2.4 Statistical analysis of sea otter prey
To test our first objective (investigate the macronutrient quality of sea otter prey across seasons), we calculated the percent protein and lipid of each prey group using the energy equivalents of 9.5 kcal/g for lipid and 5.7 kcal/g for protein . We used PRIMER v7 with a one-way analysis of similarity (ANOSIM) procedure with season as factors for each prey group, excluding crabs (α = 0.01). Snails had too few samples to conduct pairwise seasonal comparisons. Mussels were excluded from results because they were a very small portion (< 1%) of the sea otter diet around POW. Crabs were analyzed separately with season and sex as factors, using a two-way ANOSIM (α = 0.01).
To test our second objective (compare diet composition and prey macronutrient composition across different sea otter metrics), we calculated regional-level concentrations of macronutrient composition of sea otter diets using established methods . In brief, first, we converted the dry mass average (kcal, lipid, and protein) for each functional prey group to a wet mass value. Second, using the proportion of diet (also in wet mass) from SOFA outputs for each functional prey group and the average prey value (kcal, lipid, and protein), we calculated an average for each prey and macronutrient and added all individual groups together. Finally, we divided this newly calculated wet mass by dry mass to get the nutrient composition of dry mass. Statistical comparisons between sea otter metrics tested were not possible because we were not able to study known individual sea otters and all data were used at a regional scale (Western POW and surrounding islands).
Due to weather and light limitations, it was only possible to conduct visual foraging observations during spring and summer. To estimate year-round diets, we used results from stable isotope analysis . Diets estimates were made using stable isotope (δ13C and δ15N) analysis of both sea otter vibrissae and the prey present around POW using a Bayesian model to estimate percent of total diet for functional prey groups. These diet estimates were used in the calculations for seasonal sea otter macronutrient contribution and comparison to changes in sea otter prey. All datasets are archived in a publicly accessible database with the Knowledge Network for Biocomplexity .