Docstoc

Estimation of zooplankton abundance from ADCP Backscatter at .doc

Document Sample
Estimation of zooplankton abundance from ADCP Backscatter at .doc Powered By Docstoc
					      Temporal Variability of Zooplankton Biomass from ADCP
  Backscatter Time Series Data at the Bermuda Testbed Mooring Site
           Songnian Jianga*, Tommy Dickeya, Laurence P. Madinb, Deborah K. Steinbergc
       a
         University of California at Santa Barbara, Santa Barbara, CA 93106-3060, USA
            b
              Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA
      c
        Virginia Institute of Marine Science, College of William and Mary, Gloucester Pt,
                                        VA 23062, USA

SONG - THE FIRST HALF LOOKS REALLY GOOD. WE NEED TO RETHINK THE
ORGANIZATION AND PRESENTATION OF THE SECOND HALF WHEN WE DESCIRBE
THE DATA AND EVENTS. I AM THINKING WE SHOULD NOT DO DESCRIPTIONS
SEPARATE FROM DISCUSSIONS OF ANALYSES. THEN IT WILL READ MORE
COHERENLTY I THINK. THIS WILL SHORTEN THE DISCUSSION SECTION AND IT
CAN BE USED TO DISTILL THE MAJOR FINDINGS OF THE STUDY. RIGHT NOW IT
SEEMS TOO DIFFUSE AND REPETITIVE.

MAYBE DEBBIE SHOULD TAKE A FRESH LOOK AT THIS AND DO AN EDIT AND
THEN WE CAN GO BACK AT IT FOR MORE DETAILED AND FINAL EDIT AND RE-
WRITE.

Abstract

Temporal variability of acoustically estimated zooplankton biomass at the Bermuda
Testbed Mooring (BTM) site is described for time scales from minutes to the seasonal
cycle using data obtained between August 1996 and November 2000. Concurrent high
frequency BTM observations of meteorological, physical, bio-optical, and chemical
variables facilitate interpretation of the processes contributing to the zooplankton
variability. Zooplankton biomass estimates are derived from regressions of backscatter
intensity data measured with an upward looking 153 kHz acoustic Doppler current
profiler (ADCP) deployed at about 210 m depth on the BTM (at 31º43‟N, 64º10‟W), and
zooplankton net tow data collected near the BTM site. Our data show clear event-scale
variations, with daily averaged biomass peaks in spring and late fall-winter. These peaks
are associated with annual spring blooms and late fall-winter events involving mixed
layer deepening and in some cases passages of mesoscale features. During periods of
biomass peaks, zooplankton biomass extended much deeper into the water column at
night. Biomass peaks are often concurrent with maxima seen in BTM chlorophyll
fluorescence measurements (inferred phytoplankton biomass). Averaged successive 14-
day estimates of zooplankton biomass and relative vertical velocity show the vertical
structure of daily migration patterns. Seasonal variations in migration patterns are also
evident. Results of a cross correlation analysis showed high correlations with 0-2 days lag
between zooplankton biomass and chlorophyll fluorescence, and zooplankton biomass
and temperature during spring bloom and eddies. Our high temporal resolution time
series of estimated zooplankton biomass provides information on scales inaccessible
through conventional monthly ship-based sampling and have implications for the vertical


                                               1
transport of carbon through the diel migration of zooplankton. In addition, recent
analyses of sediment flux trap data (3200 m depth) collected near the BTM show strong
correlations of large mass fluxes with greater zooplankton biomass levels.


Keywords: ADCP; Backscatter; Zooplankton; Biomass


*Correspondence author. Ocean Physics Laboratory, University of California at Santa Barbara,
Suite A, 6487 Calle Real, Goleta, CA 93117, USA.

1. Introduction

Knowledge of variability in zooplankton biomass is important for understanding the
effects of climate change on ecosystems. Also, migrating zooplankton contribute to
vertical fluxes of material. In particular, they can actively transport a significant amount
of dissolved inorganic carbon and nitrogen to the deep sea (Steinberg et al., 2000; Conte
et al., 2003). Zooplankton collections in the Bermuda region have been made since 1938
(Moore, 1949), thus providing rich data sets for studying zooplankton grazing,
metabolism, feeding, migration, species composition, and biomass on diel, seasonal and
interannual time scales. However, previous studies of zooplankton in the Bermuda region
have relied upon conventional ship-based sampling methods. These data sets could not
provide high temporal resolution information for extended periods of time and thus could
not capture important processes with time scales of minutes to a month. Importantly,
variability associated with episodic events and mesoscale features on time scales of less
than a few weeks cannot be resolved even with monthly or bi-weekly shipboard sampling
(Dickey et al., 1998, 2001).


Using the relationship between acoustic backscatter intensity from acoustic Doppler
current profilers (ADCPs) and zooplankton biomass from net-collected zooplankton
samples, ADCPs have been employed to describe variability in the distribution of
zooplankton biomass and zooplankton migration by many authors for over a decade (e.g.,
Flagg and Smith, 1989; Plueddemann and Pinkel, 1989; Heywood et al., 1991; Smith et
al., 1992; Fischer and Visbeck, 1993; Roe and Griffths, 1993; Ashjian et al., 1994; Flagg
et al., 1994; Lyons et at., 1994; Zhou et al., 1994; Batchelder et al., 1995; Heywood,



                                                   2
1996; Ashjian et al., 1998; Rippeth and Simpson, 1998; Zimmerman and Biggs, 1999;
Luo et al., 2000; Ashjian et al., 2002; Hitchcock et al., 2002). Many of these observations
have been made with ADCPs deployed on moorings, bottom mounts or floating
platforms. For example, Flagg and Smith (1989) were able to correlate signals of acoustic
backscatter from a bottom-mounted narrowband 307 kHz ADCP with abundance of
zooplankton collected with a MOCNESS net system at the edge of the New England
shelf. Later, Fischer and Visbeck (1993) used moored 153 kHz ADCPs to study diel
vertical migration (DVM) over a one-year period in the Greenland Sea and described the
daily vertical zooplankton migration and its seasonality. Using 15 months of continuous
ADCP and fluorometer data in the Mid-Atlantic Bight, Flagg et al. (1994) provided a
view of the seasonal progression of zooplankton biomass and its response to physical
forcing in a coastal setting. Ashjian et al. (1998) also studied the diel vertical migration of
zooplankton biomass in the Mid-Atlantic Bight using a moored ADCP. Rippeth and
Simpson (1998) used a bottom-mounted ADCP to study the diel vertical migration of
zooplankton at a position on the Hebridean continental shelf off England. Other examples
are presented in references cited above.


,

Reviewer #2 The paragraph on p. 3 of the Introduction reads like an
advertisement for ADCPs and can be eliminated. FINE TO OMIT THIS.


The Bermuda Testbed Mooring (BTM) program was initiated in 1994 and has served
three primary functions: 1) it provides the oceanographic community with a deep-water
platform for testing new instrumentation, 2) interdisciplinary BTM data are used for
scientific studies, particularly in conjunction with other research programs including the
Ocean Flux Program (OFP) and the U.S. JGOFS Bermuda Atlantic Time-series Study
(BATS) program and 3) nearly continuous bio-optical time-series data are obtained for
calibration, validation, and algorithm development for ocean color satellites (Dickey et
al., 1998a, 2001). BTM measurements typically include: surface meteorology and optics
along with subsurface currents, temperature, salinity, and bio-optical and chemical
variables. An RDI 153 kHz ADCP has been deployed from the BTM to obtain current
and acoustic backscatter intensity (for estimating zooplankton biomass) data since late


                                              3
August of 1996. These measurements provide important information concerning periodic
and episodic processes. The BTM enables collection of nearly continuous data during
periods of inclement weather when traditional sampling is not possible and provides
otherwise inaccessible data in the important temporal spectral range of minutes to several
months. The BTM has provided important information concerning periodic and episodic
processes for over 10 years (Dickey et al., 1998a, 2001). Previous studies based on BTM
measurements have concerned physical and biogeochemical variability (e.g., Dickey et
al., 1998a,b, 2001; McNeil et al., 1999; Zedler et al., 2002; Conte et al., 2003).
Observations of zooplankton biomass as estimated from an ADCP deployed on the BTM
provide concurrent and long-term biological information about zooplankton with high
vertical spatial and temporal resolution, and facilitate the study of physical and biological
interactions.



2. Methods

2.1 Data collection

The Bermuda Testbed Mooring (BTM; Dickey et al., 1998, 2001) is located
approximately 80 km southeast of Bermuda at 31°43‟N, 64°10‟W in water about 4550
m deep and has been deployed since June 1994 (Dickey et al., 1998a, 2001). The
mooring configuration for the present study included temperature sensors at several
depths from 2 to 750 m (3.75 min sampling interval), conductivity sensors (for salinity
determinations) at one or two depths (3.75 min sampling interval), and fluorometers for
estimating chlorophyll a at several depths (3.75 min sampling interval). An upward-
looking 153-kHz Blue Water broadband ADCP manufactured by RD Instruments, Inc. of
San Diego, CA (RDI 1995) was located at a depth of 204 ~ 212 m and provided
horizontal and vertical currents and acoustic backscatter intensity (for estimating
zooplankton biomass) from 18.5 to 203.5 m (Table 1). The ADCP is configured to collect
data in 68 bins, each with a vertical bin size of 3 m and at a sampling interval of 7.5 or 15
minutes. The ADCP beam angle is 20° from the system‟s vertical axis and the acoustic
transmission beam spreads at approximately 4°. This results in a 7.5-m-diameter bin for
each beam at 72-m depth, with horizontal surface area of 44.2 m2, and with a sampling


                                             4
volume for each beam of 133 m3 with a 3 m bin size (Gilboy et al., 2000). ADCP
measurements made from 21 August 1996 (Deployment 6) to 29 November 2000
(Deployment 14) are used in our study of diel, seasonal, interannual, and episodic
variation of zooplankton biomass. Time-series of the ADCP measurements are
summarized by deployment period in Table 1. The time-series are not continuous for
several reasons: (1) new instruments were added during the program, (2) data gaps of
several days exist between mooring recoveries and redeployments, and (3) occasional
delays in mooring redeployments were caused by weather and sea-state conditions and
ship-related problems.


To our knowledge, most previous ADCP zooplankton estimates have relied upon
narrowband ADCP‟s to measure acoustic backscatter. The narrowband system‟s
automatic gain control (AGC) output is strongly temperature dependent and many studies
have been unable to determine absolute backscatter since output signal strength was
unknown. Broadband ADCP‟s (BBADCP‟s) have the advantage over narrowbands of
having much lower random fluctuations for both current and backscatter data (Deines,
1999). The Received Signal Strength Indicator (RSSI) outputs of BBADCP‟s are not
temperature dependent; moreover, the ADCP manufacturer, RDI, has provided transmit
power for the BBADCP, which is required for absolute backscatter estimation. The
broadband ADCP gives higher resolution along the profile with little reduction of
velocity precision (Gilboy et al., 2000).


Zooplankton samples were collected at the BATS site (31°50‟N, 64°10‟W) with 1-m²

rectangular, 202-m mesh nets beginning in 1994 (Madin et al., 2001). Size-fractionated
biomass (wet and dry weight) is determined from each tow by wet sieving through nested
sieves with mesh sizes of 5.0, 2.0, 1.0, 0.5 and 0.2 mm. Two replicate double oblique
tows lasting approximately 30 min were made during the day (between about 0900 and
1500 h) and at night (between about 2000 and 0200 h) during cruises. Tows were made
through the mixed layer to a depth of approximately 200 m (Madin et al., 2001).

2.2 Estimation of zooplankton biomass from acoustic backscatter data



                                            5
The echo intensity (counts) recorded by the BBADCP can be converted to a backscatter
coefficient in decibels by use of an equation given by Deines (1999).

    S v  C  10 log 10 (( Tx  273 .16 ) R 2 )  LDBM  PDBW  2R  K c ( E  E r )   (1)



where Sv is the backscattering strength in dB re (4 m)-1, LDBW is 10 log10 (transmit pulse
length, meters), PDBW is 10 log10 (transmit power, Watts), Tx is temperature of the
transducer (C), R is range along the beam (slant range) to the scatters (m), and  is the
sound absorption coefficient of water (dB/m), as calculated using the method of Francois
and Garrison (1982). The factor 2R is calculated by the method from Deines (1999). C
is an empirical constant that is required to account for some of the relevant phenomena
affecting echo intensity that cannot be measured independently. Kc is the conversion
factor for echo intensity (dB/count). The values of C, PDBW and Kc are provided by the
manufacturer. To obtain absolute backscatter data, transmit power must be estimated. In
general, this power is proportional to the input voltage (Deines, 1999). Our 153 kHz
Broadband ADCP has a high power module (with constant power) that removes this
dependence upon input voltage. Echo intensity (E) is derived from the Received Signal
Strength Indicator (RSSI) of the receivers; its real-time reference level is denoted Er, a
typical value is 40 counts. In practice, we assume that Er is equivalent to the lowest value
of E obtained in the water column during the entire data collection period. There is no
laboratory calibration for Er: instead, we use the minimum of counts for each beam when
the ADCP is in the ocean. The Er values we used for each beam were 39, 37, 43, and 33
for beams 1 through 4, respectively. If backscatter intensity of one specified beam
exceeded the mean value of the other three beams by 5 db for a given bin, then data for
this bin were discarded for averaging. This process eliminates large scatterers (such as
fish) from the data. The minimum „percent good‟ of 25% is selected for quality control.
The low RSSI values for bin 1 in Dep. 6,7,10,11,12,13 and 14 may be due to the
hardware Low Pass Filter time constant being too long. The high bin 1 RSSI values of
Dep. 8 and 9 may be due to ringing or a hard target in front of the transducer (Steve
Maier, personal communication). So, bin 1 ADCP backscatter intensity data for all


                                                  6
deployments were omitted. The acoustic signal generally decreases with distance from a
transducer. But in the near surface layer, measurements are contaminated by the lobe
effect and air bubbles, and the signal becomes stronger. We calculated the mean profiles
of intensity and then determined the averaged cutoff bin. Typically, bins 61–63 and
shallower bin data were unusable. Thus, data within about 20 (18.5 – 21.5) m of the
ocean surface were omitted from our analysis. In some cases, there are a few bad data
(may caused by air bubbles - strong wind) at depths greater than that of the averaged
cutoff bin; in those cases the data bad are edited out and interpolation was employed.


DEBBIE – CAN YOU GIVE THE FOLLOWING A GOOD READ AND MAKE ANY
CHANGES THAT CLARIFY THIS? THANX!
The relationship between zooplankton biomass and acoustic backscatter intensity was
established by comparing biomass obtained from net tows and ensembled ADCP data.
Zooplankton biomass data used for comparison with ADCP data were the sum total dry
weight per unit volume (DW) filtered for all size fractions of the individual tows taken
from September 2, 1996 to November 14, 2000. ADCP backscatter intensity data were
averaged from 18.5 – 21.5 m to the depth closest to the deepest zooplankton sampling
depth for each of the concurrent zooplankton sampling time periods. We tried two ways
to calibrate: a) Separate calibration for each deployment. b) Overall calibration using all
the deployments data combined. Table 2 lists the calibration results. The deployments
with better r-square values (> 0.5) are: 6, 8, 10, 12, and 14, SONG – NOT SURE WHAT
YOU ARE TRYING TO SAY IN FOLLOWING and only deployment 10 with much
high difference on slope and intercept compare to others. The reasons for the
deployments with weak relationship may be: the locations of zooplankton net tow
sampling were further away from the BTM; sampling depths were not close to the depth
range of the ADCP; fewer sampling points were available during the BTM deployment
and there were likely zooplankton size distribution (or patch) differences during different
deployments (or seasons). The slope and the intercept for the overall calibration are
similar to the individual deployment time periods showing good correlations (better r-
square, and lower, significant P values). The correlation (regression) between
zooplankton biomass and the backscattered ADCP signal is represented by: Reviewer #3:



                                             7
P7, par 2 (also Fig. 1) - For the purpose of clarifying the
difficulties of calibrating ADCP measurements with net tows, it would
be useful (also important!) to show calibrations based on the data
results from deployments 7-9, as well as 6. It seems disingenuous to
say simply that Deployment 6 results were "better" without showing
exactly how bad the Deployments 7-9 results were. Most of the
subsequent data interpretations are from Deployments 7-9.




         Log(DW/4) = 0.0313*S + 2.0631           (r2 = 0.25, P <<0.05, n = 156)     (2)


The linear regression calibration curve (based on equation 2) is shown in Fig. 1. For the
overall calibration, there is a "relatively weak but highly significant relationship" between
zooplankton biomass from net tows and ADCP backscatter intensity data (weak because
the regression only explains 25% of the variation, significant because P << 0.05). The r-
square is low, which may be an argument for not using it, however, it is the slope and
intercept that really affect using the calibration. Thus, we chose the overall calibration
(equation 2) and applied it to all the data.


2.2 Statistical methods


The ADCP deployed depth varied somewhat with different deployments, top bin cut off
also varied with sea state and different deployments. For consistency, depth bins for
zooplankton biomass and backscatter intensity from the ADCP were interpolated to 22 –
190 m (or 21.5 – 192.5 m), except for calibrations between backscatter intensity and net
tow zooplankton biomass data. Spectra were computed for zooplankton biomass and
chlorophyll-a using 1024-point fast Fourier transforms (FFTs) tapered with a Hanning
window, zero overlap. The 95% confidence intervals were calculated for the spectra. To
calculate cross-correlations between zooplankton biomass and chlorophyll, high-
resolution time series of zooplankton biomass, chlorophyll and temperature daily
averages were computed and then a high-pass filter (30 days) was applied to remove low-
frequency variability. SONG – I DON‟T UNDERSTAND THE FOLLOWING
SENTENCE.        We used 61 days moving from the beginning of the deployment to




                                               8
breakdown each deployment data to calculate cross-correlations and run through whole
deployment to find significant correlations.


3. Results

3.1 Overview of processes and scales of variability


The BTM site is located at the northern edge of a transition region between relatively
eutrophic waters to the north and more oligotrophic subtropical waters to the south,
where the weak surface front and energetic sub-mesoscale and mesoscale features are
often present and can affect local biology (the term “mesoscale” refers to features with
horizontal scales on the order of 100 - 200 km that pass the mooring on a time scale of
roughly a month and “sub-mesoscale” as a features on the order of 10 – 100 km that
passes the mooring on a time scale of less than a month, Dickey, et. al., 2000). The region
of the BTM site (and the Bermuda Atlantic Time Series, BATS, site, and Ocean Flux
Study, OFP, site; note all of these sites are within a few tens of kilometers of each other)
is dominated by the seasonal cycle (e.g., Michaels and Knap, 1996; Dickey et al., 1998a,
2001). The mixed layer depth and phytoplankton concentrations vary seasonally, but the
respective timing and intensities vary interannually. Synoptic scale weather patterns
typically pass every few days. In addition 15 tropical storms or hurricanes have tracked
through the vicinity of the BTM since 1995 and 153 since 1851. Several accounts of the
physical and biogeochemical variability measured at the BTM site during the past few
years are presented in Dickey et al. (1998a,b, 2001), McGillicuddy et al. (1998), McNeil
et al. (1999), Zedler et al. (2002), and Conte et al. (2003).


Time-series studies of zooplankton seasonal dynamics at the BATS site date back several
decades (Menzel and Ryther, 1961b;Deevey, 1971;Deevey and Brooks, 1977), as do
other zooplankton studies (e.g., Moore, 1949, 1950; Sutcliff, 1960; Beers, 1966). More
recently, the ZOOSWAT project compared zooplankton during the spring-bloom period
(March/April) with the summer period (August) (Roman et al.,1993,1995). The most
comprehensive time-series study of seasonal and interannual variation in zooplankton
biomass at the BATS site is presented by Madin et al. (2001). A maximum in


                                               9
zooplankton biomass occurs in March/April, generally following peaks in primary
production (after injection of nutrients from deep winter mixing), although there is
significant interannual variation. After shoaling of the thermocline and reduction of
phytoplankton stocks, zooplankton stocks decrease, although occasional summer
zooplankton biomass peaks occur (Madin et al., 2001). Zooplankton biomass is reported
to be as much as 4-6 times greater (Deevey, 1971; Menzel and Ryther (1961b) – for
upper 500 m), and more than 3 times greater (Roman et al., 1993; Madin et al., 2001; for
upper 200 m) in March/April compared to August. Roman et al. (1995) report that micro-
, meso-, and acrozooplankton together average about 30% of the integrated heterotrophic
carbon biomass at the BATS site in spring and summer. While the macrozooplankton
biomass was more than 3-fold higher in spring than summer, the mesozooplankton
biomass was approximately 3-fold higher in summer than spring (Roman et al., 1993).
Thus, smaller zooplankton dominate in the summer, and may be more efficient at
utilizing the small phytoplankton and protozoa that also dominate that time of year
(Roman et al., 1993). Average estimated flux due to macrozooplankton fecal pellets at the
BATS site in the spring was 65% of the carbon flux at 150 m (estimated from a single
cruise, Roman et al., 1993). Thus, while macrozooplankton are only a small part of the
biomass, they likely contribute significantly to carbon flux from the euphotic zone by
production of rapidly sinking fecal pellets (Fowler and Knauer, 1986; Small et al., 1989;
Altabet and Small, 1990).


Another mechanism by which zooplankton contribute to carbon and nutrient cycling and
at BATS is via active transport of biogenic material by vertical migration. Zooplankton
biomass in the upper 200 m at the BATS site on average nearly doubles at night due to
vertically migrating zooplankton (Madin et al., 2001). As elsewhere, studies in the
Sargasso Sea show that vertically migrating zooplankton can actively transport a
substantial amount of dissolved inorganic carbon and nitrogen to deep water (via
respiration and excretion), which can be significant relative to the passive flux of sinking
particulate organic matter measured with sediment traps (Longhurst et al., 1989, 1990;
Dam et al., 1995; Steinberg et al., 2001). Estimates of active transport due to vertical
migration in these studies range from 8 to 70% of the sinking POC flux and from 8% to



                                            10
more than 80% of the sinking PON flux at 150 m. Steinberg et al. (2001) demonstrated
that active transport of dissolved organic carbon by migrators is also significant at the
BATS site. They suggest that during most of the year when deep mixing does not occur,
diel migration by zooplankton makes a significant supply of DOC available for use by the
microbial community at depth. New estimates of active transport by migrants may help
resolve observed imbalances in the carbon budget at the BATS site (Michaels et al.,
1994a), but the magnitude depends highly on the biomass of the migrating community
(Debbie???? Steinberg???, 2001). In addition, recent analyses of sediment flux trap data
(3200 m depth) collected near the BTM site show strong correlations of large mass fluxes
with greater zooplankton biomass levels (REFERENCE FOR THIS?                 A CONTE
PAPER???).


(blue color here most I got from Debbie’s paper, 2001, we should rewrite) READS
WELL!! IT IS A GREAT ADDITION!!


Reviewer #3:


Section 2.1 redundancy --begins with information previously
given in Introduction.



PAPER IS IN GGOD SHAPE TO THJIS POINT – FOLLOWING PARTS NEED
SOME WORK – MORE DETAILS AND REFINEMENT – DEBBIE SHOULD BE
ABLE TO HELP. I NEED TO GO THROUGH THIS AGAIN ALSO. MAYBE
PASS WHOLE THING TO DEBBIE FOR A READ AND THEN I’LL GO
THROUGH THIS NEXT PART REALLY CAREFULLY.

3.2 Seasonal Results

Daily averaged, depth integrated (22 – 190 m) zooplankton biomass estimated using
ADCP data along with concurrent temperature, chlorophyll fluorescence, and current
measurements from late August 1996 through late November 2000 are shown in Fig. 2.
During this period, depth integrated daily zooplankton biomass fluctuated from 343 mg
m-2 to 700 mg m-2 with mean value of 497 mg m-2 (Table 3).



                                           11
MAYBE WE SHOULD ADD A COLUMN TO TABLE 3 TO GIVE DATES OF
DEPLOYMENTS AGAIN TO HELP READERS.
Table 3. Means, minima, maxima, and standard deviations of daily mean, depth-
integrated biomass (mg/m2) during BTM deployments 6-14
______________________________________________________________________________________

Deployment No.     Mean        Minimum          Standard Deviation
                                                Maximum
_____________________________________________________________________
      6            502           404           648             54
      7            505           377           700             58
      8            502           424           568             29
      9            541           385           688             64
     10            507           405           647             49
     11            481           387           618             51
     12            483           434           565             25
     13            480           343           676             69
     14            483           385           633             55
    6-14            497          343           700             57
__________________________________________________________________________________



Table 4. Seasonality of means, minima, maxima, and standard deviations of daily mean,
depth-integrated biomass (mg/m2) during BTM deployments 6-14
______________________________________________________________________________________

Season         Mean      Minimum    Maximum    Standard Deviation
________________________________________________________________________

Jan–Mar            471           343             700            62

Apr–Jun            503          440              633            36

Jul–Sep            486           385             648            51

Oct–Dec            497          343              700            57




Our data show strong event-scale variations (or fluctuation), with depth integrated (22 –
190 m) daily zooplankton biomass. Also evident are peaks in zooplankton biomass in
spring, late fall-winter and some in summer. These peaks are associated with annual
spring blooms (700 mg m-2, May, 1997; 568 mg m-2, maybe March, 1998 (increased in
zoo. Biomass, but no Chl data); 688 mg m-2, February-March, 1999); 676 mg m-2, maybe


                                           12
March, 2000 (increased in zoo. Biomass, but no Chl data) and late fall-winter bloom
events involving mixed layer deepening (648 mg m-2, December, 1996, during a fall
bloom - eddy period) and in some cases passages of mesoscale features (June, 1999; July,
2000). Some of the biomass peaks and trends determined from ADCP data are consistent
with concurrent net tow measurements by Madin et al. (2001) (dates). During periods of
ADCP-estimated zooplankton biomass peaks, zooplankton biomass with relatively higher
values (% of biomass at 22 m, DEBBIE AND LARRY - we should find the explain based
on reviewer‟s question) extended much deeper (~180 - 190 m) into the water column
(Fig. 3).
Reviewer #3:
P8, par 2 (also middle of P9) - The notion that higher zoopl biomass
values extend deeper during biomass peaks also needs to be developed or
explained more rigorously. First of all, is the observation true
statistically? Second, does this mean that deeper zooplankton values
increase disproportionately during the peaks, or does everything just
go up by the same amount? At least my eyes do not do a good job of
pulling numbers and relative distributions out of the colored contours.




3.3 Spring Bloom
Time series of chl-a fluorescence at 45 and 77 m show strong evidence of a spring bloom
in the mid-May to the end of May, 1997 (Fig. 2). The chl-a fluorescence started to rapidly
increase around May 15, 1997. During the period of the spring bloom, ADCP-estimated
zooplankton biomass increased significantly and reached its maxmia 700 mg m-2 in mid-
May. The biomass with relatively high values (greater than 6 mg/m3) extended to ~ 110 –
120 m during nights of May 17 – 19, 1997 (Fig. 3) During the period of March 3 – March
22, 1998, there were significant increases in zooplankton biomass extending much deeper
at night. On March 19, 1998, zooplankton biomass with relatively higher values went
down to ~ 190 m. These biomass increases may relate to the early spring bloom and are
similar to the biomass increases during the spring bloom that occurred in May 1997.
Time series of chl-a fluorescence at 72 m show strong evidence of a spring bloom from
late February to March 1999. The chl-a fluorescence started to rapidly increase around
February 16 and was more than doubled in magnitude on February 24. This bloom was



                                           13
mainly due to the onset of stratification that following deep mixing, which penetrated as
deeply as 100 m depth as shown in the stack plot of temperature. Around March 1, 2000,
there was a strong peak in zooplankton biomass, which showed the signal of a spring
bloom, but it is not seen in chlorophyll as no chl-a fluorescence measurements were
available.




3.4 Eddies


Mesoscale features previously have been suggested as playing an important role for new
production and phytoplankton dynamics near Bermuda (e.g., Jenkins, 1988). The
influence of mesoscale features was estimated by McGillicuddy et al. (1998) using BTM
data set, shipboard observations, and a regional eddy-resolving model. The conclusion of
McGillicuddy et al. (1998) suggests that the flux of nutrients induced by mesoscale
eddies may be sufficient to balance the nutrient budget of the Sargasso Sea.
A. Warm eddy
In late November and December 1996 (Dep. 6), the BTM sampling showed evidence of
a fall phytoplankton bloom due to passage of a warm mesoscale feature. The feature was
characterized by SONG WARM EDDY GIVES ANTI-CYLCLONIC ROTATION –
PLEASE CHECK anti-cyclonic rotation of currents and a thick, warm, low salinity
isothermal layer >180 m in depth. During the period of October 26,1996-January 4,1997,
the physical conditions were characterized by relatively high wind stress (Dickey et al.,
2001), monotonic cooling of near-surface waters, highly variable temperature at depths of
7-154 m with major warming (by ~ 4○ C) first at 71 m and about a month later at 154 m,
deepening of the mixed layer from about 50 to >175 m, directional changes in currents
and currents with magnitudes of up to 50 cm/s. Within the feature, sensors on the BTM
mooring recorded roughly two-fold increases in chlorophyll. The Oceanic Flux Program
(OFP) sediment traps recorded an abrupt, factor of 2.5 increases in mass flux at 3200 m
depth (Conte et. al, 2003). The daily averaged ADCP estimates of 22-190 m integrated
zooplankton biomass increased significantly when the feature arrived at the BTM site and
peaked (648 mg m-2) SHOULD BE MG/M2, RIGHT? I PUT IN A MINUS SIGN IN



                                            14
FRONT OF THE 2 HERE AND ELSEWHERE WHEN I NOTICED IT on December 13
- 14 when the central portion of the feature appeared to be nearly directly over the
mooring (Fig. 2). This was confirmed by zooplankton net tow data. Within the feature,
integrated zooplankton biomass in the upper 200 m averaged 785 mg m-2 at midday and
1082 mg m-2 at night (Madin et al., 2001).

B. Cold eddy

During Deployment 10, current vectors show clear evidence of a strong mesoscale
cyclonic eddy from SONG – USE DATES RATHER THAN JD‟S IN PAPER –
READERS WILL BE TOO CONFUSED. JD 360 to JD 400. Stack plots of temperature
show that temperatures between depths of 150m and 250m decreased significantly during
this period, indicating a cold eddy with counter-clockwise current rotation. Surprisingly,
chl-a fluorescence and zooplankton biomass did not change significantly during the
passage of this strong eddy. One possible explanation for this is the lack of a shallow
mixed layer and a strong thermocline, which could have restricted injection of nutrients
into the surface layer and led to depletion of nutrient that were trapped in the mixed layer.
SONG – WHAT ABOUT OTHER COLD EDDIES?                         WE PROBABLY NEED TO
EXPAND ON THIS DISCUSSION A BIT.



3.5 wind events

A. Deployment 6

During the first half of deployment 6 (GIVE DATES), three hurricane, Edouard (August
31), Hortense (September 13), and Lili (October 20), passed into the region of the BTM.
All three hurricanes are evident in time series of winds. However, the hurricanes
generated oceanic responses of different degrees. Hurricane Lili excited the strongest
oceanic signals at the BTM location because its center was only 150 km east of the site,
while the response to Hurricane Hortense was weaker but still visible in relative
humidity, short wave radiation, radiance and irradiance. Hurricane Edouard did not
generate significant responses. When Hurricane Lili passed, barometric pressure dropped
about 30 mBar from GIVE DATES JD 291 to JD 295, the greatest changes in the entire



                                             15
time series. This change was accompanied with high relative humidity, and low short
wave radiation. This hurricane also excited strong inertial oscillation in the upper ocean,
as shown in horizontal currents. Chlorophyll flourescence levels were generally low at 45
and 73 m. Zooplankton biomass was also low. (I got from data report, we should rewrite)


B. Deployment 12

A major event captured by the mooring during Deployment 12 (DATES) was the ocean‟s
response to Hurricane Gert, which passed by on September 21, 1999. During this time,
winds reached maximum values of ~18 m/s at 4.2 m (20.1 m/s at 10 m) and barometric
pressure decreased from a nominal value of 1015 mb to 985 mb. The ocean responded
with strong near-inertial currents in the upper 40 to 50 m of water (of amplitude 30-50
cm/s at 35m), a >1○ C cooling at the sea surface, and an increase in mixed layer depth
from ~20 m to 45 m. Curiously, the cooling response seems to disappear on GIVE DATE
JD 275 as abruptly as it appeared on GIVE DATE JD 264. The current response seems to
persist for a longer period of time, with decay of the near-inertial signal after GIVE
DATE JD 280. Warm surface temperatures and strong stratification of the upper ocean
persist through approximately GIVE DATE JD 264, when this behavior is interrupted by
the passage of Hurricane Gert on JD 264. On JD 272, the sea surface temperature
                           ○
abruptly increased by >1       C. Chlorophyll flourescence levels were generally low at 35
and 74 m. Zooplankton biomass was also low. (I got from data report, we should rewrite)




3.6 Diel vertical migration

Diel vertical migration is a well-known behavior of zooplankton. Vertically migrating
zooplankton play an important role as they contribute to vertical fluxes of material and
can actively transport a significant amount of dissolved inorganic carbon and nitrogen to
the deep sea (Steinberg et al., 2000; Conte et al., 2003). In particular, diel variation is
important in low biomass regions such as the Sargasso Sea (Ashjian et al., 1994; Madin
et al., 2001).




                                              16
The ADCP measures the speed of scattering particles suspended in water, rather than the
water itself. If vertical water velocities are small, the vertical velocity measured by the
ADCP can be interpreted as the vertical swimming velocity of the scattering organisms
(Heywood, 1996; Luo et al., 2000). Because of surface waves, the mooring buoy moves
vertically, and the vertical velocity measured by the ADCP is not truly an absolute
velocity of scattering organisms. To ascertain migration velocity, we define one bin (or in
some cases the averages of several bins) near the surface for reference and then subtract
its value from each bin in the same profile (Luo, 2000). This approach minimizes biases
introduced by vertical buoy motions.



A. Seasonal changes of diel vertical migration patterns from spring, 1997 through
winter, 1998

The daily cycle of zooplankton biomass and vertical movement velocity (relative to 34–
43 m depth) computed using an ensemble average of 14-day time series of data (Fig. 4)
show zooplankton biomass vertical distributions and daily migration patterns in the
spring (May 16–29, 1997), summer (August 16–29, 1997), fall (November 1–14, 1997)
and winter seasons (February 10–23, 1998) using data from Deployments 7, 8, and 9
GIVE DATES.

During the spring, biomass is high and decreases with depth; during the night, the highest
biomass has a value of SONG – SOME PLACES WE USE INTEGRATED BIOMASS
AND OTHERS RGULAR BIOMASS CONCENTRATIONS – WE NEED BE VERY
CLEAR ON THIS 5.7 mg m-3 at 22 m and biomass greater than 5 mg m-3 extend to 94 m
sometime. Interestingly, there is relatively high biomass (> 4.8 mg m-3 thin layer (22 – 31
m) near the surface from 8:00 to 17:00. Relative vertical velocity contour plots show
strong diel migration signals. Downward velocity has a maximum value of 5.4 cm s-1 at
163 m 4:45 during dawn, while upward velocity has a maximum value of 5.0 cm s -1 at
169 m 19:45 during dusk. Downward velocities have large values at depths greater than
about 100 m, while upward velocities have large values at depths greater than about 120
m. In the summer, biomass values are smaller than those in the spring. Diel migration
signals are clear, but migration speeds are slower than in the spring period. In the fall,



                                            17
there is a persistent subsurface layer with high biomass (greater than 5 mg m-3) between
~60 m and ~80 m, and the starting time of downward migration is later than that in
summer and the starting time of upward migration is earlier than that in summer. In the
winter, biomass values are small. Differences between biomass during night and day are
relatively small, but diel migration is still quite evident (Fig. 4, last panel).


The estimated maximum relative vertical velocity computed using an ensemble average
of 14-day time series of data occurred in spring with a value of 5.4 cm s-1. This value is
comparable to results presented by other researchers: 1-4 cm s-1 (Plueddemann and
Pinkel, 1989), 3-8 cm s-1 (Smith et al., 1989), 2-4 cm s-1 (Roe and Griffiths, 1993), 2-6
cm s-1 (Heywood, 1996), and maximum of 10 cm s-1 (Luo et al., 2000).


B. Seasonal progression on timing of diel migration


Averaged values for backscatter intensity and relative vertical velocity (relative to 34–43
m depth) for the depth interval between about 121 and 190 m from May 4, 1997 through
March 30, 1998 show seasonal variations of the diel migration (Fig. 5). The top panel
shows depth averaged backscatter intensity. The bottom panel shows depth averaged
relative vertical velocity; negative values indicate downward motion and positive values
represent upward motion; black lines represent sunrise and sunset times (in hour). It is
easy to see from both panels that there is a consistent diurnal migration pattern for all
seasons. Also, the times of backscatter intensity decrease (and downward movements) at
dawn and the times of backscatter intensity increase (and upward movements) at dusk
change seasonally with the seasonal light cycle progression.


Fig. 6 shows the timing of peak downward and upward migration velocity relative to the
times of sunrise and sunset. The difference in minutes between the time of the peak
vertical velocity within the periods 4:00–7:00 and 17:00–20:00 and the time of sunrise or
sunset was calculated. Negative values indicate that migration began before sunrise or
sunset. Data used for Fig. 6 were from May 4, 1997 through March 30, 1998. Backscatter




                                               18
intensity begins to decrease before sunrise and begins to increase after sunset (Fig. 6, top
panel). Downward movements begin before sunrise and upward movements begin after
sunset (Fig. 6, bottom panel). Similar results have been reported by Harrison et al.
(2001). The time differences between sunrise and the decrease (and downward
movements) of backscatter intensity, and between sunset and the increase (and upward
movements) of backscatter intensity, have values of –40 and 30 minutes with maximum
number of observation, respectively.



4. Discussion and Conclusions

Seasonal zooplankton biomass peaks are generally consistent with maxima seen in concurrent BTM
chlorophyll fluorescence measurements, which are proxies for phytoplankton biomass. During the fall
bloom eddy period of late November - December, 1996, significant increases were observed in chlorophyll
fluorescence at 45 m and ADCP-based zooplankton biomass integrated from ~ 22 to ~ 190 m (Fig. 2).
P13, line 5 - Please support this statement with correlation plot of
zooplankton        biomass      vs    concurrent        chlorophyll        fluorescence.Is          the
relationship statistically significant? Recent analyses of sediment flux trap data (3200
m depth) collected near the BTM site also show strong correlations of large mass fluxes with greater
zooplankton biomass levels (Conte et al., 2003). In May 1997, relatively large values of chlorophyll
fluorescence were evident at 45 m and 77 m, respectively, and the highest daily mean of depth integrated
zooplankton biomass was observed (Fig. 2). Zooplankton biomass with values greater than 6 mg/m3 went
down to ~120 m depth in May 18, 1997 (Fig. 3). The time-series data of chlorophyll fluorescence at 72 m
show strong evidence of an early spring bloom in late February - March 1999. The time-series data of
zooplankton biomass also show biomass increases (Fig. 2). Chlorophyll fluorescence at 24 m and depth
integrated zooplankton biomass time-series data both reveal a peak, respectively, in mid-April 1999. (may
need rewrite)




4.1 Relationship among zooplankton biomass, chlorophyll fluorescence and
temperature.

A comparison of zooplankton biomass with primary production, phytoplankton standing
stock at the BATS site indicated significant positive correlations with production and
phytoplankton stock (Madin et al., 2001). In most years the lower spring
temperatures coincide with high zooplankton stocks, and high late summer and
fall temperatures with lower zooplankton biomass. The general pattern of spring
and summer increase is seen in both phytoplankton and zooplankton biomass. A



                                                   19
lag in zooplankton biomass increase following a production peak is apparent in
some places in the time series (e.g., March/April 1995, February/March 1998).
There are also related peaks in spring and summer (April/May 1996, July 1996),
and in fall and early winter (e.g. October 1994, October to December 1995,
December 1996, December 1997), but strong maxima of production are not
always associated with increases in zooplankton biomass (e.g. December 1996
to March 1997). The same is generally true of the relationship between Chl a and
zooplankton biomass, but perhaps with fewer coincident peaks than with primary
production. A broad peak in integrated Chl a biomass occurs in February/March
of each year (except 1997), with shorter increases occurring in late spring or
summer. Most peaks in zooplankton biomass occur after this February/March
increase.

We used a cross correlation analysis described in Steinberg et al. (2001) to
determine if apparent lags were statistically significant. All parameters were
linearly interpolated on regular 1-week or 1-month time scales and cross
correlation analysis was performed on this data set. The results indicated that
zooplankton biomass followed production with a significant positive correlation
with a 1-week up to a 1-month lag (r2=0.21, 0.17,0.15,and 0.11 for 1-week, 2-
week, 3-week, and 1-month imposed lags, respectively, n=68,p>0.05). No
significant positive relationship was seen for zooplankton biomass following Chl
a, although the highest correlation occurred with a 1-month imposed lag (r2=0.09,
n=68,p>0.05). In addition, no significant positive relationship was seen for flux
following zooplankton biomass, although the highest correlation occurred with a
2-week imposed lag (r2=0.09, n =68, p>0.05). (Madin et al., 2001).

(blue color here most I got from Madin’s paper, 2001, we should rewrite)

Madin et al. (2001) pointed out that there is a significant positive correlation between
zooplankton biomass and phytoplankton standing stock at the BATS site. The general
pattern of spring and summer increase is seen in both phytoplankton and zooplankton
biomass. A broad peak in integrated Chl a biomass occurs in February/March of each
year (except 1997), with shorter increases occurring in late spring or summer. Most peaks
in zooplankton biomass occur after this February/March increase. In most years the lower
spring temperatures coincide with high zooplankton stocks, and high late summer and fall
temperatures with lower zooplankton biomass. They also used a cross correlation analysis
described in Steinberg et al. (2001) to determine if apparent lags were statistically
significant. All parameters were linearly interpolated on regular 1-week or 1-month time
scales and cross correlation analysis was performed on their data set. Their results
showed that no significant positive relationship was seen for zooplankton biomass



                                           20
following Chl a, although the highest correlation occurred with a 1-month imposed lag
(r2=0.09, n=68, p>0.05).


We used a cross correlation analysis with 1-day AVERAGED? zooplankton biomass
estimated from ADCP backscatter intensity data, chl-a and temperature to determine if
apparent lags were statistically significant. We used 61-day moving datasets instead of
whole deployment data to calculate cross-correlations (method described in 2.2) and to
find significant correlations during particular events such as spring blooms, eddies and
wind events. Fig. 7 through 9 show the results of relationship among zooplankton
biomass, chlorophyll fluorescence and temperature with high (or highest) cross
correlations during some cases of event.


A. Spring blooms
Fig. 7a shows zooplankton biomass versus chlorophyll fluorescence during the spring
bloom in 1997; both time series data are filtered with 30 days high-pass filter.
Zooplankton biomass and chlorophyll fluorescence matched very well during the period
of day GIVE DATES 143 – 172.               Fig. 7a shows the cross correlation between
zooplankton biomass (76 m) and chlorophyll fluorescence (77 m) is positive with the
maximum correlation of 0.89 at 0 lag. There was an inverse relation between zooplankton
biomass and temperature with maximum correlation of 0.79 at 0 lag (Fig. 7d).
More than one month after spring bloom in 1998, during the period of GIVE DATES day
120 – 149, just during the starting stratification, zooplankton biomass (25 m) is very low
and chlorophyll fluorescence (24 m) is also related low, we found very high cross
correlation between zooplankton biomass and chlorophyll fluorescence with the
maximum correlation of 0.89 at 0 lag (Fig. 7f). Zooplankton biomass versus temperature
was also inverse relation with maximum correlation of 0.86 at 0 lag.


B. Eddies
In late November and December 1996 (Dep. 6), a warm mesoscale feature passed the
BTM. Zooplankton biomass (46 m) and chlorophyll fluorescence (45 M) matched well
with time lag of 2 days (Fig. 8a). The cross correlation between zooplankton biomass and


                                             21
chlorophyll fluorescence is positive with the maximum correlation of 0.87 at -2 days lag
(Fig 9b). There was a positive relation between zooplankton biomass and temperature
with maximum correlation of 0.51 at 1 day lag (Fig. 8c).
During Deployment 10 GIVE DATES, physical measurements show clear evidence a
strong mesoscale cyclonic eddy (a cold eddy) from GIVE DATES JD 360 to JD 400. But
chl-a fluorescence and zooplankton biomass did not change significantly during the
passage of this strong eddy. The cross correlation between zooplankton biomass (73 m)
and chlorophyll fluorescence (72 m) calculated from day 367-396 is positive with the
maximum correlation of 0.79 at 0 lag (Fig. 8e). Zooplankton biomass versus temperature
was also inverse relation with maximum correlation of 0.70 at -2 days lag.

C. Wind events
Fig. 9 shows cross correlations between zooplankton biomass and chlorophyll
fluorescence, and zooplankton biomass and temperature for periods when hurricane
Hortense (September 13, 1996) and Hurricane Gert (September 21, 1999) passed by the
BTM. Cross correlation were not high compared to those for events of spring blooms and
eddies except for zooplankton biomass (46 m) versus chlorophyll fluorescence (45 m)
with the maximum correlation of 0.73 at –2 days lag (Fig. 9c). This may be due to the
very short periods (few days) for hurricane pass by.


4.2 Spectral analysis of zooplankton biomass, chlorophyll fluorescence and
temperature.
Spectral analysis of zooplankton biomass time series was presented for the first time
during the SEEP-II program in the southern Mid-Atlantic Bight by Flagg et al (1994).
They concluded: biomass variability decreased toward the bottom. In contrast to the
phytoplankton spectra, only the diurnal frequency band showed a consistent peak. The
diurnal peak was prominent near the surface and bottom, and significantly less so in the
middle of the water column due to the vertical migration of at least some portion of the
zooplankton population as they swam upward at dusk and downward at dawn. The semi-
diurnal peak was significant in the upper portion of the water column during summer




                                            22
when zooplankton were concentrated in and above the thermocline, and were subject to
vertical advection by internal tides.
(I got from Flagg’s paper, 1994, we should rewrite)

Zooplankton biomass spectra estimated from ADCP data are shown in Figure 10 for
depths of 45, 75, 150 and 190 m for Deployments 7, 8, and 9 GIVE DATES to allow
comparisons with respect to both depth and season. The current speed spectra, the
temperature autospectra, and the chlorophyll-a autospectra for the BTM site data for
GIVE DATES Deployments 7, 8, and 9 are discussed in detail in Dickey et al. (2001), so
they will be summarized briefly. We re-calculated the current speed spectra at 45, 75, 150
and 190 m. The current speed spectra at shallow depths show a pronounced peak in the
vicinity of the inertial frequency (the inertial period at the BTM site is 22.8 h) and the
peak is weaker below 150 m (Fig. 10). There were no semi-diurnal signals at all depths
for Deployments 9 GIVE DATES.

Energy in the near-inertial and sub-inertial frequency ranges generally decreases with
depth, especially below the mixed layer. The semi-diurnal tidal energy is typically an
order of magnitude smaller than that associated with the near-inertial currents. The
temperature autospectra (Fig. 10b in Dickey et al., 2001) show shapes similar to the
current speed autospectra in the sub-inertial range. Diurnal temperature signals in the
near surface during Deployments 7 and 8 GIVE DATES are apparent, but not for
Deployment 9 because the mixed layer is deep (100–150 m for Deployment 9 GIVE
DATES). The diurnal peak of chlorophyll-a is evident in the autospectra for
Deployments 8 and 9, but not for Deployment 7 GIVE DATES.

Strong diurnal zooplankton biomass signals in the upper 150 m are quite similar for
Deployments 7–9 and are linked to the diurnal light cycle (Fig. 11); Energy in diurnal
frequency generally decreases with depth, the maximum energy was over 100 (mg m-
3 2
 ) cpd-1 at 45 m for Deployment 8 GIVE DATES. The semi-diurnal peak is significant in
the upper 150 m and likely reflects internal tides; relatively strong semi-diurnal
zooplankton biomass signals in the upper 150 m are evident except at 45 m and 75 m for
Deployment 9 GIVE DATES. Interestingly, there are 6 and 8 hours oscillation in most



                                           23
cases for the upper 150 m. This may be linked to reverse migration during the day and
night or this could also result from a sudden influx of migrating animals into ~ 20 – ~ 200
m depth interval. Reverse migrations (i.e., zooplankton coming up during day have been
observed for cases when non-visual invertebrate predators appeared near the surface at
night (Madin et al., 1996). Further studies of diel migration need to be done. Not
unexpectedly, diurnal, semi-diurnal, 6 and 8 hours peaks are much weaker at 190 m.




4.3 ADCP performance for estimation of zooplankton biomass

Generally, ADCPs measure objects in the size range of millimeters to a few centimeters.
The minimum length of organisms detected by a 153 kHz ADCP has been reported to be
about 1 mm (Luo et al., 2000). The most likely acoustic scatterers to be detected by a 153
kHz ADCP are copepods, euphausiids and amphipods (Fischer and Visbeck, 1993; Roe et
al., 1993; Heywood, 1996; Rippeth et al., 1998). For our Bermuda site, Madin et al.
(2001) found that the smaller size classes (0.2-1.0 mm) constituted the highest proportion
of the biomass during the day and the increase in nighttime biomass is mostly due to
increases in the biomass of larger size organisms (84% of the migrating biomass > 1.0
mm in the upper 200 m, and 56% > 2.0 mm). Madin et al. (2001) and Steinberg et al.
(2000) observed that the dominant migratory species at BATS include some of the large
copepods such as Pleuromamma xiphias (3.5-6.5 mm), P. abdominalis (2-4 mm), and
Euchirella messinensis (4-6 mm); other dominant species of migrators include
euphausiids such as Thysanopoda aequalis 12-22 mm), and Nematobrachion flexipes (20-
23 mm). Other common migratory organisms include the hyperiid amphipods
Anchylomera blossevillei (6-10 mm) and Scina spp (10-20 mm), and several species of
migrating sergestid shrimps including Sergia splendens (10-12 mm), Sergestes atlanticus
(5-8 mm), and Sergestes vigilax (30 mm) (Steinberg et al., 2000).


Backscatter intensities below 150 m depth measured by our ADCP are fairly low as
compared to the top layer. Below ~180 m, backscatter intensities decrease dramatically
with depth (Fig. 12a, averaged during Deployment 7). There may be several possible
explanations for this:


                                            24
   1) Previous observations of zooplankton in the Sargasso Sea show that the
       zooplankton biomass below ~150-200 m is lower than that in the top layers
       (Ortner et al., 1978; Roman et al., 1993); High abundances of some copepods
       were sampled above 150 m at night and deeper in the daytime in the northern
       Sargasso Sea (Longhurst et al., 1989); the percent good data from our ADCP
       decreased with depth (with value less than 95% below ~140 m) (Fig. 12b, percent
       good averaged during Deployment 7), there is a minimum value of 82.8% at
       about 189 m. Lower percent good data indicated lower backscattering signals at
       deeper depths;
   2) The geometry of beams cause the ensonified volume of water near the transducers
       to be relatively smaller, and the probability of scattering organisms getting into
       the volume is lower in a volume of low backscatter background (Gwyn Griffiths,
       personal communication);
   3) Our ADCP sampling intervals are 7.5 (Deployment 6) and 15 min for other
       deployments. With a migratory speed of 5 cm/s, the depth range of a migrator
       moving during the sampling intervals would be 22.5 and 45 m. Also, our ping
       intervals are 36s (Dep. 6) and 73 s for other deployments, so the depth ranges of
       migrators moving during the ping intervals would be 1.8 and 3.7 m. Any specified
       scattering organism may be in different depth bins for different pings. Long ping
       intervals and sampling intervals may explain in part why we cannot clearly see
       migration at a deeper depth;


The separation distance between the BTM site and BATS stations where net tow data
were collected (used for calibrating our ADCP backscatter data) may cause errors in the
zooplankton biomass calibration because of the patchiness of zooplankton and
decorrelation scales. For example, the mean distance between BTM and net tow sampling
sites (BATS stations) for the time period of net tow samplings used for calibration was
6.0 km, ranging between 3.4 and 7.5 km for Deployment 6. Although there may have
been patchiness of zooplankton biomass in this region, the distribution of zooplankton
biomass is less variable, or patchy in the Sargasso Sea than in the coastal or slope waters
(e.g., Ashjian et al., 1994). To achieve more accurate calibration for zooplankton



                                            25
estimation from ADCP backscatter intensity data, specified zooplankton samplings
within few hundred meters of BTM may be required in the future.

In conclusion, moored broadband ADCP‟s can be used in the oligotrophic ocean to
monitor zooplankton biomass variability and their vertical motion. The high temporal
resolution BTM time series of zooplankton biomass estimated from ADCP data provide
information on scales inaccessible through conventional monthly ship-based sampling.
Our data show clear event-scale variations, the zooplankton biomass peaks are associated
with annual spring blooms and late fall-winter events involving mixed layer deepening
and in some cases passages of mesoscale features. P17, line 9 - Again, the
statement about "clear seasonality" is not so clear to me.                     From Fig.
7, one could make a case for highest zooplankton biomass in June/July,
not spring and fall.          You may be right in the end, after staring at
your figures for much longer than I have been able to give them.                       But
if reasonable people can "see" very different conclusions in the same
illustrations, you have to do a much more rigorous job of making your
case.


There is a consistent diurnal cycle of zooplankton and diel variation is most important in
low biomass features. Seasonal variations in diel migration patterns are also evident.
Long-term BTM observations with acoustic zooplankton biomass measurements will
play an important role in understanding diel zooplankton migrator patterns and
zooplankton biomass variability with episodic events (e.g., eddy passages, wind mixing)
in the Sargasso Sea, and can facilitate studies of physical and biological interactions and
biogeochemical cycling and fluxes.

Acknowledgements


We thank RDI for providing parameters and Steve Maier for his recommendations on the
backscatter intensity calculation and data quality control, Gwyn Griffths for his valuable
suggestions, and Derek Manov, Grace Chang, Frank Spada and Will Black from OPL for
their significant contributions. Support for this research was provided by the National
Science Foundation Ocean Technology and Interdisciplinary Coordination and Chemical
Oceanography Programs (TD: OCE-9627281, OCE-9730471, OCE-9627277), the


                                            26
National Ocean Partnership Program (TD: N40000149810803), the Office of Naval
Research Ocean Engineering and Marine Systems Program (Dan Frye: N00014-96-1-
0028), and the University of California, Santa Barbara (to T. Dickey, UCSB).


References

Ashjian, C. J., Smith, S.L., Flagg, C.N., Mariano, A.J., Behrens, W.J. and Lane, P.V.Z.,
1994. The influence of a Gulf Stream meander on the distribution of zooplankton
biomass in the Slope Water, the Gulf Stream, and the Sargasso Sea, described using a
shipboard acoustic Doppler current profiler. Deep-Sea Research I, 41, 23-50.

Ashjian, C.J., Smith, S.L., Flagg, C.N. and Idrisi, N., 2002. Distribution, annual cycle,
and vertical migration of acoustically derived biomass in the Arabian Sea during 1994-
1995. Deep-Sea Research II, 49, 2377-2402.

Ashjian, C.J., Smith, S.L., Flagg, C.N. and Wilson, C., 1998. Patterns and occurrence of
diel vertical migration of zooplankton biomass in the Mid-Atlantic Bight described by an
acoustic Doppler current profiler, Continental Shelf Research, 18, 831-858.

Batchelder, H.P., Van Keuren, J.R., Vaillancourt, R. and Swift, E., 1995. Spatial and
temporal distributions of acoustically estimated zooplankton biomass near the Marine
Light-Mixed Layers station (59°30‟N, 21°00‟W) in the North Atlantic in May 1991,
Journal of Geophysical Research 100, 6549-6563.

Conte, M.H., Dickey, T.D., Weber, J.C., Johnson, R.J. and Knap, A.H., 2003. Transient
physical forcing of pulsed export of bioreactive material to the deep Sargasso Sea, Deep-
Sea Research I, 50(10-11), 1157-1187.

Deevey, G. B., 1971. The annual cycle in quantity and composition of the zooplankton of
the Sargasso Sea off Bemuda. I: The upper 500 m. Limnol. Oceanogr., 16, 219-240.

Deines, K.L., 1999. Backscatter estimation using broadband acoustic Doppler current
profiles. IEEE, 249-253.

Dickey, T., Frye, D., Jannasch, H., Boyle, E., Manov, D., Sigurdson, D., Siegel, D.,
Chang, G., Wu, J., Knap, A., 1998a. Initial results from the Bermuda Testbed Mooring
program. Deep-Sea Research I, 45, 771-794.

Dickey, T., Frye, D., McNeil, J., Manov, D., Nelson, N., Sigurdson, D., Jannasch, H.,
Siegel, D., Michaels, T., Johnson, R., 1998b. Upper-ocean temperature response to
Hurricane Felix as measured by the Bermuda Testbed Mooring, Monthly Weather
Review, 126, 1195-1201.




                                           27
Dickey, T., Zedler, S., Yu, X., Doney, S.C., Frye, D., Jannasch, H., Manov, D.,
Sigurdson, D., McNeil, J.D., Dobeck, L., Gilboy, T., Bravo, C., Siegel, D.A., Nelson, N.,
2001. Physical and biogeochemical variability from hours to years at the Bermuda
Testbed Mooring site: June 1994–March 1998, Deep-Sea Research II, 48, 2105-2140.

Fischer, J., Visbeck, M., 1993. Seasonal variation of the daily zooplankton migration in
the Greenland Sea, Deep-Sea Research I, 40, 1547-1557.

Flagg, C.N., Smith, S.L., 1989. On the use of the acoustic Doppler current profiler to
measure zooplankton abundance. Deep Sea Research I 36, 455-474.

Flagg, C.N., Wirick, C.D., Smith, S.L., 1994. The interaction of phytoplankton,
zooplankton, and currents from 15 months of continuous data in the Mid-Atlantic Bight.
Deep-Sea Research II 41, 411-435.

Francois, R.E. and Garrison, G.R., 1982. Sound absorption based upon ocean
measurements. Part II: Boric acid contribution and equation for total absorption, Journal
of Acoustical Society America, 72, 1870-1890.

Gilboy, T.P., Dickey, T.D., Sigurdson D.E., Yu, X., and D. Manov, 2000. An
intercomparison of current measurements using a vector measuring current meter, an
acoustic Doppler current profiler, and a recently developed acoustic current meter,
Journal of Atmospheric and Oceanic Technology 17, 561-574.

Harrison, W.G., Aristegui, J., Head, E.J.H., Li, W.K.W., Longhurst, A.R., Sameoto,
D.D., 2001. Basin-scale variability in plankton biomass and community metabolism in
the sub-tropical North Atlantic Ocean. Deep-Sea Research II 48, 2241-2269.

Heywood, K.J., Scrope-Howe, S., Barton, E.D., 1991. Estimation of zooplankton
abundance from shipborne ADCP backscatter. Deep-Sea Research I 38, 667-691.

Heywood, K.J, 1996. Diel vertical migration of zooplankton in the Northeast Atlantic.
Journal of Plankton Research 18, 163-184.

Hitchcock, G.L., Lane, P., Smith, S., Luo, J. and Ortner, P.B., 2002. Zooplankton spatial
distributions in coastal waters of the northern Arabian Sea, August, 1995. Deep-Sea
Research II, 49, 2403-2423.

Holliday, D. V., Pieper R. E., and Kleppel, G. S., 1989. Determination of zooplankton
size and distribution with multi-frequency acoustic technology. J. Cons. Int. Explor. Mer
41: 226-238.

Lo, Wen-tseng and Biggs, D. C., 1996, Temporal variability in the night-time distribution
of epipelagic siphonophores in the North Atlantic Ocean at Bermuda. Journal of Plankton
Research, 18, 923-939.




                                            28
Luo, J., Ortner, P.B., Forcucci, D. and Cummings, S.R., 2000. Diel vertical migration of
zooplankton and mesopelagic fish in the Arabian Sea. Deep-Sea Research II, 47, 1451-
1473.

Lyons, M.G., Smith, P.E., and Moser, H.G., 1994. Comparison of cross-shelf trends in
acoustic Doppler current profiler amplitude and zooplankton displacement volume in
southern California. CalCOFI Report 35, 240-552.

Madin, L.P., Horgan, E.F. and Steinberg, D.K., 2001. Zooplankton at the Bermuda
Atlantic Time-series Study (BATS) station: diel, seasonal and interannual variation in
biomass, 1994-1998. Deep-Sea Research II, 48, 2063-2082.

McGillicuddy, D.J., Robinson, A.R., Siegel, D., Johnson, R., Jannasch, H., Dickey, T.,
McNeil, J., Michaels, A., Knap, A., 1998. New evidence for the impact of mesoscale
eddies on biogeochemical cycling in the Sargasso Sea, Nature, 394, 263-266.

McNeil, J.D., Jannasch H.W., Dickey T., McGillicuddy D., Brzezinski M., and.
Sakamoto C.M, 1999. New chemical, bio-optical, and physical observations of upper
ocean response to the passage of a mesoscale eddy off Bermuda. Journal of Geophysical
Research ,104, 15,537-15,548.
Napp, J.M., Ortner, P.B., Pieper, R.E., Holliday, D.V., 1993. Biovolume-size spectra of
eippelageic zooplankton using a Multi-frequency acoustic profiling system (MAPS).
Deep-Sea Research I 40, 445-459.


Ortner, P.B., Wiebe, P.H., Haury, L., and Boyd, S., 1978 Variability in zooplankton
biomass distribution in the northern Sargasso Sea: the contribution of Gulf stream cold
core rings, Fishery Bulletin, Vol. 76, No. 2, 323-334.

Pieper, R.E., Holliday, D.V., Kleppel, G.S., 1990. Quantitive zooplankton distributions
from multifrequency acoustic, Journal of Plankton Research 12(2), 433-441.

Pieper, R.E., McGehee, D.E., Greenlaw, C.F., Holliday, D.V., 2001. Acoustically
measured seasonal patterns of Zooplankton in the Arabian Sea. Deep-Sea Research I 48,
1325-1343.

Pleuddemann, A.J., Pinkel, R., 1989. Characterization of the patterns of diel migration
using a Doppler Sonar. Deep-Sea Research, 36, 509-530.

RDI, 1995. RD Instruments direct reading and self-contained broadband acoustic Doppler
current profiler. RD Instruments Tech. Manual.

Roe, H.S.J., Griffiths, G., 1993. Biological information from an acoustic Doppler current
profiler, Marine Biology 115, 339-346.




                                           29
Rippeth, T. P., Simpson, J. H., 1998. Diurnal signals in vertical motions on the Hebridean
shelf, Limnology and Oceanography, 43(7), 1690-1696.

Roman, M.R., Dam, H.G., Gauzens, A.L., and Napp, J.M., 1993, Zooplankton biomass
and grazing at the JGOFS Sargasso Sea time series station, Deep-Sea Research I, 40,
883-901.

Schott, F., 1986, Medium-range vertical acoustic Doppler current profiling from
submerged buoys, Deep-Sea Research, Vol. 33, No. 10, 1279-1292.

Smith, S.L., Pieper, R.E., Moore, M.V., Rudstam, L.G., Greene, C.H., Zamon, J.E.,
Flagg, C.N. and Williamson, C.E., 1992. Acoustic techniques for the in situ observation
of zooplankton, Acoustic techniques 36, 23-43.

Steinberg, D.K., Carlson, C.A., Bates, N.R., Goldthwait, S.A., Madin, L.P. and Michaels,
A.F., 2000. Zooplankton vertical migration and the active transport of dissolved organic
and inorganic carbon in the Sargasso Sea, Deep-Sea Research I, 47, 137-158.

Zedler, S.E., Dickey, T.D., Doney, S.C., Price, J.F., Yu, X. and Mellor, G.L., 2003.
Analysis and simulations of the upper ocean‟s response to Hurricane Felix at the
Bermuda Testbed Mooring site: August 13-23, 1995. Journal of Geophysical Research,
107 (12), p. 25-1.

Zhou, M., Nordhausen, W., and Huntley, M., 1994. ADCP measurements of the
distribution and abundance of euphausiids near the Antarctic Peninsula in winter. Deep -
Sea Research I, 41, 1425-1455.

Zimmerman, R.A., Biggs, D.C., 1999. Patterns of distribution of sound scattering
zooplankton in warm- and cold-core eddies in the Gulf of Mexico, from a narrowband
acoustic Doppler current profiler survey. Journal of Geophysical Research, 104, 5251-
5262.




Table & Figure Captions

Table 1. Bermuda Testbed Mooring ADCP measurement periods.


                                            30
   Deployment                      Time Period                                Depth Range (m)
        6              August 21, 1996 – January 10, 1997                       20.5 – 200.5
        7                 May 3, 1997 – July 30, 1997                           20.5 – 203.5
        8             August 8, 1997– November 20, 1997                         20.5 – 203.5
        9             November 26, 1997 – March 31, 1998                        21.5 – 201.5
       10             November 11, 1998 – March 19, 1999                        18.5 – 195.5
       11                 April 1, 1999 – July 21, 1999                         19.5 – 196.5
       12              July 29, 1999 – November 6, 1999                         20.5 – 197.5
       13               December 5, 1999 – May 27, 2000                         19.5 – 196.5
       14              June 1, 2000 – November 29, 2000                         20.5 – 197.5




Table 2 Summary of calibration results

   Dep      ADCP Time Periods       Net Tow Time                      Log(DW/4π) = A*Sv + B
   No                                  Periods            2
                                                          R          P             A        B     N
    6       08/21/96 – 01/10/97   09/02/96-12/14/96    0.6504   0.000161        0.0363   2.4945   16

    7       05/03/97 – 07/30/97   05/05/97-07/15/97    0.3823   0.032108        0.0282   1.9632   12

    8       08/08/97 – 11/20/97   08/11/97-01/14/97    0.6007   0.000421        0.0330   2.1306   16

    9       11/26/97 – 03/31/98   12/09/97-03/25/98    0.0803   0.179610        0.0198   1.1963   24

    10      11/11/98 – 03/19/99   11/18/98-02/24/99    0.5536   0.000031        0.0688   5.0825   24

    11      04/01/99 – 07/21/99   04/07/99-07/06/99    0.1753   0.066156        0.0233   1.4958   20

    12      07/29/99 – 11/06/99   08/03/99-10/12/99    0.6923   0.002805        0.0317   2.0181   10

    13      12/05/99 – 05/27/00   12/08/99-04/11/00    0.1856   0.213930        0.0252   1.5714   10

    14      06/01/00 – 11/29/00   06/06/00-11/14/00    0.6576   0.0000015       0.0358   2.4257   24

  Overall                                              0.2454   4.9029e-011     0.0313   2.0631   156




Table 3. Means, minima, maxima, and standard deviations of daily mean, depth-
integrated biomass (mg/m2) during BTM deployments 6-14
______________________________________________________________________________________

Deployment No.       Mean         Minimum           Maximum     Standard Deviation



                                               31
_____________________________________________________________________
      6            502           404           648             54
      7            505           377           700             58
      8             502          424           568             29
      9            541           385           688             64
     10            507           405           647             49
     11            481           387           618             51
     12            483           434           565             25
     13            480           343           676             69
     14            483           385           633             55
    6-14            497          343           700             57
__________________________________________________________________________________




Table 4. Seasonality of means, minima, maxima, and standard deviations of daily mean,
depth-integrated biomass (mg/m2) during BTM deployments 6-14
______________________________________________________________________________________

Season            Mean       Minimum           Standard Deviation
                                           Maximum
________________________________________________________________________

Jan–Mar            471          343            700             62

Apr–Jun            503          440            633             36

Jul–Sep            486         385             648            51

Oct–Dec            497         343             700            57




                                          32
Figure 1. ADCP acoustic backscatter intensity/zooplankton biomass calibration curve.
The linear regression line is shown. This relationship has been used as a calibration to
estimate zooplankton biomass.




                                             33
Figure 2. Time series of daily averaged temperature at several depths, daily averaged,
depth integrated (22–190m) zooplankton biomass estimated using ADCP backscatter
intensity data, chlorophyll a fluorescence at specific depths and horizontal current speed
measured by the ADCP at about 45 and 150 m. Years are indicated in each panel. The
color bar scale indicates depths of temperature measurements.




                                           34
a




    35
b




    36
c



    37
Figure 3. Contours of zooplankton biomass estimated from ADCP backscatter intensity
data during Deployment 6–14 (1996-2000).




                                           38
Figure 4. The daily cycle of zooplankton biomass and relative vertical velocity (relative
to 34–43 m depth) computed using an ensemble average of 14-day time series show
zooplankton biomass vertical distributions and daily migration patterns in the spring
(May 16–29, 1997), summer (August 16–29, 1997), fall (November 1–14, 1997) and
winter seasons (February 10–23, 1998).




                                           39
a




    40
b




    41
Figure 5. Backscatter intensity and relative vertical velocity (relative to 34–43 m depth)
averaged between 121 and 190 m depths from May 4, 1997 through March 30, 1998
show seasonal variations of the diel migration. The top panel shows backscatter intensity.
The bottom panel shows relative vertical velocity; negative values indicate downward
motion, and positive values represent upward motion; magenta lines represent sunrise and
sunset times (in hour).




                                           42
43
Figure6. Timing of peak downward and upward migration velocity relative to the times
of sunrise and sunset. The difference in minute between the time of the peak vertical
velocity between 4:00–7:00 and 17:00–20:00 and the time of sunrise/sunset was
calculated; negative values indicate that migration preceded sunrise/sunset. Data used for
this figure were from May 4, 1997 through March 30, 1998.




                                           44
Figure 7. Cross correlation analysis with 1-day zooplankton biomass estimated from
ADCP backscatter intensity data, chlorophyll a fluorescence and temperature. (a). Time-
series of zooplankton biomass at 76 m (solid line) and chlorophyll a fluorescence at 77 m
(dashed line). (b). Cross-correlation of zooplankton versus chlorophyll a fluorescence
using time series shown in (a). (c). Time-series of zooplankton biomass at 76 m (solid
line) and temperature at 77 m (dashed line). (d). Cross-correlation of zooplankton versus
temperature using time series shown in (c).
(e). Time-series of zooplankton biomass at 25 m (solid line) and chlorophyll a
fluorescence at 24 m (dashed line). (f). Cross-correlation of zooplankton versus
chlorophyll a fluorescence using time series shown in (e). (g). Time-series of
zooplankton biomass at 25 m (solid line) and temperature at 24 m (dashed line). (h).
Cross-correlation of zooplankton versus temperature using time series shown in (g).




                                              45
46
Figure 8. Cross correlation analysis with 1-day zooplankton biomass estimated from
ADCP backscatter intensity data, chlorophyll a fluorescence and temperature. (a). Time-
series of zooplankton biomass at 46 m (solid line) and chlorophyll a fluorescence at 45 m
(dashed line). (b). Cross-correlation of zooplankton versus chlorophyll a fluorescence
using time series shown in (a). (c). Time-series of zooplankton biomass at 46 m (solid
line) and temperature at 45 m (dashed line). (d). Cross-correlation of zooplankton versus
temperature using time series shown in (c).
(e). Time-series of zooplankton biomass at 73 m (solid line) and chlorophyll a
fluorescence at 72 m (dashed line). (f). Cross-correlation of zooplankton versus
chlorophyll a fluorescence using time series shown in (e). (g). Time-series of
zooplankton biomass at 73 m (solid line) and temperature at 72 m (dashed line). (h).
Cross-correlation of zooplankton versus temperature using time series shown in (g).




                                              47
48
Figure 9. Cross correlation analysis with 1-day zooplankton biomass estimated from
ADCP backscatter intensity data, chlorophyll a fluorescence and temperature. (a). Time-
series of zooplankton biomass at 46 m (solid line) and chlorophyll a fluorescence at 45 m
(dashed line). (b). Cross-correlation of zooplankton versus chlorophyll a fluorescence
using time series shown in (a). (c). Time-series of zooplankton biomass at 46 m (solid
line) and temperature at 45 m (dashed line). (d). Cross-correlation of zooplankton versus
temperature using time series shown in (c).
(e). Time-series of zooplankton biomass at 73 m (solid line) and chlorophyll a
fluorescence at 74 m (dashed line). (f). Cross-correlation of zooplankton versus
chlorophyll a fluorescence using time series shown in (e). (g). Time-series of
zooplankton biomass at 73 m (solid line) and temperature at 74 m (dashed line). (h).
Cross-correlation of zooplankton versus temperature using time series shown in (g).




                                              49
50
Figure 10. Power spectra computed using current speed from ADCP from depths of 45,
75, 150 and 190 m during Deployments 7–9. Inertial (I), diurnal (D), and semi-diurnal
(SD) periods are indicated.




                                         51
Figure 11. Power spectra computed using zooplankton biomass estimated from ADCP
from depths of 45, 75, 150 and 190 m during Deployments 7–9. Inertial (I), diurnal (D),
and semi-diurnal (SD) periods are indicated.




                                           52
Figure 12. Averaged backscatter intensity (a) and averaged percent good (b) data from JD
124 to 210, 1997, Deployment 7. Solid lines – midnight, from 22 to 2 hour of next day;
Dashed lines – Noon, from 10 to 14 hour.




                                           53

				
DOCUMENT INFO
Categories:
Tags:
Stats:
views:54
posted:2/3/2011
language:English
pages:53
shenreng9qgrg132 shenreng9qgrg132 http://
About