Temperature-dependent oxygen isotope fractionation in otoliths of juvenile Chinook salmon (Oncorhynchus tshawytscha)

Oxygen thermometry has become a widely used technique for reconstructing thermal history in calcifying organisms but interpretation can be subject to predictive error from vital effects. To better understand these processes in Chinook salmon (Oncorhynchus tshawytscha) we experimentally constructed a temperaturedependent, otolith-water fractionation relationship for oxygen isotopes (818O) by rearing post-yolk absorptive juvenile fish at nominal temperatures from 6 to 21 degrees C. Temperature and otolith precipitation rate had significant effects on the otolith-water 818O fractionation, but somatic growth rate did not. The slope of the 818O fractionation equation also differed significantly from that of synthetic aragonite. Our results suggest the expression of kinetic effects on 818O fractionation in otoliths of Chinook salmon that are increasingly constrained at higher temperatures by other physical and physiological processes involved in mineral formation. Species-specific 818O fractionation equations have considerable utility to reconstruct temperature history in Pacific salmon, but applications should recognize the potential for inferential uncertainty arising from interacting vital effects.

Automated summary

Oxygen thermometry is widely used to reconstruct thermal history, but interpretation can be affected by vital effects. We studied the relationship between temperature and oxygen isotope fractionation in otoliths of Chinook salmon. Our results showed that temperature and otolith precipitation rate significantly affected the oxygen isotope fractionation, while somatic growth rate did not. Kinetic effects were observed in the oxygen isotope fractionation in Chinook salmon otoliths, and these effects were constrained at higher temperatures by other processes involved in mineral formation. Species-specific fractionation equations have utility in reconstructing temperature history in Pacific salmon, but potential inferential uncertainty should be considered.