Modeling ammonia emissions from manure in conventional, organic, and grazing dairy systems and practices to mitigate emissions

Nearly 60% of all ammonia (NH 3 ) emissions are from livestock manure. Understanding the sources and magnitude of NH 3 emissions from manure systems is critical to implement mitigation strategies. This study models 13 archetypical conventional (5 farms), organic (5 farms), and grazing (3 farms) dairy farms to estimate NH 3 emissions from manure at the barn, storage, and after land application. Mitigation practices related to management of the herd, crop production, and manure are subsequently modeled to quantify the change in NH 3 emissions from manure by comparing archetypical with these alternative practices. A mass balance of nutrients is also conducted. Emissions per tonne of excreted manure for the manure system (barn, storage, and land application) range from 3.0 to 4.4 g NH 3 for conventional farms, 3.5–4.4 g NH 3 for organic farms, and 3.4–3.9 g NH 3 for grazing farms. For all farm types, storage and land application are the main sources of NH 3 emissions from manure. In general, solid manures have higher emission intensities due to higher pH during storage (pH = 7.4 for liquid, 7.8 for slurry, and 8.5 for solid manure) and lower infiltration rates after land application when compared with slurry and liquid manures. The most effective management practices to reduce NH 3 emissions from manure systems are combining solid-liquid separation with manure injection (up to 49% reduction in NH 3 emissions), followed by injection alone, and reducing of crude protein in the dairy ration, especially in organic and grazing farms that have grazing and forages as the main component of the dairy ration. This study also shows that the intensity of NH 3 emissions from manure depends significantly on the functional unit and presents results per manure excreted, total solids in excreted manure, animal units, and fat and protein corrected milk.


INTRODUCTION
Eighty percent of all ammonia (NH 3 ) emissions in the US, including emissions from natural sources and human activities, are from agricultural sources.Of these national NH 3 emissions, approximately 60% are from livestock manure and 20% from synthetic fertilizers (US EPA 2021).Concentrations of NH 3 in the atmosphere have been increasing in recent decades (Warner et al., 2017;Yu, Nair, and Luo 2018;Yao and Zhang 2019) likely due to increasing losses from agricultural systems receiving greater fertilization from manure and commercial fertilizers to increase productivity and increasing temperatures (Shen et al., 2020;He et al., 2021).Ammonia emissions can negatively impact the environment and human health.Emitted NH 3 has the potential to form particulate matter, which can lead to respiratory and pulmonary problems (Brook et al., 2010) and nitrous oxide (N 2 O), which is a potent greenhouse gas (GHG) (Krupa 2003).In addition, NH 3 can travel to farther locations and deposit in soil and water systems leading to increased soil acidity, plant toxicity and reduced diversity, and water eutrophication (Behera et al., 2013;Krupa 2003;Guthrie et al., 2018).Strategies for NH 3 emissions mitigation are needed to reduce these impacts.
Understanding the sources and magnitude of NH 3 emissions is critical to implement mitigation strategies.Quantifying NH 3 emissions and establishing emission factors from manure in livestock systems has been difficult particularly due to variability in manure management practices and environmental conditions that promote these emissions (Hristov et al., 2011).Studies have determined that NH 3 emissions from manure vary significantly with temperature, pH, and manure handling systems (Harper et al., 2009;Li et al., 2009;Sparks, 2008).For example, 2 different in situ assessments of Wisconsin dairy farms show NH 3 emissions from manure of 6.6 to 37.0 g NH 3 /day for lactating Modeling ammonia emissions from manure in conventional, organic, and grazing dairy systems and practices to mitigate emissions cows (T.K. Flesch et al., 2009) versus a comparable study showing a range of 6.7 to 18.8 g NH 3 /day for lactating cows (Powell, Broderick, and Misselbrook 2008) versus 13.4 to 25.4 g NH 3 /day for heifers (Powell, Misselbrook, and Casler 2008).Further, a study combining direct measurement and atmospheric modeling for a select number of farms that were then extrapolated to nearby farms (Harper et al., 2009) found average emissions from manure of 55.0 g NH 3 /day per cow per day, significantly higher than previous studies.These results stress the need to develop regional inventories of NH 3 emissions from manure specific to farm level and management practices.
Studies estimating NH 3 emissions from livestock manure can be classified as experimental and modeling studies.Experimental studies are the gold standard to effectively quantify NH 3 emissions.However, they are expensive, are site-specific, and therefore cannot be generalized to represent emissions across climate regions, manure characteristics, and through different management practices (Hristov et al., 2011).As a result, predictive models have been developed to estimate manure NH 3 emissions from different farm types and locations (Hafner et al., 2019;Sven G. Sommer, Webb, and Hutchings 2019;Hempel et al., 2022).Empirical models attempt to create a usable tool to estimate manure NH 3 emissions from farms utilizing emissions factors and more simple relationships between management and emissions than process-based models (Zhang et al., 2005;Wu et al., 2020;Vaddella, Ndegwa, and Jiang 2011).Process-based models integrate data from experimental studies and relate them to key chemical, biological, and physical processes that contribute to NH 3 emissions from manure (Rotz et al., 2018;Li et al., 2012;Pinder et al., 2004).In general, these models are more precise than empirical models, but they also require more inputs at the farm level.
Life cycle assessment (LCA) is a useful tool to account for the environmental impacts of different systems, including dairy (Aguirre-Villegas et al., 2022;Kim et al., 2019;Rotz et al., 2021).LCA studies generally do not report NH 3 emissions from manure separately, even though, estimation of NH 3 is included to quantify other environmental impacts such as GHG emissions and eutrophication potential.This has been changing with concern over increasing NH 3 emissions from the agricultural sector, where recent process based LCA studies modeled different environmental impacts from dairy farms across the US and reported NH 3 emissions separately (Aguirre-Villegas et al., 2022;Rotz et al., 2021).These studies evaluated differences in environmental conditions and farm management practices, but quantified emissions on a regional and national scale and did not specifically establish NH 3 emission factors for specific management practices.Moreover, these NH 3 factors are not specific to manure as they include emissions from synthetic fertilizers and feed storage.
In addition to complexities in establishing NH 3 emission factors from manure in dairy farms, methodologies for presentation of results complicates study comparisons.In broader LCA studies for milk production systems, results are commonly presented per kilogram of fat and protein corrected milk (FPCM).However, experimental studies commonly relate emissions to manure or other representative units.For example, composting studies present results in g NH 3 -N/tonne of manure (El Kader et al., 2007), g NH 3 /kg of dry manure (Bai et al., 2020), and kg of nitrogen (N) per tonne of compost (Sommer and Dahl 1999).Comparison of results requires conversion to a common unit (referred as functional unit), which adds potential for additional error and difficulty in comparing studies (Ba et al., 2020).
This study aims to i) define and model archetypical conventional, organic, and grazing dairy farms, ii) estimate NH 3 emissions from manure at the barn, storage, and after land application from these archetypical farms, iii) model different practices related to management of the herd and manure and estimate the change in NH 3 emissions from manure, while conducting a mass balance of nutrients in manure, and iv) evaluate the impact of functional unit selection on overall trends when comparing manure related NH 3 emissions from different management practices.

Description of Baseline Farms
Thirteen dairy farms with varying management representative of a range of practices in Wisconsin were modeled to assess NH 3 emissions from manure in differing dairy systems.1).Farms were modeled individually adopting the model described in Aguirre-Villegas et al. (2015) but modified to capture the practices of the farms outlined above.The farms were modeled by first determining the milk production, which then determines the number of animals (lactating, heifers, and dry cows) needed to achieve that target milk production.Daily dietary requirements and composition for each crop and animal type are considered based on NRC (2001) guidelines to effectively determine N excretion.The animal feeding ration for each farm is differentiated between young Aguirre-Villegas et al.: Ammonia Emissions from Manure on Dairy Farms stock, nonlactating cows, and lactating cows to meet maintenance, growth, pregnancy, and milk production requirements.It is assumed that for all farms, animals reach their mature weight at 21 mo.Similarly, manure excretion and total solids (TS) in manure were calculated for each cow type (Nennich et al., 2005).Nutrient flows from feed to milk and manure are estimated based on dry matter intake (DMI) and nutrient in feed crops (NRC 2001).
Farms vary in herd management (e.g., size, animal type, replacement rate, bedding used, etc.), milk production and composition, feeding strategy and diet composition, and manure management (Tables A1-A3).Farm sizes range from 50 to 1,000 lactating cows, plus the respective maintenance animals, with organic farms in the lower end of this range.There is a wide range in terms of milk production (17 to 40 kg/day) with conventional farms achieving the highest yields per cow and organic farms the lowest.The calving interval ranges between 12.5 and 14 mo and the annual replacement rate between 30 to 42% with organic farms in the lower range (Holly et al., 2019).All farms produce all feeding crops on-farm except for supplements, meaning that overall, they have land for manure application.All modeled farm scenarios have Holstein cows except for O1 that has Jersey cows that produce less milk but with higher fat and protein contents.
All farms are connected to the Wisconsin electricity grid, except for O5 that is modeled as an off-grid dairy farm with a diesel generator.One of the main differences among farm groups is the diet composition (Table A4).Organic and grazing farms rely more heavily on pasture and forage as feed components, whereas conventional farms rely more on grains and supplements.One farm in the organic group (O4) is modeled to maximize pasture in the diet (O4).The grazing season for farms with pasture as a feeding component is limited to 6 mo (3 mo for G3), with cows housed in barns the rest of the year to represent areas with a cold winter climate.Three farm sizes were evaluated for all farm types: small (1-100 lactating cows), medium (>100-200 lactating cows), and large (200-1,000 lactating cows) following the delineations from a survey conducted by (Aguirre-Villegas and Larson 2017).The model considers that small and medium farms manage solid or semi-solid manure, whereas large farms manage slurry or liquid manure.All farms land apply manure by surface broadcast.
Defining and modeling these archetypical farms captures a range of real-life dairy operations and relates their practices and characteristics to NH 3 emissions from manure (Figure 1).The initial 13 farm scenarios are referred to as base-case where integration of alternative management practices allows for evaluation of NH 3 emission mitigation strategies.It is important to note that only NH 3 emissions from manure are estimated.No embedded emissions from the use of other materials (e.g., synthetic fertilizers) or energy inputs at the farm are modeled.However, the extra N contained in bedding material is accounted for as bedding is managed along with the manure stream in dairy farms.As  a result, manure is defined as the mix of urine, feces, and bedding.

Alternative Management Practices
Alternative management practices were also modeled to quantity the change in NH 3 emissions to evaluate mitigation potential.Alternative practices were classified as those changing i) the dairy ration composition (different rates of corn/alfalfa silage, and reduction in crude protein (cp)), ii) herd management (improving feed efficiency, reducing replacement rate, and increasing milk production), iii) and manure management (storage cover, injection, and manure processing through solid-liquid separation (SLS) and anaerobic digestion (AD)).Changes in manure management affect NH 3 emissions directly as they impact the conditions (physically, chemically, or biologically) that promote NH 3 formation throughout the stages of manure management.Changes in the dairy ration composition affect NH 3 emissions from manure indirectly by altering N excretion in urine and feces, and thus, the potential formation of NH 3 .Finally, practices related to herd management can affect the intensity of NH 3 emissions from manure, depending on what denominator (functional unit) is used to present results (described further in the study).Most practices are implemented for all farm types with some exceptions (Table 2).A combination of some of the listed practices has been evaluated to determine the combined potential reduction effect in NH 3 emissions.

Estimation of NH 3 Emissions from Manure and Nutrient Balance
Ammonia emissions from manure are estimated with equations that relate key practices and manure characteristics to NH 3 emissions (e.g., manure pH, total ammoniacal N (TAN) content in manure, crust formation on top of the storage, manure type, time for manure incorporation into the soil, etc.) to capture the difference between scenarios and after the implementation of alternative management practices.When equations are not available, emission factors relating N content with NH 3 emissions are used (Table 3).Emissions are estimated on a daily step and averaged over a year to capture changes in temperature.Considering a daily step is especially important for manure storage as new N in manure is added to the storage every day, with emission calculations considering the existing and the newly added N. As shown in Table 3, temperature has a significant effect on NH 3 emissions, especially at the barn and during manure storage.The temperature of Wisconsin averaged for the last 30 years was used in the model (NOAA 2023).Given its geographical scope, emission results from this study should not be extrapolated to other regions with climates that are significantly different from Wisconsin.

Increase corn silage
Increasing corn silage by 10 to 20%, reducing 1 to 3% of alfalfa silage, and reducing corn grain by 13 to 18% while keeping dry matter intake (DMI) constant.

C1-C5
Increase alfalfa silage Reducing corn silage by 9 to 13%, increasing alfalfa silage by 18 to 27%, and reducing corn grain by 0 to 9% while keeping DMI constant.

C1-C5
Reduce crude protein (cp) Reducing crude protein in the cow's diet by 20% (resulting in a reduction in nitrogen in the cows' diet).

Herd management
Increase feed efficiency Reducing DMI by 20% but maintaining milk production, therefore, increasing feed efficiency.Manure production is also reduced based on the reduction in DMI.
O1-O5, G1-G3 Increase milk production Increasing milk production by 20% for the same amount of DMI consumed by the cow.Manure nutrients are reduced to reflect increased nutrients in milk produced.
C1-C5, O1-O5, G1-G3 Reduce replacement rate Reducing the replacement rate by 20%, which assumes a longer life for each cow with improved animal health practices.
O1-O5, G1-G3 Manure management Empty storage once per year Reducing the emptying of manure storage from two times per year (half the annual manure produced each time) to one time per year (all the annual manure produced).
C1-C5, O1-O3, G1-G3 Cover manure storage Placing an impermeable cover over the manure storage reducing losses from exposure to atmosphere and wind.
C1-C5, O1-O3, G1-G3 Inject manure Injecting manure into the soil subsurface during land application reducing losses from exposure to atmosphere and wind.Besides conducting a mass balance for N from manure excretion to manure land application (considering losses after application), this study also tracks manure phosphorus (P) reaching the land.Nitrogen availability for crop production depends on the NH 3 losses during each manure handling step.Unlike N, P is not subject to emission losses.The amount of P reaching the land will differ from farm to farm based on the implemented management practices related to dairy ration composition, herd, and manure management.In scenarios with screw press SLS, separation efficiencies (nutrients following the solid fraction) for P are 29% (Aguirre-Villegas, Larson, and Sharara 2019).Reported nutrient balances include total nitrogen (TN), TAN, organic N, and P in land applied manure and manure deposited in pastures for organic and grazing farms.
Changes in Functional Unit.When estimating NH 3 intensity, total emissions from manure are divided by a specific variable of interest, defined as functional unit by LCA methodology.LCA studies addressing the environmental sustainability of dairy farm systems usually express results per kg of fat and protein corrected milk (FPCM, corrected to 4% fat and 3.3% protein) as the function of a dairy farm is to produce milk (Thoma et al., 2013;Rotz et al., 2021;D. Kim et al., 2019).Other studies suggest that land area or number of cows better capture the benefits of dairy systems that do not focus on maximizing productivity, which might be the case for organic and grazing farms (O'Brien et al., 2012;Ross et al., 2017).Moreover, studies focusing on manure management express impacts per mass of manure produced at the farm (Aguirre-Villegas, Larson, and Reinemann 2014;Poeschl, Ward, and Owende 2012).In this study, NH 3 emissions from manure are presented per 1 tonne (1,000 kg) of excreted manure.In addition, results are expressed per kg of excreted TS, per animal unit (AU, 1 AU = 1,000 pound of animal), and per kg of FPCM to facilitate comparisons to other studies, for 100 NH 3 : emissions after application (kg NH 3 ); TAN: total ammonia N in manure (kg NH 3 -N); TS: total solids in manure (liquid: 4%, slurry: 8%, solid: 20%); days: to incorporate manure (if not incorporated, days > 7) (Jokela et al., 2004)

Anaerobic digestion
Organic Average of the following studies: Amon et al., 2001;el Kader et al., 2007;Maeda et al., 2013;Sommer & Dahl, 1999;Yang et al., 2017.the results to be used in broader assessments, and to evaluate the difference in NH 3 emission trends.

NH 3 Emissions from Manure in Baseline Farms
Estimated NH 3 emissions from manure in the 13 modeled baseline farms at the barn, storage, land application, and from manure directly excreted on pasture are presented in Figure 2.For all farm types, manure storage and land application are the main sources of NH 3 emissions.Changing the size of the farm does not significantly impact results when expressed per tonne of excreted manure as can be seen in farms C1, C2, and C3.On the other hand, the type of manure being managed influences overall NH 3 emissions.Farm C4 manages solid manure and has relatively higher NH 3 emissions from manure than the other conventional farms, mostly due to higher emissions during storage.Despite that stored solid manure has reduced surface area exposed to wind and potential volatilization of NH 3 , solid manure has higher pH levels, which increases NH 3 emissions (Kim et al., 2021;US EPA, 2004).The higher emissions from solid manure storage are greater in comparison to the liquid manure managed on C5.Liquid manure storage promotes a lower pH in manure, resulting in a reduction in NH 3 emissions despite increased surface area when compared with solid manure storage (Rotz et al., 2018).Following land application, liquid manure infiltrates more easily into the soil, facilitating the bonding of ammonium to soil particles (Jokela et al., 2020).On the other hand, solid manure remains on the soil surface where it is exposed to wind.Despite this, NH 3 emissions from application of solid manure are comparable to emissions from application of liquid or slurry manure because much of the TAN has already been lost during storage of solid manure.All baseline farm scenarios (conventional, organic, and grazing) consider surface application of manure.
For all conventional farms, the barn has the lowest contribution to NH 3 emissions from manure.C3 and C5 farms both have sand bedding, which decreases emissions (0.43 g NH 3 /tonne of excreted manure) than farms C1, C2, and C4 that use straw bedding that is another source of N. Additionally, C5 manages liquid manure with a flush system in free-stall housing, and has lower downstream storage and application NH 3 emissions compared with slurry or solid manure management, due to reduced pH and easier infiltration into the soil as explained before.
In organic and grazing farms total emissions of NH 3 also include those emissions from manure when cows are grazing.The amount of NH 3 emissions from pasture is related to the amount of manure excreted directly on the pasture.Farms O4 and O5 have relatively higher emissions than other organic farms mainly due to the use of bedded packs (hence no storage emissions are reported for O4 and O5).Bedded packs increase NH 3 emissions from manure in the barn, given that they require more straw for bedding and promote warmer conditions that increase N volatilization.As with conventional farms, emissions from land application of slurry manure (C3) are lower (faster infiltration into the soil), but comparable to emissions from land application of solid manure (O1-O2) as most of the TAN was emitted during storage of solid manure.Results mirroring organic farms are seen in grazing farms.More manure is collected in G3 than in G2 due to a shorter grazing season in G3, resulting in lower NH 3 emissions from pasture but higher emissions from storage and application.G1 has higher emissions than G2 and G3 overall due to solid manure management.
After analyzing the models to predict NH 3 emissions from manure management in this study, it is determined that the most important factors contributing to NH 3 volatilization at the barn and during storage are pH (7.4 for liquid, 7.8 for slurry, and 8.5 for solid manure), temperature, and surface area of manure exposed, which is in line with previous studies evaluating NH 3 emissions (Hristov et al., 2011;Qu and Zhang 2021).However, exposure to the atmosphere is the primary driver for NH 3 emissions after land application, as higher solids concentration in manure increases emissions due to decreased infiltration of N into the soil (S G Sommer and Hutchings 2001).

NH 3 Emissions from Alternative Management Practices
The effect of alternative management practices (Table 1) on NH 3 emissions from manure in conventional farms is presented in Figure 3.For all conventional farm baselines, the most effective practice to reduce NH 3 emissions is injection of manure into the soil (28-32% reduction) as an alternative to surface application.Practices that also reduce NH 3 emissions are installing a manure storage cover and separating manure with SLS followed by injection (up to 38% reduction in one farm).Reductions are approximately uniform over the management practices for each farm, except for C4 and C5 that have lower reductions due to solid manure being managed instead of slurry manure.This can be evaluated more clearly between C1 and C4 as all modeled components are the same between farms except that C1 manages slurry manure and C4 manages solid manure.C4 has greater emissions largely due to the increased NH 3 lost from manure storage.During Aguirre-Villegas et al.: Ammonia Emissions from Manure on Dairy Farms storage, solid manure is more susceptible to volatilization due to higher pH (Chamber, Smith, and Weerden 1997;Menzi et al., 1997;S G Sommer and Hutchings 2001).The addition of a cover limits exposure to wind but does not reduce pH in solid manure, which results in lower reduction in emissions for solid manure under a cover.C5 manages liquid manure which has lower pH, making wind exposure one of the main contributors to NH 3 emissions.When a manure cover is placed on top of the liquid manure storage, wind exposure is significantly limited.Two modeled management practices increased NH 3 emissions from manure, AD, and compost.AD mineralizes organic N into inorganic N which is more susceptible to volatilization especially when there is no cover or a natural crust on top of the stored manure.However, integrating injection with this system can mitigate the increased emissions.Aerated compost creates a combination of factors that contribute to increased NH 3 emissions including high temperatures, high pH, and exposure to the atmosphere by the aeration process.As a result, farms with manure composting have higher NH 3 emissions from manure than baseline farms.While AD and compost systems increase NH 3 emissions, they have other benefits such as reducing methane emissions, inactivating pathogens, reducing odors, degrading antibiotics, etc.There exists a tradeoff in environmental impacts when integrating these practices and additional steps need to be taken to take advantage of the benefits while mitigating NH 3 emissions (e.g., injecting manure following digestion).
The effects of management practices on NH 3 emissions from manure for organic farms are similar to conventional farms (Figure 4).SLS+injection, injection, and reduction of crude protein in the diet are the 3 most effective practices to mitigate NH 3 emissions for organic farms.Injecting manure has higher NH 3 reductions for O3-O5 than O1 and O2 as O3-O5 had higher emissions initially in the baseline farms.In the baseline, O3 stores slurry manure with crust formation, preventing volatilization during storage but making N more available for volatilization after land application.Injection restricts access to the atmosphere, more steeply reducing the potential for volatilization during application.O4 and O5 have bedded pack housing, which functions as both barn and storage for manure.Interestingly, NH 3 emissions from bedded packs housing/storage in O4 and O5 are lower than the combination of NH 3 emissions from the barn and manure storage in O1 and O2.This results in more available TAN and thus, greater NH 3 losses after land application from O4 and O5 vs O1 and O2 when manure is surface applied.Injecting manure into the soil significantly reduces these NH 3 emissions after land application.
Reductions of manure NH 3 emissions after manure processing using SLS occurs as nearly all the TAN in manure follow the liquid fraction after SLS, with nearly no TAN left in the solid fraction to promote N volatilization, despite high pH levels.As a result, when SLS is implemented, scenarios handling solid manure now handle liquid manure that have lower NH 3 emissions from stored and land applied manure.O3 has the low- est NH 3 reduction from SLS of the 3 modeled farms (O3, O1, and O4) as the baseline farm handled slurry manure.
Limiting crude protein intake in the dairy herd diet indirectly reduces manure NH 3 emissions by reducing the amount of N excreted with manure.This is especially important in organic farms as they rely on pasture and forage as major components in the dairy diet.Both pasture and forages have higher crude protein contents than grains, which then results in higher excretion of N and potential volatilization in the form of NH 3 .
As with conventional farms, composting increases manure NH 3 emissions, but the magnitude of the increase differs depending on the type of manure managed in the baseline.Bedded pack and solid stack manure management techniques (O4 and O1, respectively) are similar to composting as both result in solid manure sitting in a pile for a 6-mo period before application.However, O3 has a larger increase in emissions given that the baseline farm handles slurry manure that turns to solid manure when composting is adopted.
The effect of management practices on NH 3 emissions for grazing farms follows the same trends as for organic farms (Figure 5).For all farm types (conventional, organic, and grazing), practices that are most effective in reducing NH 3 emissions are injecting manure during land application, reducing crude protein intake, and a combination of injection and SLS.
Implementation of any of these alternative management practices would require additional investment which can be related, but not limited to capital, energy use, labor, and land use (Tan et al., 2021).For example, just the investment costs of a screw press SLS equipment can be between US$50,000 and US$150,000, without considering any costs to integrate or house the equipment (Larson et al., 2021).However, processing systems may also provide reductions in existing costs, such as a reduction in manure hauling.For example, SLS can reduce manure transportation costs for both the solid and liquid fractions as the former has increased nutrient density and reduced moisture (that can be used on farther away fields) and the later can be transported by pumps due to reduced TS reducing hauling costs (Bittman et al., 2011).Implementation of manure injection (one of the most effective practices to reduce NH 3 emissions) would also require added investment in equipment and training time to learn how  to operate the new equipment, but the farmer would benefit from having extra N not volatilized to the environment.Injection costs may not be the only limiting factor however, as integrating injection systems can be challenging for clay soils and those that are wet, reducing application periods for manure which are commonly too short already.Manure storage covers are similar in that they would require a cost to implement and agitating and hauling manure from a covered storage system can increase time and operational difficulties.However, costs may be reduced by limiting the lost N and may improve neighbor relations by reducing odors.
Changing the diet composition to reduce NH 3 emissions would pose other types of challenges in terms of cow nutrition, milk yield, and feed access, which can also add costs to the farmer.A more in-depth analysis is needed to evaluate the economic feasibility of implementing these management practices, but this is out of the scope of this paper.Nevertheless, knowing which practices reduce NH 3 emissions and by how much is useful to respond to sustainability and human health objectives in livestock farms.

Evaluation of Different Functional Units
Intensity of NH 3 emissions from manure is impacted by the functional unit used to present results.Thus far, this study has presented NH 3 emissions in g NH 3 / tonne of excreted manure, a functional unit commonly used for manure management practices.However, other functional units can be more useful in evaluating other priorities for farmers, researchers, and policy makers and/or to compare results with other studies more transparently.Results in this study have also been calculated based on milk production (per kg of FPCM), for stocking density (per AU), and for TS content in manure (kg of TS) Figure 6.
Milk productivity is extremely impactful to the magnitude of emissions when expressing results per FPCM.The difference in NH 3 intensity between highly productive and lower productivity farms (e.g., C3 and C5 vs C1 and C4, O2 and O3 vs O4 and O5) is increased when compared with other functional units.The NH 3 intensity from O4 and O5 is much higher than all other farms when expressed per FPCM due to the low pro-ductivity (24 to 30% higher than O1 and O2) when expressed per FPCM, even when compared with O1 which also has low productivity but has Jerseys cows that have increased fat content in the milk.However, emissions from O4 and O5 are comparable or even lower than O1 and O2 when emissions are expressed per tonne of manure (Tables 4 and A5).Organic farms have comparable NH 3 emissions from manure to conventional and grazing farms when comparing farms that handle the same type of manure (solid, liquid, slurry) and results are expressed per tonne of excreted manure or TS, further strengthening the importance of manure management in emissions.However, when intensity is expressed per FPCM, all organic farms have higher NH 3 emissions than both grazing and conventional farms, due to the lower milk production per cow in organic farms.The interaction of functional units is also different for each farm group, explaining some of these changes in trends.For example, the ratio of FPCM/ AU is higher for conventional farms (15.2-16.3)than for organic (7.8-8.5) and grazing farms (11.4-11.8),but the ratio manure/AU is similar between farm groups (43.7-44.5 for conventional, 39.2-45.0 for organic, and 40.7-40.9for grazing farms), showing the difference in milk production but not in manure excretion among farms.
Ammonia emissions from dairy manure can vary significantly from farm to farm based on the adopted management practices across herd, manure, and feed management.Besides having a negative environmental impact, these emissions also result in losses of valuable N that have to be replaced (most likely by commercial fertilizers) for crop production leading to further environmental impacts and costs to the farmer.Mass balances for shows that only 10% (4 to 17%) of the excreted N reaches the soil in the form of TAN for baseline conventional farms (Figure 7).Implementing management practices can improve this amount to 33% for combined liquid and solid fractions after injection and SLS.This has important consequences for dairy farms as TAN is more available than organic N and can be taken by crops immediately after land applied.
For organic and grazing farms, a significant amount of excreted N is left in pasture, reaching more than 50% in some organic farms.As a result, the amount of TAN that can be applied to growing crops in a controlled manner can be as low as 3% of excreted N and up to 10% with management practices.The availability of organic N is also important for nutrient management purposes at the farm but is plant available starting the second year from application (Laboski and Peters 2012).Even though P is not the focus of this study, it is included to guide strategies for nutrient management planning.Mass balances of P are simpler than for TN as P is not lost to the atmosphere in a gaseous form (Figure A1).As this study assumes that no P is lost through manure management, the partition of P depends entirely on the amount of manure excreted on pasture, and if the farm adopts SLS.

CONCLUSIONS
This study modeled conventional, organic, and grazing dairy farms to quantify NH 3 emissions from manure and evaluated alternative management practices that can reduce these emissions.Most NH 3 from manure is emitted from storage and after land application regardless of farm type and layout.Storage of solid manure often results in higher emissions than slurry or liquid manure due to higher pH.Solid manure limits infiltration of N into the soil, promoting NH 3 emissions after land application.The most effective practice to reduce NH 3 emissions is injecting manure into the soil, followed by reduction of crude protein in the dairy ration.The intensity of manure NH 3 emissions depends significantly on the functional unit used to express   results and the function of the system or subsystem under study.It is important to present different units so that mitigation practices can evaluate all variables important to the farm.

Figure 1 .
Figure 1.Modeled dairy farm components to quantify ammonia (NH 3 ) emissions from manure.Dashed lines show steps (i.e., manure processing) that are only included in the evaluation of alternative management practices that seek to reduce NH 3 emissions.

C1
into a solid and liquid fraction before storage (solids and liquids stored and land applied separately).C1-C3, O1, O3, O4, G1-G3 Composting Composting manure using aeration methods of turning.C1-C3, O1, O3, O4, G1-G3 Combined practices SLS+injection Both solid and liquid fractions are stored for six months and injected into the soil.C3, O1, O3, O4, G1-G3 AD+SLS Digesting manure before manure storage to produce and collect biogas.This also increases mineralization of organic nitrogen to ammonium.C3 AD+SLS+injection Digestate is separated before going to storage (six months) and then injected into the soil.C3 a Based on (Uddin et al., 2021).SLS: Solid-liquid separation.AD: Anaerobic digestion.
Aguirre-Villegas et al.: Ammonia Emissions from Manure on Dairy Farms Table 3. Equations and factors used to estimate NH 3 emissions from each manure management ammonia emissions NH 3 −N (kg/m 2 /d); TAN: total ammonia nitrogen in manure (kg N/m 2 ); c: time conversion (86,400 s/d); y: manure density (kg/m 3 ); r: resistance of NH 3 transport to the atmosphere (s/m), r barn: HSC[1-0.027(20-T)],HSC: housing specific constant (260 s/m), r storage: 75 (manure with crust), 19 (manure without crust), 10 (solid manure); M: manure solution mass per area of exposed surface (kg/m 2 ), the area exposed to manure in the barn is assumed to be 3.5 m 2 for mature animals and 2.35 m 2 for growing heifers.The area of manure storage is estimated assuming a rectangular profile for manure slurry and liquids and 5 m deep and a conical shape for solid manure stored in piles; Q: equilibrium coefficient: Kh * Ka, Kh = 10[1478 / (T + 273) − 1.69], Ka = 1 + 10[0.09018+ 2,729.9/ (T + 273) -pH], T: manure temperature (°C), pH: manure acidity (Rotz and Oenema 2006) Bedded pack 0.25 kg NH 3 -N/kg of N manure (IPCC, 2019) Dry lot 0.30 kg NH 3 -N/kg of N manure (IPCC, 2019) Manure excreted on pasture 0.20 kg NH 3 -N/kg of N manure ( the percent of nitrogen following the solids fraction, with the remining following the liquid fraction.b
Figure 3. Ammonia (NH 3 ) emissions from manure in baseline modeled conventional farms and under alternative management practices.SLS: solid-liquid separation, AD: anaerobic digestion.
Aguirre-Villegas et al.: Ammonia Emissions from Manure on Dairy Farms

Figure 4 .
Figure 4. Ammonia (NH 3 ) emissions from manure in baseline modeled organic farms and under alternative management practices.SLS: solid-liquid separation, AD: anaerobic digestion.
Figure 5. Ammonia (NH 3 ) emissions from manure in baseline modeled grazing farms and under alternative management practices.SLS: solid-liquid separation, AD: anaerobic digestion.

Figure 6 .
Figure 6.Baseline modeled ammonia (NH 3 ) emissions from manure under different functional units, where FPCM: fat and protein corrected milk, AU: animal unit, and TS: total solids.
Aguirre-Villegas et al.: Ammonia Emissions from Manure on Dairy Farms Table4.Total manure ammonia (NH 3 ) emissions per tonne of excreted manure, kg of total solids (TS) in excreted manure, per kg of fat and protein corrected milk (FPCM), and per animal unit (AU) from modeled baseline conventional, organic, and grazing farms.Detailed information for each manure management step is presented in

Figure 7 .
Figure 7. Manure nitrogen (N) mass balances for a) conventional, b) organic and c) grazing baseline farms and after alternative management practices.Data in parenthesis represents the range for all the modeled farm scenarios.

Figure A1 .Functional
Figure A1.Manure phosphorus (P) mass balances for a) conventional, b) organic and c) grazing baseline farms and after alternative management practices.Data in parenthesis represents the range for all the modeled farm scenarios.

Table 1 .
General characteristics of modeled conventional, organic, and grazing farms Only lactating cows are fed on pasture on grazing type farms.It is assumed that dry cows and heifers are only fed forages and grains during both grazing and non-grazing seasons. e

Table 2 .
Mitigation practices evaluated for NH