SCRS/2007/108 Collect. Vol. Sci. Pap. ICCAT, 62(4): 10811083 (2008)
Discards of undersized swordfish individuals
in the greek swordfish Fisheries
George Tserpes^{1}, Panagiota Peristeraki^{1}
SUMMARY
Data collected from observers onboard of commercial swordfish longliners from 20042006 were analyzed in order to estimate overall discard rates of undersized swordfish individuals as well as their temporal fluctuation. Results revealed that mean annual discard rates in terms of number of fish were in the range of 0.440.65individuals per fishing day and the probability to have discards increases in autumn and winter months. It was estimated that 45005000 undersized individuals are discarded annually, weighting around to 1517 tonnes.
RÉSUMÉ
Les données collectées auprès des observateurs embarqués à bord de palangriers commerciaux ciblant l’espadon de 2004 à 2006 ont été analysées afin d’estimer les taux de rejets totaux d’espadons soustaille ainsi que leur fluctuation temporelle. Les résultats révèlent que les taux de rejets annuels moyens, en nombre de poissons, s’élevaient à 0,440,65 spécimens par jour de pêche et que la probabilité de présence de rejets augmentait pendant les mois d’automne et d’hiver. Il a été estimé que 4.5005.000 spécimens soustaille sont rejetés chaque année, représentant 1517 tonnes.
RESUMEN
Se analizaron los datos recopilados por los observadores embarcados en palangreros comerciales dirigidos al pez espada durante 20042006 con el fin de estimar las tasas de descartes globales de ejemplares de pez espada de talla inferior a la regulada, así como su fluctuación temporal. Los resultados revelaron que las tasas medias anuales de descartes, en términos de número de peces, se encontraban en el rango de 0,440,65 ejemplares por día de pesca y la probabilidad de que se produzcan aumentos en los descartes en los meses de otoño e invierno. Se estimó que anualmente se descartan 4.5005.000 ejemplares de talla inferior a la regulada, lo que equivale a un peso de aproximadamente 1517 toneladas.
KEYWORD
Swordfish, Mediterranean, discards
Swordfish (Xiphias gladius) is a large pelagic species of high commercial value heavily exploited all over the Mediterranean basin. Greece is among the most important swordfish producers in the Mediterranean having an annual production of 16001800mt in the latest years. Greek swordfish fisheries exploit the eastern part of the Mediterranean basin covering a large area, extending from the east Ionian to the Levantine Seas. Fishing is carried out using drifting surface longlines through February to September while is prohibited by a national regulation from October to January. Aiming to the protection of juveniles, landings of fish having Lower Jaw Fork Length (LJFL) less than 100cm is prohibited through a national regulation. This in fact has resulted in discarding of a certain volume of zeroclass fish.
Since 2003, discards are monitored through regular observations onboard of commercial vessels. The aim of the present work is to examine temporal fluctuations in the discards rate and provide estimates on the total volume of swordfish discards of the fishery.
2. Materials and methods
Since 2003, in the frames of the National Data Collection Project (NDCP), regular monitoring of the Greek swordfish fishery, including seasonal observations on board of commercial longliners, has been implemented. In the present study, we analyse data from onboard observations covering 196 fishing days (sets) on board of 18 different longline vessels during three fishing periods (20042006). The data that were analyzed included information regarding the presence or absence of swordfish discards on a set (daily) basis.
The obtained Bernoullitype 0/1 measurements (0 for absence, 1 for presence) were analysed by means of Generalised linear modelling (GLM) techniques (McCullagh and Nelder 1983) assuming a binomial error structure model and a logit link function. The model included year and month as main effects. Month was considered as a continuous variable and as preliminary analysis showed that its functional relationship with the absence/presence of discards was not sufficiently explained through a linear relationship, a polynomial cubic spline was used for smoothing the “month” effect; thus the linear model was extended to a generalized additive model (GAM  Hastie and Tibshirani, 1990) with parametric terms.
Model fitting was accomplished with the statistical package Splus, following the approaches described by Venables and Ripley (1997).
3. Results
Mean annual discard rates fluctuated from 0.440.65 individuals per fishing day. Discard rates were higher at the beginning and end of the fishing season, i.e. February and September respectively (Table 1). The analysis of deviance of the GLM model demonstrated that only the month effect was significant on the 95% level (Table 2). The effect of month on the probability of having discards is shown on (Figure 1). Model estimates indicate that the probability is negligible from MayJune, while has its higher rate (86%) in February (Table 3).
4. Discussion
Our results clearly indicated that discard rates vary significantly among months. The probability of having discards start to increase after the end of the reproduction period (July) and reaches up to 86% in February. In that month there have been observed about 2 discarded individuals per fishing day. Unfortunately, there no data from October to January, as the fishery is closed aiming to the protection of juveniles. However, it is natural to assume that discard rates will be quite high in those months too. Based on the fact that the total annual effort of the Greek swordfish fishery is around to 85009500 fishing days it is estimated that 45005000 individuals are discarded annually, weighting 1517 tonnes.
References
McCullagh, P. and J.A. Nelder. 1983. Generalized Linear Models. Chapman and Hall, London.
Venables, W.N. and B.D. Ripley. 1997. Modern Applied Statistics with SPLUS, Second Edition. Springer.
Hastie, T. and R. Tibshirani. 1990. Generalized Additive Models. Chapman and Hall, London.
Table 1. Mean discard rate (number of fish per fishing day) by month.
Month

Discard rate

2

1.91

3

0.38

4

0.00

5

0.28

6

0.00

7

0.28

8

0.09

9

0.70

Table 2. Analysis of deviance table for the GLM model fitted to the presenceabsence of discards. (d.f. =degrees of freedom, Res. =Residual).
Source of variation

d.f.

Deviance

Res. d.f.

Res. Deviance

Probability (Chi)

Null



195

226.78


Year

2

2.40

193

224.39

0.30

Month

3

46.56

190

177.82

0.00

Table 3. Monthly estimates of the probability to have discards.
Month

Probability

2

0.862716733

3

0.296078581

4

0.090326401

5

0.058589457

6

0.071482698

7

0.121468416

8

0.209480339

9

0.285739698

Figure 1. GLM derived effect of month on the probability to have discards. The fitted values are adjusted to average zero and the broken lines indicate two standard errors. Months 29: FebruarySeptember.
