A review of the climate conditions off West Greenland during the past 50 years shows large variability in the atmospheric, oceanographic and sea-ice variables, as well as in fish stocks. A positive relationship is found between water temperature and the recruitment of cod and redfish, whereas the recruitment of shrimp and halibut is negatively related to temperature. Observed shifts in the hydrographic conditions during the second half of the 1990s indicate that some changes in the environment may be expected in the coming years. Relationships between the past variations in fish resources, hydrographic conditions, and the large-scale climatic conditions, as expressed by the North Atlantic Oscillation (NAO), strongly support the incorporation of environmental variability into prediction models for fish recruitment and therefore into the assessment of the fish resources.

Key words:  biomass, climate, cod, fishery, West Greenland, salinity, shrimp, temperature, variability, NAO

Introduction

In the 20th century Greenland experienced two great transitions, the first from seal hunting to a cod fishery, and then from cod to a shrimp fishery. Both affected the human population centres and the economy of West Greenland (Hamilton et al., 2000). The economic transitions reflected large-scale shifts in the underlying marine ecosystems, driven by interactions between climate and human resource use.

The marine shelf ecosystems off East and West Greenland are intermediate between the cold Polar water masses of the Arctic region and temperate water masses of the Atlantic Ocean. They are important fishing grounds and are characterized by relatively few dominant species, which interact strongly (Pedersen and Kanneworff, 1995; Rätz and Stransky, MS 1999; Pedersen and Zeller, 2001). Ocean currents, that transport water from the polar and temperate regions, affect the marine productivity on the Greenland shelves and changes in the North Atlantic circulation system therefore have major impacts on the distribution of fish species and fishery yields (Pedersen and Smidt, 2000; Pedersen and Rice, 2002).

Greenland climate has undergone some dramatic changes through the 20th century. The period 1920–70 generally experienced warm air temperatures while the subsequent 30 years have been dominated by three extreme cold periods around 1970, the early 1980s and the early 1990s. These atmospheric changes have also been reflected in the oceanographic conditions of the Greenland waters (Buch, 2000 a, b).

In the present paper we describe recent developments in the fishery, the atmospheric climate and ocean hydrography off the West Greenland and discuss possible relationships between climate variability and the fish stocks.

Data

The fisheries and fish stock data have been retrieved from a number of assessment documents available from the Northwest Atlantic Fisheries Organization (NAFO; www.nafo.int), the International Council for the Exploration of the Seas (ICES; MS 2000), and the Greenland Institute of Natural Resources.

Atmospheric sea level pressures from Ponta Delgada, in the Azores, and Reykjavik, Iceland, were obtained from the Climate Research Unit (www.cru.uea.ac.uk/cru/data/nao.htm) and updated to 2002. For a description of these data see Jones et al. (1997) and Jones (1984).

Biological Variability

A rich Atlantic cod (Gadus morhua) fishery started off West Greenland in the 1920s after a general warming of the Northern Hemisphere (Jensen, 1939; Dickson et al., 1994; Buch et al., 1994; Horsted, 2000). The West Greenland cod fishery peaked in the 1960s at annual catches between 400 000 and 500 000 tons. During the late 1960s, the cod catches declined drastically along with the catches of other commercially important fish species, including redfish (Sebastes marinus and S. mentella), Atlantic halibut (Hippoglossus hippoglossus) and wolfish (Atlantic wolfish, Anarhichas lupus, and spotted wolfish, Anarhichas minor), that were mainly taken as by-catch in the cod fishery. After 1969, catches of cod and redfish fluctuated around a much lower mean level (Fig. 2).

Except for a temporary improvement of cod during 1988–90 due to a strong 1984 year-class recruited from Iceland, annual groundfish surveys on the West Greenland Shelf (0–400 m depth), undertaken by Germany since 1982, show a dramatic decline in overall biomass and mean individual weight of the demersal fishes (Rätz and Stransky, MS 1999.

The decline in the cod catches is believed, in large part, to be a response to climate variability. Climate is thought to have led to both the center of the cod fishery moving south during the 1980s and the decrease in size-at-age (Hovgård and Buch, 1990; Riget and Engelstoft, 1998; Rätz et al., 1999; Horsted, 2000). At the same time catches of two other commercially important species, northern shrimp (Pandalus borealis) and Greenland halibut (Reinhardtius hippoglossoides) increased (Fig. 2). In recent years a new fishery for snow crab (Chionoectes opilio) shows a steep increasing trend from a few hundred tons in 1994 to 15 000 tons in 2001.

During the last two decades, northern shrimp has been the most important fishery resource in Greenland by far. Export of shrimp to Japan, has provided a high-value economic alternative to cod, and by 1995 comprised 73% of Greenland's total exports. The shrimp stock off West Greenland is distributed in NAFO Div. 0A and Subarea 1 (see Fig. 3). There is no evidence of distinct sub-populations and the entire shrimp stock is assessed as a single population. The Greenland fishery exploits the stock in Subarea 1 (Div. 1A to 1F) in offshore and inshore areas (primarily Disko Bay) (Fig. 3). Overall shrimp catches increased until 1992, varied at lower levels from 1993 to 1997 and increased thereafter (see Fig. 2).

From 1975 to 1984 the annual effort in the shrimp fishery showed a slightly increasing trend from about 75000 hr to about 93 000 hr of fishing. In subsequent years the offshore fleet increased and effort rose to 250000 hr in 1991–92 (Fig. 4). Thereafter effort decreased as a result of management measures, reduced activity in Div. 0A and a higher fishing efficiency of the participating vessels.

The catch-per-unit-effort (CPUE) for the West Greenland shrimp fishery can be used as a stock biomass index (Fig. 4). The spike in 1987 is likely the result of some very strong year-classes produced in the early 1980s. From 1990 to 2000 the CPUE index has shown an increasing trend suggesting an increasing shrimp stock biomass (Fig. 4).

The GFRI has conducted annual stratified-random shrimp trawl surveys since 1988 in the main West Greenland shrimp distribution area (Carlsson and Kanneworff, MS 2000). For the period 1988–97, biomass indices of the fishable shrimp stock in the offshore areas were stable at a level of about 250 000 tons (Fig. 5a). From 1998 the biomass indices show a significant increase to a record high biomass estimate in 2000 of 350 000 tons.

A change in geographical distribution of the commercial fishing effort for shrimp has been observed since the late 1980s (Hvingel, MS 2000). Initially, Div. 1A and 1B contained the most important fishing grounds, but from the late 1980s the fishery has gradually moved southward as indicated by the mean latitude of effort allocation (Fig. 5b). At the same time, the highest shrimp catches in the annual shrimp survey showed a movement towards shallower depths (Fig. 5c). The change in shrimp catch distributions both geographically and in depth observed during the survey period may indicate stock migration towards preferable temperature habitat due to the lower temperatures in the late-1980s and early-1990s (described in a later section). However, from 1995 to 1999, the average bottom temperature during the shrimp survey (July–August) showed a clear increasing trend (Fig. 5d) and this has not yet resulted in a movement of the center of the main fishing effort northward again (Fig. 5b).

During the 1990s there was a slight increase in catches of striped pink shrimp (Pandalus montagui) in local commercial fishing areas and during the annual shrimp survey (Fig. 5e). This shrimp species is well adapted to cold conditions and the increased catches may indicate a positive biological effect on this species of the cold sea climate from the late 1980s to the mid-1990s. The peaks in the abundance indices from the shrimp survey in 1995 and 1998 are unexplained (Kanneworff, MS 2000), but a lag between increased larval production and recruitment to the catchable stock should be expected.

From 1950 to 1984, GFRI annually collected zooplankton samples from West Greenland waters. The zooplankton displacement volume and most of the zooplankton taxa showed higher abundance indices in the generally warmer period 1950–68 compared to the colder period 1969–84 (Pedersen and Smidt, 2000). However, abundance indices of sandeel larvae were negatively correlated with sea temperature. Historical sandeel and shrimp larvae abundance indices (1950–84, in Pedersen and Rice, 2002), updated with abundance indices from zooplankton samples collected in 1996, 1999 and 2000, showed similar trends (Fig. 6) and are positively correlated (r = 0.48, p<0.05, n = 23; Spearmann rank correlation).

Climate Variability

Over the last fifty years of the 20^th century, the West Greenland area has experienced some rather dramatic fluctuations in climate, which have influenced the living conditions for all species on land, as well as in the ocean. These fluctuations may therefore be regarded as one of the reasons for the observed variability in the various fish stocks described in the previous section.

Atmospheric conditions

Over the past decade, several papers have dealt with the importance of the North Atlantic Oscillation (NAO) Index on the climate in the North Atlantic region (Dickson et al., 2000; Blindheim et al., 2000; Dickson et al., 1996; Seager et al., 2002) and thereby also in the West Greenland area (Buch, 2000b). The NAO, which is associated with changes in the surface westerlies across the Atlantic onto Europe, refers to a meridional oscillation in the atmospheric mass with centres of action near the Iceland Low and the Azores High (van Loon and Rogers, 1978). Because the signature of the NAO is strongly regional, a simple index of NAO was defined as the difference between the normalised SLP anomalies at the Azores and Iceland (detailed discussion about the historical literature can be found in van Loon and Rogers, 1978). In order to extend the NAO index back in time, measurements from Lisbon or Gibraltar are commonly used instead of the Azores (e.g. Hurrell, 1995 and Jones et al., 1997). Although the NAO signal is evident throughout the year, it is most pronounced during winter and accounts for more than one-third of the total variance of the Sea Level Pressure (SLP) field over the North Atlantic (references in Hurrell, 1995). We use wintertime (December–March) SLP difference between Ponta Delgada, Azores, and Reykjavik, Iceland, and subtract the mean SLP difference for the period 1961–90 to construct the NAO anomaly.

The variability of the NAO index since 1865 is shown in Fig. 7, where the heavy solid line represents the low-pass filtered meridional pressure gradient. Positive values of the index indicate stronger than average westerlies over the mid-latitudes associated with low-pressure anomalies over the region of the Icelandic Low and anomalous high pressures across the subtropical Atlantic.

In addition to a large amount of interannual variability, there have been several periods when the NAO index persisted in one phase for many winters, (van Loon and Rogers, 1978, Barnett 1985, Hurrell and van Loon, 1997). Over the region of the Icelandic Low, the seasonal pressures were anomalously low during winter from the turn of the century until about 1930 (with exception of the 1916–19 winters), while pressures were higher than average at lower latitudes. Consequently, the wind over Europe had a strong westerly component and the moderating influence of the ocean contributed to higher than normal temperatures over much of Europe (Parker and Folland, 1988). From the early 1940s until the early 1970s, the NAO index exhibited a downward trend into the extreme low NAO of the sixties and this period was marked by European wintertime temperatures that were frequently lower than normal (van Loon and Williams, 1976, Moses et al., 1987). A sharp reversal occurred after the early-1970s and, since the 1980s, the NAO has remained in a highly positive phase with SLP anomalies of more than 10 hPa in magnitude over both the subpolar and the subtropical Atlantic. The 1989, 1990, 1994 and 1995 winters were marked by some of the highest positive values of the NAO index recorded since 1865 (Fig. 7).

A detailed analysis by Hurrell and van Loon (1997) suggests that the temperature anomalies for the North Atlantic and surrounding land masses for the 1980–94 period are strongly related to the persistent and exceptionally strong positive phase of the NAO index (their Fig. 1a is reproduced in Fig. 8 using NCEP/NCAR reanalysis data). This clearly illustrates a strong correlation between the strength of the westerlies across the North Atlantic – (the NAO index) and the air temperatures in Greenland and Europe. It also shows that the air temperatures in Greenland and Europe are negatively correlated, a phenomenon referred to as a seesaw effect (e.g. van Loon and Rogers, 1978).

This can be illustrated further by comparing the temperatures over Greenland from a low NAO period (1960–69) to a high NAO period (1990–99) (Fig. 9). Note that especially offshore of West Greenland, it was significantly warmer in the 1960s than in the 1990s.

A time series of annual mean air temperatures from Nuuk, West Greenland for the period 1873–2002 is shown in Fig. 10. In addition to the interannual variability, it reflects the general picture of variability outlined above in the description of the NAO index (Fig. 7), i.e. high NAO conditions are normally reflected in cold conditions in Greenland (Fig. 9). The late-1990s are, however, an exception from this pattern, since both NAO and Nuuk air temperatures show relatively high values. This was due to a slight displacement of the NAO pattern towards the east or northeast (ICES, 2000, 2001, 2002).

A detailed analysis using 6 hr winds from a number of observation sites in Greenland show that the changes in the wind pattern in the Greenland area are generally minor, in large part because of the dominating influence by the local orography.

Hydrographic conditions

The waters off West Greenland are dominated by water masses advected into the region from outside (Buch, 2000a,b):

• In the surface layer close to the coast, cold and low saline Polar Water originates from the East Greenland Current
• Water below and to the west of the Polar Water derives from the North Atlantic Current

The changes in the atmospheric conditions caused by the shift from low NAO to high NAO conditions affect the ocean circulation and ocean conditions in the North Atlantic (Dickson et al, 1996, Dickson et al., 2000). The oceanographic conditions off West Greenland are also directly and indirectly affected by climate variability related to the changes in NAO.

The most well-known oceanographic time series from West Greenland is the mid-June mean temperature on top of Fylla Bank (Station 2, 0–40 m; Fig. 11), which the GFRI has maintained due to its importance to the cod stock assessment. These temperatures can vary quite drastically from one year to the next, often more than 1°C, reflecting the variability of both the atmospheric influence and the inflow of Polar Water. The 3-year running means naturally smooth out the high-frequency variations and better reflect the large-scale climatic variability.

The 50-year temperature time series reveal some very distinct climatic events:

• The 1950-68 period generally showed high temperatures around 2°C.
• Around 1970 the coldest period was experienced. The cold temperatures of this period were due to an anomalous high inflow of Polar Water, which was closely linked to the "Great Salinity Anomaly" (Dickson et al. 10988, Belkin et al., 1998). At the same time, the NAO index changed from negative to positive causing a shift from warm to cold atmospheric conditions off West Greenland.
• In the early 1980s and early 1990s two extremely cold periods were observed reflecting the cold atmospheric conditions associated with the high NAO indices during these years.
• Low ocean temperatures were observed in 1997 although the atmospheric conditions were quite warm. Together with low salinity measurements (Fig. 10) they suggest a high inflow of Polar Water.
• During the late-1990s, temperatures have been rather high despite high NAO values. This was due to a displacement of the NAO pattern towards the east or northeast (ICES, 2000, 2001, 2002).

At greater depth, there are three water masses of Atlantic origin (Buch, 2000b):

• Irminger Water – temperature around 4.5°C and salinity above 34.95.
• Irminger Mode Water – Irminger Water mixes with surrounding water masses on its way towards southwest Greenland – temperature around 4°C and salinities between 34.85 and 34.95.
• Northwest Atlantic Mode Water – temperature above 2°C and salinities between 34.5 and 34.85. In late autumn, the temperature rises to above 5°C.

At Fylla Bank Station 4, there is a similar apparent reduction in the inflow of salt and heat. The Irminger Mode Water was present in much higher quantities before the mid-1970s than after, and the three cold periods are clearly reflected in the temperatures of the upper 200 m. A weak freshening in the upper 150–200 m is observed since 1965, resulting in a less dense water mass within this layer. This freshening, however, is most dominant in the upper 50–100 m. A similar freshening during the same period was also observed in the Irminger Water component north of Iceland (Malmberg, 1985), indicating a reduction of the strength of the Irminger Current after 1965 and/or a more dominant influence of Polar Water. From the mid-1960s to the early-1970s, the freshening was caused by an anomalous high inflow of Polar Water closely linked to the "Great Salinity Anomaly", whereas afterwards it is believed to be caused by a high NAO anomaly reducing the strength of the Irminger Current by the increased windstress forcing of the North Atlantic Current eastward.

Ice conditions

Two main types of sea ice dominate in West Greenland waters:

1. "Storis", or multiyear ice of polar origin carried to southwest Greenland by the East Greenland Current, and
2. "Westice", or first-year ice formed in Baffin Bay and Davis Strait.

The shift in community structure and landing composition of fish in Greenland during the second half of the 20th century coincides in time with the large climatic changes observed in the Greenland area. It is therefore believed that the observed changes in recruitment patterns are largely driven by changes in ocean climate. In terms of mechanisms linking oceanographic factors to recruitment of fish and shellfish in West Greenland, sea temperature, drift of larvae by surface currents, and stability of the water masses (oceanographic fronts) have been proposed (Pedersen and Rice, 2002). Variability in these factors is related in turn to the inflow of water from other parts of the North Atlantic, which is strongly related to NAO variations. The individual strengths of the East Greenland Current and the Irminger Current have a dominating effect on the physical environment of the shelf areas around southern Greenland.

Based on changes in hydrographic observations, the ocean transport of salt and heat towards West Greenland appears to have decreased drastically after 1970. This occurred at the same time as the reduced heat flux into the ocean from the atmosphere. These climate changes seem to have had a negative effect on the recruitment success of the West Greenland cod stock, and a number of other boreal fish stocks, and a positive effect on the production of northern shrimp and Greenland halibut.

The drastic reduction (almost disappearance) in the West Greenland cod fishery is believed to have two causes:

• Reduction in the West Greenland spawning stock. The number of cod recruited at age 3 years have been documented to be significantly correlated with the spawning stock biomass and June water temperature on top of Fylla Bank (Hansen and Buch, 1986; Hovgård and Buch, 1990). Both factors positively affected the number of offspring and explained 51% of the observed variation in recruitment (Rätz et al., 1999).
• Reduced inflow of cod larvae from Icelandic spawning grounds. The inflow of cod larvae occurred almost every year in the 1950s and early-1960s, (Fig. 2 in Hansen and Buch, 1986) but since then has been negligible except for the 1973 and 1984 year-classes.

Northern shrimp prefer relatively cold temperatures in the range of 1–6°C and their larvae are less vulnerable to low temperatures compared to cod (Shumway et al., 1985). This may partly explain the positive reaction of the West Greenland shrimp stock to the changed climatic conditions. However, the shift in the underlying marine ecosystem at West Greenland may have been amplified by the declining cod stock due to a release in predation pressure on e.g. sandeel and northern shrimp (Koeller, 2000; Lilly et al, 2000). Additionally, by-catches in the steady growing fishery for northern shrimp during the last part of the 20th century may have played a role in reducing and keeping the mean trophic level low (Kingsley et al. MS 1999; Pauly et al., 2001).

The observed increase in the shrimp biomass during recent years is related to both increasing growth (decrease in mean length at sex change) and recruitment (Carlsson and Kanneworff, MS 1999; Siegstad, 2000). The shrimp recruitment indices (number of juvenile shrimp) show a steep increasing trend from 1997 to 2000, which is a good prospect for the shrimp fishery (Fig. 5a). This positive development is believed to be related to the favourable temperature conditions observed off West Greenland during this period, especially the increased inflow of warm Irminger Water (Fig. 14) to the area.

The relative cold period from the late-1980s to 1995, when temperatures decreased below the temperature preference of shrimp (3–4°C), seemed to have resulted in a southern migration of the shrimp stock and the fishery. The warming trend from 1995 to 2000 towards preferred temperatures seems to have favoured growth and recruitment of northern shrimp, thereby leading to an extraordinary increase in the shrimp biomass since the late-1990s.

Extensive ice cover in cold years contributes positively to the survival of larvae and juveniles in the same year and the effect can be detected in the CPUE several years later (the mean age of shrimp in the catch is about 6 years) (Parsons and Colbourne, 2000). A recent study by Ramseier et al. (2000) showed that the extent of localized sedimentation of particulate organic carbon (POC) can be derived from information about ice cover. POC likely plays an important role as food supply for shrimp, and it is possible that the explanation of the functional relationship is related more to nutrient supply than temperature-related phenomena. According to Parsons and Colbourne (2000) this would help explain the apparent inconsistencies between in situ observations, which suggest "cold conditions" are favourable for shrimp, and laboratory studies, which indicate that larval growth and survival are enhanced at higher temperatures.

• The Greenland climate since 1970 has been considerably colder than during the 1920–70 period, which can be related to a shift in the NAO index from negative to positive values.
• The redistribution of the atmospheric pressure fields has altered the surface currents of the North Atlantic, resulting in the reduction of the inflow of heat, salt and cod larvae to the West Greenland area.
• While there are good correlations between the climate changes and the observed shift in the marine ecosystem off West Greenland, scientific investigations to understand the ecological, chemical and physical processes behind the changes in the marine ecosystem are required.
• The increase in the West Greenland shrimp stock biomass is most likely not solely attributable to the changes in climate. The near disappearance of the cod stock will have reduced shrimp predation by this species. Also by-catches in the shrimp fishery would contribute to keeping the predation pressure low.
• The close relationship between climate variability and the marine ecosystem off West Greenland strongly support the incorporation of effects of environmental fluctuations in prediction models for fish stock recruitment and the assessment of the fisheries resources. This will, however, require increased research in process studies that seek to understand the processes linking fisheries recruitment to environmental factors. Such efforts must include the development of coupled ocean and ecological models, not only to increase our understanding but also to make realistic forecasts of the ecosystem changes for use in fisheries assessment work. Models of ecosystem developments under changing climatic conditions should be considered in fishery assessments in the future and they should lead to better planning for the Greenland society.

We thank Phil Jones for providing old pressure data and updating the time-series for the Azores. We also thank two anonymous referees for corrections and suggestions to improve this paper. During this work S.A.P. was funded by the Danish National Research Council project no. 9803018, and M.H.R. was funded by the Nordic Council of Ministers as part of "West-Nordic Ocean Climate Research Programme".