccur during just 1 month In this paper, we explore the causal chain of conditional at Miami, aswell as abovethe West Florida shelf.The impacts probabilitiesthat must all occurto permit land fall of large red of Saharandust can be found both throughoutFlorida in soils tides on west Florida beaches: Saharan dust; rainfall; iron- [Landinget al., 1995]and (by satellite)above the easternGulf stimulatedTrichodesmium; DON release;selective grazing on of Mexico[Carder et al., 1991;Lenes et al., 2001],where arrival fastergrowing dinoflagellate competitors; initiation of G. breve of Saharanaerosols during June-Augustmay stimulateiron- red tide; onshorecurrents; P replete regime for bloom forma- starvedcyanophyte blooms on the oligotrophicWest Florida tion. We examine ---40 years of plankton data on the West shelf followed by red tides of G. breve. Florida shelf and of concurrentdust observationsat Fort My- For example, at estimated dissolutionrates of---10% for ers, Miami, and Barbadosto test this hypothesis. iron aerosols[Moore et al., 1984; Zhuang et al., 1990] a dis- For example,background levels of 0.1-0.5 nmol Fe kg-• solvediron stockof ---12nmol Fe kg-• mightthen have been were measured[Lenes et al., 2001]with a fiber-opticspectrom- suppliedcumulatively from such1982-1983 dust eventswithin eter in surfacewaters of the outer West Florida shelfduring an a 5 m surfacemixed layer over 1 month. Recall that we ob- August 1998 Ecology and Oceanographyof Harmful Algal servedsimilar amounts at a few stationsin July 1999 [Leneset Blooms(ECOHAB) Florida cruisein the absenceof major al., 2001].A simplelinear regressionof offshoreiron stocksin cyanophyteor dinoflagellateblooms that summerand fall (Fig- the West Florida shelf duringECOHAB cruisesand monthly ure 3). In contrast,after documentationof largeaerosol optical total dust at Miami, of all events with mineral aerosol concen- thicknessevents by advancedvery high resolutionradiometer trationsof > 10/•gm -3, yieldsan r 2 of 0.92over June-October (AVHRR) imagery(see http://www.nrlmry.navy.mil/aerosol/)1999. and localradiometric data in St. Petersburg,as well asby direct The molar particulateN/Fe ratio for Trichodesmiumis only observationsof wet depositionof Saharan dust on June 27, 465 [Rueteret al., 1992], comparedto --•5000[Martin et al., 1999,at Miami, surfacevalues of 16.0 nmol Fe kg-• were 1989;Brand, 1991] for nitrate-usingdiatoms [Timmermans et found at the edgeof the shelf,above the 175m isobath,on July al., 1994]and ---16000[deBaar et al., 1990]for smaller,ammo- 6, 1999. nium-usingflagellates [Sunda and Huntsman, 1997]. On the At the same time a tenfold incremental increase of Tri- basis of an inverse relation of molar N/Fe ratios and half- chodesmiumcolonies occurred in this regionover thosefound saturationconstants for Fe uptake,with kFe of "•0.03 nmol Fe on the precedingECOHAB surveyof June 5-8, 1999, deplet- kg-• for flagellatesand ---0.12 nmol Fe kg-• for netplankton ingboth PO 4 andDOP concentrationsto <0.01/•mol P kg-•. diatoms [Leonard et al., 1999], the kFe for Trichodesmium During October 1999, after prior landward movementof the mightbe ---1.0nmol Fe kg-•. Laboratorymeasurements of diazotroph stocksand, presumably,red tide inocula during nitrogen fixation per unit biomassof these cyanophytes,in August-September,large G. brevepopulations of >5 x 106 responseto Fe enrichments,indeed suggest a kFe of "• 1.0 nmo1 cellsL -• werethen found nearshore on the 5-10 m isobaths, Fe kg-• [Rueteret al., 1990]. betweenTampa Bay and CharlotteHarbor (Figure 1). Were At backgroundlevels of dissolvediron of <0.1 nmolFe kg-• these one-of-a-kindphenomena during 1999 and 2000 or dif- on the West Florida shelf during the dust-freeperiods of May ferent realizationsof the usualsequence of eventsdefining the and October the cyanophyteswould thus be under severeFe plankton successionleading to large red tides on the West limitation. In contrast,during Saharandust events,maximal N Florida shelf?Small red tidesmay be anothermatter, primed fixationshould ensue if rainfall and phosphorusstocks are also perhaps by the vagaries of decompositionof sea grass and sufficient.We first examinethe timing of Saharandust events terrestrial plant debris. at Miami or Barbados during 1980, 1976, 1967, and 1966 in relation to phytoplanktoncell counts,as an initial test of our proposedfood web link betweenFe-starved cyanophytes and 2. Methods N-starveddinoflagellates in coastalwaters off westernFlorida. Finally, we considerinterannual variation of helicoptersight- 2.1. Dust Time Series ings of the cyanophytes,of July-Decemberabundances of G. The length of dust time seriesat Fort Myers on the south- breve,and of Julystocks of Barbadosdust over differentrecord west coastof Florida is not as extensive[Prospero, 1999a] as lengthsduring 1957-1998. thoseat either Barbados(1965-2000) or Miami (1974-2000), but concurrentdust data during 1995-1996 [Prosperoet al., 2.2. Plankton Time Series 2001] suggestthat the seasonalcycle of a summermaximum of The most extensiveset of plankton data consistsof micro- Saharanaerosols, reflecting northward movement of the pre- scopicenumeration of G. breveand T. erythraeumpopulations vailing winds [Swap et al., 1996], is the same at --•26ø30'N, obtainedby both the U.S. Fish and Wildlife Service(USFWS) 82øW, ---26øN, 80øW, and ---13øN, 60øW. The mean annual and the Florida Fish and Wildlife Conservation Commission stockof mineral aerosolsabove both Miami and Fort Myers is (FWCC) at 25-30 stations(all nearshoresymbols of Figure 1) about 40% of that at Barbados[Prospero, 1999b], yielding a within 9 km of the west Florida coast, from off Cedar Key WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES 11,601

..-1

( [-I SllOO) ( [-I SllOO) ( [-I SllOO) ([-I SllOO) ( •-I SilO:2)) 11,602 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

(29ø09'N)to near Cape Romano(25ø54'N), during 1957-1998. G. breve(Figure 4) exhibitupward movements, the sinkingloss The USFWS and FWCC assessments of red tides were based of the last term of (1) is considerednegligible. The respective on live countsof G. breveunder x45-54 magnificationof -6 grossgrowth rates #•,2 are0.70 and 0.25 d- • andthe respira- mL aliquots [Dragovichet al., 1961; Steidingerand Williams, tionrates •,2 are0.35 and 0.05 d-•; onlythe results are shown 1970]. whenthe grazingrates •/•,2 are zero.During the 1980analysis When cyanophytecounts of trichome abundanceswere un- we predictthe stocksof the two phytoplanktongroups forward available but aircraft/helicoptersurveys provided qualitative in time; for the 1957-1998time serieswe predictbackward to assessmentsof their heavy and light abundances,we assigned estimatethe date of initiationof red tides(solid hemispheres respectivestocks of 1 x 10s and2 x 106 cellsL- • (Figure3) of Figure 3). on the basisof concurrentshipboard and aerial data.An adap- tive samplingmode was usedby FWCC in which daily obser- 3. Results vations were made when and where red tides stranded once deadfish washed ashore to suggestthe presenceof brevetoxins. 3.1. Daily Time Series: 1980 To deal with suchpatchiness of both samplesand red tides,we A meanTrichodesmium population of 25,000trichomes L-•, simply sum the alongshoreabundances of G. brevecells each or an abundanceof 2.5 x 106 cellsL -• (assuming[Carpenter day, presentingthe weekly mean of theseintegrals in Figure 3. andRomans, 1991] 100 cells trichome-•), was observed within During 1976,surface samples at longer2 week intervalswere surface waters of 3 FWCC red tide stations off Charlotte taken along cross-selfsections of six stationsout to the 40 m County(open circlesof Figure 1) on July 7, 1980, at optimal isobath(pentagons of Figure 1). Live countswere then made temperaturesof >29øC. An earlier Saharan dust event of ->5 of G. breve,other dinoflagellates,Trichodesmium, diatoms, mi- /•g m-3 of mineralaerosols was found on June21-30, 1980 croflagellates,and ciliatesin relation to tenlperature,salinity, [Carderet al., 1991], similar to those measuredat Miami in chlorophyllstocks, and inorganic suspendedmatter. In the 1982-1983 [Prosperoet al., 1987]. This cyanophytestock is Hourglassstudy of 1966-1967 [Steidingerand Williams,1970] equivalentto a pigmentbiomass of -3.0/•g chlL -• (Figure5) the same live countswere made at even greater monthly in- at theircellular amounts of 1.2 x 10-6/•g chlcell -• [Borstad, tervals over three depths of two sets of stationsat the -40, 1982]. -25, and -5 m isobaths(open trianglesof Figure 1), with These light-adaptedphytoplankton have a C/chl ratio of additionaldetermination of the abundanceof copepodnauplii. -200 [Carpenterand Roenneberg,1995] and a molar C/N ratio The data of this time seriesare presentedas the meansover of -6 [McCarthyand Carpenter,1979; Carpenteret al., 1987], the water column of the paired stationsat these isobaths, suchthat their PON/chl ratio is -2.8, comparedto that of 0.4 exceptfor the cyanophytestocks, which are meansof just the assumedfor G. breve.An iron requirementof-18 nmol Fe surface data. kg-•, froma PON/chlratio of 2.8 andN/Fe ratioof 465,could Finally, during 1972-1974 at coarse approximately bi- have then been met during shorewardtransit from mainly monthly intervals,estimates of total zooplanktonabundance offshoreiron stockslike thoseeither measuredin 1999 [Lenes over the same inner-shelfregion were obtained [Houde and et al., 2001] or computedduring 1982-1983. In a Redfieldian Chitty,1976] from Bongonet haulsat -20 stationsbetween the world[Walsh, 1996] the >0.60 tzmolPO 4 q-DOP kg-• on the 50 m isobathand the coast,from off Clearwater(28øN) to near West Florida shelfin June 1999would alsoyield at least3.3/•g Naples(26øN). Another time seriesof net-caughtzooplankton chl L -• of Trichodesmium,without having to invokevertical biomass at monthly intervals is available from one station migration[Karl et al., 1992] duringonshore transport. abovethe 30 m isobathof the Florida Middle Ground (plusof The maximalgross growth rate maybe only-0.70 d-• for Figure 1) during 1969-1970 [Austinand Jones,1974]. During Trichodesmium[Carpenter and Roenneberg, 1995], compared to FWCC cruisesin March, August, and October 1992 the feed- -1.4 d-• for microflagellatesand diatoms in priormodels of ing preferencesof the dominantcopepods were studied[Klep- the Gulf of Mexico and the CarribeanSea [Walshet al., 1989, pel et al., 1996]. At very high spatial resolutiona combined 1999]. These cyanophytesand the slowergrowing dinoflagel- opticaland net samplingsystem also provided an assessmentof lateswould thus be at a competitivedisadvantage if the other zooplanktonabundances during September1998 within a 90 phytoplanktongroups had accessto either N 2 or DON. More- km section taken off Sarasota between the 10 and 50 m iso- over,their 50% respiration/excretionlosses of C/N suggestthat baths [Suttonet al., 2001]. the net populationgrowth rate of cyanophyteparticulate matter mightinstead be -0.35 d-• underno limitationand neg- 2.3. Simple Growth Model ligible grazinglosses. Such computed exponentialincrements of T. erythraeum A Lagrangian formulation of the population dynamicsof stockdo fit the biomassobservations (solid and open circles, both T. erythraeum(P•) and of G. breve(P2) is describedby respectively,of Figure 5) off Charlotte County during the OPi/Ot= #iPi- ½iPi- 'YiPi- O/OzwiPi, (1) longerSaharan dust event (>5.0/•g m-3 of aerosols)sampled at Miami on June 21-30, 1980. Our simple growth model wherethe respectiveloss terms are respiration(and equivalent suggeststhat these diazotrophpopulation changes might be DON excretionin the caseof the cyanophytes),grazing, and attributedto local unlimitedgrowth in responseto iron fertil- sinking instead of lateral mixing [Kiersteadand Slobodkin, ization.Note that a progressivevector diagram of near-surface 1953]. The first term is grossphotosynthesis (and nitrogen windvelocities at the Tampaairport (as well asat Sarasotaand fixationof the cyanophytes),while uptake of DON and DOP Fort Myers, Florida; not shown)shows steady southwesterly by the dinoflagellates,like that of PO4 by the diazotrophs,is a windsprevailed during this seconddust event (Figure 6). function of net photosynthesisto maintain Redfield ratios The nearshoretime series(Figure 5) thuspresumably sam- [Walsh,1996]. pled the historyof successiveoffshore surface Lagrangian pop- SinceTrichodesmium spp. [Villareal and Carpenter,1990] and ulations,as they were advectedonshore past the samplingsites WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES 11,603

Or • ß• ß30.5/ / ---30.0 --- •-- 29•.5I A) I . . . ! I 80.0 30•5! 3(•.0 295// I I •/ e/i / ' Temperature .05 bl .1

•0

2O 0830 1130 • 430 • 730 2030 2330 0230 0530 0830 2 - 3 October • 986 Figure4. Thevertical distributions of (a) temperature(øC), (b) phosphate(•mol kg-•), and(c) estimated biomass(•g chlL -•) of G. breveduring October 2-3, 1986,above the 20 m isobathof theWest Florida shelf. undersuch wind forcing [Yang and Weisberg,1999; Yang et al., phytoplankton;alternatively, nitrogen fixation may not have 1999].The declinein cyanophytebiomass after July 7, 1980, been at the maximalrate during the last week of July 7-14, under renewedonshore winds may be due to cessationof 1980.In anycase, a maximalnet growth rate of ---0.20d -•, at Saharaniron supplyrather than to phosphoruslimitation. Cir- the sameoptimal temperaturesas for Trichodesmium[Stei- culationmodels suggest that the Trichodesmiumpopulations dingeret al., 1998],by G. breveprovides another exponential fit would then have been retained within the nearshore waters, to most of the red tide observationsduring June-July 1980 receivingsummer estuarine influxes of phosphorus. (solidand open squaresof Figure5), suggestingthat their With a meanTrichodesmium biomass of just1.1 •g chlL -• nitrogensupplies were then adequate. over June20 to July 14, 1980 (Figure5), a 50% dailyloss of Like the cyanophytes[Guo and Tester,1994], red tides are theirnitrogen fixation as excretion(0.35 d -• x 1.1 •g chl thoughtto be unpalatableto most grazers,such that only L-• x 2.8/.rmolPON/.rg chl-•), i.e.,---1.1 /.rmol DON kg-• respirationlosses may usually apply to thistoxic group in (1), d-•, wouldyield a cumulativerelease of 27.0•mol DON kg-• unlessno other phytoplanktonare present[Turner and Tester, overthe nearshorewater column during the 25 dayperiod. Full 1997].In contrast,their dinoflagellatecompetitors are readily utilizationof the cyanophyte-releasedDON by a red tide eatenby copepodherbivores on the WestFlorida shelf, where withinthe upper5 m, where10% of the iron of Saharandust summer-fall diets of the macrozooplanktoncontain greater waspreviously assumed to dissolve,would then lead to a G. percentagesof dinoflagellatesthan thosein the sea,implying breveaccumulation there of 67.5•g chlL -• onJuly 14 if all of selectivefeeding [Kleppel et al., 1996].We shallfind that these theother dinoflagellates were either selectively grazed [Kleppel nontoxic,faster growing dinoflagellates, with a net growthrate et al., 1996]or poisoned[Freeberg et al., 1979]and no ammoni- of ---0.60d- • at 25øC,actually precede G. brevein responseto fying bacteriawere present. releaseof DON by the cyanophytes(Figures 7 and 9). The meanG. brevestock of 1.9 x 106 cellsL -• on July21 Thesimple model yields computed stocks of 1.0x 106 cells 1980(Figure 5), isequal to 19/.rgchl L -• at cellularamounts L -• onJuly 25, 1.0 x l0s cellsL -• onJuly 14, 1.0 x 104 cells of 1.0 x 10-s /.rgchl cell-• for bothcultures [Shanley and L -• on July2, and1.0 x 103 cellsL -• on June21, i.e.,back- Vargo,1993] and field populations in October1976 (Figure 7). groundlevels of 0.01/.rg chl L -• of G. brevewhen the major Thus ---72% of the computedrelease of DON in our simple dustevent began (Figure 5). At this time, T. erythraeumpre- model could have been diverted to the microbial web and other sumablyinitiated exponential growth and releasedDON, as 11,604 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

km off the Florida coast(solid diamond of Figure 1) at unde- tectableamounts of nitrateand ---10/•g chl L -• of the co- 2ot occurringnitrogen fixers. Note, again,the large red tide along -45 the centralwest coast of Florida during 1986,as well as in 1980 (Figure 3). • 14 -40 3.2. Biweekly Time Series: 1976 •o12 How often do suchcyanophyte blooms precede ones of toxic 30• •.10 dinoflagellatesat midshelf,where the phosphorussupplies may -25 • not supportlarge biomassof both phytoplanktongroups as in

o coastalwaters? During 1976,surface samples at longer2 week intervals of cross-selfsections, taken out to the 40 m isobath, -15• again indicated that populationsof T. erythraeumoccurred

-10 earlierthan those of G. breve(Figure 7a), asin 1980(Figure 5). AlthoughSaharan dust samples were not collectedat Miami 0:6•156/20 6/25 6,/30 7 /15 57/30 beforeJune during 1976,the winter-springamounts of mineral -10-5 0 5 10 T#½hodesmium(O,e)event(days of > O.1/•g Chl I-1 aerosolsin January-May of other yearssum [Prospero,1999b] _ Gymnodiniumbreve ( [3,ß) -15-10-5 0 5 10 15 2o 2•event(dayS)of>O.1/•gChll.1 to only 15% of the annualdust loading (Figure 8a). Thuslittle iron was evidentlyavailable to the first springpopulation of T. Figure 5. Mean observed (open symbols)and computed erythraeum,such that only a small red tide of G. brevewas (solidsymbols) daily biomasses (/•g chlL -•) at the surfaceof found nearshoreduring May 1976,in responseto the seasonal the nitrogenfixer T. erythraeum(open and solid circles)and onsetof downwellingfavorable winds from the south(Figure the toxic dinoflagellateG. breve (open and solid squares) within •--5 km of the west coastof Florida at three sampling 7b), i.e., onshoresurface flows during the previousApril. sitesoff Charlotte/LeeCounties (•--26ø40'N-26ø55'N) in rela- No incrementof other dinoflagellateabundance (Figure 7c) tionto Saharandust (striped columns) events (>5.0/•g m-3 of wasassociated with the first cyanophytebloom in 1976as well. mineral aerosols)observed during June 10 and 21-30, 1980,at Note that duringthe massiveamounts of Trichodesmiumoff St. Miami (25ø45'N). Cell countswere convertedto biomass,as- PetersburgBeach in May 2000, no red tide co-occurred,at a suming1.2 x 10-6/•g chlcell- • for theformer and 1.0 x 10-s time when the dissolved iron stocks on the outer West Florida /•g chl cell-] for the latter. shelfwere -<0.2 nmol Fe kg-• [Leneset al., 2001]. Sincerainfall peaksduring summeron the southwestcoast of Florida (Figure 7b and 8b), wet depositionof mineralaero- observedin 1999 [Leneset al., 2001], while herbivoresgrazed solswould alsobe favoredafter June 1976.A Julycombination on the other dinoflagellates.Macrozooplankton biomass on of greater dust loading, larger rain fall, and mainly onshore the West Florida shelf during June-Novemberis usuallyten- surfaceflows led to both a secondsummer population of T. fold that of February-March [Austinand Jones,1974; Houde erythraeumby early August 1976 (Figure 7a) and a tenfold and Chitty, 1976], such that increasedand selectivegrazing incrementof other dinoflagellates(Figure 7c), precedingthe pressureon nontoxicdinoflagellates would allow slow summer fall red tide of G. breve.As in 1980(Figure 5), a dailythreshold accumulation of G. breve in both the "real world" and our of ---5/•gm -3 of mineraldust at Miami(Figure 7a), or ---12.5 model. During August 1969 on the Florida Middle Ground /•g m-• at Barbadoscloser to the Saharansource, seems to (plus of Figure 1), however,zooplankton abundance was low, providesufficient iron for cyanophyte-inducedred tidesof G. tenfold less than that found at some stationsthere in August breve.On the basisof the regressionof offshoreiron stockson 1973, such that no red tide occurred in 1969, compared to the West Florida shelfand monthlytotal dustat Miami during thosefound in 1973 and 1980 (Figure 3). June-October1999, 5/•g dustm -3 -•0.5 nmolFe kg--•, which Upwellingfavorable winds on July 19-21, 1980 (Figure 6), approximatesour estimateof kFe for Trichodesmium. would have advectedthe coastalpopulations offshore in the Note that a short July upwellingevent fostereda brief sur- surface Ekman layer [Yang and Weisberg,1999; Yang et al., face diatom bloom, curtailedby either inadequatesupplies of 1999],leading to an exportof the nearshorered tide, as seenby nitrate, e.g., Figure 2c, and/or selectivegrazing again by her- concurrentCoastal Zone Color Scanner(CZCS) imagery(F. bivores. During a spring bloom of diatoms from estuarine Muller-Karger, personal communication,2000). When on- nutrient sources on the West Florida shelf in March 1992 shorewind forcing resumedon July 25, 1980 (Figure 6), the [Gilbeset al., 1996]this siliceous group constituted ---20-60% nearshoretime seriesrecorded another population increment of the dietsof the dominantcopepods, compared to •--10-30% of G. breve(Figure 5), buta smallernet growth rate of 0.10d -• of the availablefood [Kleppelet al., 1996]. During October for thesedinoflagellates now fits the Augustobservations, sug- 1992 the copepodsthen selectedother dinoflagellatesat mid- gestingnitrogen limitation once more. shelfduring a red tide of G. brevealong the coast(Figure 3), SinceG. brevehas the ability [l/argoand Shanley,1985; I/argo with, again,a diet of---35-65% dinoflagellates,compared to a and Howard-Shamblott,1990] to utilize DOP efficientlylike food choice of •--25-30%. Trichodesmium[Yentsch et al., 1972], the stocksof just organic When upwellingfavorable winds from the north provided phosphorus(•--40% of DOP + PO4) in coastalwaters off offshore flows in the surface Ekman layer over the rest of central Florida are sufficientto meet the daily demandsof G. Augustand most of September1976, however, the populations breveat populationlevels of up to ---6.5/•gchl L -• [Steidingerof both N2-fixing cyanophytes and nontoxic/toxic DON- et al., 1998].Indeed, duringan October1986 red tide of a mean utilizing dinoflagellatesdispersed to low backgroundlevels •--13/•gchl L- • of G. breveabove the 20 m isobath(Figure 4c), (Figure7). A diatombloom did not occurthen, suggestingthat asmuch as 0.30/•mol PO 4 kg- • werestill found (Figure 4b) 30 the selectivegrazing pressure,which allows G. breveto out- WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES 11,605

TIA June-July 1980 (0.0) - 6/01/80 2000

100

- 1000 1 2000 3000 4000 5000 Kilometers

-1000 Figure6. Progressivevector diagram, plotted head to tail,of the hourlynear-surface winds (m s-z) at TampaInternational Airport (27ø58'N)during June-July 1980. The durationof the summerdust events in 1980 are denotedby the solidsquares beginning on June 10 and on June 21-30. competeother faster growingdinoflagellates, must have re- (Figure 9b), indicatinggreater reproductivesuccess of the moved siliceous rivals as well. With return of onshore flows in adults in the secondyear. Furthermore,the stocksof other earlyOctober 1976 a red tide of G. brevewas finally found both dinoflagellateswere tenfold less in July 1967, and large red abovethe 15 m isobath(Figure 7a) andalong the coast(Figure tides of G. breveensued during October both in offshorewa- 3) until northeasterlywinds purged the littoral zone once ters (Figures9c) and alongthe coast(Figure 3). more. In contrast,after a smaller June 1966 pulse of dust, high The sum of July rainfall at Fort Myers was similar in 1976 stocksof T. erythraeumwere associatedonly with a bloom of and1980 (15.5 versus 17.8 cm month-1), as were the means of other dinoflagellates(Figure 9b), but not of G. breve(Figure Miamidust [Prospero, 1999b] that month (15 versus 18/.rg m -3, 9a).The copepodnauplii were then <102 L -z andthus also i.e., equivalentto ---38.7and 45.0/.rg m -3 off Barbados)and not plottedin Figure 9b, while a smallred tide wasfound in the subsequentweekly maxima of red tides (Figure 3) along October1966 (Figure 3). The equivalentJune dust loadings in thecoast (3.2 x 106 versus1.8 x 106 cellsl-Z). However,are the easternGulf of Mexico would havebeen monthlymeans of thesejust anomalousyears, or do the first and secondtime --•3.3tzg m -3 in 1966and of --•6.9tzg m -3 in 1967.Clearly, seriesrepresent the normal sequenceof red tide initiation, moreiron of perhaps--•0.7 nmol Fe kg-z wasavailable to the stranding,and dispersal? cyanophytesduring 1967 than in 1966(--•0.3 nmol Fe kg-Z). Similarto the smallred tide in April 1976,the smallone in 3.3. Monthly Time Series: 1966-1967 1966 may have been both indirectlyiron-limited and directly Dust data were not collected in Miami during 1966-1967; grazer-controlled,as suggestedby the differencesin copepod we thususe the Barbadosdata set [Prosperoand Nees, 1986] as abundancesduring the 2 yearsof Hourglassobservations. For a proxy(Figures 9a and9c). DuringJune 1967 dust events the example,off Sarasota,Florida, in the sameshelf region during presenceof T. erythraeumwas qualitatively noted at the Hour- September1998 the maximumestimate of grazingdemands by glassstations (Figure 9c). Cell countsin the previousyear crustaceans between the 10 and 50 m isobaths would have (Figure9a) suggestedthat the populationlevels of the cyano- consumed--•14% of the primary production[Sutton et al., phytescould have been the sameas thosein 1976(Figure 7) 2001], perhapsallowing the fastergrowing other dinoflagel- and 1980(Figure 5). The abundancesof copepodnauplii were latesto bloom,as in 1967 and 1976 (Figures7 and 9). Again, more than tenfold larger in 1967 (Figure 9d) than in 1966 a small red tide was observedin 1998 (Figure 3). Finally, 11,606 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

70 A) Trichodesmiumerythrae. um_• .... ) ./ Gymnodiniumb.revel (•) stocks(105 cell.s I-1)ß 50 •' 60 aerosol'(--"....) ' ..- / '-' ' / :: ,' ', 40 E 50 , .1 ' / ,, ,' / // 40 [ ,, , ::,; :. . :: , ', 30 qn ß ß ß e•• . , ß • .e ß ß •• ß ....,,:,• ,e,,, ,e __• /t ß, :,,, ..-"'"'• ß •• ß ,ß 25 .0 • "" .... i,,, /,' .,lOO•-•-.ß;... xx ,, ...... !i!i, ?!.- .--.•--- /.• ,,!!•' - • 'x._ -/ ,.-" 20 '• ß ß ß , e"-.In'i•,• x• ß '...e. :', e; :',: ;•,' :e.'?'•;,•,,• ', ß .•,', ß )e •e/ ß /e ß ß (•20 ,,.-", ,v ,'I• ',.lOx -, ....• • ',,, .... i:',!:'.'::',' ' ('1•';',,:_'•". '•VV ' -'-.:!"•',' • IN,_,/ ,• .,// •'E• • an '- ..... 1----C-'- -'10.:: ::: ' ',,,• :;,,; / .-' :i • 1 10 /v ß ß ß --ßl -- .ßl/ '- ,: , , [Saharan:,,,:: :, ,,,,,, : ,:, :, !;//:,,,,: ,, ,: ,,, _ ',..i!,..:",,,'i.,," Dust]: ::'-:',.:",..,' '-'",..,' i::..;,:", ...... ?',,!',,,".i ",..'v." ",,:""',,i',,. i i i i i i i 30 Feb 60 Mar 90 Apr 120 May 150 Jun 180 Jul 210 Aug 240 Sep 270 Oct 300 Nov 330 Dec

• 8 B)Upwelling/Downwelling (--)/Rainfall(...... )':i ';:: --o

4 ; {': --

;; ;: '; ', , :[ :; ;'1:1 f,: ?' ; ;;\',; 11;: 1,'4 '1 ;,',; ;;1', ', ,' ;j','; ;' ;;;;;; ; ; :', ", 0,,o o 30 ,,Feb ,,,.,,,.,. 60 Mar 90,.,,,. Apr 120,,, May 150.... Jun 180 Jul 210 Aug 240 Sep 270 Oct 300 Nov 330 Dec

70 c) Diatom (--) / Other dinoflagellates ( ...... ) stocks( 105 cells 1-1) E 60 ß ß ß...... ß ß ß ß ß ß ß ß ß ß ß ß ß

ß 50 o • 40 o ß •i '"oß ß ß o•.,..•6.• o.•ß ...... '"ß ß ß.• ...... ß e 30 ß .,,'•.),,': ß •- ß ß ß ßi'•),,,,•.•.-2...... ß ß ß ,,•--0.5.....'1 ß..... 0.1"i'!•. "'...... -:--, ...... '...... ,o.•--.1,:::...... ,,. . ß •L•F-'-/::/'/' . ....0.;1"...... •...... ! '[i;: -...... I • 20 t .0.5...... ;• ! "..... '""/':"5 " ...... '...... 1:' '"'i0 1 ."• ...... 0T ._• ,_ ...... 1.,...... ß ...... o,,,,;.., .:.1 Q 10 ß,,. 't. -t. ,. ,, ...... 1. ø.'/'i i .•"' i s'i ...... i i i ...... b..'-:.-i ...... i ', i '...o.s• i • •o.'. i ...... I i i ...... i ;,i i .....i '...... i1.---.•... I 30 Feb 60 Mar 90 Apr 120 May 150 Jun 180 Jul 210 Aug 240 Sep 270 Oct 300 Nov 330 Dec Figure7. Themonthly surface abundances (10s cells L -z) of T.erythraeum (dashed line) and G. breve (solid line)along offshore sections of sixstations (solid circles) from Tampa and Sarasota Bays (•--27ø20'-27ø40'N) to the 40 m isobathduring 1976 in relationto the dailysampling of (a) Saharandust (•g m-3) at Miami (dottedline) and (b) bothrainfall (dotted line) accumulation(mm) andupwelling/downwelling favorable winds(m s -z) atFort Myers Page Field (26ø35'N) and (c) the monthly abundances (i0s cellsL -z) ofthe sums of both otherunarmored and armoreddinoflagellates and of centricand pennate diatoms.

duringMay 1972 and 1974the nearshorezooplankton popu- southerlyand northerlyflows, as observedin 1996 [Yanget al., lationswere fivefoldto tenfoldless than thoseof 1973[Houde 1999].In contrast,during other years of 1972,1976, 1980, 1987, and Chitty,1976], and oncemore, smallred tideswere foundin and 1994 the fall bloomswere insteadexported to the east the fall of 1972 and 1974, comparedto 1973. coastof Florida [Murphyet al., 1975;Tester et al., 1991],as in 1999.The noneventsof no red tidesfound during1960-1964 3.4. Annual Time Series: 1957-1998 and 1969-1970are not shownin Figure 3. Duringthe past42 years,28 red tidesof >1 x 104 cellsL -z As sightedby aircraft,helicopters, and shipsafter 1957,the of the alongshoreintegral of G. brevepopulations were found cyanophytesco-occurred along the coasteach year (Figure3) duringmainly summer-fall at stationsalong the centralwest duringat least 87% of the 16 larger red tides,i.e., when the Floridacoast (Figure 1). At populationlevels of >5 x 103cells alongshoreweekly integral of G. breveabundance exceeded a L -z the shellfishbeds are closed.The winter occurrencesof G. valueof 1 x 106 cellsL- • overthe 42 year record. During the breve in 1968, 1974, 1982, 1984, 1989, 1995, and 1996 are mostrecent red tide of >3 x 106 cellsL -z in October2000, consideredpersistent fall bloomsof thepreceding year (Figure large populationsof Trichodesmiumco-occurred as well (C. 3), trappednear the coastby circulation patterns of alternating Heil, personalcommunication, 2000). WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES 11,607

erythraeumon the West Florida shelf is associatedwith Loop A) Current intrusions,carrying populationsfrom upstream re- gionsvia the Brazil and Yucatan Currents,i.e., from the South Atlantic [Calef and Grice, 1966] and Carribean Sea [Hulburt, 1968]. On average,the Loop Current is within the DeSoto Canyon region at --•30øNabout 5-10% of the time [Vukovich and Hamilton, 1989], i.e., 1 month of the year, when wet deposition of Saharan dust loadingsmust also occur to yield the red tides found there. For example, during 1964, no red tide was found at the standard monitoring stations(Figure 1) by either USFWS [Dragovichand Kelly, 1966] or FWCC [Steidingeret al., 1967]. B) Yet that year a large bloom occurredwithin ApalacheeBay, as during 1974 and 1982. These rare eventsrepresent 7% of the 42 year record(Figure 3), i.e., the samefrequency as the Loop Current in the northern Gulf of Mexico. Another red tide just occurred there in 1999, when AVHRR, Sea-viewing Wide 10- Field-of-View Sensor (SeaWiFS), and TOPEX imagery all showeddeep penetration of the Loop Current; this red tide IIIIllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll was also associatedwith local stocksof Trichodesmium(G. 1992 1993 1994 1995 1996 Kirkpatrick,personal communication, 2000). Figure 8. The monthlyvalues of (a) Saharandust at Miami At the edge of shelf southof Tampa Bay the Loop Current (/•g m-3) and(b) rainfall(cm) at Veniceduring 1992-1996. is insteadpresent 30-35% of the time at 27øN [Vukovichand Hamilton, 1989], providingcyanophyte seed populations more frequently.Recall that 16 large red tides,or 38% of the total 42 On the basisof the initial growth rate of the 1980 red tide year record(Figure 3), were insteadobserved off centralFlor- (Figure5), 20 of these28 red tidesfirst formedduring ---June- ida. In addition to cyanophyteprecursors of a red tide, the August,when 71% of the annual Saharandust loading arrives other biotic factor is selectivegrazing on their competitors. at Miami [Prospero,1999b] and 47% of the annual rainfall During at least 1966, 1969, 1972, 1974, and 1998, when red occursat Tampa and Fort Myers. Furthermore, 15 of the 16 tides were absentor minimal within this nearshoresampling largerred tideswere initiatedduring these months (Figure 3). region (Figure 1), low grazingstress may have allowedother Were other red tides undetectedby the FWCC adaptivesam- dinoflagellatesto competesuccessfully against G. brevefor the pling network within ---10 km of the west coastof Florida? DON of cyanophyteorigin. Once a red tide is initiated and Over a 5 year period of 1992-1996, red tides did not strand movedonshore to the samplingsites (Figure 3), however,the within this region during 1993 and 1996; they may have been amountof dust (iron) loadingin July then appearsto deter- offshoreand thus not sampled.Of theseyears, however,July mine how large a yield of toxicdinoflagellates may prevailfrom rainfall at Venice (betweenFort Myers and Tampa) was ---10 July to December (Figure 10) within phosphorusreplete wa- cmmonth -• in 1993and 1996 (Figure 8b), when Saharan dust ters. was>15/xg m-3 at Miami(Figure 8a). Underthe samedust We made no detailedstatistical analyses since past attempts loadingand rain accumulations of >20 cmmonth -•, red tides to correlate red tides with abiotic factors were fruitless exer- insteadformed during both 1994 and 1995 (Figure 3). Thus cises. Little correlation of G. breve variance was found with sufficientrainfall, aswell as Saharandust, must be presentfor hurricanes and/or rainfall [Feinstein,1956], salinity [Chew, wet depositionof particulateiron [Prosperoet al., 1987], but 1956],inorganic nutrients [Rounsefell and Dragovich,1966] and consideringthe poor correlation between past precipitation estuarineiron stocks[Ingle and Martin, 1971;Kim and Martin, eventsand red tides [Feinstein,1956], other biotic factorsmust 1974], without considerationof biotic factors. also be important. We are content to conclude from these four sets of time series in 1980, 1976, 1966-1967, and 1957-1998 that the like- lihood of a large, long red tide at the shorelineemerges from 4. Discussion a sequenceof conditionalprobabilities of (1) summerSaharan A scatterdiagram (Figure 10) of July dust at Barbadosin dust events,(2) sufficientrainfall, (3) dissolutionof aeolian relation to cumulativestocks of G. brevealong the centralwest iron, (4) seedstocks of both T. erythraeumand G. breve,(5) Florida coast(Figure 3) duringJuly-December each year ex- releaseof DON to all dinoflagellatecompetitors, (6) selective hibits thresholdphenomena, as suggestedfor the initiation of grazingstress on other dinoflagellatesas well as diatoms,and red tides of Gymnodiniumcatenatum within Australian waters (7) onshoreflows to completelandfall. [Hallegraeffetal., 1995].Below a monthlymean of 12.5p•g m -3 The origin of small red tides is more enigmatic, reflecting of BarbadosJuly dust, i.e., •5 p•gm -3 of equivalentMiami possibleadditional sediment and estuarinesources of DON. dust,or 0.5 nmolFe kg-• foundon the outerpart of thewest During the small red tide of December 1998 (Figure 3) the Florida shelf, for example,few red tides are observedoff the /5•SPONof --•5/xgchl L -• of G. breveat the 15 m isobathwas Florida coast. +4.8%0, compared to either +0.3%0 for Trichodesmium- Over the full Barbadosrange of 4.1-42.5 p•gm -3 of July dominated populations, fixing atmosphericnitrogen at the dust, however, red tides are still not found near the coast shelf break, or +7.3%o for diatoms above the 30-40 m iso- during someyears, such that Saharaniron loading and rainfall baths, after depletion of nitrate supplyfrom a Loop Current are necessarybut not sufficientinitial conditions.Seeding of T. intrusion.The intermediateisotopic value of the red tide may 11,608 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

80 25 ß ß ß ß ß ß ß ß ß ß ß • l:: 70 A)Trichodesrniurn erythraeurn(.... ) -""•f' 20E .e 6o Gymnodiniumbreve(.• ) ," , • o stocks(105 cells 1-1) during 1966 1' , 100',', • .c 50 ; 10 ,, -o o 40 ' '","i' 15u)

•- 30

.•_ 20

ß i , ,• • , , ,•, •I I • I Jan&30 Feb. ½0 Marß ½0 Apr 120 May 1•0 J•n 1•0 J•l 2•0 A•g 240 Sep 270 O•' 300Novß 330 Dec 80 ß ,o ß ß •o ß __e ß ß ß ß

E 70- / / / v / / ß 60- / / o / / .: ,50. 0.'• / "/ '/ 10. 0.1 o 40. ß --' 0.5 ' , o 30- /

._• 20. B) Otherdinoflagellate (---) stocks(105 cells I-f)-' and nauplJJ( ) abunda•n.ce(103 I-1) during 1966 ,, 10. ß . . 0.5.', . .

Jan 3 Feb 60 Mar Apr 120 May 150 Jun 180 Jul 210 Aug 240 Sep 270 300 Nov 330 Dec 80 25 '-T-' //• C) Trichodesrniurnerythraeurn ( .... ) / ', 2OE e 60 Gymnodiniumbreve ( ) / ', o stocksß (105ß cells •o1-1) •during ß1967 ß /' •,', ß ß .: ,50- .15• '• 40- /// • ...... o .10-o o 30- • ß . .T•,'; ßT •- ½½½• ._• 20. .5 • 10-

80 ß ß ß ß ß ß ß ß ß ß , ß •' 70- •, D)Other dinoflagellate (----)stocks (105 cells I-1)and .!.•X,•auplii (--)abundance (103 I-1)during1967 "

o 30- ' 1...... ,,.:

• 5)0. ._• 10. .,.,I..I/ . .,'",,.:'..',,,/ '. . Jan' 30' Feb' io Mar...... 90 Apr 120 May150 Jun 180 Jul 210 Aug240 Sep 270 &t'300 Nov' 330 ' Dec' Figure9. The monthlysurface abundances (10 s cellsL-•), or presenceT andabsence, of T. erythraeum (dashedline) in relationto boththe mean stocks over the water column (1½ cellsL -•) of G. breve(solid line) at pairedstations on the ---5,25, and 40 m isobaths(solid circles) and the monthly dust (txg m -3) sampledat Barbados(striped columns) during (a) 1966 and (c) 1967.At thesesame stations the meansover the water 5 1 columnof the sum (10 cellsL- ) of other unarmoredand armoreddinofiagellates (dashed line) and of copepodnauplii (103 L- 1 of the solidline) are alsoshown for (b) 1966 and (d) 1967. reflecta •SN-enrichedDON substrate,compared to a Tr/- bacteria may behave like zooplankton [see Checkleyand chodesmiumsource, produced presumablyby bacterial pro- Entzeroth,1985] by recyclingDON to •4N-enrichedNH4 but cessing. leavingbehind some •SN-enriched DON, i.e.,perhaps the urea We are awareof no $•sPONdata for bacteriagrowing on favored by red tides on the West Florida shelf. DON, but like phytoplankton,they prefer the lighterisotope of Unfortunately, the reprocessedDON of Trichodesmiumor- •4N suchthat their nitrificationleads to larger$•N of the igin may not have a unique isotopic signal since estuarine ammonium substrate [Mariotti et al., 1984; Horrigan et al., detritusalso has a $•sPONof + 1.1to +2.9%, [Walsh,1994]. If 1990], while denitrificationalso yields larger valuesof the re- the estuarine $•SDON has the same value as the detritus and is mainingNO3 [Clineand Kaplan, 1975]. During catabolismof palatableto bacteria[Gardner et al., 1996],G. brevemay utilize phytoplankton(PON --> DON --> NH 4 --> NO 3 --> N2) the this humic sourceof organic nitrogen as well [Carlssonand WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES 11,609

50 • References Anderson,D. M., Bloom dynamicsof toxicAlexandrium species in the 451 1965-1996 • northeasternU.S., Limnol. Oceanogr.,42, 1009-1022, 1997. Austin, H. M., and J. I. Jones,Seasonal variation of physicaloceano- 40 graphicparameters on the Florida Middle Ground and their relation to zooplanktonbiomass on the West Florida Shelf, Fla. Sci., 37, • 35 16-32, 1974. Baden, D. G., and T. J. Mende, Amino acid utilization by Gymno- dinium breve,Phytochemistry, 18, 247-251, 1979. • 30 i ß Balch,W. M., An apparentlunar tidal cycleof phytoplanktonbloom- ing and communitysuccession in the Gulf of Maine, J. Exp. Mar. '• 25 ß ß Biol. Ecol., 55, 65-77, 1981. Borstad, G. A., The influenceof the meanderingGuiana Current on •' 20 surfaceconditions near Barbados:temporal variations of Trichodes- mium (Cyanophyta)and other plankton,J. Mar. Res.,40, 435-452, 1982. 15 •' • . ß •. ' , . Brand, L. E., Minimum iron requirementsof marine phytoplankton and the implicationsfor the biogeochemicalcontrol of new produc- 10 tion, Limnol. Oceanogr.,36, 1756-1771, 1991. Calef, G. W., and G. D. Grice, Relationshipbetween the blue-green 5 alga Trichodesmiumthiebautii and the copepodMacrosetella gracilis in the plankton off northeasternSouth America, Ecology,47, 855-

I I I I I I I I I 856, 1966. 10 20 30 40 50 60 70 80 90 1 O0 Capone,D. G., and E. J. Carpenter,Nitrogen fixationin the marine environment, Science, 217, 1140-1142, 1982. CumulativeJuly-December alongshore Carder, K. L., and R. G. Steward,A remote-sensingreflectance model stocks(105 cellsI'4) of Gymnodiniumbreve of a red-tide dinoflagellateoff west Florida, Limnol. Oceanogr.,30, Figure10. TheSaharan dust (/•g m -3) sampledat Barbados 286-298, 1985. during July 1965-1996 in relation to cumulativestocks of G. Carder, K. L., W. W. Gregg, D. K. Costello, K. Haddad, and J. M. Prospero,Determination of Saharandust radianceand chlorophyll breve(105 cells L -1) foundalong the central west Florida coast from CZCS imagery,J. Geophys.Res., 96, 5369-5378, 1991. during July-Decemberof the sameyears. Carlsson,P., and E. Graneli, Availability of humic bound nitrogenfor coastal phytoplankton,Estuarine Coastal Shelf Sci., 36, 433-447, 1993. Carpenter,E. J., Nitrogenfixation by marineOsciilatoria (Trichodes- Graneli,1993], yielding a similar15N/14N ratio for their red mium) in the world'soceans, in Nitrogenin theMarine Environment, tides. Moreover, within other subtropicalhabitats, pristine editedby E. J. Carpenterand D. G. Capone,pp. 65-103, Academic, bedsof seagrasses and mangroveshave 815PON of +1.6 to San Diego, Calif., 1983. Carpenter, E. J., and T. Roenneberg,The marine planktonic cya- +2.5%0 [O'Donohueet al., 1998;Cotbert et al., 1999] derived nobacterium Trichodesmiumspp.: Photosyntheticrate measure- from nitrogen fixation by symbiontbacteria in sedimentssuch mentsin the SW Atlantic Ocean,Mar. Ecol. Prog.Ser., 118, 267-273, that these macrophytesmay provide yet another source of 1995. labile DON for G. breve of intermediate 815PON. Carpenter, E. J., and K. Romans, Major role of the cyanobacterium Trichodesmiumin nutrient cycling in the North Atlantic Ocean, We are now couplingmore complexversions of (1) to phys- Science, 254, 1356-1358, 1991. ical circulation models at 1-10 km resolution of the West Carpenter,D. J., M. I. Scranton,P. Novelli, and A. Michaels,Validity Florida shelfwith PrincetonOcean Model (POM) to explore of N 2 measurementsin marine Oscillatoria(Trichodesmium), J. the seasonalonshore transport of both T. erythraeumand G. Plankton Res., 9, 1047-1056, 1987. breve[Walsh et al., 2001], as well as their retention and fate Checkley,D. M., and L. C. Entzeroth, Elemental and isotopicfractionation of carbon and nitrogen by marine, planktonic copepods alongthe coastduring upwelling and downwellingevents. Sim- and implicationsto the marine nitrogencycle, J. PlanktonRes., 7, ilarly,the bioticdetails [Penta, 2000] of the toxicconsequences, 553-568, 1985. of the bio-opticalsignals, and of the biogeochemicalimplica- Chew,F., On the offshorecirculation and a convergencemechanism in tionsof sucha coastalfood web responseto iron alleviationof the red tide regionoff the westcoast of Florida, Trans.Eos AGU, 35, 963-974, 1955. nitrogenlimitation are now the respectivesubjects of our Na- Chew, F., A tentativemethod for the predictionof the Florida red tide tional Oceanic and AtmosphericAdministration (NOAA)- outbreaks, Bull. Mar. Sci., 6, 292-304, 1956. EnvironmentalPrediction Agency (EPA) ECOHAB:Florida, Cline, J. D., and I. R. Kaplan, Isotopicfractionation of dissolved Officeof Naval Research(ONR) HyperspectralCoastal Ocean nitrate duringdenitrification in the easterntropical North Pacific Dynamics Experiment (HyCODE), and NASA Synthesis, Ocean, Mar. Chem., 3, 271-299, 1975. Codispoti,L. A., Phosphorusversus nitrogen limitation of new and Modeling, and Prediction(SMP) field projectson the West export production, in Productivityof the Ocean: Past and Present, Florida shelf. editedby W. H. Berger,V. S. Smetacek,and G. Wefer, pp. 377-394, Wiley-Interscience,New York, 1989. Corbett, D. R., J. Chanton, W. Burnett, K. Dillon, and C. Rutkowski, Acknowledgments. This analysiswas funded by grantsNAG5-6449 Patternsof groundwaterdischarge into Florida Bay,Limnol. OCean- from NASA, NA76RG0463 and NA96OP0084 from NOAA, ogr., 44, 1045-1055, 1999. N000149615024 and N000149910212 from ONR to JJW, and R Darwin, C., The Voyageof the Beagle,pp. 1-369, Cambridge Univ. 827085-01-0from the EPA to KAS. The state of Florida supported Press, New York, 1839. JJW andKAS aswell. JoeProspero supplied the unpublisheddust data Davis, C. C., Gymnodiniumbrevis sp. nov., a causeof discoloredwater of Figure7, Bev Robertsmade most of the cell countsin Figure3, and and animal mortality in the Gulf of Mexico, Bot. Gaz. (Chicago), CarmTomas provided the nutrientdata of Figure4. We alsothank our 109, 358-360, 1948. colleaguesin the ECOHAB:Florida,HyCODE, and SMP projectsfor deBaar,H. J., A. G. Buma,R. F. Nolting, G. C. Cadee,G. Jacques,and manyideas, other data sets,and discussions.In particular,Paul Bissett P. Treguer,On iron limitationof the SouthernOcean: Experimental and Gabe Vargo made insightfulcomments on the original manu- observationsin the Weddell and ScotiaSea, Mar. Biol. Prog.Ser., 65, script,as did Joe Prospero.This is ECOHAB contribution13. 105-122, 1990. 11,610 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

Dortch, Q., and T. E. Whitledge, Does nitrogenor siliconlimit phy- 1972-74, NOAA Tech.Rep. NMFS SSRF-701, pp. 1-18, Natl. Oce- toplanktonproduction in the MississippiRiver plume and nearby anic and Atmos.Admin., Silver Spring,Md., 1976. regions?,Cont. Shelf Res., 12, 1293-1309,1992. Hulburt, E. M., Phytoplanktonobservations in the westernCaribbean Dragovich,A., and J. A. Kelly, Distributionand occurrenceof Gym- Sea, Bull. Mar. Sci., 18, 388-399, 1968. nodinium breve on the west coast of Florida, 1964-65, U.S. Fish Ingle, R. M., and D. F. Martin, Predictionof the Florida red tide by Wildl. Serv.Spec. Sci. Rep. Fish., 541, 1-15, 1966. means of the iron index: Red tide, iron and humic acid levels in Dragovich,A., J. H. Funicane, and B. Z. May, Counts of red tide steams,Environ. Lett., 1, 69-74, 1971. organisms,Gymnodinium breve, and associatedoceanographic data Joint,I., J. Lewis,J. Aiken, R. Proctor,G. Moore, W. Higman, and M. from Florida west coast, 1957-59, U.S. Fish Wildl. Serv.Spec. Sci. Donald. Interannualvariability of PSP outbreakson the north east Rep. Fish., 369, 1-175, 1961. UK coast,J. Plankton Res., 19, 937-956, 1997. Dragovich,A., J. H. Funicane,J. A. Kelly, and B. Z. May, Countsof Karl, D. M., R. Letelier, D. V. Hebel, D. F. Bird, and C. D. Winn, red tide organisms,Gymnodinium breve, and associatedoceano- Trichodesmiumblooms and new nitrogenin the North PacificGyre, graphicdata from Floridawest coast, 1960-61, U.S.Fish Wildl.Serv. in Marine Pelagic Cyanobacteria:Trichodesmium and Other Dia- Spec.Sci. Rep. Fish., 455, 1-39, 1963. zotrophs,edited by E. J. Carpenter,D. G. Capone,and J. G. Reuter, Dragovich,A., J. A. Kelly, and H. G. Goodell, Hydrologicaland pp. 219-238, Kluwer Acad., Norwell, Mass., 1992. biologicalcharacteristics of Florida's west coast tributaries,Fish. Khromov, N. S., Distribution of plankton in the Gulf of Mexico and Bull., 66, 463-477, 1968. someaspects of its seasonaldynamics, in Soviet-CubanFishery Re- Duce, R. A., and N. W. Tindale, Atmospherictransport of iron and its search,edited by A. S. Bogdanov,pp. 36-56, Isr. Prog.Sci. Trans., depositionin the ocean,Limnol. Oceanogr.,36, 1715-1726,1991. Jerusalem, 1969. Duce, R. A., et al., The atmosphericinput of tracespecies to the ocean, Kierstead, H., and L. B. Slobodkin, The size of water massescontain- Global.Biogeochem. Cycles, 5, 193-259, 1991. ing planktonblooms, J. Mar. Res.,12, 141-147, 1953. Dunstan,W. M., and J. Hosford, The distributionof planktonicblue- Kim, Y. S., and D. F. Martin, Interrelationshipof PeaceRiver param- greenalgae related to the hydrographyof the GeorgianBight, Bull. eters as a basisof the iron index: A predictiveguide to the Florida Mar. Sci., 27, 624-629, 1977. red tide, Water Res., 8, 607-616, 1974. Feinstein,A., Correlationsof various ambient phenomenawith red King, J. E., A preliminaryreport on the planktonof the west coastof tide outbreaks on the West Florida shelf, Bull. Mar. Sci., 6, 209-232, Florida, Q. J. Fla. Acad. Sci., 12, 109-137, 1950. 1956. Kishi, M., and S. Ikeda, Populationdynamics of "red tide" organisms Franks,P. J., Spatialpatterns in densealgal blooms, Limnol. Ocean- in eutrophicatedcoastal waters: Numerical experiment of phyto- ogr.,42, 1297-1305, 1997. planktonbloom in the east Seto Inland Sea, Jpn.Ecol. Model.,31, Franks, P. J., and C. Chen, Plankton production in tidal fronts: A 145-174, 1986. model of GeorgesBank in summer,J. Mar. Res.,54, 631-651, 1996. Kleppel, G. S., C. A. Burkart, K. Carter, and C. Tomas, Diets of Fraser, T. H., and W. H. Wilcox, Enrichmentof a subtropicalestuary calanoidcopepods on the West Florida continentalshelf: Relation- with nitrogen,phosphorus, and silica, in Estuariesand Nutrients, shipsbetween food concentration,food compositionand feeding activity,Mar. Biol., 127, 209-218, 1996. edited by B. J. Watson and L. E. Cronin, pp. 481-498, Humana, Clifton, N.J., 1981. Landing, W. M., J. J. Perry, J. L. Guentzel, G. A. Gill, and C. D. Pollman,Relationships between the atmosphericdeposition of trace Freeberg,L. R., A. Marshall,and M. Heyl, Interrelationshipsof Gym- elements,major ions, and mercuryin Florida: The FAMS project nodiniumbreve (Florida red tide) within the phytoplanktoncommu- (1992-1993),Water Air SoilPollut., 80, 343-352, 1995. nity, in ToxicDinofiagellate Blooms, edited by B. Taylor and H. H. Lasker, R., and F. G. Smith, Red tide, in Gulf of Mexico:Its Origin, Seliger,pp. 139-144, ElsevierSci., New York, 1979. Waters,and Marine Life, Fish. Bull., 89, 173-176, 1954. Gardner, W. S., R. Benner, R. M. Amon, J. B. Cotner, J. F. Cavaletto, Lenes, J. M., B. P. Darrow, C. Cattrail, C. Heil, G. A. Vargo, M. and J. R. Johnson,Effects of high-molecularweight dissolvedor- Callahan,R. H. Byrne, J. M. Prospero,D. E. Bates,K. A. Fanning, ganicmatter on nitrogendynamics in the MississippiRiver plume, and J. J. Walsh, Iron fertilization and the Trichodesmiumresponse Mar. Ecol. Prog.Ser., 133, 287-297, 1996. on the West Florida shelf,Limnol. Oceanogr.,in press,2001. Geesey,M., and P. A. Tester, Gymnodiniumbreve: Ubiquitous in Gulf Leonard, C. L., C. R. McClain, R. Murtugudde,E. E. Hofmann, and of Mexico waters?,in ToxicPhytoplankton Blooms in the Sea, edited L. W. Harding,An iron-basedecosystem model of the centralequa- by T. J. Smaydaand Y. Shimizu,pp. 251-255, ElsevierSci., New torial Pacific,J. Geophys.Res., 104, 1325-1341, 1999. York, 1993. Longhurst,A., EcologicalGeography of the Sea,pp. 27-29, Academic, Gilbes,F., C. Tomas,J. J. Walsh, and F. E. Muller-Karger,An episodic San Diego, Calif., 1998. chlorophyllplume on the west Florida shelf, Cont. ShelfRes., 16, Mariotti, A., C. Lancelot,and G. Billen, Natural isotopiccompositions 1201-1224, 1996. of nitrogenas a tracer of origin of suspendedorganic matter in the Glibert, P.M., and D. A. Bronk, Releaseof dissolvedorganic nitrogen Scheldtestuary, Geochim. Cosmochim. Acta, 48, 549-555, 1984. by marine diazotrophiccyanobacteria, Trichodesmium spp., Appl. Martin, J. H., R. M. Gordon, S. E. Fitzwater, and W. W. Broenkow, Environ. Microbiol., 60, 3996-4000, 1994. VERTEX: Phytoplankton/ironstudies in the Gulf of Alaska,Deep Gunther, G., R. H. Williams, C. C. Davis, and F. G. Smith, Cata- Sea Res., Part A, 36, 649-680, 1989. strophicmass mortality of marine animalsand coincidentphyto- Masserini,R. T., and K. A. Fanning,A sensorpackage for the simul- plankton bloom on the west coast of Florida, November 1946 to taneous determination of nanomolar concentrations of nitrite, ni- August1947, Ecol. Monogr.,18, 309-324, 1948. trate, and ammonia in seawater by fluorescencedetection, Mar. Guo, C., and P. A. Tester, Toxic effect of the bloom-formingTri- Chem., 68, 323-333, 2000. chodesmiumsp. (Cyanophyta)to the copepodAcartia tonsa,Nat. McCarthy, J. J., and E. J. Carpenter, Oscillatoria(Trichodesmium) Toxins, 2, 222-227, 1994. thiebautii(Cyanophyta) in the centralNorth Atlantic Ocean,J. Phy- Haddad, K. D., and K. L. Carder, Ocean intrusion: One possible col., 15, 75-82, 1979. initiation mechanismof red tide blooms on the west coastof Florida, McPherson,B. F., R. T. Montgomery,and E. E. Emmons,Phytoplank- in ToxicDinofiagellate Blooms, edited by D. L. Taylor and H. H. ton productivityand biomassin the Charlotte Harbor estuarine Seliger,pp. 269-274, ElsevierSci., New York, 1979. system,Florida, WaterRes. Bull., 26, 787-800, 1990. Hallegraeff, G. M., M. A. McCausland,and R. K. Brown,Early warn- Moore, R. M., J. E. Milley, and A. Chatt, The potentialfor biological ing of toxic dinoflagellateblooms of Gymnodiniumcatenatum in mobilization of trace elements from aeolian dust in the ocean and its southern Tasmanian waters, J. Plankton Res., ! 7, 1163-1176, 1995. importancein the caseof iron, Oceanol.Acta, 7, 221-228, 1984. Holligan, P.M., M. Viollier, C. Dupouy, and J. Aiken, Satellitestudies Murphy, E. B., K. A. Steidinger,B. S. Roberts,J. Williams, and J. W. on the distributionsof chlorophylland dinoflagellateblooms in the Jolley,An explanationfor the Florida eastcoast Gymnodinium breve westernEnglish Channel, Cont. ShelfRes., 2, 81-96, 1983. red tide of November,1972, Limnol. Oceanogr.,20, 481-486, 1975. Horrigan, S. G., J.P. Montoya, J. L. Nevins, and J. J. McCarthy, Nelson,D. M., and Q. Dortch, Silicicacid depletionand siliconlimi- Natural isotopiccomposition of dissolvedinorganic nitrogen in the tation in the plume of the MississippiRiver: Evidencefrom kinetic ChesapeakeBay, EstuarineCoastal Shelf Sci., 30, 393-410, 1990. studiesin springand summer,Mar. Ecol. Prog.Ser., 136, 163-178, Houde, E. D., and N. Chitty, Seasonalabundance and distributionof 1996. zooplankton,fish eggs, and fishlarvae in the easternGulf of Mexico, Nelson, J. R., J. E. Eckman, C. Y. Robertson, R. L. Marinelli, and WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES 11,611

R. A. Jahnke,Benthic microalgalbiomass and irradiance at the sea Smayda,T. S., Harmful algalblooms: Their ecophysiologyand general floor on the continentalshelf of the South Atlantic Bight: Spatial relevanceto phytoplanktonblooms in the sea, Limnol. Oceanogr., and temporal variability and storm effects, Cont. Shelf Res., 19, 42, 1137-1153, 1997. 477-506, 1999. Sokolik,I. N., and O. B. Toon, Direct radiative forcingby anthropo- O'Donohue, M. J., C. A. Heil, S. Lowe, J. Horrocks, E.G. Abal, S. genic airborne mineral aerosols,Nature, 381,681-683, 1996. Costanzo,and W. C. Dennison, Assessingthe impact of a flood Steidinger, K. A., Phytoplankton ecology: A conceptual overview event on Moreton Bay usingmarine plants as bioindicatorsof water based on eastern Gulf of Mexico research, CRC Rev. Microbiol., 3, quality,in MoretonBay and Catchment,edited by I. R. Tibbetts,N.J. 49-68, 1973. Hall, and W. C. Dennison,pp. 585-596, Univ. of Queensland,Bris- Steidinger,K. A., and K. Haddad, Biologicand hydrographicaspects bane, Aust., 1998. of red-tides, BioScience,31, 814-819, 1981. O'Neil, J. M., P.M. Metzler,and P.M. Glibert,Ingestion of lSN2- Steidinger,K. A., and J. Williams, Dinoflagellates.Memoirs of the 1abelledTrichodesmium spp. and ammoniumregeneration by the HourglassCruises, Tech. Ser. Fla. Dep. Nat. Resour.Mar. Res.Lab., harpacticoidcopepod Macrosatella gracilis, Mar. Biol., 125, 89-96, 2, 1-251, 1970. 1996. Steidinger,K. A., J. T. Davis, and J. Williams,Dinoflagellate studies on Paerl, H. W., Coastal eutrophicationand harmful algal blooms:Im- the inshore waters of the west coast of Florida, in Red Tide Studies, portance of atmosphericdeposition and groundwateras "new" ni- Pinellas to Collier Counties,1963-66, Fla. Board of Conserv.Prof. trogen and other nutrient sources,Limnol. Oceanogr.,42, 1154- Pap. 9, pp. 4-47, Tallahassee,1967. 1165, 1997. Steidinger,K. A., G. A. Vargo, P. A. Tester, and C. R. Tomas, Bloom Paerl, H. W., L. E. Prufert-Bebout, and C. Guo, Iron-stimulated N 2 dynamicsand physiologyof Gymnodiniumbreve with emphasison fixationand growthin naturaland cultured populations of the plank- the Gulf of Mexico, in PhysiologicalEcology of Harmful Algal tonic marine cyanobacteriumTrichodesmium, Appl. Environ. Micro- Blooms, edited by D. M. Anderson,A.D. Cembella, and G. M. biol., 60, 1044-1047, 1994. Hallegraeff, pp. 135-153, Springer-Verlag,New York, 1998. Paluskiewicz,T., L. P. Atkinson, E. S. Posmentier, and C. R. McClain, Steven, D. M., and R. Glombitza, Oscillatoryvariation of a phyto- Observationsof a Loop Current frontal eddy intrusion onto the plankton populationin a tropical ocean,Nature, 237, 105-107, 1972. West Florida shelf,J. Geophys.Res., 88, 9639-9651, 1983. Steven,D. M., A. L. Brooks, and E. A. Moore, Primary and secondary Penta, B., Phytoplanktoncompetition in the West Florida shelf: A productionin the tropical Atlantic, Prog. Rep. ONR NOOO14-67-A- simulationanalysis with "red tide" implications,Ph.D. dissertation, 0432-0001,pp. 1-124, Bermuda Biol. Stn., Bermuda, 1970. pp. 1-259, Univ. of South Fla., Tampa, 2000. Sunda,W. G., and S. A. Huntsman,Interrelated influenceof iron, light Pomeroy,L. R., H. H. Haskin, and R. A. Rogotskie, Observationsof and cell sizeon marine phytoplanktongrowth, Nature, 390, 389-392, dinoflagellateblooms, Limnol. Oceanogr.,1, 54-60, 1956. 1997. Prospero,J. M., Long-rangetransport of mineral dust in the global Sutton, T., T. Hopkins, A. Remsen, and S. Burghart, Multisensor atmosphere:Impact of African dust on the environment of the samplingof pelagicecosystem variables in a coastalenvironment to southeastern United States, Proc. Natl. Acad. Sci., 96, 3396-3403, estimatezooplankton grazing impact, Cont. ShelfRes., 21, 69-87, 1999a. 2001. Prospero,J. M., Long-term measurementsof the transport of African Swap, R., S. Ulanski, M. Cobbett, and M. Garstang,Temporal and mineral dust to the southeasternUnited States:Implications for spatial characteristicsof Saharan dust outbreaks,J. Geophys.Res., regionalair quality,J. Geophys.Res., 104, 15,917-15,927,1999b. 101, 4205-4220, 1996. Prospero,J. M., and R. T. Nees, Impact of the North African drought Taylor, F. J., and U. Pollingher,Ecology of dinoflagellates,in The and E1 Nifio on mineral dust in the Barbados trade winds, Nature, Biologyof Dinofiagellates,edited by F. J. Taylor, pp. 398-529, Black- 320, 735-738, 1986. well Sci., Malden, Mass., 1987. Prospero,J. M., R. T. Nees, and M. Uematsu, Deposition rate of Tester, P. A., and P. K. Fowler, Brevotoxin contamination of Merce- particulate and dissolvedaluminum derived from Saharan dust in naria mercenariaand Crassostreavirginica, a managementproblem, precipitationat Miami, Florida, J. Geophys.Res., 92, 14,723-14,731, in ToxicMarine Phytoplankton,edited by E. Graneli et al., pp. 499- 1987. 503, Elsevier Sci., New York, 1990. Prospero,J. M., I. Olmez, and M. Ames,A1 and Fe in PM 2.5 and PM Tester, P. A., and K. A. Steidinger, Gymnodinium breve red tide 10 suspendedparticles in south-centralFlorida: The impactof long- blooms:Initiation, transport, and consequencesof surfacecircula- range transport of African mineral dust, WaterAir Soil Pollut., in tion, Limnol. Oceanogr.,42, 1039-1051, 1997. press,2001. Tester, P. A., R. P. Stumpf, F. M. Vukovich, P. K. Fowler, and J. T. Roman, M. R., Ingestionof the blue-greenalga Trichodesmiumby the Turner, An expatriatered tide bloom: Transport, distribution,and harpacticoidcopepod, Macrosatella gracilis, Limnol. Oceanogr.,23, persistence,Limnol. Oceanogr.,6, 1053-1061, 1991. 1245-1248, 1978. Tett, P., The ecophysiologyof exceptionalblooms, Rapp. P. V. Reun. Rounsefell, G. A., and A. Dragovich, Correlation between oceano- Cons.Int. Explor. Mer., 187, 47-60, 1987. graphicfactors and abundanceof the Florida red tide (Gymno- Timmermans, K. R., W. Stolte, and H. J. DeBarr, Iron-mediated diniumbreve Davis), 1954-1961, Bull. Mar. Sci., 16, 402-422, 1966. effects on nitrate reductasein marine phytoplankton,Mar. Biol., Rueter, J. G., Iron stimulationof photosynthesisand nitrogenfixation 121, 389-396, 1994. inAnabena7120 and Trichodesmium(Cyanophyceae), J. Phycol.,24, Turner, J. T., and P. A. Tester, Toxic marine phytoplankton,zooplank- 249-254, 1988. ton grazers, and pelagic food webs, Limnol. Oceanogr.,42, 1203- Rueter, J. G., K. Ohki, and Y. Fujita, The effect of iron nutrition on 1214, 1997. photosynthesisand nitrogen fixation in culturesof Trichodesmium Tyrell, T., The relative influencesof nitrogen and phosphoruson (Cyanophyceae),J. Phycol.,26, 30-35, 1990. oceanicprimary production,Nature, 400, 525-531, 1999. Rueter, J. G., D. A. Hutchins, R. W. Smith, and N. L. Unsworth, Iron Vargo, G. A., and D. Howard-Shamblott,Phosphorus requirements in nutrition of Trichodesmium,in Marine Pelagic Cyanobacteria:Tri- Ptychodiscusbrevis: Cell phosphorus,uptake and growth require- chodesmiumand OtherDiazotrophs, edited by E. J. Carpenter,D. G. ments,in ToxicMarine Phytoplankton,edited by E. Graneli et al., pp. Capone, and J. G. Reuter, pp. 289-306, Kluwer Acad., Norwell, 324-329, Elsevier Sci., New York, 1990. Mass., 1992. Vargo, G. A., and E. Shanley,Alkaline phosphataseactivity in the Saunders,R. P., and D. A. Glenn, Diatoms:Memoirs of the Hourglass red-tide dinoflagellate,Ptychodiscus brevis, Mar. Ecol., 6, 251-264, Cruises,Tech. Ser. Fla. Dep. Nat. Resour.Mar. Res. Lab., 1, 1-119, 1985. 1969. Villareal, T. A., and E. J. Carpenter, Diel buoyancyregulation in the Shanley,E., and G. A. Vargo, Cellular composition,growth, photo- marine diazotrophiccyanobacterium Trichodesmium thiebautii, Lim- synthesis,and respirationrates of Gymnodiniumbreve under varying nol. Oceanogr.,35, 1832-1837, 1990. light levels,in ToxicPhytoplankton Blooms in the Sea, edited by T. J. Vukovich, F. M., and P.M. Hamilton, New atlas of front locations in Smaydaand Y. Shimizu,pp. 831-836, ElsevierSci., New York, 1993. the Gulf of Mexico, paper presentedat 10th Annual Gulf of Mexico Shimizu, Y., and G. Wrensford, Peculiaritiesin the biosynthesisof Information Transfer Meeting, U.S. Miner. Manage. Serv.,Metairie, brevitoxinsand metabolismof Gymnodiniumbreve, in ToxicPhyto- La., 1989. planktonBlooms in the Sea, edited by T. J. Smaydaand Y. Shimizu, Walsh,J. J., Particleexport at Cape Hatteras,Deep SeaRes., Part II, 41, pp. 919-923, Elsevier Sci., New York, 1993. 603-628, 1994. 11,612 WALSH AND STEIDINGER: DUST, CYANOPHYTES, AND RED TIDES

Walsh, J. J., Nitrogen-fixationwithin a tropical upwellingecosystem: the Sea, editedby T. J. Smaydaand Y. Shimizu,pp. 73-78, Elsevier Evidence for a Redfield budget of carbon/nitrogencycling by the Sci., New York, 1993. total phytoplankton community,J. Geophys.Res., 101, 20,607- Yanagi, T., T. Yamamoto, Y. Koizumi, T. Ikeda, M. Kamizono,and H. 20,616, 1996. Tamori, A numericalsimulation of red tide formation,J. Mar. Syst., Walsh, J. J., G. T. Rowe, R. L. Iversen, and C. P. McRoy, Biological 6, 269-285, 1995. export of shelf carbonis a neglectedsink of the global CO2 cycle, Yang, H., and R. H. Weisberg, Responseof the West Florida shelf Nature, 291, 196-201, 1981. circulationto climatologicalwind stress,J. Geophys.Res., 104, 5301- Walsh, J. J., D. A. Dieterie, M. B. Meyers, and F. E. Muller-Karger, 5320, 1999. Nitrogen exchangeat the continentalmargin: A numericalstudy of Yang, H., R. H. Weisberg,P. Niiler, and A. Sturges,Lagrangian cir- the Gulf of Mexico, Prog. Oceanogr.,23, 245-301, 1989. culation and forbidden zone on the West Florida shelf, Cont. Shelf Walsh, J. J., et al., A numerical simulationof carbon/nitrogencycling Res., 19, 1221-1245, 1999. during springupwelling in the Cariaco Basin,J. Geophys.Res., 104, Yentsch,C. M., C. S. Yentsch,and J.P. Perras,Alkaline phosphatase 7807-7825, 1999. activityin the tropical marine blue-greenalga Oscillatoriaerythaea Walsh, J. J., K. D. Haddad, D. A. Dieterie, R. H. Weisberg,Z. Li, H. ("Trichodesmium"),Limnol. Oceanogr.,17, 772-774• 1972. Yang, F. E. Muller-Karger, C. A. Heil, and W. P. Bissett,A numer- Zhu, X. R., J. M. Prospero,and F. J. Millero, Diel variabilityof soluble ical analysisof the landfall of 1979 red tide of Gymnodiniumbreve Fe(II) and solubletotal Fe in North African dustin the tradewinds along the west coast of Florida: The fall/winter upwelling mode, at Barbados,J. Geophys.Res., 102, 21,297-21,305, 1997. Cont. ShelfRes., in press,2001. Zhuang, G., R. A. Duce, and D. R. Kester, The dissolutionof atmo- West, T. L., H. G. Marshall,and P. A. Tester,Natural phytoplankton sphericiron in surfaceseawater of the open ocean,J. Geophys.Res., communityresponses to a bloom of the toxicdinoflagellate Gymno- 95, 16,207-16,216, 1990. dinium breveDavis off the North Carolina coast, Castenea,61,356- 368, 1996. K. A. Steidinger,Florida Marine Research Institute, Florida Fish White, A. W., Relationships of environmental factors to toxic and Wildlife ConservationCommission, 100 Eighth Avenue S.E., St. dinoflagellateblooms in the Bay of Fundy,Rapp. P. V. Reun. Cons. Petersburg,FL 33701. Int. ExplorMer, 187, 38-46, 1987. J. J. Walsh, Collegeof Marine Science,University of SouthFlorida, Wilson, W. B., The suitabilityof sea-waterfor the survivaland growth 140 Seventh Avenue South, St. Petersburg, FL 33701. of Gymnodiniumbreve Davis; and someeffects of phosphorusand ([email protected]. usf. edu) nitrogenon its growth,FSU Prof.Pap. Ser.,7, pp. 1-42, Fla. Board of Conserv., Tallahassee, 1966. Wyatt, T., and F. Saborido-Rey,Biogeography and time-seriesanalysis (ReceivedNovember 1, 1999;revised October 31, 2000; of BritishPSP records,1968-1990, in ToxicPhytoplankton Blooms in acceptedNovember 7, 2000.)