ICES STATUTORY ?vmETING 1993 . c.M::1993/B:19 Ref.H

A THREE LEVEL PELAGIC TRAWL FOR NEAR SURFACE SAMPLING OF JUVENILE FISH

ABSTRACT

by Olav Rune God~ and John Willy Valdemarsen

Institute of Marine Research, P.O. Box 1870 N-5024 Bergen,

Norway

In the northeast Atlantic Ocean catch rates of juvenile fish from pelagic trawls are used to calculate indices of relative abundance which are applied as a first indication of year class st;rength. A major constraint for a reliable time series of these indices may arise from a vary-@ ihg vertical fish distribution, particularly near the surface due to sampling gear behaviour.

In this paper a new three level trawl, which samples three separate depths simultaneously, is described in detail and catches are compared with catches from a standard juvenile survey trawl. The new trawl was highly more efficient in catching small juveniles in the near surface layers than the standard trawl. We concluded that changes in densities in this layer could sig­nificantly affect the reliability of the catch data from the standard trawl.

INTRODUCTION

In the northeast Arctic, estimation of recruitment of demersal fish has been derived from standardized pelagic trawl surveys run during different phases of the pelagic stage (ANON 1992). Time series of abundance indices for northeast Arctic cod and haddock have been used to tune recruitment estimates in the assessment (ANON 1993). God~ et al. (1993) sug­gested that variation in fish size, vertical distribution and migration may strongly affect the efficiency of the standard sampling trawl used in these surveys, and consequently, reduce the reliability of estimated abundance indices of the pelagic.stages of larval gadoids.

In this paper we present the design of a trawl for detailed studies of vertical distribution of juvenile fish, and some of the results from the first experiments.

MATERIAL AND METHODS

Experiments with the standard sampling trawl and a new trawl were carried out onboard R1V "Michael Sars" (47 m LOA- 1500hp) during the period May 19-23 1992 at two locations off the Norwegian coast (Fig. 1).

The standard sampling trawl (ST) is a four panel, 16 fathoms capelin trawl with 200 mm mesh size in front and 20 mm in the 40 m long codend (God0 et al. 1993). A 10 mm liner was inserted as cover in 4 m of the rear part of the codend. Doors and rigging were the same as presented by God0 et al. (1993).

The experimental trawl (ET) consists of three equal trawls (Fig. 2), mounted on the top of each other (Fig. 3) to facilitate sampling of three equal volumes of water at different water depths. The headline of the upper trawl was equipped with 80 kg flotation whereas 40 kg weights were attached along the lower ground rope. In addition 35 kg weights were attached to the lower wingtips of the lowest trawl (Fig. 3). The trawl was rigged with 2m2 Suberkrub doors connected to the wing tips by 50 m bridles, which were extended with 0.5 m for the very upper and lower pairs (Fig. 3 ).

The experiments started with 14 experimental hauls towed alternatively with the headline of the upper trawl at surface and at 20 m. The tows were carried out in a limited area and were followed by 4 hauls using the standard trawl, alternatively sampling with the headline at sur­face and 20 m. The trials where then continued with 8 hauls with ST at surface position in the southern location followed by 11 hauls with the ET, 9 at surface and 2 with upper headline at 20 m. A standard tow with both arrangements consisted of the trawl being towed for 30 min­utes at a speed of 2.8-3 knots. Those hauls taken between sunrise and sunset were considered day hauls and all others night hauls. Due to the short duration of darkness at the latitudes of the experiments in end of May, only 4 night hauls (2 at surface and 2 at 20 m) were taken in northern location and 2 night hauls at surface in the southern location.

Trawl geometry was observed with acoustic trawl instrumentation (SCANMAR and Simrad FS3300).

The experiments were carried out under favourable weather conditions.

RESULTS

Trawl geometry and performance

Trawl geometry measurements at surface position showed that the ST trawl mouth filters an area of about 300m2, from 0-30 m depth. When trawling with the headline at 20 m, the area filtered is unchanged but the depth range is reduced to about 20 m (see also God0 et al. 1993).

2

The individual trawl mouths of the ET, observed with the trawl sonar, had equal geometry,

which did not change between depths. The mouths were almost circular with a diameter of

5.5 m (Fig. 3). Trawling depth changed with warp length as illustrated in Fig. 3. Standard

depths in the experiments where the alternatives 1 and 3 in Fig. 3, i.e. upper headline at sur­face and at 20 m.

vertical distribution

Mean catches from repeated hauls with ET at 0 and 20 m are assumed to reflect density distri­

bution of juveniles at 5.5 m intervals to about 35-40 m, from 0 to 16.5m and 20 to 35 m

respectively (Fig. 3 and 4).

Saithe: Saithe was the most abundant species in the trawl catches. In the northern location the highest concentrations appeared to be at 20-25 m (i.e. in the upper trawl of the deepest hauls) during the day (Fig. 4a). Although the lowest numbers were recorded in the surface layer dur­

ing the day, these experiments demonstrated that 0-group saithe in the area are distributed in

considerable quantities right up to surface. This is more obvious from the southern location (Moere) (Fig. 4a). Here the highest concentrations were.found at surface during both day and

night, and gradually decreased with depth. As in northern experiment, catches in general

were lower at night, indicating that fish were dispersed over a wider depth range at night or that efficiency of the trawl was lower at night.

Cod: Catches of cod were much lower than saithe, particularly in the northern location. In the

south, day catches indicated reduced densities towards surface, whereas no such trend

appeared at night (Fig 4b ). Catches in north are too low to draw any conclusion, although the

almost complete absence of cod above 11 m at day, underline the same tendency as in south,

i.e more fish in surface layers at night.

Standard trawl- experimental trawl comparison The catches were first adjusted according to volume filtered by the trawl openings, i.e. the

trawl catch of ET was arrived at by multiplying the individual trawls of ET by 10 before

averaging them (God~ et al. (1993). The results showed that the efficiency of the ET exceeds

the ST many times (Fig.5). This is particularly clear for .saithe in south (Moere Fig. Sa).

Mean len~th comparisons There is a great deal of variation in mean length at the different stations and samples within

the same haul of the ET (Fig. 6). No systematic difference in mean length, indicative of a

depth dependent size distribution, were found. For the ST hauls it should be mentioned that

the substantial drop in mean length of saithe and the simultaneously peak in cod in southern

area, eo-occurred during the only night period sampled with ST (st.nos. 252-254).

3

DISCUSSION

The new three level sampling trawl functioned according to expectation with only minor adjustments during the initial trials. If avoidance is a negligible problem for sampling pelagic juveniles, as indicated by God~ et al. (1993), then the sampling trawl and experimental design used in the presented experiment give resolved fish density and size distribution down to about 20 m. This depth range can of course be expanded by performing repeated hauls at different depths. The function tests indicated no difficulties by running the upper headline down to depths of 50 m.

God~ et al. (1993) showed that the low efficiency of the ST would have the strongest effect on indices of abundance when fish were distributed in the very surface layer, and when diur­nal vertical movements occurred. In addition to the function test with the new trawl, we wanted to study such effects on results from the standard post larvae survey conducted in April-May, which covered the locations in the present study just before the start of the exper­iments (Bakkeplass et al. 1992). The main aim of that survey has been to estimate year class strength of saithe, however, the correlation with recruitment estimates from VPA has been poor (ANON 1993). The experiments with ET demonstrate an area difference in vertical dis­tribution of saithe, i.e. in south the highest densi.ties. were found at surf~ce while fish were more evenly distributed in north. No systematic .diurnal migration appeared, but the saithe seemed to be more dispersed at night With the expected lower e;fficiency of ST at surface due to its bigger mesh sizes and construction, compared to that estimated at deeper water (God~ et al. 1993), the recorded variability in vertical distribution may thus strongly reduce reliability of the estimated indices. Possibly geographic differences rather than diurnal, may represent the more important source of variability. Due to the design of the experiment, i.e. the ST hauls were performed either after or before the ET hauls, the catch comparison repre­sent a poor basis from which to do accurate comparison of efficiency. The data, however, underline the results from previous experiments that the ST has a very low efficiency (God~ et al. 1993). Since no systematic size differences.with depth was observed, then this indicates that the low and variable efficiency of ST will have a less effect on size. distribution of the indices. The day-night differences in size indicated for STin southern area, however, suggest that behavioural factors may affect ST catches. Although the material presented here is not extensive enough to draw finn conclusions on the effect of vertical distribution and migration on abundance indices, the new sampling trawl was found to represent a useful tool for such studies.

REFERENCES

ANON 1992. Preliminary report of the international 0-group survey in the Barents Sea and adjacent waters in August-September 1992. ICES C.M. 1992/G:82: 1-33.

ANON 1993. Report of the Arctic fisheries working group. Copenhagen, 25 August- 3 Sep-

4

tember 1992. ICES C.M./Assess:1: 1-169.

Bakkeplass, K., Nedreaas, K.H., Senneset, H and Smedstad, O.M. 1992. Intern toktrapport. IT 53/92: 1-25.

God~, O.R., Valdemarse, J.W. and EngAs, A. 1993. Comparison of efficiency of standard and experimental juvenile gadoid sampling trawls. ICES Mar. symp., 196: 196-201.

5

ror- ao 1rot.1- 5JCJE MAT P A~~ EL 04oo 05oo 06oo 07oo OSoo 09oo 1 0 0 0 1 1 0 0 1 2 0 0 RTEX mm # 2 - 2 13.7m 6 6 0 0 -- - -- 1\Ji'"_ -6 6 0 0

4 X 1 f48

~I , . 4 X 1T28 23 ~12 1 X 1 TlB

PA 200 -- ~~ 1242 94

19 ~12

6 5 00 J J 6500 -- - -

...,.

PA 40 148 ~12 414

PA ::ssV2 414 :4

6 4 0 0 ~ t h l6 4 0 0

P.<\ 19E: 1/2 414

:24 -PA

99 i/2 552

6 3 00 j ~~~ l6 3 0 0

-

PA ::r 300

414 10

6 2 0 0 I p f'U , I I I I I I I 6 2 0 0

04oo 05oo 06oo 07oo OSoo 09oo 1 Ooo 11 oo 1 2oo

Fig. 1. Northern and southern experimental areas. Fig. 2. The experimental trawl.

to vessel

A

10m

SUberkrUb Aluminium 2 m2

75 kg

31m Sm , __ , _____ _..::.... _______ _ 80kgl

35kg 1 40kg

Fig. 3. Rigging of the experimental trawl and and sampling depth versus warp length at 2.8 kn ..

a

211

20

b

20

211

DEPm

DEPI'H

TDIE•d AREA•Helrel

o~••~~·~•~•m••~~•m*~ ~

TDIE•a AREA• HeirS

~ ~ :;:;:·.=:::::-::::::::::::::::::::·::::·:::::~:·.; .. ·:~·:::::::::::~ill::~·.:

0 10 211 80 40 60 SI CA'roll(ll)

TlliE•d AREA•Helrai

211

TDIE•d ABEA•.._..

DKPI'R

10 20 80 40 60 SI 'Ill c.m::H(D)

'l'IKB•• ARBA-x ....

DEPI'H~~~~

o 1 z a ~ 6 e 7 8 g ~ n ~ ~ u ~ M n m w 211 CATCH (I!)

'l'IME•d AREA•Ifoaoe

DEPm_..-------------------------------------, DEPI'II

I CATCH (I!)

'l'DIE•a AREA• HeirS

ON ~ :;.::.;:::::::::::::::::::::::::::::=:::::·:::·.;:···

C&1al(ll)

20

I DEPI'H

0 1 ! I ~ 6 e 7 8 V ~ U ~ ~ U ~ CATCH (I!)

TDIE•a AREA•N-..

r~~~~~~.:~FF~~~F~~~::::•::~1~~~~~~~ 6 e 7 to n ~

CATCH(n)

Fig. 4. Vertical distribution of cod in the northern (Helgel.) and southern (Moere) area at day (Time=d) and night (Time=n), a) Saithe and b) Cod.

COD TIME=d AREA=Moere

CATCH(n) --------~ 140

130 120

no 100 90

80

~

0 m DEPTH

COD TIME=d AREA=IIclgel

CATCH(n) ..--------.. 60

40

so

m

DEPI'H

0 20

SAITHE TIME=d AREA=Moere

CATCH~)r-----------~ 600

400

300

IJEl''I'H

0 m

SAlTHE TJME=d AREA~IIclgel

CATCH~)~----------~ 1400

1300 Jm)

1100 1000

900

800

'lOO

tm 600

400

800

200

100 0

0 20

DEPI'H

Fig. 5. Comparison of STand ET catches of cod and saithe after standardization to volume filtered by the trawl openings.

3.5

3.4j 3.3

3.2

Sl

~ 3.0

z2.9 r:il ..:I

~2.8

~ 2.7

2.6

2.0

2.4

2.3

2.2

ZJO

5

4

~ z ~

~ 8

I

\ I I I I • I I I I• I it \ 1

~ 11 11 •

SAITHE

NORTH

240 250 260 2'70 280 STATION NO.

COD

. . ~ '•i I I

:'•, I I • • J i ••• I I i ••• I I

t : • •• I I :1 ; 'i- I 11 : 1'•, I I I : I '•J I I : I ''· I I : J .,

~ I I : --6" · .. I \ : t ~ I : I \

0--{l i \ I I u /~i ' I' •• I I., -- ' I J •• : \ I ; ---- ' •I ' • ~ \ I --'..:,_.--- • ,~.! ~J- ... :1 \ , ..

.- I I :1 \!." I I· • I I~ j ~

9-, SOUTH NORTH

2~~~~~~~----~~~~~~----~~----~~~----~--~ 2W 240 250 260 270 2fJ)

STATIONNQ

Fig. 6. Comparison of mean length in the individual trawl hauls and samples within the ET hauls (1) upper, 2)middel,3) lower sample) and ST (4).