Kotwicki et al : Variation in the distribution of Themgro chalcogrommo 



583 



and 12 to clusters HI and II (368.3 km and 453.7 km, 

 respectively). Similar size-dependent differences in the 

 distance of seasonal migrations were reported for Pacific 

 hake (Merluccius productus), another gadoid from the 

 north Pacific (Dorn, 1995). These observations may 

 support the length-based hypothesis of Nottestad et al. 

 (1999) for feeding migrations in pelagic fish. Focusing 

 on the energetic cost-benefit relationship of long distance 

 migration, they concluded that migration distance is a 

 function of length, weight, and age. Smaller fish may 

 undergo shorter feeding migrations because the ener- 

 getic cost of migration can exceed their total energy 

 intake resulting from the of greater hydrodynamic drag 

 associated with smaller fish size. 



Migrations of the largest pollock (>50 cm), detected 

 from the BT survey data, were of much lower magnitude 

 then those of smaller fish. Our models indicate that only 

 about 15% offish in this length category move between 

 clusters in the northeastern direction toward shallower 

 waters. These small changes detected in BT data con- 

 tradict those seen in EIT data. Whereas a small north- 

 ward shift in biomass (mostly from cluster E2 to cluster 

 E4) was detected with BT data, a southeastward shift 

 was detected with EIT data. However, because the EIT 

 survey is not well suited for estimating the distribution 

 of pollock >50 cm, we are inclined to put more weight 

 on the BT data to explain temperature-related changes 

 in biomass distribution for this length category. Larger 

 pollock (>50 cm) appear to change their migratory be- 

 havior. Shuntov (1992) noticed that the distribution 

 of larger pollock (>54 cm) fundamentally differs from 

 that of smaller pollock and that larger pollock are more 

 benthic in behavior and feeding. Stepanenko (2001) did 

 not observe any migrations to the Russian zone for pol- 

 lock six years or older. We propose that the difference 

 in the migratory behavior between pollock <50 cm and 

 pollock >50 cm is linked to a well-known shift toward 

 a diet offish with increasing pollock size (Bailey and 

 Dunn, 1979; Dwyer et al., 1987). 



Why do pollock migrate? 



Pollock feeding migrations in the EBS may be driven by 

 a combination of four factors: temperature, zooplankton 

 production, currents, and length of daylight. 



Changes in the water temperature may affect pol- 

 lock migrations. Bottom water temperature over the 

 Bering Sea shelf rises between April and September 

 (Pavlov and Pavlov, 1996; Overland et al., 1999; Khen 

 et al., 2001; Stabeno et al., 2001). Our results indicate 

 that with rising temperature pollock generally migrate 

 northward and inshore. Pollock appear to avoid tem- 

 peratures below 0°C (Swartzman et al., 1994); therefore 

 a seasonal increase in temperature above 0°C can open 

 new geographic areas for migration. Temperature was 

 presented as one of several important stimuli affect- 

 ing fish movements by Harden Jones (1968) and by 

 Wielgolaski (1990), who noticed that capelin (Mallotus 

 villosus), Atlantic cod (Gadus morhua), and haddock 

 (Melanogrammus aeglefinus) in the Barents Sea migrate 



north towards a preferred temperature, either directly 

 to satisfy metabolic requirements, or indirectly, as when 

 attracted by food organisms. 



Seasonal patterns in zooplankton production and prey 

 availability largely coincide with seasonal patterns in 

 pollock migration and distribution. The role of food 

 availability in driving fish-feeding migrations has been 

 described for other zooplanktivores such as Pacific hake 

 (Dorn, 1995), Atlantic herring (Clupea harengus), blue 

 whiting iMieromesistius poutassou), mackerel (Scomber 

 scombrus) and capelin (Nottestad et al., 1999). In the 

 Bering Sea, the abundance of zooplankton is high on 

 the EBS and NBS shelf throughout spring and sum- 

 mer, but it remains high in autumn only in the NBS 

 (Springer et al., 1989; Chuchukalo et al., 1996; Coyle 

 et al., 1996). Copepods and euphausiids are major prey 

 groups for pollock during spring and summer in the 

 northwest area of the EBS shelf, but in autumn, 30-49 

 cm pollock increase their feeding on fish and decapods 

 (Dwyer et al., 1987) which may be related to a decrease 

 in the availability of these prey (Willette et al., 1999) in 

 this area. Further north in the Navarin-Anadyr area, 

 copepods and euphausiids remain major prey compo- 

 nents in the diet of pollock <50 cm through summer and 

 autumn (Shuntov et al., 2000). The migration pattern 

 of pollock indicates they may follow their food supply as 

 the production and abundance of zooplankton proceeds 

 northward. Pollock larger than 50 cm do not undergo 

 northward feeding migrations because small pollock, 

 other fish, and benthos are the main components of 

 the diet (Dwyer et al., 1987; Yoshida, 1994; Shuntov 

 at al., 2000). 



In the area of pollock migrations northwest of Pribilof 

 Islands current speeds are in the range of 1-5 cm/s at 

 the 100 m depth and they generally run in the north- 

 west direction (Stabeno et al., 2001). Current direction 

 coincides with the direction of pollock migrations, so 

 that the cost of the migration may be offset by swim- 

 ming in the same direction as the transporting cur- 

 rent (Nottestad et al., 1999). Water currents can also 

 influence fish migration indirectly by providing visual 

 stimuli arising from the moving background (Harden 

 Jones, 1968) or by transporting food. Springer et al. 

 (1989) suggested that the transport of zooplankton by 

 the northwest current may cause greater levels of zoo- 

 plankton concentration in the NBS. Because of the lack 

 of data on current speed, he speculated that a current 

 velocity in the range of 20 cm/s was needed to explain 

 these high levels of zooplankton in the NBS if the high 

 levels of zooplankton are based only on currents. The 

 latest observations of current on the Bering Sea shelf 

 do not support these hypotheses (Stabeno et al., 2001). 

 However northwestern currents may contribute to high- 

 er zooplankton biomass in the NBS. 



Nottestad et al. (1999) suggested that light conditions 

 may play a role in fish feeding migrations because dur- 

 ing summer day-length increases the farther north fish 

 travel, thus potentially increasing feeding duration for 

 pelagic visual predators. Pollock are visual predators 

 and light conditions affect feeding efficiency of pollock 



