Quality assessment of past spawning mark estimations from a long-term survey in the Connecticut River watershed
The calcified structures of fishes provide insight into their periodic growth rates and can be combined with other biological variables to identify metrics such as size or age at maturity and mortality rates. Collecting this information on growth and life history can help evaluate the success of conservation efforts and inform future management decisions for a species in need. However, before these life history data can be applied to larger stock assessments that direct management decisions, confidence in the validity of the data needs to be reported through metrics of accuracy and precision. For this report, we assessed the bias and precision of paired reader estimations of spawning marks on scales of Blueback herring (Alosa aestivalis) collected in the U.S. Fish & Wildlife (USFWS) Annual Adult River Herring Stock Assessment for the lower Connecticut River basin. The paired reads on scales from a total of 8,698 fish over the ten years of the long term monitoring program were evaluated for the annual presence of systematic bias and precision using a combination of qualitative (i.e., frequency tables) and quantitative (i.e., Evan’s and Hoenig’s Test of Symmetry, and Coefficient of Variation (CV) calculations) analyses. While seven out of the ten survey years had systematic bias detected by the tests of symmetry, only three years (2013, 2016, 2018) had imprecision values >10% CV threshold. Data were further categorized into specific age classes within survey years to increase our resolution on where bias and imprecision was most prevalent. While the ability for accurate bias detection was limited by sufficient sample sizes (>25 fish), average imprecision values increased with age, and median age classes (4 through 6) commonly had bias detected. However, the removal of insufficient age classes prior to calculating average annual CV did not significantly change the initial average. Lastly, 2023, which was the first year to implement a standardized training procedure prior to production estimating, had the highest precision for both the annual average and specific age classes compared to all prior survey years. This standardized training procedure will continue to be used by USFWS for the lower Connecticut River tributaries, and can be modified for other river systems. Overall, this report’s results highlight the importance of assessing precision and encourage the standardization of spawning mark identification quality control and assurance for future studies. With more quality assessments and baseline information on precision and bias, there can be more beneficial discussion on defining thresholds and how to implement spawning history variability into catch curve analyses.