Corrective Measures to Improve the Accuracy of Electronic Fish Counting Systems in Estimating Run Size of Large Anadromous River Herring Populations

Automated electronic counting systems are used by resource agencies to estimate river herring (alewife, Alosa pseudoharengus, and blueback herring, Alosa aestivalis) passage throughout New England. Measuring the accuracy of these counting systems is essential to understanding the limitations of counter derived population estimates, which comprise a vital component of river herring stock assessments. Anecdotal information suggests that electronic counting systems underestimate river herring runs, especially at the high rates of passage that occur during peak migration periods. The purpose of our study was to conduct a direct visual-electronic comparison of passage to derive a better understanding of the accuracy of estimates determined from a single channel electronic counting system. From 1999-2012, we conducted over 500 direct comparisons. We explored various modeling techniques, including regression and re-sampling, to correct for any identified biases in data collected during our sampling period. Our results confirmed that our counting system consistently underestimated passage, particularly at higher passage rates. Although our results demonstrate an important limitation, they provide evidence that single channel counters provide the resolution needed to allow managers to infer population trends. Future monitoring efforts should correct any biases in single channel counters, and when possible, explore alternative technologies that provide greater accuracy.