Kotwicki et al.: Variation in the distribution of Theragra cholcogramma 



581 



biomass was particularly low in clusters Fl, F2, and F3 

 for that year. In cluster Fl we observed an increase in 

 biomass from 10% to 51%. 



For pollock 21-29 cm, changes between cluster G2 

 and Gl resembled changes between clusters F2 and Fl. 

 The percentage of total biomass in these two clusters 

 changed from 57% to 0% and from 11% to 94%, respec- 

 tively (Fig. 2B). 



A slightly different situation was observed for pol- 

 lock 30-39 cm (Fig. 2C). We identified three clusters 

 of decreasing biomass with temperature: H2, H3, and 

 H4 located, respectively, northwest of Zhemchug Can- 

 yon, northwest and east of the Pribilof Islands. Overall 

 predicted biomass change in H2, H3, and H4 decreased 

 from 65% to 2%. The offset for this negative change was 

 found in cluster HI, where we noted a positive change 

 from 4% to 75%. 



Areas with decreasing fish biomass for pollock 40-49 

 cm were located within cluster 12 (Fig. 2D). Biomass de- 

 creased from 58% in the coldest year to 4% in the warm- 

 est year. We observed temperature-related increases in 

 biomass mostly north of 12 in cluster II (0%-49%). 



A quite different situation was observed for pollock 

 >50 cm (Fig. 2E). Although pollock of this size seemed 

 to concentrate northwest and northeast of the Pribilof 

 Islands (similar to pollock 30-49 cm) during cold years; 

 in warm years they were found in EIT surveys mainly 

 in the southeast, as opposed to the smaller fish that are 

 found mainly in the north. Results for pollock >50 cm 

 should be treated cautiously because only a very small 

 part of the entire population of pollock this size can be 

 detected with the EIT survey (Ianelli et al. 3 ). Because 

 of the benthic habits of pollock >50 cm (Shuntov et al., 

 1993), most were detected in BT surveys. 



Overall, our analysis of EIT survey data indicated 

 a northward temperature-related shift of 50-80% of 

 pollock <50 cm in two major areas. With increasing 

 temperature, the density of pollock <40 cm decreased 

 northwest of Zhemchug Canyon in a large area at 100 m 

 to 200 m depths. Similarly, the density of pollock 30- 

 49 cm decreased northwest of the Pribilof Islands. Off- 

 setting these decreases, pollock density increased in the 

 northernmost area of the survey (close to the U.S. -Rus- 

 sia Convention Line). 



Although the direction of the shift was the same for 

 all length categories up to 50 cm, the mean distance 

 between the clusters with negative slopes and clusters 

 with positive slopes increased with fish size (Table 2). 



Bottom trawl survey 



For pollock <20 cm, we observed a decrease in pollock 

 biomass with temperature in cluster A3 covering the 



3 Ianelli, J. N., T. Buckley, T. Honkalehto, N. Williamson, 

 and G. Walters. 2001. Bering Sea-Aleutian Islands wall- 

 eye pollock assessment for 2002. In Stock assessment and 

 fishery evaluation report for the groundfish resources of the 

 Bering Sea/Aleutian Islands regions, p. 1-105. North Pac. 

 Fish. Manag. Council. Anchorage, AK. 



area west of the Pribilof Islands and north to Zhemchug 

 Canyon (Fig. 3A). We observed an increase in pollock 

 biomass in shallower areas north of Pribilof Island 

 (A4), as well as in the areas of 50-100 m depth east 

 from the Pribilof Islands (A5). The magnitude of change 

 was somewhat smaller than that observed for the EITS 

 survey (see Fig. 3A for details). 



For pollock 20-29 cm, we observed a decrease in 

 biomass from 70% to 5% in the area northwest of the 

 Pribilof Islands (cluster B5). A cumulative increase in 

 biomass from 7% to 52% of total biomass was observed 

 in clusters Bl and B4 north of B5, and in clusters B6 

 and B7 in shallower waters (Fig. 3B). Relatively weak 

 relationships were found between pollock biomass and 

 temperature for clusters B2, B3, and B8. 



For pollock 30-39 cm, we observed a temperature- 

 related decrease in biomass in clusters C2 and C4 (42% 

 to 1%, and 68% to 12% accordingly) (Fig. 2C). Increase 

 in biomass was observed in cluster CI (2-32%) north 

 from C2. Positive change was also observed in cluster 

 C5 (0-17%) within the shallow (<100 m) part of the 

 southeastern Bering Sea shelf. 



Clusters D2 and D4 represented areas where we ob- 

 served a significant decrease in biomass for pollock 

 40-49 cm (from 5% to 1%, and from 41% to 11%) (Fig. 

 3D). Increased biomass was detected in cluster Dl lo- 

 cated north from D4 and in D5 located to the east of 

 D4 in shallower waters. 



Very small changes were detected for pollock >50 cm. 

 Although three clusters had a relatively strong pollock 

 biomass and temperature relationship, the magnitude 

 of biomass changes within the range of observed tem- 

 peratures was quite small (Fig. 3E). 



Overall, as with the EIT surveys, northward shifts 

 in distribution in warmer years were found in the BT 

 survey data for pollock <30 cm. The magnitude of these 

 northward shifts was somewhat smaller (15-30%) than 

 those detected by EIT surveys. In addition, these data 

 suggested an inshore eastward redistribution of pollock 

 in warmer years. Changes for pollock >50 cm were evi- 

 dent but small (in the range of 15%). 



Discussion 



Inferring seasonal pollock migration from interannual 

 variations in distribution 



Interannual differences in the timing of the migration 

 from spawning grounds to forage areas are related to 

 water temperatures. The relationship between tem- 

 perature and the spatial distribution of a seasonally 

 migrating species could represent either a change in the 

 winter location of the stock or a change in the timing of 

 the migration or both (Mountain and Murawski, 1992). 

 Although the evidence is not conclusive, data suggest 

 that most pollock populations spawn in late winter 

 or early spring in the same locations year after year 

 (Bailey et al., 1999a). For example, large, prespawn- 

 ing aggregations of pollock have been surveyed around 



