Milk, fat and protein yields are the primary drivers of dairy farm revenues. Knowing yields at the herd level helps owners estimate their milk checks and guides economic decisions. Knowing yields at the cow level helps managers estimate how a cow’s health, nutrition, social environment and milking conditions are affecting her production level. Yields inform daily decisions such as milking management, operations management and culling, and yields have long-term implications for business planning and genetic progress. Arguably, milk and component yields are the most important metrics producers have for farm success and cow health. How do we measure yields, and how can we be confident in those numbers?
How do we measure milk yield?
Dairies that want to know their yields participate in dairy herd information (DHI) testing, which is performed by trained technicians using certified and calibrated meters. Typically, only one milking in a day is sampled, so this reliably measures only part of her yield. A cow’s total daily (24-hour) yield is estimated from these partial yields. Our challenge is: There are several methods of estimation currently in use.
The earliest methods date to the 1960s and are as simple as multiplying the partial yield by the number of times a cow is milked that day. While that makes intuitive sense, we realized it is more complicated. Sophisticated models that account for factors such as days in milk (DIM), milking interval and a.m. or p.m. sampling, along with a variety of statistical approaches, have since been developed. These partial yields and estimated daily yields are often not stored in herd management software but are condensed into an average yield over multiple days.
Daily yields are further extrapolated to project total lactation yields. A cow’s actual milk yield at a standard lactation length of 305 DIM is predicted from test-day data and then corrected to the mature equivalent (ME) for her breed. This correction considers factors that can affect her yields such as lactation number, age and season of freshening, previous days open and geographical region. Corrected lactation yield estimates are called 305-MEs and allow fair comparisons among cows for culling and management decisions, as well as individual genetic evaluations. Some farms have in-line milk meters that record milk weights for every milking. As a result of this continuous data flow, herd management software providers have developed tools that can even calculate new 305-MEs on a daily basis. As with daily yields, there are several methods for calculating 305-MEs.
With no standard practice for estimating yields, milking equipment manufacturers, on-farm software platforms and dairy record processing centers all use a variety of equations and assumptions to predict daily and lactation yields. This means the same cow could have completely different yield estimations depending on the sources. Often, herd management software will report 305-MEs from multiple sources for each cow, all of which measure different things in different ways with different accuracies. For management purposes, it is best to monitor one type consistently.
Another important consideration is that these daily yield and 305-ME estimates flow into the National Cooperative Database for genetic evaluations as if they represent the same thing – when in fact, they might not. This challenges the accuracy of the traditional 305-MEs and also the standardization of lactation records required for genetic evaluations administered by the Council on Dairy Cattle Breeding (CDCB).
How does this impact herd management?
Yield estimation has a “cascade effect,” where it trickles down to impact nearly all aspects of dairy production, as shown in Figure 1.
The importance of accuracy cannot be overstated, as any biases that may be introduced can occur early in this system. This is especially important for component estimation, where sampling and analytical biases, as well as less-frequent data points, can affect the predictions of fat and protein yield. Herd owners and managers make important economic and management decisions within cows’ first 120 DIM, and accurate predictions of lactation yields in early stages of lactation are a major opportunity for dairy producers.
How can we improve the situation?
Thirty years ago, standard ME correction factors for estimating yields were developed and made available to address this challenge. As the dairy industry, management practices and the cows themselves have changed, so must the parameters used in our system of estimating yields.
It is critical that we update ME factors to reflect the dairy world in 2022 and evaluate new methods to accurately estimate daily yields, predict future yields and enable fair comparisons between cows. A joint venture among the CDCB, National Dairy Herd Information Association (NDHIA) and the USDA-ARS Animal Genomics and Improvement Laboratory (AGIL) has started, with the goals to improve predictions and develop reliable tools for management purposes (Figure 2).
Click here to view the image above at full size in a new window.
The leading motive in this joint project is to fine-tune early lactation yield predictions to be as close to actual lactation yields as possible. The results of this work will be improved estimations of daily yields from partial yields, improved predictions of actual lactation yields from test day yields and improved standardization of lactation yields so cows can be comparable for selection and genetic evaluation. These new methods and correction factors will be made publicly available in the hope that a standard approach can be widely adopted. This project is just beginning and will profit all herd owners and managers by delivering accurate yield predictions for the modern U.S. dairy cow and empowering producers to confidently know their yields.
