So You Think You Can Age Fish? Inconsistent Annuli Development on Rainbow Trout Scales in the Lower Columbia River, BC, Canada

In 2001, BC Hydro established an annual monitoring program for the Columbia River downstream of Hugh L. Keenleyside Dam to detect temporal changes in fish populations and relate those changes to the dam’s operation. Rainbow Trout were captured by night-time boat electroshocking in September and October from 2001 to 2010. Captured fish were measured for fork length, weighed, and implanted with Passive Integrated Transponder (PIT) tags. Scales were collected for ageing. High growth rates in the study area allowed the identification of age-0 and age-1 individuals using modal patterns in length-frequency histograms. Rainbow Trout in the study area reached sexual maturity at approximately age-2, which resulted in reduced growth rates and overlapping of individual age-class modes in length-frequency histograms. Scales were used to age older individuals (i.e., age‑2 and older fish). Scales were digitally scanned and saved as a picture file. Three experienced individuals from 2 different organizations analyzed the scales. The analysts were provided with the scale image, fish species, date of capture, fork length, and capture location. To determine the accuracy of scale-based age‑class assignments, scales from 128 rainbow trout of known age, based on their previous capture history as age-1 individuals, were analyzed. Average error rate of the 3 scale analysts was 95.5% (range from 90.6% to 99.2%). All erroneously aged fish were over-aged. Of all age-2 fish (n = 91), 7.1% (range from 0.0% to 13.2%) were correctly aged, 35.2% (range from 29.7% to 38.5%) were over-aged by 1 year, 46.9% (range from 37.4% to 59.3%) were over-aged by 2 years, 9.9% (range from 7.7% to 13.2%) were over-aged by 3 years, and 1.6% (range from 0.0% to 2.2%) were over-aged by 4 years. Of all age-3 fish (n = 31), none were correctly aged, 28.0% (range from 19.4% to 41.9%) were over-aged by 1 year, 45.2% (range from 35.5% to 54.8%) were over-aged by 2 years, and 9.7% (range from 0.0% to 22.6%) were over-aged by 3 years. Accuracy generally decreased as the known age of the fish increased. No scales received a correct age-class assignment from all 3 analysts. The development of growth-checks on >99% of all scale samples suggests an unknown stress factor on Rainbow Trout in the lower Columbia River and renders scale analysis ineffective for age-class assignment of older individuals. Using otoliths to assign ages to older Rainbow Trout is currently being investigated.