© 2006 International Council for the Exploration of the Sea
Estimating end effects in trawl catches
a Département HGS IFREMER, 7 Place du séminaire, BP 7, 17137 L'Houmeau, France
b Département EMH, IFREMER Rue de l'Ile d'Yeu, B.P. 21105, 44311 Nantes, Cedex 03, France
*Correspondence to M-J. Rochet: tel: +33 240 374000; fax: +33 240 374075. e-mail: mjrochet{at}ifremer.fr.
The end effect in trawl catches is defined as the proportion of the fish catch taken during shooting and hauling of the net, a period excluded from that nominally referred to as haul duration. If important, this effect will lead to biased abundance estimates, because the swept area will be underestimated. An experimental survey was carried out to compare catch numbers obtained in standard research 30-min hauls with those from 0-min hauls, the latter referring to the trawl being hauled as soon as the trawl geometry stabilized on the seabed. Average catch ratios (0-min/30-min hauls) ranged from 0.05 (s.d. 0.06) for sole to 0.34 (s.d. 0.64) for hake, indicating that the end effect might be more important and more variable for highly mobile species. As a consequence, the bias in abundance indices derived from swept area estimates that ignore end effects will be species-dependent.
Keywords: abundance index, bias, trawl survey
Received 22 June 2005; accepted 14 March 2006.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
Recent studies have shown that catch rates (catch per unit effort, cpue) for fish and crustaceans can be relatively higher for 15-min hauls than for 30-min hauls (Godø et al., 1990; Walsh, 1991; Somerton et al., 2002). These observations serve to suggest that by systematically shortening haul durations during scientific surveys, the total number of samples could be increased, resulting in increased precision of abundance index estimates (Folmer and Pennington, 2000; Kingsley et al., 2002; Pennington et al., 2002). However, the reasons for the cpues being higher in shorter hauls are not obvious, and several hypotheses can be advanced: (i) fish escapement is less at the beginning of the haul when individual fish are surprised by the arrival of the net haul; (ii) fish are caught before and after the main haul, either while the gear is settling or after the hauling starts; and (iii) the net becomes saturated during longer hauls, which increases escapement as the net fishes less effectively. Evidence for the surprise effect comes from visual observations made at the trawl mouth. Albert et al. (2003) observed that proportionally more Greenland halibut entered the trawl during the first few minutes on the sea floor than later in the haul. Further, the time elapsed after arrival at the sea floor but before trawl geometry stabilizes, which is generally taken as the nominal starting time, and while hauling, has been estimated to be non-negligible for shrimp trawling in Greenland waters (Kingsley, 2001). Finally, net saturation will probably occur in certain circumstances, but should not generally be a problem in the short hauls of scientific surveys.
Therefore, the two most plausible hypotheses of those listed above, although it is acknowledged that other factors could also contribute, are the surprise effect and the end effect. Both effects are independent of haul duration and thus might be proportionally more important for shorter hauls. This would be particularly problematic if surveys with different tow duration were to be compared, or if tow duration and sinking and hauling times varied among stations within a survey. Whereas tow duration generally varies within prescribed limits, sinking and hauling times depend on depth, so surveys covering a larger depth range might be more affected. In addition, the size of fish caught while sinking and hauling might differ markedly from the sizes of those caught during the nominal towing period (e.g. for saithe, Pollachius virens; Huse et al., 2002). For these reasons, end effects might introduce significant bias in survey data, especially for low-density species. This note reports the results of an experiment designed to estimate the number and length of fish caught before and after the nominal haul duration, conducted in 2003 in the Bay of Biscay.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The experiment was conducted between 17 and 19 May and between 11 and 18 November 2003 on the RV "Gwen Drez". Nephrops twin trawls were used, with a 25.15-m headrope and a 28.60-m footrope consisting of a cable with 70-mm diameter rubber disks, and 2-m trawl doors weighing 280 kg. The same twin trawl was used on both occasions, but with different mesh size (Table 1). For technical reasons, haul duration and trawl geometry were not always recorded. In cases with measurements, average vertical opening was 1.8 m and horizontal opening 11.6 m. Towing speed was approximately 3.4 knots.
|
Six hauls of 30-min duration were carried out (Table 1). Hauls were considered to have started when the net had nearly reached a stable shape, generally a few minutes after the chain settled on the seabed. Similarly, after 30 min, the haul was considered finished when hauling begins. The time between these two points is here referred to as the nominal haul duration. For each full haul, three zero-duration hauls were also carried out, defined as when the trawl was hauled as soon as the nominal haul duration would have started in an ordinary research haul. These zero-duration hauls were located approximately at the beginning, middle, and end of the full haul.
All fish caught were identified to the lowest taxonomic level possible, weighed, enumerated, and measured. For the analysis, the catch of the two twin-trawl nets were summed. For nine species, sufficient fish were caught to allow comparison between full- and zero-duration catches. Densities were estimated as the number caught divided by the swept area, i.e. the horizontal trawl opening multiplied by the distance trawled (not including the unknown additional area swept during sinking and hauling). Length distributions of the catch for each species were compared between zero-duration and full hauls by means of a Kolmogorov–Smirnov test (Sokal and Rohlf, 1995).
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The time between the trawl settling on the bottom and the start of the nominal haul varied between 1.5 and 5 min (Table 1), in accord with the results reported by Kingsley (2001). Time between nominal and actual haul end (identified by a change in headline height) was of similar magnitude.
Estimated densities ranged between 15 and 11 000 fish km–2 (Table 1). An analysis of variance on the log-transformed catch numbers (ln(catch + 1)) showed that the location of the zero hauls did not matter, i.e. that the three zero-duration hauls could be considered as replicates. This was true for all species (Table 2).
|
Comparative plots indicated positive correlations between zero-duration and full haul catches (not shown). The catch ratio was strongly species-dependent (Figure 1). For highly mobile demersal species, such as whiting (Merlangius merlangus) and bib (Trisopterus luscus), the proportion caught in "zero-duration" hauls was high and variable, whereas for benthic species such as sole (Solea solea), it was always <10%. The exception to this statement seems to be scaldfish (Arnoglossus laterna), for which the proportion was high despite it being a benthic species. The few scaldfish caught (see Table 3) might have contributed to this result.
|
|
No significant difference was detected between length distributions in zero-duration and full hauls, except in the case of wedge sole (Dicologlossus cuneata), which were smaller in zero-duration hauls (Table 3).
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
For less mobile and truly benthic species, the catch before and after the nominal haul duration might be proportional to the duration of that period, whereas for demersal or more mobile species such as hake (Merluccius merluccius), it could be that the end effect would be more important. This latter observation sheds doubts on swept-area-based abundance indices for highly mobile species. End effects generally did not affect length distributions.
The catch of fish before and after the nominal haul, or the end effect, introduces a bias in abundance estimates that varies among species, the magnitude of which is inversely related to haul duration. This can be particularly important for surveys conducted in regions where fish density is low, as in this study. Therefore, the end effect might counterbalance the benefits expected from shorter hauls. End effects would also increase the variability in abundance indices and, more generally, in indicators derived from survey data. This adds to the uncertainty induced by spatial heterogeneity in fish density and trawl geometry. In future, this problem might be overcome by the use of dedicated technology, i.e. using trawls that are open while lowered and closed before hauling (Engås et al., 1997), although certain difficulties remain in practically operating such systems (Sarda et al., 2002).
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Albert O.T., Harbitz A., Høines Å.S. (2003) Greenland halibut observed by video in front of survey trawl: behaviour, escapement, and spatial patterns. Journal of Sea Research 50:117–127.[CrossRef]
Engås A., Skeide R., West C.W. (1997) The ‘MultiSampler’: a system for remotely opening and closing multiple codends on a sampling trawl. Fisheries Research 29:295–298.[CrossRef][Web of Science]
Folmer O. and Pennington M. (2000) A statistical evaluation of the design and precision of the shrimp trawl survey off West Greenland. Fisheries Research 49:165–178.[CrossRef][Web of Science]
Godø O.R., Pennington M., Vølstad J.H. (1990) Effect of tow duration on length composition of trawl catches. Fisheries Research 9:165–179.[CrossRef][Web of Science]
Huse I., Michalsen K., Skeide R. (2002) Estimate of towing time in surveys. ICES Document CM 2002/J: 04. 9 pp.
Kingsley M. C. S. (2001) Studies in 2001 on the end effect of the Skjervøy 3000 trawl in the West Greenland shrimp survey. NAFO SCR Document, 01/177. 7 pp.
Kingsley M.C.S., Carlsson D.M., Kanneworff P., Pennington M. (2002) Spatial structure of the resource of Pandalus borealis and some implications for trawl survey. Fisheries Research 58:171–183.[CrossRef][Web of Science]
Pennington M., Burmeister L-M., Hjellvik V. (2002) Assessing the precision of frequency distributions estimated from trawl-survey samples. Fishery Bulletin US 100:74–80.
Sarda F., Company J.B., Estevez M.A. (2002) Efficacy of a remote control closure of cod-end used in a bottom trawl experience. Scientia Marina 66:423–432.[Web of Science]
Sokal R.R. and Rohlf F.J. (1995) Biometry(W.H. Freeman and Co., New York) 887 pp.
Somerton D.A., Otto R.S., Syrjala S.E. (2002) Can changes in tow duration on bottom trawl surveys lead to changes in CPUE and mean size? Fisheries Research 55:63–70.[CrossRef][Web of Science]
Walsh S. J. (1991) Effect of tow duration on gear selectivity. NAFO SCR Document, 84. 9 pp.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||