ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on July 24, 2007
ICES Journal of Marine Science: Journal du Conseil 2007 64(6):1202-1209; doi:10.1093/icesjms/fsm110
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Diel variations in acoustic recordings of blue whiting (Micromesistius poutassou)
Institute of Marine Research, PO Box 1870 Nordnes, N-5817 Bergen, Norway
Correspondence to E. Johnsen: tel: +47 55 235355; fax: +47 55 238579; e-mail: espen.johnsen{at}imr.no
Johnsen, E., and Godø, O. R. 2007. Diel variations in acoustic recordings of blue whiting (Micromesistius poutassou). – ICES Journal of Marine Science, 64: 1202–1209.Annual landings of blue whiting (Micromesistius poutassou) in the northeast Atlantic have exceeded 2 million metric tonnes in recent years, and overexploitation is an increasing concern in terms of the sustainability of the fishery. The most important fisheries-independent dataset used for tuning the analytical stock assessment comes from the Norwegian surveys of blue whiting west off the British Isles. The survey is carried out in March/April during peak spawning, and improving its quality will have a direct positive effect on stock assessment. Here, we analyse diel effects on the abundance and vertical distribution as recorded by acoustics in 1995 and 1996 and from 1998 to 2002, and evaluate the potential effects on the survey estimates. On average, the acoustic density of blue whiting was
20% higher by day than by night. However, the diel bias varied considerably among years, and surprisingly, the acoustic density in shallow water (<350 m) was in general highest at night, when the blue whiting were distributed higher in the water column and more dispersed. The span in the vertical depth range increased considerably with bottom depth in water shallower than 550 m. In deeper water, where blue whiting had little or no bottom association, the day–night differences in vertical distribution were smaller and not affected by bottom depth. The inconsistency of the diel effect from year to year negatively affects the time-series used during annual stock assessments.
Keywords: acoustics, blue whiting, diel variation, spawning aggregation, vertical migration
Received 12 October 2006; accepted 12 June 2007; advance access publication 24 July 2007.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
Blue whiting (Micromesistius poutassou) is distributed in the Northeast Atlantic from the Mediterranean to the Barents Sea, and in the west from East Greenland to the Kola Peninsula (Bailey, 1982; Monstad, 2004). The spawning stock west of the British Isles has a southern component spawning on the Porcupine Bank and along its southern slope, and a northern component spawning west of the British Isles along the slope south to the Porcupine Bank (Monstad, 2004). Spawning is from January to May in depths of 300–600 m, with most spawning activity between 300 and 400 m deep (Bailey, 1982; Monstad, 2004). Since the 1970s, Norwegian research vessels have carried out annual acoustic surveys during peak spawning in March/April, when most of the northern stock is distributed in the known main spawning areas (Monstad, 2004). The resulting stock estimates comprise the longest time-series covering a significant part of the stock, and is therefore important for tuning the analytical assessment (ICES, 2006). In addition to the extensive horizontal migration, the vertical distribution of blue whiting has large geographical and seasonal variations (Blindheim and Jakupsstovu, 1976; Monstad, 2004). Diel variation in the vertical distribution has also been reported (Stensholt et al., 2002; Monstad, 2004). Such diel vertical migrations are common for many fish species (Neilson and Perry, 1990), and can bias the acoustic-area backscattering of both pelagic and demersal species (Huse and Korneliussen, 2000; Hjellvik et al., 2004).
Generally, a typical study on diel variation in vertical distribution and acoustic density is carried out in situ in a restricted geographical location (Engås and Soldal, 1992; Fréon et al., 1996; Michalsen et al., 1996; Aglen et al., 1999; Neilson et al., 2003). However, such an experimental design may be seriously affected by horizontal migration of fish, and further, the observations may only be valid for the location examined. Diel effects may also be studied by a more general statistical approach analysing large datasets from stock-assessment surveys for systematic day–night differences (Fréon et al., 1993; Hjellvik et al., 2001).
The main objective of our study was to examine diel variation in the historical assessment survey data of Norwegian blue whiting, using the statistical techniques developed by Hjellvik et al. (2001). Our aim was to describe and quantify day–night variation in the vertical distribution and acoustic density of blue whiting.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Data
We used acoustic data from the Norwegian assessment surveys for blue whiting west of the British Isles (Figure 1) in 1995, 1996, and 1998–2002. No survey was conducted in 1997, and 1995 was the first survey of blue whiting with a drop-keel installed. All surveys were made in March/April using a zigzag survey design, and the data were collected with a 38 kHz SIMRAD EK 500 echosounder (Bodholt et al., 1988). On a daily basis during the surveys, the acoustic recordings were interpreted using the Bergen echo integrator (BEI) (Knudsen, 1990; Foote et al., 1991). The echograms were scrutinized mile-by-mile, and the recordings were allocated to species or groups of species based on the characteristics of the recordings and the catch composition of the trawl samples. The identification of blue whiting during these surveys is relatively uncomplicated because few other fish species are present in significant quantities in the core spawning grounds. Small mesopelagic species are found in low densities throughout the area and may, in some cases, confuse interpretation when the abundance of blue whiting is low. Values by category of the nautical-area scattering coefficient [m2 nautical mile–2] (NASC) (Maclennan et al., 2002) were stored in a database with a horizontal resolution of 1 nautical mile and a vertical resolution of 10 m, referenced to the surface. Only data categorized as blue whiting were used in this study. The data from the bottom layer (0–10 m above the seabed) were stored with a vertical resolution of 1 m, referenced to the seabed. Hereafter, NASC refers to the sum of the nautical-area scattering coefficient of the whole water column by nautical mile, and S_NASC and B_NASC to the coefficients in the surface- and bottom-referenced depth layers, respectively, by nautical mile.
|
Depths below 800 m contributed less then 0.01% of the total blue whiting NASC and were excluded from the analyses. Further, because blue whiting were rarely observed in shallow water, all records collected from areas where the bottom depth was shallower than 200 m were excluded. About 1% of the total blue whiting NASC was excluded in this manner. To investigate the vertical density profile independently of bottom depth, the vertical depth distribution of the S_NASC values were normalized by dividing the depth from the surface to the bottom, or at bottom depths > 800 m into 40 equally spaced intervals (Figure 2).
|
Zero and large NASC values
The treatment of zero observations is of importance, because too many zero values in a stratum may obscure the underlying diel variation pattern (Hjellvik et al., 2002). Therefore, the influence of zero values on the diel models was tested by conducting three runs: (i) excluding all zero NASC values, (ii) excluding all strata with >50% zero values, and (iii) excluding all strata with > 90% zero values. Additionally, large values may dominate the distribution statistics and thus obscure any underlying diel pattern in the observations, so to test the influence of high values, the 1%, 2%, and 5% highest observations were excluded.
Descriptive indicators of vertical density distribution
The 10%, 50%, and 90% quantiles of the vertical density profiles of the S_NASC values by logged nautical mile were calculated for the absolute and normalized depth distributions (Figure 2). The vertical density range (VDR) was defined as the difference between the 5% and the 95% quantiles, and the average NASC by depth layer (NDL) was calculated as:
|
| (1) |
Modelling diel variation
We used the statistical methods developed by Hjellvik et al. (2001) to model the diel variation in NASC and their vertical distribution patterns. The NASC data were approximately log-normally distributed, and were log-transformed before further analysis:
|
| (2) |
The total variation in density (Hjellvik et al., 2001) is assumed to be caused by a stratum-to-stratum variation, a superimposed diel variation, and a random noise so that
|
| (3) |
i the random noise component. Hjellvik et al. (2001) estimated the value of µ by day for bottom-trawl catches in the Barents Sea, but for a systematic acoustic survey of blue whiting, it is more appropriate to stratify differently (Johnsen and Iilende, in press). Consequently, the data were stratified by survey and geographical cells of the size of 2° latitude and longitude [e.g. one stratum included data from year 1999, latitude (56°N, 58°N), and longitude (8°E, 10°E)]. The difference in levels by day and night (D) and the diel oscillation can be described by a logistic function (Hjellvik et al., 2001) for g in Equation (3):
|
| (4) |
the swiftness of this transition. Importantly, Equation (4) is a one-sided function that assumes that the sun's altitude at the night–day transition is equal to that at the day–night transition (Hjellvik et al., 2001).
Thus, negative diel amplitude (D) for the 10%, 50%, and 90% quantiles mean that the blue whiting were distributed deeper by day than by night. A positive diel amplitude (D) for the VDR means a larger span in the vertical range at night. The models were fitted to the data using the R-function nls (Venables and Ripley, 2002). In cases with small day–night differences, nls typically fails to converge when the full model [Equation (4)] was used. Keeping and/or ß fixed, an estimate of D was always obtained. Acoustic-survey data are generally serially autocorrelated, complicating the variance estimation (Legendre, 1993), and the potential consequences of such spatial dependence on our results are discussed.
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Zero and large NASC values
In all, some 25 600 nautical miles were logged during the 1995, 1996, and 1998–2002 surveys, and the percentage of nautical miles without any registration of blue whiting varied considerably between surveys; 24% (1995), 45% (1996), 35% (1998), 32% (1999), 23% (2000), 35% (2001), and 22% (2002). The annual differences in the frequency of zero observations were, at least in part, attributable to variable survey tracks in deeper water (Figure 1). The estimated diel variations in NASC were sensitive to the treatment of the zero values (Table 1), so all distances steamed without recording blue whiting were excluded from the analyses. The non-zero NASC values ranged from 0.01 to 133 644, with a skewed distribution in which relatively few observations with high NASC values contributed a large percentage of the total acoustic density (Table 2). In general, exclusion of the 1%, 2%, and 5% highest NASC values made no marked difference in the model-parameter estimates (Table 3), so all non-zero NASC values were included in the final runs.
|
|
|
Bottom depth and distribution
In deep water (bottom depth >600 m), blue whiting were mainly distributed in the water column without close association with the seabed, whereas in shallower water (bottom depth <500 m), the mean distance from the seabed tended to decrease with decreasing bottom depth (Figure 3a). The marked bottom depth dependence was also evident for the relative depth distribution (Figure 3b), suggesting that the bottom depth effect could not be normalized. Therefore, the results are presented in absolute depth only, because interpretation of the results is more easily understood when the vertical density profiles are used in terms of absolute depth. Further, to illustrate and take the marked bottom-depth dependence in vertical density profile into consideration, the parameter estimates of the diel variation model are presented by bottom depth intervals (200–249, 250–299, ... , 950–999,
1000 m).
|
Diel variation in vertical density distribution
In Figure 4, the anomalies of mean-adjusted values of the 10%, 50%, and 90% quantiles and VDR grouped in boxplots by 10° interval are plotted against the altitude of the sun. The diel variation model [Equation (4)] with
= 0.25 and ß = 5 illustrates the general diel variation patterns in the vertical density distribution of blue whiting: the vertical-density profiles changed with the altitude of the sun, and the transitions from levels associated with night distribution to those associated with a day distribution coincide with twilight. The appropriateness of the selected and ß values is demonstrated by the similarity between a depicted non-parametric model and the model of Equation (4). Further, the negative diel amplitude estimates of the 10%, 50%, and 90% quantiles and the positive diel-amplitude estimate of VDR show, respectively, that the blue whiting were distributed deeper and more contracted by day.
|
Blue whiting stayed deeper in all water layers and for all bottom depths by day, but the largest day–night variation was observed for the 10% quantiles (Figure 5). In absolute terms, the diel-amplitude estimates of the 10% and 50% quantiles and VDR increased with bottom depth from 200 to
500 m (Figure 5). For some bottom depth intervals, the full nls model [Equation (4)] failed to converge because of the small day–night differences, so or ß or both were fixed at 0.25 and 5, respectively, because these parameter values seemed to fit well with the non-parametric estimates (Figure 4).
|
Annual diel-amplitude (D) estimates of the 10%, 50%, and 90% quantiles by depth stratum varied considerably. However, it is worth noting that the three amplitude estimates were generally correlated, so large diel vertical migration in the upper layers coincided with large vertical migration in the deeper layers (Figure 6, right panel). The
s of the 50% and 90% quantiles (r2 = 0.69) were better correlated than the 10% and 50% quantiles (r2 = 0.46) and the 10% and 90% quantiles (r2 = 0.20), reflecting the greater variability in the s of the 10% quantiles (Figure 5).
|
Diel variation in acoustic abundance
When data from all the surveys were included and zero observations excluded, the diel-amplitude estimate of the log-transformed NASC values was 0.18, meaning that the average acoustic density was about 1.2 (e0.18) times higher by day than by night. The amplitude estimates showed a marked bottom depth dependence, and the acoustic density was greatest at night in shallow water (<350 m), whereas the opposite diel variation pattern was evident in deeper water (Figure 7). The diel-amplitude estimate of the log-transformed acoustic density by depth layer (NDL) was 0.40, indicating that the average day level of NDL was about 1.5 (e0.40) times higher than the night level. Diel variation in NDL had a similar depth dependence as NASC values (Figure 7). The NDL and the NASC are correlated [Equation (1)], and the relatively high diel variation in NDL is therefore caused by both diel variation in the NASC and the increased vertical distribution of blue whiting at night. Further, the depth dependence in diel variation of the NASC values varied between surveys (Figure 8). Except for two surveys, the diel amplitude increased with bottom depth, from
200 to 600 m, and thereafter either stabilized or even decreased with depth.
|
|
For all surveys, the mean diel-adjusted NASC value by survey was higher than the unadjusted mean value (Figure 8). However, the difference between the adjusted and unadjusted mean values varied considerably between survey: in 1996, for example, the adjusted mean value was 45% higher than the mean of the unadjusted NASC data, whereas this difference was <1% in 1995 (Figure 8).
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Separation of temporal and spatial variability in acoustic-survey data is not a straightforward exercise because high and low densities of fish can be registered independent of the time of day. Nevertheless, we have shown that a large number of acoustic observations combined with sound statistical methods can reveal diel-variation patterns in such data. In the case of fisheries-acoustics surveys with continuous sampling along a cruise track, the data are close in space and time and are generally autocorrelated (Rivoirard et al., 2000). The spatial structure in the data on blue whiting is not considered in the analyses and, consequently, the standard deviations may be underestimated (Legendre, 1993). On the other hand, neither the expected mean nor the global estimation is affected by such autocorrelation (Legendre, 1993). Although spatial dependence is not considered here, the results shown presumably represent the underlying behavioural and acoustic density patterns. This assumption is also strongly supported by the consistency of the observed depth and annual patterns in the estimated diel variations.
Large day–night variations in school formation are common for pelagic species, and schools generally tend to be less dense and distributed closer to the surface at night (Fréon et al., 1996). The aggregations of blue whiting west of the British Isles show a similar diel variation, with blue whiting generally distributed deeper and more concentrated during daylight. However, the vertical movement we recorded is considerably less than, for example, that in Norwegian spring-spawning herring (Clupea harengus) (Huse and Korneliussen, 2000). Moreover, the data suggest that the nocturnal and daylight behaviour are separated by a relatively long transition period, during which blue whiting gradually migrate vertically. This transition period corresponds to the change in the altitude of the sun and hence the light intensity, which is said to be the most significant regulator for such diel vertical migrations (Neilson and Perry, 1990; Helfman, 1993).
The free vertical range for a fish with a closed swimbladder (physoclist) increases with bottom-depth range (Jones, 1951; Jones and Scholes, 1981), consistent with our observation in shallow water (Figure 5) where diel variation in vertical distribution increased with bottom depth. Interestingly, the day–night differences in vertical distribution dropped markedly when the bottom depth exceeded 550 m. This is about the depth at which blue whiting seem to lose their bottom association (Figure 3). It is our hypothesis that the free vertical range is less important for diel vertical migration when the shoals are distributed off the bottom in deeper water.
In our study, we could not track the horizontal migration with the available data, but the observed diel patterns in vertical distribution in shallow water may emerge from horizontal movements. Fish on the continental shelf may perform oblique migration along the seabed, up the slope in the evening and down it at night, to compensate for the impossibility of performing purely vertical migration. Such a migration pattern will explain the large reduction in NASC during daylight in shallow water. Similarly, the drop in diurnal variation in the NASC and the NDL at bottom depths of 600–650 m may reflect horizontal migration of blue whiting towards the 350–550 m bottom depths. The migration hypothesis could be validated by in situ tracking of the migration speed and direction of individual fish over the diel cycle with advanced acoustic instrumentation. Another anomaly is the diel amplitude estimates between 900 and 950 m (Figure 7). In this case horizontal migration seems to be a less likely explanation. More detailed information from one location throughout a 24 h period is needed to understand this apparent outlier.
The vertical position of blue whiting is related to seawater temperature and food availability (Monstad, 2004). However, the greatest densities in deep water (isobath >550 m) were normally between 300 and 500 m deep in the water column at the same depth as the eggs and milt release of spawning blue whiting (Monstad, 2004). Diel variation in the 10%, 50%, and 90% quantiles, which were relatively stable from 550 to 750 m bottom depth, varied considerably in depths approximately >750 m. Exploratory analyses showed that these estimated diel amplitudes were robust to threshold changes of the NASC values included in modelling, which indicates that the patterns were not solely an artefact of possible species misidentification. The noisy diel variation in the vertical density distributions at great depth (approximately >750 m) is also reflected in the high interannual variability at those bottom depths (Figure 6). It is possible that the relative similarity in the diel patterns in shallow water is a result of the fairly regular coverage of these areas, whereas the coverage of deeper water is less consistent (Figure 1).
The effect of diel variation in the NASC values on the overall acoustic density of blue whiting varied markedly between years. It is feasible to compensate for such year-on-year differences, but the correction would be subject to the uncertainty of several parameters which negatively affect the precision of the survey estimates (Hjellvik et al., 2002). The impact of diel bias is currently ignored in the analytical assessment of this stock, but we recommend that the effect of the bias on assessment estimates should be investigated. Clearly, a diel bias may make an assessment less reliable, so perhaps the results of our study could lead to a redesign of the survey so that depth-dependent diel variation is considered. To improve survey design, a clear understanding of the causes of the observed diel variation in NASC is needed. Fréon et al. (1993) described several behavioural factors that may lead to diel bias in the acoustic density of pelagic fish. Below, we discuss the factors that most plausibly explain some of diel variation.
The bottom dead-zone may cause considerable error in acoustic estimates of demersal and semi-demersal fish, and the approximate dead-zone height for a standard 38 kHz split-beam transducer is 0.7 and 1.0 m on 200 and 350-m depths, respectively (Ona and Mitson, 1996). Where the seabed is steeply sloping, such as in the surveys for blue whiting (Figure 1), the dead-zone increases substantially (Pedersen, 2007). In shallow water, although blue whiting were distributed close to the bottom by day (Figure 5), the distribution of the vertical density profile in the bottom layers did not indicate substantial missing of blue whiting in the dead-zone because the density did not increase towards the seabed. Moreover, blue whiting were rarely registered in layers <100 m, and never in surface layers. Therefore, the observed diel variation in the NASC is not caused by blue whiting being lost in the surface dead-zone. Similarly, it is unlikely that the diel variation we recorded is caused by vessel avoidance, because >99.5% of the total NASC is shallower than 150 m in the water column. However, future experiments are needed to confirm these conclusions.
The acoustic target strength is positively correlated to swimbladder size (Gorska and Ona, 2003). Therefore, the swimbladder of physoclistous fish in general reaches its maximum size when the fish are positioned in the upper part of their vertical range. This may explain the diel variation we observed in echo abundance in shallow water (Figure 7), but it contradicts the positive diel amplitude estimates of the NASC in deeper water. An explanation of this inconsistency could be the dominance of a polarization effect in deeper water, and of a swimbladder effect in shallow water, because the target strength of gadoids changes more with depth and swimbladder size in shallow water (Hazen and Horne, 2003).
Systematic changes in the average tilt angle of fish bias the acoustic density (Hazen and Horne, 2003), because the tilt angle has a large influence on target strength (McClatchie et al., 1996). Typically, schooling fish such as herring have reduced polarization at night (Fréon et al., 1996), causing an increased tilt-angle distribution (Huse and Ona, 1996) and hence a reduction in target strength (McClatchie et al., 1996). Diel variations in the descriptive indicators of the vertical density distribution suggest that blue whiting have a wider vertical range and less dense distribution at night. Assuming that the VDR and acoustic density by depth layer are proxies for tilt-angle distribution, the diel variation of these variables can explain, at least in part, the observed diel variation in NASC values in deeper water. This is also in accord with the findings of Hjellvik et al. (2004) for cod in the Barents Sea. Further, there is an increase of the day–night NASC ratio between 200 and 400 m (Figure 7), which may relate to the greater tilt-angle sensitivity in target strength of larger fish (Hazen and Horne, 2003), because the average size of blue whiting increased with bottom depth in shallow water (Figure 9).
|
Although we have described several possible factors that may explain the diel variation we observed in the acoustic density of blue whiting, more detailed in situ studies need to be carried out to understand the proximate causes for the observed diel patterns. One that provides the detailed acoustic properties of single fish at any depth by day and by night might provide the data required to explain these patterns.
|
Acknowledgements |
|---|
We thank the Norwegian Research Council for its support through the programme "Assessing and Compensating for Uncertainty in combined Trawl and Acoustic Survey" (156251/120). We are also grateful for valuable statistical help from Vidar Hjellvik and Dag Tjøstheim, and to the two reviewers for useful remarks that improved the paper.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Aglen A., Engås A., Huse I., Michalsen K., Stensholt B. K. How vertical fish distribution may affect survey results. ICES Journal of Marine Science (1999) 56:345–360.
Bailey R. S. The population biology of blue whiting in the North Atlantic. Advances in Marine Biology (1982) 19:257–355.[Web of Science]
Blindheim J., Jakupsstovu S. H. Undersøkelser av kolmule og hydrografi i Norskehavet (1976) 2. Fisken, og Havet. 29–41. 1976 12. juli – 11. august.
Bodholt H., Nes H., Solli H. A new echosounder system for fish-abundance estimation and fishery research. (1988) ICES Document CM 1988/B: 11.
Engås A., Soldal A. V. Diel variations in bottom-trawl catches of cod and haddock and their influence on abundance indices. ICES Journal of Marine Science (1992) 49:89–95.
Foote K. G., Knudsen P., Korneliussen R. J., Nordbø P. E., Roang K. Post-processing system for echosounder data. Journal of the Acoustical Society of America (1991) 90:37–47.[CrossRef][Web of Science]
Fréon P., Gerlotto F., Soria M. Diel variability of school structure with special reference to transition periods. ICES Journal of Marine Science (1996) 53:459–464.
Fréon P., Soria M., Mullon C., Gerlotto F. Diurnal variation in fish-density estimate during acoustic surveys in relation to spatial distribution and avoidance reaction. In: Aquatic Living Resources (1993) 6:221–234.[CrossRef]
Gorska N., Ona E. Modelling the acoustic effect of swimbladder compression in herring. ICES Journal of Marine Science (2003) 60:548–554.
Hazen E. L., Horne J. K. A method for evaluating the effects of biological factors on fish target strength. ICES Journal of Marine Science (2003) 60:555–562.
Helfman G. S. Fish behaviour by day, night and twilight. In: Behaviour of Teleost Fishes—Pitcher T. J., ed. (1993) London: Chapman and Hall. 479–512. Fish and Fisheries Series, 7 715.
Hjellvik V., Godø O. R., Tjøstheim D. Modelling diurnal variation of marine populations. Biometrics (2001) 57:189–196.[CrossRef][Web of Science][Medline]
Hjellvik V., Godø O. R., Tjøstheim D. Diurnal variation in bottom-trawl-survey catches: does it pay to adjust? Canadian Journal of Fisheries and Aquatic Sciences (2002) 59:33–48.
Hjellvik V., Godø O. R., Tjøstheim D. Diurnal variation in acoustic densities. Canadian Journal of Fisheries and Aquatic Sciences (2004) 61:2237–2254. why do we see less in the dark?
Huse I., Korneliussen R. Diel variation in acoustic-density measurements of overwintering herring (Clupea harengus L.). ICES Journal of Marine Science (2000) 57:903–910.
Huse I., Ona E. Tilt-angle distribution and swimming speed of overwintering Norwegian spring-spawning herring. ICES Journal of Marine Science (1996) 53:863–873.
ICES. Report of the Northern Pelagic and Blue Whiting Fisheries Working Group. (2006) 25 August–1 September 2005. ICES Document CM 2006/ACFM: 05.
Johnsen E., Iilende T. Diurnal variation in commercial CPUE and survey catch rates. In: Can fishery data improve Namibian hake survey estimates? Fisheries Research, in press.
Jones F. R. H. The swimbladder and the vertical movements of teleostean fishes. Journal of Experimental Biology (1951) 28:553–566. 1. Physical factors.[Abstract]
Jones F. R. H., Scholes P. The swimbladder, vertical movements, and the target strength of fish. (1981) Meeting on Hydroacoustical Methods for the Estimation of Marine-Fish Populations, 25–29 June 1979. Cambridge, MA, USA: The Charles Stark Draper Laboratory, Inc. 964. 2. Contributed papers, discussion and comments.
Knudsen H. P. The Bergen echo integrator: an introduction. ICES Journal of Marine Science (1990) 47:167–174.
Legendre P. Spatial autocorrelation—trouble or new paradigm? Ecology (1993) 74:1659–1673.[CrossRef][Web of Science]
Maclennan D. N., Fernandes P. G., Dalen J. A consistent approach to definitions and symbols in fisheries acoustics. ICES Journal of Marine Science (2002) 59:365–369.
McClatchie S., Alsop J., Ye Z., Coombs R. F. Consequence of swimbladder model choice and fish orientation to target strength of three New Zealand fish species. ICES Journal of Marine Science (1996) 53:847–862.
Michalsen K., Godø O. R., Fernö A. Diel variation in the catchability of gadoids and its influence on the reliability of abundance indices. ICES Journal of Marine Science (1996) 53:389–395.
Monstad T. Blue whiting. In: The Norwegian Sea Ecosystem—Skjoldal H. R., ed. (2004) Trondheim: Tapir Academic Press. 559. pp. 263–314.
Neilson J., Perry R. I. Diel vertical migration of marine fishes. Advances in Marine Biology (1990) 26:115–168. an obligate or facultative process?[Web of Science]
Neilson J. D., Clark D., Melvin G. D., Perley P., Stevens C. The diel-vertical distribution and characteristics of pre-spawning aggregations of pollock (Pollachius virens) as inferred from hydroacoustic observations: the implications for survey design. ICES Journal of Marine Science (2003) 60:860–871.
Pedersen G. Methodology for in situ target-strength measurement of fish. (2007) Norway: University of Bergen. PhD thesis.
Ona E., Mitson R. B. Acoustic sampling and signal processing near the seabed: the deadzone revisited. ICES Journal of Marine Science (1996) 53:677–690.
Rivoirard J., Simmonds J., Foote K., Fernandes P., Bez N. Geostatistics for Estimating Fish Abundance. (2000) Oxford: Blackwell Science. 206.
Stensholt B. K., Aglen A., Mehl S., Stensholt E. Vertical density distributions of fish: a balance between environmental and physiological limitation. ICES Journal of Marine Science (2002) 59:679–710.
Venables W. N., Ripley B. D. Modern Applied Statistics with S. (2002) 4th edn. New York: Springer.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Fablet, R. Lefort, I. Karoui, L. Berger, J. Masse, C. Scalabrin, and J.-M. Boucher Classifying fish schools and estimating their species proportions in fishery-acoustic surveys ICES J. Mar. Sci., July 1, 2009; 66(6): 1136 - 1142. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||