Comparison of daily- and annual- increment counts in otoliths of bigeye (Thunnus obesus), yellowfin (T-albacares), southern bluefin (T-maccoyii) and albacore (T-alalunga) tuna

Information on the age of individuals is often required for models assessing the status of stocks. Techniques used to estimate age of tuna have varied across species and agencies, precluding meta-analyses of age and growth. Wecompared age estimates obtained fromcommonly used ageing techniques for four important tuna species: bigeye tuna, yellowfin tuna, southern bluefin tuna, and albacore tuna. Estimates of age from counts of annual increments in transverse-sectioned otoliths were generally higher than those from counts of daily increments in transverse and longitudinal sections for all species, particularly for fish older than two years. However, annual counts produced younger estimates, on average, relative to daily counts for bigeye and yellowfin tuna younger than one year. Estimates derived from daily increments in longitudinal and transverse sections were generally similar, although longitudinal sections produced relatively older age estimates for individuals older than two years. A linear or non-linear increase in the magnitude of differences between ageing methods was the bestapproximating model in all cases except when comparing daily-increment counts between transverse and longitudinal otolith sections for southern bluefin tuna. These observations are consistent with a narrowing of daily increments with increasing age, resulting in underestimates of age relative to those derived from annual increments. We conclude that (i) daily increments are unsuitable for ageing individuals over two years, especially for southern bluefin and albacore, (ii) longitudinal sections are more precise and produce older age estimates than transverse sections for daily-age estimates, (iii) there are considerable differences in these trends between species, likely dependent on longevity, and (iv) parameter estimates and/or conclusions based on meta-analyses using age data derived from different ageing methods are likely confounded withmethodological biases. This result demonstrates that greater effort is required to provide consistent, validated methods for routine age determination to support the assessment and management of these valuable populations.