Tuning the tools: Validating empirical and modelled correction factors for diet estimation in pelagic teleosts

As quantitative diet estimates from biochemical markers grow in popularity, the necessity for robust species- and tissue-specific correction factors, including trophic discrimination factors (TDFs) for stable isotopes and calibration coefficients (CCs) for fatty acids, becomes increasingly urgent. The lack of available correction factors, particularly CCs for marine teleosts, constrains most teleost diet studies to stable isotopes alone. Here we calculated, validated, and modelled correction factors for southern bluefin tuna (Thunnus maccoyii) and evaluated the performance of stable isotope mixing models in comparison with QFASA. 

QFASA produced highly accurate and precise diet estimates (96 ± 1% muscle; 94 ± 5% liver), whereas stable isotope mixing models showed lower accuracy and substantially greater individual variability (52 ± 14% muscle; 43 ± 13% liver). The transferability of empirical correction factors in diet models was further evaluated, showing that QFASA was more stable when applied across different years for muscle tissue, when using externally sourced prey data, and when estimating diets of wild individuals compared with stable isotope models. 

Finally, empirical CCs and diet estimates were compared with two modelling approaches, SimQFASA, which estimates individual CCs and diets, and SMUFASA, which estimates unified CCs and diet. Diet estimates using SMUFASA closely matched empirical results (average 67% muscle; 97% liver), whereas SimQFASA estimates were more variable (25 ± 11% muscle; 54 ± 8% liver). 

This study provides the first empirically derived CCs for pelagic marine teleosts, contributing critical data to support their broader development and application. The findings also reinforce QFASA as a robust dietary tool and indicate that model-based approaches hold potential, although further refinement and validation are needed before broader application.