Appendix 9
A NOTE CONCERNING FISH COMMUNITIES IN SOUTH MOROCCO (21–26°N)

by

F. Mennes
Associate Expert
Institut Scientifique des Pêches Maritimes
UNDP/FAO Project MOR/81/002
Casablanca, Morocco

INTRODUCTION

Since 1980 the ISPM conducted a series of trawl surveys towards the South of Morocco. The boundaries of the area have been 0–100 meters depth and 21–26°N, and some results have been presented in Mennes (1984).

This note tries to clarify the interspecific relationships and to identify species communities.

METHODS AND RESULTS

The research area has been divided into 11 strata, and for each stratum the average catch per hour by species (- group) has been calculated (Mennes, 1984). Without the pelagic species and species which are rare, 36 species (- groups) remain.

A HP 9845B computer and a software package (Hewlett Packard, 19) have been used for the treatment of the data.

For each group, for each stratum, the average catch per hour has been entered, giving a 42 x 11 matrix. Correlations have been determined for all taxonomic pairs, giving a 41 × 41 correlation matrix (Table 1).

Species (- groups) with a correlation coefficient larger than or equal to 0.8 have been selected, and plotted in a circle.

For each correlation coefficient larger than or equal to 0.8 a line has been drawn (Figure 1). The figure shows that “Communities” can be distinguished, composed of 36 species (- groups) (Table 22). All communities are completely separated from each other and only Conger links 2 communities. The communities have been compared with the relative density plots from Mennes (1984). These plots are given here in Figure 2. A strong correlation between these density plots and the communities exists. For each community a stratum can be assigned. The result is given in Figure 3.

DISCUSSION

The method used is subjective because a certain value (here 0.8) determines if a correlation line will be drawn or not. However, several other levels of correlation have been tried and confirm the classification. For example, the result of using the level 0.75 highly increases the number of “good” correlations of the community of stratum 11, whilst only stratum 2 and 3 become more linked (Conger, Sparus aurata with Sparus pagrus, Diplodus cervinus with Sparus pagrus).

For the strata 4, 5, 6 and 8 no typical community can be assigned, probably they contain a mixture of the adjacent strata, or the results have been influenced by migration.

A large scale approach of the distribution of species (- groups) is given by Garcia (1982). Here the terms “Pagellus acarne complex” and “Pagellus bellottii complex” are given. For the small scale approach presented in this note, these terms do not seem to be justified, because these 2 species do not play an important role as far as correlations are concerned. Both species are relatively abundant, just as is Diplodus bellottii (= D. senegalensis), but they are also relatively evenly distributed over the areà and do not associate with any stratum. With the stratification used in this note these species serve as (loud) noise only.

This doest not mean that a large scale approach can not give as a result that Pagellus acarne, P. bellottii or Diplodus bellottii are important for the identification of a (large) area : the results of this type of correlation depends on the stratification chosen.

REFERENCES

Garcia, S., 1982 Distribution, migration and spawning of the main fish resources in the Northern CECAF area. CECAF/ECAF series 82/85, 9pp.

Hewlett-Packard Company, 1979 Regression Analysis, 209pp.

Mennes, F., 1984 The fish population of South Morocco : A multi-species stock assessment. Mimeo, 162pp.

CORRELATION MATRIX

Table 1:

Figure 1 :

Table 2: Species composition of communities, by stratum. Number of correlations with r>=0.8 is indicated.

STRATIFICATION

Figure 2 :

Relative density plot.
Light = relative low catch/hour.
Dark = relative high catch/hour.
Casablanca, 25.385.

Figure 3 : The stratification and the communities that identify the strata.