Winship and Trites; Prey consumption of Eumetopias /ubatus 



149 



It is important to note that the model estimates of 

 daily food requirements are not estimates of daily food 

 consumption (Winship et al., 2002). Stellar sea lions do 

 not necessarily feed on a daily basis and breeding adults 

 fast for periods during the breeding season. For example, 

 the food required by a breeding male during the breeding 

 season fast (to meet its energy requirement) would have 

 been consumed before or after the breeding season (i.e. 

 outside the breeding season). On an annual basis, how- 

 ever, the model estimates of food requirements are equal 

 to food consumption if animals are consuming enough food 

 to meet their energy requirements. We assumed that the 

 amount of food consumed by the Steller sea lion popula- 

 tion in Alaska in 1998 equaled the food requirement of the 

 population. 



Bioenergetic parameters 



We used sampling distributions for bioenergetic param- 

 eters that were identical to those used by Winship et 

 al. (2002) with the exception of fecal digestive efficiency 

 and the heat increment of feeding for maintenance (for 

 nonpups). Winship et al. (2002) defined fecal digestive 

 efficiency as 1 minus the proportion of gross energy lost in 

 feces, and assumed that its value ranged from 0.90-0.96 

 for Steller sea lions (i.e. fecal digestive efficiency was sam- 

 pled from a uniform distribution). However, several stud- 

 ies have shown that the digestive efficiency of pinnipeds is 

 positively correlated with the energy density of their prey 

 (Keiver et al., 1984; Martensson et al., 1994; Lawson et al., 

 1997). In contrast, two other studies found that the diges- 

 tive efficiencies of pinnipeds did not differ significantly 

 among diets of different energy densities, although in both 

 studies the average digestive efficiency was highest for the 

 diet with the highest energy density (Fisher et al., 1992; 

 Fadely et al.M. Rosen and Trites (2000a) found that the 

 fecal digestive efficiency of captive Steller sea lions fed her- 

 ring, pollock, salmon, and squid was positively correlated 

 with the energy density of their diet. We therefore fitted 

 a logistic equation to the data in Rosen and Trites (2000a; 

 their Tables 1 and 2) using nonlinear least-squares regres- 

 sion (Nonlin; SYSTAT, Inc., 1992) and used this fitted equa- 

 tion (77=20, r'=0.75) to calculate fecal digestive efficiency as 

 a function of the energy density of prey: 



DE, 



ED^ = energy density of prey category i (kj/g wet 



mass); and 

 EDf, = 2.10 (±0.089). 



The fitted parameters {A, k. and ED^^) were randomly 

 sampled from normal distributions with the previously 

 described means and SEs (in each run of the model). 



Winship et al. (2002) defined the heat increment of 

 feeding for maintenance as the proportion of metaboliz- 

 able energy used for maintenance that is lost due to the 

 metabolic cost of digesting and processing food energy, 

 and used a uniform sampling distribution of 0.10-0.15. 

 However, there is evidence that the heat increment of 

 feeding in Steller sea lions, like fecal digestive efficiency, 

 varies with the energy density of prey (Rosen and Trites, 

 1997; 1999; 2000b). We fitted a linear equation to the data 

 from Rosen and Trites ( 1997; their Table 1, including data 

 for both meal sizes) and the raw data (Rosen-) from Rosen 

 and Trites (1999) and Rosen and Trites (2000b) using lin- 

 ear least-squares regression and used this equation (/!=22, 

 P<0.0001, /•-=0.60) to calculate heat increment of feeding 

 as a function of the energy density of prey: 



HIF. =axED, + h, 



where HIF, 



l + e 



-k(ED-ED„) 



where D£, = fecal digestive efficiency for prey category 



heat increment of feeding for prey category 

 / (as proportion of gross energy); 



a = -0.013 (±0.0023 SE); and 



h = 0.229 (±0.0173). 



The fitted parameters (a and b) were randomly sampled 

 from normal distributions with the previously described 

 means and SEs (in each run of the model). HIF was then 

 divided by fecal and urinary digestive efficiency to obtain 

 the heat increment of feeding as a proportion of metaboliz- 

 able energy (Winship et al., 2002). 



Population parameters 



We used the same sampling distributions for popula- 

 tion composition parameters (survival, maturity, and 

 reproductive rates) as outlined in Winship et al. (2002). 

 The sampling distributions for population composition 

 parameters used by Winship et al. (2002) were based 

 on life tables developed for Steller sea lions (York, 1994, 

 Trites and Larkin'*) that were based on collections done 

 in the 1970s in Alaska (Calkins and Pitcher, 1982). Those 

 life tables were developed on the assumption of a stable 

 population size. However, since the 1970s the sizes of 

 Steller sea lion populations in some regions of Alaska 



A = 0.951 (±0.0039 SE); 

 k = 1.86 (±0.016); 



Fadely, B. S., J. A. Zeligs, and D. P. Costa. 1994. Assimilation 

 efficiencies and maintenance requirements of California sea 

 lions [Zalophus californianus) fed walleye pollock iTheragra 

 chalcogramma) and herring iClupea harengusl Final report 

 to the National Marine Mammal Laboratory (NMML), 28 p. 

 NMML, NOAA, 7600 Sand Point Way N. E., Seattle, WA 98115. 



2 Rosen, D. A. S. 2001. Personal commun. Marine Mammal 

 Re.'^earch Unit, Fisheries Center, University of British Colum- 

 bia. Room 18, Hut B-3, 6248 Biological Sciences Road, Vancou- 

 ver, B.C., Canada, V6T 1Z4. 



' Trites,A.W.,andP.A. Larkin. 1992. The status of Steller sea 

 lion populations and the development of fisheries in the Gulf 

 of Alaska and Aleutian Islands. Unpubl. rep., 134 p. Marine 

 Mammal Research Unit, Fisheries Center, University of British 

 Columbia, Room 18, Hut B-3, 6248 Biological Sciences Road, 

 Vancouver. B.C., Canada, V6T 1Z4. 



