212 



Fishery Bulletin 90(1). 1992 



10.3 



9.9 



ns 

 o 



15 



c 



§ 9.1 



8 7 



8.3 



2 4 6 8 10 



Fishing Effort in 10000 HP 



12 



Figure 2 



Regression line (solid) and 95% and 99% confidence limits 

 (dotted) of the log annual commercial catch of anchovy and 

 fishing effort in horsepower (HP), in Hellenic waters, 1964-87. 



trend in monthly catches [Ln(monthly catch) = 5.38 + 

 0.0053T, n 288, r 0.38, p<0.01, where T = 1-288] is 

 most likely attributed to increased fishing effort. 



The spectrum of the resulting series (not shown 

 here), which may be postulated to be free of any an- 

 nual changes in effort, revealed a large major peak at 

 12 months (frequency 0.0833). This marked seasonal 

 pattern is most likely related to the seasonal offshore- 

 inshore migrations of anchovy and the nature of the 

 purse-seine fishery (Stergiou 1990a). Purse-seine fish- 

 ing in Hellenic waters does not occur in the open sea 

 but is mainly restricted to coastal areas where schools 

 of anchovy migrate seasonally. The anchovy starts its 

 inshore migration in early spring, but peak abundance 

 occurs in coastal waters in May-August. Offshore 

 migration probably occurs in late summer-fall. 



The smoothed spectrum of the seasonally corrected 

 and detrended series (Fig. 3) reveals a prominent peak 

 at 4.6 years (frequency 0.018) and a probable second- 

 ary peak at 1.9 years (frequency 0.043) (95% confidence 

 intervals of the spectrum for 10 df: 0.4882-3.0798 

 squared amplitude of sinusoids). In contrast, non- 

 sinusoidal periodic variability generates harmonics with 

 ■periods of less than 1 year (Fig. 3). 



Cycles of 2-3 and 4-5 years have also been identified 

 in the air temperature in the northern (Thessaloniki) 

 and western Aegean (Athens) (Table 1) and in different 

 biotic (zooplankton, phytoplankton, fish eggs/larvae, 



fish) and abiotic variables (air temperature/pressure, 

 sea temperature/salinity) in different areas of the 

 Mediterranean, Black, and Azov Seas (Table 1). These 

 cycles have also been suggested for annual anchovy 

 catches and eggs/larvae, temperature, salinity, and 

 zooplankton in the Adriatic Sea but the data set is 

 limited (annual, 1962-76) and the cycles may not be 

 statistically significant (D. Regner 1985). Correlations 

 have been found between biotic/abiotic variables (pri- 

 mary production, zooplankton biomass, winds, river 

 flow, air/sea temperature, salinity) and various abun- 

 dance indices of the Mediterranean anchovy (Azov- 

 Black Sea: Dement'eva 1987, Dekhnik and Rass 1988, 

 Porumb and Marinescu 1979; Hellenic waters: Ster- 

 giou 1986b; Adriatic Sea: S. Regner 1985; western 

 Mediterranean: Palomera and Lleonart 1989) and 

 other species oiEngraulis (see Bakun 1985). 



Cycles with periods of 2-4 and 4-7 years have also 

 been identified in the physical environment and marine 

 populations in other areas of the world (e.g., Kort 1970, 

 Shuntov et al. 1981, Colebrook and Taylor 1984, Mysak 

 1986). Such cycles have frequently been related to 

 short-term ocean-atmosphere interactions (e.g., surface 

 heat-exchange phenomena: Zupanovich 1968, Cole- 

 brook and Taylor 1984; advection: Kort 1970, Mysak 

 1986). 



A comprehensive discussion of the mechanisms 

 underlying such variability requires adequate biological 



