Salish Sea Marine Survival Project

63 Figure 4. Bi-monthly proportion of salmon in the diet of harbour seals in Cowichan Bay from 2012-2016. Data from 2012-2013 (Thomas, 2015), 2014 (Benjamin Nelson, unpublished) and 2015-2016 (Majewski and Nordstrom, unpublished). Figure provided by Sheena Majewski, DFO. 3. GPS/3D accelerometer study 2015: A second tag incorporated a GPS and 3D accelerometers and allowed for reconstruction of the fine scale movements of individuals to determine locations and ways in which seals were feeding. These data were used to create maps of spatial predation risk needed to identify predation 'hot spots' during the critical period of smolt outmigration. The same 20 seals with PIT tag scanners were also equipped with these tags: 9 seals were tagged in the mouth of the Big Q River and 11 were tagged in a nearby rocky haul-out site. The purpose of this design was to determine a Coho smolt predation rate for the seals that likely specialize on out-migrating smolts versus those seals that fed in other areas and represent the larger portion of the seal population. SUMMARY OF RESULTS TO DATE 1. Scat analysis 2012-2018: Initial analysis of scat data from Cowichan Bay for 2012-2013 showed that about 12% of the spring diet of seals was made up of juvenile salmon, mostly Chinook, Coho and Sockeye. This suggested a preference for these three species over Chum and Pink, despite the high numbers of Chum available in the Strait at that time of year. Fall diets were made up of 30% salmon, mostly adult Chum. These data were further analysed based on an average yearly estimate of Coho smolt production of 15.5 million for the period 1998-2007. A seal consumption model predicted that the harbour seal mean annual consump- tion of Coho for 2012 and 2013 was as follows: 2012: 6.2 million Coho smolts eaten, 40% of the population 2013: 10.3 million Coho smolts eaten, 67% of the population Based on an estimated 2010 Chinook smolt production of ~ 66 million, the model predicts that the harbour seal mean annual consumption of Chinook for 2012 and 2013 was as follows: 2012: 26.1 million Chinook smolts eaten, 39.5% of the population 2013: 28.3 million Chinook smolts eaten, 43.0% of the population Thus, it appears likely that harbour seal predation is responsible for a significant amount of natural mortality in the early marine stage for both Chinook and Coho Salmon. Most predation appeared to occur when juvenile fish were between 100-150 mm in length. Most Coho mortality occurred in April and May as they entered the Strait at this size range: while most Chinook mortality appears to occur in July when they have grown to the preferred size range. Diet data collected between 2012 and 2014 yielded the model estimates in Figure 2. These high rates of predation required validation using additional diet data from non-estuary sites. Age-class specific salmon consumption inside versus outside estuaries in 2016 showed minor differences in diet percentages in spring (Figure 3) with lower predation rates on salmon in non-estuary sites, but these studies need additional validation. Salmon contribution to diets was significantly higher in estuaries in the fall, consisting primarily of adult chum in both habitats. However, a longer time series for Cowichan Bay suggests that the greatest variability may be temporal differences in the proportion of salmon consumed in the diet (Figure 4). Finally, analysis of the DNA meta-barcoding method has shown that this method yields biomass estimates that are similar to the currently accepted gold standard of diet analysis (Hardparts biomass reconstruction — a quite laborious process) while also providing better taxonomic resolution (down to species) for the salmon portion of diet. INTERANNUAL VARIABILITY — COWICHAN BAY Thomas et al. 2017 Nelson Majewski et al.

• Cover