168 



Fishery Bulletin 



1990 



(1987). Data from fishing surveys carried out between 

 April 1978 and April 1979 by German and Japanese 

 research vessels (Ciechomski et al. 1979a, Cousseau 

 et al. 1979) are also incorporated in the discussion. 



Monthly maps (not shown) of fishing effort and hake 

 catch-per-unit-effort (CPUE) drawn from the commer- 

 cial fishery data were used to track the migration of 

 exploited concentrations of hake. The use of commer- 

 cial fishing statistics to follow fish migration, however, 

 has limitations. The fleet may be targeting on species 

 other than the one being considered. Fleet movements, 

 therefore, may be primarily associated with perceived 

 fluctuations in the abundance of the target species. The 

 Argentine offshore fleet targets mainly on hake; the 

 proportion (in weight) of hake in the catches was >95% 

 during most of the period studied. The proportion of 

 shortfin squid (Illex argen.tmus) in the catches increased 

 during May-July as a result of seasonal overlap in the 

 distributions of both species. Even then, the Argentine 

 fleet did not generally target on squid, which could be 

 considered as bycatch (Juanico 1982). 



The use of fishery data to investigate migration relies 

 on the fishermen's ability to locate the densest aggrega- 

 tions of hake. Several lines of evidence support this 

 assumption. The hake fishery has been taking place for 

 several years and the displacements of fish are well 

 known. Monthly histograms of standardized hake 

 CPUE show a moderately positive skewness (data not 

 shown) that may be partly associated with a widespread 

 knowledge of fish distribution (Quinn 1985). A com- 

 parison of alternative formulations for a monthly 

 CPUE index, i.e., an average-of-ratios vs. a ratio-of- 

 averages, confirms that effort was concentrated on 

 areas of higher-than-average hake abundance (Otero 

 1986). This conclusion is also supported by spatial 

 statistics of fishing effort and CPUE (Rothschild and 

 Yong 1970). 



To visualize better the patterns of hake movement, 

 monthly latitudinal and longitudinal weighted averages 

 of fishing effort and hake CPUE were computed ac- 

 cording to the procedure described by Rothschild and 

 Yong (1970). In some months, fishing effort was 

 deployed in distinctly separate regions throughout the 

 shelf. This may have reflected true operational dif- 

 ferences in the fishing fleet as a consequence of separ- 

 ate centers of fish abundance. To take into account the 

 discontinuity, I grouped all contiguous 1-degree areas, 

 calculated the total fishing effort for each cluster of 

 contiguous areas, and considered only major clusters 

 (those in which effort was >5% of the effort in the most 

 heavily exploited cluster). Most months showed only 

 one valid cluster of areas and, at most, two clusters 

 fulfilled the previous criteria. I computed the spatial 

 statistics of effort for each valid cluster. For the CPUE 

 statistics, the same clusters delimited above were used. 



(T) Boenos Atres Slope Norlti 

 ^} Buenos Aires Slope Soulh 

 (?) Rio de 10 Ploto Moutli 



(4) Pologonian Slope Norlh 



(5) Pologonian Slope Soulh 



(b) Buenos Aires Coost 



(7) Intermediole Shelf 



(e) Southern Spawning 

 Grounds 



65* 



55*W 



Figure 1 



Geographic regions in southwL'stern Atlantic Ocean used for inves- 

 tigating interannual differences in hake migration. Stippled regions 

 are discussed in the text. 



but data were included only from areas where five or 

 more trips had been reported, to reduce variability in 

 CPUE values due to low levels of sampling. 



As with most hakes (Jones 1974), tagging Argentine 

 hake has proven so far to be impossible and, conse- 

 quently, there are no estimates of migratory speed 

 derived from tag returns of individual fish. As an ap- 

 proximation, I estimated hake migration speed by using 

 the spatial statistics of fishing effort calculated above. 

 I calculated the distances between the centers of ef- 

 fort in consecutive months, and then estimated daily 

 migratory speeds by dividing the distances by 30 days 

 (regardless of the specific months involved). 



If the migratory pattern of hake is associated with 

 environmental factors, year-to-year differences in their 

 distribution may reflect fluctuations in oceanic condi- 

 tions. I examined differences in the distributional pat- 

 tern of hake between 1978 and 1979, and tried to relate 

 those differences to environmental conditions. 



A direct comparison of the commercial fishery effort 

 and CPUE data is impossible because (a) there was a 

 change in the format of the fishing tickets (the forms 

 turned in by fishermen upon return to port) between 

 1978 and 1979, and Oi) fishing effort had to be stan- 

 dardized separately for each year because of software 

 limitations (see Podesta 1987). 



A new fishing ticket was introduced in 1979 which, 

 due to its design, increased uncertainty on the reported 

 location of effort. To reduce the possibility of error, 

 I grouped the data, originally reported for 1-degree 

 squares, into eight larger geographical regions (Fig. 1). 



Because fishing effort had been standardized separ- 

 ately for 1978 and 1979, the cumulative time density 

 approach introduced by Mundy (1982) was used, which 



