Fish length back-calculation from scales: advancing methodology and correction of bias

Abstract

Fish length back-calculation from hard structures (e.g., scales, otoliths, spines, etc.) is a commonly used method to reconstruct individual growth rates and sizes at age/stage in the absence of an intensive sampling–resampling timeseries. However, reliable estimates via these methods require the empirical validation of multiple assumptions about the growth of these hard structures. Here, we focus on reducing bias in scale-based back-calculations and validating proposed improvements using archived scales from a wild population of anadromous steelhead trout ( Oncorhynchus mykiss). We first describe a series of back-calculation problems and propose solutions that can be integrated into common back-calculation methods. We then compare back-calculation accuracy, precision, and bias between proposed solutions and traditional forms of two back-calculation methods: Fraser–Lee and Body-Proportional. We discovered that the assumption that rapid growth begins immediately after juvenile steelhead pass downstream of the fish fence (300 m upstream of the ocean) was invalid and required a correction factor to prevent overestimating fish length by an average of 14%. The proposed modified methods we described performed better than traditional back-calculation methods. Overall, these findings can improve estimates of fish length from scale-based back-calculations and illustrate the importance of validating key assumptions.