Methane emission prediction models for lactating cows based on feed intake, body weight, and milk yield and...

Area of Science:

• Agricultural Science
• Environmental Science
• Animal Science

Background:

• Accurate prediction of methane (CH4) emissions from livestock is crucial for environmental management.
• Existing CH4 prediction models often rely on a constant CH4 conversion factor (Ym), which may introduce bias.
• Understanding the relationship between Ym and animal factors is key to improving emission predictions.

Purpose of the Study:

• To develop CH4 emission prediction models using a variable Ym approach, linking Ym to body weight (BW), milk yield (MY), and milk composition.
• To evaluate the predictive performance of these variable Ym models against constant Ym models.
• To assess prediction bias related to milk production levels.

Main Methods:

• Developed models using linear mixed models and generalized linear mixed models with a dataset of 266 records.
• Incorporated variables such as dry matter intake (DMI), gross energy intake (GEI), BW, MY, milk fat (MFAT), and milk protein (MPROT).
• Assessed model accuracy, precision, and bias using k-fold cross-validation.

Main Results:

• The best-performing variable Ym model was: CH4 emissions (MJ/d) = exp(-2.74 + 0.000325 × BW - 0.00883 × MY + 0.116 × MFAT - 0.142 × MPROT) × GEI, with R² of 0.30.
• Variable Ym models developed in this study showed no significant bias related to milk production levels.
• Existing Ym-based models exhibited substantial bias concerning production levels.

Conclusions:

• Variable Ym-based models offer improved CH4 emission predictions, particularly by eliminating bias associated with milk production levels.
• While the variance in Ym explained by BW, MY, and milk composition was modest, the variable Ym approach is advantageous.
• The proposed modeling methods can aid in developing country-specific Ym models when feed data is unavailable.