Both vertebrate and invertebrate predators have been thought to have significant impacts on planktonic populations of copepods and larval fishes, although the importance of these two processes relative to other biological and physical loss and gain terms has been difficult to assess. We present results from two separate but closely related ongoing projects - GLOBEC and the Coastal Ocean Program's Georges Bank Predation Project - in which we attempt to estimate the magnitude of predation mortality on two species of copepods (Calanus finmarchicus and Pseudocalanus spp.) on Georges Bank. Invertebrate predators of concern include pelagic hydroids, hyperiid amphipods, chaetognaths, and ctenophores; vertebrate predators of concern are herring and mackerel. Our approach combines direct measures of the co-distribution of predators and prey from plankton nets (1m²and 10m² MOCNESS), SCUBA observations and fishery research trawls made during broad-scale surveys and site-specific process cruises between1994-1997. Feeding rates and prey selection of the main predator species are estimated by several methods, depending on the predator, including analysis of gut contents, feeding experiments in shipboard or laboratory incubations, and energetic calculations. We contrast results from two different regions (crest vs. southern flank) and two different years ('95 vs. '96),discuss the overall impact of predation on the GLOBEC target species of copepods, and point to future work that is needed to improve these estimates of predation mortality on Georges Bank.
 

Figure 1. Map of Georges Bank, showing location of "crest" and "flank" regions sampled in 1994, 1995 and 1997. Station locations for each year indicated by numbers within the ovals. Also shown are 3 stations sampled in 1997, from which predator distribution data aregiven in Figure 2.

Our approach to estimation of predation mortality requires 1) determining the co- distribution in time and space (annual, seasonal, horizontal and vertical) of predators and prey, and 2) measuring rates of predation by specific predators on specific target species of prey. Data on abundance and distribution of invertebrate predators and prey are derived mainly from the Broad Scale Survey, which uses plankton nets (1m² and10m² MOCNESS) to sample a grid of stations on the Bank (and selected off Bank sites) on monthly cruise from January to June, every year from 1995 through 2000 (projected).

We also estimated the abundance, size composition, and diet composition of fishes on the southern flank and crest of Georges Bank during the period April 26-May 26 1995; data for 1996 are still being analyzed. Samples offish were collected using a Shuman Series 68x354 Midwater Trawl towed at7.4 km/hr for 30 minutes. A systematic sampling design was employed; 142stations were occupied during the study period. For the purposes of this analysis, we post-stratified our samples to encompass the crest and flank study sites employed in the analysis of invertebrate predators; 69 stations were used in the site-specific analysis. We computed the biomass of herring and mackerel per m² of surface area within the study sites under the assumption of no escapement of fish within the path of the net. Because the efficiency of the net is not 100%, we consider these to represent minimum biomass estimates.

The second element we are determining is the specific feeding rate of each predator species on the target species of prey. In general, we aim to express this as numbers or biomass of prey consumed per day under field conditions. We have had to use different methods to estimate these rates, depending on the feeding biology and tractability of the particular predator. Where possible, we have examined gut contents microscopically to identify and count prey items (e.g., in Clytia, Sagitta, Pleurobrachia, Bolinopsis, herring, and mackerel). To determine feeding rates from these data, we have measured digestion or gut passage times (depending on the type of gut) in shipboard or shore-based experiments. In other cases we have relied on literature data to make provisional estimates (e.g., Themisto).

As soon as sufficient distribution data are available, we expect to be able to map predation intensity, as daily percent removal of target species, over the entire Bank, month by month, from 1995-2000. In this presentation we focus on variation in species composition and abundance of the main predators, and examples of estimated predation impacts for two samples sites in May of 1995 and 1996. Our two sites are the "crest" or central well-mixed region inside the 60 m isobath, and the "flank" region between the 60 m and 100 m isobaths, south of the crest (Fig. 1).
 
 

Distribution and Abundance of Predators. Qualitative and quantitative distribution of predators collected by both MOCNESS systems varied considerably from one region of the Bank to another, as well as interannually (Fig.3). Note the differing scales of abundance (#/m²) in Fig. 3,especially when comparing 1m² and 10m² MOCNESS catches.(Although biomass data have not yet been calculated for these tows, the contrast in biomass between large and small animals caught in the two nets should be substantially less than the contrast in numbers). Estimates of vertebrate predator (herring and mackerel) biomass per unit surface area within the study area are provided in Table 1. Predator biomass was highest for herring on the crest-site; abundance of mackerel was uniformly low during the study period in 1995.

 

Predator	kg/m2	# Calanus Eaten d-1	% Calanus Eaten d-1	# Pseudocalanus Eatend-1	% Psuedocalanus Eatend-1
Herring
Crest	0.042	12.94	0.52	1,391.48	10.2
Flank	0.002	0.002	<<0.001	5.98	0.03
Mackeral
Crest	<<0.001	0.004	<<0.001	0.34	<<0.001
Flank	0.002	2.79	<<0.001	1.5	0.008

Table 1. Estimates of pelagic fish biomass and daily consumption of Calanus and Pseudocalanus by herring and mackerel on the crest and flank of Georges Bank in May 1995.

Feeding Rates. Estimates of daily ingestion for five invertebrate predator species considered here were made with the methods described above and in Madin et al. (1997), and are presented in Table 2. It was not always possible to identify Calanus or Pseudocalanus in gut contents, so predation rates are for total copepods, nauplii or eggs d^-1.Similarly, indirect estimates from energetic calculations (e.g., Themisto)refer only to total copepods. In these cases, our estimates of the impact on Calanus and Pseudocalanus assume that these species will be consumed in proportion to their abundance among all copepods. As we develop better information about selectivity for or against these species by particular predators (see below under Future Work), this default assumption of neutral selection (i.e., E=O) will be modified.
 

Predator	Feeding Rate (Prey d-1)	Predators m-2	Prey Eaten m-2 d-1	% Calanus Eaten d-1 (E=0)	% Psuedocalanus Eaten d-1 (E=0)
Clytia	0.24	1,250	300	0.1	0.1
Bolinopsis	31	0	0	0	0
Pleurobrachia	27	0.2	5.4	<<1	<<1
Sagitta	1.1	2,330	2,563	1.5	1.5
Themisto	0.6	513	308	0.2	0.2
Total Inverts		4,093.2	3,176.4	1.8	1.8

Table 2. Georges Bank, Crest region. Estimates of predation impact from 5 predator species in May 1995. Predator feeding rates derived as described in text and Table 1. Predator and prey abundances from Process cruises and/or from Broad Scale Survey samples at same stations. Prey are nauplii for Clytia; copepodites and adults for other predators. Prey densities (integrated over the water column) were Calanus:753 nauplii and 2,475 copepodites m^-2; Pseudocalanus:1065 nauplii and 13,600 copepodites m^-2, all values integrated over depth sampled.

 
 

Predator	Feeding Rate (Preyd-1)	Predators m-2	Prey Eaten m-2d-1	% Calanus Eaten d-1(E=0)	% Psuedocalanus Eaten d-1 (E=0)
Clytia	0.24	31,166	7,480	27.3	27.3
Bolinopsis	31	20	620	1.3	1.3
Pleurobrachia	27	0.1	2.7	<<1	<<1
Sagitta	0.6	1,770	1,062	2.2	2.2
Themisto	0.6	159	95	0.2	0.2
Total Inverts		33,15.1	9,259.7	31.0	31.0

Table 3. Georges Bank, Flank region. Estimates of predation impact from 5 predator species in May 1995. Predator feeding rates derived as described in text and Table 1. Predator and prey abundances from Process cruises and/or from Broad Scale Survey samples at same stations. Prey are nauplii for Clytia; copepodites and adults for other predators. Prey densities (integrated over the water column) were Calanus:27,430 nauplii and 39,344 copepodites/adults m^-2; Pseudocalanus: 30,567 nauplii and 19,030 copepodites/adults m^-2, all values integrated over depth sampled.

Estimated Predation Impacts. We have estimated the number of Calanus and Pseudocalanus consumed by the principal predators on the crest and the southern flank in May 1995 and May 1996 (Fig. 4). More complete data are presented for 1995 (Tables 1-3 ) to make clear that the estimated predation impact (% prey eaten per day) for any one predator type is the product of that predator's daily feeding rate and its density, divided by the density of prey organisms.

In comparing sites, overall predation mortality was very much higher in the crest region, with Clytia estimated to remove 27% of copepod nauplii per day in 1995 and in excess of 400% in 1996, and herring removing10% of Pseudocalanus copepodites (but a negligible % of Calanus copepodites) in 1995. In the southern flank region, predation mortality was uniformly low for nauplii and copepodites of both copepod target species, with neither vertebrate nor invertebrate predators accounting for more than a two percent daily removal rate.

In comparing years, 1996 was clearly a higher mortality year than 1995for the copepod nauplii, due primarily to the much greater impact of Clytia( 400 % prey removal in 1996; only 27% in 1995) (Tables 2 and 3; Fig.3). It is not yet possible to compare years for copepodite mortality, however, because data on herring and mackerel are not yet available for 1996.

Mortality of GLOBEC target due to predation will depend on the feeding rates and selectivities of the predators, the ability of the prey to avoid the predators, and the co- distribution of predator and prey at particular locations and times. We are attempting to understand the combined effects of predator feeding rate, selectivity and prey avoidance success by examining the end result of these processes, namely the numbers and distribution of prey articles in the guts of the predators. Because of difficulties in the identification of gut contents and the unpredictable occurrence of different predators, it may not always be possible to quantify feeding impact by this method for all predators, stations or seasons. Determinations of rates and selectivity need to be tailored to each species, and repeated with collections from different times and locations to get realistic estimates of variation in feeding behavior.

Even with the best estimates we can make for behavior of particular predator species, it is apparent that the greatest factor controlling the impact of any one predator is its abundance relative to its prey. The large variations in abundance of the hydroids (Clytia), for example, make the difference between a potential for 400% daily removal of prey at the crest region and 1% removal in the flank region a few km away (Fig. 3).In contrast, Sagitta has a relatively uniform distribution over wide areas of the Bank, with little seasonal or interannual variation (Sullivan and Meise 1996). Selectivity is probably more important than predation rate, but neither will likely change total predation mortality by more than a factor of about 5.

Changes in predator abundance may vary by up to 1000 times from one station, month or year to another. Assuming that we have identified the predator species which are consuming target species, the critical parameter for first-order estimates of predation mortality will be their co-occurrence in space and time and relative abundance with the target species. Further improvements of predation mortality estimates are desirable, however, and will require consideration of the following issues.

We point to three key issues that need to be more fully examined to refine and improve our estimates of predation mortality of GLOBEC target species on Georges Bank.

Density-dependent Feeding Behavior. Although we have such information for some predators (e.g., Clytia; Madin et al. 1996), this needs to be determined for all predator species which are thought to be significant consumers of target species on Georges Bank. This will be critical to meeting the goals of our predation project, but even more importantly, will be essential to achieving the larger programmatic goals of developing coupled biological-physical models of the population dynamics of the target species.

Feeding Selectivity. While some of our estimates (e.g., herring and mackerel) are based upon analyses of stomach contents and feeding rates, which integrate any (possible) feeding selectivity on the part of the predators, others of our estimates (e.g., Clytia) are based on the assumption of neutral selectivity of prey types. Of course, an even slightly positive(negative) selectivity for a given prey type can increase (decrease) the estimated number of prey consumed, and the resulting predation impact. Experimental determination of the feeding selectivity of these predators is therefore badly needed.

Omnivorous/Carnivorous Copepods. Because of limited resources during Phases I and II, we have not given much attention to omnivorous or carnivorous copepods as predators on earlier developmental stage of the target copepod species. This is work we hope to undertake as part of Phase III of Georges Bank GLOBEC. A simple "back of the envelope" calculation for May, 1996 illustrates the potential importance of these predators - Centropages spp. on the crest numbered 106,000 m^-2, which when combined with a feeding rate of 1.0 nauplus/day (Paffenhofer and Knowles,1980), leads to a predicted daily removal rate of 600% of the total stock of copepod nauplii (although presumably they would have switched to alternate prey types). This exceptionally high predation rate would exceed even that predicted for the hydroids Clytia. We therefore need to know much more about the dynamics of these (and perhaps other) omnivorous copepods, including the functional responses and feeding selectivities mentioned above.

This work has been supported by NSF and NOAA as part of the Georges Bank/Northwest Atlantic GLOBEC program, as well as by NOAA'S Coastal Ocean Program/Georges Bank Predation project. We wish to thank Ted Durbin for access to the broad-scale copepod data (http://globec.gso.uri.edu:81/),and S. Avent and J. Cordell for technical assistance.

Madin, L.P., S.M. Bollens, E. Horgan, M. Butler, J. Runge, B.K. Sullivan, G. Klein-MacPhee, E. Durbin, A. Durbin, A. Bucklin, M.E. Clarke. 1996.Deep-Sea Res. 43:1823-1829.

Sullivan, B.K. and C.J Meise. 1996. Deep-Sea Res. 43:1503-1520.

U. S. GLOBEC. 1992. Northwest Atlantic Implementation Plan. U.S. Global Ocean Ecosystems Dynamics, Report No. 6.

Page maintained by Sean Avent. 6-29-99