Validation of age estimates for Chum and Sockeye salmon derived from otolith and scale analysis

Abstract

Accurate age estimates are essential for developing robust stock assessment models and sustainable harvest, conservation, and management strategies. Here we use thermally marked hatchery-origin fish of known age to validate scale and whole otolith age estimation techniques for two stocks of Chum and one stock of Sockeye salmon from Washington state. Scale ages of Hood Canal summer-run Chum salmon were 95.7 % accurate (average percent error [APE] = 1.29 %) across 8 run years (RYs) and scale ages of Lower Columbia River fall-run Chum salmon were 93.0 % accurate (APE = 2.23 %) across 9 RYs. Cedar River Sockeye salmon were sampled in two locations: at the Ballard Locks during their spawning migration (scales and otoliths) or posthumously during spawning ground surveys (otoliths only). Scale and otolith ages of Ballard Locks Sockeye salmon demonstrated high and similar accuracy across 6 RYs (scales: 96.9% accurate, 0.86 % APE; otoliths: 95.3% accurate, 1.14 % APE). Otolith ages of spawning ground sampled Sockeye salmon were slightly less accurate (8 RYs: 92.9 % agreement, 1.65 % APE), largely due to a single outlying year. For both species we observed a trend of over-ageing individuals from the youngest age class and under-ageing individuals from the oldest age class. However, these age classes represented an exceedingly small proportion of sampled individuals for all stocks (always <2.6 %) and symmetry tests indicated bias was not significant in most years. To our knowledge this is the first age validation study using known age individuals for Chum and Sockeye salmon.