434 



Abstract— Along the west coast of the 

 United States, the potential impact of 

 increasing pinniped populations on 

 declining salmonid (Oncorhynchus spp. ) 

 stocks has become an issue of concern. 

 Fisheries managers need species-spe- 

 cific estimates of consumption by pin- 

 nipeds to evaluate their impact on 

 salmonid stocks. To estimate consump- 

 tion, we developed a model that esti- 

 mates diet composition by reconstruct- 

 ing prey biomass from fecal samples. 

 We applied the model to data collected 

 from harbor seals (Phoca vitulina) that 

 are present year-round in the lower 

 Columbia River where endangered 

 stocks of salmonids pass as return- 

 ing adults and as seaward-migrat- 

 ing smolts. Using the same data, we 

 applied the split-sample frequency of 

 occurrence model, which avoids recon- 

 structing biomass by assuming that 

 each fecal sample represents an equal 

 volume of consumption and that within 

 each sample each prey item represents 

 an equal proportion of the volume. The 

 two models for estimating diet compo- 

 sition yielded size-specific differences 

 in consumption estimates that were 

 as large as tenfold for the smallest and 

 largest prey. Conclusions about the im- 

 pact of harbor seal predation on adult 

 salmonids, some of their largest prey 

 species, remain uncertain without some 

 appropriate rationale or further infor- 

 mation (e.g. empirical captive studies) 

 to discrimmate between these models. 



Pinniped diet composition: 



a comparison of estimation models 



Jeffrey L. Laake 

 Patience Browne 

 Robert L. DeLong 

 Harriet R. Huber 



National Marine Mammal Laboratory 



Alaska Fishenes Science Center 



National Marine Fishenes Service, NOAA 



7600 Sand Point Way NE 



Seattle, Washington 98115 



E-mail address (for J. L Laake): |eff laakeig'noaa gov 



Manuscript accepted 15 February 2002. 

 Fish. Bull. 100:434-447 (2002). 



During the last three decades, harbor 

 seal {Phoca vitit!i/ia) and California sea 

 lion iZalophus californianux) popula- 

 tions along the west coast of the United 

 States have increased dramatically 

 (Forney et al., 2000). During the same 

 period, numerous salmonid (Oncorhyn- 

 chus spp.) stocks that are consumed 

 by these pinnipeds have declined and 

 some of these stocks have been clas- 

 sified as threatened or endangered 

 (NMFS, 1997). To evaluate the impact 

 of pinnipeds, fisheries managers need 

 species-specific estimates of salmonid 

 consumption by pinnipeds. 



In some limited situations, pinniped 

 prey consumption can be determined 

 from direct observation if the pinniped 

 brings the prey to the surface and feed- 

 ing occurs in a few predictable areas 

 (Bigg et al., 1990). However, in most 

 situations, consumption estimates have 

 relied on a less direct approach that 

 uses estimates of pinniped energetic 

 requirements, prey energy density, and 

 pinniped diet composition (Olesiuk, 

 199.3; Hammill et al., 1997; Stenson et 

 al., 1997; Nilssen et al., 2000). In their 

 simplest form, these models express 

 biomass consumption of prey species / 

 as j5, = s^^/e,. where ^ is the total en- 

 ergy requirement of the pinnipeds, re, 

 is the proportion of the energy in the 

 diet derived from species i, and e^ is the 

 energy density (kcal/g) of species ;. The 

 total energy requirement of a popula- 

 tion depends on the size of the popu- 

 lation and requirements of each seal, 

 which vary by sex, age, and status (i.e. 



whether it is molting, pregnant, lactat- 

 ing). If prey energy density is constant 

 or an average energy density is used, 

 biomass consumption can be expressed 

 as fi, = i^Tt,, where ^ is the total biomass 

 requirement and 7t, is the proportion 

 of the biomass derived from species i. 

 Valid estimates of pinniped energetic 

 requirements and prey energy density 

 are important, as well as accurate esti- 

 mates of pinniped diet composition. 



Diet composition can be determined 

 from skeletal remains in scat (feces) 

 and from stomach and intestinal con- 

 tents, or stomach lavage. Each of these 

 methods has some inherent bias (Bigg 

 and Perez, 1985; Pierce et al., 1991). 

 Examination of prey remains in scat 

 is noninvasive and allows for the larg- 

 est sample size. There are, however, 

 numerous well-recognized problems in 

 describing marine mammal diet from 

 scats (Jobling, 1987; Harvey, 1989; Har- 

 vey and Antonelis, 1994; Tollit et al., 

 1997b; Marcus et al, 1998). In particu- 

 lar, nonrandom passage of hard parts, 

 primarily otoliths, biases estimates of 

 diet composition; however, the bias can 

 be reduced by inclusion of all hard parts 

 (e.g. bones) (Browne et al., 2002). 



In 1994, we began an investigation of 

 harbor seal consumption of salmonids 

 in the lower Columbia River. Initial 

 attempts to survey the lower 110 km 

 stretch of the river to estimate adult 

 salmonid consumption by direct ob- 

 sei^ation proved infeasible. The only 

 remaining noninvasive alternative 

 was to develop a consumption estimate 



