In situ Processes Working Group
Chairman: Percy Donaghay
Participants: Joseph Katz, Uwe Kils, Gustav Paffenhofer, Rudi Strickler
The in situ rate processes working group was given the task of
identifying optical techniques that would allow direct or indirect
estimation of rate processes controlling the dynamics of key zooplankton
species. Specific tasks for optics utilization outlined in the
Background Document were: (a) obtain estimates of feeding, birth, growth
and mortality rates of planktonic animals in situ through continuous
observations of individuals or small swarms over seconds to minutes; (b)
obtain information on swimming behavior; (c) obtain predation data on
zooplankton by using 2 or 3 different scales of observation
simultaneously; (d) observe small-scale distribution of zooplankton,
especially aggregations of animals smaller than 1 mm; and (e) obtain
data on small scale physics and biophysical interactions. Each of these
tasks was evaluated both from the perspective of using optics to
directly address key scientific issues of U.S. GLOBEC (see Scientific
Problem) and as tools to routinely collect data on rates. There was a
general consensus that efforts should initially focus on developing
optical techniques to directly address key scientific issues. The
rationale for this choice is summarized below.
A fundamental objective of U.S. GLOBEC is to understand how biophysical
interactions at the individual level control zooplankton population
dynamics and fish larval recruitment. Small-scale biophysical
interactions can potentially affect zooplankton population dynamics
through their effects on feeding, swimming and predation vulnerability.
The nature of these effects is currently a matter of considerable
controversy. In particular, recent experimental and theoretical studies
have suggested that increased small-scale mixing may (1) alternatively
decrease feeding through reduction of micropatches (e.g., Lasker, 1975;
Davis et al., 1991), or increase feeding through the effects of
increased small-scale shear on encounter rates (e.g., Rothschild and
Osborn, 1988); (2) alter swimming patterns through effects on fine-scale
food structure or induction of escape responses (e.g., Donaghay, 1990);
and/or (3) modify predation vulnerability through effects on predator
encounter rates (e.g., Rothschild and Osborn, 1988), predator detection
mechanisms, and zooplankton aggregation (e.g., Kils, 1992). Although
there is sufficient theoretical and experimental evidence to indicate
that these biophysical interactions may sometimes control individual
success and population dynamics, we do not have the technical capability
to test these biophysical interaction models in situ, to evaluate the
validity of current rate measurements, or to test new techniques that
could dramatically improve in situ rate estimates (Price et al., 1988;
Marine Zooplankton Colloquium 1, 1989; Schulze et al., 1992). Just as
important, these same gaps in technical capability have limited our
ability to relate in situ responses to potentially controlling factors
and thereby identify "signature" properties that can be used in surveys
to link fine to coarse scales and to identify areas where specific
processes should dominate.
The primary goal is to develop a combination of techniques that will
allow the in situ measurement of zooplankton rate processes (feeding,
swimming and predator interactions) along with the potentially
controlling fine-scale biological structure, physical structure and
biological-physical interactions. Optical techniques will be important
in achieving this goal because they provide direct evidence of the
interaction of individual organisms with their conspecific neighbors,
their prey and potential predators.
- Measure the feeding, swimming and/or predation interactions of a group of zooplankton while simultaneously measuring in 3 dimensions the fine-scale (Ám to cm) particle and shear fields immediately surrounding these individuals. These measurements must be repeated periodically as the zooplankton move through the diel cycle.
- Interrelate the fine-scale structure and responses to small-scale (mm to m) vertical and horizontal gradients in those biological (food, zooplankton, predators), physical (shear fields, temperature, salinity, density), optical (light) and chemical (oxygen) properties that could influence the rates or induce directional responses in the fine-scale measurements. These gradient measurements must be made in conjunction with the fine-scale measurements (above) so that the two can be interrelated.
- Link fine to coarse scales by measuring biological and physical "signature" properties over a broad range of scales covered by the daily ambit of the zooplankton of interest (i.e., from scales of centimeters to 100's of meters covered by migrating zooplankton). These measurements should include the vertical structure of temperature, salinity, density and oxygen as well as the distribution and migration of foods, zooplankton and predators.
In Situ Rate Process Sampling Strategy
Two very different strategies could be used to estimate critical rate
processes in zooplankton First, a single individual zooplankter could
be followed (tracked) over time (15 minutes or so) while measurements of
individual feeding, swimming and predator interactions are recorded.
Although this technique provides ideal data for analysis of the
mechanism controlling feeding and swimming behavior, the costs of
developing equipment for individual tracking in situ appear to be
prohibitive at this time. Second, a group of zooplankton can be
followed over time while measurements of feeding, swimming and predator
interactions are repeatedly recorded for the group. This approach
provides statistical estimates of rate processes such as percent of time
feeding, percent swimming upwards, average swimming velocity, etc.
These statistical estimates are critically needed for parameterizing
models of population dynamics and for identifying factors that control
these behaviors. In contrast to tracking individuals, the acoustic
technology required for tracking groups of individuals is readily
available. The following sections therefore presume this latter
Required Sensor Technologies
Four major types of optical sensors will be required. Each of these
types of sensors are summarized briefly below. For a more detailed
discussion of these sensors see the short sensor descriptions in the
Biomass Working group report and the instrument descriptions in Schulze
et al. (1992).
- Imaging optical sensors that can provide real time and permanent 2D or 3D records of zooplankton feeding, swimming and predation interactions.
A variety of excellent two- (2D) and three-dimensional (3D) video systems have been developed for making these measurements under controlled conditions (see reviews by Price et al., 1988; Dickey, 1988; Schulze et al., 1992). These systems range from low-resolution systems designed to measure swimming behavior (Bugwatcher, e.g., Buskey and Swift, 1983; 1985), to high-resolution systems designed to measure feeding appendage motions and particle capture (Alcaraz et al., 1980). These techniques have recently been extended to allow tracking of single individuals over time (CritterCam-2D and CritterSpy-3D), as well as examination of the behavior of groups of individuals (Bugwatcher). Extensive motion analysis software has also been developed. High-resolution systems for in situ two-dimensional video measurements have recently been developed for measuring feeding (CritterCam, Bergeron et al., 1988) and predator-prey interactions (ecoSCOPE, Kils, 1992) in the ocean. The critical factor limiting application to in situ rate measurements is the lack of development of stable, nonintrusive deployment platforms. High resolution three-dimensional systems have not been developed for use in the ocean primarily because they are far more intrusive than 2D systems, although 3D videography using multiple orthogonally-oriented cameras has been recently used to study the behavioral interactions of relatively large organisms (fish and shrimp) (W. Hamner, pers. comm.).
- Imaging optical sensors that can measure particle characteristics, distributions (inter-particle distance), and motions (from swimming, sinking and small-scale mixing) in three-dimensions in the immediate vicinity of the zooplankton (20 Ám to cm).
The technology for making these measurements does not currently exist, but motion-sensing holographic systems are under development to make these measurements in the ocean and in the lab. Prototype hardware testing should begin in two years. The critical factor limiting application to in situ rate measurements is the development of the sensor hardware and data processing techniques. The lack of development of stable, non-intrusive deployment platforms will also limit some applications.
- Imaging sensors that can measure gradients in zooplankton and their resources over small- to course-scales.
Various arrays of the required video sensors are undergoing prototype testing in the field (e.g., Video Plankton Recorder (Davis et al., 1992a; 1992b); ecoSCOPE (Kils, 1992)). These systems provide sufficient quality images that visual methods can be used to identify two-dimensional spatial patterns at the species level. Routine use of these systems for spatial mapping is severely limited by image analysis software development (see Biomass Estimation Working Group Report). The holographic systems currently under development could also be used to measure gradients.
- Bulk optical sensors that can measure food distribution on both fine- and coarse-scales simultaneously (thus providing links between scales).
Bulk optical sensors measuring transmission, scattering, fluorescence and absorption at single or multiple wavelengths that could provide the required information exist or are under development . The utility of these measurements for estimating "signature" properties should greatly increase as spectral optical devices move from prototype to commercial instruments. Prototype tests have demonstrated the feasibility of developing biological/physical microstructure profiling systems that can be deployed free-fall in survey mode (laser/fiber optic profiler (Cowles et al., 1991)) or from stable platforms (bio-optical surface profiler (Donaghay et al., 1992)). The primary factor limiting routine measurement of fine structure during both surveys and rate experiments is the development of combinations of non-intrusive sensor configurations and nonintrusive sampling platforms.
The progammatic goal of in situ rate measurements can only be met by the
nonintrusive application of a combination of optical and non-optical
sensor technologies. Although most of the required sensor technologies
have been developed for use in the laboratory or for coarse-scale
sampling in the ocean, their successful application to in situ rate
process measurement will require major efforts in the five areas
- Develop in situ configurations of the required sensors.
Accurate measurement of an in situ rate process and fine-scale structure is dependent on developing non-intrusive sensor packages that do not disturb the structure and processes they are measuring. These problems are particularly critical for systems designed to collect time series data on behavior. Recent prototype work, however, has demonstrated that major advances can be made in this area by reducing instrument size through use of micro-optics and micro-electronics. Major progress can also be made by using optical relay techniques (fiber optics, relay lenses, range gating) to increase the distance between the sensor and the sensed volume. The working group strongly recommends that these technologies be exploited not only in developing new sensors, but also in reconfiguring existing and prototype sensors into operational instruments.
- Develop non-intrusive techniques for testing and routine deployment of the multiple sensor systems needed to make the required measurements.
This is a first-order problem that severely restricts the use of both existing and future optical systems for measuring in situ rates. Small sub-surface platforms (including manned submersibles and remotely operated vehicles) hold the greatest potential for rapidly developing the capability to follow a single group of zooplankton over time while making behavioral observations. This capability is essential to making the in situ rate process measurements discussed above. The working group therefore strongly recommends that non-intrusive techniques be developed and tested for deploying these sensor systems from both existing small manned submersibles and from small unmanned platforms associated with those submersibles (Donaghay, 1989).
- Develop techniques for integrating the multiple sensors into a system that can be used effectively under in situ conditions.
Two closely related sensor integration problems must be resolved. First, techniques must be developed that ensure that all recorded data can be interrelated in time and space. Although this problem sounds trivial for the systems with a fixed (i.e., Eulerian) frame of reference, the problem is more difficult when the coordinate system is constantly moving with the group of plankton being observed. Second, "smart sampling" techniques must be developed that allow real time analysis of data from one or more sensors to be used to control the timing and location of sampling by higher resolution systems. Prototype tests (bio-optical and density surface sampling, OPC triggered video camera) have demonstrated that the development of "smart sampling" techniques is an excellent way to reduce post-experiment data analysis to manageable levels, yet still insure that critical data are collected.
- Conduct cooperative engineering of sensors and sensor configurations.
Many of the imaging optical sensors under development have the potential to provide information critical to the interpretation of bulk optic and acoustic sensors. At the same time, the bulk optic and acoustic sensors have considerable potential for guiding high-resolution sampling. System integration is needed, i.e., electronic coordination of data acquisition from different sensors which cover the same water volume at the same time.
- Improve data and image analysis capabilities.
Current capabilities for analyzing images to identify species are clearly inadequate to allow automated processing of the large numbers of images that will be produced by these systems. Although it is doubtful that deficiencies in data handling and pattern recognition will limit progress in rate measurements in the near term (where problems (a)-(d) are of greatest concern), the development of these capabilities should begin immediately since they will require years to perfect.