Effect of bedding application and air change rates on environmental ammonia concentrations for intensively housed beef cattle

Introduction

Ammonia (NH₃) is an irritant gas that can be produced from manure during livestock operations, including livestock transport. Human and animal health can be compromised when exposed to high air NH₃ concentrations on a regular basis. The critical value of atmospheric NH₃ when transporting cattle, sheep and goats has been set as the human level of time-weighted average exposure tolerance of 25 ppm (17.4 mg/m³) (Costa et al. 2003). However, on the basis of immune responses (as indicated by white blood cell counts and mononuclear cell counts), it has been suggested that the maximum safe level of NH₃ exposure for cattle in live export vessels is 20 ppm (13.9 mg/m³) (Tudor et al. 2003; McCarthy and Banhazi 2016). In addition, a study on the physiology and behaviour of sheep showed that sheep exposed to 20.9 mg/m³ (range 17.2–23.7 mg/m³) NH₃ displayed discomfort when chewing because of buccal cavity irritation and increased breathing rates (Zhang et al. 2017). A study was conducted in 2007 to simulate a 12-day ship voyage to assess cattle physiological and behavioural response to air NH₃ concentration. It showed that 45 ppm (31.3 mg/m³) NH₃ increased macrophage activity, lacrimation, and nasal secretion (Phillips et al. 2010).

Ammonia volatilises from manure pads (accumulated animal manure) as a result of nitrogen (N) loss as urinary-N, which contains 69% urea (Bristow et al. 1992). Urinary urea-N is converted to NH₃ by microbial decomposition of organic N compounds and urease enzymes in the faeces (Dai and Karring 2014; Liu et al. 2017). In confined cattle housing such as in livestock export ships, NH₃ volatilisation is multifactorial and can be influenced by humidity and air temperatures levels (Ngwabie et al. 2011; Pines and Phillips 2011) and also can be affected by bedding (Banney et al. 2009), pad pH, stocking density, and air turnovers (i.e. air change rate per hour or ACH) (Costa et al. 2003).

Bedding on livestock export vessels is mainly utilised as an absorption agent for animal faeces and urine (McCarthy and Banhazi 2016). Bedding may also improve pad condition by reducing the penetrable depth and moisture of the pad (Wilkes et al. 2022) and support animal comfort as measured by lying time and number of lying bouts throughout the day (Tucker et al. 2009). Current Australian Standards for the Export of Livestock ver. 3.3 (ASEL 3.3) (Department of Agriculture, Water and the Environment 2023) require that on long-haul voyages (>10 days), bedding, such as sawdust, rice hulls or similar materials be applied at a rate of 7 tonnes (7000 kg) for every 1000 m² of cattle pen space. The Australian livestock export industry commonly utilises wood shavings and sawdust (McCarthy and Banhazi 2016) that contain lower moisture of ~30% (Banney et al. 2009) and lower odour activity value than do other materials such as corn cobs, shredded paper, or wood chips (O’Keefe et al. 2013). Even though the bedding application standard does not apply to cattle voyages leaving ports from Brisbane or north of latitude 26°S to Southeast Asia, some exporters carrying heavy cattle (over 380 kg) on short-haul voyages (<10 days) use bedding to minimise lameness incidence (Banney et al. 2009). Moreover, the use of bedding potentially mitigates the humidity issue and minimises risks of slips after washing out (MAMIC 2001) as Australian Maritime Safety Authority (2022) states ‘AMSA recommends that deck washing is conducted at intervals not more than 4 days when carrying cattle and/or buffalo.’

Air ventilation systems can affect the various gas concentrations within the environment (such as NH₃) by controlling air changes in animal housing (Ye et al. 2009). Applying higher ventilation rates by increasing air changes per hour to lower the NH₃ volatilisation (Kim and Kim 2020) may reduce the risk of disease caused by high NH₃ concentrations during sea transport (MAMIC 2001; McCarthy 2005). On livestock export vessels, the ratio of air supply to each specific pen volume is expressed as air changes per hour (ACH), which is the ventilation term used in Australian Maritime Safety Authority (AMSA) Marine Order Part 43 (MO43) regulations. According to AMSA, 20 ACH is required for decks with a ceiling height of 2.3 m or more and 30 ACH for decks with a ceiling height of 1.8 m.

In the current version of ASEL (3.3), recommendations for bedding application rates (BAR) still need a strong empirical study basis and do not include any contingency for how the bedding rate may affect the amount of NH₃ volatilised and how ventilation rates might interact in this process under specific BAR. This study aimed to assess the effects of (and interactions between) BAR and ACH on the concentration of air NH₃ and pad NH₄⁺ in simulated livestock voyages, with a view to inform refinement of ASEL 3.3 requirements. The study also monitored pad surface condition and investigated pad properties (pH, moisture, and bulk density) and associations between these properties and the concentration of air NH₃. Moreover, cattle standing and lying preferences by bedding and air change-rate treatments were also assessed.

Materials and methods

The experiment was undertaken at the Centre for Animal Research and Teaching, UNE, Armidale, NSW, Australia (30°28′57.8″S, 151°38′16.2″E). The conduct of the experiment was approved by the University of New England (UNE) Animal Ethics Committee under the Australian Code for the Care and Use of Animals for Scientific Purposes, 2013 (AEC Approval # 20-058).

Respiration chambers

Ten chambers were arranged in two rows of five, each with a single full-width entry and exit door opening onto a standard aisle between the two rows. The chambers were sealed in a recessed water-filled trench and could be raised by a pneumatic system to permit hosing out of animal waste at the end of each experiment. The outside chamber dimension was 3.6 m × 2.4 m × 2.4 m (chamber volume of 20.736 m³), and within each chamber, cattle were housed in a smaller internal steel pen structure (3.0 m × 1.8 m × 1.8 m) (Fig. 1).

Fig. 1.

Respiration chamber layout from the upper view, front view and side view and NH₃ measuring height spots (top-2.3 m from the pad (A), middle-1.8 m from the pad (B), bottom-0.3 m from the pad (C)).

Dry-bulb temperature (T_DB) and relative humidity (RH) were recorded twice daily at 07:30 hours (AM) and 14:30 hours (PM) by using an equipped SHT20 I2C temperature and humidity sensor (Core Electronic, Adamstown, NSW, Australia) located at steer head height within each chamber. In total, there were 702 observations of T_DB and RH during the trial and the means for T_DB and RH are shown in Table 2. The mean values of T_DB for PM were 7.1°C higher than those for AM. Conversely, the mean RH for AM was greater than the mean for PM. Wet-bulb temperature (T_WB) was calculated from AM and PM T_DB and RH (Table 2). The mean AM and PM T_DB and RH were equivalent, with the mean of temperature–humidity index (THI) of 62.6 and 71.3 respectively (Table 2).

Table 2.AM and PM ambient dry bulb temperatures, wet bulb temperature, temperature–humidity index and relative humidity in the respiration chambers during six 7-day experimental periods.

Variable and time point (number of observations)	Mean ± s.d.	Minimum	Maximum
TDB (°C)
AM (n = 378)	17.4 ± 2.6	11.2	23.4
PM (n = 324)	24.5 ± 2.7	18.2	30.5
TWB (°C)
AM (n = 378)	9.7 ± 1.13	7.05	12.3
PM (n = 324)	12.8 ± 1.16	10.1	15.3
THI
AM (n = 378)	62.6 ± 4.04	53.1	72.2
PM (n = 324)	71.3 ± 3.36	62.8	78.8
RH (%)
AM (n = 378)	73.9 ± 4.2	56.1	84.7
PM (n = 324)	53.0 ± 8.6	34.0	77.1

T_DB, dry-bulb temperature; T_WB, wet-bulb temperature; THI, Temperature–humidity index; RH, relative humidity.

The air was delivered into the chambers by 150-mm-diameter industrial tubing with a rotating baffle that could be closed or opened to increase or decrease the air supplied to the desired air change rate (per hour, ACH). An oscillating fan (diameter of 40 cm) located within the chamber, 2 m above the floor (Fig. 1), continuously mixed the internal air of the chamber. This was to ensure that the air was well-mixed prior, so the air exiting the chamber was representative of that throughout the chamber.

Video footage was recorded using fixed infra-red cameras (MR6822E2 and LR832, Lilin Australia, Lidcombe, NSW, Australia) positioned centrally above each chamber pen for the entire duration of each 7-day experimental period to monitor animal condition and behaviour, particularly standing and lying. The standing and lying frequency data were collected using an instantaneous-scan sampling method every 15 min for Days 1, 3 and 6. Time (h) to when each steer first lay down were measured from 20 min after the steers entered the chamber and the chamber door was closed (Time 0).

Pad properties

Daily pad-surface conditions

Pad-surface conditions were recorded every morning (AM) at the same time as air NH₃ measurement by using four variables. Each of these variables was an ordinal score (three variables) or a nominal categorical variable, where scores or categories were allocated on the basis of visual assessment. This scoring system for pad-surface conditions was developed to assess cattle pads (Supplementary material 1: Daily pad surface conditions scores), because the existing scale was for sheep pads, and these are substantially different in moisture content and other physical and visual characteristics (McCarthy and Banhazi 2016). The data were collected and recorded in the KoBo Collect application (www.kobotoolbox.org) by using scores for each category. Data were collected for four variables: moisture appearance (score 1–8; from dry to flooded respectively), area covered by moisture [score 1–3: no wet areas (0%), small wet patches (<10%), and large wet patches (10% or more)], the assumed moisture source (high humidity, water-trough spill, water-line leak, cattle playing with water, or urine), when moisture was detected in the pad (moisture-area score of >1) and estimated pad depth [score 1–6; from no pad (cm) to >20 cm depth respectively].

Pad sampling and analysis on Day 7

Pad sampling was performed on the final day of each experimental period (Day 7), because it was not feasible to conduct it before the steers were discharged from the chambers because of safety reasons, to minimise the disturbance for the steers that can lead to animal stress and to maximise the mass balance of total pad collection. Pad depth on the chamber floor was measured at 25 locations as shown in Fig. 2a. The pad was then considered as being divided into Quarters 1, 2, 3 and 4 (Fig. 2b). Within each quarter, the pad was homogenised, and five grab samples were collected to represent each quarter (2 kg per quarter in total). The pad in each quarter, including the grab samples, was weighed individually and summed to obtain total pad mass (kg) for each chamber. The grab samples from each quarter were split into subsamples and used to measure pH (pH-ORP-Temp. Meter, pH Cube, TPS Pty Ltd, Brendale, Qld, Australia) and then the remainder (250 g) was oven-dried at 105°C for 24 h or until the weight was stable for gravimetric moisture content (AS 1289.2.1.1 method) (Standards Australia 2005). Bulk density was measured using a modified compaction test (AS 1289.5.2.1 method) (Standards Australia 2017) on subsamples stored at 4°C.

Fig. 2.

Pen set up on Day 7 for 25 spot samples assigned as x, y coordinates (cm) for (a) measuring pad depth and (b) pad sampling location divided into four quarters.

The frozen (−20°C) sample was analysed for pad ammoniacal N (NH₃ (aq) and NH₄⁺), which in this study is reported as NH₄⁺ concentration via colorimetric determination through a modified Berthelot reaction (Creevey 2022). Each 1 g subsample was placed in a 50 mL polypropylene centrifuge tube and mixed with 20 mL of 2 M KCl. The mixture was tumbled for 4 h at 15 rpm, refrigerated overnight at 4°C, and then centrifuged at 3000g for 20 min (temperature 20°C) (Allegra X-15R centrifuge, Beckman Coulter USA). The supernatant was transferred to new tubes and stored at −20°C. For the assay, samples were diluted 1:10 with 2 M KCl, added to a 96-well plate in duplicate, and mixed with a working reagent. The absorbance at 650 nm was read after 30 min of incubation at room temperature by using a SPECTROstar Nano plate reader (BMG Labtech, Ortenberg, Germany), and the results were adjusted for blanks before performing a linear regression to determine NH₄⁺ concentrations. The total pad NH₄⁺ mass (g/chamber) was then calculated by multiplying the NH₄⁺ concentration (mg/g pad) by the total pad mass (kg) in the chamber.

A small portion of the pad was outside of the floor surface area accessible by the steers (Fig. 2). This pad portion was also collected for analyses of pad mass and pad NH₄⁺ concentration. This outside portion of the pad was not included in the samples for each chamber quarter.

Results

Air NH₃ concentrations by day and measuring height

Over 2106 observations [six experimental periods × (7 days AM + 6 days PM) × nine chambers × three heights], NH₃ concentration (mean ± s.d.) was 3.75 ± 1.66 mg/m³. Air NH₃ concentration increased from Day 1 to Day 3, then reached a plateau, with a minimal interaction between day number and measuring height (Fig. 3a). Concentrations were higher in AM than in the PM, with differences of similar magnitude on all days except Day 4 (Fig. 3b). Crude mean air NH₃ concentration at the middle height was higher than the means at bottom and top heights, with the latter two being similar (Table 4).

Fig. 3.

Predicted means for air NH₃ concentration by study day number (a) at three measuring heights (top, middle and bottom, 0.3, 1.8, 2.3 m above the pad, respectively) and (b) in the morning (AM) and in the afternoon (PM), with 95% confidence intervals for each predicted mean; bedding application rate (BAR), air changes per hour (ACH), measurement height and time of day (AM or PM) were included as fixed effects, as categorical variables, in both models, along with interaction between day number and measurement height (a) or time of day (b); overall P-values for interaction terms were 0.447 and <0.001 respectively.

Table 4.Air ammonia (NH₃) concentrations (mg/m³) at three measuring heights in the respiration chambers (n = 702 observations at each measurement height).

Air NH3 measuring height	Air NH3 concentration (mg/m3)
	Mean ± s.d.	Minimum	Maximum
Top (2.3 m above the pad)	3.47 ± 1.444	0.70	9.76
Middle (1.8 m above the pad)	4.34 ± 1.888	0.70	12.54
Bottom (0.3 m above the pad)	3.43 ± 1.459	0.70	12.54

Pad-surface conditions during 7 days of the experimental period

Pad moisture appearance and area covered by moisture over the 7 days of the experiment were each associated with BAR and/or ACH (Table 6). Chambers with 100% and 50% bedding rates and high and medium ACH had higher scores for moisture appearance (where higher scores indicate firmer pads) than did the chambers with no bedding treatment (P < 0.001 and 0.002 respectively) and low ACH (P < 0.001 and 0,002 respectively) where pads were more often sloppier and wetter. Relative to 0% bedding, 0% or <10% of the area as wet patches was more likely for chambers with 100% (P < 0.001) and 50% (P = 0.041) bedding treatments (and 10% or more of area as wet patches less likely). Similarly, increasing ACH reduced the odds of larger moisture-covered areas (Table 6). The presumed sources of moisture were mainly cattle urine and water-trough spills, including when the cattle played with the water (Table S2). Increasing bedding to 100% ASEL resulted in the deeper pad than did 0% ASEL (P < 0.001), whereas the changes of ACH appeared unrelated to the estimated pad depth.

Table 6.Pad-surface condition by bedding application rate (BAR) and air changes per hour (ACH) on Day 7.

Pad-surface condition	BAR (% ASEL)			ACH (number per hour)
	0	50	100	20 (low)	35 (medium)	52 (high)
Moisture appearance
Odds ratio	Reference category	5.22	12.2	Reference category	5.68	6.22
95% CI		1.80–15.15	4.59–32.2		1.921–16.8	2.42–15.9
P-value		0.002	<0.001		0.002	<0.001
Area covered with moisture
Odds ratio	Reference category	0.37	0.10	Reference category	0.35	0.38
95% CI		0.14–0.96	0.031–0.33		0.12–1.06	0.14–0.99
P-value		0.041	<0.001		0.066	0.050
Pad-depth range
Odds ratio	Reference category	1.39	2.37	Reference category	1.35	1.09
95% CI		0.74–2.59	0.031–0.33		0.69–2.62	0.53–2.62
P-value		0.294	0.018		0.821	0.384

Pad depth on Day 7

Pad depth for the 0% bedding treatment (mean ± s.d.; 3.9 ± 3.30 cm) was lower than 50% (6.1 ± 4.62 cm) and 100% (6.7 ± 4.5 cm). For all bedding treatments, most of the pad was pushed to the side of the quarters, resulting in greater depth than in the centre areas of the quarters (Fig. 4). The small portion of pad not accessible by the steer was mostly dry wood shavings (for the 50% and 100% bedding) and some manure. These outside portions of pads weighed, on average, 115.61 ± 29.30 kg (mean ± s.d.), whereas the pads inside the quarters weighed, on average, 88.94 ± 35.46 kg. The concentration of NH₄⁺ in the pad quarters (predicted mean ± s.e.m 1.11 ± 0.04 mg/g pad) was 0.42 mg/g pad higher (95% CI: 0.33–0.51; P < 0.001) than that in the outside portions of the pad (0.70 ± 0.04 mg/g pad). The total pad NH₄⁺ mass (mean ± s.d., grams per chamber) in the quarters was 100.13 ± 47.78, whereas that in the outside was 80.42 ± 39.86.

Fig. 4.

Fitted values for pad depth on the chamber floor left to right (x-coordinate as defined in Fig. 1) and back to front (y-coordinate) pooled across bedding application rates; grey band indicates 95% confidence intervals for fitted values.

Pad pH, moisture, bulk density, and NH₃ concentration on Day 7

Both pad pH and pad moisture were lower for the 100% bedding treatment (P < 0.001) than for the 50% and 0% bedding (Tables 7, S3). Predicted means for pad bulk density were similar for all BAR. Mean concentration of pad NH₄⁺ (mg/g pad) in the 0% bedding treatment was higher than in the 50% and 100% bedding treatments (predicted means 1.15, 1.01 and 0.95 respectively). However, predicted means for total pad NH₄⁺ content (g/chamber) were similar for all BARs. Pad moisture was lower for the high and medium ACH than for the low ACH. No effects of ACH on other pad properties were evident.

Table 7.Pad properties (predicted mean ± s.e.m) by bedding application rate (BAR) and air changes per hour (ACH) on Day 7.

Pad property	BAR (% ASEL)			ACH (number per hour)			Overall P-value
	0	50	100	20 (low)	35 (medium)	52 (high)	Bedding rate (B)	ACH (A)	B × A
Pad pH	7.92 ± 0.37a	7.69 ± 0.37b	7.59 ± 0.37b	7.69 ± 0.37	7.79 ± 0.37	7.72 ± 0.37	<0.001	0.188	0.376
Pad moisture (%)	80.34 ± 0.59a	78.66 ± 0.59a,b	75.34 ± 0.59b	79.34 ± 0.59a	77.42 ± 0.59b	77.58 ± 0.59b	<0.001	0.049	0.822
Pad bulk density (kg/m3)	990.0 ± 26.50	1002.3 ± 26.50	1040.7 ± 26.50	997.2 ± 28.40	1002.1 ± 28.40	1033.8 ± 28.40	0.335	0.651	0.702
Pad NH4 + concentration (mg/g pad), wet basis	1.18 ± 0.041a	1.01 ± 0.0.041b	0.95 ± 0.041b	1.04 ± 0.041	1.05 ± 0.041	1.06 ± 0.041	<0.001	0.941	0.707
Total pad NH4 + mass (g/chamber), wet basis	89.6 ± 8.21	101.2 ± 8.21	91.0 ± 8.21	100.4 ± 8.21	87.27 ± 8.21	94.8 ± 8.21	0.415	0.602	0.899

Means within a row and within a category followed by the same letter or no letter do not differ statistically significantly (at P < 0.05).

Our results showed that any correlations between pad moisture and each of pad pH (r = 0.099; 95% CI: −0.091 to 0.283; P = 0.304) and pad NH₄⁺ concentration (r = 0.051; 95% CI: −0.140 to 0.237; P = 0.604) are weak (Table S4). Pad moisture and pad bulk density were negatively correlated (r = −0.383; 95% CI: 0.020–0.382; P = 0.031), and pad moisture and total pad NH₄⁺ mass were positively correlated (r = 0.228; 95% CI: 0.040–0.399; P = 0.018). Pad NH₄⁺ concentration and total pad NH₄⁺ mass were quite closely positively correlated (r = 0.612; 95% CI: 0.478–0.718; P < 0.001).

Effects of BAR on lying behaviour

Duration of time to when each steer first commenced lying on Day 1 for all runs was shorter for chambers with more bedding. Steers first laid down at 1.14 ± 0.701 h from 20 min after the steers were fed their first meal after entering the chamber and the chamber door was closed (Time 0) for 100% bedding and at 2.06 ± 1.134 h for 50% bedding, compared with steers with no bedding that laid down for the first time at 3.89 ± 1.452 h (P < 0.001 for each of 100% and 50% bedding relative to 0% bedding).

Steers were lying most often throughout the 24-h time period on each of the three sampling days for all runs for the 100% bedding treatment (Fig. 5). On Day 1, odds of lying were 4.5 times higher (95% CI 3.4–5.9) on 100% bedding and 2.1 times higher (95% CI 1.5–2.9) on 50% bedding than in 0% BAR treatment. Effects of bedding were less on Days 3 and 6, but odds of lying were still higher where higher BAR was applied (Day 3 odds-ratio estimates 1.5 (95% CI 1.3–1.8) and 1.3 (95% CI 1.1–1.5) for 100% and 50% bedding respectively; Day 6 odds-ratio estimates 1.3 (95% CI 1.1–1.5) and 1.2 (95% CI 1.0–1.5) for 100% and 50% bedding respectively).

Fig. 5.

Steers standing and lying frequency (%) in the chambers for 0%, 50% and 100% Australian Standards for the Export of Livestock (ASEL) bedding application rate on Days 1, 3 and 6.

Effects of ACH on lying behaviour were inconsistent among days (results not shown). The steers were more likely to be lying at any particular time point during the night-time period than daytime with odds ratios for lying on Day 1 (odds ratio for night-time relative to daytime 3.3; 95% CI 2.6–4.0) and similar on Days 3 and 6.

Discussion

In general, air NH₃ concentration in the chambers increased for each sequential day of the experimental run following the accumulation of deposited manure (urine and faeces) on the floor, most likely because these were sources of urea, and urease enzyme, which resulted in the formation of NH₄⁺ that was volatilised as NH₃ gas. Of the 702 air NH₃ concentrations measured in the chambers during the study, only 4% were greater than 10 mg/m³, with a maximum concentration of 12.5 mg/m³. This concentration was lower than that reported on ships that could reach 17.07 mg/m³ (Zhang et al. 2017) and even up to 41.1 mg/m³ (Pines and Phillips 2011) and was likely due to the moderate chamber T_DB values, ranging from 17°C to 24°C. According to Adegoke and Ku (2023), urease enzyme activity increased exponentially at temperatures greater than 25°C releasing NH₃ in the gaseous phase. Given that our T_DB remained in the lower end of this range, it is likely to have led to slower urease activity, and, consequently, lower air NH₃ concentrations. These findings suggested that the lower temperatures in the current study might inhibit the full potential of NH₃ volatilisation, which would likely increase under the higher temperatures typical of live-export conditions. The relationship between T_DB and the rate of volatilisation of NH₃ has been previously demonstrated by Sommer and Olesen (1991) and Sommer et al. (1991), which showed that the rate of loss of NH₃ in the cattle slurry through volatilisation was related to T_DB increase.

In the current study, the air NH₃ concentration was lower in the morning, and more concentrated in the afternoons. In the afternoon, when the T_DB increased, the RH in the chamber decreased, which was probably related to a high water-vapour deficit (VPD) that can cause drier air (Runkle 2021). It is likely that this condition contributed to more rapid moisture evaporation from the manure pad, leaving less liquid available to retain NH₃ as NH₄⁺, and potentially increase NH₃ volatilisation from the pad surface. This result was consistent with Whitehead and Raistrick (1991) finding, which noted that greater urea hydrolysis and nitrification were noticeable when the air was warmer and drier.

In the present study, we found no evidence of an interaction between BAR and ACH in their effects on air NH₃ concentration, suggesting that the relationship between these variables might be independent under the given experimental conditions. It appeared that increasing BAR and adjusting it with high ACH was not mutually simulating the reduction of air NH₃ concentration as was hypothesised, indicating that each variable may utilise its impacts separately. This condition might be explained by the different mechanism of action between BAR and ACH in influencing air NH₃. Bedding primarily affected the initial source of air NH₃ by influencing the concentration of NH₄⁺ on the pad level. In contrast, ACH affected the removal of NH₃ gas once it volatilised from the pad surface to the higher measuring point. However, the results showed that the total pad NH₄⁺ mass per chamber was not significantly different for each BAR treatment. Therefore, under this similar condition, the concentration of air NH₃ at any measuring height was dependent mostly on the amount of air flow in the chamber.

When considering individual variables, there was a relationship between ACH and air NH₃ concentration in the chamber at each measuring height, but no marked effects for pad pH, bulk density, and pad NH₃. Contrastingly, air NH₃ concentration was similar across BAR treatments for all measuring heights; however, there were differences in pad pH, pad moisture, and pad NH₃ concentration related to the BAR treatments.

ACH was strongly associated with air NH₃ concentration (mg/m³) of all BAR treatments and at any measuring height. Lower air NH₃ concentrations were observed in the chambers with medium (35) and high (52) ACH than with low ACH (20) for all measuring heights. Meanwhile, air NH₃ concentrations were similar at high and medium ACH. Air change rate for the high ACH was 1.6 times greater than that for the low ACH. Therefore, under steady-state conditions, if volatilisation rates were unchanged and NH₃ concentration in the air was the same throughout the chamber (so that NH₃ concentration in exiting air was the same as that at any point in the chamber), air NH₃ concentrations at 52 ACH would be expected to be 38% of those at 20 ACH. In fact, the observed mean NH₃ concentrations at 52 ACH at top, middle and bottom heights respectively, were 76%, 68% and 72% of those at 20 ACH. This difference was potentially a result of the higher volatilisation rates at the higher ACH (Kim and Kim 2020; Kumar et al. 2022; Kamp et al. 2024), supported by the calculated total NH₃ (using the mean concentration of NH₃ at the middle height), which showed that the high ACH produced 1.7 times more NH₃ than did the low ACH. Furthermore, although the deposition rate of N by the cattle would be similar for both treatments and there was no observable difference in measured pad NH₄⁺ concentration among the different ACH treatments, the low ACH produced 0.283 kg less air NH₃ over the 7 days, than did the high ACH. To determine why this may be the case, first consideration was that the urease enzyme hydrolyses urea into carbonic acid (H₂CO₃) and NH₃, which remains in equilibrium with bicarbonate (HCO₃−) and ammonium (NH₄⁺) (Dai and Karring 2014). Given that the pad pH in this study was ~7.8, most N would be expected to exist in the NH₄⁺ form, but the Berthelot method measures ammoniacal N and does not appear to distinguish between NH₃ and NH₄⁺ (Creevey 2022; Alwael et al. 2024). One plausible explanation for less air NH₃ being produced under low ACH is a slower rate of urea hydrolysis, leaving more N in the form of urea-N, which had not yet undergone hydrolysis. However, further experimental investigation would be necessary to confirm this hypothesis. There may also be additional factors beyond pad pH and NH₄⁺ concentration, such as airflow dynamics, that need to be considered.

In the current study, pad moisture was lower for the high and medium ACH than the low ACH. However, even though the oscillating fans mixed the air to create a homogenous downward airflow, with empty chambers, the measured velocity of the air at pad surface was 0 and we assume that it would be the same if two animals were in each of these small spaces. According to Li (2023), when the ventilation-duct system is positioned in the upper room space of a confined space, the supply air runs down along vertical walls or columns to the floor, then shifts horizontally, forming a thin, evenly spread layer across the floor, moving with low velocity, and it is possible that that horizontal velocity was higher at higher air change rates.

Another possibility is that the rate of removal of NH₃ from the chamber does not increase in proportion to the increase in air change rate, for example, because of differences in the vertical profile of NH₃ concentrations from top to bottom. However, our results did not support this because the ratios of means for top and middle heights relative to the bottom level were similar within each air change rate. In summary, these results indicated that even though the rate of ACH increases the volatilisation of NH₃ from the pad surface, the higher ACH still reduces air NH₃ concentration. The negative association between ACH and air NH₃ concentration supports the results of previous studies, which demonstrated that NH₃ concentrations are influenced by air movement (Sheffield et al. 2007; Saha et al. 2010). This finding in the current study suggests that increasing the ACH rate in the cattle pen lowers the concentration of air NH₃. Additionally, the risk of disease on board might be reduced because the concentration of air NH₃ is lower at an increased ACH rate (MAMIC 2001).

Even though high and medium ACH reduced the air NH₃ more than did the low ACH, we found no evidence that ACH influences pad NH₄⁺ concentrations. The concentration of NH₄⁺ in the pad is influenced more by the chemical and biological processes within the bedding and manure, such as urea hydrolysis by urease-producing bacteria. Air changes less directly affect these processes, owing to airflow mainly being in contact with the pad’s surface rather than effectively penetrating the pad layer to influence these reactions. Increasing the airflow rate in the chamber over the pad surface increased the pad surface conditions, which were drier, firmer, and deeper under high and medium ACH than under low ACH. These results showed that the rate of ACH was associated with pad moisture. Applying higher ACH rates and bedding (50% and 100% ASEL) in the chamber improved the pad condition, with the pad being drier, denser, and less moist.

The opposite was found for effects of BAR treatments on air NH₃ and pad NH₄⁺. The pads were a mixture of manure (faeces + urine) and spilled drinking water for no-bedding treatment, and manure, spilled water and wood shavings for 50% and 100% ASEL bedding treatments. When bedding was applied to the chamber, the pad was built and formed deeper and denser layers. The bedding material in the pad sample has reduced pad pH, moisture, and pad NH₄⁺ concentration (mg/g pad) for higher bedding-rate treatments relative to 0% bedding, but bulk density and total pad NH₄⁺ mass (g/chamber) appeared similar for all bedding treatments. The higher concentration of NH₄⁺ per gram of pad in the no-bedding treatment was due to the pad consisting entirely of manure containing abundant organic matter and nitrogenous compounds, including ionised NH₄⁺. The absence of bedding material in this treatment to mix with or dilute the manure presumably resulted in a more concentrated N content and a lower pH. The differences in pad NH₄⁺ concentration between no bedding, 50% and 100% bedding were modest, probably owing to the slight pH range, such that BAR had no marked effect on air NH₃ concentration. Nevertheless, these results indicated that a higher bedding application rate increased the ability of bedding to absorb manure and other fluid waste and keep the pad in low pH and drier condition.

Another reason for the lack of influence of BAR on air NH₃ concentration could be attributed to the narrow range of pad pH across bedding treatments. Even though lower pH values and higher moisture values than with dry bedding (pH 8 and 91% DM) were observed in the pad on Day 7. The small variation in pad pH among BAR treatments, ranging from 7.5 to 7.9 when the bedding became saturated with manure, was insufficient to result in significant differences in pad NH₄⁺ or air NH₃ concentration. Additionally, the increase in pad moisture after 7 days (75–80% moisture), particularly for the 50% and 100% BAR treatments, was a consequence of the bedding absorbing urine and water. In this study, no correlation was evident for pad pH and pad moisture directly affecting the air NH₃ concentration. This result was in contrast with the previous study that showed pad pH and DM content influence NH₃ volatilisation (Sommer and Olesen 1991; Olesen and Sommer 1993).

This study showed no evidence that air NH₃ concentration changes by BAR treatments, which could be due to pad distribution and animal activities in the chamber. In the current study, the surface area was the same in all chambers (3.97 m²). Thus, the size of the area contaminated by the excreta was also similar. This result is supported by Webb et al. (2005), who stated that when estimating NH₃ emissions, the surface area of the facilities should be considered.

Fig. 5 shows that the pad formation in the chamber was deeper on the side of the quarter than in the centre. The pad was spread on the side of the chamber rather than the centre because of the steer activities (standing, lying and locomoting) within the central areas of the chambers. This increased animal activity in the centre of the pen area pushed the bedding and some of the manure to the side, resulting in a deeper pad depth and lower NH₄⁺ concentrations than in the centre sampling points. As a result, more pad NH₄⁺ was trapped in the undisturbed bedding layers on the side of the pen than in the centre, which the steers frequently disturbed. During the measurement of air NH₃ concentrations, the sensor was lowered into the centre of the chambers (Fig. 1). Thus, if there was minimal lateral air movement, the recorded air NH₃ would mainly have reflected concentrations in air that had escaped from the pad layer in the centre of the chamber floor instead of from the sides. As a result, the total measured air NH₃ concentrations would be expected to have been more influenced by the pad NH₄⁺ concentration (mg/g) than of the total NH₄⁺ mass in the pad. The relationships between pad NH₄⁺ concentration (mg/g pad) and air NH₃ concentration (mg/m³) in the chamber are shown in Table 8, where the air NH₃ in the chamber increased with the concentration of NH₄⁺ in the pad.

The activities of the steers in the chambers may have contributed to the differences in air NH₃ concentrations among measuring heights, where the highest concentration was at the middle height. The study of Bleizgys and Naujokienė (2023) showed that NH₃ emission from the pad surface decreased as a result of the decrease of temperature differences between the manure layer and surface and moisture evaporation. In the current study, the air NH₃ at the bottom height was potentially lower owing to the prevention of volatilisation of NH₃ from the top layer of the pad when the steers were lying down, covering the pad surface. However, when air NH₃ was measured in the chambers, the steers would stand up and disturb the pad, leading to the potential flux of NH₃ from the pad surface. Because NH₃ gas is lighter than air, the volatilisation of air NH₃ would be expected to rise to the next vertical measuring spot, middle height, where the concentration of air NH₃ was found to be the highest. At this height, the animals may have created a ‘temporary blanket effect’, with their bodies trapping NH₃ between the pad surface and the animals’ underbelly/frame, concentrating the gas in this zone before reaching the top height once the steers had moved. This may also explain why not all air NH₃ reached the top height. These patterns by height were consistent for every BAR and ACH treatment combination.

Air NH₃ was generally observed at higher concentrations at the middle height, which was approximately head height for a standing human or steer. However, most measured concentrations of air NH₃ in this study were below 25 ppm or 17 mg/m³, which is the level recommended for the livestock export industry in Australia (MAMIC 2001; Phillips et al. 2010). Therefore, the exposure of the air NH₃ in this study did not exceed the critical exposure limit to be considered toxic for humans and animals. Nevertheless, it is essential to note that these values were measured under moderate temperature conditions and may not fully represent typical conditions on live-export ships. Consequently, although the study provides valuable data, caution is needed when applying this model in actual livestock-export scenarios under warmer conditions.

When bedding was applied in the chamber, the pad was deeper and denser BAR affected the physical characteristics of the pad. The results demonstrated that pad with 100% BAR was significantly deeper than that with 50% or 0% BAR. The deeper pad was due not only to the amount of bedding applied but also to the ability of bedding particles to trap moisture and maintain pad weight. The negative correlation between pad moisture and pad bulk density suggested that as the moisture increased, the bulk density decreased. When bedding was still dry or less moist, presumably the bulk density was lower because of air space between bedding particles. In this study, because the bedding was mixed with manure, the addition of moisture in the pad generated pore-water pressures, causing the pad to be less compacted (Etim Udom and Ehilegbu 2018) and the bond between soil particles was weakened (Salman and Kiss 2018). The pad condition appeared to deteriorate as the study progressed, indicating that the ability of bedding to absorb manure was reduced over time, and additional bedding or new bedding applications could be required (Cowley et al. 2019).

Although BAR treatments showed no evidence of affecting air NH₃ concentration, bedding increased frequency of lying in this study. First, the preference of the steers to lie on the bedding compared with no-bedding treatment was shown by the shorter time it took for steers to commence lying on Day 1 than in the no-bedding (0%) chambers. From an animal-welfare perspective, being able to provide a surface that increased lying and subsequent resting behaviour may reduce the animals’ stress, fatigue, and potential for injury. Second, the drier pad conditions and low pad moisture, as a result of the application of bedding, resulted in the steers lying at more observation timepoints (indicating that they spent more time lying) when more bedding was applied (50% and 100% ASEL) than under the no-bedding treatment (not including feeding and sampling times). Presumably, the steers spent more time standing with no bedding to potentially avoid the excreta that was covering the pen floor. However, on Day 7 of observation in the experiment, most dry bedding was pushed to the side of all chambers and could not be utilised by the animals (Fig. 4). For each consecutive day, the pad was thinner, and the absorbing ability was reduced; thus, a wetter pad was formed. Despite this, the odds of steers lying on Days 3 and 6 were still higher where bedding was provided (50% and 100% ASEL); however, the reduced strengths of association between BAR and lying frequency later in the study period were presumably due to increased pad moisture over time. These results support those of the previous studies in that cattle prefer to lie down on clean, dry, and uncontaminated bedding (Crafter et al. 2006; Tucker et al. 2009; O’Connor et al. 2019; Tait et al. 2023).

Conclusions

The findings of this study emphasise the vital role of bedding application rates (BAR) on minimising potential pad NH₄⁺ and air change rates (ACH) in reducing air NH₃ concentration within a confined setting of live-export context. Air NH₃ concentrations in the current study were mostly lower than the concentrations previously reported on livestock-export ships (Pines and Phillips 2011; Zhang et al. 2017) and we suggest that this was due to the modest ambient temperature and RH in the chambers. The lowest air change rate (ACH) in this study aligned with AMSA’s minimum requirements. Under conditions such as those in this experiment, increasing ACH to more than double the standard effectively reduced air NH₃ concentration by 0.75–1.88 mg/m³ (95% CI; P < 0.001). Additionally, the bedding applications used in this study reduced pad pH, pad moisture and pad NH₄⁺ concentration and improved overall pad conditions.

Results from this study highlight the importance of strategic bedding management for those livestock-export voyages that require bedding application. Daily assessment of the pad physical conditions can inform potential washout times and need for provision of new bedding where the existing bedding has deteriorated or when the air NH₃ concentrations have risen to levels that could be harmful to both human and animals as a potential effect of pad NH₄⁺ accumulation. Future work should aim to identify the effects of BAR and ACH under other climatic conditions, including higher T_DB and RH. The results of this study will assist in predicting air NH₃ and pad NH₄⁺ concentrations in simulated livestock-export conditions and help inform optimal BAR and ACH to help improve cattle welfare when exported from Australian ports.

For practical applications, this suggests that to minimise air NH₃ concentrations in confined animal housing, focusing on optimising ACH might be more effective than adjusting BAR. However, BAR should not be neglected because it plays a crucial role in pad management, which could have other welfare and environmental implications. Although BAR may not directly affect air NH₃, it remains important for managing other pad-related factors, which can influence animal comfort and welfare, as well as long-term environmental conditions within the housing system. Proper bedding management can contribute to a healthier environment by controlling moisture, pad consistency and by potentially reducing NH₃ volatilisation at the source.

Supplementary material

Supplementary material is available online.