Development of a mechanistic model to represent the dynamics of liquid flow out of the rumen and to predict the rate of passage of liquid in dairy cattle.

A mechanistic and dynamic model was developed to represent the physiological aspects of liquid dynamics in the rumen and to quantitatively predict liquid flow out of the reticulorumen (RR). The model is composed of 2 inflows (water consumption and salivary secretion), one outflow (liquid flow through the reticulo-omasal orifice (ROO), and one in-and-out flow (liquid flux through the rumen wall). We assumed that liquid flow through the ROO was coordinated with the primary reticular contraction, which is characterized by its frequency, duration, and amplitude during eating, ruminating, and resting. A database was developed to predict each component of the model. A random coefficients model was used with studies as a random variable to identify significant variables. Parameters were estimated using the same procedure only if a random study effect was significant. The input variables for the model were dry matter intake, body weight, dietary dry matter, concentrate content in the diet, time spent eating, and time spent ruminating. Total water consumption (kg/d) was estimated as 4.893 x dry matter intake (kg/d), and 20% of the water consumed by drinking was assumed to bypass the RR. The salivary secretion rate was estimated to be 210 g/min during chewing. During ruminating, however, the salivation rate was assumed to be adjusted for the proportion of liquid in the rumen. Resting salivation was exponentially related to dry matter intake. Liquid efflux through the rumen wall was assumed to be the mean value in the database (4.6 kg/h). The liquid outflow rate (kg/h) was assumed to be a product of the frequency of the ROO opening, its duration per opening, and the amount of liquid passed per opening. Simulations of our model suggest that the ROO may open longer for each contraction cycle than had been previously reported (about 3 s) and that it is affected by dry matter intake, body weight, and total digesta in the rumen. When compared with 28 observations in 7 experiments, the model accounted for 40, 70, and 90% of the variation, with root mean square prediction errors of 9.25 kg, 1.84 kg/h, and 0.013 h(-1) for liquid content in the rumen, liquid outflow rate, and fractional rate of liquid passage, respectively. A sensitivity analysis showed that dry matter intake, followed by body weight and time spent eating, were the most important input variables for predicting the dynamics of liquid flow from the rumen. We conclude that this model can be used to understand the factors that affect the dynamics of liquid flow out of the rumen and to predict the fractional rate of liquid passage from the RR in dairy cattle.