Skip Navigation


ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on June 26, 2007
ICES Journal of Marine Science: Journal du Conseil 2007 64(6):1182-1190; doi:10.1093/icesjms/fsm088
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
64/6/1182    most recent
fsm088v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Gudmundsdottir, A.
Right arrow Articles by Sveinbjörnsson, S.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Gudmundsdottir, A.
Right arrow Articles by Sveinbjörnsson, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 2007 International Council for the Exploration of the Sea. Published by Oxford Journals. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Estimating year-class strength of Icelandic summer-spawning herring on the basis of two survey methods

Asta Gudmundsdottir, Gudmundur J. Oskarsson and Sveinn Sveinbjörnsson

Marine Research Institute, Skúlagata 4, PO Box 1390, 121 Reykjavík, Iceland

Correspondence to A. Gudmundsdottir: tel: +354 575 2000; fax: +354 575 2001; e-mail: asta{at}hafro.is

Gudmundsdottir, A., Oskarsson, G. J., and Sveinbjörnsson, S. 2007. Estimating year-class strength of Icelandic summer-spawning herring on the basis of two survey methods. – ICES Journal of Marine Science, 64: 1182–1190.

Recruitment indices at age 1 and 2 were constructed for Icelandic summer-spawning herring using data from two inherently different survey methods on their nursery areas for the years 1988–2003. The surveys were a shrimp, bottom-trawl survey and a herring acoustic survey. Indices were compared with the year-class strength at age 2 derived from an assessment model (AMCI) used for the stock. A juvenile index at age 1 from the acoustic survey and at age 2 from the shrimp-trawl survey correlated significantly with the modelled recruitment, although only marginally with each other. One index was not considered markedly better than the other, and both have shortcomings even if the acoustic survey is believed better than the trawl survey for quantifying juvenile herring. The recruitment index at age 1 derived from the acoustic survey can be improved as an indicator of recruitment if that survey were to be extended to cover the nursery areas off western and southeastern Iceland, and the same areas covered each year.

Keywords: acoustic assessment survey, herring, juvenile, recruitment

Received 5 July 2006; accepted 10 April 2007; advance access publication 26 June 2007.


   Introduction
Top
 Introduction
Material and methods
 Results
 Juvenile herring survey (SHJ)
 Discussion
 References
 
Forecasting stock sizes of fish under different exploitation rates requires an estimation of the size of the immature, growing year classes that will enter the fishing stock in future. Therefore, early information on year-class strength is critical to fisheries managers. Survey-based indices of recruitment at different ages are used in stock assessments for many stocks, e.g. in the tuning procedure of virtual population analyses (VPA) and other similar analytical methods, or to predict the development of stock size in coming years. However, for the management of Icelandic summer-spawning herring (Clupea harengus), there is no valid index of recruitment. Moreover, recruitment has traditionally only been determined from VPA at ages 2 or 3 (Jakobsson et al., 1993; Oskarsson and Taggart, 2005), after those year classes enter the fishable stock.

Spawning mainly takes place off the southern and southwestern Iceland (Fridriksson and Timmermann, 1950) from the middle of July to August (Einarsson, 1956; Oskarsson, 2005). The newly hatched larvae drift with the coastal current clockwise to their main nursery grounds in coastal waters off north and northwest Iceland, where the recruits aged 1 or 2 are generally to be found (Jakobsson and Stefánsson, 1999; Figure 1), although they are also found in other areas in certain years.


Figure 1
View larger version (36K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. The names of some fjords and banks around Iceland referred to in the text. Grey shading indicates the nursery areas, and stripes the spawning areas, and the arrows show the directions of larval drift (redrawn from Jakobsson and Stefánsson, 1999).

 
A usable recruitment index for the herring stock could possibly be constructed from two research surveys conducted on the stock's main nursery grounds and adjacent areas. These surveys are the annual, autumn coastal survey (SMG) for estimating the stock size of shrimp (Pandalus borealis) off north and northwest Iceland, and an acoustic assessment survey around Iceland aimed at assessing the quantity and distribution of herring, including juveniles and mature fish. Juvenile herring, as well as juveniles of other fish species, are common as a bycatch in the shrimp trawl in the SMG survey areas (Pálsson and Thorsteinsson, 1985). The results from the acoustic assessment surveys have not been used directly to construct an annual recruitment index for herring, but have been used occasionally for stock-size prognosis within certain assessment years. The aim of this work is to determine whether a reliable index of recruitment of the herring stock can be constructed from the shrimp-survey results using the bottom-trawl bycatches, along with the results of the acoustic survey of juvenile herring. A comparison is made of the resulting indices and also between each method and the recruitment at age 2 determined from the assessment model AMCI (ICES, 2004; Skagen, 2004).


   Material and methods
Top
 Introduction
 Material and methods
Results
 Juvenile herring survey (SHJ)
 Discussion
 References
 
A shrimp survey (SMG) has been conducted between September and November each year during the period 1988–2003 off north and northwest Iceland from research ships of the Marine Research Institute (MRI) of Iceland (Skúladóttir, 2001). The fjords surveyed annually are Arnarfjordur, Isafjordur, Hunafloi, Skagafjordur, Skjalfandi, and Axarfjordur (Figure 1). A standard shrimp trawl with diamond-mesh codend and a mesh of ~36 mm is used for the survey and the towing speed (2 knots) and station positions are standardized (Skúladóttir and Bragason, 1995). The horizontal and vertical openings of the trawl are 14.9 m and ~4.7 m, respectively.

The number of herring at each station (Ns) has always been recorded, but length has not always been measured (Table 1). Since 2000, length measurements have been standard procedure in Isafjordur and Arnarfjordur (Figure 1), but limited opportunity to collect length measurements, especially in the eastern areas, made it necessary before 2000 to restrict the collection to four areas, Arnarfjordur, Isafjordur, Hunafloi, and Axarfjordur. In those areas, measurements of total length were made on just 117 of the total of 1296 hauls containing herring during the period 1988–1999. From 2000 on, length distributions were collected in 471 of the total of 476 hauls containing herring. The number of fish measured in each of the 588 hauls analysed between 1988 and 2003 varied from 1 to 749 (mean = 46.1, s.d. = 73.1). In the absence of length distributions in some years before 1995, we used length measurements from other surveys to the same areas utilizing the same fishing gear towed at the same speed, but carried out ~1 month later (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1. Number of hauls, number of hauls with herring, number of herring counted, and number of length measurements from all hauls in the years 1988–2003 in a standardized shrimp survey (SMG) in (top panel) Arnarfjordur and Isafjordur, and (bottom panel) Hunafloi and Axarfjordur (see Figure 1).

 
The juvenile herring survey (SHJ) is an acoustic survey aimed at assessing the numbers of herring 1 and 2 years old inside the fjords of northern Iceland (northwest to northeast). It has taken place on research vessels from November to December in most years from 1980 to 2003, but information in the form of informal survey reports is only accessible from 1987. As the SMG began in 1988, we use the SHJ data from 1988 only. To identify acoustic recordings, biological samples were collected with a small Harstad pelagic trawl with ~40 mm mesh in the codend. From 1987 to 1999, the SHJ was conducted as a part of a survey aimed at assessing the adult, fishable, herring stock. From 2000 on, the autumn surveys on young herring have been conducted jointly with a survey on young capelin (Mallotus villosus). The area covered in the surveys has varied from year to year (Table 2). In some years, young herring were also found away from the traditional nursery grounds (Jakobsson and Stefánsson, 1999; Figure 1), during surveys of the fishable stock. In those cases, the numbers of herring aged 1 year were included in the indices as a part of the SHJ.


View this table:
[in this window]
[in a new window]

  		Table 2. Numbers in millions of (top panel) 1-year-old, and (bottom panel) 2-year-old Icelandic, summer-spawning herring from the acoustic juvenile surveys at different locations.

 
The SHJ data comprise length frequency distributions by area, where both the total area covered and the SA-values (area back-scattering strength) have been taken into account in the traditional way (MacLennan and Simmonds, 1992). The output data give the estimated number at length for each of the survey areas. The results for the years 1988 and 1990 were anomalous, because the results were combined over several fjords (Table 2) in the reports. Before 1993, the target strength (TS) used in the acoustic assessments was 21.7 logL –75.5 dB (where L is the length of herring), but from 1993, the relationship TS = 20 logL –72 dB was used (Jakobsson et al., 1993). Here, we make a correction for the difference in TS values.

There are no age readings of material collected from samples taken on either the SMG or the SHJ. The geographical distribution of age 1 herring and the fishable stock does not coincide, and age 2 catches by the fishery are negligible, meaning that an age–length key from the fishery cannot be applied to the survey data. Age is in years in all analyses and results here.

To make age-disaggregated indices from the SMG data, the length distributions were investigated graphically for each year and area to determine the length that separated age classes (a), and the proportion (Pa) in terms of number-at-age of all herring measured. Owing to limitations of the data in most years, all available fish length measurements were pooled from each area, assuming that the length distribution was representative for all hauls taken in the corresponding area that year. An index of age-class strength (ISMG,a) for each year was calculated as follows:


Formula 088M1

(1)
where k is the number of areas, ni the number of hauls in the ith area, Ai the size of the ith area (km2), and Sj the area swept (km2) during each haul.

Another juvenile index, ISMG,all2, was constructed by assuming that all fish counted were aged 2 years. It was calculated by setting Pi,2 = 1 and Pi,1 = 0 in Equation (1). Then data from all six fjords could be used.

When the index was calculated as in Equation (1), it was assumed that the length distribution was the same throughout an area in a given year. As the length of herring was recorded from each SMG haul made in Isafjordur and Arnarfjordur since 2000, with the number of fish measured varying from 1 to 332 (mean = 38.4, s.d. = 58.5), it was possible to test the assumption that the length distribution was invariant over an area. The index was calculated as follows:


Formula 088M2

(2)

Traditionally in SHJ data processing, the age is determined via fixed lengths over the years, where herring < 12.5 cm are assigned to age 1, herring measuring 12.5–19.5 cm are assigned age 2, and herring > 19.5 cm as age 3 and older. As the interest lies in ages 1 and 2, the indices derived from this fixed-length separation is referred to as ISHJ,a,fixed. The age of the juveniles in the stock can also be determined by examining graphically the length distribution, as for the SMG. The derived indices are referred to as ISHJ,a,graph and calculated for all available areas. The results from these methods of determining the indices-at-age were compared with a paired t-test.


   Results
Top
 Introduction
 Material and methods
 Results
Juvenile herring survey (SHJ)
 Discussion
 References
 
Shrimp survey (SMG)
Graphic examination of the length distributions suggests that the boundary between ages 1 and 2 herring varied from ~10.3 to 13.3 cm (mean = 12.1, s.d. = 0.65) and between ages 2 and 3 herring from ~17.5 to 20.3 cm (mean = 20.2, s.d. = 0.51). In some cases, for example in 1994 in Isafjordur, where the length distribution peaked at 12 cm (Figure 2), it was uncertain whether length groups belonged to age 1 or to age 2. In that case, the herring were taken to be age 1, because the length distribution in Arnarfjordur peaked at 11 cm the same year, and a year later age 2 herring were numerous in Isafjordur. Generally, the number at age 1 was less than the number at age 2.


Figure 2
View larger version (16K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 2. Length distributions of Icelandic summer-spawning herring in the shrimp bottom-trawl survey (SMG) in Isafjordur (Figure 1) from September to November of 1988–2003 (excluding 1992 and 1998).

 
Juvenile indices for age 1 (ISMG,1) and age 2 (ISMG,2) were calculated [Equation (1), Figure 3a] for the four areas combined (Arnarfjordur, Isafjordur, Hunafloi, and Axarfjordur). When index ISMG,2 was compared with the AMCI numbers-at-age 2 (AMCI2), it explained 43% of the variation. Most year classes were distributed on the right-hand side of the graph (Figure 3b), but the 1992 year class on the lower left-hand side obviously determines the slope of the regression line. Although most of the herring counted in 1994 were recorded in Hunafloi, they did not contribute to the 1992 year-class index because no length measurements were available from that area (Table 1). When the 1992 year class was excluded from the regression, the index explained ~46% of the variation (p = 0.014). Juvenile indices were also calculated for Arnarfjordur and Isafjordur alone, which contributed most of the length measurements. The index at age 2 explained 42% of the variation in AMCI2. Both indices resulted in only one exceptionally large year class (that of 1999). For the indices calculated for the different areas, there was neither a significant relationship between juvenile indices ISMG,1 and ISMG,2 (log10–log10 transformed; p = 0.32 and p = 0.10) nor between the juvenile index ISMG,1 and AMCI2 (log10–log10 transformed; p = 0.98 and p = 0.11).


Figure 3
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 3. (a) SMG juvenile index at age 2 (ISMG,2) by year class from Arnarfjordur, Isafjordur, Hunafloi, and Axarfjordur, and (b) the relationship between the log of the index (x) and the log of the AMCI-based recruitment index of herring aged 2 years (y) for Icelandic summer-spawning herring for the year classes 1986–1999 (excluding the 1994 year class). The dotted lines indicate the 95% confidence intervals of the model, and the broken line reflects the relationship when the 1992 year class is excluded.

 
A juvenile index, ISMG,all2 [Equation (1), with Pi,a = 1 for a = 2], was calculated for the same four areas as index ISMG,2, as well as for all six areas combined (Arnarfjordur, Isafjordur, Hunafloi, Skagafjordur, Skjalfandi, and Axarfjordur), assuming that all fish counted were aged 2 years. The indices ISMG,2 and ISMG,all2 correlated both for the four areas (log10–log10 transformed; r = 0.795, n = 13, p = 0.001) and for the six areas (log10–log10 transformed; r = 0.748, n = 13, p = 0.003). The index ISMG,all2 for the six areas indicated that the two year classes 1989 and 1999 were exceptionally large (Figure 4a), and the index was significantly related to AMCI2, with r2 = 0.62 (Figure 4b). Always above in the analysis of the relationship between the indices ISMG,all2 and ISMG,2 and AMCI2, the intercept was significantly different from zero (p < 0.05).


Figure 4
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 4. (a) SMG juvenile index (ISMG,all2) for herring from Arnarfjordur, Isafjordur, Hunafloi, Skagafjordur, Skjalfandi, and Axarfjordur, assuming that all were aged 2 years, and (b) the relationship between the log of the index (x) and the log of the AMCI-based recruitment index of herring aged 2 years (y) for Icelandic summer-spawning herring for year classes 1986–1999. The dotted lines indicate the 95% confidence interval of the model.

 
The index at age 2, JSMG,2 [Equation (2)], was computed for Isafjordur and Arnarfjordur separately for the period 2000–2003. For this index, the length distribution from each tow was used. Unfortunately the time-series is too short to compare with the result from the assessment model, but the indices JSMG,2 were compared with the corresponding indices ISMG,2 [Equation (1); Figure 5] for the same areas and years. In ISMG,2, it is assumed that the same length distribution existed for every tow in each area. The difference between the two indices JSMG,2 and ISMG,2 was primarily related to higher estimates of ISMG,2. Index ISMG,2 was 66% and 49% higher than JSMG,2 in Isafjordur in 2000 and 2001, respectively, and 65% higher than JSMG,2 in Arnarfjordur in 2001. This comparison indicates that the greatest difference in both areas was in 2001, a year when a spatial age distribution could be detected in both areas.


Figure 5
View larger version (12K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 5. Juvenile indices of Icelandic summer-spawning herring aged 2 years in Isafjordur and Arnarfjordur from the SMG, where the indices were calculated from a pooled length (age) distribution over all hauls (black columns) or from each haul (grey columns).

 

   Juvenile herring survey (SHJ)
Top
 Introduction
 Material and methods
 Results
 Juvenile herring survey (SHJ)
Discussion
 References
 
The lowest peaks in herring length distributions in the SHJ in the period 1988–2003 ranged from 7 to 13 cm (Figure 6), which is thought to correspond to age 1. The peak with the smallest length (7 cm) was from 1992, attributable to the 1991 year class. Usually it was easy to determine, by graphically examining the length distributions, whether the herring belonged to age 1 or age 2, but there were cases where further consideration was needed. In Faxafloi, the length distribution peaked at 12 cm in 1990, whereas 1-year-olds peaked at 10 cm in Isafjordur the same year (Figure 7). That group also spanned a relatively broad length range, from 9.5 to 14 cm. Elevated sea temperatures speed up the growth rate of herring larvae (McGurk, 1984; Heath et al., 1997). Therefore, considering the general decrease in temperature in a clockwise direction around Iceland starting from the southeast (around Stokksnes; Stefánsson, 1962) and the high zooplankton biomass in 1990 off southern Iceland (Marine Research Institute, 2004), the herring in Faxafloi probably grew faster than the herring in Isafjordur. Therefore, it was decided to assign age 1 to all herring < 15 cm in Faxafloi in 1990. This resulted in values of 387 million 1-year-olds and 12 million 2-year-olds, though the values would have been 254 million 1-year-olds and 145 million 2-year-olds if the length 12.5 cm had been used to distinguish between year classes. Similar problems arose concerning the 1993 and 1994 data in Eyjafjordur and the 2001 data from southeastern Iceland. They were all resolved by comparing the length data with length distributions from other areas and taking into consideration both the sea temperature in each area and zooplankton indices around the coast of Iceland (Marine Research Institute, 2004).


Figure 6
View larger version (15K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 6. Length distributions of Icelandic summer-spawning herring from all areas in the acoustic survey of juvenile herring (SHJ) in November of the years 1988–2003. (No surveys were carried out in 1989, 1991, 1998, or 2002.)

 


Figure 7
View larger version (10K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 7. Length distributions of Icelandic summer-spawning herring in Faxafloi and Isafjordur in 1990 during the acoustic survey of juvenile herring (SHJ).

 
The number of herring aged 1 was generally higher than the number of herring aged 2, and herring aged 1 were almost twice as often recorded as herring aged 2 (Table 2). Large numbers of age 2 herring seemed to be recorded only if very high numbers of age 1 herring had been observed the year before, for example 657 million 2-year-olds recorded in 2001 in Skjalfandi and 1450 million 1-year-olds the year before. Age 2 herring were poorly, if at all, represented in the SHJ samples (Table 2), so they were omitted from further analyses.

The difference between the two indices at age 1, ISHJ,1,fixed and ISHJ,1,graph, was only minor in most years and not significant (p = 0.61), and the indices correlated strongly (log10–log10 transformed, r = 0.945, n = 12, p < 0.001). On the basis of these results, further analyses were restricted to the index ISHJ,1,graph, by graphically examining the length distributions (Table 2) and thereby taking into account the seeming spatial variation in herring growth patterns (Figure 7).

Two different spatially restricted indices of numbers-at-age 1 (ISHJ,1,graph) were calculated. The index at age 1 for west, north, and southeast Iceland (i.e. all available data) gives a single exceptionally big year class, 1999, and two big ones, those of 1991 and 1996. The index explains ~52% of the variation when compared with AMCI2 (Figure 8b). The 1993 year class is the smallest in both the SHJ and the AMCI assessment, and clearly strongly influenced the slope of the regression line. In 1994, only northwest and north Iceland were surveyed, and only a small number of the 1993 year class (at age 1) was measured (Table 2), suggesting that this particular year class probably also largely stayed outside the areas surveyed. As this year class was such an outlier, it was considered justifiable to exclude it from the regression. When this was done, however, the index still explained a similar amount of the variation, ~55% (p = 0.02). Limiting the geographical range either to northwest and north Iceland or to Eyjafjordur, the index at age 1 was not significantly related to AMCI2 (p = 0.054 and p = 0.43, respectively). The relationship between index ISMG,all2 (all fish counted at age 2) and index ISHJ,1,graph from all areas of the SHJ survey was only marginally significant (r2 = 0.48, p = 0.055).


Figure 8
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 8. (a) The SHJ juvenile herring (age 1) index from west, north, and southeast Iceland, and (b) the relationship between the log of the index of herring aged 1 year (x) and the log of the AMCI-based recruitment index of herring aged 2 years (y) for Icelandic summer-spawning herring for the year classes 1986–1999 (excluding year classes 1986, 1988, 1990, and 1997). The dotted lines indicate the 95% confidence intervals of the model, and the broken line indicates the relationship when the 1993 year class is excluded.

 

   Discussion
Top
 Introduction
 Material and methods
 Results
 Juvenile herring survey (SHJ)
 Discussion
References
 
We have shown that two indices can be used as a juvenile index for Icelandic summer-spawning herring because they explain most of the variation in AMCI numbers-at-age 2 during the years 1988–2003. One originates from the autumn shrimp survey and is an index of numbers-at-age 2 on the assumption that all herring counted are aged 2 years, and the other stems from the herring juvenile survey and is an index of age 1 herring.

The data strongly suggest that the catchability of herring aged 1 year in shrimp trawls is less than that of herring aged 2 years. Generally, herring aged 1 year yielded less or the same number of herring aged 2 years a year later despite the likely high rate of mortality (Figure 2). This pattern was particularly evident when there were large year classes. Moreover, in areas where herring aged 2 years dominated the SMG results and herring aged 1 year were in small numbers, the SHJ survey to the same area 1–2 months later found a lot of herring aged 1 (e.g. Isafjordur 2003). The reason for the lesser catchability of herring aged 1 year in the SMG is not known, but we suggest that it might be related to trawl selection, 1-group herring escaping through the meshes of the trawl or escaping the trawl altogether proportionally more than herring aged 2 years. Different depth preferences could also be significant. Herring aged 2 years seem to stay closer to the bottom (Cardinale et al., 2003) than their younger counterparts and may therefore be more vulnerable to the shrimp trawl than the pelagic trawl used in the SHJ survey. Another possible reason for the difference in catch rates is that in the time between the SMG and the SHJ, the 2-year-old herring may, in some years, already have migrated out of the fjords to join the older component of the stock. These juvenile 2-year-old herring encountered at various sites in autumn, e.g. Latragrunn, Kogurgrunn, and Eldeyjarbanki, are usually still segregated from the mature stock.

There could, of course, be a systematic error in the SMG index of herring aged 2 years, all fish being assumed to be of that age. When all herring are generally assumed to be aged 2 years when a large year class appears aged 1 year, more herring aged 1 year will be assumed to be aged 2 than in years where there is a poor year class at age 1 year. The consequence of such misinterpretation could lead to overestimation of a year class in the year before the formation of a large year class.

A sufficient number of length measurements is critical in constructing the SMG juvenile indices. The decision to measure herring from every haul in the shrimp survey from the year 2000 onwards made a valuable addition to the database. Comparing the index for 2000–2003 in Isafjordur and Arnarfjordur obtained by pooling all data for length (age) distribution [Equation (1)] with that made using length (age) distribution from each tow [Equation (2)] showed a considerable difference in three cases (Figure 5). This implies that the indices should be based on length (age) distribution determined from each tow separately in future. Limiting the length measurements decreases the quality of the SMG indices. For example, the total lack of length distributions in some areas in the early and middle 1990s in SMG (Table 1) was responsible for the absence of those areas in the indices, even though herring were seen there in great numbers, as in 1994 in Hunafloi.

The SHJ index of herring aged 1 year covering all areas was significantly related to the AMCI index of herring aged 2, whether or not the 1993 year class was included in the analysis. The coefficient of determination (r2 = 0.55) of this relationship was slightly less than that for the index of herring aged 2 years from the SMG (r2 = 0.62). The main problem with the SHJ index of herring aged 1 year is thought to originate from interannual differences in the areas covered during the surveys, because herring aged 1 are often distributed outside the main nursery areas. Some 31% of the total index during the observation period was made up of recordings taken outside the main nursery areas, even though those areas were not so frequently surveyed. Indeed, the index outside the main nursery areas has been as high as 98% of the total in a single year. This spatial distribution of juvenile herring in the SHJ results (Table 2) indicates that the main nursery grounds of Icelandic summer-spawning herring are more extensive than implied by Jakobsson and Stefánsson (1999) and shown in Figure 1. However, we cannot reach a firm conclusion about the locations of nursery areas on the basis of the data available to us at present.

Comparisons of juvenile-herring indices derived from the SMG and the SHJ within a certain area or fjord were not possible because of gaps in the time-series. For example, Eyjafjordur, the most sampled fjord in the SHJ, is not covered in the SMG, and the fjords mainly sampled in the SMG are only occasionally covered in the SHJ.

The results imply that both survey methods, with standardized trawl stations and an acoustic assessment, can be used to provide abundance indices of young herring despite the two survey methodologies being inherently very different. Judging from a literature search, determination of the year-class strength of juvenile herring is most commonly done through acoustic measurements. For several herring stocks, both types of surveys are applied to the assessments. Examples are Norwegian spring-spawning herring (ICES, 2004), and North Sea herring, herring in Division IIIA and Subdivisions 22–24, the Celtic Sea and Division VIIJ herring (ICES, 2003b). However acoustic surveys alone are used for others (e.g. Baltic herring; ICES, 2003a). Comparison of the acoustic and bottom-trawl survey indices for North Sea herring indicates that both can be used to estimate the abundance of young herring (Simmonds, 2003), just as with the data reviewed in this paper.

However, certain factors need to be taken into account. The spatial coverage of juvenile surveys is important when considering their merit. According to the SMG and SHJ samples, juvenile herring may not be distributed uniformly within each fjord, and there are differences in the abundance of juvenile herring between fjords over time: limiting the index to the very well sampled Eyjafjordur in SHJ gave an inaccurate overview. Considering the distribution within and between fjords, both sets of survey results are suggested as useful indices of juvenile herring abundance because they cover a good portion of the nursery areas every year. However, as the SMG is targeted at shrimp distribution and does not cover the whole nursery area of herring, it must be considered less accurate than the SHJ in terms of forecasting herring recruitment. To obtain a better index from the SHJ, the survey design would need to be improved, perhaps by surveying every year the areas of Faxafloi, Breidafjordur, Arnarfjordur, Isafjordur, Hunafloi, Skagafjordur, Eyjafjordur, Skjalfandi, Axarfjordur, and Myrarbugur.


   Acknowledgements
 
We thank Páll Reynisson for making available all the survey reports covering herring, Guðmundur Skúli Bragason and Unnur Skúladóttir for providing information about the shrimp survey, and Sigfús Jóhannesson for facilitating our access to SMG data. We also thank Hrafnkell Eiríksson and two anonymous referees for suggestions on how to improve the manuscript.


   References
Top
 Introduction
 Material and methods
 Results
 Juvenile herring survey (SHJ)
 Discussion
 References
 

    Cardinale M., Casini M., Arrhenius F., Hakansson N. Diel spatial distribution and feeding activity of herring (Clupea harengus) and sprat (Sprattus sprattus) in the Baltic Sea. Aquatic Living Resources (2003) 16:283–292.[CrossRef][Web of Science]

    Einarsson H. Frequency and distribution of post-larval stages of herring (Clupea harengus L.) in Icelandic waters. Rit Fiskideildar (1956) 2:33.

    Fridriksson Á., Timmermann G. Rannsóknir á hrygningarstöðvum vorgotssíldar vorið 1950. Rit Fiskideildar (1950) 1:32.

    Heath M., Scott B., Bryant A. D. Modelling the growth of herring from four different stocks in the North Sea. Journal of Sea Research (1997) 38:413–436.[CrossRef]

    ICES. Report of the Baltic Fisheries Assessment Working Group. (2003a) 522. ICES Document CM 2003/ACFM: 21.

    ICES. Report of the Herring Assessment Working Group for the Area South of 62°N. (2003b) 438. ICES Document CM 2003/ACFM: 17.

    ICES. Report of the Northern Pelagic and Blue Whiting Fisheries Working Group. (2004) ICES Document CM 2004/ACFM: 24.

    Jakobsson J., Gudmundsdóttir Á., Stefánsson G. Stock-related changes in biological parameters of the Icelandic summer-spawning herring. Fisheries Oceanography (1993) 2:260–277.

    Jakobsson J., Stefánsson G. Management of summer-spawning herring off Iceland. ICES Journal of Marine Science (1999) 56:827–833.[Abstract/Free Full Text]

    MacLennan D. N., Simmonds E. J. Fisheries Acoustics (1992) London: Chapman and Hall. 325.

    Marine Research Institute. Environmental conditions in Icelandic waters 2003. (2004) Iceland: Marine Research Institute. 101. Fjölrit no.

    McGurk M. D. Effects of delayed feeding and temperature on the age of irreversible starvation and on the rates of growth and mortality of Pacific herring larvae. Marine Biology (1984) 84:13–26.[CrossRef]

    Oskarsson G. J. Pre-spawning factors and recruitment variation in Atlantic herring (Clupeidae; Clupea harengus, L.): a comparative approach. In: PhD thesis (2005) Halifax, NS, Canada: Oceanography Department, Dalhousie University. 251.

    Oskarsson G. J., Taggart C. T. When it comes to recruitment variation in Icelandic herring, it is the eggs of the repeat spawners that make the difference. (2005) 18. ICES Document CM 2005/Q: 22.

    Pálsson Ó. K., Thorsteinsson G. The management of juvenile fish by-catch in an Icelandic shrimp fishery. (1985) 19. ICES Document CM 1985/K: 47.

    Simmonds E. J. Weighting of acoustic- and trawl-surveys indices for the assessment of North Sea herring. ICES Journal of Marine Science (2003) 60:463–471.[Abstract/Free Full Text]

    Skagen D. AMCI version 2.4. Assessment model combining information from various sources. Model description: Instructions for installation and running File formats. (2004).

    Skúladóttir U. (2001) 56:34–37. Stofnmæling rækju [Stock size estimation of shrimp]. Hafrannsóknir.

    Skúladóttir U., Bragason G. S. The occurrence of 0-group cod and haddock as by-catch in the inshore shrimp surveys in Icelandic waters 1978–1994. (1995) 25. ICES Document CM 1995/G: 7.

    Stefánsson U. North Icelandic waters. Rit Fiskideildar (1962) 3:269.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 ICES J. Mar. Sci.Home page
G. J. Oskarsson, A. Gudmundsdottir, and T. Sigurdsson
Variation in spatial distribution and migration of Icelandic summer-spawning herring
ICES J. Mar. Sci., September 1, 2009; 66(8): 1762 - 1767.
[Abstract] [Full Text] [PDF]

Home page
 ICES J. Mar. Sci.Home page
J. F. Schweigert, D. E. Hay, T. W. Therriault, M. Thompson, and C. W. Haegele
Recruitment forecasting using indices of young-of-the-year Pacific herring (Clupea pallasi) abundance in the Strait of Georgia (BC)
ICES J. Mar. Sci., September 1, 2009; 66(8): 1681 - 1687.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
64/6/1182    most recent
fsm088v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Gudmundsdottir, A.
Right arrow Articles by Sveinbjörnsson, S.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Gudmundsdottir, A.
Right arrow Articles by Sveinbjörnsson, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?