© 2005 International Council for the Exploration of the Sea
Do climate and fishing influence size-based indicators of Celtic Sea fish community structure?
Centre for Environment, Fisheries and Aquaculture Science Pakefield Road, Lowestoft, Suffolk NR33 OHT, England, UK
*Correspondence to J. L. Blanchard: tel: +44 1502 527701; fax: +44 1502 524511. e-mail: j.l.blanchard{at}cefas.co.uk.
Ecosystem-based management requires the development of indicators that allow anthropogenic impacts to be detected against the background of natural variation. Size-based community metrics are potentially useful indicators because of their theoretical foundation and practical utility. Temporal and spatial patterns in size-based community metrics for Celtic Sea fish are described and calculated using data from the English groundfish survey of the area (1987–2003). The results reveal that the size structure of the community has changed over time, and that a decrease in the relative abundance of larger fish was accompanied by an increase in smaller fish (4–25 g). Temporal analyses of the effects of fishing and climate variation suggest that fishing generally has had a stronger effect on size structure than changes in temperature. Therefore, size-based metrics respond clearly to the effects of fishing even in variable environments, reflecting the ubiquity of size-based processes in defining community structure and responses to mortality. Spatial analyses were inconclusive, probably owing to the limited area for which fishing effort, temperature, and survey data were all available.
Keywords: Celtic Sea, climate, ecosystem indicators, fish communities, fishing effects, size spectrum
Received 1 April 2004; accepted 3 December 2004.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussionReferences |
|---|
While the ecosystem approach to fisheries management is now being pursued worldwide (Garcia and Cochrane, 2005), it has proven challenging to identify suitable metrics of communities and ecosystems that are both quantifiable and theoretically defensible (Rice, 2000, 2003; Rochet and Trenkel, 2003; Trenkel and Rochet, 2003). Size-based approaches have such a theoretical foundation and also great practical utility for detecting the effects of exploitation upon fish communities (Shin et al., 2005). Analyses of size spectra in particular have been suggested as a powerful method, because the slope changes over time in a manner consistent with exploitation-mediated changes in community structure (Pope et al., 1988; Murawski and Idoine, 1992; Bianchi et al., 2000; Zwanenburg, 2000). However, the interaction between, and relative importance of, environmental effects and exploitation on size spectra is difficult to quantify, and both factors may lead to changes in size structure. In addition, gear type, season, and the spatial extent of surveys produce significant differences in size-based metrics across given time-series (Trenkel et al., 2004; Daan et al., 2005).
For size-based metrics to be meaningful indicators in a management context, they should allow for discrimination between impacts of different forcing factors acting upon community structure, such as exploitation, pollution, and climate variation and change. Large-scale commercial fisheries expanded comparatively recently in the Celtic Sea and, unusually, their development was paralleled by the implementation of fishery monitoring and survey programmes (Pinnegar et al., 2002). The temperature and plankton community in the area is strongly influenced by decade-scale shifts in the North Atlantic Oscillation (Beaugrand et al., 2000, 2002). Consequently, the Celtic Sea represents an ideal system for studying the influences of exploitation and climate on fish community structure (Pinnegar et al., 2002).
Here, we examine temporal and spatial patterns in three commonly used size-based metrics (average weight of an individual; average maximum size, Lmax; and size spectrum slope; Rice and Gislason, 1996; Duplisea et al., 1997; Jennings et al., 1999; Bianchi et al., 2000; Rochet and Trenkel, 2003) of the Celtic Sea fish community, and relate them to measures of fishing effort and sea temperature (as a proxy for climate variability).
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Data sources
Size-abundance data by species from the annual CEFAS Celtic Sea groundfish surveys (Warnes and Jones, 1995) were used to calculate size-based metrics. Only locations sampled with the standard Portuguese high-headline trawl in >90% of years for the temporal analyses, and in all years for the spatial analyses, were used. The resultant time-series spanned 1987–2003 for 47 and 39 stations, respectively.
Winter and summer sea surface temperatures (SST; mean values January–March and June–August) for each year were obtained from http://www.cdc.noaa.gov/coads/. Gridded spatial data (1° longitude by 0.5° latitude) for SST and near-bottom temperature (NBT) for winter and summer were obtained from the ICES database.
Multispecies fishing mortality indices (F) were calculated as the biomass-weighted mean F for (i) all species assessed, and (ii) demersal species only (cf. Daan et al., 2005; F obtained from ICES, 2003b–d). Spatial fisheries surveillance data (obtained from the UK Department of Environment, Food and Rural Affairs) were standardized for sightings effort (aircraft visits per ICES rectangle per unit time).
Size-based metrics
Average weight of an individual fish in the catch was calculated as the sum of the catch weights divided by the total number of fish caught (Zwanenburg, 2000). Average maximum length of the fish in the community (Lmax) was calculated as in Jennings et al. (1999) and ICES (2003a).
The original data were standardized catch numbers by length category. Individual lengths were transformed to weights using species-specific length–weight regression coefficients, where possible (Bedford et al., 1986; Dorel, 1986; Coull et al., 1989); otherwise, a standard equation was used (W = 0.01L3). Size spectra were standardized to remove the correlation between slope and intercept by subtracting the mean from the independent variable (Trenkel and Rochet, 2003). As our interest focused on that part of the spectrum that could be approximated by a linear slope, only body-mass classes above a threshold mass of 4 g were included in the analysis. Slopes and intercepts of the normalized log2 biomass (y-axis) against the midpoints of log2 body-mass classes (x-axis) for each year were estimated by linear regression.
All metrics were calculated for all species caught and for demersal species only, and analyses were carried out separately for each group.
Analyses
Positive or negative trends in the size-based metrics and log-abundance over time were evaluated by comparing the rank correlation statistic for each time-series (Spearman's correlation coefficient, rs). To test for relative effects of F and temperature, multiple linear regression was used. All variables were normalized (subtracting the mean and dividing by the standard deviation) before analysis to ensure that each series had a mean of zero and comparable variance. For each size-based metric (y-variable) and each pair of x-variables (F; and each temperature series of winter and summer SST with 0-, 1-, and 2-year lags, to allow for delayed effects of environmental change on biological processes), a combination of forward and backward selection (ANOVA, F-test) of the following models was used:
|
| (1) |
|
| (2) |
|
| (3) |
|
| (4) |
Spatial analyses of size-based metrics were restricted to biomass spectrum slopes for all fish in the community. Survey data were pooled over time, and linear fits to normalized log2 biomass against log2 body mass were made for each station sampled over the period 1987–2003. Spatial maps of winter and summer SST and near-bottom temperature (NBT), and standardized sightings of fishing vessels (as a proxy for fishing effort), were produced by kriging (in the R package "spatial"; Venables and Ripley, 2002). Pearson's correlation coefficients between each of the predicted kriged surfaces (at the survey stations) and slopes were calculated after taking account of spatial autocorrelation, using software provided by Legendre (2000). With the exception of the latter, all analyses were carried out using R version 2.0.0.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Trends over time (Figure 1) in average weight (rs = –0.57, p < 0.05), average Lmax (rs = –0.45, p < 0.1), and slopes (rs = –0.69, p < 0.01) for all fish caught were all negative. However, all patterns exhibited a "dip" during the years 1993–1996. Metrics based on demersal fish only exhibited similar declines in average weight (rs = –0.54, p < 0.05), Lmax (rs = –0.59, p < 0.05), and slopes (rs = 0.53, p < 0.05). Winter SST increased significantly (Figure 2a; rs = 0.50, p < 0.05), whereas summer SST exhibited similar interannual variation but the temporal trend was not significant (p > 0.1). The F for all commercially exploited stocks increased significantly over the period (rs = 0.74, p < 0.01), but it appeared to be levelling off in recent years; demersal F also increased significantly (rs = 0.76, p < 0.005), particularly after 1994, and was less than total F (Figure 2b).
|
|
Relative biomass of log2 weight classes revealed an increase in small fish (all categories <100 g) over time, which was accompanied by a reduction in large fish (all categories >100 g; Figure 3). However, only the increases in the three categories below 25 g and the declines in two of the larger ones (100–121 g; 144–169 g) were significant (p < 0.05).
|
Linear regression of each normalized y against each single explanatory x-variable revealed significant negative relationships between all species F and average Lmax (r2 = 0.44, p < 0.001) as well as F and slope (r2 = 0.33, p < 0.05). For demersal fish, the relationships were similar, but weaker (Lmax, r2 = 0.21, p < 0.1; slope, r2 = 0.20, p < 0.1). Weak negative relationships between slope and 2-year lagged winter SST (all species, r2 = 0.18, p < 0.1; demersal species, r2 = 0.21, p < 0.1) and between Lmax and summer SST (demersal species, r2 = 0.21, p < 0.1) were also detected. When the temperature series were restricted to the years for which F estimates were available, the relationships were no longer significant. Serial autocorrelation was not present in any of the variables, so should not have affected significance levels.
For the whole community sampled by the survey trawl, F had a stronger effect on mean Lmax and slope than temperature. For mean Lmax, adding winter (1-year lag) and summer (2-year lag) SST series to the model increased the residual variance explained from 44% (F only) to 60% and 72%, respectively. ANOVA results from the multiple regression models showed that for each combination of SST and F, when F was the first variable, addition of F alone led to the best improvement over a null model (1) when the y-variable was slope (all species). This was also the case for Lmax, except when winter SST (1-year lag; p < 0.1) or summer SST (2-year lag; p < 0.01) was added. However, the interaction terms were not significant.
Although F did not explain a significant amount of the variance in average weight (all species), addition of F with winter SST (1-year lag, with interaction term) or summer SST (2-year lag) did. Addition of F with winter SST (1-year lag) and an interaction term led to a marginally significant improvement over the null model, because of the weak significance of the interaction alone (p < 0.1). Similarly, a model including F and summer SST (2-year lag) was marginally better than the null model (p < 0.1) when F alone was not significant (p > 0.1), but the interaction term was not significant. Adding temperature first did not affect these results. There were two cases, however, where adding temperature as the first x-variable and F as the second resulted in F no longer being a significant term (and vice versa). These were the combinations of slope and winter SST (2-year lag) for all species and demersal species only. The reduction in the residual sums of squares (RSS) was greater when F was the explanatory variable (32%) rather than SST (21%) for all species, whereas SST reduced RSS by 27% and F by 20% for demersal species alone. For all other models, when SST was added as the first variable to predict a y-variable that had strong F effects, F was still significant as a second variable.
Fitting biomass spectra spatially (by sampling station) revealed that the steepest slopes were located in the eastern and western (along the shelf edge) parts of the Celtic Sea (Figure 4). After correcting for spatial autocorrelation, there were no significant correlations between slopes and any of the predicted (kriged) values at the stations for the spatial temperature series. A proxy for thermal stratification (the difference between predicted SST and NBT at each station) in winter and summer was not significantly correlated with slopes. Also, slope was not significantly correlated with fishing intensity for positions that fell within the limited range of the UK surveillance area (Figure 5).
|
|
|
Discussion |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Patterns in size-based metrics are evident both spatially and temporally. Declines observed over time in mean weight, mean Lmax, and slopes (all species) are associated with a reduction in the abundance of large fish and an increase in small fish. For the whole community sampled by the survey trawl, F had a stronger effect on mean Lmax and slope than temperature. Although F did not explain a significant amount of the variance in average weight, addition of F with winter SST (1-year lag, with interaction term) or summer SST (2-year lag) did. For the demersal component of the fish community, F had a weaker effect on the size-based metrics, possibly reflecting the lower demersal species F and the partial sampling of interacting pelagic and demersal communities. In general, the results for the whole community demonstrate that fishing has a relatively large and consistent impact on size-based metrics. Even in the Celtic Sea, where environmental forcing is unusually strong (Southward et al., 1988), fishing effects on size-based metrics can be disentangled from environmental effects.
Changes in F and temperature will affect the size structure of communities on a range of time scales. Changes in F reflect the direct effects of mortality, but our analyses provide evidence for longer-term indirect effects, because the biomass of fish in small size classes increased with F, consistent with studies of the response of size spectra to fishing in the North Sea (Daan et al., 2005) and Fiji (Dulvy et al., 2004). Changes in temperature at a given sampling location may lead to changes in the size structure because of (i) immigration and/or emigration of species with different temperature preferences, (ii) temperature effects on the life history of resident species, and (iii) indirect effects on biological processes that support the fish community. These effects on size structure may emerge on a range of time scales, and moreover, different size classes may be affected on different time scales. For example, good recruitment of an abundant species as a consequence of improved conditions for larval survival may affect the size spectrum within 1 year, whereas the effects of temperature on asymptotic body size may not be manifest for several years. Therefore, any attempt to link changes in size-based metrics to changes in temperature assuming an instantaneous response or a fixed lag is a crude one, but this is unavoidable given the lack of detailed information on biological responses of all species and size classes to temperature. Moreover, it is questionable whether the temporal and spatial scales over which the explanatory variables were tested best reflect the effects of fishing and temperature: spatial data on fishing intensity were incomplete, and the annual set of spatial temperature data were not entirely adequate.
Because our sampling stations are fixed in space, changes in metrics attributable to the environment may represent a change in the community present at the station (owing to environmental effects on distribution) or an effect on the dynamics of the populations constituting the community. Distributional changes in response to climate are well known in the Celtic Sea. Southward et al. (1988) reported shifts in the distribution of pilchard (Sardina pilchardus; a warm-water species) and herring (Clupea harengus; a cold-water species) over the past 400 years, corresponding to long-term temperature fluctuations. Lagged responses to climate may also be driven by conditions affecting growth, survival, and food availability during early life history, including temperature and timing of algal blooms (Planque and Frédou, 1999; O'Brien et al., 2000; Platt et al., 2003), or other processes directly related to the North Atlantic Oscillation (Dippner, 1997; Attrill and Power, 2002). Determining whether and how these changes should result in altered size structure still needs to be addressed.
The body size distribution of animals in foodwebs reflects patterns of energy use and acquisition, so the slopes of size spectra are remarkably constant in many ecosystems (Boudreau and Dickie, 1992). Although temperature will have a marked effect on biomass turnover and energy flux in the system, the slope is an emergent property that is largely temperature-independent (Brown et al., 2004). This implies that the slopes of time-averaged size spectra for the entire foodweb are sensitive to size-selective mortality rather than to temperature, and should be reliable indicators of fishing impacts at the scale of the foodweb. In practice, however, trawls sample specific assemblages within the foodweb, and the size compositions of the samples reflect (i) gear selectivity, (ii) spatial distribution of individuals, (iii) short-term dynamics of populations, as well as (iv) part of the underlying structure of the foodweb. Theoretical understanding of the responses of size-based metrics to fishing is based primarily on changes in the underlying structure of the foodweb (iv) owing to effects of size-selective mortality, but these effects are increasingly more difficult to resolve at progressively smaller spatial and temporal scales because the environment affects factors (i)–(iii). This may explain the failure of size-based metrics to provide an effective indicator of fishing effects at small spatial and temporal scales (such as an ICES rectangle), even though they may be reliable indicators of fishing effects at larger scales (Piet and Jennings, 2005).
|
Acknowledgements |
|---|
We thank Niels Daan and members of the ICES Working Group on Fish Ecology for valuable discussions, Brian Rackham (CEFAS, Lowestoft) for data extraction, and the many scientists, officers, and crew that participated in the Celtic Sea groundfish surveys aboard the RV "Cirolana". Thanks are also due to Kim Stobberup and Anne Hollowed for reviewing the manuscript, to David Maxwell for statistical advice, and to Paul Eastwood and Craig Mills for providing spatially resolved standardized fisheries vessel sightings data. This research was partially funded by Defra contracts MF0731/0729 and MF0322.
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
-
Attrill M.J. and Power M. (2002) Climatic influence on a marine fish assemblage. Nature 417:275–278.[CrossRef][Medline]
Beaugrand G., Ibanez F., Reid P.C. (2000) Spatial, seasonal and long-term fluctuations of plankton in relation to hydroclimatic features in the English Channel, Celtic Sea and Bay of Biscay. Marine Ecology Progress Series 200:93–102.[ISI]
Beaugrand G., Reid P.C., Ibanez F., Lindley J.A., Edwards M. (2002) Reorganization of North Atlantic marine copepod biodiversity and climate. Science 296:1692–1694.
Bedford B. C., Woolner L. E., Jones B. W. (1986) Length–weight relationships for commercial fish species and conversion factors for various presentations. MAFF Directorate of Fisheries Research. Fisheries Research Data Report, 10. 41 pp.
Bianchi G., Gislason H., Graham K., Hill L., Jin X., Koranteng K., Manickchand-Heileman S., Payá I., Sainsbury K., Sanchez F., Zwanenburg K. (2000) Impact of fishing on size composition and diversity of demersal fish communities. ICES Journal of Marine Science 57:558–571.
Boudreau P.R. and Dickie L.M. (1992) Biomass spectra of aquatic ecosystems in relation to fisheries yield. Canadian Journal of Fisheries and Aquatic Sciences 49:1528–1538.
Brown J.H., Gillooly J.F., Allen A.P., Savage V.M., West G.B. (2004) Towards a metabolic theory of ecology. Ecology 85:1771–1789.[ISI]
Coull K. A., Jermyn A. S., Newton A. W., Henderson G. I., Hall W. B. (1989) Length/weight relationships for 88 species of fish encountered in the North East Atlantic. Scottish Fisheries Research Report 43: 81 pp.
Daan N., Gislason H., Pope J.G., Rice J.C. (2005) Changes in the North Sea fish community: evidence of indirect effects of fishing? ICES Journal of Marine Science 62:177–188.
Dippner J.W. (1997) Recruitment success of different fish stocks in the North Sea in relation to climate variability. German Journal of Hydrography 49:277–293.
Dorel D. (1986) Poissons de l'Atlantique Nord-Est, Relations taille-poids. IFREMER Report DRV.86.001/RH/Nantes. 165 pp.
Dulvy N.K., Polunin N.V.C., Mill A.C., Graham N.A.J. (2004) Size structural change in lightly exploited coral reef fish communities: evidence for weak indirect effects. Canadian Journal of Fisheries and Aquatic Sciences 61:466–475.
Duplisea D.E., Kerr S.R., Dickie L.M. (1997) Demersal fish biomass size spectra on the Scotian Shelf, Canada: species replacement at the shelfwide scale. Canadian Journal of Fisheries and Aquatic Sciences 54:1725–1735.
Fox J. (1997) Applied Regression Analysis, Linear Models, and Related Methods(Sage Publications, Thousand Oaks, CA) xxi + 597 pp.
Garcia S.M. and Cochrane K.L. (2005) Ecosystem approach to fisheries: a review of implementation guidelines. ICES Journal of Marine Science 62:311–318.
ICES. (2003a) Report of the Working Group on Fish Ecology. ICES Document, CM 2003/G: 04. 110 pp.
ICES. (2003b) Report of the Working Group on the Assessment of Mackerel, Horse Mackerel, Sardine and Anchovy. ICES Document, CM 2003/ACFM: 07. 616 pp.
ICES. (2003c) Report of the Working Group on the Assessment of Southern Shelf Demersal Stocks. ICES Document, CM 2003/ACFM: 03. 478 pp.
ICES. (2003d) Report of the Working Group on the Assessment of Southern Shelf Stocks of Hake, Monk and Megrim. ICES Document, CM 2003/ACFM: 01. 499 pp.
Jennings S., Greenstreet S.P.R., Reynolds J.D. (1999) Structural change in an exploited fish community: a consequence of differential fishing effects on species with contrasting life histories. Journal of Animal Ecology 68:617–627.[CrossRef][ISI]
Legendre P. (2000) Program Mod_t_testDepartement de sciences biologiques, Universite de Montreal Available at http://www.fas.umontreal.ca/BIOL/legendre/>.
Murawski S.A. and Idoine J.S. (1992) Multispecies size composition: a conservative property of exploited fishery systems? Journal of Northwest Atlantic Fisheries Science 14:79–85.
O'Brien C.M., Fox C.J., Planque B., Casey J. (2000) Climate variability and North Sea cod. Nature 404:142.[Medline]
Piet G.J. and Jennings S. (2005) Response of potential fish community indicators to fishing. ICES Journal of Marine Science 62:214–225.
Pinnegar J.K., Jennings S., O'Brien C.M., Polunin N.V.C. (2002) Long-term changes in trophic level of the Celtic Sea fish community and fish market price distribution. Journal of Applied Ecology 39:377–390.[CrossRef][ISI]
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., Fuentes-Yaco C., Frank K.T. (2003) Spring algal bloom and larval fish survival. Nature 423:398–399.[Medline]
Pope J.G., Stokes T.K., Murawski S.A., Idoine S.I. (1988) A comparison of fish size composition in the North Sea and on Grand Banks. In Wolff W., Soeder C.J., Drepper F.R. (Eds.). Ecodynamics: Contributions to Theoretical Ecology(Springer, Berlin) pp. 146–152.
Rice J. (2003) Environmental health indicators. Ocean and Coastal Management 46:235–259.[CrossRef]
Rice J. and Gislason H. (1996) Patterns of change in the size spectra of numbers and diversity of the North Sea fish assemblage, as reflected in surveys and models. ICES Journal of Marine Science 53:1214–1225.
Rice J.C. (2000) Evaluating fishery impacts using metrics of community structure. ICES Journal of Marine Science 57:682–688.
Rochet M-J. and Trenkel V.M. (2003) Which community indicators can measure the impact of fishing? A review and proposals. Canadian Journal of Fisheries and Aquatic Sciences 60:86–99.
Shin Y-J., Rochet M-J., Jennings S., Field J.G., Gislason H. (2005) Using size-based indicators to evaluate the ecosystem effects of fishing. ICES Journal of Marine Science 62:384–396.
Southward A.J., Boalch G.T., Maddock L. (1988) Fluctuations in the herring and pilchard fisheries of Devon and Cornwall linked to change in climate since the 16th century. Journal of the Marine Biological Association of the United Kingdom 68:423–445.[ISI]
Trenkel V.M., Pinnegar J.K., Rochet M-J., Rackham B.D. (2004) Different surveys provide similar pictures of trends in a marine fish community but not of individual fish populations. ICES Journal of Marine Science 61:351–362.
Trenkel V.M. and Rochet M-J. (2003) Performance indicators derived from abundance estimates for detecting the impact of fishing on a fish community. Canadian Journal of Fisheries and Aquatic Sciences 60:67–85.
Venables W.N. and Ripley B.D. (2002) Modern Applied Statistics with S 4th edn (Springer, New York) 512 pp.
Warnes S. and Jones B.W. (1995) Species distributions from English Celtic Sea groundfish surveys, 1984 to 1991(MAFF Directorate of Fisheries Research, Lowestoft) Report 98.
Zwanenburg K.T.C. (2000) The effects of fishing on demersal fish communities of the Scotian Shelf. ICES Journal of Marine Science 57:503–509.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. L. Blanchard, D. L. Maxwell, and S. Jennings Power of monitoring surveys to detect abundance trends in depleted populations: the effects of density-dependent habitat use, patchiness, and climate change ICES J. Mar. Sci., January 1, 2008; 65(1): 111 - 120. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. P.R. Greenstreet and S. I. Rogers Indicators of the health of the North Sea fish community: identifying reference levels for an ecosystem approach to management ICES J. Mar. Sci., January 1, 2006; 63(4): 573 - 593. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||