Gymnodinium breve red tide blooms: Initiation, transport, and consequences of surface circulation

Abstract

From its source waters in the Gulf of Mexico the led tide dinoflagellate, Gymnodinium breve is moved throughout its oceanic range by major currents and eddy systems. The continental shelf off the west coast of Florida experiences frequent G. breve blooms (in 21 of the last 22 years) where the spatially explicit phases of G. breve blooms are closely coupled to physical processes. Bloom initiation occurs offshore and in association with shoreward movements of the Loop Current or spinoff eddies. A midshelf front maintained by seasonal wind reversals along the Florida west coast may serve as a growth and accumulation region for G. breve blooms and contribute to the reinoculation of nearshore waters. Local eddy circulation in the northeastern Gulf of Mexico and in the Dry Tortugas affects the retention and coastal distribution of blooms while the Florida Current and Gulf Stream transport cells out of the Gulf of Mexico and into the U.S. South Atlantic Eight. The causes of bloom dissipation are not well known but mixing or disruption of the water mass supporting G. bp-eve cells, especially in combination with declining water temperatures, are important factors.

Full text

Gymnodinium breve red tide blooms: Initiation, transport, and consequences
circulation of surface
Patricia A. Tester
National Marine Fisheries Service, NOAA, Southeast Fisheries Science Center, Beaufort Laboratory, Beaufort, North
Carolina 28516-9722
Karen A. Steidinger
Florida Department of Environmental Protection, Florida Marine Research Institute, 100 Eighth Ave., SE, St. Petersburg,
Florida 33712
Abstract
From
its source
waters
in the Gulf of Mexico the red tide dinoflagellate,
Gymnodinium breve
is moved throughout
its oceanic range by major currents and eddy systems. The continental shelf off the west coast of Florida experiences
frequent G. breve blooms (in 21 of the last 22 years) where the spatially explicit phases of G.
breve
blooms are
closely coupled to physical processes. Bloom initiation occurs offshore and in association with shoreward move-
mcnts of the Loop Current or spinoff cddics. A midshelf front maintained by seasonal wind reversals along the
Florida west coast may serve as a growth and accumulation region for G.
breve
blooms and contribute to the
reinoculation of nearshore waters. Local eddy circulation in the northeastern Gulf of Mexico and in the Dry Tortugas
affects the retention and coastal distribution of blooms while the Florida
Current
and Gulf Stream transport cells
out of the Gulf of Mexico and
into
the US. South Atlantic Bight. The causes of bloom dissipation are not well
known but mixing or disruption of the water mass supporting G. breve cells, especially in combination with declining
water temperatures, are important factors.
Much of what is known about the distribution of the toxic
dinoflagellate Gymnodinium breve Davis [ = Ptychodiscus
brevis (Davis) Steidinger] is explained by oceanic circulation
patterns. G. breve cells are positively phototactic (or nega-
tively geotactic) (Steidinger 1975; Heil 1986) and can con-
centrate in the upper water column during the day. There
they behave like surface drifters, only smaller writ. The res-
ident population is in the Gulf of Mexico and G. breve is
transported throughout its range by the Gulf Loop Current,
the Florida Current, and the Gulf Stream, with the warmth
of the Gulf Stream fostering this subtropical species as it
moves north of 31”N. G. breve has been recorded throughout
the U.S. South Atlantic Bight (Tester et al. 1993) and be-
yond.
G. breve rarely occurs in shelf waters north of Cape Hat-
teras, North Carolina (Marshall 1982; see also Churchill and
Cornillon 1991), but the Gulf Stream may carry it farther
(see Fraga and Sanchez 1985). A drift bottle released in an
October 1966 study off the central west Florida shelf was
recovered in the Outer Hebrides a little over a year later;
another released in February 1967 reached Belgium in 187
d (averaging >50 km d-l) (Williams et al. 1977). So it seems
likely that long before
A.D.
1497, when Vespucci, credited
Acknowledgments
We thank S. Baig, who supplied the Gulf Stream Frontal Anal-
yses (GOES) and is a continuing source of information and support.
T. Lee is acknowledged for his ideas on the entrainment of G.
breve
cells in the Tortugas Gyrc. J. Turner, G. Cervetto, M. Geesey, C.
Lewis, and A. Smith helped draft figures. E. Haugen and three re-
viewers made valuable comments on the manuscript. B. Roberts
maintains and shared the G. breve bloom log. R Stegmann and K.
Carder provided CZCS imagery. The source of the AVHRR image-
ry
is B. Stone, NESDIS and the NOAA Coastwatch Program.
with being the first European to explore the Gulf of Mexico,
sailed out of the gulf in the direction of the “maestrale” for
870 leagues to Cape Hatteras and then turned eastward to-
ward Bermuda before returning to Spain (Galtsoff 1954), G.
breve had made a similar voyage.
The reports of discolored water and effects of phycotoxins
in tropical Atlantic waters were recognized and recorded in
ships’ logs by 1530-1550 (Martyr 1912). As early as 1844
popular accounts of G. breve blooms were linked with nox-
ious “gases” and massive fish kills along the west coast of
Florida (Feinstein et al. 1955). Prophetically, in that same
year an intensive study OF ocean currents was begun by Mau-
ry (1859). Data on ocean circulation, physical processes, and
the distribution and biology of G. breve would accumulate
for more than a century before this information would be
coupled, the consequences appreciated, and a model concep-
tualized to help focus research efforts,
Range
Although Lackey (1956) reported G. breve from Trinidad
in the southern Caribbean basin (cell counts were not veri-
fied), there were no recorded cases of neurotoxic shellfish
poisoning reported (S. Hall pers. comm.), and there have
been no subsequent observations of G. breve anywhere in
the Caribbean. Early interest in the circulation of the Carib-
bean, though, stemmed from its role as the source region for
water flowing into the Gulf of Mexico. In this semi-enclosed
basin the eastern gulf is characterized by anticyclonic cir-
culation and is dominated by two currents. The Yucatan Cur-
rent (75 cm s-l), entering between the Yucatan Peninsula
and Cuba, becomes the Loop Current as it extends northward
into the gulf and returns southward along the west Florida
1039

1040 Tester and Steidinger
ATLANTIC
GULF OF MEXICO
Fig. I. Generalized surface circulation of the Gulf of Mexico. Arrows denote steadiness of
current drift. (Redrawn from UNEP/CEPAL Caribbean Environment Program, Project 1037.)
continental shelf. It exits the gulf between the Dry Tortugas
and Cuba where it is known as the Florida Current (165 cm
s-l) (see Hofmann and Worley 1986) (Fig. 1). The extent of
northward penetration of the Loop Current (Maul 1977;
Vukovich et al. 1979), its spinoff eddies (Dietrich and Lin
1994; Sturges 1994), and its intrusions onto the west Florida
continental shelf vary seasonally (Huh et al. 198 1) and great-
ly affect the potential of bloom initiation, transport, and re-
tention (Haddad and Carder 1979; Haddad 1982; Lee et al.
1994). Rotating eddies can be shed from the Loop Current
and propagate westward across the gulf (Liepper et al. 1972;
Maul and Vukovich 1993). Cross-basin surface transport
also has been documented in drift-bottle studies (Williams
et al. 1977). Summer-fall bottle releases from the west Flor-
ida shelf frequently were recovered from Texas beaches
(Matagorda to Brownsville) to Vera Cruz, Mexico. The gen-
eral circulation patterns of the south central and western Gulf
of Mexico are clockwise but the velocities are less intense
(15-25 cm s I). The only exceptions to this general clock-
wise pattern are the cyclonic flows in the extreme north-
eastern and northwestern areas of the gulf (Molinari 1980;
Vastano et al. 1995)-regions where elevated background
concentrations of G. breve cells (>lOO cells liter ‘) have
been noted (Geesey and Tester 1993).
Throughout the Gulf of Mexico and the U.S. South At-
lantic Bight, G. breve is found in background concentrations
(l-1,000 cells liter- ‘) except in areas off the Texas coast and
the west Florida coast where local circulation may play a
role (Fig. 2; see Geesey and Tester 1993; Tester et al. 1993).
Although G. breve blooms have occurred in many different
areas in the Gulf of Mexico, from Yucatan in the south (Gra-
ham 1954), to the lower Laguna Madre, in the Mexican state
of Tamaulipas in the western gulf (Gunter 1952; Wilson and
Ray 1956) to Freeport, Texas (Burr 1945; Trebatoski 1988),
and around the northern gulf coast, they are most frequent
along the west coast of Florida. Blooms there are especially
frequent from Clearwater to Sanibel Island (Joyce and Rob-
erts 1975; K. A. Steidinger and B. S. Roberts unpubl.), oc-
curring in 21 01: the last 22 years. These blooms on the
southwest Florida shelf serve as a source for cells inoculat-
ing the U.S. South Atlantic Bight (Murphy et al. 1975; Stei-
dinger et al. 1995; Tester et al. 1991).
Bloom initiation
The regions of the GulF of Mexico that experience blooms
of G. breve lasting more than 2 months include the west
Florida shelf (C.earwater to Sanibel Island), the Campeche
Bay between Rio Ciatzacoalcos and Rio Grijalva (Smithson.
Inst. 1971), and the Texas coast between Port Arthur and
Galveston Bay. 411 have common features conducive to the
formation of blooms. Each of these areas is adjacent to a
continental shelf break where it intersects with the perma-
nent seasonal thl:rmocline (Fig. 3). The isothermal water on
the outer shelf results in a minimum bottom temperature of
20°C in areas off Texas and west Florida (NOAA 1985) and
Campeche Banks. Consequently these areas may provide an
important winter refuge for G. breve. These same areas also
experience either persistent, intermittent, or event-related
slope-shelf upwelling. This is best known for the west Flor-
ida shelf where blooms can occur any time of the year but
are typical in late summer and fall when >70% of the out-
breaks have begun. Bloom concentrations first appear OFF-
shore (Dragovich and Kelly 1966; Steidinger 1975; Steidin-
ger and Haddad 1981) and are associated with the fronts
caused by the onshore-offshore meanders of the Loop Cur-
rent water along the outer southwest Florida shelt Water on

G.
breve
blooms
1041
+
Fig. 2. Gymnodinium breve nonbloom, background concentrations in cells liter-‘. Samples were
taken during 1989-1991 by ships of opportunity (after Gcesey and Tester 1993; Tester et al. 1993).
Fig. 3. Winter bottom-water temperatures on the continental shelf of the Gulf of Mexico basin
(redrawn from NOAA 1985). Regions of persistent upwelling off the southwest Florida shelf, Texas-
Louisiana coast, and the Yucatan Peninsula are indicated by the arrows.

1042
Tester and Steidingrr
Fig. 4. CZCS image. 14 November 1978, providing an estimate of chlorophyll a in &.g liter
(inset) during a Gymnodinium
hrevc
bloom off the west Florida coast from soulh of Tampa Bay to
the Florida Keys. When processed ar a higher
I-csolution,
chlorophyll a is detectable at G.
hrwe
cell concenlraliuns that are not viqihlc to the human eye (i.e. Stump Pass, top arrow. 6.7X IO’ cells
hoer ‘; Doctor-Gordon Passes, hottom anow, 4.6X10‘ cells
liter-‘).
the leading edge of an eddy or meander is generally sinking
and that on the trailing edge is rising, until the feature ex-
periences bottom drag (Dietrich and Lin 1994) and forms a
midshelf front.
Fronts represent a dynamic area of nutrient regimes and
light conditions which can favor accumulation and growth
(<l div. d ‘, often 0.2-0.5 div. d-‘) of dinoflagellates.
Bloom species such as G.
breve
and
Gyrodinium
cf. aura-
lum (European waters) are well adapted to such environ-
ments and can grow throughout the euphotic zone. They
have a high photosynthetic capacity at low light and are
light adapted at varying intensities (Shanley 1985; Garcia
and Purdie 1992) although photoinhibition thresholds are
species-specific. Once growth occurs, it takes 2-X weeks to
develop into a hloom of fish-killing proportions (1-2.5X IOx
cells liter-‘) depending on physical, chemical, and biological
conditions.
Some species such as G.
breve, G.
cf. aureolum, and
Lin-
gulodinium plydrum
have growth and competitive exclu-
sion strategies that can lead to almost monospecific surface
blooms of these bpecieb (as biomass) (Sttidinga and Vargo
1988;
Morin et al. 1989). Such blooms (G.
breve, G.
cf.
aureolum) can cclvei- a surface area of up to 1.4-3.0x10’
km’ (Steidingcr and Joyce 1973; Holligan 1985; Vargo et al
1987) and although biomass concentration is patchy, chlo~
rophyll a values from >2 to > 100 mg tr’ make the rest&
tant discolored surface water detectable by color sensors 011~
board satellites (Fig. 4). In the case of G.
breve
the CZCS
sensor detected chlorophyll a from cells at densities one-
two orders of magnitude less (IO”-1 Oi) than are present when
discolored water is detectable by the human eye (>lO”) (K.
Haddad and K. Steidinger per% comm.).
Both G.
breve
and G. cf. auwulum discolor surface waters
and are phototactic; in daylight hours cells are at or near the
surface; at night they are dispersed (Hollignn 1985; Heil
1986; Geesey and Tester 1993). In addition to their ability
to exploit light regimes, both species have advantages in
nutrient dynamics. Both assimilate nitrogen at low light and
are able to utilize organic as well as inorganic nutrients (Var-
go and Shanley 1985; Steidinger and Vargo 1988; Dahl and
Tangen 1993; Shimizu et al 1995). When G.
breve
blooms
have been tracked, the zone of initiation (cell numbers
>I,000 liter ‘) develops from 18 to 74 km offshore (St&
dinger and Haddad 1981) and the strongest evidence for this
lies in the cell distribution along cross-shelf transects sam-

G. breve blooms 1043
Cells ‘I Id liter”
a > 400
igj 200 - 400
[7 >5-200
0 -=5
Fig. 5. Spatial and temporal distribution of Gymnodinium breve cell abundance along the west
Florida coast during a late summer-fall 1976 bloom.
pled by personnel from the Florida Marine Research Institute
(FMRI) and Mote Marine Laboratory.
The first example is taken from transects sampled during
a late summer bloom in 1976. A series of stations from the
inlets or nearshore to 32-70 km offshore was made between
0.3
Distance from Shore (km)
Fig. 6. Distribution and abundance of Gymnodinium hew cells
along cross-shelf transects north of Boca Grande Pass to >66 km
offshore taken bctwcen 24 September and 11 October 1976.
22 September and 15 October. Concentrations above back-
ground were first noted offshore between Sarasota and Boca
Grande Pass in late September (Fig. 5). During the following
week in the same area, cell concentrations increased and then
generally spread south. No nearshore or inlet samples during
this time were either positive for G. breve or above back-
ground until 2 weeks later. Data from north of Boca Grande
Pass are representative of the cell distributions from transects
run perpendicular to shore (Fig. 6). Typically the cells are
moved from the midshelf to onshore and then, under the
influence of the wind and(or) alongshore currents, move up
or down the coast (Fig 5). From 7 to 11 October there were
still high numbers (>l-6X 10s cells liter -I) of G. breve 3-
25 km offshore north of Boca Grande Pass. The bloom had
moved south and was parallel to the coast with lo-IOO-fold
lower cell concentrations onshore than offshore. By 13-14
October the bloom was centered south of Boca Grande Pass
and had intensified. Counts offshore of Sanibel Island were
1-1.5X
10” cells liter-‘; counts nearshore ranged frown 1.5 to
70X lo1 cells liter-‘. Farther south at Naples there were no
cells in the 25&m station, but nearer shore cells were noted.
After 15 October the inshore passes between Clearwater and
Naples had < 1,000 cells liter -I and sampling was suspend-
ed.
A similar pattern of bloom development and movement is
evident from blooms in November-December 1979 and Sep-

I 044
Tester and Steidinger
Cells I 10’ liter“
5 350
fl 200 - 350
q
>5-200
l <5
Fig. 7. Spatial and temporal distribution
of’ Gymnodinium breve
cell abundance along the weat
Florida coast during late fall 1979.
tember-November 1985 (FMRI data). The 1979 bloom was
first reported 130 km northwest of Clearwater and the ear-
liest cell counts from the Clearwater and Cedar Key areas
(26-30 November) indicated that it was a large bloom (Fig.
7). Note however, there were no cells or very low counts
(O-2,000 cells liter-‘) onshore until I O-l I December, when
offshore numbers were dropping and the bloom seemed to
have moved south loo-120 km and onshore. The third ex-
ample is from a 1985 bloom which was first sampled 16 km
off Cedar Key and southward (101-10s cells liter -I> on 10
September (Fig. 8). Again the passes and nearshore were
free of G.
breve
cells until more than a month later. The
alongshore (south) and onshore movement of the bloom is
evident; the offshore stations between Clearwater and Sar-
asota were cell-free during the first sampling period, but l-
2 weeks later there were up to lo”-10“ cells liter-‘. Within
4-5 weeks between lo5 and 10” liter-’ were observed on-
shore and south of Sarasota and Boca Grande Pass (4 No-
vember).
Bloom transport
There is evidence that some blooms can be maintained
within the midshelf zone and continually inoculate the near-
shore waters or recur in a “high occurrence zone” from
Clearwater to Sanibel Island (Steidinger and Roberts un-
publ.). One possible mechanism for this is the circulation
pattern reported by Weisberg et al. (1996). They describe
seasonal wind reversal, (northeast-southwest flow) on the
midshelf that result in zero mean Rows both in the along-
shore and crossshore directions during a 16-month period.
However, the mcnthly means can be re.latively large, and
Weisberg et al. suggested the maximum values have a bar-
oclinic origin via a thermal wind relationship. Recent re-
search on the west Florida shelf circulation describes two
basic patterns, a summer pattern (April-September) and a
winter pattern (October-March) characterized by a semiper-
manent anticyclojiic eddy (Weisberg et al. 1996) on the
northwestern Florida shelf in the Apalachee Bay-Middle
Grounds area (H. Yang and R. Weisberg :pers. comm.) This
feature dominates the northeastern shelf onshore of the 50-m
isobath (Yang and Weisberg pers. comm) and may be re-
sponsible for the ~entrainment and transport of cells north-
ward to the Florida panhandle. Haddad ( 1982) recorded an
example of a red tide preceded by the shoreward advection
of a bottom thermocline between 10 August and 3 Septem-
ber 1978 when the bloom surfaced at 19 km from shore.
Similar transport along a thermal gradienit may explain the
spring-summer bloom of 1995 when a G.
breve
red tide
apparently moved north up to Cedar Key and then onto the
Florida panhandle 55-l. 10 km offshore and subsequently in-
oculated inshore waters (FMRI unpubl. data).
From Tampa Bay south small-scale eddy features (<lo0
km; Maul 1977; Hela et al. 1955) or filaments (R. Stumpf
pers. comm.) may also play a role in the translocation of
offshore blooms. Frontal eddies (Loop Current water) and

G. breve
blooms
1045
ca Grande Pass
anibel Island
Fig. 8. Spatial and temporal distribution of
Gymnodinium breve
cell abundance along the west
Florida coast during a late summer-fall bloom in 1985.
on-offshore meanders of the Loop Current move southward
along the outer southwest Florida shelf (every 2-14 d) (Pal-
uszkiewicz et al. 1983), and there is evidence for a south-
ward mean flow over the shelf when the Loop Current is at
the shelf edge (Sturges and Evans 1983). The annual cycle
of wind stress, northward during summer and southward in
fall, is responsible for the persistent upwelling (summer) or
downwelling (fall) found over the west Florida shelf (Lee
and Williams 1988) and may concentrate or disperse blooms
depending on the site and timing of the bloom.
Another recently described eddy system operative be-
tween the Tortugas and the upper keys is dependent on a
well-developed Loop Current and the consequent offshore
position of the Florida Current. The Tortugas Gyre is a cy-
clonic recirculating feature (100-l 80 km) with a duration of
40-J 08 d that has a strong influence on the transport and
retention of zooplankton and larval fish in the lower Florida
Keys (Lee et al. 1994). In the intervals between gyre recir-
culation periods there are episodes (20-30 d) of intense east-
ward flow (Lee et al. 1994). This feature may provide insight
into the distribution and transport of G.
breve
From the west
coast of Florida to the Atlantic in 1994-1995 (T. Lee pers.
comm.). The bloom started in September 1994, between
Tampa Bay and Sanibel Island off the west coast of Florida.
In January and February 1995, extensive fish kills were re-
ported in >5,000 km2 of open water westward from the Flor-
ida mainland to the Dry Tortugas; a sample with 1.5X lo5
cells liter-’ was counted from Sanibel on 1X January. In
February, coincident with fish kills off the southwest Florida
coast, a sample from 12 km off Duck Key (14 February
1995) contained 9.6X 10” cells liter-’ (Fig. 9). Concern about
the possibility of G.
breve
transport to the Atlantic prompted
sampling off West Palm Beach from September 1994
through early March 1995. Cell counts at West Palm Beach
were only O-6 cells liter-’ from September to December
1994 but from 17 to 24 February 1995, as the Gulf Stream
impinged on the West Pplm Beach sampling site and formed
a meander, G.
breve
cell densities increased to -2X lo4 cells
liter-’ (Fig. 10) (Steidinger et al. 1995). As the Gulf Stream
moved offshore, away From the sampling area, cell numbers
decreased (27 February-3 March). From 3 to 7 March 1995,
samples from the Occulina Reef National Park (-27”53’N,
79”58’W) off Cape Canaveral, north of the West Palm Beach
site contained O-80 cells liter-.‘, but the Gulf Stream did not
impinge on the reef during the 4-d sampling period. Fourteen
days after the Gulf Stream meander migrated past West Palm
Beach, G.
breve
cells were found in 10 of 12 samples from
the outer shelf of Onslow Bay, North Carolina (33”51’N,
76”53’W, 23.4’), >l,OOO km downstream. Although the cell
counts were low (55 cells liter .I) the presence of G.
breve
in >80% of the samples is unusual in early March. Fortu-
nately, the dire oceanic menace (as depicted by the
Miami
Herald
on 22 February 1995) following the course of the
Gulf Stream was exaggerated. In early March the inner shelf
water of Onslow Bay is too cold (<12”C) to support the
growth of G.
breve
and wind mixing prevented water-col-
umn stabilization conducive to bloom development.
Prior to 1969, G.
breve
had not been recorded in the U.S.

1046
Tester and Srcidiqer
South Atlanuc Bighl, but Lackey (I 969) found one cell in
an unpreserved sample taken from Boa Katon during a wa-
ter-quality study. Nol until autumns 1972, 1976, and 1Y77
after the typical manifestations of G. hrrve blooms (e.g. eye
and respiratory irritation, fish mortality) were described by
beachgoers from Miami to Palm Beach (Murphy et al. 1915;
Robens 1979), did we undcrsland that blooms--even shorl-
lived ones (t I month)-could occur outside the Gulf of
Mexico basin. These first blooms recorded in the U.S. South
Allantic Bight were restricted in area and intcnsily. No cells
were found a\ far north as Cape Canaveral. and the count’:
wcrc neither hrgh nor persistent (2-100X 10’ cells liter ‘).
The eveas were conxdered the result of concentration and
transport of cells to the east coast by unusual Loop Current
parterns (Murphy et al. 1975).
The seasonality of the west Florida shelf circulation and
wind liclds alho contributes to the likelihood of shelf water
being advected into the Florida Current for transport to the
U.S. South Atlantic Bight (Williams et al. 1977). Late sum-
mcr-aulumn blooms have the greatest potential for transport
to the Atlantic coast bccausc summer transport rates are the
highal and “detrainment” is greatest in Pall due to low,
inconsistent trampor, (Maul and Bravo I9XY). The thrw jr-
cas of dcuainment identified by Maul nnd Bravo (19X9) ah
likely was for receiving flotsam and .jetsnm are southeast
of Cape Canaveral, eat of St. Augu\tinc, and southeast of
Onslow Bay. G. brew was TV prove itself as an apt surface-
drifter and a good test of their ideas.
Cells from a May-October 1980 war Florida shelf bloom
were transpor~cd farther than any recorded up to thai lime.
Gulf Stream Frontal Analyses (compobite GOES satellite im-
agwy) confirmed a warn-water intrusion 7-10 November
off Jacksonville (Fig. I I) and by 14 Novcmbcr local Jack-
sonville residents were suffering sore throats and watery
eyes; by 2528 November G. brew counts were 6.7X IO‘
cells lite: and the bloom had spread >I00 km south LO
Daytona where beachgoers were affected by exposure to the
surf (FMKI data). Cells were found off Cape Canaveral
(Melbourne) on 5 Decemhcr and between Jacksonville and
Cape Canaveral mcandcrs of warm water shoreward of the
western edge of the Gulf Stream were evident in GOES in-
agery of 8 December 1980 (not shown).
The 1980 east coast G.
brew
red tide dcmonaated that
Gulf Stream meanders off Jacksonville could inoculate in-

G. breve blooms 1047
14 15 16 17 23 24 25 2 3 6
Feb 95 March 95
Fig.
10. Gymnodinium breve
cell counts near West Palm Beach, FL in relation to the position of the Gulf Stream. When the western
front of the Gulf Stream was closer to the ncarshore sampling site the cell counts were I-18.8X lo3 cells liter .I. When the Gulf Stream
was seaward of the sampling site G.
breve
cell concentrations were <IO0 cells
liter
~I (after Steidinger et al. 1995). Cape Canaveral is the
cape immediately north of West Palm Beach (0).
shore areas
and be transported by countercurrents south
along the coast beyond Cape Canaveral. In 1983, another
transport occurred concurrently as a meander of the Gulf
Stream Formed south of Jacksonville on 10 October (S. Baig
pers. comm.). In the area from Daytona to below Cape Ca-
naveral (Volusia and Brevard Counties) all three signs of a
G. breve red tide were evident (i.e. human respiratory irri-
tation, fish kills, and discolored water). The cell counts were
much higher than for previous east Florida shelf blooms
(5.5X106 cells liter I, 10 October). The following day cell
counts were 1X 10’ cells liter ‘; during the next 10 d, red
tide affected areas immediately south of Cape Canaveral
(Patrick Air Force Base, Cocoa Beach, and Melbourne). By
28 October, it had moved 55 km south of Cape Canaveral
(Sebastian Inlet) and remained there until 4 November. By
23 November, the red tide had dissipated and shellfish har-
vesting areas were opened.
Prior to that east coast event, a well-developed red tide
was detected on the west coast on 6 October. Patches of dead
fish and surface-water discoloration (indicative of cell counts
23 X 10” cells liter ‘) were reported from Sarasota to Venice
from shore to 15 km offshore. Inshore G. breve concentra-
tions were 4.5 X 10” cells liter-’ and beachgoers were expe-
riencing respiratory irritation. Cell counts from 7 October
were 3X 10” cells liter I 9 km off Captiva Island in the Char-
lotte Harbor area. This is the classic source area for transport
of red tides from the west coast to the east coast of Florida.
Because it takes 2-8 weeks for a red tide to develop con-
centrations of +2.5X lo5 cells liter ’ offshore, this west Flor-
ida bloom was the likely source for cells inoculating the
Jacksonville area.
Gulf Stream transport also was implicated in an unusual
G. breve fall and winter (19X7-1988) bloom along the coast
of North Carolina which continued for 4-5 months (Tester
et al. 1991). Thirty days before this bloom there was a late
summer bloom off the southwest coast of Florida. The cou-
pling of these two events is supported by transport time of
an Argos-tracked surface drifter (60 km d I) making the
same passage in -20 d (Ortner et al. 1995) and the drift
bottles recovered from Wrightsville Beach, North Carolina,
between 3 1 and 100 d after their release off the west Florida
shelf (Williams et al. 1977). The continental shelf between
Cape Hatteras and Cape Lookout where this bloom occurred
is the narrowest of any in the U.S. South Atlantic Bight north
of the Miami area and is frequently overwashed by meanders
of Gulf Stream water, some of which nearly reach the barrier

1048 Tester and Steidinger
Fig. Il. Shoreward intrusions of Gulf Stream water onto the continental shelf off Jacksonville on 5
(solid line) and 10
(dashed line) November 1980. Note the filaments of mater stranded shoreward of the
intrusions. medrawn from Gulf Stream Frontal Analyses (GOES sea surface temperahue)
S. Baig, NOAA.]
Fig. 12. Association of Gymnodiuium breve cells with Gulf
Stream water in Onslow Bay between Cape Lookout and Cape Fear
during winter. ln a February 1991 cross-shelf transect of nine sta-
tions,
G. breve cells were found only at the two stations most close-
ly associated with the Gulf Stream meander. Water temperatures in
the Gulf Stream meander were -22°C; temperatures of the shelf
water did not exceed 14°C.
islands (Bane et #iI. 198 1; Yoder et al. 1985). These meanders
serve as nutrient pumps introducing new nitrogen from strata
beneath the Gulf’ Stream (Lee et al. 1991). After a meander
passes, parts of Ihe filament may remain on the shelf for as
long as a week before dispersing or rejoining the stream
(FRED Group 1989). The longevity (>I9 d) of the Gulf
Stream filament stranded on the continental shelf off North
Carolina in late fall 1987 was credited with sustaining this
unique G. breve bloom (Tester et al. 1991).
Dissipation
Significant qu#:stions remained in the aftermath of the
North Carolina bloom that caused the closure of major shell-
fish harvesting areas for an entire season and had an esti-
mated cost of $25 million (Tester and Fowler 1990). Samples
from ships of’ opr’ortunity were used to determine nonbloom,
background levels of G. breve for the northern and eastern
Gulf of Mexico, fhe Florida Strait, and the entire U.S. South
Atlantic Bight in:luding the Gulf Stream and western Sar-
gasso Sea (Fig. :!). G. breve has a continuous distribution
throughout this r;mge but in winter its occurrence in near-
shore waters of the U.S. South Atlantic Bight is closely as-
sociated with GLtlf Stream meanders overriding the shelf
(Tester et al. 1993). Perhaps the best example of this depen-
dency is from a cross-shelf sampling transect in Onslow Bay
(between Cape L’3okout and Cape Fear) in February 1991.
This transect bisected a Gulf Stream meander and G. breve
cells were found only in the meander and at its shoreward
edge (Fig. 12). Field studies of Rounsefell and Nelson

1049
q
5,000 - SOD00 CELLS LITER-’
n
>50.000 CELLS LITER-’
G. breve
blooms
t
0 NO CELLS •n <5,000 CELLS LITER-’
5 140T
L
z =
g c
0:
e
2
I
HATTERAS
-2
2
CAPE LOOKOUT
-2
2
CAPE FEAR
-2
7 4
3NOV I DEC IJAN I FEE 22 FEB
1987 1989
Fig. 13. Above-distribution and concentrations of Gymnodinium breve cells along the coast
of North Carolina between 2 November 1987 and 22 February 1988. Below-wind stress (dyne
cm-*) measured at the Cape Hatteras, Cape Lookout, and Cape Fear C-MAN stations (after Tester
et al. 1991).
(1966) indicated
G.
breve thrives
between 16 and 27°C
with
only a few cells surviving 7-9.9”C (Aldrich 1959). Theo-
retically the lower thermal tolerance of this species is a ma-
jor factor restricting bloom formation in the U.S. South At-
lantic Bight. There is no permanent thermocline near the
continental shelf edge and in winter nearshore bottom-water
temperatures may be <9”C when inundated by Virginia
coastal water (-5°C) (Stefansson et al. 1971).
Dissipation or termination of blooms can occur when the
offshore bloom component is entrained and transported out
of the area or when the integrity of the water mass is weak-
ened by mixing and dilution. Both declining water temper-
ature and increasing wind stress contributed to the dissipa-
tion of the 1987-1988 G.
breve
bloom off North-South
Carolina. From early November to late December 1987 most
of the cells from this bloom moved back and forth within
Onslow Bay depending on wind direction (Fig. 13). Alter
nearshore water temperatures dropped (11.O”C) and the first
winter storm (31 December 1987) was followed by inter-
mittent but strong southward winds (1-16 January 1988),
cell numbers decreased and the bloom was blown out of
Onslow Bay, around Cape Fear to the south. Samples from
5 to 30 km off Little River Inlet to Myrtle Beach, South
Carolina, from 6 to 7 February 1988 contained from 2-
11
X
10’ cells liter I, The last samples from nearshore waters
of North Carolina containing G.
breve
were collected on 11
February, and the bloom dissipated off Myrtle Beach by late
February 1988 (Tester et al. 1991).
Conclusions
G.
breve
blooms are neritic phenomena. They have a suc-
cessful survival strategy in the Gulf of Mexico ecosystem
where they can be found from the central basin to nearshore
waters (salinity >24%0). The bloom model that is most con-
sistent with observations made during the last 80-100 yr
starts with an offshore bloom initiation in late summer or
fall in conjunction with a Gulf Loop intrusion on the outer
continental shelf. Following cross-shelf transport, largely in-
fluenced by winds and wind-induced upwelling or down-
welling, cells concentrate and grow at a region approximat-
ing the midshelf front. If cross-shelf transport mechanisms

1050
Tester and Steidinger
continue to operate on the bloom, cells concentrate in near-
shore waters where movement is governed by winds and
alongshore currents.
G.
breve
is found in low concentrations (< 1,000 cells
liter -I) in the Florida Current-Gulf Stream throughout the
year. During blooms cell concentrations of >2X lo4 cells li-
ter become entrained in the Florida Current and transported
to the Atlantic. Gulf Stream meanders are known to deliver
the cells into nearshore waters. Small-scale eddy features
like the Tortugas Gyre are very productive regions for phy-
toplankton and may prove to be so for G.
breve as
well.
Strengthened transport through the Florida Strait at the dis-
sipation of the Tortugas Gyre suggests an explanation for the
infrequent, high cell counts in the Atlantic.
This conceptual model provides a framework to help de-
fine spatial and temporal scales, processes, timing, and link-
ages vital to hypothesis testing. It is essential that our un-
derstanding of’ the prerequisites for each bloom phase (e.g.
initiation, growth, maintenance, and dissipation) be as com-
plete as possible. We know little about factors that affect the
intensity OF G.
breve
blooms. It is critical that surface cir-
culation be coupled with a depth component and the model
expanded to three dimensions. It is equally important to un-
derstand what factors contribute to the reintroduction of cells
into coastal waters-a major concern during prolonged
blooms. Increased access to archived wind and temperature
data combined with cell counts from “bloom logs” will aid
in retrospective analyses of blooms in greater detail. The
advent of online, real-time environmental data and the pros-
pect of new ocean color sensors may allow enough predic-
tive capability that bloom conditions can be detected early
enough to allow for focused research efforts and provide
reliable information to the public.
References
AI.INCH, D. V. 1959. Physiological studies of red tide, p. 69-71.
In Galveston Biological Laboratory fishery research for the
year ending June 30, 1959. U.S. Fish Wildl. Serv. Circ. 62.
BANE, J. M., Ja., D. A. BKOOKS,
AND
K. R. LORENSON. 1981.
Synoptic observations of the three-dimensional structure and
propagation of Gulf Stream meanders along the Carolina con-
tinental margin. J. Geophys. Res. 86: 641 l-6425.
BURR, J. G. 1945. Science tackles a mystery. Texas Game Fish 3:
4-5.
CHUIICHII.L, J. H., AND P C. CORNILLON.
1991. Water discharged
from
the Gulf Stream north of Cape Hatteras. J. Geophys. Rcs.
96: 22,227-22,243.
DAHL, E.,
AND
K. TANGEN. 1993. 25 years experience with Gym-
nodinium aureolum in Norwegian waters, p. 15-21. In Toxic
phytoplankton blooms in the sea. Proc. 5th Int. Conf. on Toxic
marine phytoplankton. Elscvier.
DI~TRICII, D. E.,
AND
C. A. LIN. 1994. Numerical studies of eddy
shedding in the Gulf of Mexico. J. Geophys. Res. 99: 7599-
7615.
DKAC~OVICH, A., AND J. A. KELLY, JR. 1966. Distribution and oc-
currence of Gymnodinium breve on the west coast of Florida
1964-65. U.S. Fish Wild]. Serv. Rep. 54 I.
FEINSTEIN, A., A. R. CEURVELS, R. E HUTI‘ON, AND E. SNOEK.
1955. Red tide outbreaks off the Florida west coast. Univ.
Miami Mar. Lab. Conserv. Rep. 55-15.
FRED GKOUP. 1989. Frontal eddy dynamics (FRED) experiment off
North Carolina. OCS Study MMS Tech. Rep. 89-0028. U.S.
Dep. Inter.
FKAGA, S., AND
Ii
J. SANCMEZ. 1985.
Toxic and potentially toxic
dinoflagcllatc,s found in Galacian Rias (NW Spain), p. 51-54.
In Toxic dinollagellatcs: Proc. 3rd Int. Conf. Elsevier.
GAI.TSOFF, P S. .954. Historical sketch of the explorations in the
Gulf of Mex,co. Fish. Bull. 55: 3-98.
GARCIA, V. M. T, AND D. A. PURIXE. 1992. The influence of ir-
radiance on growth, photosynthesis and respiration of Gyro-
dinium cf. ahreolum. J. Plankton Res. 14: 1251-1256.
GIZSEY, M., AND l? A. TESTER. 1993.
Gymnodinium breve:
Ubiq-
uitous in Gulf of Mexico waters?, p. 251-255. in Toxic phy-
toplankton blooms in the sea. Proc. 5th Int. Conf. on Toxic
Marine Phytclplankton. Elsevier.
GRAHAM, H. W. 1954. Dinoflagellates of the Gulf of Mexico. Fish.
Bull. 55: 223-226.
GUNTER, G. 1952. The importance of catastrophic mortalities fool
marine fishcries along the Texas coast. J. Wild. Manage. 16:
63-69.
HADIIAD, K. D. ,982. Hydrographic factors associated with west
Florida toxic red tide blooms: An assessment for satellite pre-
diction and monitoring. M.S. thesis, Univ. South Florida. 161 p.
-,
AND
K. L CARDER. 1979. Oceanic intrusion: One possible
initiation mechanism of red tide blooms on the west coast of
Florida, p. 269-274. In Toxic dinoflagcllate blooms: Proc. 2nd
Int. Conf. Elsevier.
HEII., C. A. 1986. Vertical migration of Ptychodiscus hrevis (Davis)
Steidinger. M.S. thesis, Univ. South Florida. I IX p.
HEI.A, I., D. DHSYI.VA, AND C. A. CARPENTIX. 1955. Drift currents
in the red tidl: area OF the eastern most region of the Gulf of
Mexico. Univ. Miami Mar. Lab. Rep. Fla. State Bd. Conserv.
55-11.
HOFMANN, E. E., P.ND S. J. WORIXY. 1986. An investigation of the
circulation of the Gulf of Mexico. J. Geophys. Res. 91:
14,221-14,23(5.
HOI.I.I(;AN, I? M. 1985. Marine dinoflagellale blooms: Growth strat-
egies and environmental exploitation, p. 133-139. In
Toxic di-
noflagellates: Proc. 3rd Int. Conf. Elsevier.
HUII, 0. K., W. J. WIESMAN, JR., AND L. J. ROUSE, JR. 1981. In-
trusion of Lo>p Current waters onto the west Florida conti-
nental shelf. J. Geophys. Res. 86: 4 186-4 192.
JOYCE, E. A.,
ANC~
B. S. RORERTS. 1975. Florida Department of
Natural Resources red tide research program, p. 95-103. In
Proc. 1st Int. Conf. on Toxic Dinoflagellate Blooms. Mass. Sci.
Tcchnol. Found.
LACKEY, J. B. 1956. Known geographic range of Gymnodinium
hrevis Davis. Q. J. Fla. Acad. Sci. 19: 71.
-. 1969. Microbial studies in the FWPCA project area with
comparisons t#, other subtropical and tropical areas, p. 52-58.
in Dcmonstralion of the limitations and effects of waste dis-
posal on an ocean shelf. Fla. Ocean Sci. Inst. Rep. AR-69-2.
LEE:, T N., M. E. CI,ARKE, E. WILI.IAMS, A. E SZMANT, ANU T
BHIIOER. 1994. Evolution of the Tortugas gyre and its influ-
ence on recru tment in the Florida Keys. Bull. Mar. Sci. 54:
62 l-646.
-,
AND
E.
M’IIJJAMS.
1988. Wind-forced transport fluctua-
tions of the Florida Current. J. Phys. Oceanogr. 18: 937-946.
-,
J. A.
YOIER, AND L. l? ATKINSON.
1991. Gulf Stream
frontal eddy influence on productivity of the southeast U.S.
continental shelf. J. Geophys. Res. 96: 22,191-22,205.
LIE~JP~ZR, D. E, J.
COCIIRANE, AND
J. E HEWITT. 1972. A detached
eddy and subsequent changes (1965), p. 107-I 18. In L. Ca-
purr0 and J. Reid [eds.], Contributions to the physical ocean-
ography of the Gulf of Mexico. Gulf.
MARSI~ALL, H. G. 1982. The composition of phytoplankton within

G. breve blooms 1051
the Chesapeake Bay plume and adjacent waters off the Virginia
coast, U.S.A. Estuarine Coastal Shelf Sci. 15: 29-43.
MAIUY
R, I? 19 I 2. De orbo novo, the eight decades of Peter Martyr.
[E A. MacNutt transl. V. 21 Putnam.
MAUL,
G. A. 1977. The annual cycle of the Gulf Loop Part 1:
Observations during a onc-year time series. .I. Mar. Res. 35:
219-247.
, AND N. J. BRAVO. 1989. Fate of satellite-tracked buoys
and drift cards off the southeastern Atlantic coast of the United
States. Fla. Sci. 52: 154-170.
AND
E M. VUKOVICH. 1993. The relationship between
va;iations in the Gulf of Mexico Loop Current and Straits of
Florida volume transport. J. Phys. Oceanogr. 23: 785-796.
MAURY, M. E 1859. The physical geography of the sea. Harper.
MOLINARI, R. 1980. Current variability and its relation to sea-sur-
fact topography in the Caribbean Sea and Gulf of Mexico.
Mar. Geod. 3: 409-436.
MORIN, P J., J. L. BIRKDIEN,
AND
P LECORRE. 1989. The frontal
systems in the Iroisc Sea: Development of Gyrodinium aureo-
Zum Hulburt on the inner front, p. 215-221. In J. D. Ross led.],
Topics in marine biology. Sci. Mar. 53.
MURPI~Y, E. B., K. A. STEIDJN~~ER, B. S. ROBERTS,
.I. WIILIA~S,
AND
J. W. JOI,LEY. 1975. An explanation for the Florida east
coast Gymnodinium breve red tide of November 1972. Limnol.
Occanogr. 20: 481-486.
153-162. In Proc. 1st Int. Conf. on Toxic Dinoflagellate
Blooms. Mass. Sci. Technol. Found.
AND
K. HADI~AD. 1981. Biological and hydrographic as-
pe& of red tides. Bioscience 31: 814-819.
AND
E. A.
JOYCE, JR.
1973. Florida red tides. Fla. Dep.
Nit. Resour. Mar. Res. Lab. Educ. Ser. 17.
-, B. S. ROBERTS,
AND
I? A. T~!sT~R. 1995. Florida red tides.
Harmful Algae News 12/13: 1-3.
,
AND
G. A. VAR(;O. 1988. Marine dinoflagellate blooms:
Dynamics and impacts, p. 373-401.
In
C. A. Lembi and J. R.
Waaland teds.], Algae and human affairs. Cambridge.
STIJRGES, W. 1994. The licquency of ring separations from the
Loop Current. J. Phys. Oceanogr. 24: 1647-1651.
ANT)
J. C. EVANS. 1983. On the variability of the Loop
Current in the Gulf of Mexico. J. Mar. Res. 41: 639-653.
TESTI;R, P. A.,
AND
P K. FOWLI!R. 1990. Brevetoxin contamination
of Mercenaria mercenaria and Crassostrea virginica: A man-
agement issue, p. 499-503. In Toxic marine phytoplankton:
Proc. 4th Int. Conf. Elsevicr.
NOAA. 1985. Gulf of Mexico coastal and ocean zones strategic
assessment: Data atlas. U.S. Dep. Commerce.
ORTNER, P
B.,
AND OTIIEKS.
1995. Mississippi River flood waters
that reached the Gulf Stream. J. Geophys. Res. 100: 13,595-
13,601.
PAIXZKIEWICZ, T., L. I? ATKINSON, E. S. PC)SMENTIER, AND C. R.
MCCLAIN. 1983. Observations of a Loop Current frontal eddy
intrusion onto the west Florida shelf. J. Geophys. Rcs. 88:
9639-9652.
- M. E. Gu~sev,
AND
E M. Vu~ovrcrr. 1993. Gymnodinium
b&e and global warming: What arc the possibilities?, p. 76-
72. In Toxic phytoplankton blooms in the sea. Proc. 5th Int.
Conf. on Toxic Marine Phytoplankton. Elsevicr.
-,
R.
l? STUMPI’,
E M. VUKOVICH, I? K. FOWIXR,
AND
J. T
TUIINER. 1991. An expatriate red tide bloom: Transport, dis-
tribution, and persistence. Limnol. Oceanogr. 36: 1053-l 06 I.
TRI;BA~‘OSKI, B. 1988. Observations on the 1986-87 Texas red tide
(Ptychodiscus
brevis). Texas Water Comm. Rep. 88-02.
VARCO, G. A., K. L. CARI~RR, W.
GREGG, E. SIIANLI’.Y, AND C.
HEII~. 1987. The potential contribution of primary production
by red tides to the west Florida shelf ecosystem. Limnol.
Oceanogr. 32: 762-767.
,
AND
E. SHANLEY. 1985. Alkaline phosphatase activity in
the red tide dinoflagellate Ptychodiscus brevis. Mar. Ecol. 6:
25 I-264.
ROBERTS,
B. S. 1979. Occurrence of Gymnodinium breve red tides
along the west and east coasts of Florida during 1976 and 1977,
p. 199-202. In Toxic dinoflagellatc blooms: Proc. 2nd Int.
Conf, Elsevier.
VASTANO, A. C., C. N. BARRON,
JR., AND
E. W.
SHAAR, JR. 1995.
Satellite observations of the Texas Current. Cont. Shelf Res.
15: 729-754.
ROUNSE~~I,, G. A.,
AND
W. R. NELSON. 1966. Red-tide rcscarch
summarized to 1964 including an annotated bibliography. U.S.
Fish Wildl. Serv. Spec. Sci. Rep. 535.
SIIANI.EY, E. 1985. Photoadaption in the red tide dinoflagcllatc Pry-
chodiscus brevis. MS. thesis. Univ. South Florida. 122 p.
SHIMIZLJ, Y., N. WATANABJ!,
AND
G. WRENSFORD. 1995. Biosyn-
thesis of brevetoxins and heterotrophic metabolism in Gym-
nodinium breve, p. 351-375. In Harmful marine algal blooms.
Proc. 6th Int. Conf. on Toxic Marine Phytoplankton. Lavoisier.
SMITHSONIAN JNSTITU~ION. 197 1, Shortlived phenomena 1970 an-
nual report.
VUKOVICH, E M., B. M. CRI~SSMAN, M. BUSIINHLI.,
AND W. J. KING.
1979. Some aspects of the oceanography of the Gulf of Mex-
ico using satellite and in situ data. J. Geophys. Rcs. 84: 7749-
7768.
WEISHERG, R. H., B. D. BLACK,
AND
HUI.IUN-YANG. 1996. Sea-
sonal modulation of the west Florida continental shelf circu-
lation. Gcophys. Res. Lctt. 23: 2247-2250.
Wu~.rnlvls, J., W. E GREY, E. B. MurtrHY, AND J. J. CRANII. 1977.
Drift analyses of eastern Gulf of Mexico surface circulation.
Mem. Hourglass Cruises 4: 1-134.
WILSON,
W. B.,
ANI)
S. M.
RAY.
1956. The occurrence of Gym-
nodinium brevis in the western Gulf of Mexico. Ecology 37:
388.
STEFANSSON,
U., L. I? ATKINSON,
AND
D. G. BUMP~JS. 1971. Hy-
drographic properties and circulation of the North Carolina YODER, J. A., ANI) OTHERS. 1985. Phytoplankton dynamics within
Gulf Stream intrusions on the southeastern United States con-
shelf and slope waters. Deep-Sea Res. 18: 383-420. tinental shelf during summer 198 1. Cont. Shelf Res. 4: 65 I l-
STEIDINGIIR,
K. A. 1975. Basic factors influencing red tides, p.
6535.