1. Many subpopulations of sockeye in British Columbia were found to be threatened based on this assessment. Fisheries and Oceans Canada (DFO) is responsible for salmon management and is currently formulating conservation units (Holtby and Ciruna, unpublished manuscript) and initiating new policy (e.g,. Wild Salmon Policy) and procedures to stem the loss in diversity (DFO 2005, Irvine and Fraser 2008). Two locations of sockeye salmon, Sakinaw Lake and Cultus Lake, have been assessed and listed as Endangered by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC). However, the federal government has declined protection under the Species at Risk Act (SARA) due to the social and economic costs of closures to the fishery (Irvine et al. 2005). We encourage COSEWIC and DFO to focus attention on those subpopulations that we identified as threatened in this IUCN assessment.

2. We considered Ozette Lake, Lake Pleasant and Quinault Lake sockeye salmon as individual spawning sites in a larger subpopulation (SEASONAL UPWELLING). These lake systems are recognized individually as Evolutionarily Significant Units (ESUs) by National Oceanic and Atmospheric Administration (NOAA). The latter two sites were not included in the genetic baseline applied in this assessment, thus we assumed all sites are members of a larger, parent subpopulation. Further, we did not include Ozette Lake in our assessment given the recent escapement data available for the native beach-spawning population at that location is obscured by repeated attempts at establishing a tributary spawning population originating from broodstock taken from outside the basin. We encourage efforts at expanding the range-wide microsatellite-DNA baseline to include Lake Pleasant and Quinault Lake, and to investigate the status of the native beach-spawning sockeye in Ozette Lake.

3. Most of the data used in our assessment, particularly in Alaska, were from large aggregate stocks that may contain many (sometimes 100s) of individual spawning sites, and hence may mask important dynamics occurring at small scales. While a number of recent studies have shown that there is a significant degree of coherence among populations within a given region, it is important to acknowledge that a majority of the variability in vital rates are not explained by regional, environmental drivers, and may result from localized threats, for example road or other infrastructural development, or by different life history characteristics which affect productivity. Where possible and feasible, we recommend a more comprehensive monitoring approach that addresses dynamics at the scale of individual spawning sites. An excellent example of this approach is the Wood River system in Alaska’s Bristol Bay region, where a combination of sampling approaches provides a more integrated monitoring system that translates into a robust assessment of the status of the species in that basin.

4. Mixed-stock fishing is likely to be a key factor in the decline observed at many sites and subpopulations in our assessment. We feel a key priority is filling a gap in knowledge about composition of mixed-stock harvest in coastal sockeye salmon fisheries. While a great deal of resources has been invested in developing weak-stock management for Fraser River sockeye salmon through the Pacific Salmon Commission and DFO, there has been much less attention placed in other regions along the west coast of North America, particularly in the regions where we found the greatest diversity of subpopulations. We encourage continued investment in developing methods and protocols to account for subpopulation composition of coastal fisheries targeting sockeye salmon to help track harvest pressure at a biologically and ecologically meaningful scale. Further, we encourage fisheries management agencies to explore restructuring fisheries in a way that would result in shifting fishing pressure from coastal regions to more terminal locations, thus providing a more effective means of controlling fishing pressure at the scale of individual subpopulations.

5. Another leading factor threatening sockeye salmon are poor marine survival rates. This has been documented in cases where smolt-to-adult survival rates are estimated through intensive monitoring programs. This appears to be a significant factor explaining declines in adult abundance across many locations in the southern range of the species in North America. While marine conditions have been shown to cycle based on climate forcing and may, in fact, improve in future years, we feel it is important for those managing salmon in this region to acknowledge that poor marine survival may persist, particularly given projections based on global climate change. Reversing declining trends in those subpopulations affected may require increased attention to agents of mortality that are occurring at other life history stages over which we have more local control.

6. Many previous attempts at re-introductions of sockeye salmon have been unsuccessful, and we feel any effort at captive breeding or inter-basin transfers for re-introduction purposes should proceed with great caution. These efforts, to the extent that they exist, are at best stop-gap measures and are in no way a substitute for conserving the species in the wild. The lack of success from the captive breeding of endangered Sakinaw sockeye salmon by DFO, and endangered Redfish Lake sockeye by NMFS, are examples of the limitations to these expensive measures. Recent evidence of re-establishment of anadromous runs of sockeye from remnant, isolated kokanee populations following dam removal or modification (e.g., Allouette Lake in British Columbia) may offer hope in re-establishing anadromous life histories in cases where dams have prevented passage.

7. Enhancement activities (particularly hatchery releases and spawning channel construction) is likely to be a key factor in reducing abundance in neighboring, small wild populations. In our assessment, we were unable to functionally track both wild and enhanced sockeye salmon where they intermingle. We strongly recommend adoption of integrated monitoring programs that includes a robust marking program and monitoring efforts targeted toward wild sockeye salmon populations that would provide the data necessary to address the degree to which enhancement practices threaten wild sockeye salmon. It is important to note that two large basins in our assessment (Fraser and Skeena) are all strongly. influenced by enhancement activities, and these activities likely represent a key factor threatening many neighboring subpopulations.

8. While not addressed in the present assessment due to lack of data, we feel there should be more focused research attention on sea-and river-type sockeye that may serve as colonizers in the future. This line of research is particularly important given expected habitat alterations from climate change.