© 2005 International Council for the Exploration of the Sea
Variability in retention of Calanus finmarchicus in the Nordic Seas
a University of Bergen, Department of Biology PO Box 7800, N-5020 Bergen, Norway
b Institute of Marine Research Box 1870 Nordnes, N-5817 Bergen, Norway
*Correspondence to T. Torgersen: tel: +47 55584439; fax: +47 55584450. e-mail: thomas.torgersen{at}bio.uib.no.
Using a regional ocean circulation model and particle tracking, we have studied the probability of the copepod Calanus finmarchicus being retained within the Nordic Seas' population as a function of its initial location, its vertical migration pattern, and the interannual variability in physical forcing. Defining a retention index in terms of the number of particles remaining within the Nordic Seas divided by the initial number of particles released, we found that spatial location had the greatest effect on the retention index during the study period, 1988–1991. Variability as a result of differences in physical forcing among years and among different seasonal vertical migration patterns had smaller but similar effects. The seasonal vertical migration behaviours with the highest advective loss rates and the most sensitive to interannual physical forcing were those that ascended early and descended late from a shallow summer depth. Average retention within the Nordic Seas was 0.40 after one year in simulations with diffusion and advection, and 0.42 in simulations with advection only. The average retention at the end of the four-year sequence was 0.10 and 0.12 with and without diffusion, respectively. Particles located in the western areas of the Nordic Seas had the highest retention, while those along the Norwegian coast showed little or no retention after four years. Initial location has a larger influence on retention than interannual variability in advective fields. C. finmarchicus offspring tend to reside in areas different from their parents, with different probabilities of retention. This spatial variability in retention rate is also experienced as inter-generational variability by members of the population. Model results suggest that almost all of the C. finmarchicus that are advected into the Barents Sea originate from off the Norwegian coast. Thus, predicting C. finmarchicus inflow into the Barents Sea requires knowledge of their abundance on the Norwegian Shelf.
Keywords: Barents Sea, Calanus finmarchicus, models, Norwegian Sea, ocean circulation, particle tracking, retention
Received 2 July 2004; accepted 30 May 2005.
 |
Introduction |
|---|
 Top Introduction Model description and...ResultsDiscussionReferences |
|---|
Life history strategies and behaviours of organisms evolve in response to the environment. The selected strategies will be those with the highest integrated fitness over the range of variability of the environment, and therefore, populations in fluctuating and static environments will evolve differently, even if the average environmental states are the same (Murphy, 1968; Schaffer, 1974). A sudden change in the environment will often be harmful to residing organisms, and this is the background for a major concern about climate changes and ecology, i.e. populations may be unable to survive, maintain productivity, or sustain their functional roles in ecosystems if they are exposed to changes in the environmental forcing.
The copepod Calanus finmarchicus (Gunnerus) feeds and spawns in shallow water during spring to early summer. In the Nordic Seas (Figure 1), most of the population undergoes only one generation per year (Aksnes and Blindheim, 1996). The new generation descends to greater depths in summer to overwinter (most during the fifth copepodite stage, C5). The timing of seasonal ascent is more synchronous than the summer descent, and is initiated during late winter, one to two months before the spring bloom (Østvedt, 1955; Hirche, 1996; Niehoff et al., 1999). Reproduction takes place after ascent. Average realized fecundity is probably between 100 and 1000 eggs female–1, corresponding to a maximum growth rate, rmax, of 
4–6 y–1 (Aksnes and Blindheim, 1996). Some individuals initiate summer descent as early as May, while others are active in the upper layers throughout summer (Østvedt, 1955; Dale et al., 2001). During the active period, C. finmarchicus are confined to the upper 100 m, and mainly above 60 m, with depth differences both between ontogenetic stages and between day and night (Tande, 1988; Dale and Kaartvedt, 2000). Over-wintering C. finmarchicus are typically associated with the cold, and virtually predator free, Arctic Intermediate Water (AIW). Therefore, the over-wintering depth varies strongly within the Nordic Seas (Kaartvedt, 1996, and references therein). In the Atlantic domain, east of the Polar Front, where warm Atlantic Water (AW) overlays the AIW, C. finmarchicus are mainly found below 500 m, while they concentrate in the upper 300 m in the Greenland Sea Gyre, where the AIW extends to the surface (Østvedt, 1955; Hirche, 1991; Richter, 1994; Dale et al., 1999).
|
C. finmarchicus is by far the most abundant mesozooplankton species in the Nordic Seas. Early developmental stages are prey for the larvae of many commercially important fish populations in the North, Norwegian, and Barents Seas. Late copepodites and adults are the main prey of adults of several dominant fish populations, including the Norwegian spring-spawning herring (Clupea harengus L.) and mackerel (Scomber scombrus L.).
The mean near-surface circulation in the Nordic Seas is shown in Figure 1. In the eastern areas, the upper layers are dominated by AW, entering the Nordic Seas from the south, through the Faroe–Shetland Channel and across the Iceland–Faroe Ridge. In the western area, in the Iceland and Greenland Seas, the entire water column is dominated by water of Arctic and polar origin, entering by the East Greenland Current. The upper layer flow into the Nordic Seas is around 10–15 Sv (1 Sv = 106 m3 s–1), the Atlantic influx being about an order of magnitude higher than that of Arctic and Polar waters (Blindheim, 2004). One branch of the AW flows into the Barents Sea. Most of the AW branching to the west of Spitsbergen cools and sinks, and forms part of the deep water of the Nordic Sea Basins. The circulation in the Nordic Seas is largely cyclonic. Water circulates in connected gyres in the Norwegian, Lofoten, and Greenland Basins, and the main transport out of the Nordic Seas occurs close to the bottom, across the sill in the Denmark Strait, and across the Iceland–Scotland Ridge. There is also some upper layer transport out of the Nordic Seas through the East Greenland Current. Assuming that the water exchange with adjacent seas is restricted to the upper 500 m, a 12 Sv exchange as suggested by Aksnes and Blindheim (1996) corresponds to an annual water renewal rate of 0.29. For particles that are not homogenously distributed, like zooplankton, advective loss rates may deviate strongly from this value.
C. finmarchicus has two main populations, residing in the sub-polar gyre and in the Nordic Seas (Conover, 1988; Aksnes and Blindheim, 1996; Melle et al., 2004). These populations are not isolated but rather there is active exchange of individuals between them. The reproductive rates of the C. finmarchicus are, however, much higher than the exchange rates. Therefore, even if this flux of individuals was of importance on evolutionary time scales, influencing population genetics, its direct ecological significance is modest (Aksnes and Blindheim, 1996; Bucklin et al., 1996).
For C. finmarchicus that are advected out of the Nordic Seas, the probability of having descendants being repatriated is low. Those that are transported to the north into the Arctic are lost without any plausible return route. Those that enter the North Sea are unable to overwinter there due to the shallow bottom depth, which results in a potentially high predation risk from visual predators as well as high flushing rates. Consequently, the North Sea population is replenished each spring by advection from the over-wintering stocks off the shelf edge. Those that are transported south through the East Greenland Current or the deep overflow have a very small probability of being transported back to the Nordic Seas. Integrated water renewal rate of the sub-polar gyre has been estimated at 0.12 y–1 (Aksnes and Blindheim, 1996), so the repatriation rate of C. finmarchicus lost to this area should be very low. Those entering the Northwest Atlantic through the East Greenland Current would have to enter the sub-polar gyre first in order to re-enter the Nordic Seas. Hence, being advected out of the Nordic Seas can be viewed as mortality relative to the local population. Irrespective of whether the lost individuals thrive and reproduce in other areas, their contribution to the Nordic Seas' population will be zero (Iles and Sinclair, 1982; Eiane et al., 1998).
The realized fitness of organisms depends partly on strategy (genes) and partly on chance (stochasticity). C. finmarchicus' probability of being lost from the population by being advected out of the Nordic Seas domain will vary between years and within the spatial domain. Advection varies with depth, and Bryant et al. (1998) found that seasonal vertical migration was instrumental in maintaining self-sustaining, basin-scale populations of C. finmarchicus. Timing of the seasonal vertical migration and summer and winter depths may have significant effects on the risk of disappearing from the population.
Owing to the currents and gyres in the Nordic Seas, a C. finmarchicus individual will tend to reside and develop in a place other than its parents did'. Therefore, both the spatial and temporal variabilities of the advection probability may act as inter-generational variability.
The effects of spatial and temporal variabilities on zooplankton retention can only be studied effectively through ocean model simulations. We use a particle-tracking model to quantify and compare the spatial and interannual variabilities in retention of C. finmarchicus in the Nordic Seas, and to evaluate to what extent retention is determined by the seasonal vertical migration pattern. Furthermore, we monitor the variability in particle inflow to the Barents Sea, which is one of the main sinks for C. finmarchicus in the Nordic Seas.
|
Model description and simulations |
|---|
|
TopIntroductionModel description and...ResultsDiscussionReferences |
|---|
Current vectors from a physical ocean model were used to move particles, representing C. finmarchicus, around the Nordic Seas.
The physical model
Particles were advected using currents from a Princeton Ocean Model (POM) simulation (Blumberg and Mellor, 1987; Ådlandsvik and Sundby, 1994; Asplin et al., 1998). This model solves the primitive equations numerically by the finite difference method. In the vertical, terrain-following sigma-coordinates are used. For vertical mixing, a Mellor–Yamada level 2.5 turbulence closure scheme is used (Mellor and Yamada, 1982). The most important modification from the standard POM version is the use of the flow relaxation scheme as an open boundary condition. The implementation in POM is documented in Engedahl (1995).
The model domain covers the Nordic, Barents, and North Seas. The horizontal resolution is 20 km and the number of grid cells is 208 x 120. In the vertical, 14 sigma-levels are used. For the present simulations, we used a 148 x 105 grid subset (Figure 1) of the larger model. The initial description of sea surface elevation, currents, salinity, and temperature was taken from a diagnostic climatology produced by the Norwegian Meteorological Institute (DNMI) and Institute of Marine Research (Engedahl et al., 1998). At the open boundaries, this was complemented by four tidal constituents. The meteorological forcing was taken from the hindcast archive of "www.met.no" (Eide et al., 1985). Freshwater run-off from 47 rivers was included. More details on the model set-up and validation are provided by Asplin et al. (1998) and Ådlandsvik et al. (2004).
Particle tracking
C. finmarchicus were represented as particles with seasonal patterns in vertical distribution, which varied among simulations (Figure 2, Table 1). C. finmarchicus overwinter at depth, ascend in early spring for reproduction, and die after an intense reproductive period. Subsequently, their offspring grow until the C5 stage, and then descend to overwinter. Alternatively, the offspring mature and commence reproduction in the same year in which they were born, and then it is the second generation that descends for over-wintering. Since we simulated a vertical distribution trajectory over a year starting 1 January and ending 31 December, we actually simulated at least two half-lives. Thus, the descending individual in autumn could be regarded as the offspring of the same individual ascending in spring. Simulations were run over four years, and at the beginning of the year particles were initiated in the final position of the previous year. C. finmarchicus has ontogenetic changes in vertical distribution from nauplia to C5 stages. However, these changes are small in magnitude relative to the migrations related to over-wintering. They are likely to have a minor effect on transport, and were therefore disregarded.
|
|
Horizontal movement of C. finmarchicus along the x-axis can be described by:
|
| (1) |
In our simulations, 2000 particles were released on 1 January in randomly chosen cells within each of nine subareas in the Nordic Seas (Figure 1). This spatial resolution allowed us to identify large-scale differences in retention, as well as to address specific spatial patterns. The ocean circulation model was run with forcing for 1988–1991, a period of fairly high values of the North Atlantic Oscillation (NAO) Index (Hurell, 1995; Jones et al., 2001). A high NAO index corresponds to increased westerly winds and generally warmer conditions in the Nordic Seas. There was also some variation in the NAO Index, with 1988 and 1991 having lower values (1), and 1989 and 1990 the highest values at 5 and 4, respectively. Ideally, years with negative NAO values should have been included to contrast the high NAO years, but these were not available to us at the time of this work. Since C. finmarchicus has a one-year life cycle in the Nordic Seas, retention rates and variability after one year are of obvious interest. However, simulations with a one-year horizon might induce a strong bias against particles initiated in the periphery of the study area, since these need a shorter time to be transported out of the region. Therefore, all simulations were run for four years.
We ran the model both repeating the same physical forcing in each of the four years and with the forcing fields for each of the four different years in sequence in order to separate the effects of the interannual variability in physical forcing from the influence of spatial variability on the retention rate. The model was run with 24 = 16 different seasonal vertical migration patterns, varying in winter depth, summer depth, day of ascent, and day of descent (Table 1, Figure 2). The two values for each variable were chosen both to be representative and to cover much of the variability, as reviewed in the introduction. However, the observed variability in the behaviour of C. finmarchicus in the Nordic Seas is slightly greater than that covered in the simulations. During winter, particles remained at their winter depth. When bottom depths over continental shelves and slopes were shallower than the winter depth, particles were placed 1 m above the bottom. On the ascent day, they moved upwards with a vertical velocity of 1 m h–1 (Carlotti and Wolf, 1998). Their descent towards the winter depth was with the same vertical speed.
Retention was defined as the fraction of the initial particles remaining in the entire study area. Below, we present the variability in retention in terms of the coefficient of variance, CV, which is the standard deviation (s.d.) divided by the average. We investigated how variability in retention depended upon where the particles were released, the physical forcing, and their vertical migration behaviour. Since retention varied by orders of magnitude between subareas, and also between years and behaviours, the arithmetic mean of CV can be misleading (e.g. between behaviour CV may be very high in subareas with very low retention). Therefore, averages of CVs were weighted by the retention index.
|
Results |
|---|
|
TopIntroductionModel description and...ResultsDiscussionReferences |
|---|
Averaged over all behaviours, subareas, and forcing years, the retention index was 0.40 after one year in the simulations with diffusion and 0.42 in the ones without diffusion. The average retention at the end of the four-year sequence (1988–1991) was 0.10 and 0.12 with and without diffusion, respectively. Total retention through the four-year sequence with diffusion is shown in Figure 3. Average retention after four years with the same environmental forcing each year was 0.09 and 0.11 with and without diffusion, respectively, almost identical to that after the four-year sequence using the observed forcing.
|
Retention varied the greatest with initial subarea and less owing to interannual climate variability (forcing year) and behaviour. Individuals initiated in the western subareas had the highest retention index (Figure 1). Simulations with diffusion yielded zero retention of particles released in subarea 3 off northern Norway, and close to zero in subareas 1 and 2 off southern and mid-Norway after four years, respectively. The simulations without diffusion yielded some, but still very little, retention of particles released in these three subareas. The following results are from simulations with diffusion.
Average retention from simulations with spatial and temporal variabilities for each of the 16 behaviours is shown in Figure 4. Retention varied strongly among different subareas (Figures 1 and 4), and, integrated over all 16 behaviours, CV was 1.15 after the four-year sequence. The difference in retention between forcing years was about half of that among subareas (Figures 1 and 4). After four years (with the same forcing year within each simulation) and integrating over all 16 behaviours, between forcing year CV for the entire study area was 0.50. The year 1988 had the highest retention, 3.4 times higher (after four years) than 1991, the year with the lowest retention.
|
|
Total retention within the study area varied significantly with behaviour (Figures 1 and 4). After the four-year sequence, the behaviours with the highest (15 and 16) and lowest (2) retention indices differed by a factor of 8.2 (Table 1), and the inter-behaviour CV was 0.48. Behaviour 2 had a shallow over-wintering depth, early ascent to a shallow summer depth, and a late descent (Table 1). In general, behaviours with long periods at their summer depth were also more sensitive to interannual climate variability than other behaviours (behaviours 2, 6, 9, 10, and 14; Table 1, Figure 4). Variability between behaviours was less sensitive to initial subarea. However, retention of particles that descended late was slightly more sensitive to initial subarea than that of other behaviours. Generally, the subareas with the smallest retention had the highest variability, but subarea 8 was rather sensitive to behaviour in spite of its generally high average retention. After four consecutive years with the same forcing year, inter-behaviour variability was considerably lower when forced with flow fields from 1988 (Figure 5A) compared with any of the other years.
Particles retained within the Nordic Seas drifted between subareas and after four years, particles from the same initial subarea were spread over different subareas (Table 2). The final distribution of retained particles between subareas differed partly from the retention indices based on initial subareas (Table 2). The clearest differences were that subareas 5 and 9 had high final densities of particles compared with the retention indices of these subareas, while subareas 4 and 7 had low final densities compared to the retention indices.
|
A large proportion of the lost particles were transported into the Barents Sea. Variability in this transport largely follows that of loss rates from the Nordic Seas. During the four-year sequence, 22% of the initial particles were advected into the Barents Sea. Averaging over years with different physical forcing, 23% of the initial particles were lost to the Barents Sea within the first year. After four years with the same forcing-year within each simulation, the average fraction lost was only marginally higher than after one year, at 0.24. Thus, most of the particles transported into the Barents Sea entered during the first year and came from subareas 1–3 (Figure 6A). Advection into the Barents Sea was highest in 1990 and lowest in 1988 (Figure 6B). The same behaviours that had a high probability of being lost from the Nordic Seas had a high probability of ending up in the Barents Sea (Figure 6C). The transport of particles into the Barents Sea is more sensitive to their distribution within the Nordic Seas than to behaviour or interannual climate fluctuations (Figures 6 and 7).
|
|
|
Discussion |
|---|
|
TopIntroductionModel description and...ResultsDiscussionReferences |
|---|
The model transport of particles between subareas illustrates how C. finmarchicus is likely to have little population structure within the Nordic Seas. Thus, the Nordic Seas is considered the appropriate scale to study their retention and population closure. Our results suggest that C. finmarchicus experience greater inter-generational variability in risk of being lost from the Nordic Seas population owing to their variation in spatial location through transport by the mean currents than from the interannual variability in advective fields. The retention indices of the subareas are averages over large areas, and retention rates within each of these subareas are certainly not homogenous. Thus, the actual geographical variability in advective loss risk within the Nordic Sea could be even higher than outlined here. To the extent that behaviour has influence on advective loss risk, C. finmarchicus might have evolved behaviours that are robust against strong temporal variability in advection.
Contributions from processes on sub-grid scales will be lost in the output from the pure advection model. The diffusion term increases the loss from the study area compared with the pure advection model. Diffusion tends to move individuals that were initially entrapped in cyclonic gyres into export paths, such as that along the Norwegian coast. Hence, the slightly higher retention observed in the simulations without diffusion is to be expected.
The estimated overall retention estimates in this study are similar to those of Bryant et al. (1998), who, with a non-diffusive advection model, estimated retention indices of 0.52 and 0.11 after one and five years, respectively. Aksnes and Blindheim (1996) estimated the renewal rate of the water in the upper 500 m of the Nordic Seas to be 0.29 y–1, corresponding to a retention index of 0.75 y–1. Our estimated annual retention of 0.40 y–1 corresponds to a loss rate of 0.92 y–1, significantly higher than the estimated water renewal rate. However, advection is generally stronger in shallow water, and since plankton spend much of their time there, they are expected to experience stronger advection out of the area than those based on the depth-averaged estimates. Hence, differences between average water renewal rates and advective loss rates of organisms with vertical migration are to be expected. It should be noted that the subareas with the lowest retention also tended to have higher initial densities of particles. The 2000 particles were distributed in the part of the subareas that were not covered by land, resulting in higher initial particle density in the subareas in which much of the area was covered with land, which is the case for subareas 1–3 and 7. Since initial particle density varies, a direct comparison between particle loss rates and water renewal rates might be somewhat biased.
There was pronounced variability in retention between different behaviours. However, the birth and death rates of C. finmarchicus in the Nordic Seas are higher than advective loss rates, and probably more important in shaping their behaviour (Aksnes and Blindheim, 1996). It is worth noting that C. finmarchicus, remaining in very shallow waters throughout summer, which correspond to areas of high growth and high predation, also have high advective losses from the Nordic Seas (Eiane et al., 1998). Since the generally higher rates of mortality, growth, and reproduction will vary spatially within the Nordic Seas, predicting the large-scale horizontal distribution of C. finmarchicus from our particle tracking results alone is not possible. We note, however, that the areas predicted to retain particles in our study correspond well with those from the simulations of Bryant et al. (1998). The greatest biomasses of C. finmarchicus are also typically found in these areas (Melle et al., 2004). This suggests that the spatial pattern of our estimated retention reflects the actual pattern of C. finmarchicus.
Our results suggest that while transport of C. finmarchicus into the Barents Sea depends partly upon its behaviour and climate variability, almost all particles transported into the Barents Sea were from the Norwegian Shelf within the first year. Therefore, to predict C. finmarchicus inflow into the Barents Sea, one needs to determine the C. finmarchicus abundance on the Shelf. The rapid loss (within the first year) of particles initially placed off the Norwegian coast corresponds well with the simulated C. finmarchicus trajectories previously described by Pedersen et al. (2001). They found that particles initially off southern Norway in an area corresponding to our subarea 1, generally were advected northwards, and travelled distances sometimes exceeding 1600 km in 180 days.
Other aspects of the environment, besides advection that affect C. finmarchicus, include non-uniform predator and prey distributions and temperature fields. C. finmarchicus are therefore advected between areas with different growth and mortality profiles. For C. finmarchicus, the spatial variability in these environmental parameters will act as temporal variability in a way similar to that of the risk of advective loss. Therefore, we suggest that plankton living in advective, spatially variable systems are most likely to be pre-adapted to temporal fluctuations, and consequently might be robust with regard to climatic changes. Also, in order to better predict distributions of C. finmarchicus, spatially resolved growth and mortality rates must also be applied. These issues will be addressed in future studies.
|
Acknowledgements |
|---|
This work was supported by the Research Council of Norway. Comments from Bjørn Ådlandsvik, Ken Drinkwater and an anonymous referee significantly improved the manuscript.
|
References |
|---|
|
TopIntroductionModel description and...ResultsDiscussionReferences |
|---|
-
Ådlandsvik B., Gundersen A.C., Nedreaas K.H., Stene A., Albert O.T. (2004) Modelling the advection and diffusion of eggs and larvae of Greenland halibut (Reinhardtius hippoglossoides) in the north-east Arctic. Fisheries Oceanography 13:403–415.[CrossRef][Web of Science]
Ådlandsvik B. and Sundby S. (1994) Modelling the transport of cod larvae from the Lofoten area. ICES Marine Science Symposia 198:379–392.
Aksnes D.L. and Blindheim J. (1996) Circulation patterns in the North Atlantic and possible impact on population dynamics of Calanus finmarchicus. Ophelia 44:7–28.[Web of Science]
Asplin L., Ingvaldsen R., Loeng H., Ådlandsvik B. (1998) Description and validation of a three-dimensional numerical model of the Nordic and the Barents Sea. Fisken og Havet 10: Institute of Marine Research, Bergen. 35 pp.
Blindheim J. (2004) Oceanography and climate. In Skjoldal H.R. (Ed.). The Norwegian Sea Ecosystem(Tapir Academic Press, Trondheim) pp. 65–96 559 pp.
Blumberg A.F. and Mellor G.L. (1987) A description of a three-dimensional coastal ocean circulation model. In Heaps N. (Ed.). Three Dimensional Coastal Ocean Models(AGU, Washington, DC) pp. 1–16 Coastal and Estuarine Series, 4 208 pp.
Bryant A.D., Hainbucher D., Heath M. (1998) Basin-scale advection and population persistence of Calanus finmarchicus. Fisheries Oceanography 7:235–244.[CrossRef][Web of Science]
Bucklin A., Sundt R., Dale G. (1996) The population genetics of Calanus finmarchicus in the North Atlantic. Ophelia 44:29–45.[Web of Science]
Carlotti F. and Wolf K.U. (1998) A Lagrangian ensemble model of Calanus finmarchicus coupled with a 1-D ecosystem model. Fisheries Oceanography 7:191–204.[CrossRef][Web of Science]
Conover R.J. (1988) Comparative life histories of the genera Calanus and Neocalanus in high latitudes of the northern hemisphere. Hydrobiologia 167/168:127–142.[CrossRef]
Dale T., Bagøien E., Melle W., Kaartvedt S. (1999) Can predator avoidance explain varying overwintering depth of Calanus in different oceanic water masses? Marine Ecology Progress Series 179:113–121.[Web of Science]
Dale T. and Kaartvedt S. (2000) Diel patterns in stage-specific vertical migration of Calanus finmarchicus in habitats with midnight sun. ICES Journal of Marine Science 57:1800–1818.
Dale T., Kaartvedt S., Ellertsen B., Amundsen R. (2001) Large-scale oceanic distribution and population structure of Calanus finmarchicus, in relation to physical environment, food and predators. Marine Biology 139:561–574.[CrossRef]
Eiane K., Aksnes D.L., Ohman M.D. (1998) Advection and zooplankton fitness. Sarsia 83:87–93.
Eide L. I., Reistad M., Guddal J. (1985) Database av beregnede vind og bølgeparametre for Nordsjøen, Norskehavet og Barentshavet, hver 6. time for årene 1955–1981. Technical report. The Norwegian Meteorological Institute, Oslo, 38 pp. (in Norwegian).
Engedahl H. (1995) Use of the flow relaxation scheme in a three-dimensional baroclinic ocean model with realistic topography. Tellus 47A:365–382.
Engedahl H., Ådlandsvik B., Martinsen E.A. (1998) Production of monthly mean climatological archives for the Nordic Seas. Journal of Marine Systems 14:1–26.[Medline]
Hirche H-J. (1991) Distribution of dominant calanoid copepod species in the Greenland Sea during late fall. Polar Biology 11:351–362.[Web of Science]
Hirche H-J. (1996) Diapause in the marine copepod, Calanus finmarchicus – a review. Ophelia 44:129–143.[Web of Science]
Hurell J.W. (1995) Decadal trends in the North Atlantic oscillation: regional temperatures and precipitation. Science 269:676–679.
Iles T.D. and Sinclair M. (1982) Atlantic herring: stock discreteness and abundance. Science 215:627–633.
Jones P.D., Osborn T.J., Briffa K.R. (2001) The evolution of climate over the last millennium. Science 292:662–667.[CrossRef][Web of Science][Medline]
Kaartvedt S. (1996) Habitat preference during overwintering and timing of seasonal vertical migration of Calanus finmarchicus. Ophelia 44:145–156.[Web of Science]
Melle V., Ellertsen B., Skjoldal H.R. (2004) Zooplankton: the link to higher trophic levels. In Skjoldal H.R. (Ed.). The Norwegian Sea Ecosystem(Tapir Academic Press, Trondheim) pp. 137–202.
Mellor G.L. and Yamada T. (1982) Development of a turbulence closure model for geophysical fluid problems. Reviews of Geophysics and Space Physics 20:851–875.[Web of Science]
Murphy G.I. (1968) Patterns in life history and the environment. American Naturalist 102:391–403.[CrossRef][Web of Science]
Niehoff B., Klenke U., Irigoien X., Hirche H-J., Head R., Harris R. (1999) A high frequency time series at Weathership M, Norwegian Sea, during the 1997 spring bloom: the reproductive biology of Calanus finmarchicus. Marine Ecology Progress Series 176:81–92.[Web of Science]
Østvedt O.J. (1955) Zooplankton investigations from Weathership M in the Norwegian Sea 1948–1949. Hvalrådets Skrifter 40:1–93.
Pedersen O.P., Tande K.S., Slagstad D. (2001) A model study of demography and spatial distribution of Calanus finmarchicus at the Norwegian coast. Deep-Sea Research II 48:567–587.[CrossRef]
Richter C. (1994) Seasonal changes in the vertical distribution of mesozooplankton in the Greenland Sea Gyre (75°N): distribution strategies of calanoid copepods. ICES Journal of Marine Science 52:533–539.[CrossRef][Web of Science]
Schaffer W.M. (1974) Optimal reproductive effort in fluctuating environments. American Naturalist 108:783–790.[CrossRef][Web of Science]
Tande K. (1988) An evaluation of factors affecting vertical distribution among recruits of Calanus finmarchicus in three adjacent high-latitude localities. Hydrobiologia 167/168:115–126.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Daase, J. O. Vik, E. Bagoien, N. C. Stenseth, and K. Eiane The inf luence of advection on Calanus near Svalbard: statistical relations between salinity, temperature and copepod abundance J. Plankton Res., October 1, 2007; 29(10): 903 - 911. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||