ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on November 8, 2006
ICES Journal of Marine Science: Journal du Conseil 2007 64(1):69-82; doi:10.1093/icesjms/fsl018
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The dynamics of a recovering fish stock: Georges Bank herring
Department of Fisheries and Oceans, Biological Station, 531 Brandy Cove Road, St Andrews, NB, Canada E5B 2L9
Correspondence to G. D. Melvin: tel: +1 506 529-8854; fax: 506 529-5862; e-mail: melving{at}mar.dfo-mpo.gc.ca
Melvin, G. D., and Stephenson, R. L. 2007. The dynamics of a recovering fish stock: Georges Bank herring – ICES Journal of Marine Science, 64, 69–82.The decline and subsequent recovery of Georges Bank Atlantic herring (Clupea harengus) provides a rare opportunity to examine the dynamics of a recovering fish population. Moreover, the near absence of a commercial fishery on Georges Bank between 1978 and 1995 removes the confounding effects of exploitation during the recovery period. Herring abundance on Georges Bank increased and the distribution of adult spawning fish evolved from a few isolated locations to most of the northern fringe during the period 1983–1995. The distribution of recently hatched larvae also expanded in a manner consistent with progressive occupation of historical spawning grounds. Changes in the size composition and age structure of herring during the spawning season broadened from the dominance of a single age class to multiple year classes as the stock recovered and expanded. Growth, as reflected by length-at-age, decreased significantly and was correlated with the number of fish estimated to be in the stock. This and the observed difference in mean length and length at first spawning during the recovery provide strong evidence of density-dependent growth. In particular, there is a highly correlated (p<0.01) relationship between the number of 4+ herring in the stock at the start of the year and the mean length of herring aged 3 recruiting to the spawning stock in autumn of the same year. A mechanism based on an extended period of prespawning interaction is proposed to explain the density-dependence.
Keywords: age structure, density-dependent growth, fishery, herring, population dynamics, stock recovery
Received 24 February 2006; accepted 25 September 2006; advance access publication 8 November 2006.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
Atlantic herring (Clupea harengus), are distributed along the east coast of North America from Labrador to Cape Hatteras, but the major spawning grounds are limited to the middle latitude regions (Cape Cod to Newfoundland) of this broad geographical range (Scott and Scott, 1988). Three stock complexes are recognized in the Gulf of Maine/Bay of Fundy region. These are defined based on spawning affiliation as Southwest Nova Scotia–Bay of Fundy (NAFO Subdivision 4X), coastal Gulf of Maine (NAFO Subdivision 5Y), and Georges Bank (NAFO Subdivision 5Z). The latter includes spawning areas on Georges Bank and Nantucket Shoals. In the late 1970s, the Georges Bank component of the 5Z complex virtually disappeared, but it has subsequently recovered to abundance levels near or exceeding those of the 1960s (NEFSC, 1998). In 2003, the component was estimated to have contributed 85–90% of the spawning-stock biomass of the combined Gulf of Maine (5Y and 5Z) stock complexes (DFO, 2003; Overholtz et. al., 2003).
Prior to its collapse in 1977, Georges Bank supported the largest herring fishery in the Western Atlantic. Reported commercial landings by international and domestic fleets during the late 1960s and early 1970s exceeded 200 000 t annually. The fishery peaked in 1968 with reported landings of 374 000 t (Fogarty et al., 1989). Unfortunately, the high level of annual exploitation exerted by multi-national fleets, combined with several years of poor recruitment, resulted in such a dramatic decline in abundance (Anthony and Waring, 1980; Grosslein, 1987) that in 1977 just 2127 t were landed from an allocation of only 3000 t. In essence, the stock and the fishery had collapsed. Between 1978 and 1985, virtually no adult or larval herring were detected on Georges Bank east of the Great South Channel by USA autumn research surveys (Melvin et al., 1996), except for a single midwater trawl set in June of 1984 which collected more than 200 juvenile (1+) herring from the 1983 year class (Stephenson and Power, 1989; Melvin et al., 1992).
The first real signs of recovery were in 1986 when both Canadian and US research surveys reported reproductively active adult fish and recently hatched herring larvae on Georges Bank (Stephenson and Power, 1989; Stephenson and Kornfield, 1990; Melvin and Fife, 1993; Melvin et al., 1993; Smith and Morse, 1993). Throughout the late 1980s and early 1990s, multiple research surveys observed progressively stronger indications of a major recovery of this collapsed herring stock (Melvin et al., 1996; NEFSC, 1996). In 1992, both Canada and the USA documented the protraction of reproductively active herring along the northern fringe to include historical spawning grounds on the Canadian portion of Georges Bank (Melvin et al., 1993; Smith, 1993). By 1993, recently hatched herring larvae (<10 mm) were observed over almost the entire eastern bank (Melvin et al., 1994), consistent with the distribution prior to, and during, the early stages of the foreign fishery (Tibbo et al., 1958; Tibbo and Legare, 1960; Boyar et. al., 1973). About 1995, the stock began to increase at an even more rapid rate, and by 2002, it had reached its greatest abundance in 25 y (NEFSC, 1998; Overholtz and Friedland, 2002; DFO, 2003).
Two hypotheses have been advanced for the source of fish for the recovery: the resurgence of the residual spawning stock (Stephenson and Kornfield, 1990) and recolonization of Georges Bank from adjacent stocks (Smith and Morse, 1993). Stephenson and Kornfield (1990) postulate that the reappearance of herring on Georges Bank was the result of resurgence of fish from the original population/stock. They cite differences in age composition and isozyme characteristics of Georges Bank herring from that of adjacent groups and elapsed time from collapse to recovery as evidence for resurgence. Conversely, Smith and Morse (1993) argue that observations of larval drift and mixing, from 20 y of distribution data, during four seasons annually, were inconsistent with the Sinclair and Iles (1985) hypothesis of larval retention and herring stock structure in the Gulf of Maine. They suggested that the observed changes in larval distribution from the pre-collapse to the recovery makes a convincing argument for recolonization rather than resurgence of Georges Bank herring (Smith and Morse, 1993). More recently, Overholtz and Friedland (2002) concluded that the US trawl survey data support the theory that Georges Bank was recolonized from adjacent spawning components.
Regardless of the mechanism, the reappearance of mature herring on Georges Bank provides a rare opportunity to investigate the changing dynamics of a once-depleted stock during a major portion of its recovery. Moreover, the absence of a directed commercial fishery for herring on Georges Bank between 1978 and 1995 removes the confounding interaction of heavy fishing pressure on stock characteristics. Landings from the herring fishery south of Cape Cod, where Georges Bank herring are known to overwinter, were also minimal (<2000 t annually) prior to 1996 (NEFSC, 1998). For Georges Bank herring, observed changes can be attributed almost entirely to natural variability as the stock expanded.
Here we explore the transformations in geographical distribution of larval and adult herring and the biological attributes of adult/juvenile fish as the abundance of the Georges Bank herring stock increased from the early to the late stages of recovery. A number of hypotheses regarding the dynamics of fish populations in a situation of rapid expansion are investigated, and a mechanism for density-dependent growth (in terms of size-at-age) of the recruiting year class is presented.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The recovery data summarized herein were collected during the Canadian autumn herring research surveys on Georges Bank (Table 1). The survey series was initiated in 1987 to track what at the time appeared to be the early stages of recovery of Georges Bank herring. In 1988, the geographical coverage of the survey was reduced to concentrate on that area of Georges Bank which was most likely to show signs of a recovery. The study area was expanded in 1991 to incorporate the entire northeastern bank. Figure 1 shows the survey area from 1988 to 1990 and the enlarged area surveyed from 1991 to 1995, when the series ended. Additional data on the distribution of adult and juvenile fish were obtained from Canadian autumn exploratory surveys in 1986, 1999, and 2001 of Georges Bank and over four decades of published information from the US autumn bottom trawl surveys series (1963–1996), which sampled Georges Bank during the autumn spawning period.
|
|
The annual Georges Bank adult/larval autumn herring survey was conducted from several Canadian research vessels: the RV "Lady Hammond" from 1987 to 1991, the RV "Parizeau" in 1992, and the RV "Alfred Needler" from 1993 to 1995. Both the "Lady Hammond" and the "Alfred Needler" were capable of deploying plankton gear to collect herring larvae, and bottom-trawl gear to sample fish. In 1992, bottom-trawl operations were conducted by the RV "E. E. Prince", and bongo gear was deployed from the RV "Parizeau". Ichthyoplankton were sampled using a standard bongo net (555 µm mesh) and a saw-tooth tow design with a minimum retrieval time of 10 min. Plankton samples were preserved in 4% buffered formalin at sea. Specimens were identified to species where possible, and the herring larvae were measured for total length.
Adult and juvenile herring were collected in Canadian surveys using a Western IIa bottom trawl towed for the standard 30 min. Sampling stations were selected at random and opportunistically (i.e. a set was made when herring were observed on the ship's sounder). US bottom-trawl surveys deployed a Yankee 36 bottom trawl to collect fish at randomly selected stations within a stratified design. The timing of both Canadian and US surveys varied slightly throughout the series. Standard survey sampling protocol was employed for juvenile and adult fish. All herring from sets containing catches of 250 fish or less were individually measured for total length and a subsample of a minimum of two fish per 0.5 cm interval was retained for detailed examination. When a set contained more than 250 herring, a random sample of approximately 250 fish was measured for total length and a subsample retained for detailed laboratory analysis as above. Laboratory analysis included measurements of total length (mm) and weight (0.1 g), visual examination of the gonads for sex and maturity stage, and removal of the sagittal otoliths for age determination. Length frequency and age results include all measured herring, with the age distribution apportioned using an age/length key developed from the detailed samples. Data presented on length-at-age and-maturity are for only those fish that had the characteristics measured (i.e. the detailed samples).
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The Canadian survey on Georges Bank began in 1986 and expanded in 1987 to monitor the distribution and abundance of herring adults and larvae during the early stages of the stock's recovery. Annual surveys of potential spawning areas between the Great South Channel and the Northeast Peak were conducted until 1995. During this 10-y period, 23 478 juvenile and adult herring were collected by bottom trawl at 197 stations during autumn on Georges Bank (Table 1). Of these, 9044 were measured for length and 2656 examined in detail for age, weight, sex, and maturity stage. The larval component of the survey began in 1987, and it collected 127 620 herring larvae from 684 bongo tows over 9 years (Table 1). Additional data were obtained from two monitoring surveys, with reduced geographical coverage, in autumn of 1999 and 2001.
Distribution of adult and juvenile herring
The distribution and abundance of adult and juvenile herring collected by research surveys on Georges Bank and in adjacent areas changed dramatically over the four decades during which the commercial fishery operated, i.e. from 1961. Figure 2 provides a chronological overview of herring distribution on Georges Bank, as reported by Canadian (1986–1995) and US (1965–2000) autumn research surveys, for representative 5 y intervals from 1965 to 2000. Annual plots of herring abundance and distribution are presented in several reports (Melvin et al., 1996; Overholtz and Friedland, 2002; Overholtz et al., 2004). During the early and peak years of the fishery, 1961–1970, herring were sparsely distributed around the bank and adjacent areas, with concentrations in the vicinity of known spawning areas (i.e. the northern reach of Georges Bank, Nantucket Shoals, and in Massachusetts Bay). However, the number of herring per tow from the US autumn survey was relatively low during this period, when compared with recent, post-recovery, observations (Overholtz et al., 2004).
|
Between 1971 and 1977, the abundance of herring decreased and the distribution contracted to a few areas on the northern fringe of Georges Bank and around Nantucket shoals (NEFSC, 1996). The mean number per standard tow decreased rapidly throughout the entire Gulf of Maine then (Friedland, 1998; Overholtz and Friedland, 2002). Indeed, by 1979, herring had all but disappeared from Georges Bank and Nantucket Shoals during their traditional spawning season (September–November), and only a single immature herring (total length 21 cm) was taken in Massachusetts Bay during the annual survey. This trend continued into 1980, when the US autumn survey caught no herring in 121 sets, and into 1981 when just two mature herring (26 and 33 cm) were collected at two stations on Georges Bank.
In 1982, mature adult herring began to appear again in limited numbers in all three traditional spawning areas. The distribution of herring on Georges Bank was, however, restricted to sampling stations in the vicinity of Little Georges and Cultivator Shoals. Trawl stations near Nantucket Shoals and in Massachusetts Bay generally showed a wider distribution and more herring per tow during the 1982–1985 spawning seasons. The first juvenile herring, from the 1983 year class, was caught on the Canadian portion of the bank in 1984, but it was not until 1986 that any adult herring (almost entirely from the 1983 year class) were caught east of the International Boundary during the autumn surveys.
From 1985 to 1989, the distribution broadened, and the catch of herring from bottom-trawl surveys increased substantially on Georges Bank and in the surrounding area, especially in Massachusetts Bay (Figure 2). In fact, the survey catches on Georges Bank from 1988 on exceeded those of the 1960s, when the stock was being heavily exploited. The expanding spawning distribution and increasing signs of abundance continued through the 1990s and into the new millennium. Herring catches during the US autumn bottom-trawl survey since 1995 represent the broadest distribution and highest stratified mean number per tow in the entire 35 y time-series (Figure 2). Spawning herring were consistently observed in Massachusetts Bay, throughout Nantucket Shoals, and along the northern flank of Georges Bank from the Great South Channel to the Northeast Peak, in an almost continuous band during the spawning period. In essence, based on research survey information, the stock appears not only to have recovered, but also to exceed by far the abundance of the pre-collapse era. The post-recovery distribution, however, is consistent with the reported distribution from the fishery during the early 1960s (Zinkevich, 1967).
Juvenile herring (age 2) were found among the pre-spawning adults and at most sampling stations, but their numbers were relatively small compared with adult fish. For the Canadian research survey series, juvenile herring constituted 1–12% of the catch by number, except in 1991 when the 1989 year class comprised 24% of all herring caught.
Distribution of herring larvae
The Canadian larval herring data also clearly demonstrate that spawning during the period 1986–1991, as reflected by the concentrations of larvae <10 mm long, was concentrated west of the Canadian/USA International Boundary in the vicinity of Georges and Cultivator Shoals (Figure 3). Plots of the annual sampling stations and larval distribution are presented in Melvin et al. (1996). During the early years of surveying, few recently hatched larvae were taken on the Canadian portion of the bank. However, by 1992, the distribution of larvae <10 mm had expanded well into Canadian waters. A similar distribution pattern continued through to 1995, when the survey was terminated. Further, in October 2001, a very large number (3 465 per 10 m2) of 5–7 mm larvae were collected in a single tow at a plankton station on the northern portion of Georges Bank, just east of the Canada/US boundary. This represents the largest concentration of larvae observed to date, and it provides strong evidence that herring had re-established and sustained spawning on the eastern portion of the bank, with larval distribution patterns similar to historical ones. The results are also consistent with US larval survey data (Smith and Morse, 1990).
|
Age/length frequency
During the early stages of recovery (i.e. 1986), catches were dominated (89%) by a single year class (1983). As the recovery progressed, the age distribution broadened, until in 1990 ages 2–10 were represented in herring samples (Figure 4). From 1986 to 1990, research catches of herring on Georges Bank were dominated by fish 3 y old, with the exception of 1988. Between 1992 and 1995, samples were dominated by ages 3–5, with ages 8+ poorly represented. The strong presence of fish 3 and/or 4 y old throughout the entire study period (1986–1995) suggested good annual recruitment to the spawning stock. Several strong year classes can be tracked by examining the age distributions. The 1983 and 1984 year classes, which first appeared as 3-y olds in 1986 and 1987, are well represented and traceable in samples over the next 6 and 5 y, respectively. The 1990 year class was also strongly represented in samples from 1993 to 1995, and it may represent another strong year class. However, the age distribution since 1994 has shifted to slightly older fish, with ages 4 and 5 dominating the samples (Figure 4). Limited sampling in autumn of 1999 and 2001 revealed a broad age distribution (ages 2–9), with fish from the 1998 year class (age 3) dominant during the later survey.
|
Trends in the length frequency distribution follow a pattern similar to the age structure and demonstrate a shift towards smaller fish as the stock recovered (Figure 5). During the early stages of recovery, fish lengths reflected the dominance of 3-y-olds in the stock, but as time progressed the distribution widened and several length modes could be tracked. For example, fish around the 1986 mode of 270 mm were observed in the length frequencies through to 1991. The broadest distribution was observed in 1991, when herring ranging from 20 to 34.5 cm total length (ages 2–10) were found in the research samples. Between 1993 and 1995, modal lengths generally increased, reflecting the increased presence of older fish, and lengths consistent with 4- and 5-y-olds dominated catches. Note that while the model lengths progressed with time, the length distribution about a given age was more contracted and did not extend into larger length intervals (28+ cm). In 1999, most herring were between 23 and 28 cm, indicating good recruitment of age 3 and older fish to the spawning stock (Figure 5). Length frequency samples in 2001 illustrate a strong recruitment of small fish from the 1998 year class, a weak representation of herring between 26 and 28 cm (age 4), and the continued presence of larger herring (age 5+).
|
Growth and maturity
Annual growth, as reflected by mean length-at-age of herring collected during the autumn spawning season on Georges Bank, underwent marked changes during the Canadian survey series. Mean length-at-age of fish aged 2–6 by sampling year showed a general and gradual decline in size from 1986 to 1995, with some interannual deviations from the decline (Figure 6). Large differences in mean length-at-age were to be expected for the older ages, but significant differences (p<0.01), of the order of 3–4 cm, were observed for age groups first recruiting to the spawning stock, i.e. ages 3 and 4.
|
Examination of the changes in mean length-at-age of individual year classes illustrates two growth patterns from age 4 on (Figure 6). Herring originating from the 1983–1985 year classes were consistently larger at a specific age than those spawned during later years of the recovery, the 1987–1991 cohorts. The size (year class mean lengths) at age 2 demonstrated no trend over the study period, whereas by age 4, there was an almost annual decrease in mean length of the 1983–1991 cohorts. Moreover, by the time the fish had reached ages of 5 and 6, there was a difference in mean length of 2 cm or more. The 1986 year class seems to represent a transition between the two trajectories.
In the Gulf of Maine, some 50% of herring mature in their third year, and by age 4 most, if not all, are sexually mature. Although the samples collected on Georges Bank during autumn do not represent the entire stock, because the majority of juvenile fish were unlikely to be on the spawning grounds, there was a shift in the percentage of age 3 herring mature. The proportion of mature age 3 fish (mean 55.8%) declined from a maximum of 88.9% in 1986 to a minimum of 27.0% in 1995. Concurrent with this shift was a change in the mean length-at-age by sampling year for ages 3 and 4 (Figure 7). Both age groups depict a decrease in mean length from 1989 until 1994, when age 4 showed a sharp increase and age 3 a slight upward trend. The mixture and interyear variability of mature and immature herring sampled on the spawning grounds confounds the age 3 mean lengths. Partitioning this age group into mature and immature fish demonstrated a general reduction in mean length-at-age from 1986 to 1994, excluding 1987. The sharp decline for age 3 fish in 1987 can be attributed to the large number of immature fish in the samples. In 1995, mature and immature fish showed opposite trends. There was a significant difference (p<0.01) between the mean length-at-age of mature and immature herring for nearly all years sampled (except 1986 and 1988), suggesting that factors influencing the size at recruitment may have changed from the early to the later stages of the recovery. The observed difference in mean length of mature age 3 herring was 
2 cm over the 9-y period, but larger if the 1999 and 2001 data are considered (Figure 7). Unfortunately, comparable data were not available prior to 1986.
|
Abundance
Currently, Georges Bank herring are assessed as part of the Gulf of Maine stock complex, which includes spawning components from the coastal Gulf of Maine, Georges Bank, and Nantucket Shoals (DFO, 2003). Several long-term fishery-independent indices of abundance, developed from research survey data, are available for the Gulf of Maine/Georges Bank stock complex (Overholtz, 2002). During the most recent assessment of the stock (in February 2003), two approaches were examined to evaluate stock status: a delayed-difference model (KLAMZ) and a calibrated VPA model (ADAPT) (DFO, 2003). Although there were divergent views on the absolute abundance of the stock complex between the two assessment models, the trends follow a consistent pattern (Figure 8). The stock began to increase rapidly around 1983, continued its upward trend at a slow rate in the late 1980s and early 1990s, then increased at an accelerated rate from 1995 to present (DFO, 2003; Overholtz et al., 2005). The uncertainty is not with the recent increasing trend in abundance, but in how well the indices of abundance reflect the relative changes and whether or not the current biomass level exceeds those of the 1960s. In addition, although the Georges Bank component dominated the Gulf of Maine stock complex for much of the 1960s and early 1970s, its contribution to the total biomass has varied from the early to the later stages of recovery. Currently, the spawning biomass of Georges Bank is estimated to account for
85–90% of the stock complex (DFO, 2003).
|
Density-dependent growth
To investigate the potential density-dependent relationship between stock size and growth (in terms of length-at-age), the number and biomass of herring estimated for the Gulf of Maine/Georges Bank stock complex from the 2003 VPA (DFO, 2003) was taken as a proxy for abundance on Georges Bank (Figure 8). Correlation analysis of mean length-at-age for ages 2–6 herring (including age 3 mature and immature) revealed several highly significant (p<0.01) relationships when compared with start-of-the-year abundance (number of fish) estimates for the period 1986–2001 (Table 2). Pearson correlation coefficient analysis indicated no significant (p>0.05) relationships among mean lengths-at-age and individual estimates, or combined estimates of the number of juvenile fish aged 1–3. However, significant (p<0.01) correlations were found between mean length-at-age and the number of 2+, 3+, and 4+ herring in the stock. The number of 4+ herring demonstrated the strongest negative correlation (p<0.01) with mean length at ages 3 and 4, a weak (p<0.05) correlation with mean length at ages 5 and 6, and no significant (p>0.05) relationship with mean length of age 2 herring. Comparison of mature and immature age 3 length-at-age also resulted in a strong negative correlation (p<0.1) with the start-of-the-year number of 4+ herring for mature fish and a significant (p<0.05) relationship for the immature fraction (Table 2). The results indicate a strong inverse relationship between the number of adult fish and mean length-at-age, and further imply that the observations were likely a function of the overall adult stock number, not the number of fish in individual cohorts (Figure 9). Log transformation of these data improved the significance of the relationship only slightly. Comparisons of mean length-at-age with spawning-stock biomass and age 3+ biomass produced similar results to that of numbers of fish, but the correlation coefficients were less significant.
|
|
Mean length-at-age 3 was positively correlated with mean length at ages 4 and 5, but not with age 2 or 6 (data not shown). Moreover, the mean length at age 2 was not significantly correlated with the mean length of any other age group, suggesting that the density-dependent relationship may be specifically related to recruiting year classes and the number of herring in the adult population.
Inclusion of 1999 and 2001 data generated a significant decrease in mean lengths (ANOVA, p<0.05) of age 3 mature and age 4 herring compared with those observed during the period 1986–1995. Mean lengths were the lowest in the time-series (age 3 being 23.3 cm in 1999, 24.35 cm in 2001) and stock numbers among the highest since the collapse (Figure 9). The addition of these 2 years of data reduced the correlation coefficient slightly, but the relationships between mean length of age 3 (mature) and 4 herring and the number of 3+ and 4+ fish remained significant (p<0.05). The persistence of a significant relationship between number of fish in the spawning stock and mean length provides additional support to the concept of density-dependent growth. Some effects of an increase in fishing effort after 1995 may have affected the 1999 and 2001 relationships.
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The collapse and subsequent recovery of Georges Bank herring provides an opportunity to examine the dynamics of a fish stock as abundance levels vary. Stock numbers have increased from extremely low in the late 1970s and early 1980s to approach or exceed levels of the 1960s, when the fishery was active. Concurrent with the collapse was the extension of jurisdiction by Canada and the USA to a 200-mile EEZ in 1978, which prohibited foreign fishing fleets from harvesting herring during much of the recovery. Furthermore, the near absence of a Canadian or US domestic herring fishery on Georges Bank from 1978 to 1995 removed the confounding effects of heavy exploitation on the observations presented here.
A number of expectations or hypotheses have been proposed as compensatory responses associated with the collapse and recovery of a marine fish stock. Spatial distribution of feeding and spawning aggregations are expected to contract as a population declines, then to expand as the recovery progresses (Smith and Morse, 1993; Melvin et al., 1996; Toresen and Østvedt, 2000; Overholtz, 2002; Berkeley et al., 2004; Beverton et al., 2004; Overholtz et al., 2004). Age structure, which is characteristically truncated by heavy exploitation (Lambert, 1987; Berkeley et al., 2004), tends to broaden during a recovery. For a pelagic species such as herring, with its highly variable reproductive success, a broad age structure is thought to help stabilize recruitment (Leaman and Beamish, 1984; Longhurst, 2002) as well as to contribute to the overall reproductive output through size-dependent fecundity and egg viability (Anthony and Fogarty, 1985; Lambert, 1990; Trippel et al., 1997). Moreover, marine species display phenotypic plasticity in growth-related characteristics in a density-dependent manner (Sinclair et al., 1982; Anthony and Fogarty, 1985; Toresen, 1990; Zebdi and Collie, 1995; Beverton et al., 2004; Engelhard and Heino, 2004a, b). For example, age at maturity is hypothesized to decline as abundance decreases, and growth rates to decrease as abundance increases. Monitoring of trends in these characteristics is also considered indicative of a fishery's sustainability (Berkeley et al., 2004). Georges Bank herring have exhibited characteristics consistent with all the above hypotheses for a declining and expanding population.
The northeastern Georges Bank supported the largest herring spawning grounds in the Bay of Fundy/Gulf of Maine during the 1950s, 1960s, and early 1970s (Tibbo et al., 1958; Tibbo and Legare 1960; Boyar et al., 1973; Anthony and Waring, 1983; Lough et al., 1985; Grosslein, 1987). Multinational ichthyoplankton studies, initiated in 1971, found recently hatched larvae abundant in two main areas within NAFO Division 5Z: on the northeastern portion of Georges Bank, consistent with earlier studies (Tibbo and Legare, 1960), and in the Great South Channel just southeast of Cape Cod, but west of 69°W (Nantucket Shoals) (Smith and Morse, 1993). Around 1972, there was an apparent shift in the abundance and distribution of larval herring from east to west on Georges Bank. The transition was over several years, but by 1977, there was a dramatic change in larval distribution. The main concentration of herring larvae in both number and physical area was around Nantucket Shoals, not Georges Bank. By 1978, recently hatched larvae had all but disappeared from the northeastern bank and remained absent from the area until 1986 (Lough et al., 1985; Smith and Morse, 1993). Even the distribution in the western portion of 5Z contracted to a few small areas of concentrated abundance between 1978 and 1984, although there appears to have been some expansion during the early 1980s into Massachusetts Bay.
In 1987, young larvae (<10 mm) were again detected in a small area of Georges Bank, but in relatively dense concentrations (200 larvae per 10 m2), on the northern middle portion of the bank just north of Georges and Cultivator Shoals (Melvin et al., 1996). Between 1987 and 1991, the restricted distribution of larvae expanded eastwards from an area of concentration along the fringe of the bank. At the same time, there was a continuing increase in abundance and expansion of larvae from the Massachusetts Bay/Nantucket area (not covered by the Canadian survey). However, it was not until 1992 that small larvae indicative of spawning were observed east of the Canada/US International Boundary, and until 1994 that dense aggregations were found over the northeastern bank consistent with historical observations (Figure 3). In 1995, when both Canadian and US larval surveys ended, the distribution of herring larvae resembled that reported during the 1950s in the pre-collapse era. The cyclical pattern of distributional contraction during heavy exploitation, near absence during the collapse, and gradual protraction during the recovery to historical levels appeared complete. However, the mean number (per 10 m2) of 4–7 mm larvae in the Nantucket Shoal area (5Ze) was three to four times that observed on Georges Bank between 1990 and 1994 (NEFSC, 1996). It was not until more recent years (i.e. post-2000) that greater concentrations were observed over the eastern Georges Bank (DFO, 2003: Overholtz et al., 2004).
The distribution and abundance observations of adult herring from Canadian and US autumn trawl surveys on Georges Bank support the hypothesis of contraction and expansion as stock numbers declined and subsequently increased. During the collapse, late 1970s to the mid-1980s, there was a virtual absence of herring on Georges Bank in the spawning season, an early increase in abundance in the Massachusetts Bay area during the mid-to late 1980s, and a gradual build-up of herring along the northern reach from the mid-1980s to the present. In fact, although similar, the spatial distribution during the recovery and recovered period was clearly broader than the historical pattern (Figure 2). The real disparity is the low spawning season research vessel catch rates of the pre-collapse period, when the stock biomass was considered relatively high, compared with those of the post-collapse period (NEFSC, 1998; DFO, 2003). Research catch rates on the Georges Bank from 1988 on are as much as an order of magnitude greater than those of the 1960s, when the stock was heavily exploited, implying a significant increase in abundance from pre- to post-collapse (Overholtz et al., 2004). In reality, however, the catch rate increases reflect not only abundance, but also changes in catchability resulting from gear modifications and/or changes in fish behaviour (DFO, 2003; Overholtz et al., 2004).
Temporal changes in the age structure and size distribution of herring collected on Georges Bank during the autumn spawning season were also consistent with what might be expected for a recovering/expanding population. The age structure and size frequencies varied from being dominated by a single year class (e.g. in 1986, 89% were aged 3, i.e. from the 1983 year class) during the early stage of the recovery to a narrow age distribution, and finally a mixture of young and old fish. The high percentage of 3- and 4-y-old herring present on the spawning grounds each year provided evidence of good recruitment as stock abundance increased. Several productive year classes can be tracked from year to year. Coincident with the first real sign of a recovery was the recruitment of the 1983 year class, a year class that was dominant in most stocks in the region and that made a significant contribution to the early recovery of the stock. Length frequency data, however, did not display the same gradual broadening as the age structure. The length distribution of fish on the spawning grounds was wide from 1986 on, likely reflecting the phenotypic plasticity of growth rates and size at maturity under increasing abundance levels. Size modes evident in the length distributions shifted to the right as year classes grew older.
Evidence of density-dependent growth and size at maturation for Atlantic herring and other fish species during variable abundance levels is mixed. Throughout the scientific literature, there are several papers that describe the occurrence and absence of density-dependent effects on growth, the driving mechanisms behind the observations, and their role in population dynamics as a compensatory/regulating factor (Frost and Kipling, 1967; Hubold, 1978; Toresen, 1990; Post et al., 1997, 1999; Beverton et al., 2004; Engelhard and Heino, 2004a, b). Le Cren (1958) found both population density and temperature to be growth-influencing factors for perch (Perca fluviatilis), but observed no relationship between population density in the first or second year of growth. Manipulating brown trout densities in streams, Jenkins et al. (1999) reported that the average body size (length or mass) of under-yearlings was negatively correlated with trout density. Those authors further suggest that detection of density-dependent growth from purely observational data of natural populations, with relatively high fish densities, may be difficult because of the stabilizing effects of a flat-topped growth–density relationship.
For Atlantic herring, studies have shown evidence in support of, and the absence of, density-dependent growth and maturation interactions. Iles (1968) reported a significant decline in the growth of North Sea 0–group herring in the 1940s when abundance of the unfished adult stock increased. Hubold (1978) found that maturation of North Sea herring of the same age was related to size. He found a significant negative relationship between the length at age 1 and at ages 1–3 and total stock size. Recent studies on the Norwegian spring-spawning herring have found a strong size and maturation-related compensatory response to changing abundance (Toresen, 1990; Beverton et al., 2004; Engelhard and Heino, 2004a). Engelhard and Heino (2004b) concluded that, because herring mature at a threshold size, changes in maturation are a function of abundance-related growth variation. Conversely, Molloy (1984) observed no relationship between mean size and population abundance for Irish Sea herring.
For the western Atlantic, Moores and Winters (1982) found weak evidence of density-dependent growth of herring in Fortune Bay, Newfoundland. They concluded rather that density-independent factors are the primary determinant of herring growth there and in the southern Gulf of St Lawrence. Sinclair et al. (1982) observed a significant negative correlation between the size of herring aged 3 and the abundance of 2 and 3 y olds and abundance for southwest Nova Scotia. However, Lett and Kohler (1976) found no evidence of density-dependent growth for adult herring in the southern Gulf of St Lawrence. Anthony and Fogarty (1985) demonstrated a relationship between temperature, growth rate, and abundance of herring in the Gulf of Maine and used their findings as evidence for density-dependent changes in length when year classes were strong. Grosslein et al. (1980) found no evidence for density-dependent growth of adult herring from Georges Bank.
Here we have provided evidence that density-dependent compensatory mechanisms have been regulating, at several stages, the mean size of adult herring relative to the number of adult fish in the Georges Bank herring stock during its recovery. Comparison of the mean length-at-age of ages 3 and 4 herring with the estimated number of 4+ fish at the start of the year (i.e. the adult population) shows a strong negative correlation for the data for the period 1987–1995. The decrease in mean length-at-age was expected given the estimated increase in abundance. Moreover, the relationship was maintained when length data from the 1999 and 2001 spawning seasons were used to predict the number of 4+ herring during a period of rapid increase in abundance (DFO, 2003; Overholtz, 2004). The mean sizes of herring aged 3 and 4 at spawning in 1999 and 2001 were the smallest total lengths observed in the time-series.
The driving mechanism for our observations appears to be the interaction of recruiting year classes and the number of adult fish in the stock. As age 3 juvenile herring join the adult stock, likely early during the summer feeding migration, they interact with older age groups (Figure 10). The number of adult fish combined with age 3 recruits may have a significant impact on their mean lengths 6–8 month later (at spawning). Generally, we observed a significant correlation (p<0.05) between the adult population (age 4+) and the mean length-at-age for ages 3–5, especially for the mean length of herring aged 3.
|
There also appears a secondary density-dependent compensatory mechanism regulating the number of juvenile herring recruiting to the spawning stock. For Gulf of Maine (including Georges Bank) herring stocks, empirical evidence suggests that some 50% of herring mature in their third year and that by age 4, most, if not all, are sexually mature. Separating age 3 herring into mature and immature fish reveals a significant difference in the mean length of adult and juvenile fish from the same cohort in most years. During the early years of the recovery for which data are available (1986–1988), the difference was small and insignificant for two of the three years (1986 and 1988). This suggests that, during times of low population abundance, all fish physiologically capable of spawning (i.e. reaching the size threshold necessary to spawn) would be recruited to the spawning portion of the population and that, during this phase of the recovery, most fish were larger than the minimum size threshold required to become sexually mature (Engelhard and Heino, 2004b). For age 3 herring, the mean size of both mature and immature fish and the percentage of fish reaching maturation have declined dramatically since 1986. Hence, as stock abundance increased, many fish from the first recruiting year class did not reach the size threshold for maturity in their third year, and only the larger herring from the cohort matured, resulting in the observed difference in mean lengths of mature and immature fish. The one anomaly in 1987 (the 1984 year class) can also be explained in relation to density-dependent interactions. The 1983 year class was likely the largest recruiting year class between 1978 and 1995. As such, its presence as 2-y-olds with the 1984 year class (age 1) as juveniles, and/or its contribution at age 4 to the number of adult herring in the stock, had a profound effect on the growth/size of the recruiting age 3 herring in 1987.
Density-dependent interactions were found both within and between the juvenile and adult population during the stock's collapse and recovery (Figure 10). The number of herring in the juvenile and the adult population has a negative effect on the mean size of cohorts within these temporally and spatially separated groups. In essence, the more fish in a cohort, the smaller the mean size and the smaller the mean size of other age groups, which are spatially associated. However, there also appears a mechanism to transcend the boundaries between juvenile and adult fish. The number of herring in the adult population (age 4+) and the recruiting juveniles (age 3), which interact throughout the summer feeding season, appears to influence the mean size of recruited spawners in autumn. During periods of low numbers of adults, most age 3 herring will reach a size beyond the threshold for sexual maturity, and a large portion of the recruiting juveniles will contribute to the spawning-stock biomass. Conversely, during periods of high numbers of adults, the growth and mean size of the recruiting juveniles is reduced, so decreasing the number of fish recruiting to the spawning-stock biomass. By age 4, almost all herring are mature and have been recruited to the adult spawning population.
To summarize, as the abundance of herring on Georges Bank increased, the distribution of spawning adults and recently hatched larvae expanded to include a large portion of the northern fringe of the bank, consistent with historical (pre-collapse) observations. The size composition of spawning herring widened as the stock recovered, and the age structure broadened from a dominant single age class to multiple year classes. Density-dependent interactions were observed in several biological characteristics. Between 1986 and 1995, we observed a significant declining trend in the percentage of maturing age 3 herring. Mean lengths at age and first spawning were negatively correlated with the number of fish in the adult population, thus providing strong evidence for density-dependent growth. Furthermore, given the absence of a commercial fishery on Georges Bank during the study period, the results can be attributed almost entirely to natural processes.
|
Acknowledgements |
|---|
We thank the many scientific staff, officers, and crew of the research vessels for their assistance in the collection of these data. Two anonymous reviewers provided valuable comments on the manuscript.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Anthony V. C. and Fogarty M. J. (1985) Environmental effect on recruitment, growth, and vulnerability of Atlantic herring (Clupea harengus harengus) in the Gulf of Maine region. Canadian Journal of Fisheries Aquatic Sciences 42:Suppl. 1, 158–173.
Anthony V. C. and Waring G. (1980) The assessment and management of the Georges Bank herring fishery. Rapports et Procès-Verbaux des Réunions du Conseil International pour l'Exploration de la Mer 177:72–111.
Berkeley S. A., Hixon M. A., Larson R. J., Love M. S. (2004) Fisheries sustainability via protection of age structure and spatial distribution of fish populations. Fisheries 29:823–32.
Beverton R. J. H, Hylen A., Østvedt O-J., Alvsvaag J., Iles T. C. (2004) Growth, maturation, and longevity cohorts of Norwegian spring-spawning herring. ICES Journal of Marine Science 61:165–175.
Boyar H. C., Marak R., Perkins F., Clifford R. (1973) Seasonal distribution and growth of larval herring (Clupea harengus L.) in the Georges Bank–Gulf of Maine area from 1962–1970. Journal du Conseil International pour l'exploration de la Mer 35:36–51.
DFO. (2003) Atlantic herring: Georges Bank, Nantucket Shoals, Gulf of Maine Stock Complex. DFO Science Stock Status Report, 2003/028 8.
Engelhard G. H. and Heino M. (2004a) Maturity changes in Norwegian spring-spawning herring Clupea harengus: compensatory or evolutionary responses. Marine Ecology Progress Series 272:245–256.[Web of Science]
Engelhard G. H. and Heino M. (2004b) Maturity changes in Norwegian spring-spawning herring before, during, and after a major population collapse. Fisheries Research 66:299–310.[CrossRef][Web of Science]
Fogarty M. J., Almeida F. P., Chenoweth J., Idione J. S. (1989) Population dynamics of Atlantic herring in the Gulf of Maine. US Stock Assessment Working Group, SAW Working Paper 7 50.
Friedland K. D. (1998) In Mooney-Seus M. L., Goebel J. L., Tausig H. C., Sweeney M. S. (Eds.). US larval and trawl surveys in the Gulf of Maine and plans for acoustic survey work. Herring Stock Assessment and Research Priorities 74–86 New England Aquarium Report.
Frost W. E. and Kipling C. (1967) A study of reproduction, early life, weight–length relationship and growth of pike Esox lucius L, in Windermere. Journal of Animal Ecology 36:651–693.[CrossRef]
Grosslein M. D. (1987) Synopsis of knowledge of the recruitment process for Atlantic herring (Clupea harengus), with special reference to Georges Bank. North Atlantic Fisheries Organization Science Council Studies 11:91–108.
Grosslein M. D., Langton R. W., Sissenwine M. P. (1980) Recent fluctuations in pelagic fish stocks of the northwest Atlantic, Georges Bank region, in relation to species interaction. Rapports et Procès-Verbaux des Réunions du Conseil International pour l'Exploration de la Mer 177:374–404.
Hubold G. (1978) Variations in growth rate and maturity of herring in the North Sea in the years 1955–1973. Rapports et Procès-Verbaux des Réunions du Conseil International pour l'Exploration de la Mer 172:154–163.
Iles T. D. (1968) Growth studies on North Sea herring. 2. 0-group growth of East Anglian herring. Journal du Conseil International pour l'Exploration de la Mer 32:98–116.
Jenkins T. M., Diehl S., Kratz K. W., Cooper S. D. (1999) Effects of population density on individual growth of brown trout in streams. Ecology 80:941–956.[CrossRef][Web of Science]
Lambert T. C. (1987) Duration and intensity of spawning in herring Clupea harengus as related to the age structure of the population. Marine Ecology Progress Series 39:209–220.
Lambert T. C. (1990) The effects of population structure on recruitment in herring. Journal du Conseil International pour l'Exploration de la Mer 47:249–255.
Leaman B. M. and Beamish R. J. (1984) Ecological and management implications of longevity in some northeast Pacific groundfishes. International North Pacific Fisheries Commission Bulletin 42:85–97.
Le Cren E. D. (1958) Observations on the growth of perch (Perca fluviatilis L.) over twenty-two years with special reference to the effects of temperature and changes in population density. Journal of Animal Ecology 27:287–334.[CrossRef]
Lett P. F. and Kohler A. C. (1976) Recruitment: a problem of multi-species interaction and environmental perturbations with special reference to Gulf of St. Lawrence Atlantic herring (Clupea harengus harengus). Journal of the Fisheries Research Board of Canada 33:1353–1371.[Web of Science]
Longhurst A. (2002) Murphy's law revisited. Longevity as a factor in recruitment to fish populations. Fisheries Research 56:125–131.[CrossRef]
Lough R. G., Bolz G. R., Pennington M., Grosslein M. D. (1985) Larval abundance and mortality of Atlantic herring (Clupea harengus L.) spawned in the Georges Banks and Nantucket Shoals areas, 1971–1978 seasons, in relation to spawning stock size. Journal of Northwest Atlantic Fisheries Science 6:21–35.
Melvin G. D. and Fife F. J. (1992) Comparison of Canadian and United States larval herring surveys on Georges Bank: 1987–1991. Canadian Atlantic Fisheries Scientific Advisory Committee 92/67:27.
Melvin G. D. and Fife F. J. (1993) Report on the 1992 Survey of Georges Bank using commercial herring seiners. DFO Atlantic Fisheries Research Document 93/73 24.
Melvin G. D., Fife F. J., Power M. J., Stephenson R. L. (1996) The 1996 Review of Georges Bank (5Z) herring stock. DFO Atlantic Fisheries Research Document 96/29 54.
Melvin G. D., Fife F. J., Sochasky J. B., Power M. J., Stephenson R. L. (1993) Georges Bank (5Z) herring 1993 update. DFO Atlantic Fisheries Research Document 93/75 24.
Melvin G. D., Fife F. J., Sochasky J. B., Power M. J., Stephenson R. L. (1994) Georges Bank (5Z) herring 1994 update. DFO Atlantic Fisheries Research Document 94/87 25.
Molloy J. (1984) Density dependent growth in Celtic Sea herring. ICES CM/Pelagic Fish Committee 1984 13.
Moores J. A. and Winters G. H. (1982) Growth patterns in a Newfoundland Atlantic herring (Clupea harengus harengus) stock. Canadian Journal of Fisheries and Aquatic Sciences 39:454–461.
NEFSC [Northeast Fisheries Science Center]. (1996) Report of the 21st Northeast Regional Stock Assessment Workshop (21st SAW). (NOAA/NMFS/NEFSC, Woods Hole, MA)71–94 NEFSC Reference Document 96–05d:.
NEFSC [Northeast Fisheries Science Center]. (1998) Northeast regional stock assessment workshop (27th SAW). Stock assessment review committee (SARC) consensus summary of assessments. Woods Hole Laboratory Reference Document 98–15.
Overholtz W. J. (2002) The Gulf of Maine–Georges Bank Atlantic herring (Clupea harengus): spatial pattern analysis of the collapse and recovery of a large marine fish complex. Fisheries Research 57:237–254.[CrossRef]
Overholtz W. J. and Friedland K. D. (2002) Recovery of the Gulf of Maine–Georges Bank Atlantic herring (Clupea harengus) complex: perspectives based on bottom trawl survey data. Fishery Bulletin US 100:593–608.
Overholtz W. J., Jacobson L. J., Melvin G. D., Cieri M., Power M. J., Libby D., Clark K. (2004) Stock assessment of the Gulf of Maine–Georges Bank herring stock complex, 1959–2002. Northeast Fisheries Science Center, Reference Document 04-06 300.
Post J. R., Parkinson E. A., Johnston N. T. (1999) Density dependent processes in structured fish populations: interaction strengths in whole lake experiments. Ecological Monographs 69:155–175.
Scott W. B. and Scott M. G. (1998) Atlantic fishes of Canada. Canadian Bulletin of Fisheries and Aquatic Sciences 219:731.
Sinclair M., Sinclair A., Iles T. D. (1982) Growth and maturation of southwest Nova Scotia Atlantic herring (Clupea harengus harengus). Canadian Journal of Fisheries and Aquatic Sciences 39:288–295.
Sinclair M. and Iles T. D. (1985) Atlantic herring (Clupea harengus) distribution in the Gulf of Maine – Scotian Shelf area in relation to hydrographic features. Canadian Journal of Fisheries and Aquatic Sciences 42:880–887.
Smith W. G. (1993) Herring spawning is extensive on Georges Bank. 4 Northeast Fisheries Science Center, Research Highlights, October–December 1993.
Smith W. G. and Morse W. W. (1990) Larval distribution patterns: evidence for the collapse/recolonization of Atlantic herring on Georges Bank. ICES Document CM/Pelagic Fish Committee 1990/H: 87 16.
Smith W. G. and Morse W. W. (1993) Larval distribution patterns: early signals for the collapse/recovery of Atlantic herring in the Georges Bank area. Fishery Bulletin US 91:338–347.
Stephenson R. L. and Kornfield I. (1990) Reappearance of Atlantic herring (Clupea harengus harengus) on Georges Bank: population resurgence not recolonization. Canadian Journal of Fisheries and Aquatic Sciences 47:1060–1064.
Stephenson R. L. and Power M. J. (1989) Reappearance of Georges Bank herring: a biological update. Canadian Atlantic Fishery Science Advisory Commission 89/60 14.
Tibbo S. N. and Legare J. E. H. (1960) Further study of larval herring (Clupea harengus L.) in the Bay of Fundy and Gulf of Maine. Journal of the Fisheries Research Board of Canada 17:933–942.
Tibbo S. N., Legare J. E. H., Scatterwood L. W., Temple R. F. (1958) On the occurrence and distribution of larval herring (Clupea harengus L.) in the Bay of Fundy and the Gulf of Maine. Journal of the Fisheries Research Board of Canada 15:1451–1469.
Toresen R. (1990) Long term changes in growth on Norwegian spring-spawning herring. Journal du Conseil International pour l'Exploration de la Mer 47:48–56.
Toresen R. and Østvedt O. J. (2000) Variation in abundance of Norwegian spring-spawning herring (Clupea harengus L.) throughout the 20th century and the influence of climate conditions. Fisheries 1:231–256.
Trippel E. A., Kjesbu O. S., Solemdal P. (1997) Effects of adult age and size structure on reproductive output in marine fishes. In Chambers R. C. and Trippel E. A. (Eds.). Early Life History and Recruitment in Fish Populations(Chapman and Hall, New York) pp. 31–62.
Zebdi A. and Collie J. S. (1995) Effects of climate on herring (Clupea pallasi) population dynamics in the Northeast Pacific Ocean. In Beamish R. J. (Ed.). Climate Change and Northern Fish PopulationsCanadian Special Publication of Fisheries and Aquatic Sciences pp. 277–290.
Zinkevich V. N. (1967) Observations on the distribution of herring, Clupea harengus L, on Georges Bank and in adjacent waters in 1962–1965. International Commission for the Northwest Atlantic Fisheries Research Bulletin 4:101–115.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. P. Wheeler, C. F. Purchase, P. D. M. Macdonald, R. Fill, L. Jacks, H. Wang, and C. Ye Temporal changes in maturation, mean length-at-age, and condition of spring-spawning Atlantic herring (Clupea harengus) in Newfoundland waters ICES J. Mar. Sci., September 1, 2009; 66(8): 1800 - 1807. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Stephenson, G. D. Melvin, and M. J. Power Population integrity and connectivity in Northwest Atlantic herring: a review of assumptions and evidence ICES J. Mar. Sci., September 1, 2009; 66(8): 1733 - 1739. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Husebo, A. Slotte, and E. K. Stenevik Growth of juvenile Norwegian spring-spawning herring in relation to latitudinal and interannual differences in temperature and fish density in their coastal and fjord nursery areas ICES J. Mar. Sci., September 1, 2007; 64(6): 1161 - 1172. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||