The formulation and management of dairy cow diets has become increasingly more dependent on computer programming and modeling over the past two decades.
Once personal computers began appearing in the 1980s, innovative dairy scientists were quick to adapt electronic spreadsheets and database technology to the task of, first, evaluating rations and then moving on to mathematical modeling that would accurately predict milk production in a variety of environments and with a multitude of feedstuffs.
Just as quickly, economic modeling was developed using optimization programming that would accurately provide least-cost solutions for diets and milk production.
Dairy scientists and researchers came to realize that ruminant nutrition was much more complicated than monogastric nutrition and that cows actually had two sets of nutritional requirements – one for the rumen and one for the rest of the cow. Slowly, the nutritional requirements for a cow began to include various protein fractions such as rumen-degradable protein (RDP) and rumen-undegradable protein (RUP).
Crude fiber calculations gave way to acid detergent fiber (ADF) and neutral detergent fiber (NDF) and more attention was paid to mineral and vitamin levels in all feedstuffs. As more commodities became available, calculating and keeping track of a diet and the changes in that diet became increasingly more complicated.
Computerized ration modeling has evolved from simple spreadsheets with a handful of constraints to power-packed programs that now include prediction equations for rumen bacterial growth and the amounts of essential amino acids that bypass the rumen, as well as how fats impact the rumen microflora.
Nearly all feedstuffs, forages and grains alike now have assigned rates of degradation for both protein and carbohydrates. The most sophisticated computer models now quickly predict milk production based upon environmental conditions such as temperature and humidity along with metabolizable amino acids, fermentable carbohydrates and NDF digestion rates of feedstuffs.
According to educator and dairy nutritionist Dr. Charles Sniffen, the work on ration modeling continues to be a work in progress. In recent years, the primary focus on ration modeling has centered on both fiber and carbohydrate digestion. Evolving research has found that most feedstuffs do not ferment and digest at a constant rate in the rumen.
Rather, there are various rates of feed degradation occurring during digestion, resulting in multiple pools moving through the rumen during the digestion process. Each of these pools must be accounted for and assigned different rates of digestion.
As the biology of the rumen is more completely understood, nutrition models must incorporate that “new” biology in mathematical models. In order for ration modeling software to continue to be useful, newer equations are being developed that can predict how any given forage will ferment in the rumen. Grasses will ferment at different rates than legumes due to the differences in NDF.
One of the major hurdles in trying to maximize the amount of forage a cow can consume each day is predicting what the “gut fill” of any given forage will be. Recently, laboratories have begun providing nutritionists with undigestible fiber values of neutral detergent fiber (uNDF), which is used as a proxy for determining gut fill and how much digestible energy can be expected from feeding a certain level of forage. All forages have varying levels of digestible and undigestible NDF based on their stages of maturity.
Databases continue to be updated with new values as the uNDF values become available.
Because the dairy industry feeds so much corn, starch fermentation and digestibility continues to demand much attention. The degradability of starch in the rumen also varies based on how vitreous the starch is.
For corn starch, differing corn hybrids, along with maturity and moisture levels at the time of harvest, will result in different rates of starch fermentability and metabolizable energy. Here too, the various pools of starch digestion must be accounted for if ration modeling is to be accurate.
Another pressing issue confronting animal agriculture in general, and the dairy industry in particular, is that of nutrient management. Nitrogen is now regarded as an environmental pollutant. At the heart of efficient milk production is protein and amino acid nutrition, of which nitrogen is a key element. It is well known that cows are notoriously poor in nitrogen efficiencies, and the key to improving those efficiencies lies in the proper balancing of metabolizable amino acid.
Ration modeling that balances for many of the amino acids is continually being updated. The newer models will focus much more heavily on the metabolizable amino acids rather than on crude protein levels.
As continued research helps elucidate the amino acid requirements, they will be incorporated into the new, more dynamic models enabling nutritionists to predict the rumen bypass fractions of proteins and the metabolizable amino acids.
Computer modeling for dairy cow nutrition and milk production has become an industry unto itself. Scientists, researchers, economists, mathematicians and software engineers in both academia and private industry have come together to develop incredibly sophisticated models that do an excellent job of predicting milk production and the economics associated with diet changes.
At the heart of computer modeling are the mathematical equations that predict various functions of a cow’s metabolism based upon the research gathered on a cow’s biology. The data generated by feed analysis laboratories continues to aid in fine-tuning those equations to formulate more sophisticated diets.
While the models become larger, more sophisticated and complex, Sniffen advises that the commercial models should remain relatively uncomplicated for consultants to use in the field. As more inputs are needed and more calculations must be made to solve for a solution, we run the risk of making the models too laborious and slow.
In today’s high-speed world where we require information and answers almost instantly, slowing down the process, even though it may attain an added level of accuracy, may be counterproductive to the consultant in the field.
At the same time, according to Sniffen, the door is wide open for continual improvement of modeling, especially in the areas of transition, dry cow, heifer and calf nutrition. Improved submodels for amino acid metabolism and fatty acid metabolism are also in the works.
Sniffen is cautiously optimistic that updated versions of both CNCPS (Cornell Net Carbohydrate and Protein System) and the Dairy NRC (National Research Council) will be coming out in 2018.
The dairy industry worldwide will become increasingly more competitive during the 21st century. In order for the U.S. dairy industry to secure a competitive place in the growing global market place, it must figure out how to produce dairy products more efficiently.
With the multitude of different feedstuffs and commodity byproducts available for dairy cow diets, there will continue to be a growing need and reliance on computer modeling that will enable producers to arrive at the most cost-effective and efficient diets for their herds.
The U.S. is not the only country with dairy industries seeking to gain an economic advantage on the world stage. The country or countries that can come up with a better way of modeling milk production and feeding efficiencies while focusing on environmental sustainability stand to gain a distinct economic advantage.
Meeting the future food needs in our world will depend upon how efficiently that food can be produced. The use of sophisticated tools such as dairy ration modeling will be just one more way the U.S. dairy industry will remain healthy in the increasingly more competitive and growing worldwide marketplace. 
-
John Hibma
- Consulting Ruminant Nutritionist
- South Windsor, Connecticut
- Email John Hibma
