Questions: The broad-scale shipboard sampling should address the following questions:

1. What are the broad-scale distributions of the target species and their predators and prey in relation to physical variables, and how do these change from winter to summer conditions?

2. How are the broad-scale distributions of measurements/indices for growth, fertility, respiration, and starvation in the target species related to the distributions of food supply and environmental conditions?

3. How are predator populations spatially and temporally distributed with respect to their prey (target species) and environmental conditions?

4. How does the birthdate and/or life stage frequency distribution of surviving individuals change during the drift around the Bank?

5. Is cohort mortality chronic or dominated by episodic events?

6. Is the probability of survivorship associated with individuals originating from discrete stocks?

Strategy: To determine adequately the mortality curve between successive early life stages of cod and haddock, bank-wide sampling will be required at intervals of no more than 30 days for the period December to August. Monthly sampling is needed to encompass the life stage durations and to provide quantitative estimates of their population abundance and areal distribution (See Figure 8 and Figure 9). Although the monthly interval of these surveys is insufficient to resolve the population dynamics of target copepods because the generation time is less than 60 days, the data will provide broad-scale distributional information. When combined with data from the moored instrumentation and process studies, it is hoped that this will provide the information necessary to study the changes in population structure through time. Other ideas that will result in a time series that is better suited for study of copepod dynamics are encouraged.

The "episodic vs. chronic" question identifies the need to distinguish between the two contrasting time scales of recruitment dynamics. One hypothesis is that in the early lives of the larval fish in a retention environment like Georges Bank, chronic mortality rates that persist on the order of 100 days may be the important rates determing recruitment variability. A 2-3% per day change in mortality or growth rate may have a much bigger effect on recruitment level than an event such as a warm-core ring that sweeps > 50% of the larvae off the Bank in a single brief episode. Modeling studies could be most constructive here.

For the broad-scale surveys, the areal coverage should encompass the entire bank, and include a portion of the southwestern Gulf of Maine and Slope Water. One approach is to set up a sampling grid with a spacing of about 20 km, including approximately 100 stations, which can be completed within a 7-10 day period. Stations both on and off the Bank will be needed to compare physical and biological processes in contrasting domains. In addition, to determine the egg and larval abundance with sufficient precision for cohort analysis, areas of high concentration characteristically observed on the southeastern or southern flank of the bank should be re-sampled with a station spacing of 5-10 km. This will require an additional 40-50 stations which can be completed within a 3-5 day period.

Statistical methods should be applied prior to any field sampling to suggest the intensity of sampling required to give the statistical power to actually reject specific null hypotheses. One problem is that the pelagic phase of a larval/juvenile fish lasts several months whereas a copepod completes one life cycle in 30-50 days. It is hoped that a sampling design can be developed that will produce information on the different requisite time scales. A mixture of point time series and occasional spatial surveys may be an alternative.

Individuals of each target species differ in size and swimming (i.e., sampler avoidance) ability, and in abundance. Large volumes of water (100's of cubic meters) need to be filtered to collect the relatively rare fish larvae; small volumes (a few cubic meters) are needed for zooplankton. Thus an array of sampling gear is required. Larger-scale sampling for ichthyoplankton and zooplankton should be conducted with a combination of 1-m multiple net systems using 150 µm mesh and bongo nets using coarse and fine mesh. A portion of the sampling should be conducted with vertically discrete opening/closing nets and acoustical profilers to provide additional information on the vertical distribution of the organisms, particularly in areas of high larval concentration. New optical and acoustical sensors may also be deployed on the multiple net systems or in an underway mode to provide high resolution distributional information on plankton and fish biomass, size, and taxonomic composition. When sampling for animals to be used in experiments, gentle techniques must be employed -- nets must sample only while the ship drifts on station and large cod ends must be employed. In April through August, both 1-m and 10-m multiple net sampling are required to sample the pelagic juvenile stages of cod and haddock. All net systems should be equipped with CTD's and fluorometers to measure simultaneously the basic hydrographic and chlorophyll distributions on the Bank. Sampling of recently-settled juvenile fish and their epibenthic prey (if carried out) will require combined pelagic and bottom gear. ROV transects may provide the best single technique for obtaining representative data on the true abundance, distribution, and activity patterns of fish and their prey (Auster et al., 1991). Fine mesh nets (64 µm) and large Niskin bottles should be used to collect copepod nauplii, protozoans, and phytoplankton.