582 



Fishery Bulletin 103(4) 



Bogoslof Island every year since 1988 in the winter 

 (Honkalehto et al. 4 ). Further support that temperature 

 is related to the timing of the postspawning migration 

 may come from temperature effects on physiological 

 aspects of spawning. Cold water temperatures may delay 

 the onset of spawning and extend the spawning period 

 of walleye pollock as has been found for another gadid 

 (Kjesbu, 1994) and for flatfish (Lange and Greve, 1997) 

 in the Atlantic. 



The surveyed distribution of pollock in warmer years 

 should be more representative of that seen later in a 

 typical spring-summer warming cycle than the distri- 

 bution of pollock seen in colder years. Bottom tempera- 

 tures generally increased over the EBS and northern 

 Bering Sea (NBS) during spring and summer (Overland 

 et al., 1999; Khen et al., 2001; Stabeno et al., 2001). 

 Our results show that the warmer the bottom water 

 during spring-summer groundfish surveys, the farther 

 away pollock <50 cm are found from their major spawn- 

 ing grounds. Thus, we interpret areas having lower 

 pollock density with increasing temperature (clusters 

 with negative slope) to be areas from which pollock are 

 emigrating, and areas having higher pollock density 

 with increasing temperature (clusters with positive 

 slope) to be areas to which pollock are immigrating 

 (Figs. 2 and 3). 



Routes and directions of the migrations 



As the water warms during spring and summer, pol- 

 lock generally migrate northward, northwestward, and 

 inshore to shallower waters. Larger pollock (>30 cm) 

 begin their feeding migration from spawning grounds. 

 In many areas (white areas — Figs. 2 and 3) we did 

 not detect a significant increase or decrease in pol- 

 lock abundance in relation to temperature, e.g., in the 

 major pollock spawning area north of Unimak Island 

 (Hinckley, 1987; Bulatov, 1989), and this finding may 

 indicate that migration progressed beyond this area 

 before it was surveyed, even in the coldest years, or that 

 migrations were not pronounced in this area. However, 

 we observed a very large decrease in biomass with 

 increasing temperature in the Pribilof Islands area (i.e., 

 within clusters A3, B5, C4, D4, E2, H3, and 12), which is 

 another important pollock spawning location (Maeda and 

 Hirakawa, 1977; Hinckley, 1987; Bulatov, 1989; Bailey 

 et al., 1999a). An offsetting increase in biomass was 

 observed in the northernmost part of the survey area 

 (clusters Bl, CI, Dl, Fl, Gl, HI, and ID and in shallower 

 waters (clusters A4, A5, B6, B7, C5, and D5), which may 

 indicate that pollock migrate north and inshore during 

 the warming season. Echo integration trawl data indi- 



Honkalehto, T., N. Williamson, D. Hanson. D. McKelvey, and 

 S. de Blois. 2002b. Results of the echo Integration-trawl 

 survey of walleye pollock (Theragra chalcograma) conducted 

 on the southeastern Bering Sea shelf and in the southeastern 

 Aleutian Basin near Bogoslof Island in February and March 

 2002. AFSC Processed Report 2002-02, 49 p. Alaska Fish. 

 Sci. Cent., NOAA, Natl. Mar. Fish. Serv., 7600 Sand Point 

 Way NE, Seattle, WA 98115. 



cate that smaller pollock (<29 cm) probably begin their 

 migration from overwintering areas (clusters F2 and 

 G2) located mainly northwest of the Zhemchug Canyon. 

 These results agree with observations made by Bailey 

 et al. (1999b) that small age-0. age-1, and age-2 pollock 

 are distributed farther north than larger age-3 and older 

 pollock. Migrations continued generally northward to 

 the U.S. -Russia Convention Line. The near-bottom part 

 of the pollock population (detected in the BT survey) 

 also migrates northeastward into shallower waters. At 

 this point we cannot describe the exact starting and 

 ending points of migration but only the general direc- 

 tion, because surveys are performed after most of the 

 spawning has been completed, and we lacked data for 

 the NBS, where part of the pollock EBS population is 

 probably migrating. 



The direction of movements indicated by the EIT 

 survey data and the BT survey data were somewhat 

 different because of the effect of depth on the avail- 

 ability of pollock to each survey. As pollock migrate into 

 shallower water they become more available to the BT 

 survey and less available to the EIT survey. Therefore 

 the BT survey indicates greater movement into shal- 

 lower water, whereas the EIT survey indicates greater 

 movement in a northerly direction. 



Seasonal migrations by pollock in the EBS are broad- 

 ly recognized as occurring but have not been well sub- 

 stantiated; however, most of the general observations 

 and descriptions are in agreement with our results. It 

 is generally recognized that the feeding migration of 

 some EBS pollock takes them northwestward beyond 

 our survey area and into Russian waters (Shuntov et 

 al., 1992; 1993; Stepanenko, 2001). Pola (1985), in her 

 numerical simulation of pollock migrations in the EBS 

 identified two types of pollock feeding migration. One 

 was temperature induced in the northward direction, 

 and the other was seasonal in the northeastern direc- 

 tion toward shallower waters. Shuntov et al. (1993) 

 considered migrational activity to start with the on- 

 set of sexual maturity, but our findings indicate that 

 immature pollock do undergo feeding migrations in a 

 northwestward direction, but over shorter distances 

 than those traveled by mature pollock. Stepanenko 

 (2001) also recognized migration by immature pollock. 

 Only a few pollock tagged in the EBS have been recov- 

 ered (Yoshida, 1985), but the relationships between the 

 release and recovery locations are consistent with our 

 findings of a northwestward feeding migration during 

 the spring and summer over most of the EBS shelf and 

 a northeastward migration into shallower water on the 

 southeast EBS shelf. 



Length-based differences in migration patterns 



Our analysis of the EIT surveys indicates that the 

 migrations of pollock <30 cm are shorter than those of 

 pollock 30-50 cm. The distance pollock need to cover 

 from clusters F2 and G2 to clusters Fl and Gl (241.3 

 km and 217.5 km) is much shorter than the distance 

 to be covered by larger fish from clusters H4, H3, H2, 



