Simulation of nutrient digestion, absorption and outflow in the rumen: model evaluation.

A mathematical model of the rumen fermentation processes constructed to predict nutrient supply to the host animal was evaluated. Sensitivity analysis on high fiber, starch and protein diets indicated that the model responds appropriately to these types of diets and to changes in parameter values, and revealed that the model is sensitive to the availability of hexose for non-growth microbial processes and to the maximum storage rate of polysaccharides in amylolytic microbes, although sensitivity varied with diet composition. Of the parameters whose values were dependent on diet, the fraction of protozoa in the amylolytic microbial pool and the fluid and solid passage rates needed the most careful estimation. When model predictions of nutrient supply were compared with the experimental observations, those for duodenal flows of neutral detergent fiber, total non-ammonia nitrogen (NAN) and total volatile fatty acid rumen concentration were satisfactory for several feeding strategies. The partition of NAN flow into microbial and non-microbial NAN flow and the molar proportions of volatile fatty acid production and concentration were not predicted well. The representation of the complex interactions between rumen microbial populations and of their effects on the production of specific volatile fatty acids merits further study for an improvement in the prediction of nutrient supply.