154 
Fishery Bulletin 120(2) 
works by measuring the intensity of the flourogenic signal 
from the species-specific probes. The threshold for a positive 
identification was set by the StepOne software to the expo- 
nential phase of the amplification curve. Samples of salmon 
bones that had a single species read past a set threshold of 
fluorogenic signal intensity were recorded as a positive read 
for the species identified. Samples with no amplification 
or multiple signals were recorded as unidentified salmon. 
All parameters were set to best accommodate all primer 
temperature sensitivities with the optimized annealing 
temperature of 53°C (Rasmussen Hellberg et al., 2010). 
Assessment of genotyping accuracy 
We conducted a blind study of 15 samples of known Pacific 
salmon species provided by the NOAA Northwest Fisheries 
Science Center to evaluate our genotyping error rate. We 
provided the samples to our laboratory team (C. Marshall, 
R. James, and D. Shay) at the Institute of Science and Tech- 
nology of North Central High School in Spokane, Washing- 
ton, with unique identifiers that did not identify the species 
of salmon of the sample. The laboratory team ran the sam- 
ples along with the salmon bones collected from samples of 
California and Steller sea lion scat. 
Data analysis 
Species of Pacific salmon identified as prey in scat samples 
and their size classes were tabulated for each sample. We 
combined the results from our genetic analysis with the sea 
lion diet data available from Scordino et al.” to calculate 
the split-sample frequency of occurrence (SSFO) of Pacific 
salmon by size class, species, and combination of species 
and size class. We calculated SSFO twice. In the first cal- 
culation, we included a species category of unidentified 
salmon. For the second calculation of SSFO, we assumed 
that the salmon bones identified in our work were a repre- 
sentative sample of unidentified salmon bones and incor- 
porated the unidentified bones into data for identified prey 
species proportionally according to identifications within 
each salmon bone size class. 
The following formula was used to calculate SSFO: 
Yr Or IG) 4 
Ss 
SSFO; = 100, 
where O;, = 0 if taxon 7 is absent in scat sample k and 1 if 
taxon 7 is present in scat sample k, 
O,, = total number of all taxa present in scat sam- 
ple k; and 
s = total number of scat samples that contained 
prey (Olesiuk et al., 1990). 
In many studies, SSFO has been used to reconstruct pin- 
niped diets (Olesiuk et al., 1990; Tollit et al., 2015), but it 
does have the potential to overreport the importance of 
2 Scordino, J., A. Akmajian, and S. Riemer. 2021. Steller and 
California sea lion count and diet data in northwest Washington, 
2010-2013. Mendeley Data, V1. [Available from website.] 
small prey and underreport the importance of large prey 
(Laake et al., 2002; Tollit et al., 2007). The resultant SSFO 
values were multiplied by the prey consumption estimates 
for California and Steller sea lions by Scordino et al. (2022) 
to provide an estimate of annual biomass of Pacific salmon 
consumed by species, by size class, and by species and size 
class during 2010-2013 in northwest Washington. 
We used Pearson’s chi-square (”) and Fisher's exact tests 
to evaluate differences in the species and size composition of 
Pacific salmon consumed. First, we compared consumption 
of each Pacific salmon species by season and year within 
the diet of each sea lion species and between the diets of 
the 2 species of sea lions. Second, we evaluated differences 
in the size-class composition of Pacific salmon consumed by 
season and year within the diet of each sea lion species and 
between the diets of the 2 species. For yearly comparisons 
between species, only samples collected in spring, summer, 
and fall were used because no samples of California sea 
lion scat were collected during winter. Annual comparisons 
of Pacific salmon consumed for each species included only 
data for 2011 and 2012 because samples were collected 
in all seasons during those years. For analyses of the spe- 
cies composition of Pacific salmon consumed, only samples 
identified to species were used; we excluded unidentified 
salmon from these analyses. 
Comparison of salmon consumption by sea lions 
to fishery landings 
For the period from 2010 through 2013, we downloaded 
commercial salmon fishery data from the Pacific Fisheries 
Information Network APEX reporting system (available 
from website). The data comprise commercial landings 
records from both the treaty tribal fisheries and non- 
treaty fisheries. We calculated the average annual com- 
mercial catch of Pacific salmon from 2010 through 2013 
for coastal Washington (ports along the Pacific coast, estu- 
aries, and Neah Bay) and for all of Washington State to 
compare commercial catch to estimates of consumption of 
Pacific salmon by California and Steller sea lions. 
Results 
Genetic analysis was conducted on 361 samples of bones 
identified as those of Pacific salmon by Scordino et al. 
(2022). Bone samples were from 330 samples of California 
(93 bone samples, 89 scat samples) and Steller (268 bone 
samples, 241 scat samples) sea lion scat. For 4 samples of 
California sea lion scat and 27 samples of Steller sea lion 
scat, 2 size classes of salmon were identified; for all other 
scat samples, a single size class of salmon was identified. 
All data evaluated in the research described in this paper 
are publicly available (Scordino et al.*). 
3 Scordino, J., A. Akmajian, C. Marshall, S. Riemer, R. James, 
and D. Shay. 2022. Diets of Steller and California sea lions 
determined from scat collections in northwest Washington 
during 2010-2013 with genetic identification of salmon species. 
Mendeley Data, V2. [Available from website.] 
