Search

Pacific salmon are a keystone resource in Alaska, generating annual revenues of well over ~US$500 million/year. Due to their anadromous life history, adult spawners distribute amongst thousands of streams, posing a huge management challenge. Currently, spawners are enumerated at just a few streams because of reliance on human counters and, rarely, sonar. The ability to detect organisms by shed tissue (environmental DNA, eDNA) promises a more efficient counting method. However, although eDNA correlates generally with local fish abundances, we do not know if eDNA can accurately enumerate salmon. Here we show that daily, and near-daily, flow-corrected eDNA rate closely tracks daily numbers of returning sockeye and coho spawners and outmigrating sockeye smolts. eDNA thus promises accurate and efficient enumeration, but to deliver the most robust numbers will need higher-resolution stream-flow data, at-least-daily sampling, and a focus on species with simple life histories, since shedding rate varies amongst jacks, juveniles, and adults., (© 2019 John Wiley & Sons Ltd.)

Pacific salmon migration timing can drive population productivity, ecosystem dynamics, and human harvest. Nevertheless, little is known about long-term variation in salmon migration timing for multiple species across broad regions. We used long-term data for five Pacific salmon species throughout rapidly warming southeast Alaska to describe long-term changes in salmon migration timing, interannual phenological synchrony, relationships between climatic variation and migratory timing, and to test whether long-term changes in migration timing are related to glaciation in headwater streams. Temporal changes in the median date of salmon migration timing varied widely across species. Most sockeye populations are migrating later over time (11 of 14), but pink, chum, and especially coho populations are migrating earlier than they did historically (16 of 19 combined). Temporal trends in duration and interannual variation in migration timing were highly variable across species and populations. The greatest temporal shifts in the median date of migration timing were correlated with decreases in the duration of migration timing, suggestive of a loss of phenotypic variation due to natural selection. Pairwise interannual correlations in migration timing varied widely but were generally positive, providing evidence for weak region-wide phenological synchrony. This synchrony is likely a function of climatic variation, as interannual variation in migration timing was related to climatic phenomenon operating at large- (Pacific decadal oscillation), moderate- (sea surface temperature), and local-scales (precipitation). Surprisingly, the presence or the absence of glaciers within a watershed was unrelated to long-term shifts in phenology. Overall, there was extensive heterogeneity in long-term patterns of migration timing throughout this climatically and geographically complex region, highlighting that future climatic change will likely have widely divergent impacts on salmon migration timing. Although salmon phenological diversity will complicate future predictions of migration timing, this variation likely acts as a major contributor to population and ecosystem resiliency in southeast Alaska., (© 2014 John Wiley & Sons Ltd.)

To predict how climate change will influence populations, it is necessary to understand the mechanisms, particularly microevolution and phenotypic plasticity, that allow populations to persist in novel environmental conditions. Although evidence for climate-induced phenotypic change in populations is widespread, evidence documenting that these phenotypic changes are due to microevolution is exceedingly rare. In this study, we use 32 years of genetic data (17 complete generations) to determine whether there has been a genetic change towards earlier migration timing in a population of pink salmon that shows phenotypic change; average migration time occurs nearly two weeks earlier than it did 40 years ago. Experimental genetic data support the hypothesis that there has been directional selection for earlier migration timing, resulting in a substantial decrease in the late-migrating phenotype (from more than 30% to less than 10% of the total abundance). From 1983 to 2011, there was a significant decrease--over threefold--in the frequency of a genetic marker for late-migration timing, but there were minimal changes in allele frequencies at other neutral loci. These results demonstrate that there has been rapid microevolution for earlier migration timing in this population. Circadian rhythm genes, however, did not show any evidence for selective changes from 1993 to 2009.