You are here: Home / Case study areas / North Sea 

CS 2: North Sea

Background

Fisheries management and scientific advice

In 1999 the EU and Norway agreed to implement a long-term management plan for the North Sea cod stock. This was intended to constrain harvesting within ‘safe biological limits’ and to provide for sustainable fisheries and greater potential yield in the future (a ‘cod-recovery plan’). The latest proposal from the European Commission (Reg 2003/0090 (SNS)) includes both effort reduction/control and Harvest Control Rules (HCR) for setting TACs. In 2003 scientists recommended complete closure of fisheries catching cod in the North Sea, including severe restrictions on vessels targeting other species (e.g. haddock), but which catch cod as accidental by catch (ICES 2004).
In 1999, the EU and Norway agreed to implement a long-term management plan for the North Sea plaice stock (ICES 2004), and a plaice (sub-area IV) ‘recovery plan’ is now under consideration. Increased mortality as a result discarding practices appears to be a particular threat to plaice recovery.
Herring spawning stock biomass (SSB) in 2004 was estimated at 1.89 million t and this is expected to still be at 1.88 million t in 2005. SSB has increased gradually since low stock sizes in the mid-1990s. This was in response to reduced catches, strong recruitment, and strong management measures that reduced exploitation both on juveniles and adults (ICES 2005b). Poor herring year classes always occur when the stock is low, but can also occur when the stock is high. Both the 1998 year class and the 2000 year class appear to be very strong in all surveys, but the 2002, 2003 and the incoming 2004 year classes are estimated to be among the weakest on record (ICES 2005b).

Environmental impact on stock dynamics and recovery potential

In spite of the many factors that might impact stock recovery, prior experience would seem to suggest that North Sea demersal stocks can recover, even at a relatively low level of recruitment, as long as fishing mortality is greatly curtailed. Pope and Macer (1996) demonstrated that cod recruitment before and during the second-world war (WW II) was broadly comparable to the situation today (250 million in 1939 compared to 157 million in 2004). However, following an almost complete fishery closure for 5 years during WW II, SSB increased from ~80,000 t (in 1939) to ~180,000 t (in 1946). Similarly, Rijnsdorp and Millner (1996) demonstrated that plaice recruitment before WW-II was lower than the situation today (~350 million in 1933, compared to 682 million in 2004) and yet plaice SSB increased from ~200,000 t (in 1939) to ~280,000 t (in 1946) during the fishery closure.

Current recovery plans generally assume that there has been no underlying change in environmental conditions, and hence that the ‘carrying-capacity’ and the structure of the food web of the North Sea ecosystem has not changed (Myers et al. 2001). It is now widely appreciated that this might not be the case as the North Sea ecosystem has undergone a regime shift in the 1980s, centred in two periods of rapid changes (1982-1985 and 1987-1988). The changes in large-scale hydro-meteorological forcing, affecting also local hydrographic variability, have caused drastic changes in dominant zooplankton species (e.g., decrease in Calanus finmarchicus and its replacement by the closely related but smaller species Calanus helgolandicus) to zooplankton community structure (e.g., higher calanoid diversity) to fish recruitment success (e.g., decrease in gadoids and initial increase in flatfish recruitment followed by a more variable phase after the second centre period) (e.g., Beaugrand 2004; Reid et al. 2001, 2003). Generally, the period after the regime established the new state in 1988 is characterized by warmer temperature, low abundance of northern fish and zooplankton species (Beaugrand et al. 2002), and increasing abundance and diversity of southern plankton (Reid et al. 2003) and fish (Beare et al. 2004) species. These changes have been purported to have had a negative impact on North Sea cod recruitment as C. finmarchicus is a major prey for cod larvae (it is the right size and occurs at the right time of year). Consequently it has been suggested that the loss of this vital prey species could impact the ability of cod to recover because of anticipated failures in future cod recruitment (Beaugrand et al. 2003, Kell et al. (in press)). Whether a comparable but reverse regime shift in the 1960s and 1970s lead to the so-called ‘gadoid outburst’ is a possibility which could not be mechanistically explained due to lack of sufficient data (ICES 1999). It has been pointed out recently that regime shifts have profound implications and should be incorporated into management strategies - an idea consistent with viewing management in the context of ecosystems (Rothschild & Shannon 2004).

Influenced by the increasing evidence that environmental changes impact the recruitment success of key commercial North Sea stocks, a number of ICES groups have started to routinely provide process oriented expertise necessary to develop stock recruitment models that take the interactions between the species’ biology, their biotic and physical environment into account:
In recent years several ICES groups have focused their efforts specifically on North Sea cod recruitment and estimating ‘reproductive potential’. This has included two ICES Study Groups, the Study Group on Incorporation of Process Information into Stock Recruitment Models [SGPRISM] and the Study Group on Growth, Maturity and Condition Indices in Stock Projections [SGGROMAT]. The output and findings of these groups will provide the basis for the science conducted in WPs 1&2 in UNCOVER.
Reproductive potential
In current ICES assessment methods cod maturity ogives are treated as constants and were estimated using the International Bottom trawl Survey series 1981-1985. A fixed maturity ogive is also used for the estimation of SSB in North Sea plaice, but maturity at age and sex ratios are not likely to be constant over time. Grift et al. (2003) showed that the age and length at maturation have decreased over the past half century. Preliminary analyses revealed that the proportion mature of age 2 plaice varied by almost a factor of 2 over the period 1957-2003, an especially high inter-annual variability was estimated for the most recent years (ICES 2005a). In the case of herring, maturity ogives are calculated annually but again these are not sex specific. This means that in the current standard assessment, natural variation in reproductive potential of cod, herring and plaice is not considered, despite opposing evidence and existing data.
The assumption implicit in the S/R model is that female-only SSB (FSB) is equal to half of the SSB. For species that exhibit strongly dimorphic growth, maturation and mortality this is very dubious. A second dubious assumption of the current S/R models is that SSB is proportional to viable egg production (TEP) from the stock, i.e., TEP/SSB is constant.
The importance of including the dynamics of maturity into stock assessment might prove to be most valuable in terms of redefining reference points. This has been demonstrated for Baltic cod, where a previously unconvincing stock and recruitment relationship was turned into a significant relationship by substituting constant maturity with a dynamic maturity model (Köster et al. 2003). There is a high chance for similar success in North Sea cod and plaice.
Given the intrinsic importance of reproductive potential to stock/recruit (S/R) relationships, the setting of biological reference points (BRPs) and stock projections, alternative measures of reproductive potential that are currently being developed will certainly enhance the predictive capabilities of current assessment models (ICES 2004a).

Trophic interactions and spatial scales

Slow and sudden changes in the entire food web have been shown to alter the magnitude and direction of trophic control, interacting with fishery pressure:
In 2003 the ICES Study Group on Multispecies Assessments in the North Sea [SGMSNS] was tasked with evaluating the single-species recovery plan for North Sea cod by taking into account biological interactions. The 4M package was used to run MSVPA and the Commission’s HCR (harvest control rules) were applied from 2003 onwards. Seventeen different HCR scenarios were tested and cod recovery examined using both single and multispecies formulations (ICES 2003b). When the proposed HCR for cod was applied, both single and multi-species models predicted cod SSB recovery. The predicted recovery of cod SSB was slower when taking multi-species interactions into account, and Bpa was reached approximately one year later.
In terms of the impact of a cod recovery plan on other species in the North Sea, haddock SSB was predicted to decline to beyond the established Blim in all multi-species scenario simulations, but in the single-species simulations, it was predicted that haddock SSB and yield would increase. Similar reversed trends in single and multi-species predictions resulted for whiting, Norway pout, and sandeel.

The divergence in the predicted outcomes between the single-species and multi-species models indicates that much greater consideration needs to be given to multi-species interactions when considering stock recovery programmes (ICES 2003b). This is further strengthened by the results of 4M model runs simulating HCR scenarios where grey gurnard was included as a predator (as it was in the 2002 keyrun (ICES 2002)). Without (ICES 2002, 2003b), and with HCR rules applied (Kempf et. al., unpubl. data) the cod stock is predicted to go extinct due to predation of 0-group cod by grey gurnard, as the grey gurnard stock showed a fast increase after the 1988 regime shift (ICES 2002, 2003b). In the current single species cod assessment, natural mortalities are constant over time and originate from a MSVPA model run conducted in 1986 (ICES 2005a), for plaice a constant M = 0.1 is applied.

Direct predatory effects

Grey gurnard preys heavily on pelagic 0-groups cod and whiting as all three species occur in high densities in North Sea frontal areas (EU LIFECO; ICES 2002, 2003), supporting the notion that it is the pelagic-, and settlement phase that is critical to recruitment success (Heath et al., 1999). A study inside SGGROMAT revealed that the size of the realized habitat of age 1 cod was a function of stock size and influenced by their temperature preferences (Blanchard et al., in press). However, stratification maps have shown that the areas within the preferred temperature range co-occur with frontal areas in the North Sea. This highlights the importance to taking results from process models of spawning locations and larval drift routes into account when implementing spatial-temporal predator-prey overlap functions in multi-species models.

As the stock numbers and size structures of potential predators of grey gurnard (cod, benthic sharks & rays) are low and truncated due to fishing, grey gurnard may emerge as a new key predator for North Sea cod and whiting and the present regime status may therefore be stabilized (EU LIFECO; Floeter et al. 2005). This stabilizing process may be further enhanced as grey gurnard is not targeted by commercial fisheries and is a robust species that most likely has high survival rates after being discarded, suggesting that similar mechanisms could be at work as when spiny dogfish emerged as a key fish predator in the 1980s off the coast of north-eastern Canada/USA, and only declined after being heavily fished (Link et al. 2002).

There is evidence from meso-scale surveys, that predators aggregate on high density patches of fish prey (e.g., whiting and haddock on sandeel prey (Temming et al., 2004) and whiting on 0-group cod (Temming et. al. 2005)). Analyses revealed that such local “predation hot spot??? events can have a tremendous impact on the overall survival of a year class: only ten “predation hot spots??? of the observed magnitude would cover the entire number of 0-group cod consumed during the 3rd quarter of 1996 (the last strong year class) by all MSVPA model predators. As such small scale high-intensity events have a very low detection probability in large scale surveys, only simulation modeling will help to develop tools for a realistic parameterization of “predation hot spots??? in large scale models.
There are further strong field-based indications that pelagic species (herring, sandeel, mackerel, horse mackerel) predate with high intensity on cod eggs and larvae (Ellis and Nash, 1997, Seegers et al. 2005). This predation on eggs also affects another case-study species, including North Sea plaice. Studies from the 1930s onwards have shown that herring are strong predators on sandeel and are themselves heavily impacted by horse mackerel (Dahl and Kirkegaard 1987). In years when post-larval and juvenile sandeels are present in the herring stomachs (about 50% of years in the southern and western North Sea), they form an important prey item from February through to April (Hardy 1924; Savage 1937; Last,1989, Daan unpublished data). The mechanism for the interannual variability in feeding on sandeel has not as yet been determined.

Indirect predatory effects

A simulation analysis revealed that the North Sea MSVPA model is sensitive to the diet composition of grey gurnard (EU LIFECO, Floeter et al., 2005). When changes in spatio-temporal overlap lead to a decrease in predation on cod and an increase in predation of whiting, the release of predation pressure exerted by whiting on cod predicts a strong increase in the recovery potential of the North Sea cod stock. This type of 2nd order effect applies as well for other species in the North Sea fish assemblage (Pope, 1991). Further, if important prey species (sandeel, Norway pout) are depleted predation pressure on recovery species (e.g., cod) potentially increases, as the pool of alternative prey is vanishing.
In general, this highlights that the current North Sea multi-species VPA model and hence scientific advice to managers is highly sensitive to the number of predators and preys explicitly included, as well as to their diet selection and stock-recruitment parameterisation. There is further clear evidence that these processes act at different scales, and that their magnitudes of impact often depend on the prevailing environmental conditions.

Conclusion

Summarizing, stock reproductive potential, genetics, distributions and drift patterns under varying stock sizes and environmental conditions, top down and bottom up processes affecting larvae of recovery species during the drift phase, and critical trophic interactions play a key role in determining the magnitude and survival rates of incoming year classes. So understanding how the strengths and directions of these processes vary with environmental and anthropogenic changes is key to assess future stock recovery potentials and develop sustainable management and recovery strategies.

List of references

To top