Mauritania flat at the Banc d'Arguin,

Fish species composition, tidal

migration and feeding on a tidal flat at the Banc d'Arguin,

Mauritania

J.A.Vonk

Dept. of Marine Biology Rijksuniversiteit Groningen The Netherlands

Supervisors:

W.J. Wolff J-G. Hiddink

Summary

The Banc d'Arguin is an area characterised by shallow waters and tidal flats. The relationship between the fish stock on the continental shelf and the Bane d'Arguin is important for the management of the Mauritanian marine living resources. The

shallow waters act hereby as a nursery.

Nine crustacean, three molluscs and seventeen fish species were found on the tidal flats. The species most found were Callinectes niarginatus (crustacean), Sepia officinalis (cephalopod) and the fish species Sciaena umbra, Diplodus sargus, Dicentrarchus punctatus, Anus heudelotii, Ephippion guttiferum and Umbrina canariensis. These species were migrating and feeding on the tidal flat. Some of the fish specimens were canying roe.

At a low location on the tidal flat we caught significantly (p= 0.006) more fish

specimens with a sliding net compared to a location higher on the tidal flat. For the fykes there was no difference. The fish species were mainly predating on amphipods, shrimps, bivalves and gastropods. Most of the prey species were intertidal species caught at the tidal flat. The tidal flat areas of the Bane d'Arguin therefore act as feeding grounds for the fish species. It is at this moment not possible to determine a relationship between different conditions on the tidal

flats and the number and

composition of fish present.

Table of contents

Summary ¹

Table of contents ²

Introduction ³

Material and methods

Site description ⁴

Tide ⁵

Materials used ⁵

Ep/zippion guttferu,n tagging 6

Identification 7

Results

Species caught ⁸

Fykes and gill net 9

Distribution in the fykes 10

Gill net catches 11

Ep/zippion guttferunz tagging 12

Fish species stomach contents 12

Discussion and conclusions

Species composition 15

Tidal migration 16

Fish species composition and migration 17

Spawning 17

Ephippion guttferunz tagging 17

Feeding behaviour 17

Comparing studies 18

Future research 19

Acknowledgement ²¹

References ²²

Contents Appendices ²³

Appendices I -XXXIV

Introduction

The Banc d'Arguin is part of the Parc National du Banc d'Arguin, bordering the Sahara in Mauritania, W. Africa. It is an area characterised by shallow waters and tidal flats. The tidal flats of about 490 km2 consist of sandbanks and mudflats covered with seagrasses (Wolff and Smit, 1990). The area is situated in a transition zone.

Twice a year the northern subtropical hydrodynamic front passes over. This causes large variability in abiotic characteristics, like temperature and salinity. The Banc

d'Arguin fauna will therefore shows temperate, sub-tropical as well as tropical

elements (Jager, 1993).

The continental shelf of West Africa, between the Strait of Gibraltar and Cape Verde, is one of the most important fishery areas of the world (Duineveld et^a!., 1993). The Mauritanian shelf acts hereby both as a nursery for some of the most important pelagic fish species and as a source of exploitable stocks of fish, cephalopods and shrimps (Duineveld et a!., 1993). For the management of the \lauritanian marine living resources, it is important to know the relationship between the fish stocks present on the continental shelf, and those living on the Banc d'Arguin (Jager, 1993).

The potential function of the Banc d'Arguin as a nursery for larval and juvenile fish has, however, hardly been investigated. Jager (1993) found the largest number of fish in the shallow waters around Tidra. They were on average smaller than in deeper areas. Most of these fish data were acquired by beam trawls. The number of fish at the intertidal flats has not been examined, also because in really shallow waters, about two meter deep, the disturbance by a boat and a trawl is large, especially for the larger fish.

Jager (1993) concluded that more study was required to assess the importance of the Banc d'Arguin in the lifecycle of fish species. When we went to the park area our main goal was the recording of the distribution of juvenile fish on the tidal area. This was focussed on the fish species composition, tidal migration and distribution at different levels of the tidal flat. With the gear types we used, fykes and a gill net, only large specimens of fish were caught, which were mainly predators. So of these fish species composition, tidal migration and feeding was recorded. Stomach contents were analysed to identify the prey species for these fish.

Material and methods

Site description

This research was done from 21st February till ^16tt April 2001 at the Banc d'Arguin, Mauritania (fig Ia). The Banc d'Arguin is situated between 19°20' and 20°30' N.

Depth generally does not exceed 4 m, except in the eastern part where depths down to 16 m occur (Jager, 1993). Salinity is higher than normal seawater, 38-42 % (Wolff, 1989), but the salinity can reach even more extreme values close to the shore, because no inflow of fresh water from the continent exists (Wolff & Smit, 1990). Wolff &

Smit (1990) estimated that at the Banc d'Arguin 425 km2 of tidal flats exist between Cap Tafarit and Cap Timiris (fig ib) and 66 km2 in the rest of the Banc d'Arguin.

Fig la,b,c: ^Maps of the Banc d'Argum (from Altenburg et a!., 1982). la Parc National du Banc d'Arguin, Mauritania (Africa). The park consists of a land and a marine part. lb Tidal flats and isles around the isle Tidra. This is the largest tidal flat area of the park. This study was done in the Bale dAouatif, east of the peninsula of Iwik (right upper corner). Ic Baie d'Aouatif. The location of the experiment is marked with an arrow.

The research site was located at the Baie d'Aouatif near the field research station of the Parc National de Banc d'Arguin at Iwik (fig ic). The Baie d'Aouatif consists of a large tidal flat with a large gully. The tidal flat consists of mudflat, partly covered with the seagrass species Zosiera and Cymodocea, and sand flats, partly covered with large dead Anadara shells. Some creeks and small pools with permanent water were

present on the tidal flat. Smit et aL (1989) measured a water temperature that

increased from 18-19°C in mid February to about 21-22°C in mid April 1986. Winds were mostly 4 to 6 Beaufort and from northern directions (Wolff& Smit, 1990).

Tide

Tidal range is about 1 meter at neap tides and 2 meter at spring tides. Currents are about 1 m*sec' in the channels (Wolff, 1989). The tide was divided in a period of neap- and of spring tide according to the lunar cycle for the analysis of the tidal influence on the fish migration (table 1). The tidal periods consisted of about seven days.

moon date tide tidal period

lq 2-15 neap 2-14 2-21

nm 2-23 spring 2-22 3-01

fq 3-3 neap 3-02 3-08

fm 3-9 spring 3-09 3-15

lq 3-16 neap 3-16 3-23

nm 3-25 spring 3-24 3-30

fq 4-1 neap 3-31 4-07

fm 4-8 spring 4-08 4-13

lg 4-15 neap 4-14 4-21

Table 1: Tidal periods correlated to the phase of the moon. Neap or spring tide was assumed to be two days after the day of the moon stand and around this day the tidal periods of neap and spring were fixed of about seven days.

Materials used

The fish were caught using fykes (stretched mesh width 2 cm; length 3 m) and a gill

net (stretched mesh width 8 cm; length 110 m, height 40 cm). The fykes were

repositioned several times to find a good location on the tidal flat. This was found along a creek that was flooded every tide anc' branched from the main gully of the bay. The fykes were placed with their entrance towards the ebb current on the tidal flat (Jig 2). The positions were about 150 m separated from each other. Fyke I and 2 were low, 3 and 4 middle, and 5 and 6 high on the tidal flat.

In the creek there was always water, even at low water at spring tide, but the fykes fell dry every low tide, except during neap tide, when about 40 cm water stayed at the

high and middle location and 70 cm at the low location. During one and a half-month the fykes were emptied each day.

The gill net was placed parallel to the main channel at different places and heights on the tidal flat. The low location was at the spring tide low water line, which was the edge of the tidal flat, and the middle location was 150 m up on the tidal flat, at the same height of the fykes at the middle location. The net was placed at the turning of the high or low tide and was removed half a tide later. So the net was about six hours on the tidal flat.

It was not possible to measure the absolute heights of the different locations of the fykes and the gill net at the tidal flat. From observations we know that the low water level at spring tide is at the low location and the low water level at neap tide is in between the low and the middle location.

All individuals caught with both methods, were measured, weighed and for the fish species the stomach contents and the presence of roe was determined. The fish were weighed on a spring-balance with an accuracy of 10 gram during the first month, but with a precision balance, accuracy of 10 mg, during the last three weeks.

Ephippion gultiferuintagging.

E. guttiferuni ^was

always observed solitary near the waterline. Therefore the

hypothesis arose that they may have an individual territory. To determine the number

Fig. 2: Schematically map of the fykes (numbers I till 6) and gill net on the tidal flat. Two pairs of fykes are about 150 m far off each other. Fykes 1 & 2 were low, 3 & 4 middle, and 5 & 6 high on the tidal flat. The four places of the gill net are indicated with middle and low. The middle location was at equal height on the tidal flat as fyke pair 3 & 4. Thegill net was not always placed at exact the same

location low on the tidal flat, but had a range of about 2 times its own length towards both sides of the creek.

of E. gutt[erum along the tidal flats an approach can be made for the total number of E. guttiferurn along the edge of the tidal flat. We tagged the caught specimens by giving them a 'wingtag' in the dorsal fin.

The E. gutuferum tagged, were always caught with the gill net. The gill net was placed at the low level on the tidal flat in the same area. E. gutt4ferum specimens caught with the gill net were usually still in good condition. As soon as they were out of the net they were placed in a bucket with water to keep them alive. The length and weight of the fishes was taken and after that the 'wingtag' was placed in the dorsal fin. The fishes were released at the same place as they were caught. All caught E.

gutiferum were examined on tags.

Identification

The fish were identified using Les poissons de nier de Mauritanie (Maigret & Ly, 1986) and Vissen van de Europese kustwateren en de Middellandse Zee (Lythgoe and Lythgoe, 1971).

Results

Species caught

The species we caught consisted of fish, crustaceans and molluscs. Nine crustacean species, three mollusc species (table 2) and seventeen fish species (table 3) were caught with both methods.

Five crab species and four shrimp species were caught. Except for one Callinectes niarginatus all crustaceans were caught with the fykes, due to the large mesh width of the gill net. Most abundant were the crab species C. marginatus and the shrimp Penaeus ,zotialis. The crab species caught, and other crabs too, gave some trouble by eating the fish in the fykes. To get the fish out they sometimes destroyed the net by cutting holes in it.

Also the molluscs were caught with the fykes. Sepia officinalis was the most abundant cephalopod species, but also the gastropods Cynibium sp. and Semfusus mono were caught. The latter two species crawled into the fykes.

Table 2: Crustaceans and mollusc species caught in the fykes. Most abundant was Sepia officinalis of which a lot of small specimens were caught. The number of crabs caught is probably underestimated because they could escape by cutting a hole in the fykes. They also could eat the fish out of the fykes.

Crustacean

total

#

length' ^- cm

mean SE

weight

g

mean SE

Callinectes marginatus Atelecyclidaesp.

Portunidae sp.

Uca fangeri Majidaesp.

Palaemonetes i'arians Pa!aemon longirostris Penaeus notialis Pc,zac'zis kerat/zurus

15

1

5

1

1

2

1

18 2

4.5 0.8

1.3 -

2.2 0.3

2.0 -

0.8 -

4.8 0.6

4.7 -

8.3 0.4

9.5 3.3

- -

- -

- -

- -

- -

- -

- -

4.8 0.7

7.0 5.4

Mollusca Cymbiuinsp.

Se,nfusus mono Sepia ofJIcinalis

11 2 41

11.0 1.1

18.2 -

12.7 1.1

730.9 125.3

- -

2

'the width of the crabs was measured.

2 thesmall ones could not be weighted in the first month because the spring-balance was not accurate enough. Therefor mean Sepia weight overestimated the real mean weight of the S. officinalis caught.

Fourteen different fish families, belonging to seven orders were caught.

Characteristics (standard length, length and weight) of the individual fish specimens caught with fykes and gill net are listed in respectively appendix A and B. Two fish specimens could not be identified.

Table 3:All fish species caught with both methods classified by family, order and class.

Class Order Family Name

Actinopterygii Atheriniformes Atherinidae Aiherina sp.

Perciformes Gobiidae Gobius sp.

Moronidae Dicentrarchus punciatus Mugilidae Liza aurata

Sciaenidae Sciaena umbra Umbrina canariensis Serranidae Epinephelus aeneus

Epinephelus caninus Sparidae Diplodus puntazzo

Diplodus sargus Pleuronectiformes Psettodidae Psettodes belch en

Soleidae Solea senegalensis Siluriformes Ariidae Anus heudelotii Syngnathiformes Syngnathidae Syngnathus lyphie Tetraodontiformes Tetraodontidae Ephippion guttferum Elasmobrarichii Rajiformes Dasyatidae Dasyatis marmorata

Rhinobatidae Rhinobatos sp.

Fykes and gill net

The fish species showed large differences in length and weight between the methods (table 4 & 5). In the fykes more species were caught and also more fish specimens.

But the fykes were continuously on the mudflat, while the gill net was used 23 times.

The fykes caught more small fish species, while both Rajiformes were caught in the gill net.

Table 4: The total number of each fish species caught using the fykes with their mean standard length, length and weight and the total number of individuals per species which carried roe or sperm. All fish caught during the two-month research were applied. Standard length, length and weight was measured from all fishes if possible. Some individuals were too much eaten by crab to estimate their weight or even their length.

Species

Total

#

4

SL cm

mean SE

L Cm

mean SE

W g

mean SE

Gonads roe sperm

#

Anus heudelotii 42.7 1.4 49.3 1.4 1 164.6 99.8 2 -

Atherina sp. ¹ 7.2 - 8.5 - <10 - - -

Dicentrarchuspunctatus 42 15.6 0.9 19.3 1.1 84.7 17.8 5 -

Diplodussargus 31 16.4 1.3 20.2 1.7 234.8 31.7 7 1

Diploduspuntazzo 1 23.1 - 28.8 - 349.9 - -

Ephippionguttiferum 7 23.5 4.6 29.1 5.8 723.4 279.8 - -

Epinc'p/zelus aeneus 11 22.2 2.8 26.6 3.4 331.9 120.8 - -

Epinephelus caninus ¹ 30.7 - 36.5 - 600.0 - - -

Gobius sp. 12 7.3 0.6 8.6 0.8 9.4 1.6 1 -

Lizaaurata 9 15.7 2.3 ^19.2 2.8 103.6 40.6 ^{3 -}

Psettodes belcheni 2 46.8 7.0 53.2 7.9 2010.0 990.0 - -

Sciaena umbra 57 36.8 0.6 43.8 0.6 1008.3 44.5 15 1

Soleasenegalensis 4 11.0 2.5 12.6 2.9 25.0 11.9 - -

Svngnathus typhle 1 5.9 - 6.1 - <10 - - -

L,nbrina canariensis 21 33.5 0.9 40.7 1.0 743.9 63.4 6 -

Diplodus sargus, Ephippion guttferum, Sciaena umbra and Umbrina canariensis were obtained with both methods due to their size. Dicentrarchus punctatus^{and both} Epinephelus sp. were only caught in the f'kes, while Anus heudelotii was mainly caught in the gill net.

At low tide on the tidal flat we observed really small fishes in the permanent pools on the mudflat and even caught some of them with a small net (stretched mesh width 1.0 cm). They were mainly small Gobius sp., Pomatoschistus sp., and small Diplodus sp.

These specimens could however not be caught with any of the other methods used due to the mesh width of at least two centimetres.

For many of the fish species it was before spawning season, because some of the caught fish carried roe or sperm. Especially A. heudelotii and S. umbra often had roe.

Table 5: The total number of each fish species caught using the gill net with their mean standard length, length and weight and the total number of individuals per species which carried roe or sperm.

Species

total

#

SL cm

mean SE

L Cm

mean SE

W g

mean SE

Gonads roe sperm

# #

Anus heude!otii Dasyatis marmorata Diplodussargus Ephippion guttferu,n Liza aurata

Psettodesbelclieri Rhinobatos sp.

Sciaena umbra Solea senegalensis

U,nbrina canariensis

35 6 19 29

1

7 4 10

1

5

36.9 0.9 29.9 1.8

22.7 0.3 38.2 0.8

28.0 -

35.4 1.6 62.3 11.8 34.2 0.7

20.8 -

32.6 1.1

43.6 1.0

- -

27.5 0.4 47.7 0.9

34.0 -

40.9 1.8 65.0 12.5 40.8 0.8

22.3 ^-

39.1 1.1

887.6 62.4 1119.8 152.3

341.8 15.4 1667.3 84.9

332.4 -

743.4 138.6

500.0 -

833.1 47.6

103.4 -

667.5 59.7

12 13

2 -

5 -

1

- -

1 -

- -

2 3

1 -

- 3

Distribution in the fvkes

The species caught in the fykes were divided by the location, low, mid or high on the mudflat and the tidal period, spring or neap (table 6). The fyke catches were grouped pair wise at the different heights on the mudflat, because they were placed close together. The highest numbers of fish were caught at the low site during neap tide.

The highest catch, nineteen Dicentrarchus punctatus at one time, was in these fykes.

Using the log-transformed mean number of fish caught, there was no significant effect of the location (Twofactor ANOVA, p0.37) or the spring-neap tide cycle (Twofactor ANOVA, pO.59) on the number of fish caught using the fykes.

Table 6: The mean number of fishes per species, the mean number of total fish and the mean fish characteristics, (standard) length and weight, caught per location (two fykes) and two tidal periods (table I) at the different combinations of location and tide, using the fykes. The replicates are the

number of catches at each location and tide.

Fyke

location tide replicates Species

hi ^,h middle low

neap 24

spring 20

neap 24

spring 20

neap 24

spring 20

Anus heudelotii

Dicentrarchuspunctatus Diplodus sargus Diploduspuntazzo Ephippionguttferum Epinephelusaeneus Epinephelus caninus Gobius sp.

Liza aurata Psettodes beicheri Sciaenaumbra Soleasenegalensis

Umbrinacanariensis

- 0.21 0.17

-

0.04 0.04

- 0.33

- - 0.42 0.04

-

0.05 0.10 0.30

- - 0.10

- 0.05

- - 0.40

- 0.05

- 0.08 0.04 0.04 0.04 0.04

- 0.08 0.08 0.04 0.38

-

0.29

- - 0.45

- 0.05 0.10

-

0.00 0.05

- 0.35

-

0.40

0.08 0.92 0.08 - 0.08 0.08 0.04 0.04

- - 0.42 0.04 0.13

0.05 0.15 0.20

- 0.10 0.15 - 0.00 0.15

- 0.55

- 0.10 Mean catches

fish

SE

1.25 0.30

1.05 0.20

1.17 0.25

1.40 0.35

1.92 0.77

1.50 0.26 Mean fish characteristics_________

SL cm

SE

20.8 2.5

25.6 2.6

29.5 2.5

26.9 1.6

23.7 1.8

25.5 2.2

L cm

SE

25.8 3.0

31.0 3.!

34.9 3.1

33.0 1.9

28.6 2.1

29.7 2.7

W g

SE

433 96

568 108

855 142

540 66

461 75

587 93

Gill net catches

The fish caught in the gill net were split up according to the location at which the gill net was placed on the mudflat (low or middle) and the tidal period (spring or neap) during which it was placed (table 7).

It was not possible to determine from which side, mudflat or gully, the fish had entered the net.

In their struggle to free themselves they got in at

^{least two}

compartments of the net, so they were always in one big knot.

The mean number of fish caught at the low location on the tidal flats was significant higher than at the middle location on the tidal flat (Twofactor ANOVA, ^p=O.OO6), using the log-transformed mean number of fish specimens per catch. Also the number

of fish species was higher at the low location. The only species that was more

abundant on the middle level was E. guttferum. At the low site during spring tide we had one catch of 27 A. heudelotii on a total of 36 caught specimens. The influence of spring-neap tide cycle (Twofactor ANOVA, p=O.l63) and the interaction of location

and tide (Twofactor ANOVA, p=O.O85) had no significant by affecting the number of fish caught with the gill net. The number of fish species was comparable at spring tides and neap tides.

Table 7: The mean number of fishes per species, the mean number of total fish and the mean fish characteristics, (standard) length and weight, caught per catch of a half tide (max. six hours) at the different combinations of location and tide, using the fykes. The replicates are the number of catches at each location and tide. More fishes were caught at the low side of the mudflat and the catches were higher during spring tide.

Gill net location

tide replicates Species

mid low

neap

4

spring

4

neap

4

spring

11

Anus heudelotii Dasyatis ,narmorata Diplodus sargus Ephippion guttferum Lizaaurala

Psettodes beicheri Rhinobazos sp.

Sciaenaumbra Soleasenegalensis U,nbrina canariensis

- - - 0.25

- - - -

• -

- 0.25

- 2.50

- - - - - 0.25

1.00 0.25 1.25 1.50 0.25 0.25

- 0.25 0.25

-

2.82 0.36 1.27 1.45 - 0.55 0.36 0.82 -

0.36 Mean catches

fish #

SE

0.25 0.25

3.00 1.22

5.00 1.41

8.00 2.91 Mean fish characteristics

SL cm

SE

41.6

-

38.9 2.0

32.3 1.7

34.7 1.0

L cm

SE

50.4

-

47.4 2.2

39.5 2.2

41.4 1.1 g

SE

2500

-

1518 206

968 138

907 59

E. gi,ttij'erumtagging

The 15 fish we tagged during the period were never caught again.

Fish species stomach contents

Of almost all fishes caught, the stomach contents were examined. Also fish specimens that were caught during the pilot experiment of the fykes were examined, Of six

species more than twenty individuals were caught of which at least six had a full stomach (table 8). For these species the mean length and weight was calculated of the individuals with a full stomach. Sciaena umbra and Umbrina canariensis had a low % full of the number of stomachs checked. Dicentrarchus puncratus and Diplodus

sargus sometimes died in the fykes and were half eaten by crab before they were recorded. These fish had no stomach left to examine.

Table 8: Six fish species, with more than twenty individuals caught of which at least six had a full stomach. The length and weight of the species with a full stomach is also given.

caught total

#

stomach full empty full of

examined

# %

fish specimens with full stomach

SL W

mean SE ^{mean SE}

39 45 50 guttiferum 41 67 canariensis 26

29 8 78.4

31 6 83.8

30 ¹² 71.4

10 2 83.3

24 40 37.5

6 18 25.0

37.0 0.9 883.4 59.2

16.1 1.0 93.9 20.9

21.2 0.4 285.1 20.4 36.8 1.3 1547.4 108.0 35.3 0.9 894.2 65.9 35.2 2.2 867.9 186.3

Because we did not have a precise balance, it was not possible to weigh the individual fractions of the different prey species in the stomach. The absolute number of species present in the stomach was therefore determined (table 9). Some stomach contents were hard to identify, because the organisms inside were already half digested.

Table 9: The mean number of prey species for the fish found with full stomachs of six fish species.

The prey species were first identified at lowest possible taxonomically level (Appendix E). Later they were grouped. Also the total number of caught fish was added to determine the importance of the prey species for the total catch of fishes. Unfortunately it was not possible to weigh the different species as

fractions of the total stomach content.

_____________ _____________ __________ _____________

predator

# full stomach

pre

Anus heudelotii

29 mean SE

Dicenirarchus punclatus

31

mean SE

Diplodus sargus

30

mean SE

Ephippion guttiferum'

10

mean SE

Sciaena umbra

24

mean SE

Umbnina canariensis

6

mean SE

Fish Amphipods Balanussp.

Callianassa Crab Shrimp Bivalve Sepia

Gastropods Zostera2 Oilier species (worm)

- -

41.48 17.71

- -

0.31 0.17 0.28 0.10 0.31 0.15 0.83 0.22 0.17 0.09 0.07 0.05

- -

1.83 0.60

0.06 0.04 5.97 2.28

- -

0.03 0.03 0.10 0.07 1.97 1.02 0.06 0.04

- -

2.61 0.76

- -

0.10 0.07

0.03 0.03 113.24 29.17

0.27 0.27

- -

0.03 0.03 1.73 1.02 1.73 1.19 0.17 0.07 13.07 3.64 0.83 0.07 0.10 0.10

- -

- -

127.10 60.32

- -

0.10 0.10

- -

11.95 3.45

- ^-

22.30 9.46

- -

0.60 0.34

0.23 0.09 0.13 0.09

- -

- -

1.04 0.23 1.29 0.68

- -

0.04 0.04 0.04 0.04

- -

- -

- -

- -

- -

0.17 0.17 1.00 1.00

- -

1.00 0.45

- -

- -

- -

0.50 0.34

twhole intestine was examined

2 of species in which was present

Anus heudelotii caught a lot of amphipods and species living in the mudflat like Callianassa sp. and bivalves. The Dicentrarchus punctatus stomachs contained mainly gastropods, shrimps and amphipods. This was also for Diplodus sargus, but amphipods were more abundant and bivalves were also part of their diet. The number

of amphipods caught by these three fish species was high, but the size of the

amphipods was very small (<6 mm).

Ephippion guttiferum lived mainly of Balanus sp., gastropods and bivalves, which he cracked with his strong jaws and teeth. Most of these bivalves were Anadara senilis specimens. Of this species the whole intestine was examined, because it had no stomach. Sciaena umbra hunted shrimps, crabs and small fishes, but also amphipods.

Umbrina canariensis had a strongly individual choice of food (Appendix E). Each individual had one main prey species that was most present. Per individual was this prey species different. For Unibrina canariensis this consisted of shrimps, bivalves and Callianassa sp. This individual food choice was also found in other species, like D. punctatus, D. sargus and E. guttjferum. It seemed that every individual had its own speciality.

To give some indication of the relative importance of the prey species in weight; 1

crab was estimated to be equal to about 3 shrimp, and they were equal to 45

amphipods; one bivalve was equal to 10 Ba/anus sp.

The percentage of full stomach of the total number of stomachs checked per location and tide was on average lower for the fykes (table 10). Observed was that when S.

umbra or U. canariensis were removed from the fyke, sometimes crab legs and shell- piece were found in the bucket in which the fyke was emptied.

Table10: Number of fish caught with full and with empty stomach at the different locations and tides.

The full stomach percentage is calculated from the total number of stomachs checked per treatment and location. For the fykes the percentages are lower on the average.

treatment stomach tide/location

net full ^empty full

# # %

fvke full empty full

% neap/high

neap/mid neap/low spring/high spring/mid spring/low

1 0 100

8 6 57

4 0 100

53 16 77

9 10 47

10 13 43

29 13 69

6 8 43

7 15 32

10 15 40

When we look at the fishes with full stomachs caught in the different treatments and at different locations and tides, the percentages of full stomachs in the fykes is low compared to the gill net, except for the low location at neap tide.

Discussion and conclusions

Species composition

Four different shrimp species were caught in the fykes on the tidal flat, which were mainly Penaeus notialis. December 1984, Hazevoet (1985) used a two-meter beam trawl (mesh width 0.5 cm) in the same area, east of Iwik peninsula. He caught eight shrimp species. They were mainly Palaemon elegans, 773 of a total 850 caught shrimps, and a few Penaeus notialis. We caught however, no P. elegans at all. This can be caused by both mesh widths of the gear and the method used to catch the prawn.

Of the five crab species were caught, Callinectes marginatus was the most abundant.

This crab species can swim through the water and was probably migrating over the tidal flats.

The high number of small Sepia officinalis specimens caught in the f'kes correlates with Jager (1993) who suggested that the Banc d'Arguin acts as a nursery for Sepia sp. She found large numbers of cephalopods in the shallow waters. So Sepia sp. not only uses the shallow waters, but they also migrate to the tidal flat area at high tide.

322 fish specimens, seventeen species, were caught at the tidal flat. The main species caught in order of numbers were Sciaena umbra, Diplodus sargus, Dicentrarchus punctatus, Anus heudelotii, Ephippion guttferum and Umbrina canariensis. More than 25 specimens of these species were caught. Hazevoet (1985) caught mainly Pomatoschistus narnzoratus and Gobius niger (both Gobius sp.), 200 of a total 250 caught fishes, but only one E. guttferum and one D. sargus. They caught ten different fish species at 1200 m2. The difference is probably caused by the different methods used in both studies. Jager (1993) collected 45 fish species with a beamtrawl, but her survey covered a large part of the Banc d'Arguin.

Species caught in most studies done at the Banc d'Arguin are Ephippion guttferum and Diplodus sargus (Sevrin-Reyssac & Richer de Forges, 1985; Hazevoet, 1985;

Campredon & Schrieken, 1989; Jager, 1993). In this study they were the two most abundant species caught. The distribution of E. guttiferum found in this study was

comparable to that found by Campredon & Schrieken (1989). They found E.

guttferum on seagrass beds in shallow waters, whereas Jager (1993) only caught single individuals at depths between 3 and 9 meter.

The difference in number of fish species caught with both methods was mainly caused by the difference in mesh width. The smallest size of the fish caught with the gill net was limited by its relative large mesh width, while the fykes were more limited by the entrance size, 40 cm. However the largest fish, a Psettodes beicheri of 53.8-cm length and 3000 g weight, was caught in a fyke.

Tidal migration

At low tide the tidal flat was almost dry and only small fishes (<10 cm) stayed in the pools and the creek (caught with a small net; personal observations). At high tide fish were caught on the tidal flat. This proves tidal migration of the fish species on the tidal flat. The quantification of the number of fish migrating on the tidal flat at different locations and tides is, however, difficult.

The spring-neap tide cycle and the level on the flats had no effect on the total number of fish caught with the fykes. This could be caused by: 1) the creek was at a somewhat lower level compared to the other parts of the tidal flat. The flooding of this area was therefore longer and the water level higher. Fish species probably used this to enter the tidal flat early and leave late. 2) The tidal flat area beyond the fykes was very large compared to the tidal flat area between the fykes. So the relative difference between the location of the fykes was small.

The tidal migration showed, however, a significant difference (Log-transformed, Twofactor ANOVA, pO.OO6) in number of fish between the two levels on the tidal flat using the gill net. Because we do not have any information about the absolute height, ^it

is not possible to correlate these differences in catch to the tidal

^flat elevation. No difference between spring- and neap tide cycle occurred using the gill net. The difference in mean length and weight of the fish caught with the gill net was mainly caused by the relative number of E. gutt/'erum on the total number of fish.

This species was on the average larger than the other species.

We sometimes had problems with holes in the fykes and the gill net. For the fykes the crabs caused the holes when they tried to get the caught fish out of the fykes. About ten times a fyke had a hole of more than five-centimetre diameter. Most of the fish

could escape through this. Sharp dead shells, mainly Anadara senilis, stuck in the bottom caused the main damage to the gill net when it was hauled or moved by the tidal currents. Some of the compartments were damaged. No error calculation was made because of these damages.

Spawning

Some species could come to the tidal flats to spawn here. Especially specimens of Anus heudelotii and Sciaena umbra often carried roe or sperm. It is not possible to determine from this study if the fish were actually spawning on the tidal flat or they were just feeding, while the spawning was later or elsewhere.

Ephippion guttiferuin ^tagging

The fact that none of the fishes caught was ever caught back could be caused by 1) having no territory at all, 2) loss of the tags, or 3) too much stress after the catch so they swam away. The second explanation is not really possible, because the iron tag was placed around the first spine of the dorsal fin. Both other explanations cannot be

rebutted.

Feeding behaviour

As many fish had filled their stomachs with intertidal animals when caught on the tidal flat, it is possible that fish come to the tidal flat to forage. Because only the mean number of prey species caught could be estimated and not the weight, it is difficult to give an estimation of the main prey species of the six fish species examined (Anus heudelotii, Dicentrarchus punctatus, Diplodus sargus, Ephippion guttzj'erum, Sciaena umbra, ^and Uinbrina canariensis). The stomach data gives an indication of the prey composition for these six fish species. Amphipods, shrimps, bivalves and ^gastropods were the main prey species for these fish. For E. guttferum Balanus sp. were also

important. Crab was eaten by almost all species, but in low numbers, while

amphipods were eaten in large numbers, but not by all species. This, combined with

the difference in size, makes it not possible to say which species is of more

importance.

Wolff et a!. (1987) observed crushed specimens of Anadara senilis near the edges of some tidal channels at low tide. We suggest that this was probably done by E.

guttferum, because crushed A. senilis shells were found in the intestines of this species.

When Sciaena umbra or Umbrina canariensis were caught in the fyke, sometimes crab legs and shell-pieces were found in the bucket in which the fyke was emptied.

Probably they could vomit their stomach contents in stress situations. Therefore the number of individuals of these species found with full stomachs was low. This caused the low percentage of fish with full stomachs in the fykes.

Very little is known about the annual production! yearly average biomass for benthic fauna at the Banc d'Arguin (Wolff, 1989) and of the pelagic fauna. A value for the

predation pressure of the fish at the different benthic or pelagic prey species is

therefor not possible to give. Only a relative difference between the locations can be given. The difference in number of fish caught with the gill net at the low and middle location can be correlated with a difference in predation pressure of the fish species at these locations. At the low parts of the tidal flat the predation pressure of the fish species is higher, except for Ephippion guitferum, which was more abundant on the middle location.

Fish feeding and migration

Jager (1993) found that the shallow waters had a large diversity in species, and contained large numbers of juvenile fish. The shallow parts of the Banc d'Arguin are therefore considered as an important habitat for large numbers of small fish. Not only juvenile fish were present on the tidal flat area, but also a lot of adult specimens.

Some of them were predating on the juvenile fish, like Psettodes beicheri, or were feeding on intertidal organisms.

For the intertidal areas it is more difficult to assess the interaction between height and number of fish and length present. On the tidal flats the densities and composition of prey species could have a large influence on the number of migrating fish and the fish species composition found. Variance in abiotic and biotic conditions on the tidal flats can therefore have a large influence on the fish species migrating and feeding at different locations on the tidal flat. Also the distance from the main streams seems to have an influence on the species number and composition.

Comparing studies

Comparing this study with previous studies is difficult because of the different methods used, different locations within the Banc d'Arguin and different times of the year in which the studies were carried out.

The method used most in previous studies to determine the species composition at the Banc d'Arguin was a beam trawl. But in really shallow waters, about two meter deep, the disturbance by a boat and a trawl is large, especially for the larger fish (Jager, 1993). So this method is not suitable on the tidal flats.

The variation in abiotic conditions at the Banc d'Arguin is high. Taking as an example the salinity, there is a gradient of 38 %o at the seaside to quite extreme values close to the shore due to evaporation and a lack of fresh water inflow (Wolff & Smit, 1990).

This also influences the biotic conditions. Depth also correlates with the number of individuals and the mean length (Jager, 1993). So different locations can have totally different species compositions.

The time of the year is important because the area is situated in a transition zone.

Twice a year, when the northern subtropical hydrodynamic front passes over, both abiotic characteristics, like temperature and salinity, and species composition change (Jager, 1993).

Future research

This study was a pilot experiment. When we went to the park we knew almost nothing about the local conditions on the tidal flat. Our main goal was to examine the fish species composition on the tidal flats. The results of the caught fish, combined with the tide and location they were caught, can be used for further research into the number and species composition of fish at different tides, places and seasons on the tidal flats at the Banc d'Arguin.

The stomach contents are a description of the feeding of six fish species at this location and time of the year. Predation pressure estimations of the fish species on the intertidal species at the tidal flats need a lot more research on the distribution and numbers over time and place of both fish and prey species. The interaction of prey and predator at the Banc d'Arguin is almost unknown for the fish species at this time.

The influence of the large-scale fisheries on the Mauritanian shelf on the fish species and composition at the Banc d'Arguin can not be estimated at this time. Future

to prevent catastrophic shifts in this ecosystem due to human fisheries and other activities.

Acknowledgement

I thank the Mauritanian authorities, in particular the director of the Parc National du Banc d'Arguin Mr. Mohammed 0. Bouceif, for their permission to carry out the investigations. Schure-Beijerinck-Popping Fond of KNAW, Vereniging Natuurmonumenten and Marco Polo for making this study possible. Dr. Jean Worms for his advice, practical tips and supplies. Also the collaborators of the Parc National, particularly Messrs. Abou Gueye, El Hassane ould Mohammed el Abd, M'bareck, Camara, Abdellai, Ahmedou and Mohammed for advice of the fish, the net repairing and all the other help. Finally my fellow students, especially Francesc Monserrat. His help and support was, not only for the gill net, indispensable.

21

References

Altenburg, W.,Engeimoer, M., Mes, R. & Piersma, T. (1982). Wintering waders at the Banc d'Arguin, Mauritania. Stichting Veth tot steun aan Waddenonderzoek, Leiden.

Duineveld, G.C.A., Lavaleye, M.S.S. & Noort, G.J. van (1993). The trawifauna of the Mauritanian shelf (Northwest Africa): density, species composition, and biomass. Hydrobiologia 258: 165-173.

Jager, Z. (1993). The distribution and abundance of young fish in the Banc d'Arguin, Mauritania. Hydrobiologia 258: 185-196.

Hazevoet, K. (1985). Sirius West Africa expedition, Banc d'Arquin, Mauritania, 2-9 december 1984. Publ. by Sirius, Amsterdam.

Lythgoe, J. and Lythgoe, G. (1971). Vissen van de Europese kustwateren en de Middellandse Zee. Nederlandse vertaling en bewerking 1976 by Moussaolt 'S Uiigeverzj by, Baarn

Maigret, J. and Ly, B. (1986). Les poissons de mer de Mauritanie. Sciences Nat., Compiègne

Sevrin-Reyssac, J. & Richer de Forges, B. (1985). Particularités de la faune ichtyologique dans un milieu sursalé du parc national du banc d'Arguin (Mauritanie). Océanogr. Trop. 20 (1): 85-90.

Smit, C.J., Blomert, A-M. Meijboom, A., Wolff, W.J. & Zwarts, L. (1989).

Hydrographical measurements. In Ens, B.J., Piersma, T., Wolff, W.J. &

Zwarts, L (eds.). Report of the Dutch-Mauritania project Banc d'Arguin 1985- 1986: 43-56. WIWO-report 25/RIN-rapport89/6, Texel.

Wolff, W.J., Abou Gueye, Meijboon, A., Piersma, Th. & Mamadou Alassane Sail (1987). Distribution, biomass, recruitment and productivity of Anadara senilis (L.) (Moliusca: Bivalvia) on the Banc d'Arguin, Mauritania. Netherlands Journal of Sea research 21(3): 243-253.

Wolff, W.J. (1989). The interaction of bentic macrofauna and birds in tidal flat esturies: a comparison of the Banc d'Arguin, Mauritania, and some estuaries in the Netherlands. In Elliott, M. & Ducrotoy, J.P. (eds.) Proc. ECSA ^19th

Symp.: Estuaries and coasts: spatial and temporal intercomparissons, Caen (France), 4-8 sep. 1989. Internat. Symp. Ser.

Wolff, W.J. & Smit, C.J. (1990). The Bane d'Arguin, Mauritania, as an area for coastal birds. Ardea 78: 17-38.

Contents appendices

A Fyke catches

Trials ^I

Main experiment ^III

B Gill net catches X

C Fish characteristics

Anus heudelotii XIV

Dasyatis marmorata XV

Dicentrarchus punctatus XVI

Diplodus sargus XVII

Ephippion guttferunz XVIII

Epinephelus aeneus XIX

Gobiessp ^XX

Lizaaurata ^XXI

Psettodes beicheri XXII

Sciaena umbra XXIII

Umbriiza canariensis XXIV

D Stomach contents

Anus heudeloiii XXV

Dicentrarchus punctatus XXVI

Diplodus sargus XX VIII

Ephippion guttferunz XXX

Sciaena unbra XXXI

Uinbnina cananiensis XXXII

E Mean number of prey in stomachs

A. heudelotii, D. punctatus, D. sargus XXXIII E. guttferuin, S. wnbra, U cananiensis ^XXXIV

Appendix A: Fyke catches trials. The total number of trial catches with the f'kes done at the Baie d'Aoutif, from 2P' February till 2 March 2001. All caught species are listed. 1 tide is about 12 hours. An x at remarks means taht the stomach was not checked.

2

. .

0

2. 0.

'

.,

"

ee

2

(

.

date time date time

##

opening towards tidal.

current up or down

species. SL L W

cm cm g

stomach dead, roe or sperm

first try: beach in front of station, at low tide line: substrate; sand, none seagrass.

2-21 19:00 2-22 7:00 1

2-21 19:00 2-22 7:00 L

1 down x

2 both x

second try: beach, just left of station, at low tide line: substrate; sand, none seagrass.

2-22 7:30 2-22 19:00 I

2-22 7:30 2-22 19:00 1

2-22 7:30 2-22 19:00 1

2-22 7:30 2-22 19:00 I

2-22 19:00 2-23 8:00 ¹ 2-22 19:00 2-23 8:00 I

2-22 19:00 2-23 8:00 ¹ 2-22 19:00 2-23 8:00 ^I 2-22 19:00 2-23 8:00 L

I up

2 down 3 up 4 down

I up

2 down 3 up

3 4 down

x x

Liza aurala 9,5 11,3 10

x

Soleasenegalensis ^13,5 15,2 20 x

Lizaaurala 10,2 12,4 10

Diplodussargus ^{8,3 10,1 10} Soleasenegalensis 16,5 19,5 60

x

x

x x x

third try: tidal fiat in front of station, at low tide line: substrate; muddy with Zostera.

— position: 19°52,999' N 16°17,408' W

2-23 10:00 2-23 18:30 ^{1 1} both x 2-23 10:00 2-23 18:30 1 2 down x

2-23 10:00 2-23 18:30 1 3 up x 2-23 10:00 2-23 18:30 I 4 down x

2-23 I0:00 2-23 18:30 I 5 up x

2-23 10:00 2-23 18:30 I 6 both Sciaenawnbra 41,5 49,5 1520 empty 2-23 18:30 2-24 9:00 ^{I I} both Pseziodesbelclieri 39,7 45,3 1020 full 2-23 18:30 2-24 9:00 I 2 down t

2-23 18:30 2-24 9:00 1 3 up x

2-23 18:30 2-24 9:00 ¹ 4 down Diplodussargus 20,4 25,6 230 empty 2-23 18:30 2-24 9:00 ¹ 4 Diplodussargus ^10,6 13,5 40 empty 2-23 18:30 2-24 9:00 1 4 Dicentrarclzuspunciatus 12,8 15,8 40 full 2-23 18:30 2-24 9:00 ^I 5 up Dicezztrarchuspunciaius 14,2 16,8 40 empty 2-23 18:30 2-24 9:00 I

2-24 9:00 2-24 18:30 I

2-24 9:00 2-24 18:30 I

2-24 9:00 2-24 18:30 I

2-24 9:00 2-24 18:30 I

2-24 9:00 2-24 18:30 1

2-24 9:00 2-24 18:30 ¹ 2-24 18:30 2-25 9:00 ¹ 2-24 18:30 2-25 9:00 1

2-24 18:30 2-25 9:00 ¹ 2-24 18:30 2-25 9:00 I

2-24 18:30 2-25 9:00 ¹ 2-24 18:30 2-25 9:00 L

6 both

I both

2 down 3 up

4 down 5 up 6 both

I both

2 down 3 up 4 down

5 up

6 both

Dicentrarchuspunciatus 13,1 15,9 30 x

x

callinectesniarginatus 5,87 x

.r x x x x x x

x____________________________________

full

placed at small creek on tidal fiat, in small creek: sediment 1,2,3; muddy with Zostera; ^4,5,6,7; sand.

location: 19°52,999' N 16° 17,408' \V

- ^__________

2-25 9:00 2-25 18:15 2-25 9:00 2-25 18:15 2-25 9:00 2-25 18:15 2-25 9:00 2-25 18:15 2-25 9:00 2-25 18:15 2-25 9:00 2-25 18:15 2-25 9:00 2-25 18:15 2-25 18:15 2-26 9:00

r

r

I both

2 up 3 down 4 up

5 down 6 both 7 both

I both