Quantifying Methane Emission Rates Using Downwind Measurements

Definition

This entry describes the methods used to quantify methane emissions from either point or area sources using downwind methods. The methods described could be used as a practical guide to quantify emissions of any trace gas type from either a point or area emission source. Methane is a relatively strong greenhouse gas, its GWP is 25 times larger than CO₂ over a 100-year period, and an increase in methane anthropogenic emissions has been correlated to a changing global climate. Emission estimates that are calculated and used for national inventories are usually derived from bottom-up approaches, however there is now an increasing pressure for these to be validated by direct measurement. Calculating emission rates from downwind measurements has proven to be a versatile and relatively simple approach for direct measurement. Downwind measurement method descriptions are presented here as a practicable guide to quantifying point and area source emissions. Emission quantification is a two-stage process where methane concentration and meteorological data must be measured downwind of a source and then converted to emissions using an atmospheric dispersion model. Only four technology types currently measure in the range typical of downwind methane concentrations: metal oxide sensors, non-dispersive infrared sensors, tunable diode laser absorption spectrometers and optical cavity instruments. The choice of methane measurement is typically determined by the size of the emission source, location and the budget of the project. Meteorological data are essential to quantifying emissions, especially regarding wind speed and direction. In most cases, simple atmospheric dispersion approaches can be used to quantify both area and point emissions using these downwind measurements. Emissions can be generated using limited data (only methane concentration, wind speed, wind direction, and locations are necessary), but quantification uncertainty can be reduced using more input data. Site selection and location of instrument deployment are essential because quantification approaches assume a flat fetch (no aerodynamic obstructions) and constant wind fields. When modeling assumptions are violated, quantification uncertainty can range between +250% and −100% of the actual emission rate. At present there, is no happy medium between modeling complexity and computational time, and this remains the biggest challenge for downwind emission quantification.

1. Introduction

This entry describes the methods used to quantify methane emissions from either point or area emission sources using downwind measurements. The description of methods could be used as a practical guide to quantify emissions of any trace gas type (e.g., carbon dioxide, hydrogen, nitrous oxide) as gases tend to become very dilute after emission (from ~0.1% to ppm-levels after distances ~1 m) and different gases have negligible buoyancy differences when diluted [1]. Here, methane is used as the example trace gas as it has practical importance to emissions from the energy, waste, agriculture and natural sectors.

Over the past decade, it has become accepted that the changing global climate is a response to the increase in atmospheric gas species’ concentration [2,3,4]. This has largely been attributed to an increase in gas emissions from anthropogenic activities, which have resulted from changes to industrial manufacturing, agricultural processes and energy demand [2]. Due to their absorption of infrared radiation, several gases have been identified as greenhouse gases (GHGs), e.g., carbon dioxide, methane, nitrous oxide and sulfur hexafluoride. Each gas type has the capacity to absorb a different amount of infrared radiation, typically standardized against carbon dioxide, which has a global warming potential (GWP) of one, and more heat is trapped in atmosphere with a larger abundance of atmospheric GHGs.

Methane is a relatively strong greenhouse gas as its GWP is 25 times larger than CO₂ over a 100-year period [2,4]. The 100-year GWP (GWP100) is universally accepted as the standard GWP metric for comparison between gases [5]. Until recently, most GHG emissions at site, region and national levels have been calculated using bottom-up methods, where emissions are calculated by multiplying the number of emission activities within the area of study by an emission factor [6]. Emission factors are a quantitative assessment of how emissive a source is; units are typically “grams of gas per unit time”, and usually derived from a measurement study. More recently, it has become apparent that there are significant differences between bottom-up emission estimates and the amount of gas that has been emitted [7,8,9,10,11,12], with the discrepancy largely attributed to incorrect or inaccurate emission factors [10,13,14,15].

With climate goal deadlines looming, there has been an increased demand for the validation of emission estimates [16,17,18]. Typically, this has been adopted as “Measurement, Monitoring, Reporting and Verification Frameworks”, which require greenhouse gas emissions to be verified through direct measurement. Direct measurement can be made in a number of ways: enclosing the source in a chamber; summing individual sources within an area; inferring emissions from downwind measurements; using tracer flux; making eddy covariance measurements; applying a mass balance approach; or remotely sensing the plume using aircraft or satellite [7,19,20,21,22,23,24,25,26,27,28,29]. The advantages and disadvantages of these methods are described in a number of papers [23,30] but are often limited by either high quantification thresholds—10+ kg CH₄ h⁻¹ for aircraft [22] and 100+ kg CH₄ h⁻¹ for satellites [7,11,31]—or the need for direct access to the emission source.

Inferring emissions from downwind measurements does not suffer from these shortcomings and has proved to be a versatile and relatively robust approach for quantifying emissions. Sources can either be point or area sources and emissions can be calculated using relatively little data, e.g., methane concentration and wind speed, at a minimum. Uncertainty in quantification decreases with an increase in input data quality. Downwind methods have been used to quantify methane emissions from onshore oil and gas facilities [27,32], offshore oil and gas facilities [33,34], landfill [29,35,36], agriculture [37,38,39,40], and natural sources [37], with emission ranging from 10 g CH₄ h⁻¹ to 100 kg CH₄ h⁻¹ [33].

Methane concentration data are essential for calculating the emission. The size of the concentration observed is a function of distance downwind, the size of the emission, the wind speed and atmospheric stability. Typically, expected downwind methane concentrations are less than 20 parts per million by volume (ppm) [27,37,38,41]. Background methane concentrations are also required; these are typically measured upwind of the source and usually observed to be ~1.8 ppm [12,37,42,43]. Coupled with concentration data, meteorological data are also required with wind speed being essential. This can be measured using a standard rotating-cup anemometer, but better quantification accuracy can be achieved using a 3D sonic anemometer. Other meteorological data are desirable but not essential, including wind direction, temperature, relative humidity atmospheric pressure, and solar irradiance. Similarly, micrometeorological data are desirable but emissions can be quantified without quantitative data. These data are usually measured by a 3D sonic anemometer but can also be derived from lower-tech instruments. Details are provided in full below.

Point sources are defined as a single identifiable source of emission, e.g., a crack in a pipe, while an area source is a group of sources emitted over a finite region. There is a degree of fluidity between how a source may be defined, point or area, and this is usually a function of observation distance. For example, a landfill can be considered an assemblage of individual point sources when measuring on-site, an area source at a middle distance, and a single-point source when observed from even farther away. The distance at which the source transitions between types is dependent on the size and distribution of the sources and needs to be determined for each individual emission scenario. Emissions are typically presented as the emission rate in units of “grams of gas per unit time” for point sources or as a flux in units of “grams of gas per unit area per unit time” for area sources.

As mentioned above, quantification uncertainty depends on the quality of the input data. For very far-field measurements, up to 10 km away, the uncertainty has been estimated at plus/minus a factor of two (+100%, −50%) [33]. For nearer measurements (less than 1 km), downwind methods’ quantification uncertainty has been measured at ±60% [37,44] using controlled releases. At even closer distances (<5 m), downwind methods’ quantification uncertainty has been measured at less than ±15% [20,30]. The main cause of uncertainty at small distances is parameterizing the lateral dispersion of the plume, and this can be overcome by measuring directly downwind of the emission source.

In recent years, many publications have described methods that use downwind measurements to quantify methane emissions. In this entry, method descriptions are presented as a practicable guide to quantifying point and area source emissions with the inclusion of suggestions that could be used to work around missing data or overcome instrumentation shortcomings.

2. Downwind Methods Used to Quantify Methane Emissions from Either Point or Area Sources

2.1. Data Collection

2.1.1. Trace Gas Concentration Data

There are many methane sensing technologies available; however, not all detect methane at the concentrations typically observed at distances downwind. For example, catalytic combustion using Pellistor beads can be used to quantify methane and other combustible gases in air [45]; however, detection limits are relatively high (10 ppm), with a resolution of 10 ppm between measurements, and they cannot distinguish between hydrocarbon species of gas [46]. Integrated infrared (INIR) sensors measure the absorption of infrared radiation as it passes through the measurement cell [47]; however, the lowest detectable methane concentrations are ~100 ppm and sensors do not respond to methane concentrations less than this value.

There are four technologies currently available that could be used to measure methane concentrations downwind, i.e., which are able to detect methane at less than 20 ppm. These are metal oxide (MOx) sensors; nondispersive infrared (NDIR) sensors; tunable diode laser absorption spectrometers (TDLASs); and optical cavity instruments like the cavity ring-down spectrometers (CRDSs).

MOx sensors are the cheapest option, with the actual sensor costing ~$15 (as of 2025). The sensor is a tin oxide strip doped with a noble metal that changes resistance in the presence of methane [48], and the methane concentration in air can be inferred using a potential divider circuit [49,50]. In recent studies, MOx sensors have been calibrated to measure methane concentrations below 20 ppm and used to quantify emissions from oil and gas production facilities, gas pipelines, landfills and natural sources [51,52,53,54,55,56]. The MOx sensors are very sensitive to environmental change, particularly temperature and relative humidity [51], and are calibrated by comparing sensor output to a reference instrument (not using calibration gas) [57]. Recent studies have reported that both calibration and accounting for sensors’ response to environmental variability can be achieved using machine learning [58,59]. MOx sensors have been used for downwind quantification; however, there is concern over their ability to respond to rapid changes in methane concentration caused by variable, sub-minute wind fields [60]. A recent study has shown that MOx sensors can take around a minute to respond to changes in concentration [60], likely caused by a delay in methane absorption into the metal oxide strip. Metal oxide sensors are also very sensitive to temperature and, as suggested above, must be corrected for the effects of both temperature and relative humidity [51,59]. Novel ways of temperature correction are undergoing development and include the use of machine learning and neural networks [59].

NDIR sensors use three detectors to quantify the absorption of radiation at species-specific frequencies to infer the concentrations of gases [61,62]. One commercially available NDIR sensor can quantify the concentrations of nitrous oxide, methane, carbon dioxide and water to ppm levels [61]. This product is relatively new and has not been actively deployed in the field but initial laboratory-based experiments show that it can respond to sub-20 ppm changes in concentration varying at a frequency of 0.2 Hz [60]. Methane-specific NDIR sensors are priced ~$500. In comparison to MOx sensors, NDIR sensors have shown that they respond to changes in methane concentration much faster [60]. A recent study has shown that an NDIR can respond to individual methane plumes (~7 ppm above background) released 5 s apart [60]. There are currently no published data on how NDIR sensors are affected by temperature and relative humidity.

TDLAS systems typically cost between $3000 and $20,000. These instruments use a laser tuned to the absorption wavelength of methane to quantify concentration to ppb levels and can report concentrations in real-time [63]. The main shortcoming of TDLAS is that the signal-to-noise ratio is affected by ambient temperature at low concentrations as small temperature changes increase/decrease the path length of the laser. TDLAS systems are typically light enough to be mounted on aerial drones and have been used to quantify methane concentrations downwind of landfill and oil and gas production facilities [64,65,66].

Optical cavity-based instruments come at a higher price point, $30,000 to $50,000. These instruments are laboratory-grade and can measure methane-to-ppb level concentrations with high accuracy and precision [67,68]. As such, these are often used in tall-tower measurement, where precision is essential to quantifying landscape emissions by measuring changes in boundary layer methane concentrations on the order of tens of ppb [12,69,70,71]. Optical-cavity instruments overcome the temperature shortcomings of the TDLAS system in various, instrument-specific ways by either having a cavity ring-down spectroscopy (CRDS) or Off-Axis Integrated Cavity Output Spectroscopy (OA-ICOS) technology [72,73,74]. As these instruments are highly sensitive, they can also be fragile (despite ruggedized packaging).

The choice of instrument in the field often comes down to project budget, but there are other considerations like power consumption and measurement resolution (

Table 1. Summary of current technologies that can be used to make volumetric mixing ratio measurements of methane. Instrumentation includes metal oxide (MOX), nondispersive infrared (NDIR) sensors, tunable diode laser absorption spectrometers (TDLASs), and optical cavity analyzers. Minimum and maximum indicate likely ranges in quantification. Power consumption, cost and range of methane detected varies between manufacturers; this is only meant to be suggestion of likely values.

Technology	Minimum (ppmv)	Maximum (ppmv)	Resolution (ppmv)	Power	Methane Specific	Cost (USD)
MOX	1	200	0.5	210 mW	No	$15
NDIR	1	2500	0.5	500 mW	Yes	$500
TDLAS	0.1	1000	0.1	1.5 W	Yes	$10,000
Optical cavity	0.001	100	0.001	35 W	Yes	$40,000

). All systems listed here are theoretically capable of making methane concentrations suitable for quantification emission rates; however, most suitable instruments will likely be affected by the parameters of the experiment (e.g., emission rate, power availability, and accessibility of the site) and environmental conditions (e.g., temperature, precipitation, and variability in wind field). Care should be taken to ensure that the instruments used can measure within the range of likely downwind concentrations, and pilot studies are often the best way to achieve this.

Ideally, instrumentation is calibrated before, during and after deployment, depending on the length of the experiment. Calibrations are typically performed using at least three concentrations of the highest-quality calibration gas, which bookend the expected measured concentrations: a high concentration above the highest expected concentration (20 ppm); low concentration around the background (1.8 ppm); and a target concentration around the median expected concentration (5 ppm). The instrument response should then be used to generate a calibration curve that can show any offset (±X ppm), non-linear behavior of the instrument, or instrument drift. These calibration curves should be used to correct all measured data in post-processing. It is not recommended that the calibration of field instruments is performed using the instrument software as it does not allow the user to independently track instrument variability over time.

Once collected and calibrated, the methane concentration data should be reviewed, and data removed for periods when the instrument is either not working or affected by external influences, e.g., while being maintained, human breath can increase concentrations near the inlet. Typically, to calculate an emission concentration, data will be averaged over a period of time, usually 15 or 20 min, short enough to account for periods of atmospheric turbulence but long enough to include real concentration changes caused by varying emission rates [39].

2.1.2. Meteorological Data

Meteorological data matching the concentration are essential for emission quantification as dispersion models require an understanding of how the gas will disperse as it travels downwind [75,76]. At a minimum, wind speed and wind direction measured at a known height in clear air (i.e., with no effects of aerodynamic obstructions) are required. Data should be collected at as high a sampling rate as possible and meteorological instruments should be calibrated (i.e., make sure the wind vane is pointing the correct direction by comparing it to a compass). Note, wind direction in dispersion modeling is the direction wind has come from.

Secondly, an understanding of solar insolation is important. This is usually measured using a pyranometer, which reports the global radiation in Watts per square meter (W m⁻²). Solar insolation is used to generate an understanding of the atmospheric stability (described in Section 2.1.4). Other environmental data are good to have but not strictly necessary for emission quantification; these include temperature (air and surface), atmospheric pressure, relative humidity and precipitation. For use in atmospheric dispersion models, all environmental data should be time-synced with the methane concentration data and averaged over the same time period.

2.1.3. Micrometeorological Data

Micrometeorology relates to near-surface atmospheric phenomena and processes. in terms of modeling, it explains what affects the dispersion of air traveling between the source and the detector. With a better understanding of micrometeorology comes a more accurate estimate in terms of emission quantification. Typically, the micrometeorology can be explained using two key variables: the atmospheric stability and the roughness length.

Atmospheric stability

Atmospheric stability is a measure of the vertical and horizontal movement of the air between the emission source and detector [23,40,76,77,78]. With high solar heating, air at the surface becomes buoyant and rises vertically and if the vertical movement is greater than the horizontal movement (i.e., in light winds), the atmosphere is described as unstable [76]. In unstable air, gas emitted from source will travel vertically quickly, resulting in a tall plume. With low solar heating and low wind speeds, the vertical movement of air is towards the surface. This is described as a stable atmosphere and emitted gas is typically trapped near to the surface [76]. At higher horizontal wind speeds (>5 m s⁻¹), the atmosphere is described as neutral and the plume disperses in a classical conical shape, reflecting off the surface and the top of the boundary layer [76].

In some situations, it is good enough to describe the atmospheric stability using the Pasquill Gifford Stability Classes (PGSCs). The PGSC categorizes the atmosphere from A to G depending on the wind speed and solar insolation and can be determined using look-up tables presented in many publications [37,76,77]. PGSC A denotes highly unstable conditions (low wind speed and high solar insolation), PGSC D denotes neutral conditions (higher horizontal wind speeds), and PGSC G denotes highly stable atmospheric conditions (low wind speeds and low/no solar heating). The PGSC can be directly used in most atmospheric dispersion models.

In other situations, it is necessary to quantify the stability of the atmosphere. This is typically achieved by calculating the Monin–Obukhov length (L, m) of the atmosphere using a sonic anemometer. The Monin–Obukhov length is calculated from the surface friction velocity (u_*, m s⁻¹), mean potential temperature (Θ, K), von Kármán’s constant (k, 0.41), gravitational acceleration (g, 9.8 m s⁻²) and the surface turbulent flux of sensible heat H (Equation (1)), with the Θ, u_*, and H output from the sonic anemometer [79,80].

$$L=-{\displaystyle \frac{\Theta u_{*}^3}{kgH}}$$

(1)

Roughness length

The roughness length (z₀, m) is a measure of the mechanical turbulence caused by aerodynamic obstructions on the ground between the emission point and the detector, i.e., the fetch. The roughness length can be estimated as one tenth of the height of an obstruction in the fetch or from look-up tables based on a description of what is in the fetch [76,81,82,83]. Roughness lengths vary from open ocean (z₀ = 0.0002 m) and featureless terrain (z₀ = 0.005 m) to bushes/small trees (z₀ = 0.5 m), mature forest (z₀ = 1 m) and buildings (z₀ > 2 m).

The roughness length can also be quantified using multiple anemometers at different heights. When the natural log of the anemometers’ heights (y-axis) is plotted against the wind speed measured by the anemometers (x-axis), the intercept on the y-axis is the roughness length. Of note, the gradient of the line of best fit multiplied by the von Karman constant (k = 0.41) is the friction velocity, which is used to calculate the Monin–Obukhov length.

2.1.4. Instrument Deployment

One major assumption of both downwind approaches used here is that the fetch between the emission point and the detector is relatively flat and does not have tall aerodynamic obstructions in it. This assumption will determine where the detector is to be placed. Ideally, the fetch will be very flat; however, this is unlikely to be the case. The effects of obstructions in the fetch can be overcome by placing the instruments farther downwind by a factor of between 20 and 100 times the height of the obstruction [30].

Once a suitable location for downwind measurement is identified, the instrument should be located downwind of the inlet, assuming air is being drawn to the instrument. The inlet, which is carrying air to the methane instrument, should be suitably high off the ground so that it is not affected by surface turbulence (1 to 2 m above ground level), and is protected against water/dust inclusion. The instrument/inlet package should be designed to create as little aerodynamic resistance as possible. One important note is to measure all locations/distances between emission sources, instruments and any aerodynamic obstructions.

Systematic error is also an important consideration when deploying instrumentation as methane sources, including when research staff can affect the reported methane concentration. This is of particular concern when monitoring small changes in concentration with high-precision instruments, e.g., tall tower measurements, as in, e.g., [12,84], where even small changes in the signal can indicate large sources at a distance. Care should be taken when locating analyzer inlets and visits to sites documented for QC/QA purposes.

2.1.5. Practical Considerations

The main practical considerations with running downwind measurements are as follows: (1) locating the methane instrument; (2) continuously powering the instrument; (3) weather; and (4) data collection and back up. As described in Section 2.1.4., locating the methane instrument to avoid aerodynamic obstructions is essential. However, it is also important to locate the instrument downwind of the emission source enough to answer the following research questions: “What is the instantaneous emission of the source?”; “How does the emission change over time?”; and “How does environmental variability affect emissions?”, with changes in the deployment location depending on the wind direction on the day of measurement, or use of a more long-term approach to maximize data collection downwind by considering long-term prevailing wind directions. Considering this can avoid the time-consuming task of moving the instrumentation at a later date.

Power is a major consideration and often the difference between a success and failure. Long-term deployment (i.e., longer than 24 h) or using an optical cavity instrument will likely require access to mains power, which greatly reduces the options for deployment location; however, it does mean that interruptions in data capture will be reduced. The lower-cost methane instruments can be powered by battery with solar/wind backup, but these instruments still require a non-trivial amount of energy (MOx—42 mA; NDIR—30 mA) which makes continuous measurement a challenge.

Rain, snow, wind, temperature and sunlight are all problematic for data collection. Some equipment can be protected from water incursion using IP6X enclosures; however, summer highs and winter lows (+40 °C to −30 °C in Colorado) mean that instruments (especially TDLASs and optical cavity instruments) need to be deployed in a climate-controlled environments. This impacts both power and location (see above). While it is often easiest to locate an instrument in a vehicle or outbuilding, aerodynamic considerations mean they are poorly suited for the task. Some experiments have been conducted in small climate-controlled enclosures [85,86], but these can be expensive and require mains power for heating/cooling.

Data backup for storage away from the measurement location is essential as instrument failure, fire, strong winds or theft can result in total data loss. This can typically be performed via a mobile phone network, but not in all cases. For fieldwork in remote locations away from these networks, data should be physically backed up on a daily basis and stored in a safe location.

2.2. Quantifying Emissions

Even though there are many different modeling approaches available, the two most commonly used methods for quantifying emissions using downwind measurements are 1. the Gaussian Plume approach and 2. the backwards Lagrangian stochastic model. Equations derived in the first half of the twentieth century described the lateral and vertical dispersion of a pollutant plume [87,88], including reflection at the surface, and formed the basis for the Gaussian Plume approach [78]. Backwards Lagrangian stochastic models simulate the random path that air parcels take from the point of emission to where they are detected as they are acted upon by horizontal and vertical forces [75]. For the purposes of this study, the Gaussian Plume approach will only be used to describe the quantification of point source emissions, while the Backward Lagrangian stochastic model will be used to quantify area source emissions.

2.2.1. Point Source Emissions—Gaussian Plume Approach

The Gaussian Plume approach uses an empirically derived equation which can be used to calculate the concentration of a gas downwind of a point emission source. The concentration of a gas (Χ, g m⁻³) at a point in space, x meters downwind of the source, y meters laterally from the plume and z meters above ground level can be calculated from the emission rate of the gas (Q, g s⁻¹), wind speed (u, m s⁻¹), the height of the source (h_s, m), the height of the boundary layer (h, m) and the Pasquill Gifford Stability Class (Equation (2)). The PGSC is used to describe the dispersion of the plume using lateral (σ_y, m) and vertical (σ_z, m) mixing ratio distributions which are obtained using EPA-published look-up tables, Table 1 and Table 1-2 [78]. Of note in Equation (2) are the following: 1. the first exponential (outside the brackets containing y), which refers to the lateral dispersion and is equal to one when directly downwind of the emission point; 2. the first and second exponentials inside the brackets that model the reflection of the plume by the ground; and 3. the last three exponentials in the bracket that model the reflection of the plume at the bottom of the boundary layer.

$$\left(x,y,z\right)={\displaystyle \frac{Q}{2\pi u{\sigma }_y{\sigma }_z}}{e}^{{\displaystyle \frac{-y^2}{2{{\sigma }_y}^2}}}\left(e^{{\displaystyle \frac{-{\left(z-h_s\right)}^2}{2{{\sigma }_z}^2}}}+e^{{\displaystyle \frac{-{\left(z+h_s\right)}^2}{2{{\sigma }_z}^2}}}+e^{{\displaystyle \frac{-{\left(z-2h+h_s\right)}^2}{2{{\sigma }_z}^2}}}+e^{{\displaystyle \frac{-{\left(z+2h-h_s\right)}^2}{2{{\sigma }_z}^2}}}+e^{{\displaystyle \frac{-{\left(z-2h-h_s\right)}^2}{2{{\sigma }_z}^2}}}\right)$$

(2)

To determine σ_y, the user uses Table 1-1 in [78] to determine values for c and d for the PGSC. The value of σ_y is then calculated from c, d and the distance downwind in kilometers (x, km) in Equation (3). To determine σ_x, the user uses Table 1-2 in [78] to determine values for a and b for the PGSC in the table for the correct distance between the source and the detector. The value of σ_x is then calculated from a, b and the distance downwind in kilometers (x, km) in Equation (4).

$${\sigma }_y=465.11628\times x\times tan\left(0.017453293\left[c-dln\left(x\right)\right]\right)$$

(3)

$${\sigma }_x=a\times x^b$$

(4)

The Gaussian distribution generated by the equation assumes (1) a constantly emitting source; (2) perfect reflection of gas at the surface and the top of the boundary layer; (3) wind speed is constant; (4) uniform vertical mixing; (5) the fetch is flat and the roughness length does not matter; (6) the fetch is more than 100 m; and (7) a point source emission. In many practical cases these assumptions are transgressed, if not completely violated, but this does not matter, as an answer will always be calculated (even though it may be wrong). This is both a strength and weakness of the approach and note must be taken of how these assumptions are accommodated for when validating the quality of an emission estimate generated using this approach.

Despite concerns over quantification estimates, the Gaussian Plume approach has been validated using controlled release experiments with a quantification uncertainty of ±60% at 30 m downwind and ±15% at 5 m downwind [30,37]. The main cause of uncertainty at small distances is parameterizing the lateral dispersion of the plume, and this can be overcome by measuring directly downwind of the emission source. This approach performs less well in more aerodynamically complex environments. In the case of measuring 30 m downwind of an oil production site (maximum obstructions of 6 m above ground level), the quantification uncertainty was measured at −100 and +250% for single-point source release [85].

Practical implications

Even though the Gaussian Plume approach has been used to quantify emissions from landfill and agriculture, it is most suited to practically quantifying the point source emission associated with oil and gas production and transmission. This method has been used extensively for quantifying onshore production emissions [27,37,44,53,89,90] and methods are now being developed to measuring offshore emissions [33,34]. As described above, the main shortcoming with the Gaussian Plume approach is that it cannot account for aerodynamic obstructions.

2.2.2. Area Source Emissions—Backward Lagrangian Stochastic Model

Backward Lagrangian stochastic models can be used to estimate emissions from area sources. The WindTrax model v2.0 (www.thunderbeachscientific.com (accessed on 24 April 2025)) is open source software that can be used to quantify emissions using downwind measurements from agricultural, waste and natural sources up to 1 km away [39,40,75,91,92,93]. The minimum data input to WindTrax is the size and shape of the emission source, the downwind methane concentration, location of the detector, wind speed, and wind direction. The model then makes assumptions on stability, roughness length and atmospheric temperature to generate a flux estimate (Q, g m⁻² s⁻¹). Additional data can be added, including PGSC, L, air temperature, atmospheric pressure, altitude and z₀, but it is not clear how much this changes the output.

Similar to the Gaussian Plume approach, WindTrax assumes constant emission rates and constant atmospheric conditions over space and time. WindTrax also assumes that the fetch has a maximum z₀ of 15 cm (low crops), a distance between the source and detector of less than 1 km, and the source is a homogeneous emitter with no emission hot spots. Again, it is likely that one or more of these assumptions is violated in the real world; however, WindTrax will generate a flux estimate if the detector is downwind of the source and the distance is less than 1 km. Fluxes from multiple sources can be calculated, providing there are enough sensors downwind (at least an equal number of sources and detectors). WindTrax has an advantage that batch simulations can be run through the graphical user interface, with data read in, simulations completed and fluxes written to an output file. To date, there have been no controlled release experiments to validate emission estimates for area sources.

Practical implications

As backward Lagrangian stochastic models have been developed to quantify area source emissions, they have been used extensively to calculate emissions from the agriculture, waste and natural emissions sectors [23,35,37,39,92,94]. These sources are typically aerodynamically flat with defined emission areas which help to reduce the uncertainty in emission estimates. As with the Gaussian Plume approach, the backward Lagrangian stochastic approach has difficulty with aerodynamic obstruction, but it does have the potential for further development, unlike the Gaussian approach, to account for air movement over larger obstructions.

3. Conclusions

3.1. Summary of Downwind Approaches

This entry details how to quantify methane emissions from either point or area sources using downwind measurements. This is a two-stage process where methane concentration and meteorological data must be measured downwind of a source first and then converted to an emission using an atmospheric dispersion model. Downwind methane concentrations will typically be tens of ppm and only four technologies can currently measure in this range: metal oxide sensors; non-dispersive infrared sensors; tunable diode laser absorption spectrometers; and optical cavity instruments. Typically, the lower the cost the of sensor, the less responsive to small changes in methane concentration and the more the sensor is affected by environmental conditions. However, the higher-precision instruments are more expensive and require mains power. The choice of methane measurement is typically determined by the size of emission source, location and budget of the project. Meteorological data are essential to quantify emissions, especially wind speed and directions. More meteorological instruments mean that more parameters can be measured (including micrometeorological variables); however, these are not always essential and the number of instruments used typically depends on the aims of the experiment.

In most cases, simple atmospheric dispersion approaches can be used to quantify both area and point emissions using downwind measurements. Emissions can be generated using limited data (methane concentration, wind speed, wind direction and locations) and confidence in calculated emission rates can be increased by using more input data. Site selection and location of instrument deployment are essential as quantification approaches assume a flat fetch (no aerodynamic obstructions) and constant wind fields.

3.2. Future of Downwind Quantification

Growing societal pressures on the energy, agricultural and waste industries mean that the verification of greenhouse gas emission estimates through measurement are likely to become more important in the coming years. For methane, the technology to quantify downwind concentrations is developed enough to generate representative emission estimates. The major current shortcoming in quantifying emissions is the atmospheric dispersion models that are being used. Even though emission estimates can be calculated, several assumptions crucial to the models’ operation, including an aerodynamically flat fetch, are being violated and result in uncertainties between +250% and −100% of the actual emission rate. While this may could be viewed as acceptable for academic purposes, this is unlikely to satisfy the needs of the Paris Agreement on methane intensity rules for natural gas production. Even though more complex modeling has been applied to the problem, including computational fluid dynamics, computational time is prohibitively long and unlikely to be adopted. At present, there is no happy medium between modeling complexity and computational time, and this remains the biggest challenge in the field of emission quantification.