ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on December 13, 2007
ICES Journal of Marine Science: Journal du Conseil 2008 65(3):339-350; doi:10.1093/icesjms/fsm171
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
This article appears in the following ICES Journal of Marine Science issue: 4th International Zooplankton Production Symposium: Human and Climate Forcing of Zooplankton Populations [View the issue table of contents]
Characteristics of Calanus finmarchicus dormancy patterns in the Northwest Atlantic
1 Department of Fisheries and Oceans, Bedford Institute of Oceanography, PO Box 1006, Dartmouth, NS, Canada B2Y 4A2
2 NOAA, Southwest Fisheries Science Center, 1352 Lighthouse Avenue, Pacific Grove, CA 93950-2097, USA
3 School of Marine Sciences, University of Maine, and Gulf of Maine Research Institute, 350 Commercial Street, Portland, ME 04101, USA
4 Department of Fisheries and Oceans, Northwest Atlantic Fisheries Centre, PO Box 5667, St John's, NF, Canada A1C 5X1
5 Department of Fisheries and Oceans, Institut Maurice Lamontagne, 850 Route de la Mer, CP 1000, Mont-Joli, QC, Canada G5H 3Z4
6 Graduate School of Oceanography, University of Rhode Island, Narragansett, RI 02882, USA
Correspondence to C. L. Johnson: tel: +1 902 4260753; fax: +1 902 4269388; e-mail: JohnsonC{at}mar.dfo-mpo.gc.ca
Johnson, C. L., Leising, A. W., Runge, J. A., Head, E. J. H., Pepin, P., Plourde, S., and Durbin, E. G. 2008. Characteristics of Calanus finmarchicus dormancy patterns in the Northwest Atlantic. – ICES Journal of Marine Science, 65: 339–350.Demographic time-series from four fixed stations in the Northwest Atlantic Ocean demonstrate variable timing of entry into and emergence from dormancy in subpopulations of the planktonic copepod Calanus finmarchicus. A proxy for timing of entry was established as the date each year when the proportion of the fifth copepodid stage (CV) in the subpopulation rose to half its overall climatological maximum CV proportion at that station. The proxy for timing of emergence at each station was set as the first date when adults were more than 10% of the total abundance of copepodid stages. An alternate emergence proxy date was determined by back-calculating the spawning dates of the first early copepodid stages appearing in spring, using a stage-structured, individual-based model. No single environmental cue (photoperiod, surface temperature, or average surface-layer chlorophyll a concentration) consistently explained entry or emergence dates across all stations. Among hypotheses put forward to explain dormancy in Calanus species, we cannot eliminate the lipid accumulation window hypothesis for onset of dormancy or a lipid-modulated endogenous timer controlling dormancy duration. The fundamental premise of these hypotheses is that individuals can only enter dormancy if their food and temperature history allows them to accumulate sufficient lipid to endure overwintering, moult, and undergo early stages of gonad maturation.
Keywords: Atlantic zonal monitoring programme, Calanus finmarchicus, copepod population dynamics, diapause, dormancy, individual-based model, lipids
Received 7 July 2007; accepted 19 September 2007; advance access publication 13 December 2007.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussionReferences |
|---|
Arthropods have evolved a variety of ontogenetic strategies to survive environmental adversity. In general, these responses involve a dormant phase, which may range from a functional slowdown of metabolism and growth (resting or quiescence) when conditions are unfavourable, to a genetically determined arrest of development at a fixed point in development (diapause) in response to one or more environmental cues. In the Copepoda (reviewed by Dahms, 1995), dormancy may occur in the egg, naupliar, or copepodid stages. Open-ocean calanoid copepods typically undergo dormancy as a late copepodid stage, in which the dormant phase is characterized by reduced metabolic and developmental rates, and residence for several months much deeper in the water column than the normal feeding zone (Conover, 1988; Dahms, 1995). The dormant stages re-emerge and migrate to the surface in winter or spring, when they take advantage of the spring bloom to support high reproductive rates (Conover, 1988; Dahms, 1995; Ohman et al., 1998).
Because of its predominance and role in pelagic foodwebs of the North Atlantic, Calanus finmarchicus has been the subject of substantial research on its population dynamics, including the nature of its dormancy. The dormancy exhibited by C. finmarchicus is seasonal (entry in summer/autumn, arousal in early to midwinter) and occurs primarily during a single developmental stage, the fifth copepodid stage (CV; reviewed by Hirche, 1996). Here, we refer to the dormant or resting phase in Calanus as "diapause", although it is not clear that all individuals or populations enter into the same level or intensity of diapause state. The CV diapause state is characterized by sluggish behaviour, seasonal vertical migration to deep water, arrested (or slowed) development rates, reduced metabolism (some 15–30% of active level), cessation of feeding, arrested jaw development, and reduced RNA:DNA ratio (Hirche, 1983, 1989; Miller et al., 1991; Wagner et al., 1998). There is also evidence of a reduction in levels of ecdysteroids, which are involved in the hormonal control of moulting (Johnson, 2004).
The timing of the active and overwintering periods varies among locations across the range of C. finmarchicus (e.g. Planque et al., 1997; Hind et al., 2000). We focus here on characterizing dormancy patterns of this species in the Northwest Atlantic. Seasonal patterns in "canonical" dormancy timing in C. finmarchicus may differ by several months in this region. It has been reported that C. finmarchicus CVs emerge from dormancy in late December in the Gulf of Maine (Durbin et al., 1997), in February on the Scotian Shelf (SS) (McLaren et al., 2001), and in March/April in the St Lawrence Estuary (Plourde et al., 2001). In each of these areas, many first generation (G1) copepods enter dormancy when they reach the CV stage, although the transition of a population to dormancy may be gradual as some individuals continue development and reproduction. Most of the population enters dormancy and is represented by CVs by late summer in the Gulf of Maine/Georges Bank region (Meise and O'Reilly, 1996) and on the SS (McLaren et al., 2001), and by early autumn in the St Lawrence Estuary (Plourde et al., 2001). Even locations that are spatially proximate can exhibit differences in dormancy timing on time-scales of months where hydrographic gradients are steep, e.g. C. finmarchicus in slope water south of Georges Bank begin emerging from dormancy 2–3 months later than copepods in the Gulf of Maine. They return to dormancy after a single generation (Miller et al., 1991), whereas the Gulf of Maine populations often produce second or more generations.
The condition, abundance, and seasonal timing of emergence from dormancy, relative to the timing of the spring transition–bloom period, are critical for the population dynamics of these species (Carlotti and Radach, 1996; Speirs et al., 2005, 2006). The timing and the magnitude of the spring bloom in the Northwest Atlantic can vary significantly on interannual and interdecadal time-scales (Sameoto, 2001; Durbin et al., 2003; Platt et al., 2003), so this interaction is likely to be an important factor driving interannual variability in copepod abundance and seasonal timing. The timing of migration to deep water in autumn and the emergence in spring are also critical for copepod predators, such as herring, and for the larvae of fish that feed on the naupliar stages produced following the emergence of dormant copepods in spring. However, the factors that trigger entry to dormancy by C. finmarchicus remain unknown, as do the factors controlling the duration of the dormant state (Dahms, 1995; Ohman et al., 1998; Hind et al., 2000). Understanding these controls is a major challenge for modelling copepod population dynamics (Head et al., 2001; Runge et al., 2005), particularly when evaluating the role of climate forcing in driving interannual and longer term population variability.
Seasonal cues including photoperiod and temperature trigger the onset of diapause in many insect species (Danks, 1987), as well as several copepod species that diapause as resting eggs (reviewed by Dahms, 1995). Although photoperiod has been used as a triggering signal in modelling studies of Calanus dormancy strategy (e.g. Fiksen, 2000), field observations and simulation experiments examining the onset of dormancy do not support the hypothesis that dormancy is triggered by photoperiod alone (Hind et al., 2000; Johnson, 2004). For populations with large spatial ranges, such as C. finmarchicus, photoperiod may not predict times when departure from the surface would benefit the population throughout its range. Induction by a combination of seasonal cues, as observed with photoperiod and temperature in copepodid-stage dormancy in a fresh-water copepod (Watson and Smallman, 1971) and in resting-egg dormancy in several fresh-water and marine copepod species (Johnson, 1979; Marcus, 1982; Uye, 1985; Hairston et al., 1990; Hairston and Kearns, 1995), is an alternative hypothesis. Watson (1986) developed an ecological classification of copepod diapause responses in which he categorized C. finmarchicus as Type IV, i.e. days shorter than a particular value allow CVs to continue moulting to adulthood, and low temperature enhances (increases the proportion in the population) continuous development. Several theoretical studies have indicated that the optimal timing for onset of diapause is typically at least one generation before the arrival of seasonally predictable unfavourable conditions (Hairston and Munns, 1984; Norrbin, 1996). In the Northwest Atlantic, properties of the water column that indicate increasing warming (i.e. temperature indices) or stratification may be more generally predictive of impending unfavourable conditions, attributable to decreasing surface primary production, across the range of the population.
Using a modelling approach to simulate contrasting seasonal population dynamics patterns in C. finmarchicus in the northeastern Atlantic, Hind et al. (2000) concluded that copepods were most likely not responding to photoperiod but to a decrease in food supply, i.e. a direct cue rather than a seasonally predictable one. In this analysis, the switch to enter dormancy is determined by the ambient food concentration when the copepodid stage CIV moults to CV; if at this moment the ambient food is less than a critical threshold level, the CV sinks and enters a dormant state. A similar study of populations of northwestern Atlantic C. finmarchicus, which are genetically distinct from eastern Atlantic populations (Bucklin et al., 1996), has not been carried out yet.
In their development of a three-dimensional, physical–biological model of the distribution and seasonal dynamics of C. finmarchicus across the North Atlantic, Speirs et al. (2006) ruled out the use of a fixed food-threshold concentration as a general cue for dormancy initiation. Instead, they adopted the rule that a fixed fraction (70%) of each generation enters dormancy, the rest continuing on to moult to the adult stage. This procedure follows a practice employed in several modelling studies of Calanus population dynamics (e.g. Carlotti and Wolf, 1998; Miller et al., 1998; see review in Speirs et al., 2006). The fixed-fraction concept yielded the best fit to ocean-scale data on timing and the distribution of CV and adult life stages.
The food-limitation hypothesis put forward by Hind et al. (2000) does not explicitly take into account the need for copepodids to accumulate a large quantity of wax esters, the lipids stored in the internal oil sac, in preparation for dormancy (Lee et al., 1970; Hirche, 1996; Miller et al., 1998). The potential role that accumulation of wax esters might play in controlling entry into dormancy has recently been considered explicitly (Rey-Rassat et al., 2002; Irigoien, 2004; Hassett, 2006). Rey-Rassat et al. (2002) proposed that there is a threshold amount of wax ester that is needed to achieve moulting and gonad maturation and the energetic requirements associated with dormancy. Accumulation of this threshold amount (estimated to be somewhere between 25% and 50% of dry weight for a 150–300 µg dry weight stage CV) would trigger physiological responses, likely hormonally mediated (Irigoien, 2004), to descend to deep water and enter dormancy. Individuals that did not attain this threshold would remain at the surface. Observations of "fat" and "thin" (related to oil sac volume) Calanus in deep and surface waters, respectively, of Georges Bank and the Gulf of Maine (Miller et al., 2000; Hassett, 2006) are consistent with this hypothesis.
What factors might control the duration of and arousal from dormancy? Photoperiod has been hypothesized to induce emergence from dormancy, analogous to the situation in many insect species. Miller et al. (1991) proposed that the mechanism for arousal from dormancy (they did not examine factors that may induce dormancy) in a population of C. finmarchicus in the slope water off southern New England may be the accelerated change in daylength in February, which would serve as a direct environmental cue. On the other hand, perhaps some time during winter, C. finmarchicus may simply stop responding to a diapause-maintaining factor and emerge shortly thereafter, perhaps dependent on the ambient temperature at the depth of dormancy or triggered by photoperiod. Speirs et al. (2005) also proposed a photo-awakening hypothesis for dormancy termination. In their implementation of a spatially explicit model in which copepods with different life histories may be transported to the same location, they concluded that a photoperiod cue (12 h) provided a better fit to synchronous emergence of adults in the surface 100 m in time-series of C. finmarchicus abundance in the Northeast Atlantic than an endogenous timer controlled by a reduced temperature-dependent development rate during dormancy.
Another factor that may be involved in dormancy termination is energy consumption, particularly metabolism of storage lipids, during dormancy (Miller et al., 1991; Hirche, 1996; Ohman et al., 1998; Visser and Jónasdóttir, 1999; Irigoien, 2004; Saumweber and Durbin, 2006). The length of dormancy might be controlled by the quantity of lipid reserve built up before entering dormancy, i.e. a copepod might become active when it has depleted its lipid reserve to a certain level. Although the amount used up during the dormancy period is small, it may be critical to retain a fairly large quantity of lipid for use in egg production or adult metabolism on emergence from dormancy, especially in species that emerge before the primary spring bloom (Rey-Rassat et al., 2002; Irigoien, 2004). If this hypothesis is correct, it also suggests that the onset of dormancy may require a minimum level of lipids. This mechanism is not seasonally predictive, but rather reflects the copepod's recent feeding history. Recent metabolic rate analysis of C. finmarchicus in the Gulf of Maine supports this hypothesis: the length of dormancy is limited to a few months at the warm deep-water temperatures of the Gulf of Maine, and copepods entering dormancy early must emerge several months before the spring bloom (Saumweber and Durbin, 2006). Under this hypothesis, high interannual variability in deep-water temperature in the Northwest Atlantic could result in variability in emergence timing (Bugden, 1991; Petrie and Drinkwater, 1993).
A third hypothesis is that the length of dormancy is controlled by an endogenous developmental trigger (Miller et al., 1991; Hirche, 1996). After copepods enter dormancy as CV, development may proceed at a reduced rate compared with active CV development. The timing of emergence from the dormant phase may be determined by development to a given point in the moult cycle, after which copepods proceed with moulting and active development. This type of scenario, in which the development rate during dormancy is a fraction of the temperature-dependent active-phase development rate, provided good agreement among C. finmarchicus seasonal populations cycles at several Northeast Atlantic time-series stations in a model developed by Hind et al. (2000), and it provides a mechanism for emergence in copepods from deep water, where seasonal signals may not be detectable. Alternatively, there may be an endogenous, long-term timer that triggers the maturation of gonads and subsequent arousal from dormancy in late winter.
Here, we investigate the control of dormancy timing and duration in populations of C. finmarchicus in the shelf regions of the Northwest Atlantic. We compile environmental and C. finmarchicus demographic data from the Atlantic Zone Monitoring Programme (AZMP) time-series, taken over a 5–12 year period at four stations spanning 
5° of latitude. We find, not unexpectedly, that no single environmental cue explains the observed seasonal patterns of dormancy entry and exit. However, the available data do not allow us to eliminate the lipid accumulation window hypothesis or a lipid-modulated endogenous timer as factors controlling dormancy duration, which are based on concepts put forward by Rey-Rassat et al. (2002), Irigoien (2004), and Saumweber and Durbin (2006), discussed earlier.
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Data sources
The timing of entry into and emergence from dormancy and associated environmental conditions were determined using data from the AZMP (Fisheries and Oceans Canada) fixed stations on the Newfoundland Shelf (NS; station S27), the Anticosti Gyre (AG), the lower St Lawrence Estuary (LSLE; station RIM), and the SS (station H2; Figure 1). Stations were sampled approximately every 2 weeks between the late 1990s and the mid-2000s (Table 1), with less-frequent sampling in winter during periods of ice cover or bad weather. Zooplankton was collected using 0.75 m diameter ringnets equipped with 200 µm mesh. Nets were towed vertically from
10 m above the bottom to the surface. Between 1994 and 2003 at the RIM station, zooplankton was sampled on a weekly basis with a net of 1 m diameter and 333 µm mesh (for stages CIV–VI) and a 1 m diameter 73 µm net (for CI–III and naupliar stages; Plourde et al., 2001). Zooplankton abundance was estimated by enumerating subsamples of zooplankton from the whole tow.
|
|
Environmental conditions were described at the fixed stations using CTD (conductivity, temperature, depth) cast data and chlorophyll a measured from Niskin-bottle samples, and averaged over the upper 50 m (http://www.meds-sdmm.dfo-mpo.gc.ca/zmp/main_zmp_e.html). Temperature at 5 m was used as an index of surface mixed-layer temperature. Deep-water temperature indices were calculated by averaging temperatures over the estimated depth range of dormant copepods at each of the four stations (NS, 100–160 m; AG, 150–250 m; LSLE, 150–240 m; SS, 100–140 m). Daylengths at each station were calculated as the time between sunrise and sunset (i.e. the sun 0.83 degrees below the horizon), as a function of latitude and day of year (Jarmo Lammi photoperiod calculator; http://personal.inet.fi/cool/jjlammi/stuff.html).
Proxies were developed to estimate dormancy entry and exit dates from time-series of developmental stage proportions in the AZMP zooplankton samples. Because samples were collected from almost the whole water column, the stage proportions represent both the active and deep, dormant parts of the population, so stage proportions can be used as an index of population life history status. The beginning of the dormancy onset period was defined as the date when the proportion of CVs in the population rose to half its overall maximum, calculated as each year's maximum CV proportion averaged over all years at that station. The start of the emergence period was defined as the first date when adults were more than 10% of the population of stages from copepodid stage 1 to adult. Both these proxies were seasonally persistent, i.e. once the threshold was crossed, values usually remained high for several months. However, in some cases the proportion of adults was transiently >10% in autumn or winter, so the proxy for the first emergence date was constrained to exclude dates followed by two or more dates when adults were <10% of the population.
Emergence dates were also estimated by back-calculating the spawning dates of the first early copepodid stages to appear in spring. The first appearance date for early copepodid stages was defined as the first date after the index emergence date when copepodid stages 1–3 (CI–III) exceeded 5% of stage CI to adult abundance. The development time from egg to CIII was estimated by running the individual-based model (IBM) described later, starting eggs at each date of the year, and identifying the appropriate development time for CIIIs at the first appearance date. The back-calculation was run using observed temperature and either observed chlorophyll a concentrations ("ambient") or chlorophyll a concentrations that are food-saturating for C. finmarchicus growth and development ("saturating") at each station. Spawning dates were calculated by subtracting the appropriate development time and seven additional days, the time required for egg production to begin after moulting (Runge, 1984; Plourde and Runge, 1993), from the first appearance date.
One-way analysis of variance (ANOVA) was used to test whether environmental factors at onset and emergence were different among stations. In cases where variance was unequal among stations, a one-way analysis of means, not assuming equal variance, was performed (Welch, 1951). A pairwise t-test with Bonferroni adjustment was used to test for differences in environmental factors among particular stations; however, a pairwise t-test with non-pooled standard deviation was used where variance was unequal among stations.
Stage-structured IBM
Because the field samples only provide abundance data for later copepodid stages, we required a method to back-calculate from the appearance of these later stages the date at which they would have been spawned. As food and temperature affect the growth and development of this species, and both are highly variable temporally and spatially, we used an IBM of C. finmarchicus to make these back-calculations. In the model, individuals develop through five defined "stages": an aggregated egg through third nauplius stage (E–NIII), an aggregated fourth through sixth naupliar stage (NIV–VI), and copepodid stages I–III. Each "stage" has unique parameters controlling the growth and development of the individuals within that stage as a function of temperature and prey concentration, using relationships derived from laboratory measurements (Campbell et al., 2001). In the model, individuals forage for food throughout a one-dimensional vertical water column, with 12 dynamic vertical layers (two layers from the surface to the bottom of the mixed layer, and ten layers spaced evenly from the base of the mixed layer to the bottom). The base of the mixed layer was determined as the depth at which there was a 0.5°C temperature difference from the surface. Therefore, as temperature and food concentrations vary with depth, each individual can have different growth and development rates at any moment and integrated over their lifetime, depending on the date on which they were spawned and their depth within the water column. Each day, five eggs were released into the model domain, with a vertical location chosen at random from throughout the mixed layer. Each individual was then tracked from this point until it reached the CIV stage.
In the model used here, we are mainly interested in development rate, which is somewhat decoupled from overall growth rate, so only this portion of the model is described. The development rate, Di (d–1), for each individual within each of the stages, i, is determined by a B
lehrádek function, modified by a "penalty" function which lengthens the time within a stage when there is food limitation, of the form
|
| (1) |
i and mi are coefficients that affect the development rate as a function of food concentration, derived empirically from laboratory studies (Campbell et al., 2001). The penalty function (the second term within square brackets) is essentially an Ivlev, Type-II relationship as a function of food. For the non-feeding stages, development is the same as Equation (1), without the feeding penalty function, i.e. simply the Blehrádek portion of the equation.
Beginning with the NIV stage (the second stage that feeds), the copepods were allowed to forage vertically through the one-dimensional water column. At each time-step, a copepod moves randomly either up or down, with the distance moved per model time-step, S, a function of the concentration of food at the copepod's current location and the size of the copepod, such that
|
| (2) |
|
Bimonthly vertical temperature data were obtained from a standard run of the CANDIE physical circulation model (Sheng et al., 1998, 2001). These data were then interpolated to provide a daily temperature field within which the copepods can develop. As the chlorophyll a observations were taken at irregular dates, and large gaps existed in data from some years and stations, a climatological daily chlorophyll a series was created using all data from a particular station. Any remaining gaps were filled via linear interpolation. Although there was vertical structure in some of the chlorophyll a profiles, for simplicity, the vertical structure was forced such that there was always a slight subsurface chlorophyll maximum (SCM) located at half the depth of the base of the mixed layer. Surface concentration was set to 75% of the SCM, increasing linearly to the depth of the SCM, decreasing linearly to 50% of the SCM at the base of the mixed layer, then linearly decreasing to 0 at the edge of the next deeper layer. For all stations, the minimum chlorophyll level was scaled to 0.3 µg Chl a l–1, the half-saturation constant for growth for the NIV–VI stage, using an assumption that there is always at least some minimum food resource (phyto- or microzooplankton) available to these stages.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
The timing of onset of dormancy differed by up to 3 months among the four stations (Figure 2, Table 3). Onset of dormancy started earliest on the SS (mean date, 10 June), and latest on the NS (mean date, 24 September). There was also considerable interannual variability within regions, as illustrated in Figures 3–5, which show the proxy dates of onset and emergence for each year at each of the stations arranged along data for each environmental variable. The standard deviation of the onset date ranged from 17 d in the LSLE, similar to the 2-week sampling resolution, to 26 d at the AG. The appearance of adults and small peaks in early copepodid stages after the proxy onset date at some stations in some years suggests that CVs enter dormancy over a broad range of times.
|
|
|
|
|
Emergence from dormancy, as estimated by increasing proportions of adults in the population, began earliest on the SS (mean date, 10 January), NS (mean date, 19 February), and in the AG (mean date, 23 February), and latest in the LSLE (mean date, 8 May; Table 3), where emergence dates were nearly always on the first sampling date of the season. The standard deviation of the emergence date ranged from 11 d in the LSLE to 51 d on the NS. The proportion of adults in the population ramped up steadily after exceeding 10% in all years and at all stations to a relatively high level, often more than half. During this period of increasing adult dominance, the total abundance of copepodid and adult stages was low, and this period likely represents the time when CVs leave dormancy, egg production has begun, and naupliar stages develop and increase in abundance. Dormancy emergence dates were estimated by back-calculation from the appearance of early copepodid stages for both unlimited food conditions and for ambient food conditions, estimated by a climatological daily average surface layer chlorophyll a concentration (Figure 6). The ambient and unlimited food estimates should bracket the actual emergence date; in most instances, the actual date is likely to be closer to the food-unlimited case. The back-calculations corresponded well in most years to estimates from our stage-proportion-based proxy (Figure 6), except the NS. There, the back-calculated emergence dates were considerably earlier, from 3 weeks to
4 months, depending on the state of food limitation. Early copepodid production persists sporadically through autumn and winter in this region, complicating the estimation of dormancy timing using the back-calculation method. The good correspondence between the two methods provides support for estimating emergence dates using increasing adult proportions, even at the LSLE station, where emergence dates are at the beginning of the sampling season, and this method was applied for the analyses presented below.
|
The four stations span
5 degrees of latitude, across which maximum and minimum daylengths at the stations differ by at most 45 min. The range of daylengths at which CVs started to enter dormancy was different among the four sites (Figure 3, Table 3). Daylength at onset of dormancy was highest (mean ± s.d., 15.3 ± 0.3 h) on the SS, where C. finmarchicus begins to enter dormancy close to the summer solstice. Onset was during a period of declining daylength at all other stations, and C. finmarchicus entered dormancy at the shortest daylengths on the NS (mean ± s.d., 12.1 ± 1.3 h). Emergence from dormancy was generally during periods of increasing daylength, but in several years, emergence began before the winter solstice. The mean daylength at emergence was longer at the LSLE station than at the three other stations (Figure 3, Table 3). Surface temperatures exhibit high seasonal variability throughout the region, with minimum temperatures in February or March coinciding with high vertical mixing, and maximum temperatures in late summer coinciding with strong stratification (Loder et al., 1998). Minimum temperatures at 5 m in winter were lower at the more northerly stations, reaching about –1°C in the St Lawrence Estuary and on the NS. The Gulf of St Lawrence is mostly covered with sea ice from January to late March (Koutitonsky and Bugden, 1991; Saucier et al., 2003), the NS is only ice-covered in some years. Minimum temperatures at 5 m dip to around 0.4°C on the SS, which does not have seasonal ice cover. The timing and amplitude of the seasonal 5 m temperature cycles at the four sites were very different (Figure 4). The highest maximum temperatures were observed on the SS in September and the lowest were in the LSLE in August. The onset of dormancy began during different phases of the seasonal temperature cycle at the four stations. On the SS, onset began while temperature was increasing. Onset began around the seasonal temperature peak in the LSLE and after the seasonal peak, when temperature was declining, on the NS and in the AG. The temperature at onset was not significantly different on the SS, NS, and in the AG, but was significantly lower in the LSLE. Temperatures at 5 m at the start of emergence from dormancy were higher on the SS than at the NS and AG stations (Table 3). Seasonal cycles of stratification, not shown here, were similar to the 5 m temperature at the four stations.
The seasonal chlorophyll cycles on both the NS and the SS were dominated by a spring peak in late March or April, and a much smaller peak in chlorophyll in late summer or autumn (Figure 5). Spring chlorophyll peaks occurred later in the LSLE and the AG, where chlorophyll concentrations remained higher through summer and early autumn. At the Gulf of St Lawrence stations, onset of dormancy began during periods when climatological chlorophyll concentrations remained relatively high, but onset nearly always began when chlorophyll was lower than annual maxima. The seasonal cycles of chlorophyll concentration on the Newfoundland and SS were similar, but whereas the onset of dormancy began about a month after the spring bloom on the SS, it did not begin until after the small autumn bloom on the NS. Chlorophyll concentrations experienced by copepods as the onset of dormancy began were not significantly different among stations (Table 3). The overall mean chlorophyll concentration at onset was 0.92 mg Chl a m–3 (s.d. = 0.85 mg Chl a m–3), lower than the threshold for maximum growth and development rates (Runge et al., 2006; R. G. Campbell, pers. comm.). Data on microzooplankton were not collected at the AZMP stations, so the contribution of this food source cannot be evaluated. Emergence always began either before the spring chlorophyll peak or when chlorophyll concentrations were increasing before the peak (Figure 5).
The duration of dormancy, estimated as the difference between the proxy dates of entry emergence, varied among regions (Figure 7a). The longest duration, 8–9 months, was estimated for the LSLE. Dormancy duration varied between 4 and 7 months in the AG and on the SS. Estimates of duration were shortest, 3–6 months, at the NS station, which also had the coldest deep-water temperature during the dormancy period. There was no inverse relationship between deep-water temperature and dormancy duration (Figure 7a), as would be expected from a temperature-dependent endogenous timer if initial lipid content was the same across regions. However, there was a significant inverse relationship between the duration of dormancy and the surface-layer temperature on the date of onset of dormancy, as would be expected if the quantity of lipid storage were temperature-dependent. As C. finmarchicus grow larger at lower temperatures, their total lipid content is also likely to be larger at low temperatures. Irigoien (2004) suggested that C. finmarchicus need to overwinter below the winter mixed-layer depth to avoid physical transport to the surface during dormancy. If dormant copepods rely on signals from the surface to trigger emergence, then a deeper mixed layer might imply longer dormancy duration, because surface signals might take longer to propagate into deeper water. However, such a mechanism seems unlikely, because the late winter and early spring mixed-layer depth is shallowest in the LSLE, where dormancy duration is longest.
|
|
Discussion |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
The data collected at the AZMP fixed stations constitute valuable time-series of plankton conditions representative of different coastal regions of the Northwest Atlantic. Nevertheless, there are limitations to interpreting dormancy patterns from the time-series. At all stations, advection becomes a significant source of error if "new" individuals entering the sampling area have different environmental histories, which could change the stage structure between sampling dates. The assumption in the analysis presented here is that each station is representative of the mean population dynamics over a broad regional area (Ouellet et al., 2003). This assumption is likely compromised to a varying extent across stations.
Cross-shelf advection from deep ocean basins is considered a primary mechanism for seeding C. finmarchicus onto continental shelves, and it may bias our estimates of the emergence or onset of dormancy in the data from the NS. However, AZMP surveys conducted in April on the NS provide little evidence of substantial differences in the relative stage composition of C. finmarchicus among the coastal areas, the continental slope, and the Labrador Sea (PP, unpublished data). If emergence had been earlier in the deep ocean basin, we might expect the C. finmarchicus community in deeper water to be in a more advanced state of development than on the continental shelf. Although existing circulation models indicate that cross-shelf transport of plankton would be limited because of the strong offshore branch of the Labrador Current, Helbig et al. (1992) argued that short-term variations in windforcing associated with winter storms could cause significant short-term cross-shelf mixing and exchange of plankton. Any bias in estimates of the emergence from dormancy may be reduced because of strong cross-shelf mixing, because C. finmarchicus in the Labrador Sea emerges during winter when storms are frequent.
The frequency of sampling is a constraint on precision and therefore contributes to the variability in the estimates of proxy dates of both onset and emergence. Ideally, the precision is 2 weeks, reflecting the nominal sampling frequency. In practice, the interval between sampling dates was sometimes longer at critical times of the year, particularly in winter. This would shorten apparent dormancy duration for estimates based on onset date and lengthen the duration for estimates based on emergence date. Estimates from multiple years at each station are therefore valuable for averaging out the particular sampling patterns of any given year.
The sampling frequency and dynamics at the Rimouski station in the LSLE and the AG in the northwest Gulf of St Lawrence need special consideration. Because of ice conditions and sampling logistics at the Rimouski station, sampling each year does not commence earlier than April, and it terminates in autumn. This is sufficiently early to capture the dynamics associated with the spring phytoplankton bloom, which is typically in mid-June (Plourde and Runge, 1993), and subsequent summer production, but it does not provide coverage during winter when emergence from dormancy and moulting to females would be expected. However, the C. finmarchicus demographics at the LSLE (i.e. RIM) station are likely a complex interaction of the dynamics of two subpopulations, one responding to the particular primary production cycle in the estuary, and one responding to the more classic production cycle present in the northwest gulf.
Recent research (SP, JAR, B. Zakardjian, J. J. Dodson, and P. Joly, unpublished data) indicates that a component of the LSLE subpopulation is at least phenotypically adapted to the late timing of the local spring bloom: it emerges from dormancy late (i.e. April or later), and produces eggs in June and throughout the summer period of high phytoplankton food availability (Figure 5). Only late-emerging adults and their offspring are maintained in the lower Gulf of St Lawrence Estuary and northwest Gulf of St Lawrence; individuals that emerge earlier, in late winter and early spring, are transported towards the central and southern gulf in the strong buoyancy-driven spring surface flow (Saucier et al., 2003; SP, F. Maps, B. Zakardjian, and F. Saucier, unpublished coupled physical–biological model simulations). In summer, adults and early stages are transported slowly from the LSLE into the northwest gulf and, according to simulations, largely retained and recirculated around the AG. They are transported back into the LSLE either in surface waters along the northern shore or, after entering dormancy in the northwest gulf, in the slower, deep inflowing current. Hence, most of the animals entering dormancy in the northwest gulf would originate in the LSLE, explaining the similar timing of onset of dormancy between these two regions. This local adaptation in the lower Estuary could also explain why, in most years, the emergence dates from the stage proportion and back-calculation methods correspond so well, despite the late start of sampling at the Rimouski station (Figure 6).
The slow upstream advection of CV from the northwest gulf to the lower Estuary also means that the overwintering stock in the northwest gulf is gradually replaced by animals originating in regions to the east (SP, F. Maps, B. Zakardjian, and F. Saucier, unpublished coupled physical–biological model simulations). These dynamics likely contribute to the much shorter duration of dormancy estimated for the northwest gulf compared with the LSLE, despite the similarity in environmental conditions in the two regions (Figure 7a, Table 3); populations with an early (originating in the central gulf) and late (originating in the lower Estuary) timing were sampled in the northwest Gulf of St Lawrence in winter/spring and summer/autumn, respectively. However, the early life stages present in April of some years at the RIM station (Figure 2) may be offspring of females that emigrated from the gulf and produced eggs in response to transient early phytoplankton blooms. Whether the summer subpopulation in the lower Estuary and northwest gulf is also genetically separated from the C. finmarchicus population in the rest of the gulf is not yet known.
We tested whether any of the environmental conditions were the same at the date of onset and emergence across all stations (Table 3). The results indicated that chlorophyll a concentration was the only environmental cue of the four that we examined that was relatively constant across all four stations. However, variability in chlorophyll a concentrations during summer, between June and September, was low at all stations, so chlorophyll a concentration alone did not offer a strong signal for induction of dormancy. Not only were daylength, sea surface temperature, and photoperiod different among stations, there was also considerable interannual variability in the timing of onset and emergence within stations. Moreover, as discussed earlier, analysis of the data presented in Figures 3–5 indicated no consistent pattern in the gradients of environmental cues that could explain the observed dormancy entrance and exit patterns. We conclude that a mechanistic understanding of dormancy transitions must involve interaction of multiple environmental factors.
Among the hypotheses put forward (see Introduction) to explain dormancy patterns in Calanus, only the lipid accumulation window hypothesis, which we expand on here, provides a mechanistic explanation involving a physiological response of the animals to their environmental history. Under this hypothesis, individuals can only enter dormancy if their food and temperature history allows them to accumulate sufficient lipid stores to sustain metabolism and support moulting and gonad maturation costs on emergence from the dormant phase (Rey-Rassat et al., 2002; Irigoien, 2004). The capacity for lipid store accumulation is controlled by the functional relationship between the production of wax esters that constitute the lipid stores (Miller et al., 1998) and ambient food availability and temperature. Presumably, these functional relationships resemble the asymptotic relationships of growth rate to food and temperature (Campbell et al., 2001). Superimposed on the accumulation rate of lipid is the development time–temperature relationship, such that lipid accumulation, similarly to growth, is constrained by the moulting rate at higher temperatures (Miller et al., 1977). There is, consequently, a seasonal window where temperature and food conditions are propitious for accumulation of lipids above the dormancy threshold (likely to be 30% on a weight-specific basis; Irigoien, 2004); if the threshold is not attained by, say, midway through the CV stage, the internal process to prepare for dormancy is not initiated and the individual continues on to moult into the adult stage. It is likely that the "decision" to prepare for dormancy is not made at the CV stage, but at a previous stage, probably CIV, in which case the relative amount of lipid stored during stage CIV could trigger a shift in energy allocation to lipid accumulation, rather than reproductive development, during the CV stage before descent and entry into dormancy. Such a lipid-mediated shift could be hormonally controlled, as suggested by Irigoien (2004).
Although lipid data are not available for C. finmarchicus collected at the AZMP stations to test this hypothesis directly, the lipid accumulation window hypothesis is consistent with several elements of the observed dormancy pattern. The interannual and among-station variability in timing of dormancy onset would be explained by variability in the timing and duration of the food and temperature conditions that form the lipid-accumulation window at each station in each year. The frequent observation of a small number of females present in autumn and winter would be the consequence of moulting through to the adult stage of a fraction of individuals that did not attain the lipid-store quota during the stage CIV/CV feeding period. This offers a mechanistic explanation for the empirical rule set by Speirs et al. (2006) that a fixed fraction of each generation enters dormancy. Large lipid stores observed in the LSLE subpopulation (40–70% dry weight: SP et al., unpublished data) are consistent with the favourable temperature and feeding conditions for lipid accumulation during summer in the lower Estuary (Figures 4 and 5), and allow stage CVs to sustain dormancy for up to 9 months (Figure 7).
The process for controlling dormancy duration and emergence would follow from concepts presented in Hind et al. (2000) and Saumweber and Durbin (2006). Development through stage CV continues at some fraction of the development time–temperature relationship determining CV stage duration during its active phase. However, the dormancy state can also be terminated early if the metabolic use of wax esters during dormancy levels depletes lipid stores below some critical level. Under this hypothesis, the negative relationship between dormancy duration and surface temperature (Figure 7b) would indicate that variable relative amounts of lipid stores are accumulated during the lipid-accumulation window, driven by constraints on lipid accumulation by higher temperatures (and perhaps lesser food availability). The absence of a negative relationship between dormancy duration and deep temperature at which the dormant individuals reside (Figure 7a) would be a consequence of the smaller relative lipid stores available to the NS individuals, forcing comparatively early exit from dormancy, despite the colder temperatures there.
As the next step in analysis of the validity of the lipid accumulation window hypothesis, we intend to extend the stage-based IBM for Calanus to include processes controlling lipid accumulation and dormancy duration. The full model can be applied to test the assumption that a single parameterization of demographic and physiological processes reproduces the observed dormancy patterns across regions. Although this is a necessary step towards quantitative formulation of the hypothesis for diagnostic and predictive purposes, the true test of the hypothesis will require new observations at sea and experiments for refining lipid-accumulation parameters in the laboratory. For example, measurement of lipid stores in CVs collected at or near the proxy date of dormancy onset would test the prediction that the inverse relationship between dormancy duration and temperature (Figure 7b) is attributable to different levels of lipid stores at the time of dormancy initiation. Field and laboratory experiments are needed to investigate further the nature of the relationship between food availability and feeding and lipid accumulation rates (Hakanson, 1984, 1987).
The advancement of a mechanistic understanding of dormancy patterns consistent with observations is critical for understanding impacts of climate variability on C. finmarchicus population dynamics. For example, the mechanism controlling emergence from dormancy, whether it is an inflexible function of photoperiod (Speirs et al., 2005) or a variable function of physiological processes influenced by previous temperature and feeding conditions, may have first-order repercussions on the population response to changes in ambient water temperature and the timing and duration of phytoplankton production cycles.
|
Acknowledgements |
|---|
We thank Michael Bates for assistance with data analysis. The AZMP time-series are maintained by funding from Fisheries and Oceans Canada. Support for the research was contributed by National Science Foundation and NOAA awards to JAR, AWL, and CLJ as part of the US GLOBEC programme. CLJ was also supported by a National Science Foundation International Postdoctoral Research Fellowship. This paper is US GLOBEC contribution number 559.
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
-
Bucklin A., Sundt R., Dahle G. Population genetics of Calanus finmarchicus in the North Atlantic. Ophelia (1996) 44:29–45.[Web of Science]
Bugden G. L. Changes in the temperature–salinity characteristics of the deeper water of the Gulf of St Lawrence over the past several decades. In: The Gulf of St Lawrence: Small Ocean or Large Estuary?—Therriault J-C., ed. (1991) 113. Canadian Special Publication in Fisheries and Aquatic Sciences. 139–147. 359.
Campbell R. G., Wagner M. M., Teegarden G. J., Boudreau C. A., Durbin E. G. Growth and development rates of the copepod Calanus finmarchicus reared in the laboratory. Marine Ecology Progress Series (2001) 221:161–183.[CrossRef][Web of Science]
Carlotti F., Radach G. Seasonal dynamics of phytoplankton and Calanus finmarchicus in the North Sea as revealed by a coupled one-dimensional model. Limnology and Oceanography (1996) 41:522–539.[Web of Science]
Carlotti F., Wolf U. A Lagrangian ensemble model of Calanus finmarchicus coupled with a 1-D ecosystem model. Fisheries Oceanography (1998) 7:191–204.[CrossRef][Web of Science]
Conover R. J. Comparative life histories in the genera Calanus and Neocalanus in high latitudes of the northern hemisphere. Hydrobiologia 167/168 (1988) 127–142.
Dahms H. U. Dormancy in the copepoda—an overview. Hydrobiologia (1995) 306:199–211.[CrossRef][Web of Science]
Danks H. V. Insect dormancy: an ecological perspective. In: Biological Survey of Canada (Terrestrial Arthropods), (1987) Ottawa. 439.
Davis C. S. Zooplankton life cycles. In: Georges Bank—Backus R. H., ed. (1987) Cambridge, MA: MIT Press. 256–267. 593.
Durbin E. G., Campbell R. G., Casas M. C., Ohman M. D., Niehoff B., Runge J. A., Wagner M. Interannual variability in phytoplankton blooms and zooplankton productivity and abundance in the Gulf of Maine during winter. Marine Ecology Progress Series (2003) 254:81–100.[CrossRef][Web of Science]
Durbin E. G., Runge J. A., Campbell R. G., Garrahan P. R., Casas M. C., Plourde S. Late fall–early winter recruitment of Calanus finmarchicus on Georges Bank. Marine Ecology Progress Series (1997) 151:103–114.[CrossRef][Web of Science]
Fiksen Ø. The adaptive timing of diapause—a search for evolutionarily robust strategies in Calanus finmarchicus. ICES Journal of Marine Science (2000) 57:1825–1833.
Hairston N. G., Dillon T. A., De Stasio B. T. A field test for the cues of diapause in a freshwater copepod. Ecology (1990) 71:2218–2223.[CrossRef][Web of Science]
Hairston N. G., Kearns C. M. The interaction of photoperiod and temperature in diapause timing: a copepod example. Biological Bulletin (1995) 189:733–751.
Hairston N. G., Munns W. R. The timing of copepod diapause as an evolutionarily stable strategy. The American Naturalist (1984) 123:733–751.[CrossRef]
Hakanson J. L. The long and short term feeding conditions in field-caught Calanus pacificus, as determined from lipid content. Limnology and Oceanography (1984) 29:794–804.[Web of Science]
Hakanson J. L. The feeding condition of Calanus pacificus and other zooplankton in relation to phytoplankton pigments in the California Current. Limnology and Oceanography (1987) 32:881–894.[Web of Science]
Hassett R. P. Physiological characteristics of lipid-rich "fat" and lipid-poor "thin" morphotypes of individual Calanus finmarchicus C5 copepodites in nearshore Gulf of Maine. Limnology and Oceanography (2006) 51:997–1003.[Web of Science]
Head E. J. H., Pepin P., Runge J. A. Proceedings of the workshop on "The Northwest Atlantic ecosystem—a basin scale approach". CSAS Proceedings Series, 2001/23. (2001) 112.
Helbig J. A., Mertz G., Pepin P. Environmental influences on the recruitment of Newfoundland/Labrador cod. Fisheries Oceanography (1992) 1:39–56.[CrossRef]
Hind A., Gurney W. S. C., Heath M., Bryant A. D. Overwintering strategies in Calanus finmarchicus. Marine Ecology Progress Series (2000) 193:95–107.[CrossRef][Web of Science]
Hirche J. Overwintering of Calanus finmarchicus and Calanus helgolandicus. Marine Ecology Progress Series (1983) 11:281–290.[CrossRef][Web of Science]
Hirche J. Spatial distribution of digestive enzyme activities of Calanus finmarchicus and Calanus hyperboreus in Fram Strait, Greenland Sea. Journal of Plankton Research (1989) 11:431–443.
Hirche J. Diapause in the marine copepod, Calanus finmarchicus—a review. Ophelia (1996) 44:129–143.[Web of Science]
Irigoien X. Some ideas about the role of lipids in the life cycle of Calanus finmarchicus. Journal of Plankton Research (2004) 26:259–263.
Johnson C. L. Seasonal variation in the molt status of an oceanic copepod. Progress in Oceanography (2004) 62:15–32.[CrossRef][Web of Science]
Johnson J. K. Effects of temperature and salinity on production and hatching of dormant eggs of Acartia californiensis (Copepoda) in an Oregon estuary. Fishery Bulletin US (1979) 77:567–584.
Koutitonsky V. G., Bugden G. L. The physical oceanography of the Gulf of St Lawrence: a review with emphasis on the synoptic variability of the motion. In: The Gulf of St Lawrence: Small Ocean or Large Estuary?—Therriault J-C., ed. (1991) 113. Canadian Special Publications in Fisheries and Aquatic Sciences. 57–90. 359.
Lee R. F. J., Nevenzel J. C., Paffenhoffer A. Wax esters in marine copepods. Science (1970) 167:1510–1511.
Leising A. W., Franks P. J. S. Copepod vertical distribution within a spatially variable food source: a simple foraging-strategy model. Hydrobiologia (2000) 22:999–1024.
Loder J. W., Petrie B., Gawarkiewicz G. The coastal ocean off northeastern North America: a large-scale view. In: The Sea—Robinson A. R., Brink K., eds. (1998) 11. Cambridge, MA: Harvard University Press. 105–133. 1062.
Marcus N. H. Photoperiodic and temperature regulation of diapause in Labidocera aestiva (Copepoda: Calanoidea). Biological Bulletin (1982) 162:45–52.
McLaren I. A., Head E. J. H., Sameoto D. D. Life cycles and seasonal distributions of Calanus finmarchicus on the central Scotian Shelf. Canadian Journal of Fisheries and Aquatic Sciences (2001) 58:659–670.
Meise C. J., O'Reilly J. E. Spatial and seasonal patterns in abundance and age-composition of Calanus finmarchicus in the Gulf of Maine and on Georges Bank. Deep-Sea Research II (1996) 43:1473–1501.[CrossRef]
Miller C. B., Cowles T. J., Wiebe P. H., Copley N. J., Grigg H. Phenology in Calanus finmarchicus: hypotheses about control mechanisms. Marine Ecology Progress Series (1991) 72:79–91.[CrossRef][Web of Science]
Miller C. B., Crain J. A., Morgan C. A. Oil storage variability in Calanus finmarchicus. ICES Journal of Marine Science (2000) 57:1786–1799.
Miller C. B., Johnson J. K., Heinle D. K. Growth rules in the marine copepod genus Acartia. Limnology and Oceanography (1977) 22:326–335.[Web of Science]
Miller C. B., Morgan C. A., Prahl F. G., Sparrow M. A. Storage lipids of the copepod Calanus finmarchicus from Georges Bank and the Gulf of Maine. Limnology and Oceanography (1998) 43:488–497.[Web of Science]
Norrbin F. M. Timing of diapause in relation to the onset of winter in the high-latitude copepods Pseudocalanus acuspes and Acartia longiremus. Marine Ecology Progress Series (1996) 142:99–109.[CrossRef][Web of Science]
Ohman M. D., Drits A. V., Clarke M. E., Plourde S. Differential dormancy of co-occurring copepods. Deep-Sea Research II (1998) 45:1709–1740.[CrossRef]
Ouellet M., Petrie B., Chassé J. Temporal and spatial scales of sea-surface temperature variability in Canadian Atlantic waters. Canadian Technical Report of Hydrography and Ocean Sciences (2003) 999:30.
Petrie B., Drinkwater K. F. Temperature and salinity variability on the Scotian Shelf and in the Gulf of Maine 1945–1990. Journal of Geophysical Research (1993) 98:20079–20089.[CrossRef]
Planque B., Hays G. C., Ibanez F., Gamble J. C. Large scale spatial variations in the seasonal abundance of Calanus finmarchicus. Deep-Sea Research I (1997) 44:315–326.[CrossRef]
Platt T., Fuentes-Yaco C., Frank K. T. Spring algal bloom and larval fish survival. Nature (2003) 423:398–399.[Medline]
Plourde S., Joly P., Runge J. A., Zakardjian B., Dodson J. J. Life cycle of Calanus finmarchicus in the lower St Lawrence Estuary: the imprint of circulation and late timing of the spring bloom. Canadian Journal of Fisheries and Aquatic Sciences (2001) 58:647–658.
Plourde S., Runge J. A. Reproduction of the planktonic copepod, Calanus finmarchicus, in the lower St Lawrence Estuary: relation to the cycle of phytoplankton production and evidence for a Calanus pump. Marine Ecology Progress Series (1993) 102:217–227.[CrossRef][Web of Science]
Rey-Rassat C., Irigoien X., Harris R., Carlotti F. Energetic cost of gonad development in Calanus finmarchicus and C. helgolandicus. Marine Ecology Progress Series (2002) 238:301–306.[CrossRef][Web of Science]
Runge J. A. Egg production of Calanus pacificus Brodsky: laboratory observations. Journal of Experimental Marine Biology and Ecology (1984) 74:53–66.[CrossRef][Web of Science]
Runge J. A., Franks P. J. S., Gentleman W. C., Megrey B. A., Rose K. A., Werner F. E., Zakardjian B. Diagnosis and prediction of variability in secondary production and fish recruitment processes: developments in physical–biological modeling. In: The Sea, 13, The Global Coastal Ocean: Multi-Scale Interdisciplinary Processes—Robinson A. R., Brink K., eds. (2005) Cambridge, MA: Harvard University Press. 413–473. 1033.
Runge J. A., Plourde S., Joly P., Durbin E. G., Niehoff B. Characteristics of egg production of the planktonic copepod, Calanus finmarchicus, on Georges Bank: 1994–1999. Deep-Sea Research II (2006) 53:2618–2631.[CrossRef]
Sameoto D. D. Decadal changes in phytoplankton color index and selected calanoid copepods in continuous plankton recorder data from the Scotian Shelf. Canadian Journal of Fisheries and Aquatic Sciences (2001) 58:749–761.
Saucier F. J., Roy R., Gilbert P., Pellerin H., Ritchie H. Modeling the formation and circulation processes of water masses and sea ice in the Gulf of St Lawrence. Journal of Geophysical Research (2003) 108:3269–3289.[CrossRef]
Saumweber W., Durbin E. G. Estimating potential diapause duration in Calanus finmarchicus. Deep-Sea Research II (2006) 53:2597–2617.[CrossRef]
Sheng J., Greatbatch R. J., Wright D. Improving the utility of ocean circulation models through adjustment of the momentum balance. Journal of Geophysical Research, (2001) 106:16711–16728.
Sheng J., Wright W. R., Greatbatch R. J., Dietrich D. E. CANDIE: a new version of the DieCAST ocean circulation model. Journal of Atmospheric and Oceanic Technology (1998) 15:1414–1432.[CrossRef][Web of Science]
Speirs D. C., Gurney W. S. C., Heath M., Horbelt W., Wood S. N., de Cuevas B. A. Ocean-scale modeling of the distribution, abundance, and seasonal dynamics of the copepod, Calanus finmarchicus. Marine Ecology Progress Series (2006) 313:173–192.[CrossRef][Web of Science]
Speirs D. C., Gurney W. S. C., Heath M., Wood S. N. Modeling the basin-scale demography of Calanus finmarchicus in the north-east Atlantic. Fisheries Oceanography (2005) 14:333–358.[CrossRef][Web of Science]
Uye S. Resting egg production as a life history strategy of marine planktonic copepods. Bulletin of Marine Science (1985) 37:440–449.[Web of Science]
Visser A. W., Jónasdóttir S. H. Lipids, buoyancy and the seasonal vertical migration of Calanus finmarchicus. Fisheries Oceanography (1999) 8:100–106.[CrossRef]
Wagner M. M., Durbin E. G., Buckley L. J. RNA: DNA ratios as indicators of nutritional condition in the copepod Calanus finmarchicus. Marine Ecology Progress Series (1998) 162:173–181.[CrossRef][Web of Science]
Watson N. H. F. Variability of diapause in copepods. In: Proceedings of the Second International Conference on Copepoda—Shriever G., Schminke H. K., Shih C-T., eds. (1986) 58. Syllogeus. 509–513. 662.
Watson N. H. F., Smallman B. N. The role of photoperiod and temperature in the induction and termination of an arrested development in two species of freshwater cyclopoid copepods. Canadian Journal of Zoology (1971) 49:855–862.
Welch B. L. On the comparison of several mean values. Biometrika (1951) 38:330–338.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Plourde, P. Pepin, and E. J. H. Head Long-term seasonal and spatial patterns in mortality and survival of Calanus finmarchicus across the Atlantic Zone Monitoring Programme region, Northwest Atlantic ICES J. Mar. Sci., October 1, 2009; 66(9): 1942 - 1958. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||