Robustness of length-class interval in length based analysis for the estimation of growth parameters - A simulation study for Sardinella longiceps

Abstract

In length based fish stock assessment using von Bertalanffy growth model, the estimates of growth parameters Lm and K play a key role. A widely used estimation method is the one given by Pauly and David, using lengthfrequency data. It is not known whether the estimates of L_ and K are robust to the class interval used for grouping length samples. Here, an attempt is made to study this feature through simulation by using length samples simulated for 12 months with input parameters of Sardinella longiceps. Simulated data were used to estimate growth parameters by grouping and estimating for different class intervals ranging from 5 to 25 mm and the results then compared with the actual inputs. Results indicated that 5 mm class interval is optimum for this species.