Nordic dairy cow model Karoline in predicting methane emissions: 1. Model description and sensitivity analysis

• Updated digestion and methane sub-models of Karoline dairy cow model was presented.
• Conducted sensitivity analysis showed the importance of the accuracy of input data.
• Accurate parameters of NDF digestion kinetic are important for model performance.
• Predicted responses in methane emissions were similar to those stated in literature.

Decreasing methane (CH4) emissions is necessary both environmentally, as CH4 has a strong greenhouse gas effect and nutritionally as CH4 represents a loss of feed energy. Karoline is a whole dairy cow mechanistic, dynamic model predicting nutrient supply and milk production. The objectives of this study were to revise the digestion and CH4 emissions modules of the Karoline model. In addition, a sensitivity analysis was conducted to evaluate the importance of the accuracy of input data required in predicting CH4 emissions. Modifications were made in the equations predicting digesta passage kinetics, microbial cell synthesis, digestion in the hind-gut, and utilization of hydrogen. The Karoline model predicted similar decreases in both organic digestibility (OMD) and neutral detergent fibre digestibility (NDFD) and improvements in the efficiency of microbial nitrogen synthesis with increasing dry matter intake (DMI) as reported in the literature. The proportion of ruminal digestion of total NDFD (0.95) and fecal metabolic and endogenous output (98 g/kg DMI) also agree with the literature data. Predicted total CH4 emissions increased with a diminishing rate by increased DMI. Predicted CH4 emissions as a proportion of GE intake decreased linearly with increased DMI. The relationships between feeding level and CH4 emissions (a decrease of 7.8 kJ/MJ gross energy per multiple of maintenance) were in good agreement with experimental data. The sensitivity analysis suggested that feed variables related to digestion kinetics of NDF [indigestible NDF (iNDF) and digestion rate of potentially digestible NDF] have a strong influence on predicted CH4 emissions; for example, predicted CH4 emissions decreased with increasing iNDF concentration. Digestion rates of starch and insoluble protein had smaller effects on predicted CH4 emissions than NDF digestion rates. Fat had a strong negative influence on predicted CH4 emissions (0.27 kJ/MJ gross energy per 1 g fat/kg DM). The sensitivity analysis suggested that accurate values of digestion kinetic variables are required for satisfactory predictions of CH4 emissions with mechanistic models. Based on reliable predictions of digestibility, microbial protein synthesis and CH4 emissions, it can be concluded that the revised Karoline model is a promising tool for predicting CH4 emissions and understanding the underlying mechanisms.