Comparison of Methods Used to Age Spring–Summer Chinook Salmon in Idaho: Validation and Simulated Effects on Estimated Age Composition

Abstract

Validation of aging methods with known-age individuals is rarely done with wild fish. We used samples collected from carcasses of adult Chinook salmon Oncorhynchus tshawytscha that were tagged as juveniles to (1) compare accuracy and precision of ocean ages determined from scales and fin rays and (2) simulate the effects of aging errors on run reconstruction (assignment to migratory cohort) under two age composition scenarios. Scale age had an overall accuracy of 81.8% and was biased high (χ² = 8.67; P = 0.014). Fin ray age had an overall accuracy of 98.6% and was unbiased (χ² = 2.00; P = 0.34). Precision of fin ray readings was higher than that of scale readings (coefficient of variation = 2.1% versus 7.8%, respectively). Accuracy of fin ray ages was greater than that of scale ages (Z = 2.198; P = 0.03). For two age composition scenarios, age classification errors were greatest when using scales and least when using fin rays. Aging errors inflated the size of weak cohorts in simulated run reconstructions. We showed how this can cause large errors in estimates of smolt-to-adult return rate when run strength and age composition vary among years. The amount of error associated with scale aging then becomes problematic for tracking the status of both threatened and healthy salmon stocks. Data based on fin ray readings provided the most accurate, unbiased estimates of age structure. Correction for methodological bias is important if age data are to be used in rigorous analyses.