366 



Fishery Bulletin 103(2) 



Timing of peak zooplankton biomass occurs later in 

 the year in northern regions, but zooplankton biomass 

 typically declines during summer and fall (Batten et al.. 

 2003). Declining zooplankton biomass in epipelagic wa- 

 ters is related, in part, to the ontogenetic migration to 

 deep waters of some major zooplankton species, such as 

 Neocalanus spp. (Mackas and Tsuda, 1999). Declining 

 zooplankton biomass during summer likely enhanced 

 the effect of competition exerted by pink salmon during 

 odd-numbered years. July through at least September 

 is a period of high potential salmon growth (Ishida et 

 al., 1998); therefore sockeye salmon may be especially 

 influenced by prey reduction during this period. During 

 early spring, when scale growth of sockeye salmon was 

 great and did not differ between odd- and even-num- 

 bered years, prey availability was apparently sufficient 

 to minimize the effects of competition. Walker et al. 

 (1998) reported that density-dependent growth of Asian 

 pink salmon occurred after late June — a finding that is 

 consistent with our study. 



The transition from foraging on zooplankton to for- 

 aging on squid for both pink and sockeye salmon may 

 also contribute to the alternating-year pattern of sock- 

 eye salmon growth. Aydin (2000) suggested that pink 

 and sockeye salmon may begin to feed intensively on 

 micronekton squid after reaching sufficient size dur- 

 ing their second growing season. Pink salmon report- 

 edly begin feeding on squid during spring, whereas 

 sockeye salmon may not begin to feed on squid until 

 summer because sockeye salmon are smaller. During 

 odd-numbered years, pink salmon may have reduced 

 the availability of squid to sockeye salmon and influ- 

 enced the observed differences in scale growth after 

 spring. In support of this hypothesis, sampling of sock- 

 eye and pink salmon during a recent 10-year period in 

 the Bering Sea (June and July) indicated a 58% reduc- 

 tion among sockeye salmon and 32% reduction among 

 pink salmon in the weight of squid consumed during 

 odd- compared to even-numbered years (Davis, 2003). 

 Few annual estimates of squid abundance are available, 

 but Sobolevsky (1996) estimated that epipelagic squid 

 biomass in the western Bering Sea was approximately 

 five times greater in an even-year (1990) than in an 

 odd-year (1989). Population dynamics and life history 

 of squid are not well known (Nesis, 1997; Brodeur et 

 al., 1999), but their apparent one- or two-year life his- 

 tory, in conjunction with predation by pink salmon, may 

 lead to an alternating-year pattern of squid abundance 

 that re-enforces the alternating-year pattern of sockeye 

 salmon growth. 



Ruggerone et al. (2003) reported that Bristol Bay 

 sockeye salmon that inhabited the ocean in odd-num- 

 bered years of their second year at sea experienced 

 lower smolt-to-adult survival compared with sockeye 

 salmon that were present during even-numbered years. 

 Lower survival was believed to be related to competi- 

 tion with Asian pink salmon. Our findings suggest that 

 this mortality was likely related to reduced growth 

 during late spring through fall, rather than during 

 the first winter. We hypothesize that reduced sockeye 



salmon growth during the second year at sea led to 

 lower energy reserves and to greater mortality during 

 the second winter, but predation on smaller salmon may 

 also be an important factor (Nagasawa, 1998). Bioener- 

 getic modeling of salmon by Aydin (2000) indicated the 

 greatest difference between the need for prey and prey 

 availability is during winter. Nagasawa (2000) reported 

 exceptionally low prey availability and corresponding 

 low lipid content for salmon in the North Pacific Ocean 

 during winter. Ishida et al. (1998) examined salmon 

 on the high seas and determined that condition factor 

 of all salmon species was lowest during late winter. 

 Beamish and Mahnken (2001) provided evidence that 

 relatively low growth of salmon during summer and 

 fall can lead to significant growth-related mortality 

 during the first winter at sea. Growth-related mortal- 

 ity appears to occur among Bristol Bay sockeye salmon 

 in response to competition with pink salmon, but this 

 competition-related mortality primarily occurs during 

 the second winter at sea. 



Bristol Bay sockeye salmon are broadly distributed 

 across the North Pacific Ocean and Bering Sea. They 

 occur in several oceanographic regions in which domi- 

 nant prey may vary (e.g., the Bering Sea (euphausiids, 

 squid, fish], subarctic current [squid], ridge domain 

 (small zooplankton], the Alaska stream [small zoo- 

 plankton, squid, fish], and the coastal domain [fish, 

 euphausiids]) (Pearcy et al., 1988; Aydin, 2000). The 

 alternating-year pattern of scale growth was persistent 

 among adult Kvichak and Egegik sockeye salmon of all 

 age groups returning to Bristol Bay even though many 

 of these fish likely inhabited different ocean habitats. 

 Thus, the observed scale growth pattern is either highly 

 persistent in most of these ocean habitats or it is es- 

 pecially important in certain key regions inhabited by 

 Bristol Bay sockeye salmon. 



Salmon growth in relation to the regime shift 

 of the mid-1970s 



Several studies indicate that a significant change in the 

 species assemblage of the North Pacific Ocean began 

 near 1977 and concurrent with a dramatic shift in 

 physical oceanic regimes (Francis et al., 1998; Anderson 

 and Piatt, 1999). Pacific salmon abundance, including 

 Bristol Bay sockeye salmon, more than doubled after 

 this period (Rogers 1 ). Zooplankton and squid biomass 

 have appeared to increase substantially, especially in 

 coastal regions, since the mid-1970s (Brodeur and Ware, 

 1992; Brodeur et al., 1996). Furthermore, Mackas et al. 

 (1998) reported that the period of maximum zooplank- 

 ton biomass shifted one or two months earlier after 

 the mid-1970s. In comparison, seasonal scale growth 

 of Kvichak and Egegik sockeye salmon during the first 

 and second years at sea tended to be high after the 

 regime shift. This pattern was also observed in annual 

 scale measurements of sockeye salmon (Ruggerone et 

 al., 2002). Spring scale growth of sockeye salmon after 

 the regime shift was relatively high immediately after 

 entry of sockeye salmon into Bristol Bay and during 



