Abstracto
- This study compared the performance of five nonlinear growth models (Gompertz, Richards, von Bertalanffy, Logistic, and Michaelis-Menten) for describing cattle growth patterns across five sub-regions of Córdoba, Colombia. Body weight data from 731 calves across 16 beef farms were analyzed from birth to weaning (0–240 days). Three analytical approaches were employed: individual-level fitting, regional grouping, and regional averaging. Model selection used a voting method based on AIC, BIC, MSE, MSEP, MAE, and R2 criteria. Birth weight averaged 32.3±9.9 kg, consistent with crossbred tropical cattle. Individual-level analysis favored Gompertz models (66% selection), followed by von Bertalanffy (35%) and Logistic (2%). Regional and average analyses showed different model preferences, with von Bertalanffy dominating regional analysis and Gompertz preferred in average analysis. Asymptotic weights ranged from 428-563 kg across analytical approaches. Maximum growth rates for Gompertz models (700-900 g·day⁻1) were more biologically plausible than von Bertalanffy estimates (<300 g·day⁻1). Inflection points occurred at 200-250 days for Gompertz models, aligning with expected puberty timing. Regional variations were substantial, with San Jorge showing superior growth performance and Costanera exhibiting constraints (204 kg asymptotic weight in regional analysis). Data aggregation methodology significantly influenced model selection outcomes, suggesting that individual-level modeling suits genetic evaluation while population-level approaches better serve management recommendations. Enhanced growth monitoring protocols and consideration of regional environmental factors are essential for implementing growth models in tropical cattle production systems. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.