A single dairy cow emits roughly 300–450 g of methane per day — about 12–19 g/hr on average. From the standpoint of airborne gas leak detection, this is one of the weakest methane point sources that any aerial system might attempt to register. No published study has ever recorded methane from an individual cow using an aircraft-mounted sensor. The thresholds of validated airborne detectors start at 100 g/hr and go up to several kilograms per hour, meaning they need anywhere from 7 to several hundred co-located cattle to trigger.
Yet field data collected during pipeline inspections across Europe with the ALMA G4 instrument tell a different story: the device has repeatedly picked up methane from individual cows and small groups of animals at altitudes of 37–46 m and helicopter speeds of 40–75 km/h. Typical registered concentrations were 7–15 ppm·m, and the cloud from each cow measured less than 2 m across. The estimated gas flow rate for these detections was 0.1–0.9 l/min, equivalent to 4–36 g/hr.
These numbers form a practical field test for the suitability of an airborne methane detector for pipeline inspection: if the instrument sees methane from every cow in a field, its sensitivity is more than sufficient to catch real leaks from an underground gas pipeline.

The peer-reviewed literature converges on a narrow range for enteric methane from a typical North American or European lactating dairy cow (~650 kg body weight, 25–30 kg milk/day). The EPA Cattle Enteric Fermentation Model yields ~140–148 kg CH₄/head/year, which translates to 380–405 g/day or ~16 g/hr on average. The USDA baseline calculation using the Niu et al. (2018) equation gives 401 g CH₄/head/day for a lactating cow. The IPCC Tier 1 default is 128 kg/head/year for North American dairy cattle and 53 kg/head/year for other cattle.
Respiration chamber studies report 322 g/day (Grainger et al., 2007) to 587 l/day (Holter et al., 1997) — equivalent to ~385 g/day at STP. The SF₆ tracer technique yields 331 g/day (Grainger et al., 2007) and 405 g/day with SD of 156 (Hristov et al., 2016). GreenFeed automated systems measure 328–373 g/day across multiple studies (McGinn et al., 2021; Hristov et al., 2016). Garnsworthy et al. (2012) measured instantaneous methane flux during milking at 2.07 g/min (SD 0.629), with daily estimates of 278–456 g/day. Danish farm-scale tracer dispersion measurements gave 26 ± 8.5 g per livestock unit per hour — 35% above IPCC inventory estimates, due to the manure contribution.
Beef cattle emit substantially less. A typical ~550 kg grazing beef cow produces 200–300 g/day (~85–100 kg/year per EPA Tier 2). Feedlot cattle on high-concentrate diets emit only 100–150 g/day (35–50 kg/year), reflecting a methane yield of just 8–13 g/kg dry matter intake versus ~20 g/kg DMI for forage-fed animals. The IPCC methane conversion factor (Ym) captures this: 6.5 ± 1.0% of gross energy for most cattle, dropping to 3.0 ± 1.0% for >90% concentrate feedlot diets.

Emissions are not steady-state. About 99% of methane exits through the mouth and nose via eructation (belching) and exhalation; only ~1% from flatus. The pattern follows a diurnal biphasic rhythm peaking mid-morning and late afternoon (during active feeding), with peak hourly rates reaching 3× the daily minimum. Eructation frequency in dairy cows averages ~54 events per hour, creating a pulsed rather than continuous signal.

No published study has directly mapped the full spatial plume from a single grazing cow in open field, but a consistent picture emerges from near-field laser measurements, open-path instruments, and Gaussian dispersion modelling.

• In the breath zone (0–0.3 m), sniffer analyzers at automated milking stations record 50–200 ppm during normal respiration, spiking to 500–1,000+ ppm during eructation events. Background barn methane runs 26 ± 10 ppm (Haque et al.).
• In the near-head zone (1–3 m), handheld Laser Methane Detectors (LMDs) report mean column densities of 97 ppm·m, with eructation peaks reaching 218–457 ppm·m (Sorg et al. 2018; Pereira et al.; Difford et al.).

Catrina cattle on Azores pastures showed breath readings of 33 ± 29 ppm·m and eructation peaks of 219 ± 67 ppm·m at 3 m distance. Standard Gaussian plume dispersion for a single cow emitting 10.8 g/hr (0.003 g/s) at 1 m source height under neutral stability (Pasquill–Gifford Class D) yields the following centreline ground-level concentrations:

Plume width (2σ_y) expands from ~1.6 m at 10 m downwind to ~16 m at 100 m. Column-integrated concentrations relevant to remote sensors drop rapidly: approximately 1.2 ppm·m at 10 m, 0.24 ppm·m at 50 m, and just 0.12 ppm·m at 100 m downwind (Class D, 3 m/s wind). These modelled values align well with field data: Weerasekara et al. (2024, Atmos. Meas. Tech.) measured a 24 ppb enhancement from 10 simulated cattle at 60 m downwind, closely matching the per-cow model scaled up.

The physics make individual cow detection from altitude effectively impossible for all existing airborne systems — except ALMA G4.

The California methane survey using AVIRIS-NG (Duren et al., 2019) flew at 3 km altitude with ~3 m pixel resolution and a detection threshold of 2–5 kg/hr — equivalent to 130–300 co-located cows. Dairies accounted for 26% of detected point sources, but exclusively from manure management infrastructure (lagoons, digesters), not enteric emissions. The California Air Resources Board explicitly noted that emissions from dairy cattle are mostly diffuse and may not be detectable by plume imaging technologies.
GHGSat’s satellite (Fabry-Perot spectrometer at 500 km orbit) first detected feedlot methane from space in 2022 — five plumes of 361–668 kg/hr southeast of Bakersfield, California. McLinden et al. (2024) independently validated GHGSat’s MDL at 100–240 kg/hr, corresponding to emissions from roughly 7,000–16,000 cows. Hacker et al. (2016) detected methane from a 17,000-head Australian feedlot up to 25 km downwind using aircraft-mounted quantum cascade lasers and could resolve individual pen rows — but not individual animals.
The EPA’s ground-based OTM-33A method (Picarro CRDS + 3D anemometer) has an MDL of approximately 36 g/hr — theoretically sufficient for a group of 3–5 cows under favourable conditions, but no livestock-specific validation has been published.

The airborne methane detection landscape spans helicopter-based TDLAS/DIAL instruments for pipeline patrol to high-altitude spectrometers for regional surveys. The most comprehensive independent comparison is El Abbadi et al. (2024), a blind test of five major airborne platforms at Stanford’s METEC facility across more than 700 measurements.

The sensitivity gap is an order of magnitude and more. ALMA G4 with an MDL of ~5 g/hr operates below the emission level of a single dairy cow (~16 g/hr), which is exactly what the field data confirm. Adlares CHARM at ~100 g/hr needs at least ~7 cows for minimum detection. Bridger Photonics at an MDL of 500–3,000 g/hr needs the equivalent of at least ~30 cows. Systems like AVIRIS-NG and GHGSat operate at the scale of large farms and industrial facilities.

When airborne methane detectors patrol pipeline rights-of-way through agricultural areas, biogenic sources create interference ranging from nuisance false positives to serious data confusion.

The scale of the problem is quantifiable. Weller et al. (2018) found in ground-level surveys that roughly 25% of all methane leak indications did not correspond to natural gas leaks — they were biogenic (sewers, wetlands) or unidentifiable. A single dairy farm emits 4–8.6 kg CH₄/hr (Vinković et al., 2022), comparable to many gathering and distribution pipeline leaks. A 139,000-head feedlot in California’s Imperial Valley was identified as the single largest methane point source in the entire state, exceeding every oil well, refinery, and landfill.
Drone-based TDLAS surveys (MDPI Sustainability, 2024) recorded false methane peaks exceeding 5,000 ppm·m from surface reflectance artefacts, with background levels varying dramatically between agricultural, suburban, and industrial areas and between sunny and overcast conditions. Pätzold et al. (2024) explicitly warned that agricultural emissions may mix with leak signals and trigger false alarms in automated systems.

• Ethane co-detection is the primary discrimination method. Biogenic sources contain less than 0.2% ethane, while pipeline-grade natural gas contains 5–15%. Commercial instruments (Aerodyne SuperDUAL, Aeris MIRA Ultra LDS) achieve sub-ppb ethane sensitivity at 1 Hz, enabling real-time source classification. Yacovitch et al. (2014, Environ. Sci. Technol.) established the threshold values now widely used: ethane/methane ratios below 0.2% indicate biogenic origin; above 1% indicates thermogenic gas.
• Spatial localisation helps at high resolution. Bridger Photonics achieves ~2 m geolocation accuracy, enabling GIS-based filtering of known agricultural facilities.

In the case of ALMA G4, methane registrations from cows during European pipeline inspections are deliberately included in reports — gas companies forward the total volume of agricultural methane registrations to environmental organisations. High instrument sensitivity does not create a false-alarm problem; instead, it generates an additional stream of useful data.

The ~5 g/hr MDL estimate for ALMA G4 is based on field data from real inspections. Formal compliance with OGMP 2.0 Level 4 and Level 5 requirements calls for an MDL calculation at 90% probability of detection (POD), which requires controlled releases at a test facility with a wide corridor and a scanning system suitable for statistical analysis.
Additionally, TDLAS and LiDAR integration is under development — following the approach of Bridger Photonics, where LiDAR provides spatial referencing and TDLAS/DIAL provides concentration analysis. This combination will enable reporting compatible with OGMP 2.0 and Carbon Mapper monitoring frameworks.