SCRS//. ANALYSIS OF THE SIZE STRUCTURE AND LENGTH-WEIGHT RELATIONSHIPS OF SWORDFISH (Xiphias gladius) CAUGHT BY THE MOROCCAN DRIFTNET FISHERY IN THE MEDITERRANEAN SEA DURING 7. N. Abid and M. Idrissi 1 ABSTRACT An analysis of the size frequencies distribution of swordfish caught by the Moroccan driftnet fishery during 7 was carried ou. The variability in the length-weight relationships of this species was also analyzed using a GLM approach. KEY WORDS Driftnet, swordfish, Mediterranean Sea, size frequencies, length-weight relationship, GLM. 1
1- Material and methods The size and weight data of swordfish for the year 7 were collected through the biological sampling network of the Regional Centre of the National Institute of Marine fisheries (INRH), at the port of Tangier located in the Strait of Gibraltar (Moroccan Mediterranean Sea). In 7, a total of 97 swordfishes were sized. Individual size and weight data were also obtained from 5 individuals to estimate the length-weight relationship for this species. As recommended by the SCRS, a conversion factor of 1.1 was used to convert the gutted weight (GW) into the round weight (RW). A General Linear Modelling approach (GLM) to analysis of variance (McCullagh, 199) was used to analyse the variability in the length-weight relationships estimated for the period -7, which is formulated as follow: RW=aLJj-FL b, the response variable analyzed was the natural logarithm of RW (kg). The effects of factors year, month and all the interactions between them were analyzed. The type III Wald tests was used to test the significance of all the factors implemented in the model at the 1 level (Dobson, 1999). -Results and discussions In 7, the size of the sampled fish ranges from cm to 37cm LJFL, corresponding to a round weight from 11 to kg (Table 1 and Figure 1). The mean size is about 15 cm LJFL, corresponding to 53 kg RW. The mean size and the mean weight increased by 7 and 3 with respect to, respectively. It is important to highlight that the average size of fish has generally shown an increase trend during the last 5 years as it was previously documented (Abid and Idrissi, ). In 7, the proportion of fish less than cm LJFL represents about 1 against 3 in, which means that the proportion of the immature fish decreased by compared with. Some differences in the size frequencies distribution can be noticed from a month to another (Figure 1). But according to the monthly analysis of the SWO size frequencies in the last years, no clear trend can be noticed. The type III Wald test indicates that all the factors are significant at the 1 level, including the interactions between them; which means that there are differences in the shape of the length-weight relationships depending on the month and the year. The differences affect both the slope and the origin of coordinates (Table and Figure ). These differences in the length-weight parameters indicate the variability in the condition factor of the fish according to the month and the year. These results are similar to those presented in the previous paper (Srour et al, ) 3-Literature cited ABID, N and M. Idrissi.. Analysis of the size data of swordfish (Xiphias gladius) caught by the Moroccan driftnet fishery operating in the Mediterranean Sea. Period 1999-. DOC SCRS //33 SROUR, A, N. Abid, J. M. de la Serna, J. Mª Ortiz de Urbina and D. Macías.. Preliminary analysis of the temporal and spatial variability in the length-weight relationship of the swordfish (Xiphias gladius) caught by the Moroccan drift net fishery. Col. Vol. Sci. Pap. ICCAT, 5(3): 99-931
DOBSON, A. J. 1999. An introduction to generalized linear models. Chapman and Hall/CRC McCULLAGH, P. and J. A. NELDER. 199. Generalized Linear Models, nd edition. Chapman and Hall, NY Table 1. Size frequencies distribution of swordfish caught by the Moroccan driftnet fishery in the Mediterranean during 7 LJ-FL (cm)/month 5 Total 1 1 5 3 11 1 1 5 9 115 3 15 9 57 13 5 15 79 135 1 5 1 7 9 1 15 3 15 5 3 3 155 1 3 17 1 3 15 1 1 17 1 175 11 13 1 3 1 19 3 15 3 9 1 19 1 9 1 195 1 13 7 13 5 1 1 3 3 15 3 5 3 5 5 3 1 3 1 1 35 1 1 3
Table. ANOVA Results for GLM analysis. Swordfish Length weight relationship for Moroccan driftnet fishery Class level information Class level values Year 5 7 Month 5 5 7 Number of observations 171 Dependent variable: lnrw F-statistic: 39 on 15 and 151 degrees of freedom, the p-value is R-Squared Coeff.var Resid.std.error LnRW.9553 3.97.1359 3. Type I Sum of Squares Df Sum of Sq Mean Sq F Value Pr(F) lnljfl 1 717.971 717.971 3911.1.e+ Year 3 1.377.57 5..e+ Month 3.395.999 9.71.e+ Year:lnLJFL 3.15.5.77.959e- Month:lnLJFL 1.355.339 1.5.e+ Type III Sum of Squares Df Sum of Sq Mean Sq F Value Pr(F) lnljfl 1 1.159 1.159 91.7. Year 3.97.1359 7.9.57571 Month 1.397.399 1.9335. Year:lnLJFL 3.39.1315.75.133 Month:lnLJFL 1.3555.339 1.5155.
1 1 April n=9 1 13 15 17 1 19 3 1 1 May n=57 1 13 15 17 1 19 3 August n=7 1 1 1 13 15 17 1 19 3 Total n=97 1 1 1 13 15 17 1 19 3 5
Figure 1. Monthly size frequencies distribution of swordfish caught by the Moroccan driftnet fishery in the Mediterranean Sea in 7.. 35. 3. 5. RW(kg). 15.. 5.. 5 7 9 1 13 15 17 1 19 3 5 7 5 7 RW=x - LJ-FL 3.13 RW=x -7 LJ-FL 3.55 5 RW=x - LJ-FL 3.3 RW=x - LJ-FL 3.3 Figure. Length-weight relationships for swordfish estimated for each year