Concentrations and Emissions of Ammonia from Different Laying Hen Production Systems of Conventional Cage, Aviary and Natural Mating Colony Cage in North China Plain

Ammonia (NH3) concentrations in summer were continuously monitored from three typical laying hen houses of CC (conventional cage), AV (aviary), and NM (natural mating colony cage) with manure belt systems in North China Plain to quantify their emission levels, to characterize the diurnal variations, and to investigate the impact of environmental factors. Diurnal profiles were acquired by hourly measurements, and the effect of environmental factors on NH3 emissions was presented by correlation analysis. The results showed that house-level NH3 emissions in summer were the highest in the NM at 27.16 ± 13.12 mg/h·hen, followed by the AV at 4.08 ± 3.23 mg/h·hen and the CC at 3.43 ± 1.46 mg/h·hen within a complete manure removal cycle, which were significantly affected by manure accumulation inside the houses. After manure removal, NH3 concentrations were reduced by 64.29%, 28.57%, and 35.71% in CC, AV, and NM, and consequently their emissions were lowered by 67.12%, 71.36%, and 55.69%, respectively. It was suggested that the manure should not be stored on the belt for more than 4 days in NM. A positive impact of indoor and outdoor temperature and ventilation rate on NH3 emissions from AV and NM were found, while indoor and outdoor relative humidity had a negative effect. However, the above five factors did not significantly affect the emissions from CC.


Introduction
Livestock production is one of the major sources of NH 3 to the atmosphere, which is estimated to contribute 34-40% of global NH 3 emissions [1]. NH 3 can irritate the eyes and respiratory membranes, and develop chronic stress, which is a threat to workers and animals and has been classified as a hazardous substance under EPCRA [2]. Besides, NH 3 can be a precursor of secondary particulate matter [3,4] and is also the key substance and indicator of malodor from livestock facilities [5]. Consequently, NH 3 emission is considered as an issue of political and scientific concerns because of its serious influence on health, environment, and ecosystem [6,7].
Current research has revealed that about 20% of global NH 3 came from China [8], which was mainly from livestock activities [9]. China is the No.1 layer producer in the world, with a share of Appl. Sci. 2020, 10, 6820 3 of 12

Description of the Monitoring Sites
The studied sites included a CC, an AV, and a NM house located in the NCP. The basic information of the three laying hen houses can be founded in Table 1.

Sampling and Measurement Methods
Gas concentrations, temperature, and RH were continuously measured in and out of CC, AV, and NM houses from 20 September 2019 to 24 September 2019, 7 June 2020 to 12 June 2020, and 10 July 2019 to 31 July 2019, respectively. NH 3 concentration was monitored by using a portable monitoring unit (PMU) configured with an NH 3 sensor (ME3-NH 3 , with a measurement range of 0-50 ppm; Winsen, China, every 20 min. CO 2 concentration was detected in the AV at an interval of 20 min, with a CO 2 sensor (MH-Z14, with a measurement range of 0-5000 ppm, accuracy ±100 ppm; Winsen, China) installed in the PMU. Two PMUs were set in CC and NM at the height of 1.5 m, which were fixed close to the air inlet and central exhaust fan. In the AV, three PMUs were located at a height of 1.5 m, were close to the air inlet and cage walls of two aviary cages near the exhaust fans. Temperature and RH were measured by T/RH data loggers (U23-001, HOBO, Bourne, USA), with a measurement ranges of −40 • C to 70 • C and 0-100%, and accuracies of ±0.18 • C and ±2.5%, respectively. Three T/RH data loggers were located in the center of the three houses at the height of 1.5 m, and another three loggers were set outside, which were 2.0 m away from the head of laying hen houses. Temperature and RH were continuously monitored by the data loggers.
Before the experiment, the air velocity of each fan for CC and NM was measured by a hot-wire anemometer (model KA41 L, Kanomax, Osaka, Japan). When measuring, the fan was divided into 4 equal parts by cross-marking, and 2 of them were tested randomly. Each part was evenly distributed with 16 sampling points ( Figure 1). The probe of an anemometer was attached to the protective net in front of the fan to measure the air velocity of every sampling point. When the air velocity at a sampling point was comparatively stable, 20 values were recorded continuously by an anemometer, and the average of all sampling velocities was used as the average air velocity through the fan.
Appl. Sci. 2020, 10, 6820 4 of 12 equal parts by cross-marking, and 2 of them were tested randomly. Each part was evenly distributed with 16 sampling points ( Figure 1). The probe of an anemometer was attached to the protective net in front of the fan to measure the air velocity of every sampling point. When the air velocity at a sampling point was comparatively stable, 20 values were recorded continuously by an anemometer, and the average of all sampling velocities was used as the average air velocity through the fan.

Ventilation Rate and Emission Rate Calculation
For the CC and NM houses equipped with exhaust fans, the ventilation rate (VR) was calculated as follows: where VR was the ventilation rate of the house, m 3 /h; V i was the average air velocity of the fan i, m/h; S i was the area of the fan i, m 2 ; n was the number of the operating exhaust fans. For the AV house, its VR was estimated by a CO 2 mass balance method recommended by CIGR [25], and the equations were listed below: where (CO 2 ) P was CO 2 production per heat production unit based on a 24-h period, 0.185 m 3 /h; (CO 2 ) i and (CO 2 ) o were CO 2 concentrations inside and outside of the tested house, mg/m 3 ; ø tot was the total heat production of the birds under 20 • C, W; a was a constant to express the amplitude with respect to the constant 1, 0.61; h was the hours after midnight; h min was the hours after midnight with minimum activity, −0.1; T i was the temperature inside the house, • C. The NH 3 emission rate was calculated as follows: where ER was NH 3 emission rate, mg/h; ∆C was NH 3 concentration difference between the air outlets (at the exhaust fan end) and inlets of the tested laying hen houses, mg/m 3 .

Statistical Analysis
To study the diurnal NH 3 concentrations and emissions, the data at the same stage of the manure removal cycle was pooled together to calculate the mean value and analyzed by least significant differences using repeated measurement analyses of variance (ANOVAs) with a significance level of 0.05 in SPSS 21.0 (IBM, New York, NY, USA). When comparing the difference of diurnal NH 3 concentration and emission rate during a complete manure removal cycle, the average of the whole day data was used. The correlation impact of temperature and VR on NH 3 emission was developed by SPSS 21.0, where p ≤ 0.05 means significant correlation, and p ≤ 0.01 means a very significant correlation.

Bird Performance and Diurnal Air Temperature and RH
Age, feed intake, and crude protein (CP) of laying hens in different houses during the experiment were listed in Table 2. Feed intake in AV was the highest because of the higher energy requirement with more activities. The CP content in the three houses was similar.

CC and AV
The diurnal patterns of average VR, NH3 concentrations, and emissions in CC and AV, were plotted in Figure 3. The VR ranged from 1.21 to 5.15 m 3 /h·hen for CC, and 0.84 to 12.16 m 3 /h·hen for AV. In CC and AV, VR presented a variation pattern following that of outdoor temperature, which showed a peak within 9:00 to 20:00 in both days of a complete manure removal cycle, because of the high temperature, and it did not change greatly at the time of manure removal (11:00-12:00, and

CC and AV
The diurnal patterns of average VR, NH 3 concentrations, and emissions in CC and AV, were plotted in Figure 3. The VR ranged from 1.21 to 5.15 m 3 /h·hen for CC, and 0.84 to 12.16 m 3 /h·hen for AV. In CC and AV, VR presented a variation pattern following that of outdoor temperature, which showed a peak within 9:00 to 20:00 in both days of a complete manure removal cycle, because of the high temperature, and it did not change greatly at the time of manure removal (11:00-12:00, and 16:00-17:00).
Appl. Sci. 2020, 10, x FOR PEER REVIEW 6 of 12 It could be seen that NH3 concentrations kept stable on the second day after manure removal in AV. Before 10:00, the VR increased gradually as the indoor temperature increasing to emit the NH3 out of the house to keep a stable indoor NH3 concentration. The phenomenon occurred again after 20:00. As known, the hens were more active comparatively during the daytime, while, the indoor temperature, VR, and NH3 concentration in AV remained fairly constant during 12:00-20:00 on the second day, which illustrated that animal activity did not have a significant effect on NH3 concentration after separating the manure from hens kept inside the aviary units since the manure was the main NH3 source [28,29]. NH3 concentrations in AV ranged from 0.54 to 1.30 mg/m 3 (0.77 ± 0.19 mg/m 3 in average), which were significantly lower than that in summer (about 2.12 mg/m 3 ) in a similar type of house with lower VR (maximal VR of 0.75 m 3 /h·hen) and longer manure accumulation (3 to 4 days) inside the operation that Shepherd et al. [27] studied. On the day after manure removal in CC and AV, NH3 emission rates were higher in the daytime and lower in the nighttime, which agreed with Wang-Li et al. [16] and Alberdi et al. [30]. It was likely influenced by the house temperature and VR. High temperatures can accelerate microbiological degradation of manure to promote bigger generation rates of NH3 [16]. Liang et al. [14] also observed  Figure 3(a2,b2) depicted the temporal variations in NH 3 concentrations during manure removal (day 1) and the day after manure removal (day 2). On both days, concentrations of NH 3 in CC were higher during the night, due to cooler temperature and lower VR, which has been supported by many studies with the same production system [12,18,26]. NH 3 concentrations in the CC ranged from 0.63 to 3.79 mg/m 3 (2.01 ± 0.93 mg/m 3 in average), which were comparable with the average NH 3 concentration in similar facilities with the ambient temperature over 25 • C and stock density of 516 cm 2 /hen in the US midwest (near to 38 • N) [27,28].
It could be seen that NH 3 concentrations kept stable on the second day after manure removal in AV. Before 10:00, the VR increased gradually as the indoor temperature increasing to emit the NH 3 out of the house to keep a stable indoor NH 3 concentration. The phenomenon occurred again after 20:00. As known, the hens were more active comparatively during the daytime, while, the indoor temperature, VR, and NH 3 concentration in AV remained fairly constant during 12:00-20:00 on the second day, which illustrated that animal activity did not have a significant effect on NH 3 concentration after separating the manure from hens kept inside the aviary units since the manure was the main NH 3 source [28,29]. NH 3 concentrations in AV ranged from 0.54 to 1.30 mg/m 3 (0.77 ± 0.19 mg/m 3 in average), which were significantly lower than that in summer (about 2.12 mg/m 3 ) in a similar type of house with lower VR (maximal VR of 0.75 m 3 /h·hen) and longer manure accumulation (3 to 4 days) inside the operation that Shepherd et al. [27] studied.
On the day after manure removal in CC and AV, NH 3 emission rates were higher in the daytime and lower in the nighttime, which agreed with Wang-Li et al. [16] and Alberdi et al. [30]. It was likely influenced by the house temperature and VR. High temperatures can accelerate microbiological degradation of manure to promote bigger generation rates of NH 3 [16]. Liang et al. [14] also observed that VR had a significant effect on NH 3 emissions and caused higher emission rates during the daytime. After manure removal, NH 3 emission rates in both houses sharply dropped down. The NH 3 emission rates within a whole manure removal cycle ranged from 0.92 to 6.52 mg/h·hen (3.43 ± 1.46 mg/h·hen in average) in CC, which was comparable with reported values of 4.04 and 2.83 mg/h·hen in laying hen houses with CC and similar management in the U.S. Midwest [27] and 3.60 mg/h·hen for a traditional cage with manure belt-natural drying system with a 492 cm 2 /hen stocking density in the laboratory [31]. The average NH 3 emissions from laying hen houses in California (36.46 • N) with a stocking density of 332 cm 2 /hen [2] and North Carolina (35.26 • N) with a 400 cm 2 /hen density [16] in the U.S. were 39.58 mg/h·hen, and 45.00 mg/h·hen, respectively, which showed similar operational information and environmental variables having similar NH 3 emission characteristics. NH 3 emission rates in AV averaged at 4.08 ± 3.23 mg/h·hen, ranging from 0.66 to 17.50 mg/h·hen, which was comparatively lower than the value reported by Shepherd et al. [27]. They documented an average rate in a similar facility with the house temperature of 20-25 • C was 6.29 mg/h·hen. Meanwhile, the VR (maximal VR of 0.75 m 3 /h·hen) was lower and the manure accumulation (3 to 4 days) was longer than this study.

NM
The diurnal average VR in NM ranged from 11.70 to 20.80 m 3 /h·hen (Figure 4b), which was significantly higher than that in CC and AV (p < 0.05). The daily VR change trend in NM was similar to CC, and the peak of VR occurred from 9:00 to 18:00 because of the high temperature. It was seen that the NH 3 concentration gradually increased with the manure accumulation extending (Figure 4a). When the manure was accumulated for 1-3 days, the change in NH 3 concentration was contrary to that of VR, and the higher concentrations of NH 3 in NM occurred during the night, which was similar to the situation of CC. While it changed reversely from the fourth day, NH 3 concentration did not decrease with the increasing VR. In this case, after the manure accumulated on the belt for 4 days, NH 3 volatilized from manure with the temperature increasing was more than that exhausted by fans. As shown in Figure 5, the NH 3 concentration on day 7 was 3 times higher than that on the second day after manure removal. Mendes et al. [32] reported NH 3 concentration at air outlet with manure accumulated for 6 days was about 21 times higher than that for 2 days storage in a laboratory with the hen age of 23 to 36 weeks. Thus, manure accumulation significantly affected NH 3 emission rates. It was recommended that the manure should not be stored on the belt for more than 4 days in NM, which was consistent with the findings of Mendes et al. [32]. Otherwise, higher VR was demanded to maintain a good inner environment. NH 3 concentrations in NM ranged from 0.93 to 5.65 mg/m 3 with an average of 3.06 ± 1.29 mg/m 3 , which was substantially higher than those in CC and AV (p <0.05) mainly due to much longer manure accumulation time inside the house.
by the Ministry of Agriculture and Rural Affairs of the People's Republic of China [35]. Furthermore, it was found that NH3 concentrations and emission rates were sharply dropped down after manure removal. In CC, AV, and NM, NH3 concentrations were reduced by 64.29%, 28.57%, and 35.71% after manure clearance, and consequently, their emissions were lowered by 67.12%, 71.36%, and 55.69%, respectively. It is suggested that the manure belt could efficiently reduce NH3 concentrations and the emissions as well by thoroughly and timely removing manure outside of the laying hen houses, which was also supported by Huang and Guo [18] and Nicholson et al. [36].

Influence of Environmental Factors on NH3 Emission Rate
The correlation coefficients for the NH3 emission rate in summer against environmental factors are listed in Table 3. It was found that temperature, RH, and VR did not significantly correlate NH3 NH 3 emission rates in NM gradually increased with manure accumulation, which agreed with the report by Ning [33]. It also varied with time, similar to that of CC and AV, in which relatively higher NH 3 emission rates occurred during the day and lower emission rates occurred at night [16]. NH 3 emission rates in NM ranged from 1.48 to 78.93 mg/h·hen (27.16 ± 13.12 mg/h·hen in average), which was similar to the value of 20.83 mg/h·hen reported by Hayes et al. [34]. The NH 3 emission rates in NM were much higher than previously reported value (2.68 mg/h·hen in average) from a caged-hen house with a 620 cm 2 /hen stocking density in the laboratory, inside which the manure was accumulated for 6 days [32]. In Mendes et al. [32], the minimum VR of laying hen house was regulated based on the maximum allowed indoor CO 2 concentration level of 4000 ppm v . The average VR inside NM was 16.03 m 3 /h·hen, highest among the three studied laying hen houses, which had a significant and positive impact on NH 3 emissions [14,16], and it was also mentioned in CC and AV houses. Besides, differences also occurred in manure belt or manure pan cleanliness between the two studies. In this study, there was inevitably some manure residue on the belt even after manure removal, which contributed to higher NH 3 emissions. In Mendes et al. [32], all the facilities were cleaned and the existing manure pans were replaced with new ones after each trial. NH 3 concentrations in CC, AV, and NM with manure belt systems were well controlled, and all were under the recommended threshold of 15 mg/m 3 to maintain a good environmental quality given by the Ministry of Agriculture and Rural Affairs of the People's Republic of China [35]. Furthermore, it was found that NH 3 concentrations and emission rates were sharply dropped down after manure removal. In CC, AV, and NM, NH 3 concentrations were reduced by 64.29%, 28.57%, and 35.71% after manure clearance, and consequently, their emissions were lowered by 67.12%, 71.36%, and 55.69%, respectively. It is suggested that the manure belt could efficiently reduce NH 3 concentrations and the emissions as well by thoroughly and timely removing manure outside of the laying hen houses, which was also supported by Huang and Guo [18] and Nicholson et al. [36].

Influence of Environmental Factors on NH 3 Emission Rate
The correlation coefficients for the NH 3 emission rate in summer against environmental factors are listed in Table 3. It was found that temperature, RH, and VR did not significantly correlate NH 3 emission rate in CC (p > 0.05), which might attribute to the indoor temperature and RH were controlled well within narrow ranges, and NH 3 emission rates in CC during the day changed slightly (Figure 3(a3)). This agreed with the results of Huang and Guo [18]. Outdoor and indoor temperatures and VR had significantly positive correlations with the emissions in AV and NM (p < 0.01). A significant and negative effect of outdoor and indoor RH on NH 3 emission rate was also revealed for AV (Outdoor RH: p < 0.01, indoor RH: p < 0.05) and NM (p < 0.01), which aligned with the findings of Huang and Guo [18]. While it was contrasted to the results reported by Wang-Li et al. [16], who found the RH had significant positive impact on NH 3 emissions. In Wang-Li et al. [16], the correlations between NH 3 emission rates and RH were analyzed using the data from 24 September 2007 to 31 December 2009, making no distinction between seasons. High RH would increase NH 3 emissions by promoting manure fermentation [37], especially in cold seasons with low VR, which might cause the difference between the two studies. Besides, the manure dropped directly into the manure pit, and was cleaned out annually to make the moisture in manure volatilized difficultly. Table 4 illustrated the average NH 3 emission rates on the first and second days of manure accumulation in CC, AV, and NM. On these two days of manure accumulation, the NH3 emission rate in NM was greatly higher than that in CC and AV (p < 0.05). The RH of NM was the highest compared with that of CC and AV, which could cause higher moisture content in the manure of NM to improve microbial activity to produce more NH 3 [27,37]. Besides, this discrepancy might also be caused by the different levels of manure residue on the belts. Additionally, the highest VR inside NM would also promote higher NH 3 emissions.