Accuracy and Precision of Estimates of Back-Calculated Channel Catfish Lengths and Growth Increments Using Pectoral Spines and Otoliths

Abstract

Back-calculation of lengths at previous ages using measurements of growth increments on the bony structures of fish is a common practice, yet studies validating this procedure are few. Our objective was to determine the accuracy and precision of back-calculated lengths and the last growth increment for channel catfish Ictalurus punctatus using pectoral spines and otoliths. In recent studies, the last growth increment has been related to the length of the fish at the start of the growing season to determine size-specific growth. Growth increments on basal sections of pectoral spines were measured along three radii—posterior (SP), anterior (SA), and lateral (SL)—and otoliths were measured along the ventral radius by three independent readers. Back-calculated lengths derived from eight models (biological intercept, body-proportional hypothesis [BPH] linear, BPH nonlinear, Dahl–Lee, Fraser–Lee, Morita–Matsuishi, scale-proportional hypothesis [SPH] linear, and SPH nonlinear) for each bony structure radius were compared with the actual growth histories of captive channel catfish to determine the best model and radius (for pectoral spines) for the two bony structures. The SPH linear model for SL provided the most accurate data for pectoral spines and was also appropriate for otoliths. Back-calculated lengths and last growth increments were considerably more accurate and precise for SL than for otoliths. The SPH linear model provided an initial length–last growth increment relationship that was similar to the observed relationship for SL but not for otoliths. We also estimated last growth increments and the initial length–last growth increment relationship from the traditionally used Dahl–Lee model for SP and found that these estimates were similar to those derived from the SPH linear model for SL. We recommend using the SPH linear model for SL to estimate length-at-age and growth increment data. However, the Dahl–Lee model for SP also provides reasonable estimates of the last growth increment.