The main goals of this project are to reconstruct trends in salmon abundance in the California Current and Northern Gulf of Alaska systems over the past 500 to 2000 years, to compare trends in abundance between the systems, and to determine relationships between climate change and salmon abundance. This project will use newly developed techniques to reconstruct salmon abundance trends from stable isotopic analysis of lake sediment cores. Several sites will be studied in each of the California Current and Northern Gulf of Alaska systems to compare trends within and between regions. The records of salmon abundance will be compared with paleoclimatic data determined from studies of tree rings, glacial advances and other sources.

The main objectives are:

• To reconstruct records of salmon abundance over the past 500 years at three sites in the California Current region.
• To reconstruct records of salmon abundance over the past 500 years at three sites in the Northern Gulf of Alaska region (Kodiak Island).
• To extend the record of sockeye salmon abundance estimates for Karluk Lake (Kodiak Island, Northern Gulf of Alaska region) back to c.a. 2000 yr BP.
• To compare trends in salmon abundance between the California Current and Northern Gulf of Alaska regions, and to determine relationships between climate change and salmon abundance over the last 500 to 2000 years.

  Cores were in hand for this project from the Alaska sites, and fieldwork was carried out during the summer of 1998 for sites in the Pacific Northwest. A total of 14 Columbia River sockeye systems (present and extinct) were successfully cored, and preliminary work has been conducted on the cores to select sites for further study. In addition, through collaboration with DFO, we are working on cores from Shuswap Lake, Fraser River, BC. Salmon reconstructions for the Alaska sites are nearly complete, and preliminary data is available for several sites in the Pacific Northwest/BC.

  Figure 1 shows preliminary N-15 data for Redfish Lake, Idaho. The N-15 data is interpreted as a proxy for sockeye salmon escapement, with higher N-15 values indicating higher levels of marine-derived nutrients in freshwater ecosystems resulting from higher escapement (Finney, 1998). The most conspicuous feature is the sharp decline following construction of the Sunbeam dam on the Salmon River downstream of the lake in 1910, and the subsequent slide towards extinction. Prior to this event, N-15 variability may be related to climate/ocean influences. A core GLOBEC hypothesis is that "Production regimes in the Coastal Gulf of Alaska (GOA) and California Current (CC) systems covary, and are coupled through atmospheric and ocean forcing". A preliminary evaluation of this hypothesis is presented in Figure 2, which compares N-15 records from Redfish Lake (CC) to Becharof Lake, Bristol Bay, AK (GOA). The 20th century portion of both records is strongly impacted by human influence. As discussed, the Sunbeam dam and subsequent changes downstream on the Columbia River had a large impact on Redfish Lake sockeye. The decline in Becharof Lake around 1950 probably reflects the shift to a cooler, less productive regime for GOA salmon. Since that time the fishery has been managed for a constant escapement goal. Thus the sediments do not record the large increase in salmon production in this system following the regime shift in the late 1970s. Interestingly, the records from about 1750 to the early 1900s suggest opposite trends during some periods, consistent with the hypothesis of out of phase behavior. Work is in progress to relate these records to paleoclimatic changes, derived from tree ring data.

  

  

  Finally, a 1700 year-long record from Karluk Lake, Kodiak Island, AK (Fig. 3) provides a long term view on salmon production in the GOA. The significant low frequency, long term trend suggests that salmon abundance in this system was reduced during the Medieval warm period, and reached higher overall levels during the Little Ice Age. This pattern is opposite to what has been observed for GOA salmon during the 20th century, in which periods of warmer temperature (corresponding to a stronger Aleutian Low or a positive phase of the PDO) seem to result in higher salmon abundance. Records from other sites are clearly needed to determine if this trend is representative. However, the data raise several speculative hypotheses regarding climatic influence on salmon populations. It is possible that some physical or ecological threshold may be reached during warm periods that result in lower salmon productivity. Alternatively, it is possible that some other mode of circulation or ecosystem state may exist in the North Pacific, that is not evident from observations made during the 20th century.
  
  Reference

  Finney, B.P. (1998). Long-term variability in Alaskan sockeye salmon abundance determined by analysis of sediment cores. North Pacific Anadromous Fish Commission Bulletin 1: 388-395.