University of South Florida
Scholar Commons
Marine Science Faculty Publications
College of Marine Science
3-31-2009
Fish Effects on Ocean Current Observations in the
Cariaco Basin
Jyotika I. Virmani
Robert H. Weisberg
University of South Florida, [email protected]
Follow this and additional works at: http://scholarcommons.usf.edu/msc_facpub
Part of the Marine Biology Commons
Scholar Commons Citation
Virmani, Jyotika I. and Weisberg, Robert H., "Fish Effects on Ocean Current Observations in the Cariaco Basin" (2009). Marine
Science Faculty Publications. 138.
http://scholarcommons.usf.edu/msc_facpub/138
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].
Click
Here
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114, C03028, doi:10.1029/2008JC004889, 2009
for
Full
Article
Fish effects on ocean current observations in the Cariaco Basin
Jyotika I. Virmani1 and Robert H. Weisberg1
Received 24 April 2008; revised 13 January 2009; accepted 27 January 2009; published 31 March 2009.
[1] Multiple years of moored current meter observations from the Cariaco Basin show
low-frequency variations along with near-inertial waves and further imply the
persistent diurnal movement of fish species known to populate the basin. In agreement
with short-term observations from 1979, the more recent observations with acoustic
Doppler current profilers provide evidence of the multidecadal presence and behavior of
these species. An unwanted corollary, however, is a bias in both the vertical and
horizontal components of velocity due to the fish movements. Removal of this fish bias
results in large data loss (approximately 72%); however, an interpolated, non-biased
data set is developed with depth-averaged horizontal velocities comparable to the
observations, demonstrating successful elimination of the bias. Further comparisons show
that the interpolated data result in minimal variance density loss at low frequencies and a
reduction of variance density at high frequencies such that the interpolated data in the
internal wave range more closely fit the Garrett-Munk spectrum. The net result is a data
set appropriate for further analysis. A mean downward velocity of 0.18 cm s1 is a
reflection of a biogenic particle flux and some residual fish contamination. The mean
settling speed of particles in the Cariaco Basin is calculated, via Stokes law, to be smaller
than 0.04 cm s1. Velocity observations from acoustic current meters at depths greater
than 400 m are impacted by the water clarity; therefore alternate methods should be used
to make velocity measurements at depth.
Citation: Virmani, J. I., and R. H. Weisberg (2009), Fish effects on ocean current observations in the Cariaco Basin, J. Geophys. Res.,
114, C03028, doi:10.1029/2008JC004889.
1. Introduction
[2] The deep bathymetry and anoxic environment of the
Cariaco Basin (hereafter, the Basin), located off the northern coast of Venezuela, makes it a suitable location to study
the long-term climate variability of the Atlantic Ocean,
Caribbean Sea, and surrounding landmasses [Black et al.,
1999; Peterson et al., 2000; Haug et al., 2001]. The
dominant climate influence on the Basin is the easterly
Atlantic trade winds, which reflect the large-scale atmospheric pressure variability and Intertropical Convergence
Zone over the Atlantic Ocean. The proximity to the low
pressure and associated strong convection over South
America also plays a role. Biogenic particles, proxies of
surface climate variability, sink to the basin floor [Goñi et
al., 2003] and are preserved in an undisturbed anoxic
environment. Anoxic conditions prevail in the deeper part
of the basin because communication with the rest of the
Caribbean Sea is restricted by two sills to the north and
west, at 135 m [Febres-Ortega and Herrera, 1975] and
146 m [Richards and Vaccaro, 1956] depths, respectively.
Above the sill depths, water from the Caribbean Sea is less
inhibited from entering the Basin. The flux of these
1
College of Marine Science, University of South Florida, St. Petersburg,
Florida, USA.
Copyright 2009 by the American Geophysical Union.
0148-0227/09/2008JC004889$09.00
biogenic particles is affected both by the rate at which they
are produced and decompose [e.g., Thunell et al., 2000] and
the ocean circulation within the basin. Previous studies of
the physical characteristics of the water in the basin have
concentrated on hydrographic properties [e.g., Astor et al.,
2003] and implications from local wind-driven upwelling
[e.g., Muller-Karger et al., 2001, 2004]. In summer and
winter, wind-driven coastal upwelling may result in
changes in primary productivity and may affect the depth
of the anoxic-oxic boundary. Observations of the ocean
circulation within the Basin, especially below the sill
depths, however, have been lacking.
[3] To address this, we initiated ocean current measurements using moored acoustic Doppler current profilers
(ADCP), first from 1996 –1998, and again from 2002 to
present. A persistent feature in the current velocity observations are the diurnal vertical migrations of fish, also
observed by Love et al. [2004], seen every day throughout
this multiyear data set. Herein, we describe the mooring
design and give a cursory description of the observed
currents. The fish migrations bias the ADCP observations.
Therefore a scheme to remove the fish bias from the
velocity data is implemented and the validity of the new,
interpolated data set is tested. As an outcome of this
process, estimates of the biogenic particle’s downward
speeds are made. Finally, vertical mixing in the Basin is
discussed. A more detailed description of the physical
characteristics of the Cariaco Basin water, using the non-
C03028
1 of 12
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
C03028
C03028
Table 1. ADCP Deployment Details
Deployment Times
8 May to 6 November 1996
7 November 1996 to 13 May 1997
14 May 1997 to 11 November 1997
12 November 1997 to 1 June 1998
25 June 2002 to 5 May 2003
6 May 2003 to 16 May 2004
20 May 2004 to 12 May 2005
ADCP
Orientation
Deptha of Transducer
Head (m)
Depth of First
Good Bin (m)
Upward
Upward
Upward
Upward
Upward
Downward
Upward
Upward
Downward
250
262
215
211
289
316
289
289
316
237
249
202
198
277
329
276
277
329
a
All depths are relative to mean sea level.
biased interpolated data set, will be provided in a subsequent paper.
2. Data
[4] The first set of observations, from May 1996 to June
1998, had four sequential deployments with the same
mooring configuration. Sub-surface foam spheres carried
an upward looking 150 kHz RDI narrowband ADCPs to
measure horizontal and vertical currents. A Seabird Inc.
Seacat was mounted on the line 4m below the sphere to
measure temperature and salinity. The second set of observations, used in this study, from June 2002 to May 2005 had
three sequential deployments. Again, a sub-surface sphere
housed an upward looking 150 kHz RDI narrowband
ADCP. In addition to this ADCP, a second downward
Figure 1. Hourly horizontal and vertical velocity components observed in the Cariaco Basin and NCEP
daily winds from 25 June 2002 to 31 July 2002.
2 of 12
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
C03028
Figure 2. Hourly horizontal and vertical velocity components observed in the Cariaco Basin from
30 June 2002 to 5 July 2002. Time stamp is local time (GMT4 hours).
looking 150 kHz narrowband ADCP was mounted in an inline cage beneath the sphere. To measure temperature and
salinity, a Seacat was mounted 4m below the sphere and an
additional Seabird Inc. Microcat was mounted with the
downward looking ADCP in the cage.
[5] The depths of the ADCP transducer heads and the
first bin with data closest to the transducers for each
deployment are given in Table 1. The blanking distance
from the transducer heads was 4 m, and the bin sizes were
8.7 m. Sidelobe reflections contaminated the near-surface
observations resulting in the shallowest good data value at
approximately 40 m below sea level. All ADCPs sampled
every hour. The Seacats and Microcats sampled every
20 minutes from which hourly averages were calculated.
The temperature and salinity observations will be discussed
separately.
3. Hourly Currents and Fish Bias
[6] A representative sample of hourly horizontal and
vertical velocities, from 25 June 2002 to 31 July 2002,
and corresponding NCEP reanalysis daily winds [Kalnay et
al., 1996] are shown in Figure 1. The east (u) and north (v)
components show low-frequency current reversals with time
and depth. A pattern at higher frequency (approximately
every 2 – 3 days) with upward propagation, beginning at
depths of approximately 200 m, is often seen in the east
component (for example on 11 July 2002) and occasionally
in the north component. These upward propagating current
reversals are not present at all times, however they occur
regardless of the lower frequency current variations. The
local inertial frequency is 0.018 cph which corresponds to
2.3 days. Therefore these upward propagating patterns may
be due to near-inertial waves.
[7] The vertical (w) component shows a pattern at high
frequency (diurnal) variations oriented vertically. There are
also discrete vertical pattern excursions seen in all velocity
components extending below 400 –450 m. These observations occur every day from 1996 – 2005 with remarkable
clockwork regularity, and they are a consequence of fish
migration. Love et al. [2004] used data from biological
trawls and acoustic scattering measurements taken over
17 days in January and February 1979 to determine that
three different species of diurnally migrating fish reside in
the Cariaco Basin. In mid-to-late afternoon an upward
migration begins and by dusk all species are in the upper
200 m. At dawn, a downward migration begins, but a split
in the species occurs at the oxic-anoxic interface such that
two of them remain in the oxygen-rich environment, while
the third continues down to depths of 800– 1000 m. Details
3 of 12
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
C03028
Figure 3. Horizontal and vertical velocity components at 100, 200, 400, and 500 m depths from 25 June
2002 to 31 July 2002.
of the vertical migration from our more recent multiyear in
situ observations can be seen in Figure 2, which shows five
days of hourly data from 30 June to 5 July, 2002. The times
have been adjusted to reflect local time (LT; GMT4 hours).
In agreement with Love et al. [2004], these observations
show upward velocities increasing between 2 pm and 4 pm
LT. Positive vertical velocities diminish at 8 pm LT, when
the fish are in the upper 200 m. Downward migration begins
between 5 am and 7 am LT and continues until 12 pm to
1 pm LT. The consistent timing between the Love et al.
results and our observations show that these fish species
have been swimming in this basin with the same diurnal
rhythm for at least the past 27 years.
[8] Below 400– 450 m there are velocity variations in all
components during daytime, between 7 am and 7 pm LT,
when there are sufficient fish at those depths to generate an
acoustic return. However, during the night the fish are near
the surface and the ADCP echo amplitudes are too low to
discern water motion at depth. The lack of a good acoustic
signal at depths greater than approximately 400 m suggests
that there are insufficient scatterers, implying that the water
is very clear and locally homogeneous. This makes the
measurement of water velocity below 400– 450 m difficult
using acoustic current meters; however, we assume that the
currents do not change drastically from the daytime when
fish are at depth and relatively quiescent, and therefore both
horizontal and vertical current velocities are small at this
location below approximately 400– 450 m.
[9] The diurnal fish migration is also apparent in both
horizontal velocity components. For example, from 25 June
to 2 July the southward flow between 50 – 150 m
depths (Figure 1) has vertical striations of reduced speed
corresponding to times when the fish are closest to the
surface. This can also be seen in Figure 2 where horizontal
velocities at all depths vary synchronously with variations
in vertical velocity.
[10] Velocity time series at 100 m, 200 m, 400 m, and
500 m for 25 June to 30 July 2002 (Figure 3) show the
speeds at which the fish move. The intermittent east (U),
north (V), and vertical (W) components at 500 m confirms
that at depth most of the observed velocity is entirely due
to fish movement, with maximum horizontal speeds of
approximately 17 cm s1 and vertical speeds of approximately 8 cm s1. Velocity changes resulting from fish
movement decrease at shallower depths such that at 100 m
the maximum upward velocity is 5 cm s1, and the
4 of 12
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
C03028
C03028
Figure 4. Fish-biased (gray) and interpolated (black) vertical variance density spectra at 150 m using
cut-off values of 0.7, 0.5, and 0.3 cm s1.
maximum fish-related horizontal velocity is 5 cm s1 (north
component) to 7 cm s1 (east component).
4. Data Processing
4.1. Fish Bias Removal
[11] Schools of fish that swim in a coherent manner have
been known to bias acoustic measurements [e.g., Wilson
and Firing, 1992]. The distinct diurnal movements of fish
are biasing the Cariaco Basin ADCP current observations.
Further data processing is required before these observations can be used to make any calculations regarding the
ocean circulation in the Basin.
[12] The fish biased currents were removed by identifying
all observations in which the vertical velocity was greater
than a cut-off velocity of ±0.5 cm s1. Sensitivity tests were
made using cut-off values between 0.1 cm s1 to 1 cm s1
with 0.1 cm s1 increments. Data loss from these values
ranged from 99% to 32%, respectively. The auto spectra for
the observed vertical component at 150 m (Figure 4)
showed a large diurnal peak of 70.7 cm2 s2/cph. Interpolating over the gaps, the corresponding spectra from the
interpolated vertical components with cut-off values of
0.7 cm s1, 0.5 cm s1 and 0.3 cm s1 were 2.5 cm2 s2/cph,
1.2 cm2 s2 /cph, and 0.2 cm2 s2 /cph, respectively.
Although a cut-off value of 0.3 cm s1 almost eliminated
the diurnal peak associated with fish movement, the
corresponding data loss was 86%. Cut-off values of more
than 0.5 cm s1 continued to show a clear bias in all three
velocity fields. 0.5 cm s1 was chosen as a value which
removed as much bias as possible without crippling data
loss. Eliminating all corresponding observations in the east
and north component and interpolating over the gaps gave a
new, relatively unbiased data set (hereafter, the interpolated
data). This does not remove all the fish bias, but reduces it
considerably. Although the smallest gap was one hour, the
largest interpolation covered a 22 hour gap. Therefore
sections of the interpolated data need to be considered
carefully before studying high frequency variability at timescales of less than 24 hours. Figure 5 shows an example of
the observed east, north, and vertical components of velocity (gray) and the interpolated values (black) from June 26
to June 30. During the deployment covering June 2002 to
May 2003, the percentage of data lost by removing values
corresponding to vertical speeds greater than ±0.5 cm s1
was approximately 72%, which results, on average, in about
6 good samples per day as opposed to 24 good samples per
day. Despite this large loss of data due to fish bias, statistical
comparisons of the interpolated velocity field with the
original velocity observations show good agreement at
periodicities longer than diurnal.
4.2. Downward Particle Flux Velocity
[13] The depth-averaged velocity, time-averaged velocity
profile, variance with depth, and auto spectra at 150 m for
all velocity components, calculated from hourly values of
observed (gray) and interpolated (black) data between June
2002 to May 2003 are shown in Figure 6. The observed
depth-averaged vertical velocity component (Figure 6a, top)
shows the large diurnal variation from fish movement. The
vertical velocities at individual depths are comparable to the
depth-averaged vertical velocity. The mean depth-averaged
vertical velocity isnegative in both the observed (0.91 cm s1)
and interpolated (0.18 cm s1) data, presumably resulting
from the downward flux of biogenic particles in the Cariaco
Basin [Thunell et al., 2000; Scranton et al., 2001; MullerKarger et al., 2001]. The interpolated vertical velocity,
which no longer shows diurnal variability, is comparable
to velocities observed in the Sargasso Sea (0.12 cm s1)
[Deuser et al., 1990].
[14] Assuming that the particles are spherical, Stokes law
may be used to estimate the falling or terminal velocity, Vt,
of a particle:
Vt ¼
gd 2 rp rm
18m
ð1Þ
where g is gravity (9.8 m s2), d is the particle diameter (m),
rp is the particle density (assumed to be the bulk wet density
of the sediment, 1.13 103 kg m3, from Scranton et al.
[2001]), rm is the density of seawater (1.025 103 kg m3),
and m is the dynamic viscosity of seawater (1.09 103 kg
m1 s1 at 18°C). Sediment traps deployed in the Cariaco
Basin [e.g., Thunell et al., 2000] show a tri-modal particle
size distribution, with peaks at 3, 22, and 80 mm [Elmore et
al., 2004]. Using Stokes law, these correspond to particle
terminal velocities of 4.7 105 cm s1 (0.04 m day1),
5 of 12
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
C03028
Figure 5. Observed (gray) and interpolated (black) horizontal and vertical velocity components from
25 June 2002 to 30 June 2002.
2.5 103 cm s1 (2.2 m day1), and 3.4 102 cm s1
(29.4 m day1) respectively, which are similar to observed
sinking speeds of particles of comparable size [e.g., Knutsen
et al., 2001]. These values are smaller than the depthaveraged vertical velocity of 0.18 cm s1 (156 m day1)
observed by the current meters after fish removal. There are a
number of explanations for this discrepancy. This may be
because particle behavior differs in the natural environment
compared to the theoretical arena required for Stokes law to
apply. For instance, Chase [1979] showed an order of
magnitude increase in velocity for observed particles with
diameters 28.5 mm. Another factor may be that particles
collected in the sediment traps are smaller than those in the
water column. Observations of marine particles in situ prior
to their accumulation in a sediment trap [Sternberg et al.,
1999] have shown that the sediments in the traps may be an
order of magnitude smaller than those in the water column.
Other studies also suggest that the ocean contains larger
particles [e.g., McCave, 1975; Fowler and Knauer, 1986].
Assuming that the particle diameter is an order of magnitude
6 of 12
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
C03028
Figure 6a. The observed (gray) and interpolated (fish bias removed; black) W component: depthaveraged (top), time-averaged (lower left), and variance (lower center) distributions and variance density
spectra at 150 m (lower right).
too small, recalculating the terminal velocity for particles of
200 mm gives a terminal velocity of 0.21 cm s1 which is
comparable to the observed depth-averaged vertical
velocity after fish removal.
4.3. Interpolated Data Validity
[ 1 5 ] The time-averaged vertical velocity profile
(Figure 6a, lower left) shows a difference between the
observed and interpolated values, with the interpolated
values being closer to zero at all depths. The mean
interpolated vertical velocity profile above 300 m is
0.2 cm s1, which is similar to the mean depth-averaged
velocity. Below 300 m, the mean interpolated vertical
velocity is an order of magnitude smaller.
[16] The maximum observed variance (Figure 6a, lower
center) is between 450 –500 m, which is to be expected
from the fish movement. The interpolated variance is almost
zero at all depths except between 330 – 450 m, because all
the fish contaminated data could not be removed.
[17] The observed spectrum of vertical velocity (Figure
6a, lower right) at 150 m shows large energy peaks at
diurnal and higher frequencies, as a result of fish movement.
There is a diurnal peak in the interpolated data because
removing all the fish bias would result in a greater loss of
observations. However, the interpolated spectrum at higher
frequencies is more realistic than the observed spectrum,
with energy decreasing down to the Nyquist frequency
(0.5 cph) in the interpolated spectrum. This is in agreement
with other spectra of vertical velocity in the ocean [e.g.,
Lien et al., 1998]. At the highest frequencies, the spectra for
the interpolated data levels out such that near the Nyquist
frequency this measure of the instrument background level
variance is approximately 0.05 (cm s1)2 for the horizontal
components and 0.002 (cm s1)2 for the vertical component, giving root-mean-square (rms) values of 0.2 cm s1
and 0.04 cm s1, respectively. For the original observed
data, the rms values are 0.6 cm s1 (horizontal), and 0.3 cm
s1(vertical). The interpolated values are comparable to the
manufacturer’s rms values of about 0.1 cm s1 for the
horizontal and 0.04 cm s1 for the vertical components.
[18] The depth-averaged observed and interpolated horizontal velocities (Figures 6b and 6c) agree well with each
other, with mean values differing by less than 0.4 cm s1.
The observed and interpolated horizontal velocity vertical
profiles, especially above 300 m, also agree with each other.
There is a small reduction in the interpolated variance
compared to the observed variance. The variance density
spectra at 150 m in both velocity components show a
decrease at high frequencies in the interpolated data relative
to the observed data, but no significant change in the low
frequencies. There are peaks in both components, including
at 2.3 days and 24 hours. The energy at 24 hours can be
accounted for by residual fish bias because not all the fish
7 of 12
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
Figure 6b. As in Figures 6a but for the U component.
Figure 6c. As in Figures 6a but for the V component.
8 of 12
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
C03028
C03028
Figure 7. The high-frequency variance density spectra of observed (gray) and interpolated (black)
horizontal velocities with the Garrett-Munk 2 spectrum (red line).
movement could be removed, otherwise the loss of data
would be too large. The 2.3 days energy peak is consistent
with near-inertial waves. The remaining frequencies suggest
current variability due to other ocean dynamics.
[19] The decrease at higher frequencies in the interpolated
spectra in all velocity components is representative of
spectra from other geophysical observations [Gould, 1971;
D’Asaro et al., 1995; Henjes, 1999; Virmani, 2005; Mayer
et al., 2007]. At sub-inertial frequencies, in the internal
wave range of the spectra, the slope of the interpolated
spectra follows the Garrett-Munk spectrum [Garrett and
Munk, 1972], with a 2 slope (Figure 7). To further verify
the interpolated spectrum at high frequencies, an ensemble
spectrum was calculated using 15 segments from the
original data (ranging from 10 hours to 41 hours) when
there was minimal fish bias. At high frequencies, the
ensemble average spectra in the horizontal velocity components resembles the interpolated spectra. The agreement
between the interpolated data at these high frequencies
and the Garrett-Munk spectrum confirms that the cut-off
velocity is reasonable and the removal of fish-biased observations improved the data quality.
ical quantities in the Basin have increased since the 1950s
[Scranton et al., 1987; Zhang and Millero, 1993], therefore
some surface water properties are reaching these depths.
Are the persistent fish migrations responsible for this mixing, or is it due to ocean dynamics alone?
[21] Experiments in tanks have demonstrated that mixing
is greater with fish present than without [Rasmussen et al.,
2005]. Ocean mixing by the marine biosphere may be
comparable to physical mixing mechanisms such as wind
and tides [Dewar et al., 2006]. Huntley and Zhou [2004]
demonstrate that the turbulent energy generated by motile
fish is comparable to wind-driven turbulence and the
cumulative generation of turbulent energy by assorted
schools of fish in a region may be substantial. Following
the method used by Huntley and Zhou, we estimate the
energy production produced by fish in the Cariaco Basin.
[22] The Reynolds numbers (Table 2) for two fish species
in the Cariaco Basin (B. cantori and S. argentea) was
calculated using
5. Vertical Mixing
where L is the characteristic length of the fish, uf is the
speed at which the fish are moving and n is the kinematic
molecular viscosity of seawater, 1.4 106 m2 s1 [Pond
and Pickard, 1989]. From Figure 3, a conservative value of
0.03 m s1 was chosen for uf. For S. argentea, the Reynolds
[20] Richards and Vaccaro [1956] and Zhang and Millero
[1993] find that vertical mixing at depth greater than 150 m
(the sill depth) is inhibited, predominantly because of the
thermal stratification. However, various physical and chem-
Re ¼
9 of 12
uf L
v
ð2Þ
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
Table 2. Fish Parameters and Reynolds Number
Bregmacerous cantori
Steindachneria argentea
Average
Lengtha, L (m)
Average Wet
Massb, M (kg)
Reynolds
Number, Re
2.9 102
7.6 102
8.3 104
4.22 103
621
1629
a
From Love et al. [2004].
Wet mass = 5 dry mass [Huntley and Zhou, 2004]. Dry mass from
Love et al. [2004].
b
number is greater than 1000, which indicates turbulent
mixing.
[23] Assuming that the total rate of turbulent energy
dissipation by fish per kilogram of seawater, E, is equal to
the rate at which they expend kinetic energy, this may be
estimated using,
E¼
ed N
hr V
[26] The remaining term from (3) that needs to be
determined is the number of fish per unit volume, N/V.
From trawls conducted in 1979, Love et al. [2004] estimate
that during the night, when the fish are close to the surface
and active, the average density of adult B. cantori is 2.7 104 m3 and S. argentea is 8.2 105 m3. These values
are an order of magnitude smaller than the catches during
the day, when the fish were lethargic. However, these trawls
did not catch all the fish in their path and no fish from the
third species, Diaphus taaningi, were captured, possibly
because the trawls were towed too slowly. Additionally, fish
are known to avoid nets and trawls, resulting in an underestimate of these densities [Fréon and Misund, 1999]. An
alternate way of estimating N/V [Huntley and Zhou, 2004]
is,
N
¼ 0:64M 1:2
V
ð3Þ
where ed is the rate of work expended per fish to overcome
drag, h is the propulsive efficiency per fish, r is the density
of seawater, 1.025 103 kg m3, and N/V is the number of
fish per unit volume (m3). Fish work to overcome the drag
force and propel forward; therefore the rate of work per fish,
ed, is,
ed ¼ Duf
ð4Þ
where uf is the speed of the fish (3 cm s1), and D is the
drag force exerted on the fish by the water. The drag, D, is
calculated by,
1
D ¼ rSw u2f cd
2
ð5Þ
where cd is the drag coefficient calculated from [Hoerner,
1965],
cd ¼ 0:072Re0:2
ð6Þ
and Re is the Reynolds number (Table 2). Sw is the total
wetted surface area of the fish, which may be estimated as,
Sw ¼ 2WL
ð7Þ
where L is the length of the fish, and W is the width and is
assumed to be L/10 [Huntley and Zhou, 2004].
[24] The propulsive efficiency, h, in equation (3) is
estimated empirically [e.g., Weihs, 1977] using,
h ¼ 0:39u0:24
:
f
ð8Þ
[25] The values obtained from this equation fall within
the range of values for propulsive efficiency of swimmers
with similar mass [Torres, 1984; Lee and Lee, 1996].
C03028
ð9Þ
where M is the wet mass of the fish (Table 2). There is a
large difference between the numbers of fish per cubic
meter from the trawls versus those calculated theoretically
from equation (9). From acoustics, some of the trawls were
conducted partially or totally at depths and times where
there were no fish [Love et al., 2004], therefore the numbers
of fish were estimated. The trawl information implies 3
fish per 1000 m3 (B. cantori), which is too sparse to create
the observed acoustic signal during the peak vertical
velocities: in areas where there is a large acoustic signal,
we expect a greater density of fish. Further, Euphausia
superba (krill; 2 104 kg) and Engraulis japonicus
(2 103 kg) have comparable mass to B. cantori and
S. argentea. Values obtained from equation (9) are reasonable given that observed school volumes of E. superba and
E. japonicus are 4000/m3 [Hamner, 1984] and 4600/m3
[Lee and Lee, 1996], respectively.
[27] The values from equations (3) to (9) for each species
of fish are given in Table 3. The energy dissipation rate by
each species is O(107) W kg1. The largest uncertainty in
this calculation is the number of fish per unit volume. In the
mixed layer, wind-generated turbulence is O(107) W kg1
and turbulence due to convective overturning may be
O(105) W kg1 [e.g., Moum et al., 1995; Brainerd and
Gregg, 1995; Horne et al., 1996; Lozovatsky et al., 1999].
The calculated fish-induced dissipation rates are comparable
to the dissipation rates in the mixed layer due to physical
processes. Therefore it seems plausible that the persistent
diurnal fish migrations are contributing to vertical mixing in
the Cariaco Basin.
6. Summary
[28] Current meter observations made in the Cariaco
Basin, from 1996– 2005, show variability over a range of
timescales. In addition to low-frequency current reversals,
near-inertial waves are also seen in the horizontal velocity
Table 3. Values Used to Compute the Turbulent Energy Dissipation by Each Fish Species
(9) N/V (m3)
Bregmacerous cantori
Steindachneria argentea
3186
453
(8) h
(7) Sw (m2)
0.17
0.17
4
1.7 10
1.2 103
(5) D (kg m s2)
(6) cd
2
2.0 10
1.6 102
10 of 12
6
1.6 10
9.1 106
(4) ed (kg m2 s3)
8
4.8 10
2.7 107
(3) E (W kg1)
8.9 107
7.2 107
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
components. The largest vertical velocities result from the
diurnal migration of fish. Previous studies, using data from
1979, found three species of fish in the basin [Love et al.,
2004]. In agreement with our observations, these fish
migrate upward during the afternoon, reside close to the
surface at night, and descend early in the morning. The
agreement between these observations shows that these fish
species in the Basin have been following the same diurnal
movement for at least the past 27 years and hence throughout the modern climate era. Our estimates of the rate of
turbulent energy dissipation by these fish suggest that they
are contributing to the vertical mixing within the Basin.
[29] Diurnal fish migration is also seen in the horizontal
velocity components and heavily biases these observations.
At 500 m, for example, the horizontal velocity due to fish
migration is 17 cm s1. A simple scheme to remove the fish
bias is developed: all velocity measurements that occur
when the vertical velocity is greater than 0.5 cm s1 are
removed, and the data are interpolated to cover these gaps.
Although this results in a large loss of data, statistical
analysis shows an improved and more realistic data set
after the interpolation. The validity of this fish removal
scheme is confirmed by comparing the auto spectrum of the
fish-biased data with that of the interpolated data. At low
frequencies, there is good agreement between the horizontal
components of both data sets. The greatest change in all
velocity components occurs in the high frequencies (diurnal
and smaller timescales), such that the energy in the interpolated data decreases in accordance with other geophysical
observations, approximately following a Garrett-Munk
spectrum for horizontal velocities in the internal wave
range.
[30] At depths above 400 m, there appears to be a net
downward movement of water. This results from the flux of
particles, whose downward terminal velocity may be calculated using Stokes law and particle size measurements
[Elmore et al., 2004]. The observed depth-averaged vertical
velocity (after fish removal) of 0.18 cm s1 (156 m day1)
is larger than the calculated particle terminal velocities,
which range from 4.7 105 cm s1 to 3.4 102 cm s1
(0.04 m day1 to 29.4 m day1). This discrepancy may be
because of non-spherical particles or the differences in size
between particles in the water column and sediment trap.
[31] At depths greater than 400 – 450 m, the ADCPs only
detect an acoustic signal large enough to estimate a velocity
during the day, when fish are present. At other times, the
dearth of scatterers implies very clear water. Measuring
currents at those depths (and deeper) requires techniques
other than acoustic-based. However, when fish are present
the observed current velocity is very small, and we assume
that below 400– 450 m the currents are also weak at this
mid-basin location.
[32] Although the extent to which fish are biasing ADCP
data in the Cariaco Basin is substantial, the cyclic nature
of fish movement allow for easy identification of this
problem and subsequent removal of the biased data. The
edited data will be used to investigate the ocean dynamics in
the Basin.
[33] Acknowledgments. This work was supported by the National
Science Foundation, grant OCE-0326268. Dr. F. Muller-Karger is the lead
C03028
P.I. for this Cariaco Basin Project. Special thanks for logistical support and
essential mooring work assistance are given to personnel from the Fundación La Salle de Ciencias Naturales, Estación de Investigaciones Marinas
de Margarita (FLASA/EDIMAR) and the crew of the R/V Hermano Ginés
(FLASA). Fieldwork and computer support was provided by R. Cole,
J. Donovan, J. Law, L. Lorenzoni, D. Mayer, D. Rueda Roa, J. Scudder, and
P. Smith from USF; Dr. R. Thunnell and E. Tappa from USC; and R. Varela
and Y. Astor from EDIMAR. NCEP reanalysis data provided by NOAA/
OAR/ESRL/PSD, Boulder, Colorado, USA, from their Web site at http://
www.cdc.noaa.gov/. We also thank the anonymous reviewers for their
constructive comments.
References
Astor, Y., F. Muller-Karger, and M. I. Scranton (2003), Seasonal and interannual variation in the hydrography of the Cariaco Basin: Implications
for basin ventilation, Cont. Shelf Res., 23, 125 – 144.
Black, D. E., L. C. Peterson, J. T. Overpeck, A. Kaplan, M. N. Evans, and
M. Kashgarian (1999), Eight centuries of North Atlantic Ocean atmosphere variability, Science, 286, 1709 – 1713.
Brainerd, K., and M. Gregg (1995), Surface mixed and mixing layer depths,
Deep Sea Res. I, 42, 1521 – 1543.
Chase, R. R. P. (1979), Settling behavior of natural aquatic particulates,
Limnol. Oceanogr., 24, 417 – 426.
D’Asaro, E. A., C. C. Eriksen, M. D. Levine, P. Niiler, C. A. Paulson, and
P. Van Meurs (1995), Upper-ocean inertial currents forced by a strong
storm. part I: Data and comparisons with linear theory, J. Phys. Oceanogr.,
25, 2909 – 2936.
Deuser, W. G., F. E. Muller-Karger, R. H. Evans, O. B. Brown, W. E.
Esaias, and G. C. Feldman (1990), Surface-ocean color and deep-ocean
carbon flux: How close a connection?, Deep Sea Res., 37, 1331 – 1343.
Dewar, W. K., R. J. Bingham, R. L. Iverson, D. P. Nowacek, L. C. St.
Laurent, and P. H. Wiebe (2006), Does the marine biosphere mix the
oceans?, J. Mar. Res., 64, 541 – 561.
Elmore, A. C., R. C. Thunell, D. E. Black, R. W. Murray, and N. C.
Martinez (2004), Sedimentological signatures of transient depositional
events in the Cariaco Basin, Venezuela, Eos Trans. AGU, 85(47), Fall
Meet. Suppl., Abstract OS33A-0569.
Febres-Ortega, G., and L. E. Herrera (1975), Interpretacion dinamica de
algunas de las caracteristicas oceanograficas de la Fosa de Cariaco, Mar
Caribe, Bol., Inst. Oceanogr. Univ. de Oriente, 14, 3 – 29.
Fowler, S. W., and G. A. Knauer (1986), Role of large particles in the
transport of elements and organic compounds through the oceanic water
column, Prog. Oceanogr., 16, 147 – 194.
Fréon, P., and O. Misund (1999), Dynamics of Pelagic Fish Distribution
and Behaviour: Effects on Fisheries and Stock Assessment, Fishing News
Press, Oxford.
Garrett, C. J. R., and W. H. Munk (1972), Space-time scales of internal
waves, Geophys. Astrophys. Res., 80, 291 – 297.
Gould, W. J. (1971), Spectral characteristics of some deep current records
from the eastern North Atlantic, Philos. Trans. R. Soc. London Ser. A,
270, 437 – 450.
Goñi, M. A., H. L. Aceves, R. C. Thunell, E. Tappa, D. Black, Y. Astor,
R. Varela, and F. Muller-Karger (2003), Biogenic fluxes in the Cariaco
Basin: A combined study of sinking particulates and underlying
sediments, Deep Sea Res. I, 50, 781 – 807.
Hamner, W. M. (1984), Aspects of schooling Euphausia superba, J. Crustacean Biol., 4(1), 67 – 74.
Haug, G. H., K. A. Hughen, D. M. Sigman, L. C. Peterson, and U. Röhl
(2001), Southward migrations of the intertropical convergence zone
through the Holocene, Science, 293, 1304 – 1308.
Henjes, K. (1999), On probing the inertial subrange, Boundary Layer
Meteorol., 91, 367 – 384.
Horne, E., J. Loder, C. Naimie, and N. Oakey (1996), Turbulence
dissipation rates and nitrate supply in the upper water column on Georges
Bank, Deep Sea Res. II, 43, 1683 – 1712.
Hoerner, S. (1965), Fluid-Dynamic Drag, SF Hoerner, Bricktown, N. J.
Huntley, M. E., and M. Zhou (2004), Influence of animals on turbulence in
the sea, Mar. Ecol. Prog. Ser., 273, 65 – 79.
Lee, M., and K. Lee (1996), The larval anchovy (Engraulis japonicus)
fishery in relation to the environmental factors in coastal waters of
Fangliao, Taiwan, Fish. Res., 26, 37 – 48.
Lien, R.-C., E. A. D’Asaro, and G. T. Dairiki (1998), Lagrangian frequency
spectra of vertical velocity and vorticity in high-Reynolds-number
oceanic turbulence, J. Fluid Mech., 362, 177 – 198.
Lozovatsky, I., T. Dillon, A. Erofeev, and V. Nabatov (1999), Variations of
thermohaline structure and turbulent mixing on the Black Sea shelf at the
beginning of autumn cooling, J. Mar. Syst., 21, 255 – 282.
11 of 12
C03028
VIRMANI AND WEISBERG: OBSERVATIONS IN THE CARIACO BASIN
Kalnay, et al. (1996), The NCEP/NCAR 40-year reanalysis project, Bull.
Am. Meteorol. Soc., 77, 437 – 470.
Knutsen, T., W. Melle, and L. Calise (2001), Determining the mass density
of marine copepods and their eggs with a critical focus on some of the
previously used methods, J. Plankton Res., 23, 859 – 873.
Love, R. H., R. A. Fisher, M. A. Wilson, and R. W. Nero (2004), Unusual
swimbladder behavior of fish in the Cariaco Trench, Deep Sea Res. I, 51,
1 – 16.
Mayer, D. A., J. I. Virmani, and R. H. Weisberg (2007), Velocity comparisons from upward and downward profilers on the West Florida Shelf,
J. Atmos. Ocean Technol., 24, 1950 – 1960.
McCave, I. N. (1975), Vertical flux of particles in the ocean, Deep Sea Res.,
22, 491 – 502.
Moum, J., M. Gregg, R. Lein, and M. Carr (1995), Comparison of turbulence kinetic energy dissipation rate estimates from two ocean microstructure profilers, J. Atmos. Ocean Technol., 12, 346 – 366.
Muller-Karger, F., et al. (2001), Annual cycle of primary production in the
Cariaco Basin: Response of upwelling and implications for vertical export, J. Geophys. Res., 106, 4527 – 4542.
Muller-Karger, F., R. Varela, R. Thunell, Y. Astor, H. Zhang, R. Luerssen,
and C. Hu (2004), Processes of coastal upwelling and carbon flux in the
Cariaco Basin, Deep Sea Res. II, 51, 927 – 943.
Peterson, L. C., G. H. Haug, K. A. Hughen, and U. Röhl (2000), Rapid
changes in the hydrologic cycle of the tropical Atlantic during the Last
Glacial, Science, 290, 1947 – 1951.
Pond, S., and G.L. Pickard (1989), Introductory Dynamical Oceanography,
Pergamon Press, Oxford.
Rasmussen, M. R., J. Laursen, S. R. Craig, and E. McLean (2005), Do fish
enhance tank mixing?, Aquaculture, 250, 162 – 174.
Richards, F. A., and R. F. Vaccaro (1956), The Cariaco Trench, an anaerobic basin in the Caribbean Sea, Deep Sea Res., 3, 214 – 228.
C03028
Scranton, M. I., F. L. Sayles, M. P. Bacon, and P. G. Brewer (1987),
Temporal changes in the hydrography and chemistry of the Cariaco
Trench, Deep Sea Res., 34, 945 – 963.
Scranton, M. I., Y. Astor, R. Bohrer, T.-Y. Ho, and F. Muller-Karger (2001),
Controls on temporal variability of the geochemistry of the deep Cariaco
Basin, Deep Sea Res. I, 48, 1605 – 1625.
Sternberg, R. W., I. Berhane, and A. S. Ogston (1999), Measurement of size
and settling velocity of suspended aggregates on the northern California
continental shelf, Mar. Geol., 154, 43 – 53.
Thunell, R. C., R. Varela, M. Llano, J. Collister, F. Muller-Karger, and
R. Bohrer (2000), Organic carbon fluxes, degradation, and accumulation in an anoxic basin: Sediment trap results from the Cariaco Basin,
Limnol. Oceanogr., 45, 300 – 308.
Torres, J. J. (1984), Relationship of oxygen consumption to swimming
speed in Euphausia pacifica. II: Drag, Efficiency and a comparison with
other swimming organisms, Mar. Biol., 78, 231 – 237.
Virmani, J. I. (2005), Ocean-atmosphere interactions on the West Florida
Shelf, Ph.D. dissertation, Univ. of South Florida, St. Petersburg, Fla.
Weihs, D. (1977), Effects of size on sustained swimming speeds of aquatic
organisms, in Scale Effects in Animal Locomotion, edited by T. J.
Pedley, pp. 299 – 313, Acad. Press, London.
Wilson, C. D., and E. Firing (1992), Sunrise swimmers bias acoustic
Doppler current profilers, Deep Sea Res., 39, 885 – 892.
Zhang, J.-Z., and F. J. Millero (1993), The chemistry of the anoxic waters in
the Cariaco Trench, Deep Sea Res. I, 40, 1023 – 1041.
J. I. Virmani and R. H. Weisberg, College of Marine Science, University
of South Florida, 140 Seventh Avenue South, St. Petersburg, FL 33701,
USA. ([email protected])
12 of 12