16 



Fishery Bulletin 104(1) 



humpback whales currently feeding in the study ar- 

 ea, these whales would be removing nearly 8.83x10^ 

 (6.41x106 to 1.36x10") kg annually, including 3.26xl0« 

 {2.37x10*5 to 5.01x10*5) kg of pollock, nearly 2.55 xlO*^ 

 (1.85x10*5 to 3.92x10*5) kg of capelin, and 6.71x105 

 (4.88x105 to 1.03x10*5) kg of eulachon. If the same diet 

 were consumed by a population of humpback whales al- 

 lowed to return to prewhaling abundance, the projected 

 population would remove 1.9x10" (1.86x10" to 2.12x10") 

 kg of prey annually, including approximately 7.13x10*5 

 (6.88x10*5 to 7.82x10*5) kg of pollock, 5.57x10*5(5.38x10*5 

 to 6.11x10*5) kg of capelin, and 4.25x10*5 (4.10x10*5 to 

 4.65x10*5) kg of euphausiids (Table 4). 



Discussion 



Consumption rate 



Estimating the energy requirements of large cetaceans is 

 inherently difficult and values presented in the present 

 study may be subject to substantial uncertainty. Previ- 

 ous studies in which consumption rates for cetaceans 

 were estimated have used a range of values to adjust 

 BMR (£=70M0 '5) for active metabolism. These values 

 generally range from approximately 1.5 to 3 times BMR 

 (Hinga, 1979; Lockyer, 1981; Sigurjonsson and Vikings- 

 son, 1998). Our value of 192 is 2.7 times larger than 70 

 and is, therefore, a reasonable estimate because it fits 

 within this range and is based on the observed oxygen 

 consumption rates of baleen whales. However, the con- 

 sumption estimates are highly sensitive to perturbations 

 of model input; a 5% error in this value would cause 

 deviation of the same percentage (5%) in final consump- 

 tion values. Further, all values in our consumption model 

 are assumed to be constant when body mass, physiologi- 

 cal status, and assimilation efficiency are likely subject 

 to large seasonal fluctuations (Innes et al., 1987; Perez 

 and McAlister, 1993, Kenney et al.. 1997; Trites et al., 

 1997; Sigurjonsson and Vikingsson, 1998). Our model, 

 however, did account for seasonal changes in the energy 

 density of local prey sources; previous models, on the 

 other hand, did not account for these changes (Perez 

 and McAlister, 1993). Further research is necessary to 

 obtain reliable field estimates of metabolic rates if model 

 uncertainty is to be reduced. 



The historic prevalence of surf smelt in diet A could 

 imply a dramatic change in surf smelt availability, 

 misidentification, or an overestimation of smelt found 

 in stomachs. Thompsons (1940) analysis resulted from 

 "samples of stomach contents" obtained from catcher 

 vessels; therefore, these samples may have completely 

 missed less prevalent species. Further, stomach samples 

 may have only reflected the most recent meal of the 

 whale and therefore be biased toward a single species. 

 This potential bias, however, could have been minimized 

 by sampling stomachs throughout the season (May 30- 

 August 09) (Thompson, 1940). Diet B was dominated by 

 walleye pollock, a species not present in historic diet A. 

 The increased importance of juvenile pollock in contem- 



porary humpback whale diet B could reflect changes in 

 prey species availability and use, foraging selectivity, 

 or reflect our diet reconstruction method. 



Diet B is considered provisional for two reasons. First, 

 it is assumed that humpback whales eat prey species 

 in proportion to their availability within foraging ar- 

 eas. Humpback whales select preferred prey species 

 and consumption, therefore, may be disproportional 

 to availability. That is, they may be selectively forag- 

 ing from all available prey sources. Previous foraging 

 studies have described humpback whale distribution 

 as being correlated with areas of capelin (Whitehead 

 and Carscadden 1985; Piatt et al. 1989) and sandlance 

 abundance (Payne et al. 1986; Kenney et al. 1996) and 

 this correlation may indicate a possible preference for 

 small forage fish species. Given that in the decades 

 since whaling, the Gulf of Alaska has shifted from a 

 system dominated by forage fish to one dominated by 

 pollock and other groundfish (Merrick 1997; Anderson 

 and Piatt 1999; Benson and Trites 2002), a shift in 

 prevalence from surf smelt in the historic diet to pol- 

 lock in the current diet is not unexpected. Pollock have 

 been shown to be a dominant prey source of humpback 

 whales harvested in Russia (Klumov, 1963). Addition- 

 ally, humpback whales in southeastern Alaska have 

 been observed near schools of juvenile pollock and are 

 believed to eat pollock to an unknown, but potentially 

 large, extent in some years (Gabriele^). 



The second source of uncertainty in diet B stems 

 from the assumption that our mid-water trawl surveys 

 provide unbiased samples of all available prey. Because 

 these surveys were not designed to sample zooplankton, 

 they may have produced a biased estimate of euphausiid 

 availability. This bias may not be significant, however, 

 because the 22% value we used in diet B was based on 

 historic usage and falls within the range of euphausiid 

 consumption (5-30% of the total diet) estimated in 

 other humpback whale studies (Perez and McAlister, 

 1993; Kenney et al., 1997). 



Further, diet B was constructed from the results of 

 mid-water trawl surveys that may underestimate the 

 availability of some forage fishes, particularly Pacific 

 sandlance. Pacific sandlance are often small enough 

 to swim through the meshes in the net or are found in 

 benthic habitats and cannot be captured by mid-water 

 trawl methods. To minimize this potential sampling 

 bias, we supplemented our trawl surveys with purse 

 seine sampling in the nearshore subarea. Despite this 

 effort we may have underestimated the prevalence of 

 Pacific sandlance in the area because it was found to 

 dominate the diets of other coastal piscivores; stomach 

 contents of 34 coho salmon (Oncorhynchus kisutch) and 

 Pacific halibut iHippoglossus stenolepis) in 2002 (Wit- 

 teveen^) and regurgitants from blacklegged kittiwakes 



= Gabriele, C. 2001-2002. Personal commun. Glacier Bay 

 National Park. P.O. Box 140. Gustavus, AK 99826-0140. 



'5 Witteveen, B. H. 2002. Unpubl. data. Fishery Industrial 

 Technology Center, University of Alaska Fairbanks, Kodiak, 

 AK 99615. 



