C. A. Rotz 2
USDA, ARS, Pasture Systems and Watershed Management Research Unit, University Park, PA 16802-3702
Reduction of nitrogen loss in animal production requires whole-farm management. Reduced loss from one farm component is easily negated in another if all components are not equally well managed. Animal excretion of manure N can be decrerased by improving the balance of protein or amino acids fed to that required by individual animals or animal groups or by improving production efficiency. Management to increase milk, meat, or egg production normally improves efficiency by reducing the maintenance protein required per unit of production. Large losses of manure nitrogen occur through the ammonia and nitrous oxide that are emitted into the atmosphere and the nitrate leached into groundwater. Up to half of the excreted nitrogen is lost from the housing facility, but this loss can be decreased through frequent manure removal and by avoiding deep litter systems and feedlots. Techniques such as acid treatment of manure, scrubbing of ventilation air, and floor designs for separating feces and urine substantially reduce ammonia emissions, but these practices are often impractical or uneconomical for general use. Manure storage units improve nutrient utilization by allowing better timing of nutrient application with crop needs. At least 70% of the nitrogen entering anaerobic lagoons is typically lost, but a less than 10% loss can be maintained using slurry storage with a natural crust or other cover, or by drying poultry manure to at least 50% dry matter. Irrigation and surface spreading of manure without soil incorporation often ensures the loss of all remaining nonorganic nitrogen (typically, 20 to 40% of remaining nitrogen). Rapid incorporation and shallow injection methods decrease this loss by at least 50%, and deep injection into the soil essentially eliminates this loss. For grazing animals, excessive loss can be avoided by not overstocking pastures and avoiding late fall and winter grazing. Reducing emissions between the animal and the soil can lead to greater leaching and denitrification losses from the soil if this additional nitrogen is not used properly. The use of a crop rotation that efficiently absorbs these nutrients and applying nitrogen near the time it is needed by crops reduce the potential for further loss. Maintaining the proper number of animals per unit of land available for manure application is always crucial for efficient recycling of nitrogen. Our understanding of nitrogen loss processes is improved through modeling, and computer models assist in the development of integrated systems for efficient and economical nitrogen use in animal production.
Key Words: Ammonia • Denitrification • Farm Production • Leaching • Management • Nitrogen Loss
Source: J. Anim. Sci. 2004. 82:E119-E137
Nitrogen is an essential element in animal production. Largequantities of N are required for the growth of feed crops. CropN, primarily in the form of protein, is then an essential feedcomponent for animal growth and development. Most of the N consumedis excreted by animals, providing manure nutrients needed forcrop growth. The problem in this cycling of N is that largelosses normally occur that contribute to the degradation ofour environment (Figure 1). The challenge is to manage the animals,crops, and other farm components to efficiently use availablemanure N, and thus reduce the potential loss to the environment.
Figure 1. Major nitrogen flows in animal production within the farm and between the farm and its environment.
Reducing N loss from the farm must begin with proper animalfeeding and management to reduce N excretion. Even with goodmanagement though, large quantities of N are in the manure.A major portion of this excreted N can quickly transform intoammonia, which may readily volatilize and move into the atmosphere.Volatile loss begins soon after excretion, and it continuesthrough all manure handling processes until the manure nutrientsare incorporated into soil. If steps are taken to maintain theN until it is incorporated, leaching and denitrification lossesof soil N will increase if that N is not applied at the appropriateamount and time for crop uptake. Only by properly managing allfarm components can we truly increase N use efficiency in animalproduction and reduce N escape into our environment (Figure1; Rotz et al., 1999b).
Management processes are available or under development thatreduce N loss, but implementation often remains a challengedue to various constraints. The major constraint is usuallyeconomic. New management practices often require large investmentsor greater operating costs that are difficult to justify. Currentprofit margins are low in most animal operations, and the directeconomic return for saved nutrients is small considering theirfertilizer replacement value. Labor can also be a constraint.Available labor is often heavily used in animal production,so changes that require more time, particularly for the farmmanager, will not be readily accepted. Environmental issuesthemselves may also be a constraint. For example, reducing ammoniaemissions may lead to greater emission of nitrous oxide. Dueto the potential long-term impact of this greenhouse gas, thenet benefit to society may not be positive, and implementationof the ammonia-conserving technology will be constrained.
The goal of this review is to quantify N losses for variousanimal production strategies and to discuss the major managementoptions available for reducing these losses. Although the materialpresented generally applies to all animal species, the emphasiswill be placed upon dairy, beef, swine, and poultry production.
Strategies for Reducing Nitrogen Excretion
Conservation of N in animal production must begin by improving the N use efficiency of the animals. On dairy farms today, 20 to 30% of the N consumed by the herd is in the protein of the milk and meat produced, and the remainder is excreted in manure (Dou et al., 1996; Kohn et al., 1997; Oenema et al., 2001a). Pasture-fed dairy animals are at the lower end of this range, and pasture-produced beef have a N use efficiency of less than 10% (Hutchings et al., 1996). When finishing beef in a feedlot, about 10% of the N intake is retained in body tissue (Bierman et al., 1999). In poultry or swine production, where the protein needs of the animals can be more closely met, this efficiency may average 30 to 35% and even approach 40% (Mohan et al., 1996; Jongbloed et al., 1997; Lee et al., 1998; Lindberg and Andersson, 1998; Lenis and Jongbloed, 1999; Han et al., 2001). Through various management techniques, these utilization efficiencies can be increased. The maximum possible efficiency varies with animal species, age, stage of lactation, and so on, but this theoretical limit is about 50%.
Thus, large amounts of N are excreted in the production of all animal species (Table 1). By reducing excreted N, losses in all manure handling processes throughout the N cycle are potentially reduced. Nitrogen excretion is directly related to the animal’s N (protein) intake, so less protein must be fed per unit of production. Two general strategies can be used to reduce N excretion. The first is to reduce the protein fed by improving the match between the protein quality fed and that required by the animal. The other is to improve animal productivity. As more milk, meat, or egg product is produced per animal, the maintenance requirement of protein per unit of production is reduced. Thus, the animal product can be produced with less N consumed and excreted. Although improved productivity can increase N use efficiency, greater improvements are generally obtained through strategies that improve protein-feeding efficiency.
Feed protein generally refers to crude protein. Crude protein is defined as N content times 6.25, which assumes 16 g of N/100 g of protein in feeds. Proteins are composed of amino acids, which are required for the maintenance, growth, and productivity of animals (NRC, 1994; NRC, 1998; NRC, 2000; NRC, 2001). Matching the amino acid levels in rations to that required by the animal is very complex. There are 20 primary amino acids in proteins. Different amounts of each amino acid are required, and these amounts vary with animal age and other characteristics. Providing the precise amount of each amino acid required at any point in the animal’s life is essentially impossible, but improved feeding management can approach this goal.
Amino acids are generally classified as either essential or nonessential. Essential amino acids are those that must be obtained directly from the feed. Nonessential amino acids can be synthesized by the animal from other amino acids fed in excess in the diet. Since nonessential amino acid requirements are more easily met, the emphasis in formulating diets must be given to the essential group. Although there are similarities between ruminant and nonruminant species, different systems are used to balance feed protein levels with animal needs.
In nonruminants, such as poultry and swine, the optimal dietary pattern among essential amino acids that corresponds to the needs of the animal is often referred to as the ideal protein (NRC, 1998). This ideal protein or ideal amino acid pattern varies with the characteristics of the animal, such as sex, age, genotype, and production function (Han and Lee, 2000). The digestibility of each amino acid also affects its bioavailability to the animal, which further complicates the matching of available protein to animal needs. This has led to the use of the apparent ideal digestibility of various amino acids in feeds to match the animal’s requirements for those amino acids.
Dietary crude protein can be reduced through the supplementation of synthetic amino acids to reduce N excretion from pigs and poultry (Jongbloed et al., 1997; Sutton et al., 1999; Nahm, 2002). Reductions of 2 to 4 percentage units of dietary protein have been made without decreasing weight gain or feed conversion (Han and Lee, 2000). In a review of swine production data, amino acid supplementation with low-CP diets was found to reduce N excretion by 3 to 62%, depending on the size of the pig, on the reduction in dietary CP, and on the initial CP content in the control diet (Kerr, 1995; Sutton et al., 1999; Han et al., 2001). An average reduction in N excretion per unit of dietary CP reduction is about 8.5%. In a review of poultry data, dietary changes made to reduce CP content using synthetic amino acids reduced N excretion by 10 to 27% in broiler production and 18 to 35% in chick and layer production (Nahm, 2002). The production and use of synthetic amino acids for animal feed has grown rapidly over the past few years, with an average annual growth rate of 6% (Han and Lee, 2000).
With less N excreted, the potential for N loss is reduced. Most excess excreted N is in a form that is easily transformed into ammonia (Han et al., 2001), so the potential reduction in N emission during manure handling and storage is at least equal to and likely greater than the reduction in excretion. With a 41% reduction in N excretion, ammonia emission from a swine facility was reduced 47 to 59% (Kay and Lee, 1997).
Dietary carbohydrates can also be manipulated to reduce ammonia emission from swine manure. Including bacterially fermentable carbohydrates in diets has reduced the ratio between urinary N and fecal N (Lenis and Jongbloed, 1999). Because fecal N is less easily degraded to ammonia, emission is reduced. Slurry pH is also reduced, which further reduces the potential for urease activity and ammonia volatilization. Use of raw potato starch in the diets of growing pigs reduced ammonia emission by 13% (Lenis and Jongbloed, 1999).
A linear relationship was found between the intake of dietary nonstarch polysaccharides and ammonia emission (Cahn et al., 1998). For each 100-g increase in the intake of the nonstarch polysaccharides, slurry pH decreased by about 0.12 unit and ammonia emission decreased 5.4%. Thus, replacing cornstarch in the diet with components high in fermentable carbohydrates appeared to increase volatile fatty acid concentration in manure, which lowered pH. Urinary pH can also be reduced by manipulating the dietary cation-anion difference. By adding acidifying Ca salts to the diet instead of CaCO3, urinary pH was reduced by 1.6 to 1.8 units. This reduced ammonia emission by 26 to 53% (Lenis and Jongbloed, 1999).
As animals grow, their nutrient needs change. Phase feeding provides a strategy for improving the match of the diet to the growth stage of the animal. The largest benefit is attained going from a single phase to a two-phase strategy where a 10% reduction in N excretion was found (Hartung and Phillips, 1994; Han et al., 2001). Adding an additional phase to a two-phase feeding strategy for growing pigs was predicted to provide a 6% reduction in N excretion (Lenis, 1989). Multiphase feeding can be accomplished by mixing a diet high in N with a low-N diet in decreasing proportions each week throughout the growth period. Multiphase feeding reduced urinary N excretion in swine production by 15% (van der Peet-Schwering et al., 1996). With less excreted N, ammonia emission was reduced 17%. A combination of three-phase feeding and the use of synthetic amino acids reduced N excretion by more than 30% (Hartung and Phillips, 1994). Using separate diets for pregnant and lactating sows has also shown as much as a 24% reduction in N excretion over a yearly cycle (Lenis, 1989).
Other feeding strategies that have been studied to reduce N emissions include feed additives to inhibit urease activity or to bind ammonia (Lenis and Jongbloed, 1999). In theory, urease inhibitors will increase N utilization, reduce urea degradation, and thus reduce ammonia emission. Binding agents, such as clay minerals, should reduce the volatility of ammonia. Thus far, experimental work with these additives has shown small and inconsistent effects on ammonia emission (Lenis and Jongbloed, 1999).
The digestive process in ruminant animals includes the added step of rumen fermentation. Ruminally degraded protein (RDP) provides a mixture of peptides, free amino acids, and ammonia for microbial growth and protein synthesis (NRC, 2001). Ruminally synthesized microbial protein typically supplies most of the amino acids passing to the small intestine. Ruminally undegraded protein (RUP) is the second most important source of amino acids that are absorbable within the intestinal tract. Thus, the rate of degradation of proteins is very important in supplying the right amount and type of amino acids. Crude protein can be divided into five fractions (A, B1, B2, B3, and C) according to degradation rate (NRC, 2001). The A fraction, which includes nonprotein N, is essentially immediately available for degradation in the rumen, and the C fraction is considered nondegradable or never available. The remaining B fractions are potentially degradable true protein that fall within three ranges of degradation rate. Rumen degradable protein is defined to include the A component and a portion of the B components, and RUP includes the remainder of the B components and the entire C component.
The goal in feeding ruminants such as cattle is to supply the right amount of protein with the proper balance of degradation rates to provide the appropriate amino acids to the intestinal track. Most forage protein is highly degradable, so there is normally little problem meeting the RDP requirement. Total protein can be overfed to meet the minimum RUP requirement, but this leads to the excretion of considerable excess N. This excess N also requires energy to digest, which may reduce animal performance. The better solution is to feed rumen-protected proteins, which bypass the rumen but degrade within the intestinal tract. Industrially produced amino acids in pure form may also be fed to meet specific amino acid needs.
A direct relationship between N excretion and protein intake is well documented. Protein needs of cattle, particularly lactating cows, are normally best met by reducing the total protein in the diet through the supplementation of feeds high in RUP (NRC, 2001). As the balance between animal needs and feed availability is improved, animal performance is normally maintained and perhaps improved. With all diets balanced to meet the RDP and RUP requirements of lactating dairy cows, urine N excreted from a high-protein diet (18% CP) was 2.3 times greater than that from a low-protein diet (12% CP; Tomlinson et al., 1996). Fecal N excretion was only 25% greater using the high-protein diet, which illustrates that excess protein is primarily excreted in urine. In a similar experiment comparing low- (14% CP) and high- (19% CP) diets, laboratory-measured ammonia emissions were about three times greater from the manure excreted on the high-protein diet (Frank et al., 2002). In an experiment in which ammonia emissions were monitored through various storage practices, emissions were reduced 70% using a low (12.5% CP) diet compared to a high (17.5% CP) diet (Külling et al., 2001). Nitrous oxide emissions from stored manure were also reduced using the diet of lower protein.
A simulation study by Rotz et al. (1999b) demonstrated the potential whole-farm benefit of improved protein feeding. Compared to the use of soybean meal as the sole protein source, the use of a feed mix with lower RDP reduced N excretion by 39 kg/cow per year. This reduced volatile N loss from the farm by 27%, with a small reduction in N leaching loss. Using the more expensive but less degradable protein feed improved the annual net return of the farm by $45 to $70/cow depending on other management strategies used. Using a more precise or ideal protein feeding strategy only reduced volatile N loss an additional 9%.
Group-feeding cattle of similar age or stage of lactation provides a strategy to better match the diet to the animal’s nutrient needs. An optimum of six milking animal groups was found for a simulated farm (St-Pierre and Thraen, 1999). Compared to feeding the milking herd in a single group, the use of the six-group strategy reduced N excretion by 8%. In a full lactation study with high-producing dairy cows, a reduction in dietary protein from 17.5 to 16% was possible around wk 30 of lactation without an adverse affect on production (Wu and Satter, 2000). This reduction in protein fed reduced N excretion by about 14% over the full lactation.
The type and amount of forage fed can also affect protein intake and N excretion by cattle. Alfalfa silage is normally very high in degradable protein. In contrast, corn silage is low in total protein and this protein is generally less degradable. Dairy cow diets balanced to meet protein and energy needs using moderate to high amounts of corn silage have provided milk productions similar to those attained with alfalfa-based diets with lower total protein intake and greater protein use efficiency (Dhiman and Satter, 1997). In this full lactation study, high-producing dairy cows on a diet in which one-third of the forage was corn silage excreted 6% less N compared to cows with all of their forage coming from alfalfa silage. With two-thirds of the forage from corn silage, N excretion was reduced by 15%. In a whole-farm simulation study, substituting corn silage for alfalfa silage provided up to a 30% reduction in N excretion for the dairy herd, improving the overall farm balance of N available to that required for crop uptake (Borton et al., 1997).
Diet can also affect the portion of the total N excreted in feces vs. urine. Feedlot cattle fed a diet with 7.5% roughage had 7% more of the total N excreted in the feces, and feeding wet corn gluten feed gave 12% more of the total N in the feces compared to an all-concentrate diet (Bierman et al., 1999). Fecal N has the advantage of being less volatile in the feedlot and during manure handling. In this experiment, however, total N excretion and volatile N loss were lowest with the all-concentrate diet.
Improved Production Efficiency
Steps taken to improve the productivity of animals (rate of gain, milk, or egg production) will normally improve feed efficiency. Thus, growing animals will reach their weight goal faster or gain more weight over a fixed period. Either way, more product is obtained per unit of feed expended on maintenance of the animal. Van Heugten and van Kempen (2000) determined that improving feed efficiency in swine production by 0.1 point (lowering feed per unit of gain from 3.0 to 2.9) resulted in a 3.3% reduction in nutrient excretion (assuming similar growth and nutrient retention). For dairy cows, increasing milk production yields more milk per unit of feed required for animal maintenance. Although increased production often requires increased feed intake, the net result is normally less N intake and excretion per unit of milk produced. A simulated 25% increase in milk production was found to reduce N excretion per unit of milk by 8% (St-Pierre and Thraen, 1999). Reductions in N excretion that are achieved through improved production efficiency are generally small, so such strategies have a lesser role for potential reductions in the N that is lost to the environment.
Productivity increases in dairy production can be obtained using numerous strategies, including genetic, feeding, and handling improvements. Hormone and other injections and various feed additives can also be used. Bovine somatotropin injection, increased milking frequency, and extended photoperiod have all improved milk production. Simulation of these three strategies predicted reductions in herd N excretion per unit of milk produced by 7.8, 7.0, and 3.6%, respectively (Dunlap et al., 2000). Combining all three strategies decreased excreted N per unit of milk by about 15%. If this N is efficiently cycled through the farm, the reduction in what is lost to the environment should be similar or slightly greater than the reduction in excretion. On a simulated dairy farm, a 20% increase in milk production decreased volatile N loss per unit of milk produced by 12% and leaching loss by 16% (Rotz et al., 1999b).
Improving the overall longevity of animals in the herd or flock can also improve N use efficiency a small amount. Steps taken to reduce the replacement rate or age at first production in dairy, cow-calf, and layer operations will allow a given production with the growth and maintenance of fewer replacement animals. For example, reducing the replacement rate of a dairy herd from 35 to 30% will reduce the required replacement animal numbers by 15%. On a farm where all replacements are raised, the manure N excretion for the herd will be reduced by about 5% with about 6% less N loss from the farm (Rotz, unpublished simulation data).
Feed processing can influence feed intake, digestibility, animal production, and the excretion of N. Common processing techniques include grinding and pelleting. Grinding to obtain the proper particle size is particularly important in maximizing feed efficiency in swine and poultry feeding (Nahm, 2002). Reducing particle size from 1,000 to 400 µm was found to improve nutrient digestibility, which reduced N excretion in the manure of finishing pigs by about 30% (Han et al., 2001; Nahm, 2002). For each 1% improvement in digestibility, N waste per kilogram of meat produced was decreased about 1.4% (Han et al., 2001). Pelleting of feeds may also improve average daily gain and the feed efficiency in swine production by about 6% (Hancock et al., 1996; Van Heugten and van Kempen, 2000; Han et al., 2001). This processing of the feed has reduced DM and N excretion in the feces by about 22%. Considering that 20 to 25% of the total N excreted is in the feces, overall N excretion is reduced about 5%.
Feed additives, such as enzymes, antibiotics, probiotics, organic acids, and growth hormones, may also reduce N excretion. Enzymes can improve feed digestibility and nutrient availability. They have been used in swine diets to improve feed efficiency, but their effect on N excretion appears to be small (Han et al., 2001). Antibiotics, probiotics, and organic acids have all been used to improve feed efficiency in swine production. Their effect on N excretion was often a 5% reduction or less, but reductions of up to 25% have sometimes been reported (Han et al., 2001). Growth hormones have had greater effects, reducing N excretion by 12 to 38% (Han et al., 2001).
Housing and Manure Removal
Nitrogen loss begins soon after manure is excreted by the animal. For swine and cattle, urine N is primarily in the form of urea. In poultry, excretions are high in uric acid, which is transformed into urea through aerobic decomposition. When mixed with urease enzymes prevalent in fecal material, urea N can quickly transform into ammonia, which is highly volatile and easily diffused into the surrounding air (Monteny and Erisman, 1998). The rate of this transformation is a function of the total ammoniacal (ammonia plus ammonium) N content, temperature, moisture content, and the pH of the manure (Hartung and Phillips, 1994; Sommer and Hutchings, 1995). As just presented, the ammoniacal N portion excreted is primarily related to the way animals are fed. Urease activity is relatively low below a temperature of 10°C, but it increases exponentially over higher temperatures. If the pH of manure is held below 6, ammonia release is low. At normal pH levels of 7 or more, the ammoniacal N is readily volatilized as ammonia. Urease activity and ammonia loss are also reduced by drying manure to less than 40% moisture content (Groot Koerkamp, 1994; Sommer and Hutchings, 1995). Under typical temperature, pH, and moisture conditions, urea transformation is very rapid, reaching a maximum rate within 2 h of deposition (Monteny and Erisman, 1998). Because of the extra transformation steps required for uric acid, ammonia loss from poultry manure is slower, requiring a couple days to reach maximum transformation and volatilization rates (Groot Koerkamp, 1994).
As ammonia is produced, other factors that control ammonia loss include the exposed manure surface area and the air movement across this surface (Hartung and Phillips, 1994; Sommer and Hutchings, 1995). Depositing manure to a greater depth creates less surface area. With less exposure, ammonia has less opportunity to escape. The volatilization rate from the manure surface is a function of the ammonia mass transfer coefficient and the difference in concentration or partial pressure of gaseous ammonia between that at the surface and that in the ambient air. This mass transfer coefficient is related to the temperature and velocity of the air at the surface. With air movement, more ammonia is carried away from the microenvironment near the surface, increasing the volatilization rate.
Housing design and manure removal practices have a large effect on the rate of N transformation and loss. Strategies are available that reduce this N loss, and new techniques are being developed. Because of the differences in housing design and animal handling among poultry, swine, and cattle, each species will be discussed separately.
A common poultry housing system is the high-rise laying hen house. Normally, excreted manure drops from the cage to the floor or into a pit. This manure is often removed annually, so the housing system includes long-term storage. Composting can occur during this storage period, stimulating greater N emissions (Groot Koerkamp, 1994). Average annual N losses from this type of facility are reported to be about 50% of the total N excreted (Table 2). Use of a deep litter housing system also produces relatively high N loss, but this loss may be half that of the high-rise structure (Groot Koerkamp, 1994; van Horne et al., 1998).
Ammonia loss in the housing facility can be greatly reduced by frequently removing manure from the facility. Manure can be removed by belts or flushing. Belt removal twice a week with or without drying can reduce ammonia emission from the housing facility by 60 to 90%, and more-frequent removal can provide an additional reduction (Groot Koerkamp, 1994; Hartung and Phillips, 1994; van Horne et al., 1998). Flushing provides an estimated 80% reduction (van Horne et al., 1998). Once removed from the structure, though, scraped or flushed manure must normally be stored up to a year, during which continuing loss will occur. Thus, these loss reductions do not describe the full system. Storage losses will be discussed further in the next section. Through simulation of full production systems, the use of manure removal belts with partial manure drying was found to reduce total N emission by 21% (van Horne et al., 1998).
Aviary housing systems have been developed in Europe to promote improved welfare for laying hens. These systems provide greater freedom of movement for the hens, nest boxes, and a dust bathing area with litter. A disadvantage is that these housing systems have up to three times the ammonia emission compared to traditional battery cage systems with daily manure removal (Groot Koerkamp and Bleijenberg, 1998). The increased emission is primarily due to ammonia volatilization from the litter. A litter drying system was tested that reduced ammonia emission to a level slightly lower than that of the traditional cage system (Groot Koerkamp et al., 1998).
Filtration to remove ammonia from air exiting the housing facility is another option for reducing ammonia emissions. Emission reductions of up to 80% can be achieved through biofiltration or scrubbing of the air (van Horne et al., 1998). Simulation of this approach predicted a 26% annual reduction in ammonia emission from a typical layer facility. Practical application of this technology is constrained due to a relatively high cost and technical problems created by the large amount of dust in poultry houses (Groot Koerkamp, 1994).
Swine and cattle can be produced on a deep-litter system, in which bedding such as straw or sawdust is used to absorb and cover urine and feces. A complex decomposition process occurs in the deep litter. The primary microbial processes include aerobic and anaerobic degradation of organic matter, urea hydrolysis, nitrification, denitrification, and N immobilization (Jeppsson, 1999). The bedded pack normally provides 3 to 12 mo of manure storage, with ammonia volatilization occurring throughout this period (Table 2). Microbial processes of nitrification and denitrification can also convert ammonia into inert dinitrogen gas. Normally, conditions are such that these processes do not run to completion, and nitrous and nitric oxides are produced and emitted (Groenestein and Van Faassen, 1996).
Reports on the N emissions from deep-litter or bedded systems are inconsistent. Some report less loss whereas others report more compared to systems that do not use bedding (Hartung and Phillips, 1994). Proper management in tending the bedding can likely reduce ammonia emission. Two deep-litter systems using various amounts and types of sawdust were compared to a traditional slatted floor in fattening pigs (Groenestein and Van Faassen, 1996). In this study, the deep-litter systems were able to reduce ammonia emission by as much as 50% compared to the use of slatted floors; however, total N loss was greater due to the loss of nitrous oxide. Thus, deep litter systems were not environmentally beneficial.
The most common swine housing facility uses a slatted floor to allow fecal and urine excretions to drop into a pit below the floor for removal. The portion of the floor area that is slatted affects ammonia emission. Reducing the slatted floor area from 50% of the pen floor to 25%, reduced ammonia emission by 20% in the rearing of pigs (Aarnink et al., 1996). During the fattening of pigs, emission was reduced by only 10% with this change in floor design. Aarnink et al. (1996) concluded that reducing the slatted floor and slurry pit area in swine housing decreased ammonia emission from the slurry pit but increased the fouling and emission from the floor. Combining a partially slatted floor over a slurry channel with a sloped floor that is flushed several times a day reduced ammonia emissions by 30% (Hartung and Phillips, 1994). By maintaining flushing liquid in the channel, up to a 70% reduction in ammonia emission was maintained relative to a traditional slatted floor (Hartung and Phillips, 1994).
The ventilation rate of the housing facility can affect the rate of ammonia emission. Emission from pig slurry stored under a slatted floor was approximately doubled by increasing the air exchange rate from two air changes per hour to seven (Hartung and Phillips, 1994). Air temperature must also be considered, though, because warmer temperatures in the facility due to low ventilation rates will increase volatile loss. Thus, optimal management of air temperature and ventilation can reduce N loss while maintaining or improving animal productivity.
Ammonia concentrations in swine confinement buildings have been reduced using an acid and oxidation treatment (Jensen, 2002). The treatment included spraying a water and sulfuric acid mixture on the floor, adding sulfuric acid to the manure, and oxidizing the manure. Ammonia concentrations in the facility were reduced from 8 to 10 ppm down to 1 to 2 ppm and pig performance was improved. The reduction in ammonia emission kept more N in the manure, improving its fertilizer value.
Common housing systems for cattle include tie stall barns, free stall barns, and open feedlots. The lowest ammonia emission is normally found in a tie stall facility (Table 2). Because animal movement is limited, most manure is deposited in a relatively small area and normally in a deep gutter. With less exposed surface, volatile loss is reduced. In addition, manure is normally removed on a daily basis, allowing less time for loss to occur. In an evaluation of 34 dairy farms in Sweden, ammonia concentrations in tie stall barns with solid manure handling were about half those in free stall barns with liquid manure (Swensson and Gustafsson, 2002).
Deep litter or a bedded pack is sometimes used, particularly for housing young stock. In a comparison of a bedded area and a flat solid floor with manure removed three times a week, ammonia emission from the solid floor was 40% of that from the bedded area (Jeppsson, 1999). Use of a 60% peat and 40% chopped straw bedding mixture reduced ammonia emission by 57% compared to long straw, which provided an emission similar to that from the solid floor. As stated earlier, a major concern with deep litter is that incomplete nitrification and denitrification processes produce two environmentally undesirable gases, nitrous oxide and nitric oxide.
The most common housing for dairy cows, particularly on larger farms, is the free stall system. Cubicles are provided for resting, and animals have the freedom to move to the feeding area on an open floor. Manure is primarily deposited on the open floor. Common floor designs are a solid floor that is frequently scraped or flushed or a slatted floor from which manure drains between the slats into a pit below. Nitrogen emission from these floor systems is nearly all in the form of ammonia. Under warm conditions with fecal and urine materials well-mixed on the floor, N loss is high, with most of the urea transforming into ammonia and volatilizing to the atmosphere. Under cold winter conditions, N loss is relatively low. On average, about 16% of the excreted N is lost from the free stall area (Table 2). Loss from a slatted floor may be a little greater than that from a scraped floor (Kroodsma et al., 1993; Monteny and Erisman, 1998), but the pit underneath the floor in this type of facility normally provides long-term manure storage. Thus, a complete comparison must include storage loss as well.
On solid floors, floor shape and surface characteristics can influence ammonia loss. A small, 3% slope of the floor allows urine to drain away from the feces, reducing ammonia emission by 21% compared to solid or slatted level floors (Braam et al., 1997a). A double-sloped floor with a urine gutter in the center reduced ammonia emissions by 50% (Braam et al., 1997b). Spraying this floor with water following scraping provided further reduction, with a total emission reduction of 65% compared to the reference floors. A grooved solid floor system was evaluated that included perforations through the floor. The grooves enabled urine to move away from fecal material and then drain through the floor perforations. Compared to a traditional slatted floor system, the grooved and perforated floor reduced ammonia emissions by 46% (Swierstra et al., 2001). Using smoother, coated floors reduced urease activity but had little effect on overall ammonia emission (Braam and Swierstra, 1999).
Manure is removed from solid floors by mechanical scraping or flushing with water. Scraping of solid or slatted floors appears to have little effect on ammonia emission (Kroodsma et al., 1993). The spreading and mixing of feces and urine caused by scraping maintains rates of urease activity and ammonia emission that are similar to those of a manure-covered floor. Thus, increasing the scraping frequency had little effect on ammonia emission (Braam et al., 1997a). Flushing reduced ammonia emission up to 70% immediately after the flush, and short and more-frequent (every 2 h) flushing cycles provided the greatest reduction (Kroodsma et al., 1993). Further analysis indicated that flushing could reduce average annual ammonia emissions from the facility by 14 to 17% (Ogink and Kroodsma, 1996). Compared to traditional scraped or slatted floor systems, the amount of manure was roughly doubled through the inclusion of the flush water.
Greater reductions in ammonia emission can be obtained with less flushing solution by adding a disinfectant or acid to deactivate or slow urease activity. Spraying a dilute formaldehyde solution on slatted or sloped solid floors has reduced long-term ammonia emissions by 50% and 87%, respectively (Ogink and Kroodsma, 1996; Monteny and Erisman, 1998). Reductions of up to 60% were achieved by a combination of the acidification of slurry in a shallow pit and regular flushing of the slats with the acidified slurry (Monteny and Erisman, 1998). Animal and human safety is a major concern when using these treatments, so a safe flushing solution must be developed to allow practical application of this kind of technique.
The greatest housing N loss occurs when cattle are on a feedlot (Table 2). Losses of 40 to 90% of the excreted N are reported by the time feedlot pens are cleaned (Eghball and Power, 1994; Bierman et al., 1999). Most of this loss is emitted into the atmosphere, but portions are also lost through runoff from rain and leaching into the soil below the feedlot. In the drier climate of the Great Plains, 3 to 6% of the excreted manure N is lost in runoff (Eghball and Power, 1994). Although nitrate leaching into groundwater does not appear to be a problem, very high nitrate-N levels have been measured in soils under abandoned feedlots (Eghball and Power, 1994). In a comprehensive feeding experiment, Bierman et al. (1999) found that 9 to 19% of the N excreted by cattle on various finishing diets was removed in the manure scraped from the lot at the completion of the study. Nitrogen lost in runoff was 5 to 19% of the excreted N, with 10 to 16% leached into the soil. The remaining 57 to 67% was assumed to be lost by volatilization. Volatile loss would primarily be in the form of ammonia, but denitrification products would also occur.
Long-term storage of manure is an important step in improving nutrient utilization and reducing N loss in animal production systems. Storage allows more timely application of manure nutrients. When N is applied just before seeding or when the crop is actively growing, it is better utilized by the crop. Applying large amounts of manure within a short time period also provides more convenience for rapid incorporation of the nutrients into the soil. With rapid incorporation, volatile N loss in the field and the potential loss in surface runoff are reduced.
A number of storage methods are used, which can greatly affect the amount and type of N lost. A major factor affecting N loss is manure DM content. Storage methods can be categorized as solid, slurry, and liquid storage, and typical DM contents are greater than 15%, 7 to 15%, and less than 7%, respectively. Other manure characteristics that affect N loss include total N concentration, ammoniacal N concentration, and pH. With a higher concentration of N in the manure and, in particular, ammoniacal N, the rate of loss and the potential total loss of N increase. The transformation and loss of ammonia are very sensitive to manure pH; relatively low loss occurs below a pH of 6 and very high loss occurs when the pH exceeds 8 (Muck and Steenhuis, 1982). Environmental factors, such as ambient temperature, wind velocity, and solar radiation, also affect the rate of loss from open storages (Sommer, 1997). These effects on loss and methods for reducing loss will be addressed for each storage type.
Solid manure is produced when large amounts of bedding are used to absorb manure moisture or when the manure is partially dried. Solid manure can be stored in a stack or heap. During storage, some level of aerobic decomposition or composting will occur. The amount of decomposition and the N loss that occurs is largely related to the manure DM content, the carbon-to-N ratio, and the amount of aeration that the manure stack receives. A stack of undisturbed poultry manure at a DM content of 50% or more is reasonably stable, with relatively small N loss (Table 3; Rodhe and Karlsson, 2002; van Horne et al., 1998). Undisturbed stacks of cattle and swine manure typically have a lower DM content and a greater range in N loss compared to poultry manure (Table 3). In general, less N loss occurs as the DM of cattle and swine manure decreases, but manure pH and environmental conditions also have an effect (Petersen et al., 1998).
Slurry storage in a tank or earthen pond is commonly used on dairy farms and in some other animal production systems. A combination of feces, urine, and wash water provides slurry with a DM content normally in the range of 7 to 12%. These wastes can be pushed onto the surface of stored manure or pumped into the bottom of the storage unit. Nitrogen loss from slurry storage is primarily a function of the amount of mixing that occurs throughout the storage period. With top surface loading, more mixing occurs and relatively fresh manure remains on the surface. Ammonia volatilizes more readily from this fresh manure, which leads to greater N loss. With bottom loading, a crust of dried manure and bedding material can form on the surface. This crust greatly restricts volatile loss (Table 3; Muck et al., 1984). Due to the anaerobic conditions of slurry storage, essentially all of the N loss is in the form of volatilized ammonia.
With a top-loaded slurry tank or pond, the rate of N loss varies widely as influenced by manure pH, ambient temperature, wind speed, and loading rate (Muck and Steenhuis, 1982; Olesen and Sommer, 1993). Little loss occurs at temperatures below freezing or with a manure pH less than 6, but the rate of loss increases rapidly with increasing temperature and pH (Muck and Steenhuis, 1982). When averaged over the weather conditions typical of the northern United States, a loss of about 30% of the initial N stored can be expected (Table 3).
A number of covering techniques have been tested to reduce N loss and odor from slurry storages. The most effective cover is a permanent lid. When a good seal is maintained, this cover nearly eliminates storage loss (Sommer et al., 1993). Other options include the use of a floating foil or plastic cover, or floating layers of peat, Leca, oil, straw, or polystyrene spheres (Sommer et al., 1993; Hartung and Phillips, 1994). Any of these covers can reduce N loss from slurry storage by 80 to 90%, if a complete cover can be maintained throughout the storage period. When surface cracks develop or the covering material sinks into the slurry, the effectiveness is greatly reduced. Overall, the development of a natural crust appears to be about as effective as the use of other covering materials.
Liquid storage in large lagoons is a common practice on large livestock, swine, and poultry production operations. Large amounts of water are added to the manure to obtain a DM content of about 5% or less. Some type of liquid and solid separation is often used to reduce the solids content of the manure entering the lagoon. A flush system is normally used to remove manure from the housing facility where recycled lagoon effluent provides most of the flush water. With this approach, N loss is normally very high (Table 3). High loss occurs because a surface crust does not form, and wave action provides constant mixing throughout the lagoon.
When liquid and solid separation is used, about 25% of the total manure N is removed in the manure solids with the remainder in the liquid portion (Sutton et al., 1994). If these solids are applied to cropland, the crop can use a large portion of this organic N. Often these solids are used as bedding material, which recycles this N back into the manure. To reduce the potential spread of disease organisms, these solids may be processed through composting. As discussed earlier, about 40% of this N may be lost during composting.
Most of the N entering the lagoon is lost to the atmosphere or recycled in the manure flush system. Normally, a series of lagoons are used where the effluent from the first becomes the influent of the next. Nitrogen concentration and N volatilization decrease as the fluid flows through each successive lagoon. Nitrogen emission rates are correlated with N concentration, pH, and surface temperature of the lagoon liquid and the wind speed across the lagoon (Harper et al., 2000; Aneja et al., 2001; de Visscher et al., 2002). At least 70% of the N entering a series of lagoons is lost to the atmosphere (Sutton et al., 1994; van Horne et al., 1998). Harper et al. (2000) found that less than 1% of the initial N entering the first lagoon was recovered from the final lagoon and applied to cropland. Most of the effluent from the final lagoon was used as flush water, thus carrying remaining N back through the barn and into the first lagoon. With this type of recycling, nearly 100% of the N is emitted into the atmosphere from some point in the cycle. When a single or two-stage lagoon system is used, where a greater portion of the liquid is applied through irrigation, the N loss from the lagoon is expected to be less, probably around 50%.
Of the large amounts of N emitted, about half of this N is in the form of ammonia (Harper et al., 2000). Most of the remaining half seems to be in the form of N2, with a small emission of N2O. Since N2 emission is not detrimental to the environment, the large N loss is not as great a problem as may be first assumed. Ammonia losses, however, are still high from large lagoon systems, which can adversely affect the environment of the region.
A number of additives have been promoted or tested to reduce the ammonia emission from stored manure (McCrory and Hobbs, 2001). Additive types include digestive additives, acidifying additives, and adsorbents.
Digestive additives consist of selected microbial strains and/or enzymes that are intended to enhance the biodegradation of manure. Such microbial treatments provide a promising solution due to their regenerative nature, but those evaluated thus far have not provided consistent or substantial benefit (McCrory and Hobbs, 2001). Current products appear to have been developed without a thorough understanding of the microbiological processes occurring in livestock manure. Although current products are not recommended for routine use, potential for the development of effective products exists. More research is needed to investigate known strains of bacteria or enzymes with known modes of action. Effective organisms must be able to grow, reproduce, and become a dominant or at least a major part of the indigenous community.
Acidifying additives that reduce the pH of the manure can greatly reduce ammonia volatilization. Potential treatments include acids, base precipitating salts, and labile carbon (McCrory and Hobbs, 2001). A number of acids can be used to decrease manure pH, but problems that deter their use include high cost, corrosiveness, and hazards to animal and human health. Base precipitating salts, such as chloride and nitrate salts of magnesium and calcium, can also decrease pH, but they are less effective than acids in reducing and sustaining low pH levels. They offer a treatment for short-term reductions in ammonia volatilization. Labile carbon treatments, such as sucrose and potato starch, have reduced pH by stimulating indigenous anaerobic microorganisms to produce organic acids. Tests with these treatments have shown a reduction in ammonia volatilization of 42 to 98% (McCrory and Hobbs, 2001). Large quantities of these additives are required though, which makes current treatment options uneconomical.
A variety of additives can be used to adsorb the ammonia and/or ammonium N in manure. The most common of those tested or used are clinoptilolite and peat (McCrory and Hobbs, 2001). Clinoptilolite can be applied as a feed additive, but it is more effective when applied directly to the manure. Application to broiler litter has reduced aerial ammonia concentrations by 35%, but total reductions in N loss have not been documented. Peat can adsorb 2.5% of its dry weight in ammonia N. Covering heaps of solid manure or floating a layer of peat over stored slurry can reduce N loss by up to 80%. A major challenge with this technique is maintaining a complete cover throughout the storage period.
Manure provides a valuable resource when applied to crop or grassland. Organic matter and nutrients are returned to the soil to help rebuild soil texture and biological activity and to provide nutrients required for growth of the next crop. Nitrogen losses can be very high following application, reducing the fertilizer value of the manure and adding to the degradation of the environment. Management used during and following application can have a large impact on the N losses that occur.
Nitrogen losses during this phase of manure handling include ammonia volatilization, nitrate leaching, N runoff, and N emissions that result from nitrification and denitrification processes. The greatest loss normally occurs through ammonia volatilization. When manure is surface-applied through broadcast spreading, most of the ammonium N can be lost within a few days of application. Of the total N in manure, roughly half is in an ammonium form or another form that can easily transform into ammonia. Loss occurs rapidly following slurry application, with 30 to 70% of the total loss occurring within the first 6 to 12 h (Sommer and Hutchings, 1995; Meisinger and Jokelo, 2000). The rate of loss then slows due to a lower ammonium N concentration, a drop in manure pH, infiltration into the soil, and formation of a surface crust. For application of drier manure, such as poultry litter, the initial loss rate is lower and more constant through time (Meisinger and Jokelo, 2000). Thus, initial loss from solid manure may be less than half that of slurry. However, when manure is not incorporated into the soil, total loss over time can be similar between manure types.
The rate and amount of ammonia loss is related to the weather conditions and characteristics of the manure and the soil to which it is applied. The rate of loss is very low at 0°C, but it increases exponentially with increasing temperature (Sommer et al., 1991). Loss increases with wind speed up to a speed of about 2.5 m/s. Air humidity may have an effect, but this effect appears to be small. Precipitation soon after application can greatly suppress ammonia emission by moving the ammonium N into the soil (Meisinger and Jokelo, 2000). Soil conditions including the moisture content, texture, cation exchange capacity, pH, and plant or residue cover all affect the rate and amount of ammonia loss (Meisinger and Jokelo, 2000).
Manure characteristics that affect ammonia loss include ammonium N content, DM content, and pH. Loss generally increases in proportion to the amount of ammonium in the manure. A lower DM content allows the manure to absorb into the soil more quickly, thus reducing loss (Meisinger and Jokelo, 2000). Ammonia emission is reported to increase by 5 percentage units of ammonium N for each 1% increase in slurry DM content between 1 and 9% DM (Chambers et al., 1999). At DM contents greater than 12%, there is little effect. Manure pH can have an effect, but this effect is normally small because the pH of slurry increases rapidly after application (Meisinger and Jokelo, 2000).
Nitrogen can be lost through runoff when manure is surface-applied before or following frozen soil conditions. This loss should be relatively small, around 3% of the total N applied, but runoff losses up to 10% of applied N are reported (Gangbazo et al., 1995). When manure was incorporated into the soil, runoff loss of N was found to be no greater than that from unfertilized land (Gangbazo et al., 1995).
Organic and ammonium N incorporated into the soil can be taken up by a crop, lost through nitrate leaching, or lost through denitrification. There is considerable interaction among these processes, which makes it difficult to predict the fate of the N. Normally, leaching loss is relatively small when manure is applied to a growing crop or just before the establishment of a crop. Loss of less than 2% of total N can be expected (Stout et al., 2000). When manure is applied to fallow land in the autumn, excessive leaching loss of 10 to 30% of applied N can occur (Carey et al., 1997; Beckwith et al., 1998; Weslien et al., 1998; Di and Cameron, 2002a). Compared to fallow soil, seeding a cover crop, such as winter rye, can reduce N leaching loss by 40 to 80%, depending on the precipitation patterns over the winter and soil characteristics (Beckwith et al., 1998). A cover crop reduces N loss by reducing the moisture drained through the soil profile and by taking up nitrate N moving through the profile.
Nitrification and denitrification processes in the soil cause N emissions in various forms, with the primary forms being N2O and N2. The amount of denitrification that occurs is a function of the amount of total N in the soil, available carbon in the soil, and the anaerobic conditions within the soil (Carey et al., 1997). Although total denitrification loss of N can be large (Carey et al., 1997), limited data indicates that N2O loss is normally less than 2% of the total N applied (Weslien et al., 1998; Sherlock et al., 2002).
Numerous methods are used to apply and incorporate manure, and N loss varies widely depending on the method used. The major methods include irrigation, broadcast spreading, band spreading, and some form of injection into the soil (Table 4). The largest N loss usually occurs through irrigation. During the irrigation process, a portion of the N is volatilized or otherwise made airborne before contacting the plant or soil surface. Very high values of 15 to 43% of the total N applied have been reported for this loss (Safley et al., 1992), but more typical values appear to be in the range of 5 to 10% (Sommer and Hutchings, 1995; Meisinger and Jokela, 2000). Ammonia volatilization continues from the field surface, causing an additional loss of up to 35% of total applied N (Meisinger and Jokela, 2000). Loss is influenced by ambient temperature, precipitation, and other weather conditions. Due to wet soil conditions following irrigation, rapid incorporation of the manure is difficult, which contributes to greater loss.
A number of techniques have been used to apply manure slurry in bands (Meisinger and Jokela, 2000). This technique is primarily used on grassland or other established crops. When bands are well formed and maintained on the surface, N loss can be reduced 30 to 70% compared to broadcast spreading on the same crop. Other studies have shown band application to be less effective, particularly when the band spreads, covering more of the field surface.
Direct injection of slurry into the soil is the most effective method for reducing N loss during and following application. With deep injection (>10 cm in depth), so that the soil totally covers the manure, ammonia N loss is less than 5% of the total N applied (Table 4). A problem with deep injection in grassland or other established crops is that the injector causes root damage that can reduce crop growth (Mattila and Joki-Tokola, 2003). More shallow injection techniques (5 cm in depth) allow less root damage. These are not as effective in stopping ammonia emissions, but they are more effective than broadcast or band spreading (Table 4). Since injection conserves more of the manure N, greater leaching and denitrification losses can be expected from the soil, but this depends on the timing of the application and crop, soil, and weather conditions following application (Dosch and Gutser, 1996; Weslien et al., 1998).
Management to Reduce Application Losses
A number of management strategies or techniques can be used to conserve N during and following field application. From an N conservation perspective, irrigation of manure should be avoided (Table 4). Broadcast spreading of slurry on grassland or on heavy crop residue without incorporation should also be avoided. Injection techniques provide the best option in cropping systems where they can be practically used. Rapid incorporation of manure within a few hours of application can provide substantial benefit. Band spreading with a trailing hose or similar device can provide some benefit particularly when slurry is applied to an established crop or grassland where incorporation is not possible. Tilling the soil before irrigation, broadcast, or band application may also improve the infiltration of liquid or slurry manure and thus reduce loss slightly.
Application rate and slurry DM content can be manipulated to reduce N loss. A higher application rate may reduce the fraction of the applied N that is lost, but this depends on competing forces of adsorption into the soil and volatilization (Meisinger and Jokela, 2000). Less loss would occur with one heavy application as compared to two or more lighter applications that provide the same total applied N. Because slurry with a low-DM content is absorbed more rapidly, dilution can reduce loss (Sommer and Hutchings, 1995). This strategy is not usually practical because a 50% reduction in DM content will double the amount of manure handled. Aeration of slurry has been used to reduce DM content by up to 16% (Leinonen et al., 1998). About 10% of the total N can be lost during the aeration process, which offsets any reduction in application loss obtained through the decrease in DM content.
Dropping the manure pH below 7 before it is applied can reduce ammonia emission (Sommer and Hutchings, 1995; Meisinger and Jokela, 2000). Nitric or sulfuric acid treatments used to drop the pH to 6.5 have reduced ammonia loss by up to 75%. Use of amendments, such as alum or ferrous sulfate, can also have an acidifying effect on manures, which may reduce N loss.
The best strategy to reduce leaching and denitrification losses is to apply the right amount of manure close to the time the nutrients are needed by the crop (Beckwith et al., 1998). Autumn application, particularly on fallow soil, can lead to large losses. About 50 to 70% of the nitrate accumulated in the soil profile by late autumn is leached during the winter (Di and Cameron, 2002a). Likewise, excess N beyond crop needs will move through the soil profile and be lost. Large amounts of N in the soil also increase denitrification loss. Nitrification inhibitors have been used to reduce the N lost by denitrification with mixed results. Use of an inhibitor almost eliminated an increase in denitrification following the injection of cattle slurry into grassland soil (de Klein et al., 1996). In a study of N2O loss in different crop rotations, use of an inhibitor had little effect (Stoeven et al., 2002). Use of an inhibitor also had small and inconsistent effects on the amount of N leached through the soil profile over winter (Beckwith et al., 1998).
Substantial N losses from volatilization, leaching, and denitrification occur from manure deposited by grazing animals, and the loss processes are somewhat different from those of slurry or solid manure application. Fecal and urinary deposits normally occur at different times and locations in the field. Most fecal N is organic and is thus relatively stable following deposition. About 5% of the fecal N is lost by volatilization (Ryden, 1986). Most of the excreted N (55 to 75%) is in the urine, in which higher levels are associated with the overfeeding of protein (Jarvis et al., 1989). Urinary N (urea) is rapidly hydrolyzed to form ammonia, which is then nitrified at a slower rate. Ammonia volatilization loss can be very high, but rapid absorption of the urine into the soil surface can reduce this loss. Reported losses vary from 5 to 66% of the total urinary N with greater loss under hot and dry weather and soil conditions and relatively low loss under cool and moist conditions (Ball and Ryden, 1984; Jarvis et al., 1989). Average total loss is approximately 10% of the excreted N (Table 4; Oenema et al., 2001b). Ammonia emission is greatest during and immediately after a grazing event. Rain, poor drying conditions, and low wind all help reduce this emission rate (Ryden, 1986). Limited data on outdoor swine production indicate that N loss may be twice that found with grazing cattle (Eriksen et al., 2002).
Leaching loss of N can be much higher under grazing conditions than occurs for spread manures. Nitrogen concentrations under a urine patch are very high, equivalent to an application rate of 300 to 1,000 kg N/ha. Much of this N is in excess of crop needs and is leached down through the soil profile. Reported loss ranges from about 10 to 60% of the urinary N deposited (Garwood and Ryden, 1986; Stout et al., 1997; Silva et al., 1999; Di and Cameron, 2002b). Factors affecting this loss are soil type or texture, rainfall following deposition, and the time of the year the deposit is made. Urine N deposited in the spring is more likely to be taken up by a growing crop and thus provides about half the loss of that deposited in the fall (Stout et al., 1997). Leaching loss from fecal N is small, about 2% of that deposited (Stout et al., 1997). Combined leaching losses are 10 to 30% of the total N excreted on the pasture (Table 4).
Runoff loss of N also occurs from pastures, but this loss is small. Preliminary data from Ahmed et al. (2002) indicated an average loss of about 1% of the excreted N on continuously grazed pasture. Rotational grazing provided up to an 80% reduction, and the use of a vegetative filter strip reduced this loss by 50% or more. On poorly drained soils, though, runoff loss may be much greater, with less leaching (Garwood and Ryden, 1986).
Denitrification losses from pastures can also be substantial, particularly under urine deposits. Reported losses range from about 5 to 30% of the applied urinary N (Garwood and Ryden, 1986; Fraser et al., 1994; Di et al., 2002). Most of this loss appears to be in the environmentally benign form of N2, but some portion will be in the form of N2O. Available data indicate that less than 8% of the applied N will be lost as N2O with a typical loss around 2% (Ryden, 1986; Clough et al., 1996; Oenema et al., 2001b).
Management can be used to reduce N loss from grazing animals, but the benefit of these changes may be small and implementation may be impractical. One practical step that should always be considered is to feed supplemental protein feeds efficiently, and thus reduce urinary N excretion. Overstocking of animals along with a large import of forage and other supplemental feeds should also be avoided. Movement of watering and supplemental feeding areas will improve nutrient distribution, thus increasing plant uptake and reducing loss. Volatile loss may be reduced by irrigating the paddock immediately after grazing to wash the N into the sod and soil. Spreading of crushed manure natural zeolites has been suggested to increase the cation exchange capacity at the base of the sward (Ryden, 1986). Leaching loss can best be reduced by avoiding grazing in the late autumn or winter when plant uptake of N is low. Removing the autumn growth through silage harvest can help reduce the accumulation of excess soil nitrate, which at that time of the year will likely be lost by leaching (Stout et al., 1997). Less use of N fertilizer with greater use of clover and other legumes to supply needed crop N can also reduce soil N levels and leaching loss (Garwood and Ryden, 1986). Di and Cameron (2002c) decreased leaching loss 60% and decreased denitrification loss by 82% by applying a nitrification inhibitor, but practical application of this technology would be difficult.
Management to reduce N losses in animal production requires a whole-farm approach. As discussed above, many changes can be made to reduce N losses in each step of manure management between animal excretion and crop uptake. However, the benefit for reducing the loss in any one component is low if steps are not taken to reduce losses occurring in other components. For example, reducing ammonia emission in the housing facility has little benefit if that retained N is simply lost due to poor management during manure storage and field application. Reducing ammonia emissions also may not provide overall benefit if that additional manure N leads to greater losses through denitrification and leaching. These losses of nitrous oxide emitted to the atmosphere and nitrate in groundwater may have a greater long-term cost to society than the ammonia emission. Only by providing similar levels of management to animal feeding, housing, manure storage, and field application can production systems be developed with reduced or optimum environmental impact.
Considerable effort is being given to farm level assessment and management of N. A comprehensive review is beyond the scope of this paper, but a brief review is included to demonstrate the type of work being conducted. Research efforts can generally be divided between actual farm studies and modeling studies. Although these two categories generally represent the type of work done, in practice these two approaches often join to provide the most comprehensive assessment and application of research on farming systems.
An awareness of the environmental problem in animal production began with the measurement of the potential nutrient imbalance in production systems (Bacon et al., 1990). Economic incentives during the last half of the last century drove farms toward greater animal densities per unit of land with less integration between crop and animal production. Numerous studies have illustrated the potential for excess N on the farm where feed and fertilizer inputs exceed the N contained in products sold off the farm. Over the long term, this excess N must be lost in some form to the environment. Case studies have taken an in-depth look at specific farms to predict where this N is lost (e.g., Hutson et al., 1998).
Experimental farms have been developed to closely monitor N flows through animal production systems. To evaluate grassland management systems for beef production, three farmlets were compared at two sites in South West England (Laws et al., 2000). Tactics including timely manure application, slurry injection, early housing of cattle, and the use of a mixed grass and white clover sward were successful in reducing surplus N, but animal and herbage production were reduced. The Karkendamm farm in Germany was established by the University of Kiel to measure N fluxes in the soil-plant-animal system (Taube and Wachendorf, 2000). Various management strategies in crop production and animal feeding were evaluated at the farm level with a goal of measuring and reducing all N flows and losses in dairy production.
A prototype farm in The Netherlands (De Marke) has been in operation for over 10 yr to research and demonstrate efficient nutrient management on Dutch dairy farms (van Keulen et al., 2000). The farm was designed to meet stringent environmental norms for N, P, and foreign substances, such as pesticides. Goals included reducing the annual surplus N on the farm to 128 kg/ha, with annual ammonia emission limited to 30 kg N/ha, nitrous oxide emission limited to 3 kg N/ha, and nitrate in upper groundwater below 50 mg/L. The farm was intentionally placed on soil that was among the driest and most leaching prone in The Netherlands.
Several techniques or strategies were used at De Marke to increase nutrient use efficiency and reduce losses to the environment. This began with efficient feeding of the animals. A unique floor system was used in the housing facility to separate urine from feces, which reduced the transformation of N to volatile ammonia and its subsequent loss. Urine and feces were combined and stored up to 6 mo in a covered storage tank. Manure was applied through shallow injection on grassland and deep injection on arable land. Minimum amounts of N fertilizer (120 kg N/ha) were used on grassland to maintain adequate yields and protein contents with no N or P fertilizer applied to other crops. A "catch crop" of annual ryegrass was seeded during the cultivation of corn land about 8 wk after the corn was established. In autumn, following corn harvest, this crop took up residual mineralized N to reduce potential leaching to groundwater. Implementing these N-conserving technologies reduced fertilizer N input by 74%. The farm has successfully demonstrated that the stringent environmental goals of The Netherlands can be met on a dairy farm (van Keulen et al., 2000).
A number of research efforts have focused on developing models to predict N loss during manure handling and following field application. Both empirical and mechanistic approaches have been used to develop models of ammonia emission during animal housing, manure storage, and field application (Hutchings et al., 1996; Ni, 1999; Sogaard et al., 2002). After the manure is incorporated into the soil, mechanistic models have been used to simulate N transformations through nitrification and denitrification and to predict nitrate leaching into groundwater and N emissions into the atmosphere (Shaffer et al., 1991; Eckersten et al., 1998; McGechan and Wu, 2001). These models have improved our understanding of the dynamic processes involved, and they provide ways of predicting or quantifying losses in each component of animal production.
Component models of N loss have been linked or combined to predict or assess N losses in whole-farm production systems. Bussink and Oenema (1998) compared dairy production systems in The Netherlands and found that up to threefold reductions in ammonia loss were possible along with marked reductions in mineral fertilizer use. In a comparison of dairy farming systems in the United Kingdom, Jarvis et al. (1996) found that using a tactical approach to fertilizer application, injecting slurry, or using 50% corn silage provided substantial reductions in ammonia, nitrous oxide, and nitrate losses. Dou et al. (1996) developed a computer worksheet to compare efficiencies of N utilization and nutrient balances among dairy production systems.
More sophisticated models have combined N loss with other production components to predict both environmental and economic implications of management changes. Schmit and Knoblauch (1995) used a linear programming approach to determine the economically optimal dairy herd intensities, manure application rates, and crop mix for unrestricted and restricted scenarios of N loss on New York dairy farms. They found that optimal cow numbers per hectare decreased by nearly 35% with a restriction on N loss. Stonehouse et al. (2002) used a mixed integer programming approach to generate optimal whole-farm plans for specialized swine finishing enterprises in Ontario. They found tradeoffs between economic and environmental goals, and environmental goals could only be reached at some loss in farm net return.
Computer simulation provides a powerful tool for evaluating the long-term impacts and interactions of management changes in animal production. McGechan and Wu (1998) developed a weather-driven simulation to compare slurry management options in dairy production in the United Kingdom. Environmental benefits were shown for long-term manure storage and injection, but neither practice could be economically justified. Kuipers et al. (1999) showed a similar result for The Netherlands, but manure storage and injection have become obligatory in that country.
A comprehensive farm simulation model was developed by Rotz et al. (1999b) to evaluate and compare the economic and environmental impacts of alternative dairy production systems. They demonstrated that including a low-RDP feed in ration formulation for a typical dairy herd reduced volatile N loss from the farm by 13 to 34 kg/ha of cropland with a small (1 kg/ha) reduction in leaching loss, and a $46 to $69/cow improvement in annual farm net return. Environmental and economic benefits for efficient protein feeding were generally greater with more animals per unit of land, higher milk production, more sandy soil, or daily manure hauling. The model was also used to evaluate and compare the long-term environmental and economic effects of various alfalfa, corn, soybean, and small grain cropping strategies on dairy farms (Borton et al., 1997; Rotz et al., 2001; Rotz et al., 2002). This model has been expanded to form the Integrated Farm System Model, which includes crop, beef, and dairy production options (Rotz, 2003).
Modeling has also been carried beyond the farm level. De Vries et al. (2001) modeled all of agriculture in The Netherlands to study the impacts of both structural measures and improved farming practices on major N fluxes, including ammonia and nitrous oxide emissions, leaching, and runoff. Improved farming practices were able to provide significant reductions in N loss, but these improvements were not enough to reach all targets set for those fluxes. The current ammonia emission target suggested for the year 2030 could not be met without eliminating all poultry and swine farming from The Netherlands and restricting all cattle to low-emission stables.
Whole-farm management is necessary to decrease nitrogen lossesin animal production. If steps are taken to reduce loss in onecomponent of the farm, the nitrogen saved will likely be lostelsewhere if all components are not equally well managed. Managementmust focus on improving the nitrogen use efficiency of the animalsto reduce nitrogen excretion, retaining that nitrogen in themanure until it is incorporated into the soil, and applyingthe appropriate amount of manure in a timely manner to enhancecrop uptake. Management of all the factors involved in the nitrogencycle of the farm is complex. Whole-farm research on nitrogenmanagement is needed, but such research is difficult and expensive.Computer simulation and other farm models provide research toolsthat integrate farm components to assess and compare long-termeffects of production systems. These models can provide helpwith strategic farm planning and tactical operations to reduceadverse environmental effects, while maintaining or improvingfarm profit.
2 Correspondence: Building 3702, Curtin Road (phone: 814-865-2049; fax: 814-863-0935; e-mail: firstname.lastname@example.org ).
Received for publication July 8, 2003. Accepted for publication August 26, 2003.
Table 1. Typical values for the annual excretion of N by various animal types expressed as percentage of body weighta
Table 3. Typical N losses for the major types of long-term manure storage used in animal production expressed as a percentage of total N entering storagea