© 2005 International Council for the Exploration of the Sea
Cod recruitment is strongly affected by climate when stock biomass is low
ICES H. C. Andersens Boulevard 44–46, DK-1533 Copenhagen V, Denmark
*tel: +45 33 33 38 67 00; fax: +45 33 93 42 15. e-mail: keith{at}ices.dk.
Stocks of Atlantic cod (Gadus morhua) have been declining over much of the North Atlantic for the past 30 years, owing to a combination of overfishing and adverse changes in their environment. In a previous study, environmental effects were introduced as an extra parameter in the stock-recruit relationship, where they act as a multiplier, independent of the level of spawning-stock biomass (SSB). Using a non-parametric pooled analysis of all cod stocks on the European Shelf south of 62°N, it is shown here that environmental variability (as represented by the North Atlantic Oscillation) only has a significant effect on recruitment when the spawning stock is low. This has implications for fisheries management strategies, and for rates of stock recovery, which will be very dependent on environmental conditions.
Keywords: climate change, cod, NAO, recruitment, stock recovery
Received 1 April 2004; accepted 26 July 2004.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
The poor state of exploited demersal fish stocks in European waters has been a subject of concern for some time (Cook et al., 1997). Spawning-stock biomass of cod (Gadus morhua) in the North Sea and other European waters is at a low level (Figure 1), and this has given rise to debate in scientific journals (Schiermeier, 2004), and in the media (Urquhart, 2003), about the causes of the decline and the measures required to reverse it. Unwarranted claims that fishing is not responsible for the decline (Urquhart, 2003) were made following recent work (Beaugrand et al., 2003), which showed that recruitment of North Sea cod is affected by changes in the abundance, seasonality, and size composition of zooplankton.
|
, and – represent high, intermediate, and low recruitment, respectively): (a) Eastern Baltic; (b) North Sea; (c) Western Baltic; (d) West of Scotland; (e) Irish Sea; (f) Celtic Sea; (g) NAO index (categories high, intermediate, and low are demarcated by the dashed lines). Note that the NAO categories shown are based on ranking the full time-series. Where the time-series available for a stock is shorter, this will affect the ranking and categorization of the NAO.Cod in European waters south of 62°N are assessed and managed as six separate stocks (Eastern Baltic, Western Baltic, West of Scotland, Irish Sea, Celtic Sea, and North Sea), which vary considerably in their spawning-stock biomass (Figure 1) and annual landings (ICES, 2002). The stocks are independent of each other, with separate spawning areas, different migration patterns, and low exchange rates between them, based on evidence from tagging and genetic studies (Brander, 1994). However, annual recruitment to the six stocks shows strong similarity: in an analysis using paired correlations, 8 of 15 were significant, four at p < 0.01 and four at p < 0.05 (Brander, 2003a). This is thought to be due to their common response to environmental forcing, as represented by a negative relationship with temperature (Planque and Fredou, 1999; Brander, 2000), or with the North Atlantic Oscillation (NAO) index (Brander and Mohn, 2004), during their pelagic phase.
The NAO is an indicator of climate variability, with well-documented effects on the physical and biological environment (Hurrell et al., 2003). It affects the recruitment of most cod stocks, including those in European Shelf seas south of 62°N (i.e. around the British Isles and in the Baltic; Ottersen et al., 2001; Brander and Mohn, 2004), via temperature effects on growth and survival of larvae (Ottersen et al., 1994), or less directly by altering the timing, magnitude, and size composition of zooplankton production on which cod larvae depend (Beaugrand et al., 2003).
A previous study modelled recruitment for each stock separately, using a modified form of the Ricker stock-recruitment relationship (Ricker, 1954), with an extra parameter for the environmental variable (Brander and Mohn, 2004). The environmental effect acts as a multiplier, which is independent of the size of the spawning stock. It is difficult to test alternative, more complex models, because the time-series available for each stock are relatively short (longest for the North Sea, with 38 years), and data on spawning-stock biomass and recruitment are noisy. However, as Brander and Mohn (2004) showed that recruitment for the six European stocks exhibits a common response to the NAO, a pooled analysis seems justified, in order to test whether the effects of environmental variability are indeed independent of the level of spawning-stock biomass.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The area occupied by the six cod stocks varies by more than a factor of ten, and their spawning-stock biomass (SSB) and annual number of recruits, which were estimated by virtual population analysis, vary by comparable amounts (ICES, 2002). Ranking and categorization were used to overcome the difference in scale between stocks (Rothschild and Mullen, 1985). SSB and recruitment for each stock were ranked, and the rankings were divided into three equal categories, i.e. high, intermediate, and low levels of SSB and recruitment. The observed frequencies, based on raw counts of ranked data, were analysed using a simple, non-parametric method.
The NAO index was used to represent climate variability because it affects recruitment of these European Shelf stocks in a consistent way (Brander and Mohn, 2004). The time-series used (Figure 1) is the winter mean (December–March) pressure difference between southwest Iceland and Gibraltar (Osborn, 2004). Data for the six stocks all fall within the range of years from 1963 to 2001, and use the same annual winter NAO values, so there is a degree of pseudo-replication. For each stock, the winter NAO values were ranked over the range of years for which data on spawning-stock biomass and recruitment are available, and the rankings were divided into three equal categories (high, intermediate, low).
The choice of a frequency-based method of analysis, and the way in which the results are shown, was influenced by presentational criteria to make them easy to understand and to show the observational base from which the conclusions are drawn. Research in cognitive psychology has shown that information is more easily understood when presented as natural frequencies (Gigerenzer, 2002), and its quality is more transparent to lay judgement. More sophisticated statistical analysis may be justified if the increase in interpretative insight outweighs the loss in comprehensibility; the aim here is to provide results that are useful to fisheries managers.
The time-series structure of SSB, level of recruitment, and NAO (Figure 1) has not been taken into account in the analysis. Only one of the six stocks (Eastern Baltic) has significant autocorrelation of recruitment at a lag of 1 year (Brander, 2003a). The six stocks provide 200 years of data, which were analysed as a 3*3*3 table of frequencies. The frequencies of high, intermediate, and low recruitment are shown in Figure 2 for each category of spawning-stock biomass and NAO level. For example, the top left cell in Figure 2a shows that, of 200 observations, the combination of low SSB and high NAO occurred 34 times, and high recruitment in only one of those 34. The top left cell of Figure 2b shows that, for the same 34 occurrences of low SSB and high NAO, intermediate recruitment occurred nine times, and the top left cell of Figure 2c shows the remaining 24 observations, which were of low recruitment.
|
A
2 test was applied to each of the three levels of spawning-stock biomass in the 3*3*3 table of observed and expected frequencies, to test whether recruitment was independent of NAO. The lowest expected value in the 2 test (3.67) is sufficient for accurate test statistics (Snedecor and Cochrane, 1971). |
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The
2 test gives the probability that recruitment is independent of the NAO; these probabilities are <0.001, <0.1, and >0.5 for low, intermediate, and high SSB, respectively. The hypothesis of independence from NAO effects is therefore rejected when SSB is low, but it is not rejected when the SSB is intermediate or high. High recruitment is more frequent at medium or high levels of SSB than at low levels (Figure 2a), and low recruitment is more frequent at low levels of SSB (Figure 2c). At low SSB, the frequency of high recruitment is greater when the NAO is low, and conversely for the frequency of low recruitment. Although the effects of the NAO are not statistically significant at intermediate and high levels of SSB, they appear to be present to some degree. |
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The environment, as represented by the NAO winter index, affects recruitment of European Shelf cod stocks more strongly when their spawning biomass is low. This has consequences for their management, and for models that are used to carry out short- and long-term projections. Models of spawning stock and recruitment in which the environmental effects are independent of SSB are probably not appropriate. The strategy for managing European cod stocks should take into account their increased sensitivity to environmental variability when SSB is low, and this is an additional, powerful justification for avoiding low stock biomass.
A semi-quantitative short-term projection of recruitment can be made using the natural frequency information given in Figure 2. For example, the NAO for the winter of 2003/2004 is –0.63 (average for December–February; Osborn, 2004), which puts it in the low category (Figure 1g). The current levels of SSB for European cod stocks are generally low (Figure 1). Recruitment to these stocks has been high in 3 years of 11 when SSB and NAO were both low. In contrast, recruitment was high in just 1 year of 34 when SSB was low and the NAO high. Therefore, when SSB was low, high recruitment was about nine times more frequent when the NAO was low than when it was high. The prospects for the 2004 year classes of cod in European waters are less bad than they might be.
The long-term prospect for recruitment to these stocks is not favourable, because the underlying trend in the NAO has been upwards over the past 40 years. Most climate models simulate a continuing increasing trend in the NAO (Gillett et al., 2003). Models of fish populations that do not take account of systematic changes in the NAO (or other environmental indicators, such as temperature) may provide misleading information for developing fisheries management strategies, or for assessing the prospects for stock recovery. There are well-developed frameworks for carrying out evaluations of the risk and uncertainties arising from climate change (Willows and Connell, 2003). Regularly updated scenario models are used to assess future management options in many areas of economic activity (e.g. flood and coastal defence, insurance, water supply, health, tourism), and fisheries management would benefit from applying such frameworks and scenarios.
Myers (2001) used a meta-analysis of more than 700 fish populations to show that recruitment variability increased at low stock biomass, which he ascribed to strong density-dependent mortality at the juvenile stage. The current analysis provides a different, environmental explanation for European cod stocks. A number of complementary processes may be acting to increase the sensitivity of survival during the early life stages to environmental variability, when SSB is low. Increased mortality (by fishing or other causes) reduces SSB, life expectancy, and the number of older, larger fish that make a greater contribution to reproductive output (Marteinsdottir and Thorarinsson, 1998). The resultant spawning population, with a lower mean age of spawners, and fewer age classes, has a shorter spawning season and a smaller range in specific gravity of eggs. Both effects reduce the distribution of early life stages in space and time, and may make them more vulnerable to variability in environmental conditions. Smaller spawning stocks may also occupy a reduced spawning area, which again reduces the distribution of early life stages, and increases vulnerability to environmental factors (Begg and Marteinsdottir, 2002).
The analysis presented here does not attempt to estimate the relative impact of fishing and environmental variability in bringing about the current low levels of SSB in these six cod stocks; both have played a part. However, it does show that the frequency of good recruitment, which is a major contributor to stock recovery, is now much more dependent on favourable environmental conditions, specifically low values of the winter NAO index.
The upward trend in the NAO over the past four decades, in combination with heavy exploitation, has contributed to a downward trend in recruitment. An analysis by Sparholt (2002) of ICES stock assessments and forecasts of North Sea and Baltic cod since 1988 found that SSB was consistently greatly overestimated. Those stock assessments use estimates and forecasts of recruitment based on data for previous years, and will therefore generally overestimate recruitment during a period in which it has been declining. Sparholt (2002) questioned whether climatic factors might be partly implicated in the systematic bias that he described, and this paper provides one possible process by which this could happen.
Short- and long-term assessments of European cod stocks should take account of the current level and trends in the NAO and/or related environmental indices. Further work is needed on parametric or non-parametric forms of the stock-recruitment relationship, in order to ensure that environmental effects are modelled in an appropriate manner (Ciannelli et al., in press). Development of fisheries management strategies should use the framework and scenario models, which have been developed for assessing risk and uncertainty arising from climate change in many other economic activities (Willows and Connell, 2003). The climate scenarios that are currently produced for evaluating terrestrial activities (agriculture, water supply, etc.) often lack the detail and information needed to evaluate impact on marine populations (Parry, 2000). The marine science community needs to make a case to climate modellers for such information (e.g. sea temperature, location of mesoscale features) to be provided routinely as part of future scenarios, if they are to be useful for evaluating climate impacts on marine systems (Brander, 2003b). Ignoring the effects of environmental variability and climate change on fish stocks and fisheries management is not an option, or rather, is equivalent to making the implicit assumption of status quo. This is not a sensible assumption in the case of European cod stocks, for which our knowledge of recruitment dynamics has been obtained during a period of unprecedented increase in the NAO, from its lowest state during the 1960s (Gillett et al., 2003).
|
Acknowledgements |
|---|
This work was funded by the US National Science Foundation, the UK Department of the Environment, Food and Rural Affairs, the Norwegian Research Council, and the Canadian Department of Fisheries and Oceans. I also thank Henrik Sparholt, Brian Rothschild, and Bob Mohn for helpful suggestions.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Beaugrand G., Brander K.M., Lindley J.A., Souissi S., Reid P.C. (2003) Plankton have effect on cod recruitment in North Sea. Nature 426:661–664.[CrossRef][Medline]
Begg G.A. and Marteinsdottir G. (2002) Environmental and stock effects on spatial distribution and abundance of mature cod (Gadus morhua). Marine Ecology Progress Series 229:245–262.[Web of Science]
In Brander K. M. (Ed.). Spawning and life history information for North Atlantic cod stocks. (1994) ICES Cooperative Research Report, 205: 150 pp.
Brander K.M. (2000) Effects of environmental variability on growth and recruitment in cod (Gadus morhua) using a comparative approach. Oceanologica Acta 23:485–496.[Medline]
Brander K.M. (2003) What kinds of fish stock predictions do we need and what kinds of information will help us to make better predictions? Scientia Marina 67:Suppl. 1, 21–33.[Web of Science]
Brander K. (2003) Fisheries and climate. In Wefer G., Lamy F., Mantoura F. (Eds.). Marine Science Frontiers for Europe(Springer, Berlin) pp. 29–38.
Brander K. and Mohn R. (2004) Effect of North Atlantic Oscillation (NAO) on recruitment of Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 61:1558–1564.
Ciannelli L., Chan K. S., Bailey K. M., Stenseth N. C. Non-additive effects of environmental variables on the density-dependent survival of a large marine fish population. Ecology (in press).
Cook R.M., Sinclair A., Stefansson G. (1997) Potential collapse of North Sea cod stocks. Nature 385:521–522.[CrossRef]
Gigerenzer G. (2002) Adaptive Thinking: Rationality in the Real WorldOxford University Press.
Gillett N.P., Graf H.F., Osborn T. (2003) Climate change and the North Atlantic Oscillation. In Hurrell J.W., Kushnir Y., Ottersen G., Visbeck M. (Eds.). The North Atlantic Oscillation: Climatic Significance and Environmental Impact(American Geophysical Union, Washington, D.C.) pp. 193–209.
Hurrell J.W., Kushnir Y., Ottersen G., Visbeck M. (2003) The North Atlantic Oscillation: Climatic Significance and Environmental Impact(American Geophysical Union, Washington D.C.) 279 pp.
ICES. (2002) Report of the ICES Advisory Committee on Fishery Management, 2002 (Parts 1–3). ICES Cooperative Research Report, 255: 948 pp.
Marteinsdottir G. and Thorarinsson K. (1998) Improving the stock-recruitment relationship in Icelandic cod (Gadus morhua L.) by including age-diversity of spawners. Canadian Journal of Fisheries and Aquatic Sciences 55:1372–1377.
Myers R.A. (2001) Stock and recruitment: generalizations about maximum reproductive rate, density dependence and variability. ICES Journal of Marine Science 58:937–951.
Osborn T. (2004) NAO update. www.cru.uea.ac.uk/~timo/projpages/nao_update.htm.
Ottersen G., Loeng H., Raknes A. (1994) Influence of temperature variability on recruitment of cod in the Barents Sea. ICES Marine Science Symposia 198:471–481.
Ottersen G., Planque B., Belgrano A., Post E., Reid P.C., Stenseth N.C. (2001) Ecological effects of the North Atlantic Oscillation. Oecologia 128:1–14.[CrossRef][Web of Science]
In Parry M.L. (Ed.). Assessment of Potential Effects and Adaptations for Climate Change in Europe: The Europe ACACIA Project (2000) (Jackson Institute, University of East Anglia, Norwich, U.K.) 320 pp.
Planque B. and Fredou T. (1999) Temperature and the recruitment of Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 56:2069–2077.
Ricker W.E. (1954) Stock and recruitment. Journal of the Fisheries Research Board of Canada 11:559–623.
Rothschild B.J. and Mullen A.J. (1985) The information content of stock-and-recruitment data and its non-parametric classification. Journal du Conseil International pour l'Exploration de la Mer 42:116–124.
Schiermeier Q. (2004) Climate findings let fishermen off the hook. Nature 428:4.[Medline]
Snedecor G.W. and Cochrane W.G. (1971) Statistical Methods(Iowa State University Press, Ames, Iowa).
Sparholt H. (2002) Quality of ACFM advice: how good have forecasts been since 1988? Working Document 5 in Report of the Working Group on Methods of Fish Stock Assessment. ICES Document, CM 2002/D: 01.
Urquhart F. (2003) Fishing "not to blame for cod shortage". The Scotsman, 12-12-2003.
Climate Adaptation: Risk, Uncertainty and Decision Making. In Willows R. I. and Connell R. K. (Eds.). UKCIP Technical Report (2003) (UKCIP, Oxford).
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. G. Pitois and C. J. Fox Empirically modelling the potential effects of changes in temperature and prey availability on the growth of cod larvae in UK shelf seas ICES J. Mar. Sci., December 1, 2008; 65(9): 1559 - 1572. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Howden, J.-F. Soussana, F. N. Tubiello, N. Chhetri, M. Dunlop, and H. Meinke Climate Change and Food Security Special Feature: Adapting agriculture to climate change PNAS, December 11, 2007; 104(50): 19691 - 19696. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. O Hjermann, B. Bogstad, A. M. Eikeset, G. Ottersen, H. Gjosaeter, and N. C. Stenseth Food web dynamics affect Northeast Arctic cod recruitment Proc R Soc B, March 7, 2007; 274(1610): 661 - 669. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Pitois and C. J. Fox Long-term changes in zooplankton biomass concentration and mean size over the Northwest European shelf inferred from Continuous Plankton Recorder data ICES J. Mar. Sci., January 1, 2006; 63(5): 785 - 798. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Babcock, E. K. Pikitch, M. K. McAllister, P. Apostolaki, and C. Santora A perspective on the use of spatialized indicators for ecosystem-based fishery management through spatial zoning ICES J. Mar. Sci., January 1, 2005; 62(3): 469 - 476. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||