Deep-Sea Research I 49 (2002) 321–338
The annual cycle and biological effects of the Costa Rica Dome
Paul C. Fiedler*
NOAA/National Marine Fisheries Service (NMFS), Southwest Fisheries Science Center, P.O. Box 271, La Jolla, CA 92038, USA
Received 20 October 2000; received in revised form 17 May 2001; accepted 4 September 2001
The Costa Rica Dome is similar to other tropical thermocline domes in several respects: it is part of an east–west
thermocline ridge associated with the equatorial circulation, surface currents flow cyclonically around it, and its
seasonal evolution is affected by large-scale wind patterns. The Costa Rica Dome is unique because it is also forced by a
coastal wind jet. Monthly climatological fields of thermocline depth and physical forcing variables (wind stress curl and
surface current divergence) were analyzed to examine the structure and seasonal evolution of the dome. The annual
cycle of the dome can be explained by wind forcing in four stages: (1) coastal shoaling of the thermocline off the Gulf of
Papagayo during February–April, forced by Ekman pumping on the equatorward side of the Papagayo wind jet; (2)
separation from the coast during May–June when the intertropical convergence zone (ITCZ) moves north to the
countercurrent thermocline ridge, the wind jet stops, and the North Equatorial Countercurrent extends toward the
coast on the equatorward flank of the ridge; (3) countercurrent thermocline ridging during July–November, when the
dome expands to the west as the countercurrent thermocline ridge shoals beneath a band of cyclonic wind stress curl on
the poleward side of the ITCZ; and (4) deepening during December–January when the ITCZ moves south and strong
trade winds blow over the dome. Coastal eddies may be involved in the coastal shoaling observed during February–
March. A seasonally predictable, strong, and shallow thermocline makes the Costa Rica Dome a distinct biological
habitat where phytoplankton and zooplankton biomass are higher than in surrounding tropical waters. The physical
structure and biological productivity of the dome affect the distribution and feeding of whales and dolphins, probably
through forage availability. Published by Elsevier Science Ltd.
Keywords: Equatorial circulation; Trade winds; Ekman pumping; Thermocline; Biological production; Annual variations
The Costa Rica Dome, located off the coast of
Central America, is a shoaling of the generally
strong (large dT=dz) and shallow thermocline of
the eastern tropical Pacific Ocean. It was first
observed in 1948 (Wyrtki, 1964) and first described
by Cromwell (1958). The dome has been observed
and studied several times since the late 1950s,
when a productive tuna fishery began to develop in
the region. However, it has been sampled inten-
sively, only a few times: the Costa Rica Dome
Expedition by the Scripps Tuna Oceanography
Research program in 1959 and the Mexican
DOMO surveys in 1979–1982 were exceptional
for applying much sampling effort in repeated
surveys of the dome.
E-mail address: firstname.lastname@example.org (P.C. Fiedler).
0967-0637/02/$-see front matter Published by Elsevier Science Ltd.
PII: S 0967 -0637(01 )0 0057-7
The mean position of the dome is near 91N
901W, at the end of a thermocline ridge which
shoals from west to east across the Pacific, between
the westward North Equatorial Current (NEC)
and the Eastward North Equatorial Counter-
current (NECC) (Fig. 1). This ridge and the dome
extend below the thermocline, to a depth of more
than 300m (Fig. 2). Tropical thermal domes also
exist in the eastern Atlantic (the Guinea Dome to
the north of the equator and the Angola Dome to
the south; Mazeika, 1967) and the western Pacific
(Mindanao Dome; Wyrtki, 1961). There is a sub-
thermocline Peru Dome in the southeastern
tropical Pacific (Voituriez, 1981).
The Costa Rica Dome is the peak at the
terminus of the countercurrent thermocline ridge.
The top of the thermocline is about 15m deep
at the dome, compared to 30–40m to the north
andsouth (Fig. 2, top).
ridge shoals gradually from west to east, then
drops off sharply between the dome and the
coast (Fig. 2, bottom). Thus, the dome is a
minimum in the annual mean field of thermocline
depth, 300–500km in diameter and centered
300km off the Gulf of Papagayo between Costa
Rica and Nicaragua (Fig. 1).
Surface winds and currents in the region of the
Costa Rica Dome change seasonally as the
intertropical convergence zone (ITCZ) between
the trade wind belts moves north and south with
the sun (Fig. 3). The relative strengths of the
northeast and southeast trade winds in the region
vary considerably during the year. Winter high-
pressure systems over the Gulf of Mexico and
Caribbean Sea force strong winds through low
altitude passes in the mountainous backbone of
southern Mexico and Central America. Intense
and narrow wind jets blow offshore at the Gulfs of
Tehuantepec, Papagayo, and Panama (McCreary
et al., 1989; Chelton et al., 2000a,b). In summer,
the southeast trade winds blow strongly across the
equator as far as 81N.
100˚W 95˚W 90˚W 85˚W 80˚W
Equatorial Thermocline Ridge
Fig. 1. Annual mean thermocline depth (201C isotherm depth) in the region of the Costa Rica Dome. Data from World Ocean
Database 1998 (Conkright et al., 1999).
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 322
The westward NEC and eastward NECC flow in
geostrophic balance along the poleward and
equatorward slopes, respectively, of the counter-
current thermocline ridge. The westward South
Equatorial Current (SEC) flows along the equa-
torial thermocline ridge. Part of the NECC is
deflected north at the coast of Central America
and joins the Costa Rica Coastal Current (CRCC),
which flows into the NEC. This pattern of cyclonic
flows around the Costa Rica Dome. The NECC
does not extend east of 1001W during February–
Cromwell (1958) published the first description
of the ‘‘Costa Rican Thermal Dome’’ and noted its
relationship to ‘‘cyclonic current shear’’. Wyrtki
(1964) elaborated Cromwell’s observations using
data collected by the Scripps Tuna Oceanography
Research program in the late 1950s. He proposed
that the steady-state heat balance of thermocline
doming was maintained by upwelling ‘‘caused by
the cyclonic flow around the dome’’. Upwelling
was explained as the result of an adjustment of the
geostrophic balance of the NECC as it turned
northward at the coast, around the eastern end of
the countercurrent thermocline ridge. Although he
Equatorial Thermocline Ridge
Countercurrent Thermocline Ridge
?COSTA RICA DOME
?COSTA RICA DOME
110˚W105˚W 100˚W 95˚W 90˚W 85˚W
Fig. 2. Annual mean meridional (top) and zonal (bottom) temperature sections through the Costa Rica Dome. Data from World
Ocean Database 1998 (Conkright et al., 1999).
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 323
had observations only for May–December, Wyrtki
predicted that doming would be weak in early
spring when the NECC was weak.
Wyrtki (1964) explicitly dismissed consideration
of ‘‘an effect of the local wind on the circulation
and the upwelling in the dome’’. In contrast,
Hofmann et al. (1981) proposed that upwelling in
the dome is forced by Ekman pumping beneath a
local, seasonal (May–October) maximum in cyclo-
nic wind stress curl associated with the seasonal
northward migration of the ITCZ. Umatani and
Yamagata (1991) considered remote wind effects
on the Costa Rica Dome. Based on the results of
experiments with a high-resolution regional ocean
circulation model, they concluded that the dome is
eroded by anticyclonic eddies generated on the
poleward side of the strong winter wind jet
blowing across the Nicaraguan lake district just
north of the Gulf of Papagayo in November–
March. Note that this wind jet, as well as the
resulting eddies, are conventionally labeled ‘‘Pa-
pagayo’’ although the core of the jet is about
70km north of the Gulf of Papagayo (Fig. 3;
Chelton et al., 2000b). Regeneration of the model
dome during February–April requires vorticity
input by a weaker cyclonic eddy formed on the
other side of the wind jet. Additional model
experiments showed that weak positive local wind
stress curl contributes to doming at the eastern end
of the countercurrent thermocline ridge during the
summer. However, these summer winds alone are
not sufficient to generate the dome.
Like the Costa Rica Dome, all tropical thermal
domes are associated with a cyclonic turning of
predominantly zonal tropical surface currents, as
expected in geostrophic balance. The possible
25 cm s-1
100˚W 95˚W 90˚W 85˚W 80˚W
25 cm s-1
100˚W95˚W 90˚W85˚W 80˚W
Fig. 3. Mean monthly fields of surface wind velocity (top) and surface current velocity (bottom), representing seasonal extremes in the
region of the Costa Rica Dome. Surface winds from COADS (Roy and Mendelssohn, 1995), surface currents from windage-corrected
ship drift (NOAA/NODC Ocean Current Drifter Data). Shading indicates surface wind divergence o?0.5?10?7s?1(intertopical
convergence zone). NEC=North Equatorial Current, SEC=South Equatorial Current, NECC=North Equatorial Countercurrent,
CRCC=Costa Rica Coastal Current.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 324
influence of winter coastal eddies is unique to the
Costa Rica Dome (Umatani and Yamagata, 1991).
Voituriez (1981) proposed that subthermocline
doming was the result of subsurface cyclonic flow,
however the doming of isotherms below the
thermocline dome known as the Costa Rica Dome
is not considered here.
This paper will examine monthly climatological
fields of thermocline depth and physical variables
(wind stress curl and surface current divergence)
that might be related to the annual cycle or
seasonal evolution of the Costa Rica Dome.
Comparison with studies of other tropical thermal
domes, in particular, the Guinea and Mindanao
Domes which are also located in northern
equatorial current systems and are arguably the
closest analogs to the Costa Rica Dome, provides
some insight. Finally, biological effects of the
dome on phytoplankton production and the
distribution of zooplankton and other animals
are briefly reviewed.
Hydrographic variables were derived from
MBT, XBT, and CTD profiles in the NOAA/
National Oceanographic Data Center World
Ocean Database 1998 CD-ROM Data Set Version
2.0 (Conkright et al., 1999), incorporating the
corrections available on-line as of July 20, 2001. A
total of 30,030 hydrographic profiles were avail-
able in the study area from 1941 to 1996, with 93%
from the period 1955–1996. Thermocline depth
was defined as the depth of the 201C isotherm,
which is often used as an index of thermocline
depth in the tropical ocean (Kessler, 1990).
Monthly thermocline depth observations ranged
from 1582 in July to 3656 in November (Fig. 4).
Surface current velocities were derived from ship
drift data, because available CTD data are not
adequate to resolve monthly patterns of geopo-
tential anomaly and geostrophic currents. A total
of 130,576 ship drift observations were available in
the study area from 1901 to 1987, with 80% from
the period 1922 to 1941. Monthly ship drift
observations ranged from 5457 in July to 20,793
in March (Fig. 4). Although ship drift is not an
ideal estimator of surface currents, Arnault (1987)
found that surface currents estimated from ship
drift in the tropical Atlantic were nearly equal to
the sum of geostophic currents and Ekman drift.
Surface current velocities derived from ship drift
(NOAA/NODC Ocean Current Drifter Data, CD-
ROM NODC-53) were corrected for windage by
the method of Richardson (1997).
Wind velocities were extracted from the Com-
prehensive Ocean Atmosphere Data Set in the
‘‘COADS on CD-ROM’’ data product (Roy and
Mendelssohn, 1995). A total of 984,602 wind
observations were available in the study area from
1857 to 1990, with 80% from the period 1954 to
1990. Monthly wind observations ranged from
76,372 in December to 88,534 in May (Fig. 4).
Data were gridded with the objective analysis
algorithm used to produce the World Ocean Atlas
(da Silva et al., 1994), with some modifications for
this study area: (1) The radii of influence used in
the global analysis were reduced to resolve smaller
scale features. (2) The radii of influence were
MBT, XBT, CTD
Fig. 4. Monthly totals of temperature profile (MBT, XBT and CTD), ship drift, and COADS wind observations in the Costa Rica
Dome study area, 0–171N, 100–771W.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338325
anisotropic (meridional/zonal ratio of 0.6, based
on variogram analysis of the hydrographic data).
Five passes of the objective analysis scheme were
performed at scales of 17.51 (10.51), 13.71 (8.21),
10.01 (6.01), 6.21 (3.81), and 5.01 (3.01) longitude
(latitude). A third modification was made for this
study area, in which interannual variability can be
as great as variability between months. (3) Raw
data were averaged by month–year before being
averaged by month, to minimize bias that might
arise at a grid point if many observations were
made in one or more anomalous years.
Sources of error in 11 monthly fields derived
from objective analysis of ship observations are
discussed in da Silva et al. (1994). Sampling error
for climatological monthly means of temperature
and wind speed in the eastern tropical Pacific, due
to the concentration of ship observations along
well-traveled shipping lanes, should be much less
than the root-mean-square errors of B0.31C and
estimated for individual monthly
means in the tropics (da Silva et al., 1994). No
wind speed corrections for Beaufort scale esti-
mates, anemometer height, or stability were
attempted, assuming that these errors are noise
that will not substantially affect the climatological
wind stress fields analyzed here.
Wind velocity was converted to wind stress
(Nm?2) using an air density of 1.2kgm?3and a
drag coefficient equal to 1.2?10?3for wind speeds
(ms?1)?1for stronger winds (Large and Pond,
1981). For calculation of both wind stress curl
(¼ dv=dx?du=dy) and surface current divergence
(¼ du=dxþdv=dy; where u is eastward and v is
northward wind stress or surface current), u and v
were gridded at 0.51 resolution. The derivatives
were then approximated as linear regressions
through grid values within 71.51 latitude or
longitude from each grid point. For surface
current divergence, flow was set to zero at the
Biological data were obtained from several
sources. Monthly mean SeaWiFS (Sea-viewing
Wide Field-of-view Sensor, http://seawifs.gsfc.
nasa.gov/SEAWIFS.html) chlorophyll fields were
calculated from Level 3 monthly standard mapped
image products (chlorophyll a concentration,
October 1997–July 2001, 9-km resolution at the
equator, version 3 reprocessing). The SeaWiFS
data cover one extreme El Ni* no year and two
moderate La Ni* na years: while the monthly
climatologies may change somewhat as the data
set is extended, they are the best available at this
time. The spatial and temporal coverage of the 3-
to 4-yr SeaWiFS chlorophyll climatologies are
considerably more complete than the 8- to 9-yr
CZCS climatologies (1978–1986). Historical zoo-
plankton volume and nitrate data were obtained
from NOAA/National Oceanographic Data Cen-
ter. The sample coverage of these data sets were
not sufficient for monthly analysis, so only mean
fields are presented.
The annual cycle of the Costa Rica Dome is
illustrated by maps of monthly mean thermocline
depth (Fig. 5). In January, the dome is far offshore
(921W) and relatively deep; thermocline depth is
433m. In February, the thermocline begins to
shoal at the coast on the equatorward side of the
Gulf of Papagayo. Coastal shoaling increases
during March and April. The elevation of the
thermocline begins to look like a dome in April,
but remains connected to the coast. During May
and June, the dome separates from the coast. The
dome deepens somewhat during June, but by July,
it is a distinct offshore dome. From July through
October, the dome increases in size, primarily to
the west along the countercurrent thermocline
ridge. The dome remains shallow and extends
slightly northward towards the Gulf of Tehuante-
pec in November, then decreases in size as the
countercurrent thermocline ridge deepens in De-
cember. Thus, the annual cycle of the dome
consists of: (1) coastal shoaling in February–April,
(2) separation from the coast in May–June, (3)
countercurrent thermocline ridging in July–Novem-
ber, and (4) deepening in December–January.
Annual cycles of winds and wind stress curl
(Fig. 6) and ship drift surface currents and surface
current divergence (Fig. 7) are presented along
with the cycle of doming (35m thermocline depth
contour) to examine interrelationships. The annual
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338326
cycle of doming and concurrent physical forcing
variables can be summarized as follows.
In January, the countercurrent thermocline
ridge is relatively deep. A weak doming at about
81N, 921W (Fig. 5) is a remnant of the preceding
year’s Costa Rica Dome. The Papagayo wind jet
has been blowing strongly to the WSW since
December (Fig. 6). The wind jet produces two
lobes of wind stress curl extending off the coast: a
negative (anticyclonic) lobe on the poleward side
of the jet and a stronger positive (cyclonic) lobe on
the equatorward side of the jet. The same pattern
of wind stress curl is produced by the Tehuantepec
wind jet, which begins blowing strongly in
November. An important difference between the
Tehuantepec and Papagayo wind stress curl
patterns is that the lobe of cyclonic curl on the
equatorward side of the Papagayo jet is contiguous
with the band of cyclonic curl on the northern side
of the ITCZ.
70 60 55 50 45 40 35 30 25
Thermocline Depth (m):
Fig. 5. Monthly mean fields of thermocline depth (201C isotherm depth) in the region of the Costa Rica Dome. Grid lines are at 91N
and 901W. Data from World Ocean Database 1998 (Conkright et al., 1999).
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338327
From February through April, thermocline
shoaling occurs at the coast directly beneath the
cyclonic wind stress curl associated with the
Papagayo wind jet. Strong surface current diver-
gence is evident in the ship drift data at this
location (Fig. 7). The ITCZ is at its annual
southernmost extreme in February. As coastal
shoaling continues during March and April, the
ITCZ moves to the north with two important
results: (1) the Papagayo wind jet weakens, and (2)
the zonal band of cyclonic curl on the northern
side of the ITCZ moves north to a location
over the countercurrent thermocline ridge. The
thermocline ridge deepens to the west of 901W,
however, as it shoals near the coast at this time
The dome begins to separate from the coast in
May, when the Papagayo wind jet stops and the
associated cyclonic curl and surface current
divergence weaken. At the same time, the SE trade
Apr May Jun
Wind Stress Curl (x10-7 N·m-3):
-1.0 -0.5 -0.2 0 +0.2 +0.5 +1.0
Winds: 10 m·s-1
Fig. 6. Monthly mean fields of winds and wind stress curl in the region of the Costa Rica Dome. Dashed line marks the meridional
minimum in surface wind divergence (ITCZ). Monthly positions of the Costa Rica Dome (35m thermocline depth contour, thick lines)
from Fig. 5. Data from COADS (Roy and Mendelssohn, 1995).
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338328
winds extend northward and the NECC begins to
flow to the east past the southern side of the dome.
The dome is clearly separated from the coast in
June. It is relatively small and near the coast; the
thermocline ridge to the west of 901W is only
beginning to shoal at this time.
The countercurrent thermocline ridge shoals
from July through October. The dome expands
as the thermocline shoals, primarily to the west
along the ridge. The east–west dimension of the
dome increases from 300km in June to 1000km in
November. This lengthening seems to coincide
more with the band of cyclonic wind stress curl
associated with the ITCZ (Fig. 6) than with the
(Fig. 7). Surface current convergence appears
over parts of the dome in some months, but
this may reflect deficiencies of the ship drift
data for estimating surface current divergence.
The SE trade winds and NECC remain strong on
the southern side of the dome and ridge into
Ship Drift Divergence (x10-7 s-1):
-10 -7 -4 -2 0 +2 +4 + 7 +10
Ship Drift: 50 cm·s-1
Fig. 7. Monthly mean fields of ship drift surface currents and surface current divergence in the region of the Costa Rica Dome.
Monthly positions of the Costa Rica Dome (35m thermocline depth contour, thick lines) from Fig. 5. Data from NOAA/NODC
Ocean Current Drifter Data.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338329
The Tehuantepec wind jet begins to blow in
November as the ITCZ moves to the south.
Shoaling of the thermocline on the north side of
the dome at the southern extreme of this jet
appears to be forced by cyclonic wind stress curl
on the equatorward side of the jet. The dome
deepens during December and January as the band
of cyclonic curl moves south of the thermocline
ridge and the NE trade winds blow strongly over
Monthly climatologies of SeaWiFS chlorophyll
concentration and doming show winter–spring
blooms near the coast at the Gulfs of Tehuantepec,
Papagayo and Panama (Fig. 8). At the Gulf of
Papagayo, this bloom is more closely associated
with the wind jet itself than with the thermocline
shoaling on the equatorward side of the wind jet.
When the Papagayo winds have abated in May,
the coastal blooming ends. From May through
September, chlorophyll increases offshore along
the shoaling thermocline ridge, with a local
maximum corresponding very closely to the Costa
Rica Dome. Coastal blooming begins again in
October at Tehuantepec and December at Papa-
gayo and Panama, when the wind jets initiate
mixing of surface waters near the coast.
A clear annual cycle of the Costa Rica Dome is
seen in the monthly climatologies of thermocline
Fig. 8. Monthly mean fields of SeaWiFS chlorophyll concentration in the region of the Costa Rica Dome. Each monthly mean is the
mean of three or four Level 3 monthly Standard Mapped Image products (reprocessing #3) from September 1997 to July 2001.
SeaWiFS data produced by NASA SeaWiFS Project and distributed by the Distributed Active Archive Center at NASA/Goddard
Space Flight Center. Monthly positions of the Costa Rica Dome (35m thermocline depth contour, thick lines) from Fig. 5.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338330
depth (Fig. 5). The dome may be ‘‘permanent’’ in
the sense that thermocline doming and associated
cyclonic circulation are always present within the
region. However, the dome’s location and magni-
tude (the height and breadth of the dome atop the
thermocline ridge) vary considerably during the
year. Doming begins near the coast at the Gulf of
Papagayo in February–March, moves offshore
(WSW) during April–July, intensifies from July
through November, then diminishes in December–
January. The annual cycles of winds and currents
presented here suggest that the observed cycle of
doming is closely related to the variability of
winds. However, doming and surface circulation
are also related.
It is well known that surface currents flow
cyclonically around the Costa Rica Dome, and all
tropical thermocline domes, as must occur in
geostrophic balance. For northern hemisphere
domes (Costa Rica, Guinea, Mindanao), this
circulation consists of the NECC on the south
side, the NEC on the north side, a coastal current
between the dome and the coast, and some return
flow on the opposite side of the dome. The Costa
Rica and Guinea Domes are at the shallow,
eastern end of the thermocline ridges between the
NEC and NECC in the Pacific and Atlantic; the
Mindanao Dome is at the western end of the
Pacific countercurrent thermocline ridge, where
the thermocline is much deeper at B140m
The observed stages in the annual cycle of the
Costa Rica Dome are closely related to migration
of the ITCZ and associated wind stress curl
patterns. Local Ekman pumping is consistent with
the February–April shoaling of the thermocline at
the coast (Fig. 6). Hofmann et al. (1981) noted a
correspondence between the mean summer posi-
tion of the dome and the large-scale wind stress
curl field. However, for much of the annual cycle
(after the February–April coastal shoaling) in the
finer-scale monthly fields presented here, the
spatial maximum in the band of positive (cyclonic)
wind stress curl associated with the ITCZ is
located to the east of the dome. This point has
been made previously by Barber! an et al. (1986)
and Umatani and Yamagata (1991). Spatial
mismatches between thermocline domes and posi-
tive wind stress curl maxima are also apparent at
the Guinea Dome (Bakun, 1987; Siedler et al.,
1992) and Mindanao Dome (Harrison, 1989).
Accordingly, these authors relate doming to
large-scale wind patterns.
The Costa Rica, Guinea, and Mindanao Domes
all exhibit seasonal variability that appears to be
related to the large-scale wind field, in particular,
the convergence of the trade winds in the ITCZ
and the associated cyclonic wind stress curl. The
Guinea Dome, like the Costa Rica Dome, is more
shallow in summer-fall, when the ITCZ lies over
the dome and local winds are weak, than in winter,
when the ITCZ is south of the dome and strong
NE trade winds blow over the dome (Siedler et al.,
1992). In a similar manner, the Mindanao Dome is
more shallow in winter (December–March), when
the NE Asian winter monsoon to the north
converges into the ITCZ over the dome, than in
summer when weaker winds converge to the north
of the dome (Masumoto and Yamagata, 1991).
Although all tropical domes seem to be related in
some way to the large-scale wind field, the winter
Papagayo wind jet is a small-scale feature of the
wind field that is unique to the Costa Rica Dome.
The observations summarized and analyzed
here show that doming is initiated near the coast
in February on the equatorward side of the
Papagayo wind jet. Ekman pumping velocities
calculated from wind stress curl according to the
algorithm of Xie and Hsieh (1995) can reach
values as high as 0.5md?1at this location (Fig. 9,
only February and August are presented here
because the spatial patterns are very similar to the
wind stress curl patterns in Fig. 6). Wyrtki (1964)
B0.1md?1from the steady-state heat balance of
the Costa Rica Dome. The dome deepens in
December when the NE trade winds begin to
blow strongly and the thermocline is eroded by
wind mixing. Mean monthly temperature profiles
(Fig. 10) illustrate that the surface layer over the
dome becomes mixed and the thermocline be-
comes both deeper and weaker between November
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 331
In contrast to the effects of local winds apparent
in the results presented above, remote forcing by
winds has been shown to cause seasonal changes in
doming in a series of model studies of the three
tropical thermal domes. Coastal downwelling
Kelvin waves weaken the Guinea Dome in
November–December and April–July (Yamagata
and Iizuka, 1995). Downwelling long Rossby
waves initiate decay of the Mindanao Dome in
March–April (Masumoto and Yamagata, 1991).
As described above, model coastal eddies weaken
the Costa Rica Dome in December–January and
then initiate regeneration of the dome in Febru-
ary–March (Umatani and Yamagata, 1991).
4.2. Coastal eddies
Both anticyclonic and cyclonic eddies have been
observed propagating from the coast near the Gulf
of Papagayo during late fall and winter (Giese
et al., 1994; Willett, 1996; M. uller-Karger and
Fuentes-Yaco, 2000). Willett (1996) conducted a
comprehensive study of anticyclonic eddies gener-
ated by winter wind jets at the Gulfs of Tehuante-
pec and Papagayo. In seven years of satellite
altimeter data, she detected two to three Papagayo
eddies per year generated between October and
February. These eddies moved west–southwest,
then west, at a mean speed of 11.8cms?1for 2–8
months before dissipating west of 1001W in the
seasonally strengthening NECC.
The Papagayo and Tehuantepec wind jets also
generate cyclonic eddies on the opposite, equator-
ward side of the jets. These cyclonic eddies are
Ekman Pumping Velocity (m day-1):
-1.0 -0.5 0 +0.5 +1.0
Fig. 9. February and August monthly mean fields of Ekman pumping velocity in the region of the Costa Rica Dome, calculated from
wind stress curl (Xie and Hsieh, 1995). Monthly positions of the Costa Rica Dome (35m thermocline depth contour, thick lines) from
Fig. 10. Monthly mean temperature profiles near the center of
the Costa Rica Dome (8–101N, 89–911W). Data from World
Ocean Database 1998 (Conkright et al., 1999). N ¼ 183
(November), N ¼ 85 (February).
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338332
weaker and shorter lived than the anticyclonic
eddies (McCreary et al., 1989), but have been
observed in sea surface temperature and ocean
color imagery (Stumpf and Legeckis, 1977; M-
. uller-Karger and Fuentes-Yaco, 2000). M. uller-
Karger and Fuentes-Yaco (2000) observed both
anticyclonic and cyclonic ‘‘mesoamerican’’ eddies
with translation speeds of 14–16cms?1.
The model of Umatani and Yamagata (1991),
and the present results, show regeneration of the
dome in late winter near the coast. In a model
experiment driven by May winds year-round (no
winter wind jets), the countercurrent thermocline
ridge was formed along 101N, but not the Costa
Rica Dome at the eastern end of the ridge
(Fig. 11(b) in Umatani and Yamagata, 1991).
Umatani and Yamagata (1991) argued that this
experiment demonstrated that a coastal cyclonic
eddy is necessary to ‘‘seed’’ the regeneration of the
The coastal shoaling and separation from the
coast stages in the seasonal evolution of the dome
(Fig. 5) may reflect, at least in part, the generation
and offshore movement of cyclonic coastal eddies.
At a nominal speed of 12cms?1, a cyclonic eddy
would move 310km in one month, approximately
equal to the distance between minima in the
thermocline depth climatologies at the coast in
February and offshore in March. However, the
role of cyclonic coastal eddies in the late winter to
spring evolution of the Costa Rica Dome cannot
be resolved in the monthly climatologies presented
This study has addressed the annual cycle of
doming and the factors that influence seasonal
variability of the Costa Rica Dome. The data and
analysis used here were not appropriate for
examining interannual variability. However, the
National Center for Environmental Prediction’s
Ocean Data Assimilation System uses a dynamic
ocean model to assimilate sparse temperature data
and produce subsurface temperature fields from
which thermocline depth can be calculated (Beh-
ringer et al., 1998). The results for months of July
100˚W 95˚W 90˚W 85˚W 80˚W
Costa Rica Dome - July 1980-1990
Normal years (1980,1984,1985,1986,1989,1990)
Warm years (1981,1982,1987)
Cold years (1988)
Fig. 11. July 1980–1990 positions of the Costa Rica Dome (35m thermocline depth contour). For the warm years of 1981, 1982, and
1987, J marks the peak of a thermocline dome deeper than 35m (no dome was evident in 1983). Data from National Center for
Environmental Prediction’s Ocean Data Assimilation System (Behringer et al., 1998).
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 333
from 1980 to 1990 show that the location and
magnitude of the Costa Rica Dome can vary
almost as much between years as within a year
4.3. Biological effects of the dome
Biological sampling of the Costa Rica Dome
region has not been adequate for resolving
monthly climatologies, with the exception of
satellite ocean color data (Fig. 8). The SeaWiFS
climatologies show patterns that can be related to
variability in thermocline depth and associated
processes implied by the physical data. A regional
maximum in chlorophyll concentration is present
in surface waters over the dome from May to
September. The association between high chlor-
ophyll and the thermocline dome is very close
during these months.
High chlorophyll levels off the Gulf of Tehuan-
tepec in October–March and Gulf of Papagayo in
December–April could be due to nutrient enrich-
ment of the surface layer by wind mixing and/or
Ekman pumping associated with the winter wind
jets (Trasvi* na et al., 1995). Winter winds at the
Gulf of Papagayo lag those at the Gulf of
Tehuantepec by one to two months (monthly
mean winds exceed 3ms?1from October to
February at the Gulf of Tehuantepec and from
December to April at the Gulf of Papagayo). The
high winter productivity associated with both the
Papagayo and Tehuantepec wind jets is centered
directly beneath the jets, indicating the importance
of wind mixing. In contrast, doming occurs on the
equatorward side of the Papagayo jet where wind
stress curl is positive, reflecting the importance of
Ekman pumping or upwelling in this process.
High primary productivity at the Costa Rica
Dome, and along the thermocline ridge to the west
of the dome, is supported by nutrients brought to
the surface by wind mixing and/or upwelling
(Broenkow, 1965; King, 1986). The thermocline
in the eastern tropical Pacific is relatively strong
and permanent: there is no deep winter mixing to
120˚W 110˚W100˚W 90˚W80˚W
0-20m Nitrate, mg m-3
Fig. 12. Mean 0–20m nitrate concentration in the region of the Costa Rica Dome. Data from 1960–1989 profiles in the on-line NODC
Oceanographic Profile Data Base (n ¼ 1276). Monthly positions of the Costa Rica Dome (35m thermocline depth contour, thick lines)
from Fig. 5.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 334
inject nutrients into the surface layer. Shoaling of
the thermocline puts cold, nutrient-rich water close
to the surface. Since turbulent motions driven by
wind stress on the sea surface decrease exponen-
tially below the surface, shoaling of the thermo-
cline can have a large effect on the input of
nutrients into the surface layer. The nitrate
climatology (Fig. 12) shows that the surface layer
in the region of the dome is enriched (see Pe* na et al.
(1994) for a surface nitrate climatology of the
entire tropical Pacific).
Nitrate is the limiting nutrient in tropical Pacific
waters where iron is adequate for nutrient uptake
(Barber and Chavez, 1991). Iron availability is not
as low at the Costa Rica Dome as it is in the high-
nitrate low-chlorophyll regions of the equatorial
and subarctic Pacific and Southern Ocean (Fung
et al., 2000). Mixing and upwelling of waters near
the coast by the Papagayo wind jet could be a
source of iron. Recent research may elucidate the
role of iron and other factors in the productivity of
the Costa Rica Dome (Bruland, K. pers. comm.,
University of California Santa Cruz).
Fig. 13. Mean daytime zooplankton volume from 333mm mesh nets on EASTROPAC surveys, 1967–1968. Monthly positions of the
Costa Rica Dome (35m thermocline depth contour, thick lines) from Fig. 5.
? Research vessel sightings
? U.S. tuna boat sightings
Fig. 14. Sighting locations of blue whales (Balaenoptera muscu-
lus) from research vessels (1976–1999, n ¼ 327) and US tuna
boats (1971–1990, n ¼ 191) in the NOAA/NMFS/SWFSC
sightings database. Monthly positions of the Costa Rica Dome
(35m thermocline depth contour, thin lines) from Fig. 5.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338335
Zooplankton data from the 1967 to 1968
primary production in the region of the Costa
Rica Dome supports a high zooplankton biomass
(Fig. 13). Mean biomass there is as high or higher
than in the equatorial upwelling zone. Larger
animals at higher trophic levels are also affected by
the dome. Blue whales observed year-round at or
near the dome (Fig. 14) may be feeding on large
standing stocks of euphausiids (Reilly and Thayer,
1990). Common dolphins feed on small pelagic fish
and squid in upwelling-modified waters off Baja
California, along the equator, and both at and east
of the Costa Rica Dome (Fig. 15). In contrast,
spotted dolphins are relatively rare at the dome.
Spotted dolphins feed on fish and squid in the
warmest waters of the eastern tropical Pacific,
where the thermocline is very strong and slightly
deeper than at the dome (Fiedler, 1992). They have
evolved a complex feeding association with yellow-
fin tuna and birds and apparently depend on the
tuna to drive prey from the thermocline up to the
surface. Perhaps, this association does not func-
tion or provides no advantage at the dome where
show that the high
the thermocline is shallower and weaker than to
This study has improved the description of the
annual cycle of the Costa Rica Dome, clarified the
wind forcing of this cycle, and demonstrated
pervasive biological effects of the dome. Further
study is needed on the influence of coastal eddies
on primary production at the dome, the season-
ality of secondary production and the exploitation
of this production by predators.
I thank the hundreds of scientists and techni-
cians whose observations over the years have made
this study possible. David Behringer, Russ Davis,
and Billy Kessler provided helpful comments, but
bear no responsibility for remaining shortcomings.
Several anonymous reviewers have prompted
substantial improvements. Linda Stathoplos of
NODC provided the EASTROPAC zooplankton
data and Al Jackson of SWFSC provided the
marine mammal sightings data. SeaWiFS data
Fig. 15. Sighting locations of common dolphins (Delphinus delphis) and spotted dolphins (Stenella attenuata) from research and tuna
vessels in the NOAA/NMFS/SWFSC sightings database (1971–1999). Monthly positions of the Costa Rica Dome (35m thermocline
depth contour, thin lines) from Fig. 5.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 336
were produced by the SeaWiFS Project (Code
970.2) and distributed by the Distributed Active
Archive Center (Code 902) at the Goddard Space
Flight Center, Greenbelt, MD 20771, sponsored
by NASA’s Mission to Planet Earth Program.
Finally, I thank Steve Reilly of SWFSC for
continued support of marine mammal habitat
Arnault, S., 1987. Tropical Atlantic geostrophic currents and
ship drifts. Journal of Geophysical Research 92, 5076–5088.
Bakun, A., 1987. Applications of maritime data to the study of
surface forcing of seasonal and interannual ocean variability
in eastern boundary regions. Ph.D. Thesis, Oregon State
Barber, R.T., Chavez, F.P., 1991. Regulation of primary
productivity rate in the equatorial Pacific. Limnology and
Oceanography 36, 1803–1815.
Barber! an, J.M., Gallegos, A., Fern! andez, A., Mee, L.D., 1986.
Subsurface origin of the Costa Rica Dome, unpublished.
Behringer, D.W., Ji, M., Leetmaa, A., 1998. An improved
coupled model for ENSO prediction and implications for
ocean initialization. Part I: the ocean data assimilation
system. Monthly Weather Review 126, 1013–1021.
Broenkow, W.W., 1965. The distribution of nutrients in the
Costa Rica Dome in the eastern tropical Pacific Ocean.
Limnology and Oceanography 10, 40–52.
Chelton, D.B., Freilich, M.H., Esbensen, S.K., 2000a. Satellite
observations of the wind jets off the Pacific coast of Central
America. Part I: case studies and statistical characteristics.
Monthly Weather Review 128, 1993–2018.
Chelton, D.B., Freilich, M.H., Esbensen, S.K., 2000b. Satellite
observations of the wind jets off the Pacific coast of Central
America. Part II: regional relationships and dynamical
considerations. Monthly Weather Review 128, 2019–2043.
Conkright, M.E., Levitus, S., O’Brien, T., Boyer, T.P.,
Stephens, C., Johnson, D., Baranova, O., Antonov, J.,
Gelfeld, R., Rochester, J., Forgy, C., 1999. World ocean
database 1998 CD-ROM data set documentation, Version
2.0. NODC Internal Report 14, 116pp.
Cromwell, T., 1958. Thermocline topography, horizontal
currents and ‘‘ridging’’ in the eastern tropical Pacific.
Bulletin Inter-American Tropical Tuna Commission 111,
da Silva, A.M., Young, C.C., Levitus, S., 1994. Atlas of surface
marine data 1994. Algorithms and Procedures, Vol. 1.
NOAA Atlas NESDIS 6, US Government Printing Office,
Washington D.C. 83pp.
Fiedler, P.C., 1992. Seasonal climatologies and variability of
eastern tropical Pacific surface waters. NOAA Technical
Report NMFS 109, 65pp.
Fung, I.Y., Meyn, S.K., Tegen, I., Doney, S.C., John, J.G.,
Bishop, J.K.B., 2000. Iron supply and demand in the
upper ocean. Global
Giese, B.S., Carton, J.A., Holl, L.J., 1994. Sea level variability
in the eastern tropical Pacific as observed by TOPEX and
tropical ocean-global atmosphere tropical atmosphere-
ocean experiment. Journal of Geophysical Research 99,
Harrison, D.E., 1989. On climatological monthly mean wind
stress and wind stress curl fields over the world ocean.
Journal of Climate 2, 57–70.
Wind generation of the Costa Rica Dome. Science 214,
Kessler, W.S., 1990. Observations of long Rossby waves in the
northern tropical Pacific. Journal of Geophysical Research
King, F.D., 1986. The dependence of primary production in the
mixed layer of the eastern tropical Pacific on the vertical
transport of nitrate. Deep-Sea Research I 33, 733–754.
Large, W.G., Pond, S., 1981. Open ocean momentum flux
measurements in moderate to strong winds. Journal of
Physical Oceanography 11, 324–336.
Masumoto, Y., Yamagata, T., 1991. Response of the western
tropical Pacific to the Asian winter monsoon: the generation
of the Mindanao Dome. Journal of Physical Oceanography
tropical Atlantic Ocean. Limnology and Oceanography 12,
McCreary, J.P., Lee, H.S., Enfield, D.B., 1989. The response of
the coastal ocean to strong offshore winds: with application
to circulations in the Gulfs of Tehuantepec and Papagayo.
Journal of Marine Research 47, 81–109.
M. uller-Karger, F.E., Fuentes-Yaco, C., 2000. Characteristics of
wind-generated rings in the eastern tropical Pacific Ocean.
Journal of Geophysical Research 105, 1271–1284.
Pe* na, M.A., Lewis, M.R., Cullen, J.J., 1994. New production in
the warm waters of the tropical Pacific Ocean. Journal of
Geophysical Research 99, 14255–14268.
Reilly, S.B., Thayer, V.G., 1990. Blue whale (Balaenoptera
musculus) distribution in the eastern tropical Pacific. Marine
Mammal Science 6, 265–277.
Richardson, P.L., 1997. Drifting in the wind: Leeway error in
shipdrift data. Deep-Sea Research I 44, 1877–1903.
Roy, C., Mendelssohn, R., 1995. CODE (Comprehensive
Ocean Dataset Extraction). User’s Guide. NOAA/NMFS/
Pacific Fisheries Environmental Group, Monterey, CA.
Siedler, G., Zangenberg, N., Onken, R., 1992. Seasonal changes
in the tropical Atlantic circulation: observation and simula-
tion of the Guinea Dome. Journal of Geophysical Research
Stumpf, H.G., Legeckis, R.V., 1977. Satellite observations
of mesoscale eddy dynamics in the eastern tropical
Pacific Ocean. Journal of Physical Oceanography 7,
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 337
Trasvi* na, A., Barton, E.D., Brown, J., Velez, H.S., Kosro, Download full-text
P.M., Smith, R.L., 1995. Offshore wind forcing in the
Gulf of Tehuantepec, Mexico: the asymmetric circu-
lation. Journal of Geophysical Research 100, 20649–
Umatani, S., Yamagata, T., 1991. Response of the eastern
tropical Pacific to meridional migration of the ITCZ: the
generation of the Costa Rica Dome. Journal of Physical
Oceanography 21, 346–363.
Voituriez, B., 1981. Northern and southern equatorial under-
currents and the formation of tropical thermal domes.
Oceanologica Acta 4, 497–506.
Willett, C.S., 1996. A study of anticyclonic eddies in the eastern
tropical Pacific Ocean with integrated satellite, in situ, and
modeled data. Ph.D. Thesis, University of Colorado, 127pp.
Wyrtki, K., 1961. Physical oceanography of the southeast Asian
waters. Naga Reports 2, 195pp.
Wyrtki, K., 1964. Upwelling in the Costa Rica Dome. Fishery
Bulletin 63, 355–372.
Xie, L., Hsieh, W., 1995. The global distribution of wind-
induced upwelling. Fisheries Oceanography 4, 52–67.
Yamagata, T., Iizuka, S., 1995. Simulation of the tropical
thermal domes in the Atlantic: a seasonal cycle. Journal of
Physical Oceanography 25, 2129–2140.
P.C. Fiedler / Deep-Sea Research I 49 (2002) 321–338 338