Application of a hand-held laser methane detector for measuring enteric methane emissions from cattle in intensive farming

Abstract

The hand-held laser methane detector (LMD) technique has been suggested as an alternative method for measuring methane (CH₄) emissions from enteric fermentation of ruminants in the field. This study aimed to establish a standard procedure for using LMD to assess CH₄ production in cattle and evaluate the efficacy of the protocol to detect differences in CH₄ emissions from cattle fed with diets of different forage-to-concentrate (FC) ratios. Experiment 1 was conducted with four Hanwoo steers (584 ± 57.4 kg body weight [BW]) individually housed in metabolic cages. The LMD was installed on a tripod aimed at the animal’s nostril, and the CH₄ concentration in the exhaled gas was measured for 6 min every hour for 2 consecutive days. For the data processing, the CH₄ concentration peaks were identified by the automatic multi-scale peak detection algorithm. The peaks were then separated into those from respiration and eructation by fitting combinations of two of the four distribution functions (normal, log-normal, gamma, and Weibull) using the mixdist R package. In addition, the most appropriate time and number of consecutive measurements to represent the daily average CH₄ concentration were determined. In experiment 2, 30 Hanwoo growing steers (343 ± 24.6 kg BW), blocked by BW, were randomly divided into three groups. Three different diets were provided to each group: high FC ratio (35:65) with low-energy concentrate (HFC-LEC), high FC ratio with high-energy concentrate (HFC-HEC), and low FC ratio (25:75) with high-energy concentrate (LFC-HEC). After 10 d of feeding the diets, the CH₄ concentrations for all steers were measured and analyzed in duplicate according to the protocol established in experiment 1. In experiment 1, the mean correlation coefficient between the CH₄ concentration from respiration and eructation was highest when a combination of two normal distributions was assumed (r = 0.79). The most appropriate measurement times were as follows: 2 h and 1 h before, and 1 h and 2 h after morning feeding. Compared with LFC-HEC, HFC-LEC showed 49% and 57% higher CH₄ concentrations in exhaled gas from respiration and eructation (P < 0.01). In conclusion, the LMD method can be applied to evaluate differences in CH₄ emissions in cattle using the protocol established in this study.

The hand-held laser methane detector (LMD) technique has been suggested as a potential method for measuring methane (CH₄) emissions from enteric fermentation of ruminants in the field. The LMD method can be applied to evaluate differences in CH₄ emissions in cattle efficiently using the protocol established in this study.

Introduction

Methane (CH₄), which has 25 times more global warming potential than carbon dioxide (IPCC, 2014), is the primary greenhouse gas that needs to be reduced in animal agriculture. Ruminants, in particular, produce a large amount of CH₄ via rumen fermentation, and CH₄ emissions from the enteric fermentation of ruminants account for 17% of global anthropogenic CH₄ emissions (Knapp et al., 2014). Enteric CH₄ production also contributes to a loss of 2–12% of gross energy intake of ruminants (Johnson and Johnson, 1995). Therefore, reducing enteric CH₄ emissions is of great interest in the field of cattle production (Beauchemin et al., 2008). To reduce enteric CH₄ emissions, quantification of CH₄ emissions is essential. Although various methods have been proposed for measuring enteric CH₄ emission in ruminants (Garnsworthy et al., 2019), the use of mobile or portable devices is critical for quantifying CH₄ emissions and evaluating the effectiveness of mitigation strategies in field situations. The use of a hand-held laser methane detector (LMD) has been proposed as a potential method for on-farm quantification of enteric CH₄ emission from ruminants (Chagunda et al., 2009).

The LMD uses a semiconductor laser based on infrared absorption spectroscopy and measures CH₄ concentrations using the second harmonic detection of a wavelength-modulated spectrometer. Using LMD, CH₄ concentration (in ppm-m) in the exhaled gas from the nostril of an individual animal is measured non-invasively and in real-time, without gas sampling; the use of LMD could, therefore, potentially reduce labor, time, and costs associated with measuring CH₄ emissions from multiple sites. Since its first proposition, several studies have demonstrated the feasibility of using LMD to quantify CH₄ emission from ruminants. For example, high correlation coefficients (0.8 [Chagunda and Yan, 2011] and 0.86–0.98 [Sorg et al., 2017]) were observed between the CH₄ measurements obtained from LMD use and the respiration chamber method. A relatively strong repeated-measures correlation (r = 0.66) was also observed between the measurements taken by an LMD and a GreenFeed (Sorg et al., 2018). Despite its advantages, the repeatability of LMD measurements was relatively low (0.14–0.27; Sorg et al., 2018), and standardized protocols for measurement and data analysis of LMD measurements have not been developed (Sorg, 2022).

The time, duration, and frequency of LMD use in a day and the number of days of LMD measurement vary by study. In dairy cows, measurements were typically obtained after feeding or milking (Ricci et al., 2014; Cameron et al., 2018; Rey et al., 2019). In studies of beef cattle, LMD measurements have been conducted before or after feeding, or sparsely according to an animal’s activity, such as standing, ruminating, and lying (Ricci et al., 2014; Denninger et al., 2020). In other studies, authors did not specify the criteria for choosing the time of the day to take measurements (Mapfumo et al., 2018; Roessler et al., 2018; Niero et al., 2020). In addition, across studies, the duration per measurement has varied from 1 min (Grobler et al., 2014) to 6 min (Denninger et al., 2020), and the frequency of measurement has varied from one to three (Roessler and Schlecht, 2021) per day over mostly three consecutive days (Sorg, 2022). Data processing, which is critical for analyzing and interpreting LMD data, has also varied across studies. Enteric-fermented CH₄ in ruminants is emitted through the following pathways: respiration and eructation (Murray et al., 1976). Owing to the cyclic nature of respiration and eructation, as well as equipment fluctuations, there are peaks and troughs in the LMD data. The peaks are commonly assumed as all the points whose values are greater than the preceding and following points, and the means of the peaks are used for subsequent analyses (Chagunda et al., 2009; Ricci et al., 2014; Sorg et al., 2018). Recent studies have divided the peaks by respiration and eructation; however, the threshold to differentiate between the two events has also varied across studies. For instance, one study defined a threshold based on two normal distributions after logarithmic transformation (Ricci et al., 2014), another classified the eructation peaks by calculating outliers (Sorg et al., 2018), and a third separated datasets by fitting a double normal distribution (Pinto et al., 2020; Reintke et al., 2020; Kobayashi et al., 2021). Owing to the periodic characteristics of respiration and eructation, and diurnal fluctuations of CH₄ concentration in the exhaled gas, the procedures for processing LMD data significantly affect the values measured by the LMD method. Thus, a standardized protocol must be established for the LMD method to be more broadly applicable.

The objective of this study was to establish a standard procedure (i.e., for measurement and data processing) for the use of an LMD to measure CH₄ production in cattle. Two trials were conducted for this study. Experiment 1 was conducted to determine the most appropriate time of day to take measurements and the least number of time-series measurements required to represent daily averages of CH₄ concentration in the gas exhaled through respiration and eructation. In experiment 2, we applied the protocol established in experiment 1 and evaluated its efficacy in detecting the differences in CH₄ emissions from cattle that had been fed with diets of different forage-to-concentrate (FC) ratios.

Materials and Methods

Experiment 1

Experiment 1 was conducted from June 2020 to July 2020 at the National Institute of Animal Science, Korea. The use of animals and protocols for this experiment were reviewed and pre-approved by the National Institute of Animal Science Ethics Committee (NIAS 2020-499).

Four 35-month-old Hanwoo steers (584 ± 57.4 kg body weight [BW]), individually housed in metabolic cages, were used in this trial. A mixture of oat hay and concentrate mix was offered, limited to 1% of BW, at 0700 and 1500 hours. The steers were fed diets with an FC ratio of either 30:70 or 15:85. The feed ingredients of the concentrate mix and chemical compositions of the forage and concentrate mix are described in Tables 1 and 2, respectively.

Table 1.

Diet composition (g/kg as fed) of the concentrate mix in experiment 1

Ingredients	Concentrate mix
Wheat grain	140
Corn grain	105
Corn DDGS†	150
Corn gluten feed	140
Soybean hull	115
Alfalfa pellet	40
Rice bran	25
Soybean meal	67
Palm kernel meal	130
Molasses	50
Limestone	30
Salt	4
Others	4

Corn distillers dried grain with soluble.

Table 2.

Analyzed chemical composition (g/kg DM or as stated) of the diets in experiment 1

Items†	Concentrate mix	Oat hay
DM, g/kg as fed	899	910
OM	926	951
CP	196	71
SOLP	74	32
NDICP	21	6
ADICP	6	6
aNDF	257	621
ADF	98	381
DL	28	63
Ether extract	29	13
sh	74	49
Ca	16	3
P	6	1
K	12	11
Na	4	5
Cl	5	7
S	3	1
Mg	735	572
TDN	899	910
ME, MJ/kg DM	2.82	2.09
NEm, MJ/kg DM	1.90	1.23
NEg, MJ/kg DM	1.26	0.68
Total carbohydrates	701	867
NFC	466	252
Carbohydrate fraction, g/kg carbohydrate
CA	67	136
CB1	511	16
CB2	87	138
CB3	242	536
CC	95	173
Protein fraction, g/kg CP
PA+B1	378	451
PB2	516	465
PB3	76	4
PC	30	80

DM, dry matter; OM, organic matter; CP, crude protein; SOLP, soluble CP; NDICP, neutral detergent-insoluble CP; ADICP, acid detergent-insoluble CP; aNDF, neutral detergent fiber analyzed using a heat-stable amylase and expressed inclusive of residual ash; ADF, acid detergent fiber; ADL, acid detergent lignin; TDN, total digestible nutrients; NEm, net energy for maintenance; NEg, net energy for growth; NFC, non-fiber carbohydrate; 2CA, carbohydrate A fraction, ethanol-soluble carbohydrates; CB1, carbohydrate B1 fraction; starch; CB2, carbohydrate B2 fraction, soluble fiber; CB3, carbohydrate B3 fraction, available insoluble fiber; CC, carbohydrate C fraction, unavailable carbohydrate; 3PA+B1, protein A and B1 fractions, soluble CP; PB2, protein B2 fraction, intermediate degradable CP; PB3, protein B3 fraction, slowly degradable fiber-bound CP; PC, protein C fraction, unavailable CP.

The CH₄ concentrations in the exhaled gas of the steers were measured after 19 d of feed adaptation. Two trained operators performed the measurements every hour for 2 consecutive days, alternating at 12-h intervals. A Laser Methane Mini (Tokyo Gas Engineering Solutions Co. Ltd., Tokyo, Japan) having a visible laser (Class 3 R laser, 532 nm) for guiding an invisible laser (Class 1 laser, 1653 nm) to aim at the desired site was used in the present study. The LMD was stably installed on a tripod, aiming the visible laser at the beginning of one of the animal’s nostrils from a distance of 1.2 m. The operator continuously adjusted the LDM on the tripod to point the visible laser on the same spot, following the animal’s head movement. CH₄ concentrations were measured every 0.5 s for 6 min. Eructation, the primary pathway of CH₄ emission, coincides with the B-sequence of rumen contractions, which occur irregularly once every 1–3 min (Waghorn and Reid, 1983). Thus, the measurement duration was set to 6 min, which was long enough to capture two to three eructations. This was similar to the suggested duration (minimum 2 min) to keep an animal in the GreenFeed system to ensure the inclusion of at least one eructation event (Velazco et al., 2014).

LMD data processing

The CH₄ concentration measured by LMD shows fluctuations owing to the periodic nature of respiration (inhalation and exhalation), eructation, and equipment fluctuations. Because the CH₄ concentration in the exhaled gas better represents CH₄ emission from enteric fermentation (Chagunda et al., 2009), peak detection is essential for data analysis. Unlike previous approaches (Ricci et al., 2014; Sorg et al., 2018), which identified the peaks simply based on the difference between the preceding and following values, we used the automatic multi-scale peak detection (AMPD) algorithm for peak detection. The AMPD algorithm effectively detects the peaks of noisy periodic and quasi-periodic signals by calculating and analyzing the local maxima scalogram, which linearly detrends the signal to determine the local maxima using a moving-window approach (Scholkmann et al., 2012). The peaks of CH₄ concentration were identified by detecting the oscillations of CH₄ concentration due to periodic respiration and eructation using the AMPD package in the R software, version 4.0.2 (R Core Team, 2020).

For each 6-min measurement, the distribution of the peaks from the two pathways (respiration and eructation) was separately identified by the mixdist R package, which allows fitting a mixture of distributions by maximum likelihood using a combination of a Newton-type algorithm and the Expectation-Maximization algorithm (Macdonald and Du, 2018). A total of 16 combinations of two of the four distribution functions (i.e., normal, log-normal, gamma, and Weibull) were tested to determine the optimal mixture of distributions to separate and fit the peaks from respiration and eructation. After identifying the distributions, the expected value of each distribution was assumed to be the CH₄ concentration in the pathway. A smaller number of peaks belonged to eructation, but they were of higher value. The correlation coefficient between the two pathways was calculated for all combinations assuming that the CH₄ concentration in the exhaled gas from respiration is correlated with those from eructation.

Additionally, an analysis was conducted to determine the appropriate time of day to take measurements and the least number of time-series measurements required to represent the 24-h averages of CH₄ concentration in exhaled gas through respiration and eructation. The estimated means of the CH₄ concentration peaks at each hour were averaged over 24 h (from 1100 to 1000 hours on the following day) to determine the 24-h mean, for respiration and eructation separately. The means from two to five subsequent hours were calculated (i.e., 2 h, 3 h, 4 h, and 5 h means) and compared with the 24-h mean. For both respiration and eructation, the square root of the sum of squares of differences (RSSD) between the 24-h mean and the composite mean of CH₄ concentration over 2 d was calculated for each hour as follows:

$${\displaystyle \begin{array}{ll}{\displaystyle \mathrm{R}\mathrm{S}\mathrm{S}{\mathrm{D}}_{i\times t}=}& \sqrt{{({{\displaystyle \overline{\left[\mathrm{C}{\mathrm{H}}_4\right]}}}_{\mathrm{d}\mathrm{a}\mathrm{y}1}-{{\displaystyle \overline{\left[\mathrm{C}{\mathrm{H}}_4\right]}}}_{\mathrm{i}\times \mathrm{t}\times \mathrm{d}\mathrm{a}\mathrm{y}1})}^2+}\\ & {({{\displaystyle \overline{\left[\mathrm{C}{\mathrm{H}}_4\right]}}}_{\mathrm{d}\mathrm{a}\mathrm{y}2}-{{\displaystyle \overline{\left[\mathrm{C}{\mathrm{H}}_4\right]}}}_{i\times t\times \mathrm{d}\mathrm{a}\mathrm{y}2})}^2\end{array}}$$

where $\mathrm{R}\mathrm{S}\mathrm{S}{\mathrm{D}}_{i\cdot t}$ is the square root of the sum of squares of differences between the 24-h mean and the composite mean of i subsequent hours (i = 2, 3, 4, and 5) of CH₄ concentration (in ppm) at t h, ${\overline{\left[C{H}_4\right]}}_{day1}$ is the 24-h mean of CH₄ concentrations measured using the LMD on day 1 (and from 1100 to 1000 hours on the following day), ${\overline{\left[\mathrm{C}{\mathrm{H}}_4\right]}}_{i\cdot t\cdot \mathrm{d}\mathrm{a}\mathrm{y}1}$ is the average CH₄ concentration measured using the LMD for i subsequent hours (i = 2, 3, 4, and 5) at t h on day 1, ${\overline{\left[\mathrm{C}{\mathrm{H}}_4\right]}}_{\mathrm{d}\mathrm{a}\mathrm{y}2}$ is the daily mean CH₄ concentration measured using the LMD on day 2 (and from 1100 to 1000 hours on the following day), and ${\overline{\left[\mathrm{C}{\mathrm{H}}_4\right]}}_{i\cdot t\cdot \mathrm{d}\mathrm{a}\mathrm{y}2}$ is the average CH₄ concentration measured using the LMD for i subsequent hours (i = 2, 3, 4, and 5) at t h on day 2.

Experiment 2

This study was conducted at the Center for Animal Science Research, Chungnam National University, Republic of Korea. The use of animals and protocols for this experiment were reviewed and pre-approved by the Chungnam National University (CNU) Animal Research Ethics Committee (202103A-CNU-027).

Thirty 10-month-old Hanwoo steers (343 ± 24.6 kg BW) were used in this experiment. Steers, grouped according to initial BW and estimated breeding values for carcass weight, were arranged in a completely randomized block design (Seo et al., 2018). Two steers of similar BW were grouped within a block and housed together in a pen (5 m × 5 m) equipped with a forage feed bin, which allowed us to measure individual feed intake automatically by identifying each animal using a radio-frequency identification tag attached to them (Dawoon, Co., Incheon, Korea). The experiment lasted for 25 days, 7 d after the adaptation period.

Timothy hay and two concentrate mixes (a low-energy concentrate mix [LEC] containing 346 g/kg neutral detergent fiber [NDF] and 564 g/kg total digestible nutrients [TDN] on a dry matter [DM] basis, and a high-energy concentrate mix [HEC] containing 232 g/kg NDF and 708 g/kg TDN on a DM basis) were used to prepare the dietary treatments in this study. There were three dietary treatments: 1) high FC ratio with low-energy concentrate mix (HFC-LEC; 35% Timothy and 65% LEC), 2) high FC ratio with high-energy concentrate mix (HFC-HEC; 35% Timothy and 65% HEC), and 3) low FC ratio with high-energy concentrate mix (LFC-HEC, 25% timothy, and 75 HEC). The diets were formulated to meet nutrient requirements according to Korean feeding standards for Hanwoo steers (NIAS, 2017). The formulation and chemical composition of the experimental diets are presented in Tables 3 and 4, respectively. Each group of steers was randomly allocated to one of the three dietary treatments and fed twice daily at 0800 and 1800 hours. Forage and drinking water were freely accessible to the animals throughout the experiment. The concentrate mix was given individually, and its amount was restricted to maintain the designed FC ratio of each treatment according to the forage intake of the steer.

Table 3.

Diet formulation of the diets in experiment 2

Item	Treatment†
	HFC-LEC	HFC-HEC	LFC-HEC
Ingredients, g/kg DM
Timothy hay	350	350	250
Wheat, ground	108	231	266
Corn, ground	78	7	8
Hydrogenated fat	0	14	16
Corn gluten feed	199	115	133
Rice bran	29	58	66
Soybean hull	7	0	0
DDGS	49	122	141
Soybean meal	0	35	41
Palm kernel meal	132	25	29
Molasses	11	11	14
CMS	11	11	14
Limestone	20	16	18
Salt	3	3	3
Vitamin and mineral mix‡	2	2	3

HFC-LEC, a high forage-to-concentrate ratio (0.35) with a low-energy concentrate mix; HFC-HEC, a high forage-to-concentrate ratio (0.35) with a high-energy concentrate mix; LFC-HEC, a low forage-to-concentrate ratio (0.25) with a high-energy concentrate mix.

33,330,000 IU/kg vitamin A, 40,000,000 IU/kg vitamin D, 20.86 IU/kg vitamin E, 20 mg/kg Cu, 90 mg/kg Mn, 100 mg/kg Zn, 250 mg/kg Fe, 0.4 mg/kg I, and 0.4 mg/kg Se.

Table 4.

Analyzed chemical composition (g/kg DM or as stated) of the diets in experiment 2

Items†	Concentrate mix		Timothy
	High energy (HEC)	Low energy (LEC)
DM, g/kg as fed	904	890	896
OM	911	913	927
CP	201	191	97
SOLP	72	68	39
NDICP	15	32	15
ADICP	7	15	10
aNDF	232	346	636
ADF	102	187	402
ADL	32	31	53
Ether extract	20	12	19
Ash	89	87	73
Ca	11	17	3
P	8	7	2
K	11	11	23
Na	3	2	0
Cl	5	6	13
S	4	4	2
Mg	4	3	2
TDN	708	665	564
ME, MJ/kg DM	11.3	10.5	8.6
NEm, MJ/kg DM	7.5	6.7	5.1
NEg, MJ/kg DM	4.8	4.2	2.7
Total carbohydrates	689	710	811
NFC	472	396	189
Carbohydrate fraction, g/kg carbohydrate
CA	55	56	74
CB1	397	235	5
CB2	232	266	154
CB3	203	338	608
CC	113	104	158
Protein fraction, g/kg CP
PA+B1	358	356	402
PB2	570	474	447
PB3	36	90	42
PC	36	80	108

DM, dry matter; OM, organic matter; CP, crude protein; SOLP, soluble CP; NDICP, neutral detergent-insoluble CP; ADICP, acid detergent-insoluble CP; aNDF, neutral detergent fiber analyzed using a heat-stable amylase and expressed inclusive of residual ash; ADF, acid detergent fiber; ADL, acid detergent lignin; TDN, total digestible nutrients; NEm, net energy for maintenance; NEg, net energy for growth; NFC, non-fiber carbohydrate; CA, carbohydrate A fraction, ethanol-soluble carbohydrates; CB1, carbohydrate B1 fraction, starch; CB2, carbohydrate B2 fraction, soluble fiber; CB3, carbohydrate B3 fraction, available insoluble fiber; CC, carbohydrate C fraction, unavailable carbohydrate; PA+B1, protein A and B1 fractions, soluble CP; PB2, protein B2 fraction, intermediate degradable CP; PB3, protein B3 fraction, slowly degradable fiber-bound CP; PC, protein C fraction, unavailable CP.

The individual daily concentrate intake was recorded by measuring the feed offered and refused. Daily feed intake was averaged for 22 d, excluding the sampling period, which might have affected accurate intake measurements. When averaging, daily intakes that were less than 0.5 kg, or greater or less than three times the standard deviation (SD) were treated as outliers and omitted. BW was measured before morning feeding at the beginning and end of the experiment.

Measuring CH₄ emissions using the LMD

After 10 d of feeding the steers the experimental diets, the CH₄ concentration in the exhaled gas of the animals was measured using the LMD according to the protocol established based on the results of experiment 1. Briefly, the LMD was installed on a tripod, aiming at the animal’s nostril from a distance of 1 m. The CH₄ concentrations were measured every 0.5 s for 6 min. Daily measurements were performed four times (−2, −1, +1, and +2 h after the morning feed). The measurements were performed for all 30 steers for 5 consecutive days and duplicated for an additional 5 d.

For data analysis, the peaks of the CH₄ concentration measured by the LMD were detected using the AMPD R package. The peaks were divided into two pathways (respiration and eructation) by fitting a double normal distribution using the mixdist R package. A larger number of peaks belonged to respiration, but they were of lower value. The mean of the normal distribution was assumed to be the representative CH₄ concentration of the exhaled gas from the pathway for the hour. The four time values of a day were averaged to represent the mean CH₄ concentration of the day.

Chemical analysis

Sampled forage, concentrate mix, and dried feces were ground using a cyclone mill (Foss, Hillerød, Denmark), fitted with a 1 mm screen, prior to chemical analysis. The nutrient composition of the samples was analyzed at Cumberland Valley Analytical Services Inc. (Hagerstown, MD, USA). The nutrient content of the samples was analyzed using the methods described by Jeon et al. (2016).

The content of TDN, net energy for maintenance, and net energy for growth were estimated using the National Research Council (NASEM, 2016) equations. Dietary carbohydrate and protein fractions were estimated according to the Cornell Net Carbohydrate and Protein System (Fox et al., 2004) with the following modifications. For the carbohydrate fractions, carbohydrate A fraction (CA) was sugars and organic acids assumed to be equal to ethanol-soluble carbohydrate, carbohydrate B1 fraction (CB1) was starch, CB2 was soluble fiber calculated as non-fiber carbohydrate (NFC)–CA–CB1, CB3 was available NDF estimated by aNDF–neutral detergent-insoluble crude protein (CP) minus 2.4 times acid detergent lignin (ADL), and carbohydrate C fraction (CC) was unavailable carbohydrate estimated by 2.4 times ADL. In the protein fractions, PA+B1 was a soluble protein that was equal to soluble CP (SOLP), PB2 (Protein B2 fraction) was intermediate degradable CP estimated by 100–NDICP–SOLP, protein B3 fraction (PB3) was slowly degradable fiber-bound CP estimated by NDICP–acid detergent-insoluble CP (ADICP), and protein C fraction (PC) was unavailable CP, which was equal to ADICP. All carbohydrate and protein fractions were expressed as grams per kilogram of total carbohydrates and CP, respectively.

Statistical analysis

All statistical analysis was performed using R software (R Core Team, 2020; version 4.0.2, R Foundation for Statistical Computing, Vienna, Austria). The correlation of the CH₄ concentrations between respiration and eructation in experiment 1 was analyzed using cor() function.

In experiment 2, data were analyzed according to a completely randomized block design using the lmerTest R package (Kuznetsova et al., 2017). The blocks (i.e., initial BW and breeding value for carcass weight) were treated as random effects. For this analysis, no structure was assumed for the variance–covariance matrix. Pairwise comparisons of the least square means were conducted using the lsmeans R package (Lenth, 2016) with the Tukey–Kramer adjustment when there was a significant overall treatment effect. Significance was declared at P < 0.05, and a trend was discussed at 0.05 ≤ P < 0.1.

Results

Experiment 1

The Pearson correlation coefficients between the mean CH₄ concentrations in the exhaled gas from respiration and eructation within the same measurement ranged from 0.41 (log-normal and log-normal) to 0.79 (normal and normal) (Fig. 1). Because the highest value was observed when a normal distribution was assumed for both pathways, subsequent analyses were performed by fitting the LMD peaks with a normal distribution. The mean of the distribution of lower value peaks (which were more in number) was assumed as the CH₄ concentration of the exhaled gas from respiration, whereas the mean of the distribution of the higher value peaks (which were fewer in number) was assumed as the CH₄ concentration of the exhaled gas from eructation.

Figure 1.

Pearson correlation coefficient between CH₄ concentrations, measured by LMD, in the exhaled gas from respiration and eructation differentiated by fitting a two probability distributions (respiration–eructation). N, normal; L, log-normal; G, gamma; W, Weibull distribution functions.

As expected, the more hours used to calculate the composite mean, the smaller the RSSD between the 24-h mean and the composite mean of CH₄ concentrations observed in both respiration (Fig. 2) and eructation (Fig. 3). For respiration, the smallest mean RSSD between the 24-h mean and the composite mean of CH₄ concentrations was observed at 0500 hours, with the means of four subsequent hours, corresponding to two (0500) and one (0600) hours before, and one (0800) and two (0900) hours after morning feeding, and five subsequent hours with an additional three (1000) hours after morning feeding (Fig. 2). The mean RSSD at 0500 hours with four (4.1) and five (3.7) subsequent hours were significantly smaller than that with two (19.4) or three (8.7) subsequent hours; however, they were not different from each other (P > 0.1). The same trend was observed in eructation, although the smallest mean RSSD overall (15.7) was observed at 1200 hours, with two subsequent hours (1200 and 1300 hours). Four and five subsequent hours had the smallest mean RSS of 17.1 and 17.5, respectively, at 0500 hours. Since a fewer number of hourly measurements is preferred, it was concluded that the most appropriate measuring time and number of days to represent the daily mean CH₄ concentration of the exhaled gas is four time measurements, rather than five-time measurements, at 2 h and 1 h before, and 1 h and 2 h after morning feeding.

Figure 2.

The square root of the sum of squares of differences (RSSD) between the 24-h mean and the composite mean of CH₄ concentrations in the exhaled gas from respiration (in ppm). The mean (point) and standard deviation (error bars) of four steers of the composite means of two (circles), three (triangles), four (diamonds), and five (squares) subsequent hours at each hour of the day over 2 d.

Figure 3.

The square root of the sum of squares of differences (RSSD) between the 24-h mean and the composite mean of CH₄ concentrations in the exhaled gas from eructation (in ppm). The mean (point) and standard deviation (error bars) of four steers of the composite means of two (circles), three (triangles), four (diamonds), and five (squares) subsequent hours at each hour of the day over 2 d.

Experiment 2

As designed, the actual FC ratio of the low FC group (LFC-HEC) was significantly lower (26% of dietary DM) than that of the high FC group (34% of dietary DM), but there was no significant difference in total dry matter intake (DMI) and growth performance among the treatments (P > 0.05; Table 5). The forage DMI did not differ within the high FC ratio group (P > 0.05), but the LFC-HEC consumed 0.79 kg (28%) less forage than HEC-LEC (P < 0.05). In contrast, the concentrate DMI was 0.68 kg greater in the LFC-HEC group than that in the high FC ratio group (P < 0.05). Also, as planned, within the HEC-fed group, no statistically significant difference was observed in NDF intake (P = 0.615), which was 0.83 kg (31%) greater than HFC-LEC (P <0.01). Forage NDF intake did not differ between HFC-LEC and HFC-HEC or between HFC-HEC and LFC-HEC; however, there was a 0.5 kg difference between HFC-LEC and LFC-HEC in forage NDFI (P < 0.01).

Table 5.

Effects of dietary treatments on growth performance in Hanwoo growing steers

Item†	Treatment‡			SEM	P-value
	HFC-LEC	HFC-HEC	LFC-HEC
Initial BW, kg	346	346	338	11.2	0.193
Final BW, kg	375	372	365	11.1	0.155
ADG, g	1,033	956	977	61.4	0.630
FCR	7.65	8.19	8.11	0.516	0.587
Intake, kg
Total DMI	7.81	7.53	7.72	0.271	0.527
Forage DMI	2.82a	2.52ab	2.03b	0.276	0.014
Concentrate DMI	5.00b	5.01b	5.68a	0.048	<0.001
Total NDFI	3.52a	2.77b	2.61b	0.172	<0.001
Forage NDFI	1.79a	1.60ab	1.29b	0.175	0.014
Fiber content in the diet consumed, % DM
Forage	35.6a	33.0a	25.7b	2.63	<0.001
NDF	44.9a	36.5b	33.6c	1.02	<0.001
Forage NDF	22.6a	21.0a	16.4b	1.67	<0.001

BW, body weight; ADG, average daily gain; FCR, feed conversion ratio; DMI, dry matter intake; NDF, neutral detergent fiber; NDFI, neutral detergent fiber intake.

HFC-LEC, a high forage-to-concentrate ratio (0.35) with a low-energy concentrate mix; HFC-HEC, a high forage-to-concentrate ratio (0.35) with a high-energy concentrate mix; LFC-HEC, a low forage-to-concentrate ratio (0.25) with a high-energy concentrate mix.

Means that do not have common superscripts significantly differ within the treatments (P < 0.05).

The CH₄ concentrations in the exhaled gas of the steers differed significantly by treatment (P < 0.001). The CH₄ concentration in the exhaled gas from both respiration and eructation (ppm-m and ppm-m/kg of DMI) was significantly higher in the HFC-LEC group than that in the LFC-HEC group (P < 0.001; Table 6). Compared with the LFC-HEC, the HFC-LEC showed 49% (8.1 ppm) and 57% (52.6 ppm) higher CH₄ concentrations in the exhaled gas from respiration and eructation, respectively. Within the HEC-fed group, the CH₄ concentration in the exhaled gas from respiration did not differ (P > 0.05); however, HFC-HEC had a 31% (28.3 ppm) higher CH₄ concentration in the exhaled gas from eructation than LFC-HEC (P < 0.05). Within the high FC ratio group, no difference in CH₄ concentration in the exhaled gas was observed for both respiration and eructation.

Table 6.

Effects of dietary treatments on enteric CH₄ emissions in Hanwoo growing steers

Item‡	Treatment†			SEM	P-value
	HFC-LEC	HFC-HEC	LFC-HEC
CH4 from respiration
ppm	24.5a	20.6ab	16.4b	1.36	<0.001
ppm/kg of DMI	3.2a	2.8a	2.1b	0.18	<0.001
ppm/kg of FDMI	9.2	8.7	8.8	1.06	0.868
ppm/kg of NDFI	7.1	7.6	6.3	0.55	0.122
ppm/kg of fNDFI	14.5	13.7	13.8	0.18	0.868
CH4 from er.uctation
ppm	145.2a	120.9a	92.6b	8.27	<0.001
ppm/kg of DMI	18.6a	16.1a	11.9b	1.02	<0.001
ppm/kg of FDMI	53.8	50.7	49.4	5.74	0.698
ppm/kg of NDFI	41.6	44.3	35.5	2.84	0.058
ppm/kg of fNDFI	84.6	79.7	77.7	9.02	0.698

HFC-LEC, a high forage-to-concentrate ratio (0.35) with a low-energy concentrate mix; HFC-HEC: a high forage-to-concentrate ratio (0.35) with a high-energy concentrate mix; LFC-HEC, a low forage-to-concentrate ratio (0.25) with a high-energy concentrate mix.

DMI, dry matter intake; FDMI, forage dry matter intake; NDFI, neutral detergent fiber intake; fNDFI, forage neutral detergent fiber intake.

Means within the same row with different superscripts are significantly different (P < 0.05).

When the CH₄ concentration was expressed as per intake (ppm/kg), such as DMI, forage intake, NDF intake, and forage NDF intake, the CH₄ concentration of the respiration gas did not differ between treatments (P > 0.05). However, the CH₄ concentration of the eructed gas was significantly altered by the treatments (P < 0.001; Table 6). The CH₄ concentration per DMI in the exhaled gas from eructation was 6.7 ppm/kg (36%) lower than that in the high-FC ratio group (P < 0.001). This difference diminished when the CH₄ concentration was expressed per forage or NDF intake. (P > 0.05; Table 6).

The CH₄ concentration in the exhaled gas was significantly affected by the fiber content of the diet (P < 0.01; Figs. 4 and 5). The most significant variable that could explain the variations in CH₄ concentration was dietary NDF concentration (%), followed by NDF intake (in kg/d). The dietary NDF concentration explained 39% and 53% of the variation in CH₄ concentration in the exhaled gas from respiration and eructation, respectively.

Figure 4.

Plots of the CH₄ concentration in the exhaled gas from respiration against forage NDF(%), dietary NDF(%), forage as a percentage of dietary DM (%), and NDF intake (in kg/d). The lines represent the best-fit regression.

Figure 5.

Plots of the CH₄ concentration in the exhaled gas from eructation against forage NDF(%), dietary NDF(%), forage as a percentage of dietary DM (%), and NDF intake (in kg/d). The lines represent the best-fit regression.

Discussion

Although various methodologies have been proposed to measure CH₄ emissions from ruminants, a simple and readily applicable technique in practical situations is still needed to quantify and mitigate CH₄ emissions from enteric fermentation of ruminants in field settings. In this regard, measuring the CH₄ concentration in the exhaled gas using a hand-held LMD is a promising technique. The LMD technique does not require specialized facilities or restrict the normal behavior of animals. Moreover, the LMD method is time- and cost-effective and less laborious because no sampling, chemical analysis, or animal training is required. Owing to its potential, the LMD technique has been examined by various authors (Chagunda et al., 2009; Ricci et al., 2014; Reintke et al., 2020; Kobayashi et al., 2021), and a comprehensive review article has recently been published (Sorg, 2022). However, the LMD method suffers from low repeatability (Garnsworth, 2019), and a more sound and standardized protocol for measurement procedures and data analysis is required. Therefore, this study aimed to establish a standard procedure for the LMD technique to quantitatively assess enteric CH₄ emissions from cattle (experiment 1) and to evaluate its efficacy in detecting differences in CH₄ emissions from cattle fed different FC ratios (experiment 2).

The CH₄ concentration peaks of the exhaled gas can be successfully differentiated into respiration and eructation with a measurement duration of 6 min according to this study. The duration per measurement in the literature varied from 1 min (Grobler et al., 2014) to 6 min (Denninger et al., 2020) and between 3 and 5 min long in most studies (Sorg, 2022). When deciding on the measurement duration, there is a trade-off between the quality of data and workload, considering the number of animals measurable within limited work hours. Eighty-one percent of the CH₄ produced from enteric fermentation is emitted through eructation and the remainder through respiration (Murray et al., 1976). Therefore, the measuring time of the LMD technique should be sufficiently long to capture the CH₄ concentration of the gas exhaled by eructation. Considering the frequency of eructation events, a measurement once every 1–3 min is required (Waghorn and Reid, 1983), and a measuring duration of less than 6 min may be insufficient to represent the normal cyclic pattern of CH₄ concentration in the exhaled gas, due to errors.

This study proposes the use of the AMPD algorithm for peak detection. As the LMD technique measures the concentration of CH₄ gas around the nose at a specific time without gas sampling, the measured value fluctuates periodically with exhalation and inspiration during respiration. Therefore, it is necessary to consider only the peak concentration of CH₄ released by exhalation when measuring CH₄ emissions (Chagunda et al., 2009). Ricci et al. (2014) introduced the concept of mini-peaks and mini-troughs in LMD measurements, and Sorg et al. (2018) followed this approach. However, fluctuations in the LMD measurements occur not only due to inhalation and exhalation but also other unknown factors, including wind and equipment conditions. Thus, a simple algorithm based on the difference between the preceding and following values may be inappropriate for identifying the peak values of CH₄ released by the exhalation events. Therefore, in this study, we used the AMPD algorithm to effectively detect the peaks of a noisy periodic and quasi-periodic signal using a moving-window approach (Scholkmann et al., 2012). Because respiration and eructation are periodic events, it may be more appropriate to detect peaks assuming the periodic nature of time-series measurements using LMD, although a direct comparison was not performed in this study.

Point measurements for the CH₄ concentration in the exhaled gas from respiration and eructation were determined separately by fitting a normal distribution to the LMD measurement peaks. Although most previous studies assumed a normal distribution for respiration and eructation peaks (Chagunda et al., 2009; Ricci et al., 2014; Reintke et al., 2020; Kobayashi et al., 2021), no attempt has been made to test whether this assumption is valid, and there is no consensus on the separation of peaks according to the two events, and determination of representative point measurements. A comprehensive summary and discussion of this topic can be found in a recent review by Sorg (2022). We assumed that 1) the peaks from respiration and eructation could be separated by fitting a double distribution function and 2) the mean of each distribution is a representative point measurement. The adequacy of the determination of point measurements by distribution means can be evaluated only by the actual values, and this was performed in vivo in experiment 2. Whether a normal–normal distribution function adequately represented the distribution of the respiration and eructation peaks was assessed by comparing the correlation coefficients between respiration and eructation of the point measurements, assuming that there should be a correlation in the CH₄ concentrations between respiration and eructation. Among the 16 combinations of four distributions (normal, log-normal, gamma, Weibull), the highest correlation between point measurements of respiration and eructation was found with the normal–normal distribution (0.79), followed by Weibull–gamma (0.70), and normal–gamma (0.67). Since we do not have enough evidence that the normal–normal distribution is inadequate, and the normal–normal distribution has the highest correlation coefficient, we conclude that it is appropriate to fit both respiration and eructation with a normal distribution.

The most appropriate times to measure the daily CH₄ concentration of cattle fed twice daily were as follows; 2 h and 1 h before, and 1 h and 2 h after morning feeding. The smallest number of measurements is preferred when applying the LMD technique in the field. The difference between the mean CH₄ concentration at 2 h and 1 h before and after the morning feeding and the 24-h mean was small, and it was similar to the difference between the mean of five measurements and the 24-h mean. There is a diurnal pattern in CH₄ emission of cattle, which is cyclic according to their feeding patterns (Wang et al., 2015). Previous studies, which used respiration chambers, indicated that the emitted CH₄ concentration was the lowest before morning feeding, increased rapidly for 1 to 2 h after feeding, and then decreased constantly until next feeding (Crompton et al., 2011; Jonker et al., 2014; Hammond et al., 2015; Wang et al., 2019). This periodic pattern occurs because enteric CH₄ production is related to ruminal methanogenesis linked to the fermentation of digestible organic matter and volatile fatty acid production (Baldwin et al., 1987). These dramatic variations in enteric CH₄ production before and after feeding seem to be encompassed by the four measurements at 2 h and 1 h before, and 1 h and 2h after morning feeding. However, since the diurnal pattern in ruminal fermentation and methanogenesis is dependent on the number of feedings (Crompton et al., 2011), the appropriate measurement times may vary according to the number of feeds offered in a day.

In this study, the LMD technique successfully detected variations in CH₄ emission in cattle depending on the level of forage intake. Increasing the level of concentrate in the diet reduces enteric CH₄ production in cattle (Beauchemin et al., 2008). A high-fiber diet increases acetate synthesis, which enhances hydrogen (H₂) production. A higher H₂ concentration in the rumen leads to increased CH₄ production (Janssen, 2010). On the other hand, starch in concentrates enhances propionate production, which acts as an H₂ sink and reduces methanogenesis in the rumen. Although DMI is the single most significant factor affecting enteric CH₄ production, the inclusion of dietary forage content (in %%) and NDF intake (in kg/d), along with DMI, in a CH₄ prediction equation greatly improves its predictability (Lee et al., 2012). Although the total DMI were similar, there were significant differences in the CH₄ concentrations between the treatments. The difference in the CH₄ concentrations between the treatments was mainly due to forage DMI, which was supported by the small difference in the CH₄ concentration per forage DMI (in ppm/kg) between treatments. In addition, all fiber-related variables (e.g., dietary NDF concentration, dietary forage NDF concentration, and forage as a percentage of dietary DM) were significantly correlated with the CH₄ concentration in the exhaled gas from respiration and eructation of steers. The single most significant factor that could explain the variations in the CH₄ concentrations of steers was the dietary NDF concentration. These results were consistent with those of previous studies using respiratory chambers, which, in turn, indicated the capability of the LMD technique to detect the difference in CH₄ production manipulated by dietary treatment. The mean CH₄ concentration in the exhaled gas from eructation was 5.8 times greater than that from respiration.

Similar to the present study, the differences in CH₄ emissions between high-and low-forage diets using the LMD technique were also reported in a previous study (Ricci et al., 2014). Ricci et al. (2014) fed Aberdeen Angus × Limousin-crossed steers a diet with a high (48:52) or low (8:92) FC ratio and measured CH₄ emissions using LMD. Compared with a high forage diet, they found significant reductions in the CH₄ concentration in the exhaled gas from respiration (1.7 ppm; 18%) and eructation (33.6 ppm; 44%) in a low forage diet. Considering the greater difference in the FC ratio between the two treatments in their study compared with our study (35:65 vs. 25:75), the LMD protocol proposed in the present study may be more powerful for evaluating the effect of dietary treatment on CH₄ concentrations in cattle.

Conclusions

In this study, we proposed a protocol to use the LMD technique to measure enteric CH₄ emission from cattle, which includes measuring CH₄ concentration in the exhaled gas every 0.5 s by aiming LMD at the animal’s nostril, with four time measurements (−2, −1, +1, and +2 h after morning feeding) daily and duplicated measurements. We also proposed a data processing and analysis procedure for detecting CH₄ concentration peaks using the AMPD algorithm and fitting the peaks with a double normal distribution to separate the peaks into two emission pathways: respiration and eructation. The mean of the normal distribution was assumed to be the representative point measurement of CH₄ concentration for each event.

We conclude that the LMD method can be applied to evaluate differences in CH₄ emission in cattle using the protocol established in this study, which is less invasive, portable, cost-effective, and less laborious overall.

Glossary

Abbreviations

ADICP

acid detergent-insoluble crude protein

ADL

acid detergent lignin

AMPD

automatic multi-scale peak detection

BW

body weight

CA

carbohydrate A fraction

CB1

carbohydrate B1 fraction

CB2

carbohydrate B2 fraction

CB3

carbohydrate B3 fraction

CC

carbohydrate C fraction

CH₄

methane

CNU

Chungnam National University

CP

crude protein

DM

dry matter

DMI

dry matter intake

FC

forage-to-concentrate

LMD

laser methane detector

NDF

neutral detergent fiber

NDICP

neutral detergent-insoluble crude protein

NFC

non-fiber carbohydrate

PA

protein A fraction

PB1

protein B1 fraction

PB2

protein B2 fraction

PB3

protein B3 fraction

PC

protein C fraction

RSSD

square root of the sum of squares of differences

SD

standard deviation

SOLP

soluble crude protein

TDN

total digestible nutrients

Conflict of Interest Statement

The authors declare no conflict of interest.