Trophic Relationships Among Juvenile Salmon During a 16-Year Time Series of Climate Variability in Southeast Alaska

Early marine survival of juvenile Pacific salmon (Oncorhynchus spp.) has been linked to growth related to marine temperatures and feeding conditions during the first few months in the ocean (Mortensen et al. 2000; Farley et al. 2007; Saito et al. 2009). These relationships for juvenile salmon have been contrasted among coastal regions (Brodeur et al. 2007) and variable environmental conditions in the coastal waters of the Bering Sea (Karpenko et al. 2007; Andrews et al. 2009; Moss et al. 2009), Prince William Sound (Armstrong et al. 2008; Cross et al. 2008), and Southeast Alaska (Landingham et al. 1998; Mortensen et al. 2000; Weitkamp and Sturdevant 2008; Sturdevant et al. 2012c). However, less is known about these relationships for most juvenile salmon over an extended time series characterized by climate change (Miller et al. 2013; Orsi et al. 2013). To address how trophic linkages and nutritional condition may shift over long-term periods of environmental change, we examined a 16-year time series in groups of warm and cold years for juvenile pink (O. gorbuscha), chum (O. keta), sockeye (O. nerka), and coho (O. kisutch) salmon. Trophic data included July diet composition and feeding intensity, and nutritional condition (energy density and body size) from Icy Strait (58°N, 135°W), Southeast Alaska. Typically, an n ≤ 10 diet and n ≤ 20 energy subsamples of average-size fish per species and year were selected from annual Southeast Coastal Monitoring (SECM) project trawl sampling in Icy Strait during daylight hours (0700-1900) between 21-31 July, 1997-2012 (Orsi et al. 2012). However, to meet sample size criteria, a few coho salmon diet and energy samples were collected from inshore and coastal stations in 1997, but no coho salmon energy samples were available in 1999. A few pink and chum salmon diet samples were supplemented from night trawls in 2004. Warm and cold years were defined by annual deviations from the long-term average of the Icy Strait Temperature Index (ISTI, mean °C, 20-m integrated water column) calculated across the months of May, June, July, and August at n = 8 SECM stations in Icy Strait (Fig. 1). The ISTI is significantly correlated with the climate Multivariate ENSO Index (MEI, averaged from November to March, lagged by one year; Sturdevant et al. 2012a; Orsi et al. 2013a), indicating that broad scale climate patterns can be detected in regional signals. We identified nine warm years (mean 9.6°C; range 9.3–10.3°C) and seven cold years (mean 8.8°C; range 8.3–9.0°C) in the 16-year time series (Fig. 1), although monthly anomalies were variable (Sturdevant et al. 2012b).