Key points are not available for this paper at this time.
The structure and characteristics of a model suitable for estimating digestion within the rumen and rates and patterns of nutrient entry in lactating cows are presented. The model consists of 12 state variables comprising a large particle pool, small particle pools representing insoluble dietary nutrients, soluble pools representing soluble dietary nutrients, and fermentation intermediates and end products. The model was constructed assuming continuous feeding, using Michaelis-Menten or mass action kinetics. The computer program was written in ACSL to run on a VAX computer. A fourth-order Runge-Kutta procedure was used for numerical integration. Sensitivity and behavioural analysis demonstrated that overall stability and sensitivity of the model to individual parameters was generally satisfactory, but the need to improve the description and parameterization of aspects such as particle size in relation to availability, rate and affinity constants for amino acid degradation and rate constants for particle outflow from the rumen was established. Adjustments of the model to examine discontinuous feeding regimes were undertaken and initial results with respect to changes in fermentation rates, rumen acetate levels and microbial metabolism were considered realistic. Comparisons with experimental data were considered satisfactory on forage-based and medium concentrate-containing diets, but with diets comprising 90% cereal, some inconsistencies, especially with respect to predictions of volatile fatty acid production rates, were observed. Reasons for this are put forward and suggestions for improvements in the model are discussed.
Baldwin et al. (Sun,) studied this question.