ECOBE
Effects of North Atlantic Climate Variability On the Barents Sea Ecosystem
Project coordinator: Svein Sundby, Institute of Marine Research

This is a multidisciplinary proposal that addresses all three foci of NFRs Polar Climate Research Programme. The main emphasis is on the focus “Ecological effects of climate change”. In support of this main focus, parts of the proposal also address the two other foci, “Marine climate” and “Technology and methods for earth observations and oceanographic measurements”.

Background
In arcto-boreal marine ecosystems the reproduction, recruitment and growth in fish stocks show large interannual variations. These variations are basically linked to ocean climate fluctuations that influence the fish stocks partly directly and partly indirectly through the lower trophic levels of the food web. The Arcto-Norwegian cod in the Barents Sea is a typical example of a fish stock in such an ecosystem, and historically Norwegian and Russian fisheries biologists and oceanographers have contributed considerably to enlighten the relations between ocean climate and fish production (e.g. Sars 1879; Helland-Hansen and Nansen 1909; Hjort 1914; Ishevskii 1961). Over the recent 20 years, and particularly during the 1990s, increased focus has been put on the study of the functional mechanisms behind the relations between ocean climate and physical and bio-physical processes causing growth in fish stocks (see in particular the proceedings of the Symposium on Cod and Climate Change (eds. Jakobson et al. 1994); Symposium on Climate Change and Northern Fish Populations (ed. Beamish 1995); GLOBEC First Open Science Meeting (eds. Coombs et al. 1998); Investigation of Calanus finmarchicus migrations between oceanic and shelf seas off north-west Europe ICOS (eds. Heath and Coombs 1999), SCOR 1997;1999). This implies that the “ocean climate” concept, which traditionally has been mainly considered as sea temperature, has been extended to include wind mixing, turbulence, vertical stability, light conditions and advection of water masses.
Examples of direct climate effects on fish are temperature which influences digestion, metabolic rates, migration and distribution, light which influences fish behaviour, and turbulence which influences encounter rates between early stages of fish and their prey. Examples of indirect effects through lower trophic levels are temperature which influence metabolic rates in phytoplankton and zooplankton, light which influence photosynthesis in phytoplankton and vertical behaviour in zooplankton, turbulence which influence nutrient uptake and plankton contact rates, vertical stability which influence nutrient supply and plankton vertical distribution, and advection of water masses which influence plankton concentrations. Hence, it becomes evident that understanding the population dynamics of fish stocks is a multidimensional problem that demand for interdisciplinary approaches accounting for both biological and physical processes through the food web.
In the Barents Sea ecosystem there are three key fish species in strong interaction: cod, capelin and herring. In addition the production of copepods, particularly the Calanus finmarchicus, is a major component in fuelling the growth of the three fish stocks. Also, the population dynamics of the Northeast-Arctic haddock is an interesting fish species to consider, since it has many common features with cod recruitment in relation to ocean climate variability, but with a considerably larger range of recruitment variability than the cod. The ocean climate parameters defined above strongly influence the production of all these organisms, but there are also internal interactions between the species where the climate effects are subordinated: The growth of the adult cod is strongly dependent of the abundance of capelin. In years of high abundance of young herring in the Barents Sea the predation pressure on larval and juvenile capelin from the young herring is strong resulting in poor recruitment to the capelin stock and bad feeding conditions for cod (Hamre and Hatlebakk 1998). Also, there are periods of strong interaction between young cod and older cod when cannibalism is important (Bogstad and Mehl 1997). The food consumption of sea birds indicate that early juveniles of the key fish species might be an additional contribution to reduce the year-class strengths of the key fish species (Sakshaug et al. 1994). Consequently, a change in a climate signal might result in a change in the recruitment in one of the stocks which might have cascading effects by internal interaction to other stocks long after a climate event occurred.

Scientific rationale
The effects of climate fluctuations on the interannual variability in reproduction, recruitment and growth of the Arcto-Norwegian cod stock are strong, but the functional mechanisms behind these variations are far from fully understood. By long time series of sea temperature, particularly the time series from the Kola section, it has been documented that good recruitment generally occurs in warm years and that cold years always result in poor recruitment (Sætersdal and Loeng 1987; Ellertsen et al. 1989; Ottersen and Sundby 1995). However, as pointed out in the above paragraph the functional relationship to temperature might as well be indirectly through the lower trophic levels. Skjoldal and Rey (1989) and Helle and Pennington (1999) showed that the role of temperature for cod recruitment in the Barents Sea is a proxy for the advection of zooplankton-rich Atlantic water from the Norwegian Sea into the Barents Sea, as the temperature and the influx of Atlantic water to the Barents Sea is strongly correlated (Ådlandsvik 1989; Ådlandsvik and Loeng 1991). Recently, it has been suggested a generic mechanisms on a pan-Atlantic scale for the relationship between Atlantic cod recruitment, temperature and advection of water masses (Sundby 2000). Recruitment of C. finmarchicus to the Northern Norwegian shelf and to the Barents Sea is probably governed by the same physical mechanisms. In early spring the Norwegian shelf waters are almost devoid of C. finmarchicus (Melle et al. 1993). Later, during spring and summer, however, an abundant population of C. finmarchicus inhabits the shelf waters, an indispensable prey for larval cod (Ellertsen et al. 1995) and herring (Fossum 1996). The most probable source for replenishment of the shelf stock of C. finmarchicus is the abundant stock residing in the depths of the Norwegian Sea during winter (Slagstad and Tande 1996, Harms et al. 2000).
Wind-induced turbulence and light conditions are other climate parameters that might influence cod recruitment (Sundby et al. 1994; Fiksen et al. 1998). These investigations indicate that in the cod recruitment-temperature linkage sea temperature might be a proxy for a number of other climate parameters. Therefore the mechanisms need to be explored by first-principles models. Much effort has recently been allocated to investigation of feeding and growth processes in larval fish (herring and cod, in particular) and to quantitative formulations of these processes (e.g. Fiksen et al. 1998, in press; Fiksen and Folkvord 1999, Otterlei et al. 1999, Werner et al. 2001, Fiksen and MacKenzie in press). These studies have provided the tools necessary to realistically model the impact on growth (and hence on recruitment success) of larval fish from large-scale climatic changes. Here, we will apply these models to predict the consequences imposed by changes in sea temperature, irradiance, turbidity, turbulence and currents.
The initial egg production is a necessary boundary condition for the first-principles models described above. Both recruitment research and fisheries management have traditionally assumed that spawner biomass is directly proportional to total egg production. Recent studies have shown that the individual fecundity (egg number) and egg quality are highly variable in long-lived species such as Atlantic cod (Kjesbu et al. 1998; Lambert and Dutil 2000). Scaling these individual-based data up to the stock level using the demographic information available from assessments and surveys has shown that spawner biomass is not proportional to total egg production (Marshall et al. 1998; Köster et al. 2001). There are also strong indications in the literature that age structure (Marteinsdottir and Thorarinsson, 1998) and proportion of repeat vs. recruit spawners (Kjesbu et al. 1996) in the spawning stock of Atlantic cod significantly affect subsequent recruitment success. Although these results are relevant to this project, few studies have actually focused per se on effects of variation in the physical environment on reproductive performance.

Over the recent two decades focus has been set on effects of large-scale weather pattern on marine and terrestrial ecosystems, particularly the effects of the El Niño-Southern Oscillation (ENSO). Similarly, there has been rapidly growing literature on the effects of the North Atlantic Oscillation Index (NAO) on the ocean climate and on marine ecosystems of the North Atlantic sector. The NAO, which essentially is an indication of the relative strengths of the Icelandic Low and the Azores High, has been shown to influence water mass characteristics, volume fluxes and heat exchange (e.g. Kwok and Rothrock 1999; Häkkinen 1999; Stein 1999). It influences marine ecosystems on the level of primary production (Belgrano et al. 1999), on zooplankton production (e.g. Fromentin and Planque 1996; Planque and Fromentin 1996; Fromentin et al. 1998; Reid et al. 1998; Piontkovskii et al. 2000) and on the level of fish production, particularly cod (Rodionov 1995; Stein et al. 1998). There are also indications of the influence of other large-scale climate indices, for example the Arctic Oscillation (AO) (Thompson et al. 1998), the Euro-Siberian Oscillation (Kelly et al. 1999), and Polar-Eurasian and Scandinavian Patterns. In the Arctic region there are also examples of inter-decadal variability that indicate linkage to the earth’s nutation (Yndestad 1999a, 1999b, 2002). Such studies have clearly demonstrated the profound effects on these highly aggregated climate events on marine ecosystems. However, they often leave us with confusion about the functional relationships to ecosystem change because of the wide range of effects of the various climate parameters. A change in NAO index might for example be associated with a change in sea temperature in northern Northeast Atlantic, a change in wind pattern including wind-induced mixing, a change in advection of Atlantic water, change in cloud cover and rainfall, change in light conditions, and change in mixed-layer depth and heat exchange across the sea surface. All these parameters will in various ways influence biomass production at each trophic level.
In order to understand the functional relationships between fish reproduction, recruitment and growth and the climate parameters we, therefore, need to investigate the linkages between the various climate parameters, how they are spatially and temporally interlinked, how they influence production at trophic levels below fish production, at trophic levels above fish, and how they influence trophic transfer. Secondly, we will improve the basis for medium-term to long-term ocean climate forecasts by improving coupled ocean-atmosphere numerical models (e.g. Drange 1999) and by improving statistical forecasting by utilising recent knowledge on persistence, lags and cycles in the ocean climate (e.g.Yndestad 1999, 2002, Furevik 2000a, Furevik 2000b). These investigations will form the scientific basis for improved assessment of the fish stocks in the Barents Sea, particularly of cod, haddock, herring and capelin, with emphasis on predicting the development in the stocks on a medium-term to long-term time scale. These challenges call for a joint effort by Norwegian marine institutions of the present project.

Goal
The over-all goal of this integrated project is to understand and quantify the impacts of Arctic climate variability on trophic transfer and ecosystem structure of the Barents Sea in order to improve the prediction of growth and recruitment on key fish species.

The over-all goal will be implemented by addressing objectives of five modules/work packages.

1. Explore the linkages between large-scale weather patterns, such as the NAO, and the regional and local climate, and investigate how such patterns cascades into spatio-temporal changes in the ocean climate parameters that are of importance for biomass production.

2. Explore the effects of ocean climate and circulation on the production and advection of Calanus finmarchicus onto the northern Norwegian Shelf and the Barents Sea.

3. Develop an integrated model system based on first-principles physics and biology to simulate distribution, transport, growth and survival of fish larvae from the spawning areas in spring to 0-group distribution in autumn when year-class strengths are largely determined.

4. Develop egg production models for the key fish species, with special focus on Arcto-Norwegian cod, Arcto-Norwegian haddock and Norwegian spring-spawning herring, based on the combined effects of food abundance and temperature on gonad production and maturation.

5. Develop a trophodynamic model system that integrates the models described above to simulate growth and recruitment of Barents Sea fish stocks. The trophodynamic model system will form the basis for sensitivity analysis to explore quantitatively the effects of the range of physical and biological parameters and processes of importance to the general problem of fish recruitment.

The figure illustrates the linkage between the objectives of the present proposal and specifies the key physical parameters and processes of the interaction in the trophic transfer. Module 1 comprises the climate analysis and prediction that feed into the physical part of the ecosystem. This influences each trophic level and the interactions between the levels. Module 2 addresses the copepods and trophic transfer to larval and juvenile fish. Module 3 addresses the transformation from eggs and larvae to the recruiting 0-group fish. Module 4 addresses the egg productions as input to the pelagic offspring in Module 3. Module 5 comprises the integration of the entire model system components to a trophodynamic model system. The final modelling output of 0-group fish will be evaluated by comparing it with the historical time series of 0-group fish abundance and distribution, and model sensitivity analyses will be used to explore the effects of the range of bio-physical parameters and processes.