In NPAFC catches for 1997-2025, warm PDO years did not raise Alaska salmon catch and the Alaska-WOC inverse pattern stopped; only the negative PDO link of WOC coho held as a forecast
We tested three published Pacific salmon catch-climate claims on NPAFC commercial catch data for 1997-2025, which the authors did not use: warm Pacific Decadal Oscillation (PDO) phases raise Alaska catch and lower Washington-Oregon-California (WOC) catch (Mantua et al. 1997), Alaska and WOC catches move in opposite directions (Hare et al. 1999), and total catch follows the Aleutian Low (Beamish and Bouillon 1993). We used 17 regional series (Alaska, Southeast Alaska, Canada, WOC), a pre-set plan, effective-sample-size intervals, random-walk forecast skill, and 1,000 surrogate PDO series. Detrended Alaska catch correlated with the marine-residence PDO in 1925-1996 (r +0.26 to +0.50 for five species) but not in 1997-2025 (-0.07 to +0.24), except sockeye (+0.43, 90% interval +0.08 to +0.68). The PDO kept its link to regional sea temperature, and regional SST also lost its link to Alaska pink and chum catch. The 1925-1996 Alaska link came from few regime contrasts and was confounded with the growth of Alaska hatchery releases after 1977. The negative link held in WOC: the year-to-year coho correlation was -0.41 (-0.64 to -0.12), and a 1925-1996 slope forecast 1997-2025 coho changes better than a random walk (skill +0.052, +0.019 to +0.088). It was the only one of 17 series that met the pre-set rule. The Alaska-WOC correlation changed from -0.47 to +0.13 (p = 0.03). The North Pacific Index gave no negative correlation with total catch in 1950-1992 and a positive one after 1992. PDO lags selected in training kept a small forward skill above the surrogate null (mean +0.016, p = 0.009), about 1.6% of the random-walk error. Limitations: catch is not abundance and depends on management; the NPI is a proxy for the ALPI; the forward period has few PDO phase changes. NPAFC data are not bundled; code, aggregate results, hashes and a retrieval script are included.
References
15- A sea-ice curve for Adélie penguin colony trends is confounded with region in 1984-2017; after 2017 it held on the Antarctic Peninsula, but Ross Sea colonies did not grow as it predicted
- Night light did not delay autumn leaf colour in US phenology records in 2009-2025; earlier spring leaf-out at bright sites followed regional, not local, light after site and observer controls
- A 1998-2020 farmland-versus-forest heat contrast in US NestWatch nest success did not repeat in 2021-2025 nests and depended on the habitat code and heat measure
- Colder winter minimum temperatures did not decrease the next-year growth of mountain pine beetle tree-mortality area in four western US regions, 1999-2025
- reviews, challenges, supportsdoi:10.1175/1520-0477(1997)078%3C1069:APICOW%3E2.0.CO;2
- reviews, challengesdoi:10.1577/1548-8446(1999)024%3C0006:IPR%3E2.0.CO;2
- reviews, challengesdoi:10.1139/f93-116
- supportsdoi:10.1098/rspb.2018.1475
- doi:10.1073/pnas.1921266117
- usesdoi:10.1139/f98-104
- doi:10.1139/f02-020
- useshttps://www.npafc.org/statistics/
- useshttps://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/index/ersst.v5.pdo.dat
- usesdoi:10.7289/V5T72FNM
- usesdoi:10.1175/1520-0477(1996)077%3C0437:TNYRP%3E2.0.CO;2