Sensor Technology for Zooplankton Assessment

Abstract

The composition and structure of marine food webs and it's variation in space and time influence the quality and quantity of carbon exported from surface water, the concentration of inorganic nutrients, and the elemental ratios of particulate and dissolved pools in the ocean. Marine phytoplankton, microzooplankton and larger zooplankton important to the biological pump, range over several orders of magnitude in size, occur over widely varying vertical and horizontal scales, exhibit different behaviors (e.g., vertical migration), and may occur seasonally or episodically. Consequently, assessing the efficiency and variability of the biological pump remains a major challenge. Two different sensor systems used to assess the abundance and distribution of marine biota will be described. The first approach uses a split-beam digital echo sounder system and the second approach uses a high-speed digital line scan imaging system. The merits, limitations, and recent developments of these systems will be presented. Although hardware and software used to assess marine biota have improved significantly during the last few decades, researchers still face serious technical issues in relating the detected image or signal to a biologically useful measure. Recommendations from a recent workshop on sensor technology are discussed.

Full text

SENSOR TECHNOLOGY FOR ZOOPLANKTON ASSESSMENT
Daly, K.L., S. Samson, T. Hopkins, A. Remsen, T. Sutton1, & L. Langebrake
University of South Florida, St. Petersburg, FL U.S.A., kdaly@marine.usf.edu
1Woods Hole Oceanographic Institution, Woods Hole, MA 02543 USA
Abstract
The composition and structure of marine food webs and it’s variation in space and
time influence the quality and quantity of carbon exported from surface water, the
concentration of inorganic nutrients, and the elemental ratios of particulate and dissolved
pools in the ocean. Marine phytoplankton, microzooplankton and larger zooplankton
important to the biological pump, range over several orders of magnitude in size, occur
over widely varying vertical and horizontal scales, exhibit different behaviors (e.g.,
vertical migration), and may occur seasonally or episodically. Consequently, assessing
the efficiency and variability of the biological pump remains a major challenge. Two
different sensor systems used to assess the abundance and distribution of marine biota
will be described. The first approach uses a split-beam digital echo sounder system and
the second approach uses a high-speed digital line scan imaging system. The merits,
limitations, and recent developments of these systems will be presented. Although
hardware and software used to assess marine biota have improved significantly during the
last few decades, researchers still face serious technical issues in relating the detected
image or signal to a biologically useful measure. Recommendations from a recent
workshop on sensor technology are discussed.
Introduction
Knowledge about the role of the biological pump in carbon and associated
biogeochemical cycles in the ocean is limited due to undersampling of temporally and
spatially varying processes. It is well known, however, that food web structure is an
important determinant of ocean carbon cycling and vertical fluxes (e.g., Longhurst, 1991;
Legendre and Rasoulzadegan, 1996; Wassmann, 1998). The magnitude of biologically
mediated elemental transport between the surface layer and the ocean interior may be
dependent on particular key groups of phytoplankton and zooplankton, in which the
interactions of these organisms govern the partitioning and rate of exchange between
particulate and dissolved carbon pools (Daly et al., 1999). Although significant progress
towards understanding the function of the biological pump has been made in recent years,
the relative contributions of the biological pump versus the solubility pump and the
processes that control the structure and efficiency of the biological pump remain poorly
known. The vertical and horizontal spatial variability of key components of the
biological pump need to be assessed on space and time scales relevant to the controlling
biological, chemical and physical processes (Daly and Smith, 1993).
Marine phytoplankton, microzooplankton and larger zooplankton important to the
biological pump, range over several orders of magnitude in size, occur over widely
varying vertical and horizontal scales, exhibit different behaviors (e.g., vertical
migration), and may occur seasonally or episodically. Consequently, assessing the

2
efficiency and variability of the biological pump remains a major challenge. The Second
Marine Zooplankton Colloquium noted that the role of zooplankton in biogeochemical
cycles is an important focus issue, which urgently needs to be addressed (Bathmann et
al., 2001). To date the spatial and temporal variability of zooplankton has been assessed
using water bottles and pumps for small microzooplankton and nets, pumps, diver
observations, hydroacoustic systems, optical counters, and cameras for larger
zooplankton. Bottles, pumps, nets, and diver observations have limited sampling
capability, while acoustics and imaging systems offer a greater opportunity for
advancement in understanding the affect of small-scale zooplankton distribution and
behavior on large scale population biology and community ecology. Here we discuss
recent developments in high-frequency acoustics and high-resolution imaging systems,
which are being used to quantify the abundance, distribution and interaction of marine
particles and organisms with ocean physics.
Developments in Hydroacoustic Technology
Hydroacoustic technology has long been used to observe physical phenomena in
the ocean, such as internal waves and bubbles, and to detect bathymetry, suspended
sediments, and deep-scattering layers (reviewed in MacLennan and Holliday, 1996).
This technology was initially applied to fisheries investigations during the 1930s to track
schools of fish and has developed to become an essential and accepted tool used by
commercial and sports fishermen and fisheries managers today. Acoustic sensors may be
hull mounted on ships, deployed off ships in towed bodies (Fig. 1), or mounted on
moorings and AUVs.
Figure 1. Example of an acoustic towed
body containing 38 and 120 kHz split
beam transducers (Hydroacoustics
Technology, Inc) mounted in a down-
ward-looking mode.
The foundation of acoustic methodology is based on knowing the speed of sound
in seawater and the time of sound travel (interval between transmission and reception) in
order to calculate the distance to a sound scatterer. The amplitude of sound energy
reflected or backscattered by a target (i.e., target strength) is proportional to the
orientation, size, and material properties (sound speed and density contrast relative to
seawater) of the target. The returned sound is detected by a receiver, amplified, and sent
to an output device and/or an echoprocessor. If the returned signal exceeds a set
threshold, a mark appears on the graphic display (Fig. 2). The distance from the top of
the display to the mark is proportional to the roundtrip of the sound traveled and therefore

3
the range or distance from the transducer. The range of sound penetration and the size of
the target detected are dependent on the frequency of the transmitted sound. Frequencies
Figure 2. Chart echogram (left) and digital display (right) of 120 kHz acoustic backscattering
illustrating spatial patchiness of zooplankton distribution in the Antarctic. The zooplankton
layers centered at about ca. 30 m (left) and at ca. 50 and 100 m (right) are primarily composed
of the euphausiids, Euphausia superba and E. crystallorophias. Digital display from
Hydroacoutic Technology, Inc. Model DES244; volume backscattering increases from blue to
red.
appropriate for zooplankton range from about 50 kHz to 10 MHz. Multiple frequencies
discriminate zooplankton by size and can be effectively used to estimate a zooplankton
size-abundance spectrum (Holliday et al., 1989).
The development of dual beam and split-beam technology (Ehrenberg, 1983;
Traynor and Ehrenberg, 1990; Ehrenberg and Torkelson, 1997) made it possible to
estimate the target strength of individuals (for non-overlapping echoes) in situ and to
determine swimming speed and direction of individuals through the acoustic beam.
The split beam transducer has four quadrants, which combine to transmit, and two split-
beam pairs for receiving, which allows the measurement of the angular location of the
target in two right angle planes. Split-beam technology also reduces noise sensitivity
compared to other systems.
The echo integration technique, developed during the 1960s, is based on the
assumption that the echo energy from multiple targets is equivalent to the sum of the
echo energies from the combined individual targets. This technique is important for
estimating the numbers of fish in schools and for zooplankton abundance, since they
typically occur too close together to be resolved as individual echoes. The conversion
of echo-integrated volume backscattering to a biologically meaningful parameter, such
as biomass or abundance, requires ground-truth information and knowledge of the
target strengths of the species. Backscattering by fish has been well studied, but
acoustic signatures of zooplankton are poorly known because they are difficult to
obtain in situ. Most work has focused on developing theoretical models of zooplankton
target strength, which account for body shape and orientation (e.g., Stanton et al.,
2000). Because there are few data on material properties of zooplankton, models
assume that the properties are similar to that of the surrounding seawater or properties
are varied to examine the sensitivity of acoustic scattering. A growing number of

4
investigations of material properties of zooplankton (e.g., Foote, 1990; Chu et al, 2000)
and average orientation in situ will help to improve target strength models.
The advantages and limitations of acoustic technology are listed in Table 1.
Table 1. Advantages and limitations of hydroacoustic methods
Advantages
Limitations
Cost effective, low field man-hours Lack of species identification
Non-invasive, rarely affects organism
behavior Difficulty obtaining information near
surface and bottom boundaries
Nonselective, minimizes sampling bias
towards size or behavior High initial capital investment
Relatively complex to use, requires
Established theory and methods
Real time results training and experience
Possible impact on some marine
Collects large quantity of data for statistical
comparison mammals and fish
Samples large volumes of water
After Ransom et al. (1999)
One important advantage is that acoustics is the only technology available that can
provide real time images or visualizations and quantitative analyses of the spatial
distribution of biological organisms over large volumes of water in the ocean. The
large quantity of data collected is important for statistical considerations. In addition,
acoustic systems sample all sizes of organisms and generally are not biased by
avoidance behavior, problematic for nets and cameras. The cost of hydroacoustic
systems and relative complexity of the technique have limited the number of users.
Continuing developments in computer and acoustic hardware and software, however,
are decreasing costs and making systems more user-friendly. Validation of target
identification through net tows or camera images continues to be necessary.
Nevertheless, some success has been achieved in identifying species using behavior and
discriminant analyses. Development of models, analytical methodology, and field
validations are being actively pursued to improve this vital technology.
Developments in High Resolution Imaging Systems
A number of different types of imaging systems are currently in use to assess
the abundance, distribution, and behavior of zooplankton (e.g., Davis et al., 1996; Jaffe
et al., 1998; Tiselius, 1998; Sieracki and Sieracki, 1999). Here, we describe a high-
speed digital line scan camera that was developed for the collection of continuous
pictures of marine particles ranging in size from 200 µm to many centimeters, including
fragile particles, such as marine snow and entire siphonophore colonies, that normally
are destroyed in nets (Samson et al, 2001; http://www.marine.usf.edu/sipper/). The
Shadowed Image Particle Profiling and Evaluation Recorder (SIPPER) system (Fig. 3)
records images from two orthogonal directions and profiles a 96 mm2 sampling area or
33 m3 of water per hour at 1 m s-1. SIPPER can be operated from an autonomous

5
Figure 3. SIPPER mounted on the High Resolution Sampler towed vehicle (left) and on a 53
cm diameter AUV (right). The black square sampling inlet can be seen forward of the camera
housing, as indicated by an arrow.
vehicle, a profiling platform, or from a towed body. This system has several advantages
over other cameras (Table 2). For example in video systems, high magnification limits
the depth of focus to a few millimeters, images must be digitized before computer
processing, and recording on analog tape introduces noise. Holographic methods allow
high-resolution imaging of relatively large volumes of water, but require relatively large
and expensive laser illumination and one-time use film.
Table 2. Advantages and limitations of digital line scan imaging systems
Advantages
Limitations
Large size range of organisms imaged Invasive, possible avoidance
Identification possible
Relatively simple to use Automated ID software not yet widely
available
Gives in situ spatial distribution
between organisms Requires post-processing; not real time
Relatively high capital investment
Resolution dependent on water flow
Particles imaged by SIPPER may be collected by 20 remotely triggered nets to
ground-truth the images when mounted on the High Resolution Sampler (HRS) as shown
in Fig. 3 (Sutton et al., 2001). Particle flow through SIPPER is unidirectional, with
uniform velocities (± 5%) within ca. 5 mm of the sampling duct walls (Samson et al.,
2001). One imaging axis of the camera scans 15,000 lines s-1 at 2048 pixels per line.
The second imaging axis scans at 7,500 lines s-1 at 4096 pixels per line. Binary image
thresholds are preset using field programmable gate array (FPGA) based processing
hardware. Images are compressed and stored on a 50 GB SCSI disk drive. Over 50
hours of high-resolution images may be stored in this fashion, a practical consideration
for AUV deployments. Power consumption ranges from 8 W up to 75 W during data
recording. A 240-Wh lead-acid battery pack enables up to 3 h of operation on AUVs.
Longer deployments could be accommodated with more modern battery technology.

6
The current system was designed to operate at a maximum depth of ca. 500 m. It
has been tested in both the Bahamas and the Gulf of Mexico as part of a towed platform,
which included a suite of environmental sensors and 1.2 L Niskin bottles (Sutton et al.,
2001). Imaged particles included Tricodesmium, diatom chains, radiolarians, copepods,
lobster phyllosome larvae, medusae, euphausiids, midwater shrimp, salps, larvacean
houses, and ca. 10 cm long siphonophores (Fig. 4).
Figure 4. Images of organisms from the Gulf of Mexico acquired by the 2048-pixel SIPPER
camera. The representative sample includes Trichodesmium spp. (A), protists (B), a copepod (C),
pteropod (D), a ctenophore with tentacles extended (E),a midwater shrimp (F), a doliolid (G), a
small squid (H), a colonial siphonophore (I), and a larvacean (J). Scale bars indicate size ranges
of organisms.
Although larger zooplankton were collected by this system, avoidance of the 9.6 x 9.6 cm
sampling opening may limit the accuracy of estimates for relatively large taxa. Recent
developments include increasing the camera resolution to 4096 pixels for both arrays
with scanning at 21,000 lines per second, which would permit imaging of 50 µm sized
particles. Other improvements include an increased data storage capacity to enable
longer deployments and100 Mb s-1 data offload rates via Ethernet. In addition,
automated image processing software is being developed as well as a scaled down
version of SIPPER (about one third the size) for imaging particles as small as 15 to 25
µm. This upgraded mini-SIPPER will be a valuable tool for microplankton studies.

7
Future of Biological Sensors
A recent workshop on sensor technologies noted that the development and
validation of biological sensors is urgently needed (Daly, 2000;
http://sensors.marine.uef.edu). Many biological processes in the ocean remain poorly
known owing to under-sampling or the inability to sample important processes by
traditional shipboard techniques. Although hardware and software used to assess marine
biota have improved significantly during the last few decades, researchers still face
serious technical issues in relating the detected image or signal to a biologically useful
measure. Acoustic and imaging systems often include other environmental sensors so
that biological data may be collected on the same scales as physical and chemical
parameters. The lack of inexpensive and reliable sensors, such as those available to the
physical oceanographic community, however, hinders extensive chemical and biological
observations over broad regions of the ocean. Biological sensors also have suffered from
a lack of rigorous field validation, limiting confidence in data interpretation. The
accuracy and interpretation of sensor data must be improved with standards for in situ
calibrations of the systems and standard protocols for analyzing the data. Continuing
development and validation of biological sensors to investigate a spectrum of basic
processes in the ocean will require sufficient support by funding agencies, a trained
workforce of users, and data quality controls.
Acknowledgments. We thank Drs Koh Harada and Tommy Dickey for inviting us to
participate in the workshop. We also thank B. Ransom, P. Nealson, S. Johnston, and J.
Ehrenberg of Hydroacoustic Technology, Inc. and the University of Washington for
informative discussions. Preparation of this manuscript was partially supported by
National Science Foundation grants OPP-9816594, OPP-0196490, and OPP-0196489.
References
Bathmann, U., M.H. Bundy, M.E. Clarke, T.J. Cowles, K. Daly, H.G. Dam, MM.
Dekshenicks, P.L. Donaghay, D.M. Gibson, D.J. Gifford, B.W. Hansen, D.K.
Hartline, E.J.H. Head, E.E. Hofmann, R.R. Hopcroft, R.A. Jahnke, S.H. Jonasdottir,
T. Kiorboe, G.S. Kleppel, J.M. Klinck, P.M. Kramer, M.R. Landry, R.F. Lee, P.H.
Lenz, L.P. Madin, D.T. Manahan, M.G. Mazzocchi, D.J. McGillicuddy, C.B. Miller,
J.R. Nelson, T.R. Osborn, G.A. Paffenhofer, R.E. Pieper, I. Prusova, M.R. Roman, S.
Schiel, H.E. Seim, S.L. Smith, J.J. Torres, P.G. Verity, S.G. Wakeham, and K.F.
Wishner. 2001. Future marine zooplankton research - a perspective. Marine Ecology
Progress Series 222: 297-308.
Chu, D., P. Wiebe, and N. Copley. 2000. Inference of material properties of zooplankton
from acoustic and resistivity measurements. ICES Journal of Marine Science 57:
1128-1142.
Daly, K.L. and Walker O. Smith. 1993. Physical-biological interactions influencing
marine plankton production. Annual Review of Ecology and Systematics 24: 555-
585.
Daly, K.L. 2000. Sensor technology for remote, interactive aquatic experiments. EOS 81:
580.

8
Daly, K.L., D.W.R. Wallace, W.O. Smith, Jr., A. Skoog, R. Lara, M. Gosselin, E. Falck,
and P. Yager. 1999. Non-Redfield carbon and nitrogen cycling in the Arctic: Effects
of ecosystem structure and function. Journal of Geophysical Research 104, No. C2:
3185-3199.
Davis, C.S., S.M.Gallager, M. Marra, and W.K. Stewart. 1996. Rapid visualization of
plankton abundance and taxonomic composition using the Video Plankton Recorder.
Deep-Sea Research II 43: 1947-1970.
Ehrenberg, J.E. 1983. A review of target strength estimation techniques. FAO Fish. Rep.
300: 90-95.
Ehrenberg, J.E. and T.C. Torkelson. 1996. Application of dual-beam and split-beam
target tracking in fisheries acoustics. ICES Journal of Marine Science 53: 329-334.
Foote, K.G. 1990. Speed sound contrast in Euphausia superba. J. Acoust. Soc. Am. 87:
1405-1408.
Holliday, D.V., R.E. Pieper, and G.S. Kleppel. 1989. Determination of zooplankton size
and distribution with multifrequency acoustic technology. J. Cons. Int. Explor. Mer
46: 52-61.
Jaffe, J.S., M.D. Ohman, and A. De Robertis. 1998. OASIS in the sea: measurement of
the acoustic reflectivity of zooplankton with concurrent optical imaging. Deep-Sea
Research 45: 1239-1253.
Legendre, L. and F. Rasoulzadegan. 1996. Food-web mediated export of biogenic carbon
in oceans: Hydrodynamic control. Marine Ecology Progress Series 145: 179-193.
Longhurst, A.R. 1991. Role of the marine biosphere in the global carbon cycle.
Limnology and Oceanography 36: 1507-1526.
MacLennan, D.N. and D.V. Holliday. 1996. Fisheries and plankton acoustics: past,
present, and future. ICES Journal of Marine Science 53:513-516.
Ransom, B., T.W. Steig, S.V. Johnston, J.E. Ehrenberg, T.C. Torkelson, and P.A.
Nealson. 1999. Using Hydroacoustics for fisheries assessment. Hydroacoustic Tech.,
Inc., Seattle.
Samson, S., T. Hopkins, A. Remsen, L. Langebrake, T. Sutton, and J. Patten. 2001. A
system for high resolution zooplankton imaging. IEEE Journal of Oceanic
Engineering 26:671-676.
Sieracki, C.K. and M.K. Sieracki. 1999. Results of continuous and discrete flowcam
measurements of plankton community structure, Boothbay Harbor, Maine, USA.
Ocean Sciences abstract, Santa Fe, NM.
Sutton, T., T. Hopkins, A. Remsen, and S. Burghart. 2001. Multisensor sampling of
pelagic ecosystem variables in a coastal environment to estimate zooplankton grazing
impacts. Cont. Shelf Res. 21: 69-87.
Stanton, T.K. and D. Chu. 2000. Review and recommendations for the modeling of
acoustic scattering by fluid-like elongated zooplankton: euphausiids and copepods.
ICES Journal of Marine Science 57: 793-807.
Tiselius, P. 1998. An in situ video camera for plankton studies: design and preliminary
observations. Mar. Ecol. Prog. Ser. 164: 293-299.
Traynor, J.J. and J.E. Ehrenberg. 1990. Fish and standard target strength measurements
obtained with a dual-beam and split-beam echo-sounding system. Rapp. P. -v. Reun.
Cons. Int. Explor. Mer 189: 325-335.

9
Wasmann, P. 1998. Retention versus export food chains: processes controlling sinking
loss from marine pelagic systems. Hydrobiologia 363: 29-57.