© 2004 International Council for the Exploration of the Sea
Running the gauntlet: the predation environment of small fish in the northern Gulf of St Lawrence, Canada
Fisheries and Oceans Canada, Institut Maurice-Lamontagne Mont-Joli, QC, G5H 3Z4 Canada
*Correspondence to D. E. Duplisea: tel: +1 418 775 0881; fax: +1 418 775 0740. e-mail: Duplisead{at}dfo-mpo.gc.ca.
Predation size spectra were constructed for the northern Gulf of St Lawrence, covering prey size ranges that include pre-recruit cod. Predation by fish and harp seals was modelled with a log-normally distributed predator–prey size ratio along with a relationship between predator body size and the energy required. Fish concentrate predation on prey of weight 0.5–2 g, whereas harp seals prefer prey of 60–125 g. It is speculated that predation caused by harp seals on pre-recruits could be a major factor limiting cod recruitment in the system. The northern Gulf of St Lawrence is a cold boreal system with a large predatory seal population, and cod recruit older than elsewhere. Therefore, cod recruitment may be more strongly affected by predation in the northern Gulf of St Lawrence than in warmer systems such as the North Sea, where recruitment is strongly influenced by temperature.
Keywords: depensation, foodweb, size spectrum, stock recovery
Received 1 April 2004; accepted 25 October 2004.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussionReferences |
|---|
Body size is a fundamental characteristic of organisms that is a good indicator of an individual's role in community dynamics (Peters, 1983). Predation is a key process in fish communities, and is strongly related to body size (Dickie et al., 1987). Fish predators commonly swallow their prey whole, constraining them to eat prey smaller than themselves and often in fixed proportion to their own size (Floeter and Temming, 2003). The constraint on maximum prey size for a predator is balanced by another constraint on minimum prey size, where foraging costs are too high if their local density is too low (Kerr and Martin, 1970). Given these upper and lower constraints and empirical evidence from stomach sampling, the predator–prey body size ratio for fish can be approximated using a lognormal distribution (Floeter and Temming, 2003; Benoît and Rochet, 2004).
Metabolic rate and energy demand are also a function of body size (Peters, 1983). Therefore, in the ideal situation where sufficient prey abundance of the appropriate size is available, food consumption of a predator can be modelled as a function of its own size.
These two simple and generally accepted allometric relationships, predator–prey size ratio and energy demand, after parameterization, allow us to determine what prey sizes a predator is most likely to target to satisfy its demand. This information, combined with data on the abundance of predators of different size in a community, permits the calculation of a predation size spectrum (PSS) for the community. Specifically, this PSS can be plotted with body size of prey on the x-axis, and the potential consumption rate of each prey size by all predators in the community on the y-axis. Such a PSS may reveal the sizes of prey in a system that would be most heavily targeted by predators, given the predator size distribution. Time-series of predator size and abundance can be used to show how a PSS changes over time. A PSS models community predation with only few assumptions, compared with complicated community and ecosystem models such as Ecopath (Pauly et al., 2000).
An annual PSS for the northern Gulf of St Lawrence was calculated using predator size and abundance data for the fish community from surveys. Data were included for harp seals (Phoca groenlandica), because they are by far the most abundant marine mammal in the system and include a large proportion of fish in their diet. Changes in the PSS are then related to known changes in the structure of the fish community, especially with reference to trends in cod (Gadus morhua) abundance. The main purpose of constructing a PSS is to examine if changes in the Gulf of St Lawrence system, characterized by declines in predatory fish and increases in seals, led to changes in the sizes of prey targeted over time. For fish stock assessments, natural mortality of cod is assumed to have doubled after 1986, yet there is little information independent of cohort reconstruction models to verify this assumption. The PSS is a simple method to examine how predation on different sizes of pre-recruit cod has changed.
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
An annual summer stratified random sampling survey (Doubleday, 1981) has been conducted in the northern Gulf of St Lawrence since 1984. Gear and vessel were changed in 1991, after a calibration year in 1990. Data used here are from 1984 to 2003, and corrected data from 1984 to 1990, for differences in sampling efficiency between the two vessels. The survey usually consists of some 200 hauls annually, each of 20-min duration, with a small-mesh (19 mm) otter trawl. Despite differences in length-based catchability among species, the trawl is considered to yield fairly good estimates of relative abundance of fish >10 cm (Harley and Myers, 2001; Savenkoff et al., 2004).
Harp seal data were updated from Hammill and Stenson (2000). This data set (1960–2002) is a reconstruction of harp seal abundance, but takes harvest and modelled pupping rates into account.
Predator size spectra were constructed (Equations 1–6) by first calculating (Equation 1) the food required by a single average predator in each size class (EW), then multiplying the abundance of predator size (NW) to obtain an estimate of the total food required (QW). The range of W is taken as the size range of fish caught in the survey. The proportion of QW derived from prey size 
is then determined (Equation 4) from the lognormal predator–prey size ratio (
) following the general form of Equation (2), where the sum of preference over the prey size range is normalized over the prey size range considered (Equation 3, where 0.01 g < I < W). This is tantamount to saying that a predator W eats only within this prey size range. Predation on each prey size i by all predators W (Qi) is then calculated by summing 
over all W (Equation 5). Finally, a relative Qi (RQi) is obtained by dividing each Qi by the sum of all Qi (Equation 6), the PSS being the RQi values plotted against Wi:
|
| (1) |
|
| (2) |
|
| (3) |
|
| (4) |
|
| (5) |
|
| (6) |
) = 2 was assumed, well within the observed range for North Sea cod (Floeter and Temming, 2003). For harp seals, I used R = 10 (=log2 1000; Hammill and Stenson, 2000), with s.d. = 3.32. Standard deviations were chosen such that the range of possible ratios was large, but the probability of ratios near unity was low. In the PSS, size of individuals is presented as log2 because of precedents set in seminal size spectrum studies (Sheldon et al., 1972; Platt and Denman, 1977). One of the useful aspects of the use of log2 integer bins is that it gives the greatest number of size classes (and hence resolution) in a logged size spectrum, while retaining an integer base value. Also, because the nominal size of a class is equal to its width only in a log2 size spectrum, it simplifies normalizing size spectra with a single matrix operation.
Daily consumption rate (E; kg individual–1 d–1) for fish was based on an allometric formulation (Magnússon, 1995) used in MSVPA for North Sea cod (ICES, 1998):
|
| (7) |
The daily consumption rate for harp seals was taken from Hammill and Stenson (2000):
|
| (8) |
Feeding preferences for each predator size were assigned on the basis of the log-normal R curve. The range of prey sizes considered was 0.01 g–100 kg wet weight, to cover all possible fish sizes, and these were split in integer log2 prey size class bins (g). The preference for each prey size class was then normalized to total prey consumption.
For each predator size group encountered in the data, the calculated food consumption of one predator was distributed over a prey size range according to the corresponding preference function. This assumes that each predator meets its required ration by feeding according to its prey size preference. The final (relative) PSS was constructed by multiplying the predation for each predator of each size by the relative abundance of that predator from either survey data or harp seal abundance data. The relative PSS can easily be transposed into an absolute PSS if reliable estimates of absolute predator abundance at size are available.
The model was implemented in Splus 6.1, and is available from the author.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Most predation by fish is targeted at two prey size classes, spanning sizes of about 0.5–2 g (Figure 1a), which roughly corresponds to 4–6.5 cm for a cod-like fish. Before the collapse of the northern Gulf of St Lawrence cod stock in 1993 and the fishing moratorium in 1994, predation pressure on the small fish size range was considerably greater than thereafter. Total predation by fish peaked in 1987 and, though steeply declining, remained quite high until about 1990.
|
Most seal predation is targeted at prey of 60–125 g (Figure 1b), roughly corresponding to 18–23-cm cod-like fish. This accords well with the cod sizes reconstructed from otoliths found in harp seal stomachs (Hammill and Stenson, 2000). Figure 1b also suggests, however, that harp seals would inflict a substantive though lesser predation pressure on larger and smaller prey.
|
Discussion |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
The PSS method provides an alternative to detailed modelling of predation, and may aid directly in single species assessment by informing decisions on changing natural mortality values input in cohort reconstruction approaches. Furthermore, predation size spectra represent a general indicator of the predation environment experienced by all fish that can be used to uncover major changes in the predation patterns of an ecosystem. This could be a valuable tool for evaluating if and when a community has made a large-scale state transition (Scheffer et al., 2001), because the predation environment would likely change in conjunction with state.
In this particular application, however, there are three main players, cod, harp seals, and the fisheries. The northern Gulf of St Lawrence cod fishery was closed in 1994, and only small fisheries have been allowed ever since (CSAS, 2004), although F may still be high, because the stock biomass is small. For the period from 1991 onwards, recruitment has been consistently low, even though spawning-stock biomass (SSB) varied almost sixfold over the period (CSAS, 2004), suggesting that recruitment is a function of more than SSB. In stock assessment, natural mortality of post-recruit cod is assumed to have doubled since 1986, to M = 0.4 (CSAS, 2004). It is conceivable that the reasoning behind an increased M for post-recruits also applies to pre-recruits. For instance, the elevated predation attributable to increasing seal numbers could potentially create a significant increase in pre-recruit mortality. Although predation by fish on smaller pre-recruits has probably decreased considerably since 1984, owing to general declines in predatory fish abundance, the increase in the numbers of harp seals could have shifted the predation mortality to larger pre-recruits.
Cod begin to recruit to the fishery only at age 3, some 400–500 g, and are not fully recruited until age 8. Therefore, the high potential predation on pre-recruits shown by the PSS in the late 1980s may have contributed to depressed recruitment when the stock collapsed in 1993. Exacerbating the large F on an already depleted stock, predation of pre-recruit cod could have precluded the stock from making a large-scale recovery.
Predation size spectra represent simple mass-balance calculations. There is no starvation for predators similar to MSVPA (Magnússon, 1995), but perhaps more importantly, there is no prey switching. That is, a PSS models simply the sizes at which predators would have focused their predation if all prey were sufficiently abundant. In reality, predators may concentrate their effort on prey sizes that are smaller or larger than "ideal", should the local density of those alternative prey be sufficiently attractive for predators to switch (Kerr and Martin, 1970). The plots presented are perhaps best thought of as reflecting a "predation risk", rather than actual predation. Despite this limitation, the concept of predation risk is useful to interpret how a fish might experience its predation environment. In fact, this static predation risk may be more relevant than a dynamic model in making inferences on adaptive and evolutionary strategies of fish to avoid predation.
Fisheries are size-selective, and one of the most common observations on exploited stocks is that average size of fish decreases with increasing exploitation (Hilborn and Walters, 1992). Because predation is generally concentrated on the smaller sizes, the proportion of a fish stock that is vulnerable to predation is likely to increase with exploitation. This could be especially true when a major predator is a seal targeting relatively large pre-recruits that would have a high potential to recruit to the fishery. In systems with a great abundance of predators targeting smaller pre-recruits, the chance that those recruit to the fishery is anyhow lower, because they must endure a high mortality environment for a longer time. Hence, different units of predator biomass may affect recruitment disproportionally, and a nominally equal amount of predation changing from small to large prey is likely to reduce recruitment.
In contrast to the northern Gulf of St Lawrence, cod in the North Sea begin to recruit to the fishery at age 1, but around the same size of 400–500 g. Consequently, a North Sea cod spends considerably less time in the window of heavy predation. It is therefore conceivable that recruitment in the northern Gulf may be more strongly determined by the predation environment than say North Sea cod recruitment, which appears to depend more on sea temperature (Planque and Frédou, 1999).
If the prey size preference of harp seal is constant from year to year, and prey are always selected between 60 and 125 g, such peaks in predation could very well create a selection pressure on prey size. This suggests that it would be evolutionarily advantageous for prey either to avoid growing to the most intensively preyed upon size range, or to grow quickly out of that range. As fish predators strongly target prey between 0.5 and 2 g, there would be a small respite in predation mortality between 2 and 60 g (though both seals and fish will eat prey in this range, with lower preference). Therefore, the best strategy is for a fish to grow out of the pre-recruit size range as quickly as possible. Fisheries, on the other hand, create a selection pressure for fish that mature early and remain small (Olsen et al., 2004). Locked between the threats of a natural predation environment and selective fisheries, fish such as cod experience a variety of mortality gauntlets during their lives that may influence patterns of growth, size, and maturity in different directions.
|
Acknowledgements |
|---|
Red Méthot provided a table of survey abundance by species corrected for relative catchability. Mike Hammill provided data on the abundance, growth, and food requirements of harp seals in the Gulf of St Lawrence. I am also grateful for the valuable comments on the draft manuscript made by Niels Daan, Sarah Kraak, and Jason Link.
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
-
Benoît E. and Rochet M-J. (2004) A continuous model of biomass size spectra governed by predation and the effects of fishing on them. Journal of Theoretical Biology 226:9–21.[CrossRef][Web of Science][Medline]
CSAS. (2004) The northern Gulf of St Lawrence (3 Pn, 4RS) cod in 2003. Fisheries and Oceans Canada, Canadian Science Advisory Secretariat, Ottawa Stock Status Report, 2004/19.
Dickie L.M., Kerr S.R., Boudreau P.R. (1987) Size-dependent processes underlying regularities in ecosystem structure. Ecological Monographs 57:233–250.[CrossRef][Web of Science]
Doubleday W.G. (1981) Manual on Groundfish Surveys in the Northwest AtlanticScientific Council Studies, Northwest Atlantic Fisheries Organisation. 2.
Floeter J. and Temming A. (2003) Explaining diet composition of North Sea cod (Gadus morhua): prey size preferences versus prey availability. Canadian Journal of Fisheries and Aquatic Sciences 60:140–150.
Hammill M.O. and Stenson G.B. (2000) Estimated prey consumption by harp seals (Phoca groenlandica), hooded seals (Cystophora cristata), grey seal (Halichoerus grypus) and harbour seals (Phoca vitulina) in Atlantic Canada. Journal of Northwest Atlantic Fisheries Science 26:1–23.
Harley S.J. and Myers R.A. (2001) Hierarchical Bayesian models of length specific catchability of research trawl surveys. Canadian Journal of Fisheries and Aquatic Sciences 58:1569–1584.
Hilborn R. and Walters C.J. (1992) Quantitative Fisheries Stock Assessment: Choice, Dynamics and Uncertainty(Chapman and Hall, New York).
ICES. (1998) Report of the Multispecies Assessment Working Group. ICES Document, CM 1997/Assess: 16.
Kerr S.R. and Martin N.V. (1970) Trophic-dynamics of lake trout production systems. In Steele J.H. (Ed.). Marine Food Chains(Oliver and Boyd, Edinburgh) pp. 365–376.
Magnússon K.G. (1995) An overview of the multispecies VPA – theory and applications. Reviews in Fish Biology and Fisheries 5:195–212.[Web of Science]
Olsen E.M., Heino M., Lilly G.R., Morgan M.J., Brattey J., Ernande B., Dieckmann U. (2004) Maturation trends indicative of rapid evolution preceded the collapse of northern cod. Nature 428:932–935.[CrossRef][Medline]
Pauly D., Christensen V., Walters C. (2000) Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impacts of fisheries. ICES Journal of Marine Science 57:696–706.
Peters R.H. (1983) The Ecological Implications of Body Size(Cambridge University Press, Cambridge, UK).
Planque B. and Frédou T. (1999) Temperature and the recruitment of Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 56:2069–2077.
Platt T. and Denman K. (1977) Organisation in the pelagic ecosystem. Helgoländer Wissenschaftliche Meeresuntersuchungen 30:575–581.[CrossRef][Web of Science]
Savenkoff C., Bourdages H., Swain D.P., Despatie S.P., Hanson J.M., Méthot R., Morissette L., Hammill M.O. (2004) Input data and parameter estimates for ecosystem models of the southern Gulf of St Lawrence (mid-1980s and mid-1990s). Canadian Technical Report of Fisheries and Aquatic Sciences 2529:.
Scheffer M., Carpenter S.R., Foley J.A., Folke C., Walker B. (2001) Catastrophic shifts in ecosystems. Nature 413:591–596.[CrossRef][Medline]
Sheldon R.W., Prakash A., Sutcliffe W.H. (1972) The size distribution of particles in the ocean. Limnology and Oceanography 18:719–733.[Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. M. De Roos, T. Schellekens, T. Van Kooten, and L. Persson Stage-specific predator species help each other to persist while competing for a single prey PNAS, September 16, 2008; 105(37): 13930 - 13935. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. E. Duplisea and D. Robert Prerecruit survival and recruitment of northern Gulf of St Lawrence Atlantic cod ICES J. Mar. Sci., September 1, 2008; 65(6): 946 - 952. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Jennings, F. Melin, J. L Blanchard, R. M Forster, N. K Dulvy, and R. W Wilson Global-scale predictions of community and ecosystem properties from simple ecological theory Proc R Soc B, June 22, 2008; 275(1641): 1375 - 1383. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Jennings and A. S. Revill The role of gear technologists in supporting an ecosystem approach to fisheries ICES J. Mar. Sci., November 1, 2007; 64(8): 1525 - 1534. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Persson, P.-A. Amundsen, A. M. De Roos, A. Klemetsen, R. Knudsen, and R. Primicerio Culling Prey Promotes Predator Recovery--Alternative States in a Whole-Lake Experiment Science, June 22, 2007; 316(5832): 1743 - 1746. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||