Recent Harvest Trends of Pink and Chum Salmon in Southeast Alaska: Can Marine Ecosystem Indicators Be Used as Predictive Tools for Management?

Pink and chum salmon are the principal salmon harvest components in Southeast Alaska (SEAK) commercial fisheries and were valued at greater than $75 million from 2008 to 2010 (ADFG 2011). The combined harvests of both species in SEAK generally increased from about 10 to 50 million fish from 1960 to 2000. However, production trends have also varied annually, and over the last decade there has been evidence for a downward trend in harvest for both species. Stock structure also differs dramatically between species. Of the salmon harvested from 1997 to 2010, 98% of pink salmon were unmarked, presumably wild stocks, whereas 74% of chum salmon were identified as hatchery stocks (McNair 1998, 1999, 2000, 2001, 2002; Farrington 2003, 2004; White 2005, 2006, 2007, 2008, 2009, 2010, 2011). Recent variations in SEAK salmon production could be linked to ocean or freshwater conditions, species interactions, management or enhancement practices, or a combination thereof. To better understand links between marine ecosystem factors and recent pink and chum salmon production in SEAK, data from a regional ocean sampling survey (the Southeast Coastal Monitoring [SECM] project) and corresponding ocean basin-scale indices were examined over the years 1997–2010. Pink and chum salmon have in common the life history trait of migrating to sea as fry, with strong dependence on the littoral marine habitat followed by rapid migration through seaward corridors as juveniles en route to the Gulf of Alaska (GOA). Regional and ocean basin metrics associated with salmon were compared to the harvest and survival response variables for pink and chum salmon in SEAK. Metrics were chosen to align with salmon early ocean life, which has been identified as a critical period of high and variable mortality (Parker 1968; Karpenko 1998; Kaeriyama et al. 2007; Wertheimer and Thrower 2007). At this time, juvenile salmon are vulnerable to marine conditions that influence year-class strength through factors such as growth, predator abundance, and seaward migration size and timing. Identifying a set of marine ecosystem indicators useful as predictive tools for forecasting returns would benefit salmon management in SEAK, particularly when sibling models are not possible, as is the case for pink salmon. Objectives of this study were to determine whether (1) ecosystem metrics associated with juvenile pink and chum salmon are related to recent trends in SEAK salmon production, and (2) ecosystem metrics can be used as predictive tools for salmon management. Metrics used for this study were obtained primarily from the SECM project and ocean basin-scale data sources. These ecosystem metrics were also selected to reflect time scales that affect each cohort prior to harvest. The SECM project metrics were obtained from surface trawl samples of juvenile salmon migrating to the ocean (yr1), plankton nets, and oceanographic data taken at monthly intervals from May to August in the northern region of SEAK (58°N, 135°W; Orsi et al. 2000, 2011). The SECM project metrics used were: upper 20-m integrated water temperature in May, average mixed layer depth (MLD) in June-July, average zooplankton biomass displacement volume (Bongo net 333-μm mesh) in June-July, peak fish CPUE (Ln[catch+1]) in June or July, fish energy density (kcal·g-1 WW) in July, fish size (FL, mm) on 24 July (growth proxy), seasonality (peak migration month June–August), stomach fullness (percent body weight) in July, and fish condition factor residuals in June–July. The ocean basin and regional metrics were selected as potential factors affecting juvenile salmon prior to and during their ocean residence, such as in winter (November–March, yr-1) and summer (June–Aug, yr1). These metrics included GOA (45–58°N, 130–156°W) sea surface temperatures in winter and summer (Kalnay et al. 1996), the North Pacific Index (NPI; winter and summer; Trenberth and Hurrell 1994), the Pacific Decadal Oscillation (PDO; winter and summer; Mantua et al. 1997), the Multivariate El Nino Southern Oscillation Index (MEI; winter and summer; Wolter and Timlin 1993, 1998), and the US Geological Survey monthly river discharge levels for the Mendenhall River, SEAK, in spring/summer (March–Aug, yr1; USGS 2011). Harvest and survival data were the response variables chosen for pink and chum salmon production in SEAK (Figs. 1 and 2). The harvest data was assumed to represent total return for each species and was lagged appropriately: conditions affecting juvenile salmon during their ocean entry year (yr1) were lagged one year to adults in harvest for pink salmon (yr1+1) and three years to adults in harvest for chum salmon (yr1+3). This permitted comparisons of 13 years of pink harvests and 10 years of chum salmon harvests. Marine survival data from three hatcheries’ salmon releases and one wild salmon