Straying, stress, and potential for reproductive interactions between hatchery-produced and wild chum salmon (Oncorhynchus keta) in Southeast Alaska

Abstract

Approximately 1.5 billion juvenile hatchery-produced Pacific salmon (Oncorhynchus spp.) are currently released each year into Alaskan waters with goals of enhancing important fisheries and minimizing detrimental impacts on wild stocks. As the abundance of hatchery-produced salmon has increased, so have concerns about hatchery-origin strays entering wild systems and interactions with wild individuals on the spawning grounds. The influx of non-native strays and their associated fitness-related traits can reduce the resilience and productivity of recipient wild stocks, and is likely to be most deleterious when disparities in population sizes and heritable phenotypic characteristics between wild and hatchery fish exist. Thus, understanding the ecological and life-history mechanisms that regulate gene flow between hatchery and wild populations is crucial for conservation and management strategies in areas where hatchery enhancement is common. Currently, the ecology of strays on the spawning grounds and proximate physiological factors associated with straying (e.g., stress) are not well known. In this thesis I examine, 1) differences and similarities in several fitness-related phenotypic traits between naturally produced (presumably wild local individuals) and stray hatchery-produced chum salmon (Oncorhynchus keta) that died on the spawning grounds of Sawmill Creek, a small watershed near Juneau, Alaska, and 2) physiological differences in cortisol concentrations and the frequency of crystalline (vaterite) structure of otoliths between straying and correctly homing salmon. Hatchery-strays comprised 51.4% of the adult chum salmon that returned to Sawmill Creek during the 2015 spawning season. Hatchery males and females returned approximately seven days later, were consistently smaller (10% for males, 6% for females) in length, and younger on average than their naturally-produced counterparts. Additionally, hatchery-produced females lived fewer days on the spawning grounds during the spawning season, and retained a higher proportion of their eggs than did naturally produced females. To explore the potential role of stress on straying, I compared cortisol samples and frequency of vaterite formation in otoliths among groups of hatchery-produced fish that homed to the hatchery, hatchery-produced fish that strayed to Sawmill Creek, and naturally produced chum salmon that presumably homed to Sawmill Creek. No significant differences in cortisol concentration were found among any groups, though differences between the sexes were detected. Males of all groups had significantly lower cortisol concentrations than did females. No differences in frequency of vaterite occurrence were found between hatchery-stray and hatchery-home groups, though both hatchery groups were higher than naturally produced groups, which is consistent with findings of other studies. Thermal marking while at the hatchery during early development was not associated with vaterite formation, and no difference in frequency of vaterite formation was observed among groups of varying mark intensities. Overall, these results revealed there was ample opportunity for reproductive interactions between stray hatchery-produced and naturally produced chum salmon in Sawmill Creek during the 2015 spawning season, and consistent differences in phenotypic traits suggests the potential for gene flow to alter population-level phenotypic variation. However, despite the potential for gene flow, these results also reveal potential barriers to introgression and indicate that at least some of the presumed locally adapted traits of the natural stock remain intact. It remains unknown what the characteristics of the wild stock were prior to regional hatchery production and the extent to which the traits of this population are reflections of genetic differences between the hatchery and wild groups or phenotypic plasticity. To the extent these results are generalizable, observed differences in fitness-related traits between naturally produced and stray hatchery-produced fish may underlie the reduced reproductive success often reported in the literature. There were no differences in cortisol concentrations and frequency of vaterite occurrence between hatchery chum salmon that strayed and those that homed correctly, and the frequency of vaterite occurrence of hatchery chum salmon did not change as thermal mark intensity increased, which suggests that thermal marking may not directly alter homing ability of adults or development of juveniles, at least via otolith formation. Despite not having an effect on straying, the consistent findings of higher frequency of vaterite occurrence in hatchery-produced fish compared to naturally produced counterparts highlight the need for future work to uncover the causal underlying mechanisms and implications of vaterite on survival of the 1.5 billion salmon released each year in Alaskan waters.