Nutritional Modeling and Its Impact on Managerial Goals in Dairy Production

Modeling, as defined by Mertens (59), involves development of a set of mathematical equations to predict changes in the real world. Modeling is a recognized part of the scientific method and is as old as science itself. Regression equations, fat-corrected milk formulas, and total digestible nutrients (TDN) equations all provide examples of modeling in its simplest form. Only the complexity which current models are taking is new. Modeling should not represent the endpoint of scientific research but be used as an important scientific tool. The endpoint of modeling research must lie in simulation. Mertens (59) defines simulation as use of models to study dynamics of a system. Models such as those developed to evaluate nutrient requirements of dairy cows (62) are used daily in simulation to predict energy, protein, and mineral requirements. While this paper deals specifically with modeling, it is difficult and undesirable to separate modeling from simulation, and, consequently, discussion will overlap. Recently there has been a rapid increase in interest in nutritional modeling (2, 12, 20, 25, 54, 67, 68). To understand the impact such models may have, it is important to understand the concepts involved in model development. An objective of all model developers is better to understand input/output relationships of a system. As an example, many milk production models have been developed to help interpret effects of a specific variable (i.e., crude protein) on milk yield. Often models include a second objective of understanding and resolving interactions between independent variables. For example, protein and energy intake interact to affect milk yield. Increasing both simultaneously results in greater production than the sum of results if each are raised independently. Baldwin et al. (2, 67, 68) demonstrated another major