Inter-method differences (SIMS vs. IRMS) in oxygen isotope fractionation: insights from Chinook salmon otoliths

Stable oxygen isotopes (δ 18 O) in biogenic carbonates serve as a valuable proxy for reconstructing thermal history. Fish otoliths (ear stones) are particularly useful, as they precipitate continuously throughout life, creating a temporally resolved archive of water temperature. Here, we calibrate the temperature-dependent oxygen isotope fractionation equation for Chinook salmon ( Oncorhynchus tshawytscha ) using two analytical methods, secondary ion mass spectrometry (SIMS) and isotope ratio mass spectrometry (IRMS), to assess method-specific effects on fractionation equations and their implications for otolith-based thermometry. Juvenile fish were reared for 15 weeks under controlled freshwater conditions (salinity < 0.1 ppt) with a stable ambient water δ 18 O of − 5.54 ‰ (VSMOW) (±0.10, 1 SD) at three temperatures (11, 16, 20 °C). Otolith δ 18 O values measured by SIMS showed a significant linear inverse relationship with ambient water temperature: 1000 l n α = 11.51 ± 1.39 , 1 S E × 10 3 T ( K ) - 1 - 10.94 ± 4.80 , 1 S E δ 18 O otolith ( V P D B ) - δ 18 O w a t e r V S M O W = - 0.14 ± 0.02 , 1 S E × T ( ° C ) + 0.64 ( ± 0.27 , 1 S E ) . Applying this equation yielded water temperature reconstructions with an accuracy of ± 1.97 °C and a precision of ± 0.70 °C (1 SD). A paired comparison revealed SIMS δ 18 O values were on average 1.97 ‰ lower than IRMS values, likely due to matrix effects and organic content. This offset produced large differences in equation intercepts, leading to reconstructed temperatures from IRMS-based equations that deviated ∼ 10 °C from observed temperatures when applied to SIMS data. In contrast, the slopes (thermal sensitivity) of SIMS and IRMS equations were highly consistent, indicating that relative temperature changes can still be reliably inferred from SIMS δ 18 O values using IRMS-based equations. Greater variability in SIMS δ 18 O values compared to IRMS may partly reflect fine-scale isotopic heterogeneity within otoliths, suggesting that SIMS-based temperature reconstructions may require larger sample sizes or additional calibration tailored to specific contexts. Across-species comparison of fractionation equations revealed that inter-method differences exceeded inter-species differences, highlighting the need for method-matched equations for accurate absolute temperature reconstructions. Despite these challenges, once calibrated, SIMS-based otolith thermometry provides a powerful tool for reconstructing fine-scale fish thermal histories and assessing habitat refugia and resilience to climate change.