Improving Feed Efficiency to Promote a Sustainable Dairy Industry

Summary

Dairy feed efficiency, as defined by the fraction of feed energy and protein captured in products, has more than doubled in many countries over the past 100 years. This increase occurred mostly because of a focus on improving production per cow, but, for the future, we must focus more directly on efficiency. Efficiency can be improved further by using genomic technologies to select for more efficient cattle and by feeding cows to more precisely meet nutrient requirements during different stages of the lactation cycle.

Introduction

Dairy feed efficiency in North America has doubled in the past 100 years, largely as a byproduct of selecting and managing cows for increased milk production (VandeHaar and St-Pierre, 2006, Capper et al., 2009). Milk synthesis is an efficient process, and increasing milk per cow results in a greater percentage of total feed intake being used for milk instead of cow maintenance. The current average milk production of dairy cattle in North America is ~10,000 kg/cow/year, and elite dairy cattle currently partition >3 times more feed energy toward milk than toward maintenance over their lifetime (VandeHaar and St-Pierre, 2006). Much of the gain in feed efficiency from increasing productivity has already occurred in these cattle. Future increases in feed efficiency will still rely on increasing productivity, but we must specifically target feed efficiency as a farm goal. More detailed reviews of dairy feed efficiency have recently been published (Berry and Crowley, 2013, Connor, 2015, Pryce et al., 2014, VandeHaar et al., 2016). In this paper, methods to further improve feed efficiency in the modern dairy cow will be briefly discussed. These include 1) selection for efficient genetics using genomics and milk output relative to body size, and 2) management to take advantage of the genetic potential of superior cattle.

Defining feed efficiency

Feed efficiency is a complex trait for which no single definition is adequate (refer to VandeHaar M.J. et al., 2016 open access). Feed efficiency should be considered over the lifetime of a cow and include all feed used as a calf, growing heifer, and dry cow and all products including milk, meat, and calves. At the farm level, feed efficiency also should account for feed that is wasted and for products that are not suitable for human-consumption. We should also consider inputs of human-consumable vs other foods, fossil fuels, water, and land, and outputs of greenhouse gasses, pollutants, fertilizers and other products not used for human consumption. How we feed dairy cattle also impacts ecosystem services (such as availability of wildlife habitat), rural aesthetics and sociology, soil conservation, food quality and healthfulness, food security, animal well-being, the need for imported oil, and how many beef cows are needed to produce calves. Developing a metric that includes all relevant factors for feed efficiency would be difficult. Thus, we use simple metrics such as milk to feed ratio, feed cost per unit milk, and income over feed cost. Whereas protein could be considered the most important component of milk, energy intake generally limits milk production, and feed energy includes the energy of protein. Thus, this paper focuses on energetic efficiency.

Gross energy (GE) is the total chemical energy of a feed but some of it is lost as the chemical energy in feces, gasses, and urine, and some is lost as the heat associated with the metabolic work of fermenting, digesting, and processing nutrients. The remaining chemical energy is known as net energy (NE). Some NE is used to support maintenance functions and is subsequently lost as heat. Some NE is the chemical energy of secreted milk or accreted body tissue and conceptus. For this paper, Gross Feed Efficiency (GEff) is defined as the energy captured in milk and body tissue divided by the GE consumed by a cow in her lifetime, and it is highly correlated to milk energy output per unit body weight (BW). GEff is a useful way to examine how well the dairy industry is stewarding resources.

The major components affecting feed efficiency can be divided into 1) those that alter maintenance and the dilution of maintenance, or the portion of NE that is captured in milk or body tissues instead of used for maintenance, and 2) those that alter the conversion of GE to NE, which include diet and cow effects.

Selecting for cows that capture NE more efficiently

The typical Holstein cow has a maintenance requirement of ~10 Mcal of NE/day (equivalent to ~25 Mcal of GE and ~6 kg of feed). This feed is used for basal life-sustaining functions even if the animal is not producing milk, growing, working, or pregnant. Any extra feed consumed above that needed for maintenance, can be converted to milk or body tissues. If the cow eats twice as much feed as she needs for maintenance, so 2X maintenance, then only half of her feed NE intake is used for maintenance with the remainder used for milk. As she eats even more feed, the portion used for maintenance becomes a smaller fraction of total feed intake. This “dilution of maintenance” increases efficiency and has been known for a long time (Freeman, 1975, VandeHaar et al., 2016). However, the marginal increase in efficiency from diluting maintenance diminishes with each successive increase in feed intake, and efficiency likely plateaus at ~5X maintenance intake (Figure 1). Production relative to maintenance can be increased by increasing production or by decreasing maintenance. Maintenance is highly correlated to a cow’s body weight, and, over the past 50 years, the body size of dairy cattle has increased. Because of this, the US genetic base for body size traits in all dairy breeds is continually being adjusted up. However, our latest analysis on 5000 Holsteins in mid-lactation (using the dataset of Tempelman et al., 2015) demonstrated no genetic correlation between BW and milk energy output (VandeHaar et al., 2014); moreover, BW was genetically correlated negatively with GEff. In a smaller subset of that data, Manzanilla-Pech et al. (2015) showed that milk energy output had zero or negative genetic correlations with BW and that stature was genetically correlated negatively with GEff. The fact that cows have gotten larger over the past 50 years is counter to the goal of increasing efficiency. We suggest that the best way to improve efficiency is to select cows for greater milk production and smaller BW together.

Figure 1. Change in Gross Feed Efficiency (GEff) as intake increases. Each intake multiple of maintenance will be about 10 Mcal of NEL and 5 to 6 kg of feed DM. The dashed line depicts the expected GEff based on the NRC (2001) maintenance requirement and demonstrates that GEff diminishes with increasing intake. This expected response assumes no depression in digestibility as intake increases. Thus, the actual response should plateau even sooner and at a lower level. The symbols depict the measured GEff in ~5000 mid-lactation Holstein cows assuming a feed GE value of 4.5 Mcal/kg (unpublished data from the study of Tempelman et al., 2015). The solid line is the trend line for GEff and demonstrates that the diminishing response is GEff = -0.10 + 0.13 x MM - 0.0094 x MM2, where MM = Multiple of Maintenance.

Using the projected curve for the dilution of maintenance, efficiency should begin to plateau as cows achieve about 5X maintenance. However, this projection is overly optimistic because as cows eat more, the percentage of feed that is digested is depressed (NRC, 2001, Huhtanen et al., 2009). According to the equations used in the NRC (2001), efficiency peaks at ~4X maintenance intake, which is ~45 kg milk (3.5% fat) per day for a 680-kg-cow. The NRC (2001) model discounts digestibility too much at high intakes, as shown in Huhtanen et al. (2009) and supported by Figure 1, so efficiency might not be close to maximal until cows produce at least 60 kg for 680 kg BW. Importantly, however, feed efficiency should be considered on a lifetime basis, so we must account for body tissue gain and the feed consumed as a heifer and dry cow. Based on the theoretical model of VandeHaar (1998), lifetime GEff is ~20% for cows producing 10,000 kg milk/year (the current US average), 25% for cows at 20,000 kg/yr, and would likely never exceed 30%. Certainly, substantial gains can still be made in lifetime feed efficiency from increasing production relative to body size. However, top North American herds are at a point where the return in efficiency from further gains inproductivity will be smaller than they have been in the past. Thus, along with continuing to breed for more milk per unit BW, we should also develop new methods to select for feed efficiency directly and focus on ways to save on feed inputs through better feed management and nutritional grouping.

Selecting for cows that convert GE to NE more efficiently.

One way to select for feed efficiency as a breeding goal, independently of production level, is to select for low Residual Feed Intake (RFI). RFI is a measure of actual versus predicted intake for an individual and is essentially “unjustified feed intake” and has been previously described (Berry and Crowley, 2013, Pryce et al., 2015, Connor, 2015, Tempelman et al., 2015, VandeHaar et al., 2016). Usually, RFI is determined statistically as the deviation of actual dry matter intake (DMI) of a cow from the average DMI of other cows that are fed and managed the same (Cohort) after adjusting for the major energy sinks of BW (related to maintenance), milk energy output, and body energy change, where the residual error term is RFI. Thus, RFI includes error that is true variation amongst cows due to genetics, true variation that is due to permanent environmental effects, and variation from measurement error. Cows that eat less than expected have negative RFI, and thus are desirable when comparing animals for selection purposes as long as RFI is only seen as one factor to use in selecting for efficiency. Selecting for high milk production relative to BW also remains an important selection criteria.

Based on our data examining the GEff of cows compared to their level of production (Figure 1), efficiency varies considerably among cows within a production level. This variation can also be examined in intake units, or RFI. Whereas part of the variation in RFI is error in measurements, some RFI is biological with a heritability of 0.17 based on 4900 cows (Tempelman et al., 2015). During the 20th century, selection of superior genetics relied heavily on quantification of the phenotype in daughters of young sires; sires with outstanding daughters were deemed genetically superior. Because DMI cannot be measured easily and routinely on individual cows in commercial farms, direct selection for feed efficiency was impossible. Genomic selection enables selection for traits like feed efficiency for which daughter phenotypes are unknown. Excellent reviews on the general methodology of genomic selection are Eggen (2012) and Hayes et al. (2010).

The biological basis for variation in RFI is not clear. This variation is associated mostly with the conversion of GE to NE, and thus is due to differences in digestibility, methane production, urinary energy losses, and metabolic pathways involved in processing nutrients. Variation among cows in the actual maintenance requirement relative to MBW also contributes to RFI. Additional thoughts on the variation amongst cows can be found in VandeHaar et al. (2016) and Herd and Arthur (2009). However, regardless of the reasons that some genotypes are more efficient, genomic selection can and will be used.

Evidence that genomic selection for RFI can work in the dairy industry has been demonstrated by Davis et al. (2014). In their study, cows selected to be in the top 10%, compared to the bottom 10%, for RFI genotype needed 0.6 kg less feed to produce the same amount of milk. This is similar to the expected savings in feed for maintenance in a cow weighing 80 kg less. The use of genomics in selection against RFI or DMI is already beginning in Australia (Pryce et al., 2015) and the Netherlands (Veerkamp et al., 2014) and will likely occur in North America in the near future. If selection for efficiency is to be realized by selection for RFI, RFI should be a repeatable trait across diets, climate conditions, lactations, stages within a lactation, and even stages of life. Data to date suggest that it is (Tempelman et al., 2015, Potts et al., 2015, Connor et al., 2013, MacDonald et al., 2014). It is important to note that RFI is only part of feed efficiency. Selection for efficiency must also consider the optimal levels of milk production relative to body weight. The approach used by Pryce et al. (2015) seems reasonable, with an index to select against body size and against RFI while also selecting for milk yield and composition. Improvements in feed efficiency must not occur at the expense of health and fertility of dairy cows. Many traits must be optimized as we consider the ideal cow of the future to promote profitability of farms and sustainability of the dairy industry (refer to VandeHaar M.J. et al., 2016 open access).

Managing for feed efficiency

Using the model described in VandeHaar (1998), the impacts of various management changes on efficiency were predicted. Increasing average daily milk production by 10% or increasing cow longevity from 3 to 4 lactations is expected to increase lifetime energetic efficiency ~0.7%. Reducing feed use by 2% with no change in milk production, by selecting against RFI or for smaller cows, or decreasing feed wastage, would improve energy efficiency ~0.5%. Reducing the age at first calving by 2 months, or reducing calving interval by 1 month would increase lifetime efficiency ~0.3%. How cows are fed and managed at each stage of life can alter milk yield per day of life and thereby dilute maintenance and increase efficiency. These management changes promote similar improvements in the efficiency of converting feed protein to milk or body protein. However, the single biggest impact farms could make on efficiency of protein use is to simply quit overfeeding protein, as is often done in late lactation. Feeding cows past 150 days postpartum a diet with 2 percent less protein (15 vs 17% CP) would increase efficiency of lifetime protein use by ~1.3%.

Nutrient requirements vary as lactation progresses, and the optimal diet for maximum efficiency and profitability changes as well (NRC, 2001, Allen and Piantoni, 2014). The widespread adoption of totally mixed ration (TMR) feeding in North America has improved productivity and efficiency because cows eat a consistent diet, but cows are less likely to receive a diet that matches their individual requirements. This is especially true if all lactating cows are fed the same TMR; feeding a single TMR across lactation can never maximize production and efficiency. A single TMR is usually formulated for the higher producing cows and is more nutrient-dense than optimal for cows in later lactation, resulting in inefficient use of most nutrients for these cows. In addition, although a single TMR is formulated for the high producers, it will not maximize milk income over feed costs for the herd because forages, grains, and expensive supplements cannot be allocated optimally.

Contreras-Govea et al. (2015) found the two major constraints to nutritional grouping on commercial farms were that “It makes things too complicated” and “Low diets decrease milk yield”. Thus, in my opinion, the job of a nutritionist is to 1) develop diets that consistently meet needs optimally for fresh, peak, and maintenance groups and demonstrate their benefits, 2) use supplements, metabolic modifiers, feed additives, and low cost alternative feeds to improve efficiency within groups, 3) help farms make rules based on milk and BCS for moving cows and design systems to track cows, and 4) develop protocols for feeding an extra diet.

The optimal number of rations on any farm depends on many factors, but we recommend at least three based on feeding goals and cow biology. The regulation of voluntary feed intake must be considered in diet formulation. Intake is likely limited by hepatic oxidation of fuels in fresh cows and perhaps late lactation cows, but by rumen fill throughout much of the duration of lactation (Allen and Piantoni, 2014). The goal in feeding cows around parturition is optimal health, so expensive supplements are warranted. The goal in feeding cows for the first half of lactation, which includes peak lactation, is maximal milk, rebreeding, and health; these cows should be fed minimum fiber diets with plenty of digestible starch to maximize energy intakes, and again expensive supplements, including amino acids, might be warranted. Cows in later lactation, after replenishing their body stores to a body condition score of 3, should be fed to optimize milk and maintain body condition; they should be fed less fermentable starch and more fermentable fiber to promote partitioning of nutrients toward milk instead of body tissues (Allen and Piantoni, 2014, Boerman et al., 2015). Late lactation cows also should be fed lower protein diets to maximize efficiency of protein use (NRC, 2001); expensive supplements are almost never needed. Nutritional grouping and multiple TMR undoubtedly do increase capital, management, and labor costs; however, feeding cows according to requirements enhances production, efficiency, profitability, and sustainability of the industry (VandeHaar and St-Pierre, 2006).

Should we choose diets that enhance the conversion of GE to NE?

Diet can significantly and directly alter GEff. Fats are energy-dense, and some feeds are simply more digestible than others. If attaining the highest GEff was the goal of the dairy industry, all dairy cattle would be fed diets high in grains and fats with minimal forage and byproduct feeds. However, for purposes of efficiency, sustainability, and profitability, feeding to maximize GEff is illogical. Diet composition can alter a cow’s voluntary feed intake, (Allen and Piantoni, 2014), and therefore, because feed intake alters the dilution of maintenance, the effect of diet on GEff is not always easy to predict. Moreover, energy-dense feeds are often expensive, even on a per unit of energy basis, so energy diets might be less profitable even if they increase efficiency. Finally, one of the values of the ruminant system is its ability to obtain energy from fibrous feeds, such as forages and high-fiber byproducts. From a global perspective, forages and byproduct feeds high in fiber should be fed to dairy cattle when possible and grains and fats should be fed only as needed to optimize production and health. For making decisions about feeding and management, GEff is almost never useful; instead metrics such as income over feed costs on a farm basis, the efficiency of using human-consumable foods, or milk per acre, seem more reasonable. For making decisions about animal selection, however, GEff and IOFC or milk per acre would be highly correlated.

Efficiency and stewardship

Life Cycle Analysis (LCA) is used to best assess the environmental impact of management decisions. Thomassen et al (2008) compared conventional and organic Dutch dairy farms. Conventional farms used 60% more energy and caused 50% more eutrophication per unit of milk produced. However, with no difference in climate change gasses. However, organic farms required 40% more land to get the same amount of milk. Climate change gasses were similar per unit of milk, but in our view, the higher productivity of the conventional farms would spare land for biofuel production or carbon sequestration. Thus, the conventional dairies would have had less negative climate change impact. This is consistent with an LCA study by Capper et al. (2009) showing that the US dairy industry has decreased greenhouse gas emissions 60% per unit of milk in the last 60 years, mostly because of the enhanced feed efficiency from higher productivity. Thus, increased lifetime productivity increases efficiency, and increased efficiency generally is good for the environment – more people can be fed with less resources and less negative environmental impact. A recent FAO-report (FAO, 2010) shows that even scientists who are not part of the US dairy science community agree with this view. Improving efficiency of milk production by using new technologies seems the responsible thing to do for the environment, at least in the foreseeable future.

Conclusion

We have made major gains in feed efficiency in the past 50 years as a byproduct of selecting, feeding, and managing cows for increased productivity, which dilutes maintenance. Average production is currently ~10,000 kg/yr in North America, and most cows have the genetics for even higher production. We must harness the genetics of the current dairy cattle population to improve feed efficiency even further and help feed people sustainably. Better feeding and management may be especially helpful for many lower producing herds to help them achieve at least 15,000 kg/yr. To enhance efficiency further, we should take advantage of new genomic tools that will enable us to select for cows that require less feed per unit of milk by using a selection index that favors greater milk production and components, smaller cow size, and negative RFI.

Stay naturally informed with the latest from BIOMIN!