Using Bayesian Method and Deviance Information Criterion to Verify Uncommonly Used Fish Growth Functions

Validating age and growth estimates is always important in the exploited fisheries researches. Von-Bertalanffy growth model (VBGM] has been widely used in modeling fish growth but has been found not the most appropriate growth function in some cases. In this study beside VBGM, another uncommonly three growth functions were included to model possible alternative growth patterns and they are Gompertz growth model (GGM], Schnute-Richards growth model [SRGM] and Generalized-von-Bertalanffy model (G-VBGM]. The standard growth functions prescribe monotone increasing growth and may fail to model alternative changes in growth rates. Based on a Bayesian approach, the performances of growth models were compared using deviance information criterion (DIC) and the convergence of the MCMC was monitored using the Scale Reduction Factor (SRF). Our study found that among the five examples GGM was selected for one fish species and G-VBGM was selected for the rest four examples species. DIC remains effective for Bayesian model selection based on our further simulation study. Our study suggest that alternative fish growth models may be considered beyond commonly used VMGB.