Describing the growth pattern of domestic geese (Anser anser) using nonlinear models.

This study aimed to characterize the growth trajectory of domestic geese (Anser anser) by fitting six classical nonlinear growth models (Brody, Bridges, Janoschek, Richards, Schumacher, and Morgan) and one alternative sinusoidal function to body weight data and comparing their performance. Models were fitted using nonlinear regression, and their adequacy was evaluated based on adjusted R², root mean square error (RMSE), Akaike's Information Criterion (AIC), and Bayesian Information Criterion (BIC). The dataset comprised body weight records from hatch to 63 days of age. All models showed good agreement with observed data (adjusted R² of 0.9812 to 0.9996, 0.9781 to 0.994, and 0.9800 to 0.9996 for males, females, and all geese, respectively). However, the sinusoidal model consistently provided the best fit, yielding the lowest RMSE (27.88 to 30.74), AIC (92.48 to 94.43), and BIC (70.66 to 72.61) values compared with alternative models such as Brody (RMSE = 183.00 to 195.60, AIC = 129.65 to 130.98, BIC = 107.53 to 108.87). Durbin-Watson (DW) values varied from 0.98 (Brody) to 2.66 (sinusoidal) for males, 0.94 (Brody) to 2.90 (sinusoidal) for females, and 0.95 (Brody) to 2.86 (sinusoidal) for the pooled dataset. Across sexes and the pooled dataset, the sinusoidal function demonstrated superior predictive accuracy and model efficiency. These results indicate that the sinusoidal model offers a flexible and precise approach for describing goose growth dynamics and may serve as a robust alternative to conventional nonlinear growth models in poultry science.