The relationship between body and scale growth proportions and validation of two back-calculation methods using individually tagged and recaptured wild Atlantic salmon

Using individually tagged wild Atlantic salmon Salmo salar from two Icelandic rivers, we evaluated scale growth in relation to body growth and validated smolt length estimates derived from the Dahl-Lea and Fraser-Lee back-calculation models. Smolts were measured and individually tagged during their seaward migration, and scale samples were taken. The scale reader's accuracy was tested, and the bias was found to be less than 5% on average. Spring freshwater growth measurements (i.e., tiverine growth during the spring before seaward migration) were validated using scales from smolts and again from the same individuals as adults. Spring growth was largely underestimated (average=31-34%) when back-calculated from adult scales. Isometry in body growth and scale growth depended on the fish growth rate. The Dahl-Lea back-calculation model overestimated smolt lengths in both populations (average=0.76-1.43 cm). Use of the Fraser-Lee back-calculation model resulted in slightly more extensive overestimation of smolt length (average=1.36-1.71 cm). The overestimation decreased with increased smolt size, especially in the Vesturdalsa. The results indicate that the accuracy of the Dahl-Lea back-calculation model is acceptable and that this model is less biased than the Fraser-Lee model for back-calculating smolt length. The effect of over-or underestimating freshwater growth in spring should be kept in mind when back-calculating from adult scales, even if the final error will be proportionally small.