Assessing Ammonia (NH₃) Emissions, Precursor Gas (SO₂, NO_x) Concentrations, and Source Contributions to Atmospheric PM_2.5 from a Commercial Manure Composting Facility

Abstract

Increased ammonia (NH₃) emissions from intensive agriculture negatively affect environmental and ecosystem health, contributing to formation of particulate matter (PM) and the potent greenhouse gas, N₂O. Better understanding NH₃ emissions from the manure composting process and their behavior as a constituent of the atmospheric aerosol load is a crucial element in creating better farm management systems, improving public health outcomes, and mitigating the broader environmental and climatic impacts of agriculture. Retarded generation of PM with a major constituent source of NH₃ is a primary mechanism for evaluating the effects of agricultural contribution to PM. This study aimed to quantify NH₃ emissions, examine the influence of environmental factors, and investigate the relationship between precursor gases (SO₂, NO_x, NH₃) and PM_2.5 at a modern manure composting facility in Paju, South Korea. Over 35 days, average internal concentrations of NH₃, SO₂, and NO_x were significantly higher than external levels. NH₃ concentrations reached 3.64 ± 0.06 mg m⁻³ at 3 m height and 2.43 ± 0.16 mg m⁻³ at ground level, while the total NH₃ flux from the facility was 24.47 ± 1.39 NH₃-N kg d⁻¹. Internal PM_2.5 concentrations (36.9 ± 2.6 µg m⁻³) were about 50% higher than external levels (23.7 ± 2 µg m⁻³), with a moderate correlation (r = 0.341) suggesting some contribution of external PM_2.5 to internal levels. Despite large quantities of internal emissions, the facility’s sealed design with a negative pressure ventilation system effectively minimized external emissions. These results suggest that while manure composting facilities are significant sources of NH₃ and PM_2.5, advanced systems like high-volume ventilation and scrubbing technologies can effectively reduce their impact on regional air pollution, contributing to better environmental management in agriculture.

1. Introduction

Agriculture-related emissions, including greenhouse gases (GHGs), ammonia (NH₃), particulate matter precursors, and toxic pollutants, can travel across regions, adversely affecting air quality and posing significant risks to human health [1]. These far-reaching impacts highlight the urgent need for effective emission management strategies in agricultural systems to mitigate their environmental and health consequences [2].

In 2017, agricultural activities were estimated to be responsible for approximately 80% of NH₃ emissions from the Republic of Korea [3], in line with similar figures from the United States, China, and the European Union [4,5,6,7]. Livestock manure is thought to be the single largest contributor to NH₃ emissions [8], estimated to account for around 50% of the total in the USA and China [9,10]. Increased NH₃ emissions as a result of intensive agriculture are known to negatively affect environmental and ecosystem health through eutrophication [11] and as an indirect source of the potent greenhouse gas, nitrous oxide (N₂O) [5,12].

Per capita meat consumption in South Korea is projected to continue increasing over the next decade, with consumption of chicken and pork having doubled since the year 2000 and beef consumption increasing by more than 40% [13]. Available grazing and arable land is limited in South Korea [14], and pressure on agriculture production means that concentrated animal feeding operations (CAFOs) are common. Effectively managing the manure produced in these facilities is important for controlling NH₃ and odor pollution [10,15]. Manure composting is an appealing management technique involving the growth of aerobic micro-organisms within the manure that break it down into a nutrient-rich organic fertilizer that can be applied to agricultural land [7]. When composted, biological degradation of the manure results in the production of sulfur dioxide (SO₂) and nitric oxide (NO_x) gases, as well as GHGs such as carbon dioxide (CO₂), methane (CH₄), and N₂O [16,17]. Ammonium (NH₄⁺) is also created through microbial processes within the manure and although it is not volatile itself, it exists in equilibrium with NH₃ gas [12]. Factors that affect NH₃ emissions from manure compost include temperature, moisture content, pH, initial nitrogen content, and air turbulence [7,12]. Aerating compost (via forced aeration, mechanical turning, and the use of organic bulking agents such as wood chips) has also been found to reduce direct emissions of the GHGs CH₄ and N₂O, lowering the impact of manure composting on climate change, but this causes increases in the pH and temperature of the manure, shifting the equilibrium towards gaseous NH₃, increasing emissions, and leading to interest in capturing and scrubbing NH₃ from the air [18,19,20].

NH₃ emissions are thought to be an important contributor to the formation of secondary particulate matter. NH₃ gas reacts with acidic compounds (such as SO₂ and NO_x) in the atmosphere, resulting in PM_2.5 formation [21,22]. Studies have investigated the role of NH₃ interactions with SO₂ and NO_x in contributing to severe PM_2.5 episodes in both South and East Asia [6,23,24], as well as in Europe [25,26] and North America [9,27]. Poor air quality is among the greatest issues currently facing East Asian societies, negatively affecting health outcomes [28,29] and economic productivity [30] across the region.

Better understanding NH₃ emissions from the manure composting process and their behavior as a constituent of the atmospheric aerosol load is a crucial element in creating better farm management systems, improving public health outcomes, and mitigating the broader environmental and climatic impacts of agriculture. This study aimed to characterize and quantify SO₂, NO_x, NH₃ (SNA), and PM_2.5 emissions from a modern manure composting facility in South Korea in order to (a) calculate NH₃ flux, (b) assess the role that physical and environmental variables play in NH₃ emission, and (c) investigate any relationship between SNA and PM_2.5 formation within the facility. In addition, this paper aims to evaluate whether this modern facility, equipped with an NH₃ scrubbing system, is able to minimize external emissions that may contribute to regional air pollution.

2. Materials and Methods

A multi-instrument monitoring system was set up to monitor SNA emissions and PM concentrations at a manure composting facility located in Papyeong-myeon, Paju City, Gyeonggi Province, South Korea. The chosen facility began operations in 2017 and is a purpose-built, modern manure composting facility installed with a high-volume ventilation and scrubber system. The monitoring instruments were operated over a 5-week period between 11 June and 16 July 2020. Internal concentrations of SNA and external concentrations of NH₃ and PM_2.5 were monitored continuously. In addition, internal PM_2.5 concentrations were monitored for shorter durations during the monitoring period.

2.1. Monitoring Site

The Paju manure composting facility (PMCF) has been designed to process animal manure from animal feeding operations (AFOs) into organic fertilizer for agricultural use. The facility is composed of several distinct sections. Fresh manure is delivered from local poultry and livestock farms and stored for 25 days in an atrium sectioned into ten large bays for bulk manure storage prior to composting. The manure is then mixed at a 85:15 manure–sawdust ratio and added to one of two halls, to agitate the manure during the composting process. In each hall, the manure is mixed daily by a conveyor-type agitator that lifts, aerates, and deposits the manure along three windrows. During the daily agitation, the manure is gradually moved from the input to the extraction end of each hall over a period of 18 days. Following the composting agitation process, the manure is extracted and taken to a composted manure storage hall where it is allowed to rest for approximately 33 days, before being moved to a facility for packaging composted manure for sale as organic fertilizer. The facility also contains a separate underground tank for processing the liquid/slurry component of the manure. This study focused on monitoring precursor gas and PM emissions from one of the manure composting agitation halls (Figure 1). It was expected that the greatest NH₃ losses would occur during this stage of the manure composting process, due to the regular agitation and aeration of the manure.

The building is installed with two 1400 m³ min⁻¹ capacity extraction fans which remove the air in the facility, maintaining a negative internal pressure, and pass it through scrubbers designed to remove precursor gases and PM before it is emitted externally. Two ceiling-mounted vents leading to the extractor fans are located in each manure composting agitation hall.

2.2. Monitoring System

A weatherproof station to house the monitoring instruments was installed outside the PMCF, adjacent to one of the manure composting agitation halls. The instruments were housed externally to reduce the possibility of contamination and corrosion of the instruments due to high concentrations of SNA precursor gases inside the facility. Filtered gas samplers were placed along a walkway inside the manure composting agitation hall. The instruments were connected to the filtered gas samplers with 10 m lengths of 4 mm diameter Teflon tubing, wrapped in a heated trace to reduce condensation, and run through holes in the building’s window fittings. Filtered gas samplers were used to ensure that the tubing did not become blocked with dust particles entrained during the daily manure agitation cycles.

2.2.1. Precursor Gas Measurement

NH₃ gas concentrations were monitored via a 6-channel gas sampling doser (INNOVA1403) connected to an INNOVA1512i photoacoustic NH₃ gas analyzer (INNOVA1403 & 1512i, LumaSense Technologies, Ballerup, Denmark). Four channels were used to monitor NH₃ concentrations along the length of the manure composting agitation hall. In addition, two gas-sampling channels were installed outside the facility, one located at a height of 3 m above ground level adjacent to a window on the outer wall of the manure composting agitation hall and one located at ground level next to the instrument station. The INNOVA gas-sampling system provided NH₃ concentrations (measurement range: 0–500 ppm, sensitivity: 1 ppb) sequentially across the 6 channels at approximately 45 s intervals. Except where explicitly mentioned, internal NH₃ emissions are reported as averages of the data from the four internal channels.

Sulfur dioxide (SO₂) and nitrogen oxide (NO_x) gas concentrations (measurement range: 0–500 ppb, sensitivity: 0.1 ppb) were sampled at five-minute intervals at single points inside the manure composting agitation hall, using KENTEK gas analyzers (KENTEK Co., Ltd., Daejeon, South Korea). NO_x was measured on a KENTEK MEZUS 210 instrument via chemiluminescence. SO₂ was measured by a KENTEK MEZUS 110 instrument, using an ultraviolet fluorescence method. NO_x gas concentrations were only available prior to 4 July 2020. While heated traces were used to avoid moisture condensation inside the Teflon sampling tubes, condensation buildup after this date caused abnormal NO_x data readings after 4 July.

The presented precursor gas concentration datasets measured at the facility were transformed to 15 min averages to reduce noise and to aid cross comparison between data sampled at different time intervals. For comparison with environmental variables and regional PM and gas concentration measurements, precursor gas concentrations were converted to hourly average values to match the data resolution of these externally sourced data.

Regional hourly SO₂ and nitrogen dioxide (NO₂) gas concentration data for the duration of the study, measured at local government-operated air quality monitoring sites in Paju, were downloaded from the Air Korea government website (https://www.airkorea.or.kr/index (accessed on 6 September 2020)).

This study investigated the difference in gas concentrations during daytime and nighttime and between weekdays and weekends. This was because manure agitation only occurs during the daytime at weekdays. Daytime was defined in this study as between 7 am and 7 pm, while nighttime was from 7 pm to 7 am. Weekends were defined as the 48 h from 7 am on Saturday until 7 am on Monday.

2.2.2. Particulate Matter Measurement

An automatic beta-ray-absorption ultrafine-dust measuring device was installed at ground level outside the facility next to the instrument station, to monitor external PM concentrations. The β-ray instrument conducted hourly air sampling for the duration of the sampling period.

A GRIMM 11-D Aerosol Spectrometer Optical Particle Counter (GRIMM OPC 11-D, Grimm Aerosol Technik Ainring GmbH & Co. KG, Ainring, Germany) instrument was used to sample internal PM concentrations in the composting conveyor hall for two separate periods, from 24 to 26 June and from 8 to 16 July 2020.

Regional particulate matter concentration data for the duration of the study, measured at local government-operated air quality monitoring sites in Paju, were downloaded at an hourly resolution from the Air Korea government website (https://www.airkorea.or.kr/index).

2.2.3. Environmental Variables

Hourly local temperature and humidity data for the duration of the study, measured at a government-operated meteorological station in Paju, were downloaded from the Korean Meteorological Association website (https://data.kma.go.kr/ (accessed on 6 September 2020)).

2.3. Ammonia Emissions Flux Calculations

For calculation of the flux of NH₃ emissions from the manure being processed inside the manure composting agitation hall, the extraction fans were disabled, allowing gaseous NH₃ concentrations to rise until they reach peak concentration, probably as a result of equilibrium in the concentration gradient at the manure–air boundary. The difference between normal (ventilated) NH₃ concentrations and saturated NH₃ concentrations was divided by the time taken for this increase to occur, and the results were averaged across the four internal NH₃ sampling channels.

E_NH3 = (C_NH3(sat.) − C_NH3(vent.))/Δt

(1)

Equation (1) demonstrates how the NH₃ emission rate was calculated, where C_NH3(sat.) is the point at which NH₃ concentrations in the air reach equilibrium with NH₃ concentrations at the manure surface (i.e., there is no longer a concentration gradient, leading to a stabilization of NH₃ concentration in the air), C_NH3(vent.) is the ambient concentration of NH₃ under ventilated conditions (i.e., when there is negative pressure and emitted NH₃ is being constantly removed and scrubbed by the ventilation system during ordinary operation), and Δt is the change in time in minutes between these two states following the disengagement of the ventilation system.

2.4. Statistical Analyses

The bivariate correlations between gas concentrations and environmental variables were tested using Pearson’s correlation coefficients, while the significance of differences between precursor gas concentrations under day, night, weekday, and weekend conditions was tested using a one-way ANOVA test at a 95% confidence level, using R package version 4.0.2.

3. Results and Discussion

3.1. Environmental Variables: Temperature and Relative Humidity

Data for environmental variables showed that temperature and relative humidity maintained a relatively consistent diurnal pattern throughout the monitoring period (Figure 2). While separate internal data for temperature and humidity were not measured, it is expected that these were closely related to local meteorological conditions. The agitation hall was not climate controlled but the air was constantly ventilated to maintain a negative pressure situation, ensuring that the internal environment remained similar to prevailing local environmental conditions.

The average temperature during the monitoring period was 22.3 °C, with a coldest temperature of 16.1 °C degrees early in the morning of 17 June and an overall high of 34 °C on the afternoon of 22 June (

Table 1. Summary of average, minimum, maximum, and standard deviation values of temperature and relative humidity, along with daily maximum and minimum averages.

	Temperature (°C)	Relative Humidity (%)
Average	22.3	82
Min.	16.1	32
Max.	34.1	100
St. Dev.	3.6	15.9
Daily Max. Average	27.1	98.4
St. Dev.	2.9	2.9
Daily Min. Average	18.3	60.2
St. Dev.	1.3	13.7

). Daily minimum temperatures (mean = 18.3 °C, standard deviation = 1.3 °C) varied slightly less than daily maximum temperatures (mean = 27.1 °C, standard deviation = 2.9 °C). As expected, relative humidity showed an inverse pattern compared with temperature, with lower humidity during the day and higher humidity during cooler nights (Figure 2). Relative humidity reached dew point (100% RH) most nights. The relative humidity data attest to two periods of rain during the monitoring period from 23–26 June and 12–14 July. The variable environmental data show that the main variation was diurnal, with no discernible trends over the whole monitoring period.

3.2. Gas Concentrations

3.2.1. Ammonia Monitoring Results

NH₃ was the most abundant of the monitored gases, with internal NH₃ concentrations averaging 57.1 ± 1.3 mg m⁻³ across the entire monitoring period (excluding the two occasions when the extractor fans where disabled—these periods are discussed separately). There was substantial variation in NH₃ concentrations depending on the timing of the measurements. The highest average NH₃ concentrations were recorded during weekend daytime at 66.3 ± 8.2 mg m⁻³ (

Table 2. Average internal and external concentrations of monitored SNA gases and PM_2.5.

		NH3 Int	NH3 Ext Hi	NH3 Ext Lo	NO Fac	NO2 Fac	SO2 Fac	PM2.5 Int	PM2.5 Ext
		mg m−3			μg m−3
Week Day	Mean	63.951	3.850	3.155	34.007	9.000	371.004	-	23.570
	Std. Err.	2.530	0.070	0.366	4.052	1.846	30.452	-	3.832
Week Night	Mean	54.217	3.555	2.053	31.928	4.991	321.118	-	22.171
	Std. Err.	2.074	0.084	0.092	5.421	1.052	23.671	-	3.277
W/end Day	Mean	66.278	3.793	2.079	27.939	10.332	264.985	-	22.337
	Std. Err.	8.157	0.179	0.092	6.616	1.363	45.646	-	3.529
W/end Night	Mean	37.911	3.180	1.736	49.723	7.032	147.155	-	23.319
	Std. Err.	2.558	0.112	0.056	13.629	1.328	15.669	-	2.644
Over all	Mean	57.063	3.642	2.425	33.350	7.276	307.224	36.901	23.713
	Std. Err.	1.316	0.056	0.159	3.153	0.846	17.426	2.599	1.959

), although this was not significantly different (p < 0.05) from the average weekday daytime concentrations of 64 ± 2.5 mg m⁻³ (Figure 3). Average weekday nighttime NH₃ concentrations were found to be slightly lower at 54.2 ± 2.1 mg m⁻³, and they were statistically significantly different from the daytime concentrations. Weekend nighttime average NH₃ concentrations were measured to be less than 60% of the average daytime levels, at 37.9 ± 2.6 mg m⁻³. The differences in average concentrations of NH₃ during separate time periods probably reflect differences in environmental variables (i.e., temperature and humidity) and manure agitation activity between night and day, and the absence of manure agitation activities at weekends.

The diurnal fluctuations in NH₃ concentrations and the differences in concentration patterns between weekdays and weekends were clearly seen across all four internal NH₃ sampling points (Figure 4). On weekdays, all internal sampling points showed concentrations peaking sharply, three times, during the daily agitation events, matching the travel of the agitator along the three manure windrows (Figure 1). At weekends, a single broader NH₃ concentration peak was observed. It is likely that agitation itself resulted in shorter, higher magnitude NH₃ concentrations that were quickly cleared by the ventilation system, reducing peak NH₃ emissions outside of the agitation period. By contrast, at weekends, there was a lower but broader NH₃ concentration peak during the day (Figure 4). On weekdays, the lowest concentrations of NH₃ were usually recorded immediately after manure agitation had ceased, while at weekends, the lowest concentrations of NH₃ are observed overnight, in the early morning prior to sunrise. The pattern of emissions at the weekend may suggest a greater role of temperature in NH₃ emissions at this time.

Turning of the manure during agitation cycles aerates the compost and provides a supply of oxygen to the microorganisms that are performing aerobic digestion [20]. The biological oxidation of C to CO₂ releases heat, which raises the temperature of the compost into a thermophilic (>40 °C) state [20]. In addition, aeration exposes the manure to air, allowing the release of dissolved CO₂, which raises the pH of the manure [12]. Increases in either temperature or pH promote the emission of NH₃ by increasing the dissociation of NH₄⁺ [12,18,19]. The process of turning/aeration itself increases the capacity for NH₃ vitalization through air turbulence across the greatly expanded manure surface area [12], leading to the high concentrations of NH₃ recorded during the agitation events. However, alongside volatizing the reservoir of NH₃ gas in the manure, this process also temporarily lowered the temperature of the manure, which probably led to the lower NH₃ concentrations that were recorded directly after agitation. Following aeration, the increased availability of oxygen led to temperature increases within the compost over the next day [17]. Following aeration that occurred during the week, the temperature of the undisturbed manure at weekends remained higher, leading to increased NH₃ concentrations during the warmer daytimes and lower NH₃ concentrations during the cooler nights.

The agitation events can be clearly recognized in the raw NH₃ concentration dataset based on the internal sampling points (Figure 5). Three periods of raised concentrations are apparent, representing the agitator travelling along each of the three windrows. The first period of raised concentrations with the highest peaks (at 275–500 mg m⁻³) shows the agitation of the windrow closest to the internal sampling points and distinctly displays sequential concentration peaks that arose as the agitator passed each of the four sampling points.

External NH₃ concentrations were generally an order of magnitude lower than internal concentrations. Average NH₃ concentrations at 3 m height above ground level adjacent to the manure agitation hall window were 3.64 ± 0.06 mg m⁻³ across the whole monitoring period, while average external NH₃ concentrations at ground level were slightly lower at 2.43 ± 0.16 mg m⁻³ (Figure 6; Table 2). There was no substantial diurnal/weekday–weekend variation in external NH₃ concentrations at 3 m height, although at ground level, weekday daytime NH₃ concentrations (3.1 mg m⁻³) were approximately 50% higher than during the other time periods (1.7–2.1 mg m⁻³). Although the data showed that the NH₃ emitted by the facility was low compared with internal concentrations, increases in external NH₃ concentrations were detected coincident with increases in internal concentrations, demonstrating that some leakage did occur. In addition, external NH₃ concentrations at 3 m height were comparable to average concentrations of 3.7–5 mg m⁻³ that were measured inside a Chinese commercial manure-belt layer house [31], and as such, they are not negligible. It was expected that the external sampling point at 3 m height would record higher average NH₃ concentrations than the ground-level sampling point, due to the potential for emissions through the window fitting. However, external ground-level NH₃ concentrations, although lower on average, regularly displayed greater peak concentrations of NH₃ during manure agitation activity (Figure 6), although it is unclear why this might have been the case.

3.2.2. SO₂

Recorded concentrations of SO₂ gas shared a similar pattern to NH₃ concentrations (Figure 7), displaying high but variable concentrations during manure agitation and more broadly elevated concentrations during weekend daytimes when agitation did not occur. This similarity was reflected by a strong positive correlation (r = 0.751, p < 0.001) between the concentrations of NH₃ and SO₂ (Figure 8,

Table 3. Calculated 15 Min/1 h Correlations—NH₃ Int, NH₃ Ext 1, NH₃ Ext 2, SO₂, NO, NO₂, Temp, RH, Beta-Ray PM, Paju PM.

	NH3 Int	NH3 Ext Wi	NH3 Ext Gr	NO Fac	NO2 Fac	SO2 Fac	PM2.5 Int	PM2.5 Ext
Temp	0.101	0.086	0.071	−0.191	0.476	0.073	0.277	0.033
RH	−0.072	−0.012	−0.063	0.059	−0.600	−0.047	−0.322	0.078
PM2.5 Paju	0.191	0.154	−0.139	0.072	0.096	0.053	0.341	0.666
NO2 Paju				0.023	−0.025		0.084	0.412
SO2 Paju						−0.204	−0.017	0.404
NH3 Int
NH3 Ext wi	0.831
NH3 Ext gr	0.429	0.345							* = 0.05
NO Fac	0.191	0.140	0.107						** = 0.01
NO2 Fac	−0.084	−0.139	−0.102	0.290					*** = 0.001
SO2 Fac	0.751	0.702	0.400	−0.008	−0.201
PM2.5 Int	0.047	0.117	0.026	0.063	0.051	0.155
PM2.5 Ext	0.096	0.072	−0.110	−0.030	0.028	−0.065	0.088

* Light blue cell means p = 0.05, ** blue cell means p = 0.01, *** pink cell means p = 0.001.

). Although sharing a similar profile, SO₂ concentrations were consistently two orders of magnitude lower than NH₃ concentrations, averaging 307.2 ± 17.4 µg m⁻³ over the monitoring period. Unlike NH₃, the average concentrations of SO₂ during weekdays (371 ± 30.5 µg m⁻³) and nights (321.1 ± 23.7 µg m⁻³) were not found to be significantly different (p < 0.05). In addition, unlike NH₃, average weekend daytime SO₂ concentrations (265 ± 45.6 µg m⁻³) were lower than both weekday daytime and nighttime concentrations although also not statistically different (p > 0.95). However, average SO₂ concentrations on weekend nights (147.2 ± 15.7 µg m⁻³) were found to be significantly lower (p < 0.05) than SO₂ concentrations at all other times. During weekday manure agitation events, SO₂ concentrations shared a similar pattern with NH₃, with three main concentration spikes related to the sequential agitation of the three manure windrows. Weekday SO₂ concentrations were consistently measured to be lowest in the period directly after agitation activity ceased. However, again showing similarities to NH₃ concentration patterns, weekend daytime concentrations of SO₂ were elevated for longer durations throughout the day and the lowest concentrations occurred during the night, prior to sunrise.

These data suggests that both NH₃ and SO₂ were emitted during the manure composting process, and the similarity in the patterns of the recorded concentrations may suggest similar emission mechanisms. SO₂ can result from incomplete oxidation during manure breakdown by the activity of microorganisms, perhaps as a result of intermittently aerobic conditions [17]. However, unlike this study, research by Wang [31] found that indoor SO₂ concentrations in a belt-layer house Beijing related predominantly to the ambient air. While that study found a positive correlation between internal and regional SO₂ concentrations, data from the current study rule out any substantial effect of ambient air on internal SO₂ concentrations at the PMCF (Table 3). Indeed, the strong positive correlation between NH₃ and SO₂ concentrations suggests that manure composting is the predominant source of both. Two existing studies both found that SO₂ was emitted during the manure composting process; however, they produced widely different results for sulfur mass loss, ranging from <1% [32] to approximately 20% [17], and the SO₂ emission mechanism is not well understood.

3.2.3. NO_x

NO_x represents the total combination of NO and NO₂, which were both measured during this study. In this study, NO was found to be the predominant component of NO_x (Table 2), with average concentrations (33.4 ± 3.2 µg m⁻³) four-and-a-half times greater than those of NO₂ (7.3 ± 0.8 µg m⁻³). Unfortunately, the NO_x concentration data were contaminated after 4 July 2020, due to a build-up of condensation in the sampling tube, meaning that NO_x data for the final 12 days of the study were not used. Weekday (34 ± 4.1 µg m⁻³) and night (32 ± 5.4 µg m⁻³) concentrations of NO were similar, and weekend daytime concentrations were slightly lower (28 ± 6.6 µg m⁻³). Average NO concentrations were greatest on weekend nights (49.7 ± 13.6 µg m⁻³); however, these difference were not found to be statistically significant (p < 0.05). Similar to NH₃ and SO₂, NO concentrations recorded increases during manure agitation (Figure 7); however, NO concentrations at the weekends appeared to display an approximately inverse pattern to NH₃ and SO₂. That is, when NH₃ and SO₂ concentrations were elevated during the weekend daytime, NO concentrations were depleted, and this situation was reversed during weekend nights. NO is produced as a result of nitrification and denitrification processes during composting [17,31]. Similar to the increased concentrations of NH₃ and SO₂ during agitation, NO concentrations may have been increased as a result of manure turning, by exposing more manure to the air, increasing the potential for diffusion [17]. The lower concentrations of NO at weekend daytimes may be explained through the inhibition of nitrobacteria activity by high temperatures [17,19], with greater concentrations of NO overnight when temperatures were lower. Fillingham [18] reported an inverse relationship between NH₃ emissions and NO emissions, with greater NH₃ emissions in well-aerated and higher-temperature conditions, while NO emissions increased during manure storage. Previous research has found that NO emissions substantially increase only later in the composting process (beyond 15–30 days), after the thermophilic phase, when mesophilic nitrifying bacteria are able to grow [33,34].

Unlike NO, concentrations of NO₂ displayed a pattern more similar to NH₃, with elevated concentrations during weekdays and weekend daytimes, and lower concentrations at night (Table 2). Internal NO₂ concentrations were not correlated with ambient regional NO₂ concentrations (Table 3), suggesting that the main source of NO₂ in the facility was the composting manure. However, analysis showed that NO₂ concentrations had a strong negative correlation with relative humidity (r = −0.6, p < 0.001) (Figure 9, Table 3), and a moderate positive correlation with temperature (r = 0.476, p < 0.001), although the latter is likely to reflect only the relationship between temperature and humidity. It is possible that greater humidity could have contributed to the indirect relationship between NO₂ and NO concentrations by providing water for the reaction from NO₂ to nitric acid and NO_x (Equation (2)), thus also explaining the negative correlation with humidity.

3 NO₂ + H₂O + → 2 HNO₃ + NO

(2)

3.3. Particulate Matter Concentrations

Average external PM_2.5 concentrations during the monitoring period were 23.7 ± 2 µg m⁻³ (Table 2), and unlike internal NH₃ concentrations, there was no significant variability between day–night or weekday–weekend periods. The hourly external PM_2.5 concentrations measured at the PMCF closely matched regional PM_2.5 concentrations (Figure 10), with which they had a strong positive correlation (r = 0.67, p < 0.001).

There were two exceptions, on the 25 and 26 June 2020, respectively, when external PMCF PM_2.5 concentrations greatly diverged from regional concentrations (Figure 10). However, internal NH₃, SO₂, and NO_x concentration data for these two dates did not appear to show any large anomalies that might have corresponded with the high PM_2.5 concentrations. This suggests that the elevated PM_2.5 concentrations measured outside the facility on the 25 and 26 June did not originate from the manure composting function of the facility and were from a different source. Together with the high correlation with regional PM_2.5 concentrations, these data suggests that the PMCF is unlikely to be a significant local or regional source of PM_2.5.

Average internal PM_2.5 concentrations were 36.9 ± 2.6 µg m⁻³. Similar PM_2.5 concentrations (39–56 µg m⁻³) were found during monitoring of slatted-floor hog houses in China [35], and these were are slightly lower than PM_2.5 concentrations (40–80 µg m⁻³) at a commercial laying hen house in Canada [36]. However, they were much lower than the PM_2.5 concentration (100–140 µg m⁻³) measured by Wang in Chinese manure-belt laying hen houses [31], and in a range of swine (75–177 µg m⁻³) and poultry houses (233–389 µg m⁻³) in the mid-western USA [37]. Elevated periods of PM_2.5 concentrations appear to have coincided with manure agitation (Figure 11) and could have been a result of mechanical lofting of fine-fraction dust particles and/or secondary aerosol formation from elevated concentrations of precursor gas. Secondary aerosol formation can occur when NH₃ gas reacts with acidic species SO₂ and NO_x to form NH₄⁺ salts like (NH₄)₂SO₄ and NH₄NO₃ [9,21]. Reaction between NH₃ and sulfuric acids occurs preferentially over nitric acids under normal atmospheric conditions [21], and the reaction rate between NH₃ and sulfuric acids is faster than for nitric acids [22]. Peak PM_2.5 concentrations were recorded at 889 µg m⁻³ and 1479 µg m⁻³ on the 14 and 15 July 2020, respectively (Figure 11), showing that very high internal PM_2.5 concentrations occasionally occurred for brief durations coincident with agitation activity. However, PM_2.5 concentrations rarely rose above 200 µg m⁻³ and it is likely that the high-volume ventilation system at the PMCF was an important factor in the relatively low internal PM_2.5 concentrations, removing fine-fraction dust particles and precursor gases before PM_2.5 formation was able to occur.

Due to the shorter duration (approx. 10 days) over which internal PM_2.5 was measured, it was not possible to assess statistical differences between weekend and weekday concentrations. Although internal concentrations of PM_2.5 were observed to reach much higher levels on weekdays following manure agitation, concentrations appeared to show comparatively little variability at weekends. A weak but significant correlation was found between regional and internal PM_2.5 concentrations (r = 0.341, p < 0.001), suggesting that PM_2.5 in the ambient air was responsible for a portion of the internal PM_2.5 concentrations; however, no correlation existed between external and internal PM_2.5 (Table 3).

3.4. Ammonia Emission Rate

An experiment was undertaken during a 5 h period between 12:30 and 17:30 on Saturday 11 July 2020, in which the high-volume ventilation/scrubber system was turned off in order to measure the flux of NH₃ from the composting manure (Figure 12). Prior to the experiment, under ventilated conditions, internal NH₃ concentrations were stable at approximately 100 mg m⁻³, rising to around 150 mg m⁻³ as all entrances and exits to the facility were closed. After the ventilation system was disabled, NH₃ concentrations were recorded to drop to an average of ~30 mg m⁻³ initially, before rising quickly to reach values of ~250 mg m⁻³ within 30 min. Peak concentrations of ~325 mg m⁻³ were reached after approximately 40 min. Concentrations of SO₂ followed a similar pattern to NH₃, initially dropping before exceeding the monitoring instrument’s detection limit at 1.31 mg m⁻³ (500 ppb). It is not clear why NH₃ concentrations dropped after the ventilation system was disabled, but we hypothesize that the negative pressure environment created by the ventilation system caused a greater flux of NH₃ from the manure surface due to the pressure gradient, and therefore, when the ventilation system was disabled, NH₃ was emitted at a reduced rate. Alternatively, the decrease could have been related to lower air turbulence across the manure surface, which would be expected to reduce convective mass transfer of NH₃ from the manure surface to the air [12].

The emission rate of NH₃ was calculated according to Equation (1), based on the time taken for NH₃ concentrations to rise from their minima at ~30 mg m⁻³ to the inflection point at ~250 mg m⁻³, after which the concentration increase slowed down. The results showed that under stable non-ventilated conditions, the NH₃ emission rate inside the manure agitation hall was between 6.16 and 8.02 (mean = 7.28 ± 0.41) NH₃-N mg min⁻¹ m⁻³ (

Table 4. NH₃ emission rates inside the Paju Manure Composting Facility under non-ventilated conditions.

	Emission Rate	Scaled Emission Rates
	mg min−1 m−3	g h−1 m−2 Manure	g h−1 m−3 Manure	g kg−1 Manure *
Mean	7.28	2.77	1.85	1.52
Std. Error	0.41	0.16	0.1	0.09

* over the entire 18-day composting period.

); extrapolated, the total mass of NH₃ emitted daily was between 20.7 and 26.96 (mean = 24.47 ± 1.39) NH₃-N kg d⁻¹.

There is little standardization of the units used to report NH₃ emission rates across various studies. The data from this study were compared with other studies in the units in which those studies published their data. The scaled NH₃ emission rates in the PMCF were high compared with studies investigating emissions from AFOs. Hristov [12] presented a compilation of data from multiple dairy farms and beef feed lots (n = 29), where all but one study found NH₃ emissions between 0.01 and 0.9 g h⁻¹ m⁻², compared with 2.77 ± 0.41 g h⁻¹ m⁻² NH₃-N in this study. Daily NH₃ emissions from a composting facility at a manure belt poultry house were estimated to be much greater, at 596 kg d⁻¹ [7], approximately 20 times higher than in this study. However, Zhao [7] also relayed data from conference proceedings presented by Matsusada [38], reporting NH₃ emissions of 0.32–2.84 g kg⁻¹, 0.02–2.74 g kg⁻¹, and 0.04–0.46 g kg⁻¹ for composting of swine, poultry, and dairy manure, respectively. These figures are comparable to the NH₃ emissions of 1.52 ± 0.09 g kg⁻¹ from the mixed manure at the PMCF. A separate study of cattle manure composting reported NH₃ emissions of 0.07–0.19 g d⁻¹ kg⁻¹ during the first 50 days [39], again similar to the figure of 0.084 ± 0.005 g d⁻¹ kg⁻¹ for this study. As expected, the NH₃ emission rate at the PMCF was more similar to measurements from other manure composting operations than those from AFOs.

Internal PM_2.5 concentrations were also measured during the ventilation-off experiment and were found to initially rise from approximately 3–10 µg m⁻³ under ventilated conditions to 18 µg m⁻³ when the ventilation system was disengaged, before decreasing again to ~5 µg m⁻³ during the experiment. At the end of the experiment, when the ventilation system was re-engaged, PM_2.5 concentrations peaked at 65 µg m⁻³. It was expected that the high concentrations of NH₃ and SO₂ during the ventilation-off experiment would lead to secondary inorganic particulate formation and higher PM_2.5 concentrations. The presence of PM_2.5 concentration peaks at the beginning and end of the experiment, but not during it when NH₃ and SO₂ concentrations were highest, suggests that the processes affecting PM_2.5 formation were more complex than the co-existence of precursor gases alone. This could be a productive area for future research.

4. Conclusions

The concentrations of SNA precursor gases and PM_2.5 at a modern manure composting facility in Paju, South Korea were continuously measured over a period of 35 days in June and July 2021. Average internal NH₃ concentrations were 57.1 ± 1.3 mg m⁻³, compared with external NH₃ concentrations of 3.64 ± 0.06 mg m⁻³ at 3 m height and 2.43 ± 0.16 mg m⁻³ at ground level. Internal concentrations of NH₃ were found to differ significantly between daytime (higher) and nighttime (lower), and weekend nights had concentrations that were significantly lower than weekday nights. These concentration differences are likely to have been caused by manure agitation activity, which occurred during weekday daytimes, and temperature differences between day and night. Sulfur dioxide (SO₂) concentrations followed a similar pattern to NH₃, and SO₂ was found to be the second most abundant precursor gas, with average concentrations of 307.2 ± 17.4 µg m⁻³. Weekend nighttime concentrations of SO₂ were found to be significantly lower than at all other times. The predominant component of NO_x gases was determined to be NO, with average concentrations of 33.4 ± 3.2 µg m⁻³, compared with NO₂ concentrations of 7.3 ± 0.8 µg m⁻³. Although there were no significant differences between NO_x concentrations at different times, the NO_x concentration profile was markedly different to both NH₃ and SO₂, suggesting that the processes governing emissions were different. Average PM_2.5 concentrations inside the facility (36.9 ± 2.6 µg m⁻³) were about 50% higher than external concentrations (23.7 ± 2 µg m⁻³), and a moderate correlation (r = 0.341, p < 0.001) between the two suggested that ambient PM_2.5 contributed part of the internally measured concentrations. The NH₃ emission rate from the facility’s manure agitation hall was calculated to be 7.28 ± 0.41 NH₃-N mg min⁻¹ m⁻³, corresponding to scaled emissions rates of 1.85 ± 0.1 NH₃-N g h⁻¹ m⁻³ for manure compost and 1.52 ± 0.09 NH₃-N g kg⁻¹ for manure compost over the full 18-day composting period. Based on these figures, total daily NH₃ flux from the manure agitation hall was approximately 24.47 ± 1.39 NH₃-N kg d⁻¹. The data from this study suggest that despite the large fluxes of NH₃ and other precursor gases emitted inside the manure composting facility, the sealed building design with a negative-pressure high-volume ventilation/scrubber system is effective at minimizing NH₃ and PM_2.5 emissions to the atmosphere.