Shimada and Kimura: Seasonal movements of Gadus macrocephalus 



801 



ment was to evaluate the feasibility of a tag- recapture 

 program for Pacific cod and walleye pollock 

 (Shimada 1 ). During the second year, as tag-recovery 

 information accumulated, field efforts focused exclu- 

 sively on Pacific cod because few tagged walleye pol- 

 lock were recovered. Growth-increment data obtained 

 from these tag returns have been analyzed by Kimura 

 et al. (1993). This paper presents new information 

 on the seasonal movements and long-range migra- 

 tion of Pacific cod in the eastern Bering Sea and ad- 

 jacent waters. 



Materials and methods 



Between 1982 and 1990, Pacific cod were success- 

 fully captured and released from AFSC-chartered 

 fishing vessels engaged in summer bottom trawl sur- 

 veys off Alaska. Pacific cod were tagged throughout 

 their eastern Bering Sea distribution (Fig. 1). This ef- 

 fort was augmented by tag releases from cooperating 

 Japanese, Korean, and U.S. research vessels operat- 

 ing in the Aleutian Islands and Gulf of Alaska (Fig. 1 ). 



Capture gear included bottom trawls, pots, and 

 hook-and-line. For the bottom trawl, predetermined 

 stations were sampled each year across the eastern 

 Bering Sea shelf (Bakkala, 1993). Thirty-minute 

 trawl hauls and biological samplings were performed 

 at each station. Occasional opportunistic hauls of 10 

 to 30-minute durations were made to obtain addi- 

 tional tag releases. On retrieval of the trawl net, 

 Pacific cod were taken from the unprocessed portion 

 of the catch and placed in on-deck holding tanks sup- 

 plied with running sea water. After a recovery pe- 

 riod (typically 1-2 h), fish were removed with a dip 

 net and examined for visible signs of injury or stress. 

 Those not seriously harmed during capture were 

 placed in a padded cradle, tagged, and measured for 

 fork length to the nearest 0.5 cm. General condition 

 was noted, and fish were quickly returned to the sea. 



Two tag types were used in this study: 3.5-inch 

 anchor tags and 8-inch lock-on spaghetti tags. 2 Both 

 types were constructed from international orange #20 

 vinyl tubing and labeled with a tag serial number 

 and return address. The majority of releases were 

 made with the spaghetti tag (69%). This tag was 

 applied through the dorsal musculature, behind the 

 head and anterior to the first dorsal fin, with a hol- 

 low needle applicator and secured by interlocking 



1 Shimada, A. M. 1982. Cruise results, NOAA ship Chapman, 

 Cruise CH-82-06, 7 p. Available: Alaska Fish. Sci. Cent., NOAA, 

 NMFS, 7600 Sand Point Way NE„ Bin C15700, Seattle, WA 

 98115-0070. 



2 Floy FD68BC anchor tag and FT-4C lock-on spaghetti tag. 



plastic terminals. Anchor tags were inserted between 

 individual fin rays at the base of the first or second 

 dorsal fin. 



Pacific cod were tagged across the entire size range 

 available to the capture gears (Fig. 2); priority was 

 placed on the release of fish less than 55 cm (i.e. 

 younger than about age 5 yr; conversion from length 

 to age in this paper is based on growth data in 

 Kimura and Lyons [1990]). Pacific cod smaller than 

 about 42 cm (i.e. age 3 yr or younger) were not re- 

 cruited to the commercial fisheries and were of par- 

 ticular interest as the preexploited population com- 

 ponent; however, owing to availability larger fish 

 made up the majority of tagged fish. Data recorded 

 at release included date, haul or set number, gear 

 type, depth fished, bottom water temperature, re- 

 lease position, fork length, and general fish health. 



Tag information from Pacific cod recovered by com- 

 mercial trawl, longline, and pot fisheries (Table 1) 

 through the first quarter of 1992 were used in this 

 paper. The Bering Sea groundfish fishery was domi- 

 nated by foreign fishing until 1987, when harvest 

 allocations shifted to joint- venture and domestic fish- 

 eries. By early 1991, the Pacific cod fishery had 

 evolved into an exclusively U.S. enterprise. Mirror- 

 ing this transition in fleet involvement, tag recover- 

 ies initially came from foreign and joint- venture trawl 

 and foreign longline fisheries. More recent tag re- 

 turns have come from the domestic trawl and longline 

 fisheries. Tag recovery reports provided information 

 regarding capture date, catch location, and body 

 length. Some tag returns also included capture 

 method, depth fished, sex, body weight, maturity, and 

 a collection of scales or otoliths. 



Tag recovery data were analyzed by three comple- 

 mentary methods: 1) mapping, which described the 

 location of release or recovery or the movement of 

 fish from the area of tagging to the area of recovery; 

 2) multiway contingency table analysis (Fienberg, 

 1977), which was used to analyze the strength of re- 

 lationships between the season and area of tagging 

 and the season and area of recovery; and 3 ) direct popu- 

 lation dynamics modeling of the tagged population. 



Three primary areas of interest (Fig. 3) were de- 

 fined for use in the mapping, contingency table analy- 

 sis, and the population dynamics model: 1) the inner 

 shelf from depths <30 m to 100 m [Area 1]; 2) the 

 outer shelf between >100 m and 200 m and incorporat- 

 ing the upper continental slope at depths greater than 

 200 m [Area 2]; and 3) winter spawning grounds sur- 

 rounding Unimak Pass and adjacent waters [Area 3]. 



To emphasize the main features in the data, 

 release-recovery positions were plotted individually 

 (Fig. 4) and as mean-movement vectors based on tags 

 released over blocks of 2° latitude x 5° longitude (Fig. 



