Comparison of nonlinear mathematical models for lamb growth analysis

Area of Science:

• Animal Science
• Agricultural Mathematics
• Quantitative Genetics

Background:

• Accurate modeling of animal growth is crucial for effective livestock management and economic viability.
• Previous growth models have varying applicability and accuracy across different breeds and environments.
• Crossbred sheep present unique growth dynamics requiring robust mathematical descriptions.

Purpose of the Study:

• To identify the optimal mathematical model for describing the relationship between age and live weight in crossbred sheep.
• To assess the applicability of the selected model across diverse farm settings.
• To develop a performance classification system for lambs based on predicted growth trajectories.

Main Methods:

• Fitted five growth models (Brody, von Bertalanffy, Richards, Logistic, Gompertz) to 9985 live weight records from 837 crossbred lambs.
• Compared model performance using statistical criteria: adjusted R², Syx%, chi-square, and residual analysis.
• Validated the best-fit model on an independent dataset of 2898 weights from six different farms.

Main Results:

• The Richards model demonstrated the best fit to the lamb weight data, accurately capturing the sigmoid growth curve.
• Model validation confirmed consistent performance across multiple farms, indicating broad applicability.
• Lambs were reliably classified as high, medium, or low performance relative to model predictions.

Conclusions:

• The Richards model is the most suitable for describing live weight gain in crossbred lambs.
• This model provides a reliable tool for classifying lamb performance, aiding management decisions.
• The model's cross-farm applicability supports its use in diverse sheep production systems.