Evidence: Validation of landfill methane measurements from an unmanned aerial system: Project SC160006

Abstract

Methane is an important greenhouse gas and emission controls for methane are a part of the international Paris Agreement and UK national strategies to reduce greenhouse gas emissions.
Landfill gas is mainly composed of methane and carbon dioxide, in near equal measure. While modern UK landfills capture and use much of the methane gas produced, some methane is emitted to the atmosphere. However, the precise amount of methane arising from UK landfills remains highly uncertain. The work described in this report represents a new method for precisely quantifying landfill methane emissions that has the potential to be widely used.
A feasibility study commissioned by the Environment Agency in 2013 identified that the use of unmanned aerial systems (UAS) to quantify methane emissions from landfills was a viable new measurement approach. A field trial at a UK landfill in 2015 demonstrated that it was possible to derive an emission flux of methane with a known uncertainty using in situ UAS-mounted instrumentation.
This report presents the results of a subsequent validation field trial of the UAS technology and flux-calculation approach. The aim of the field trial was to release controlled fluxes of methane gas in order to test how well the UAS approach evaluated the controlled flux.
Methane fluxes were emitted at a rate below that typically expected of UK landfills: the maximum emission rate of methane was just over 10 kg/h. This emission rate allowed the system to be tested and validated at the lowest limit of sensitivity needed for UK landfills and allowed for the characterisation of flux uncertainty to be improved in order to inform future operational use of the method.
The validation field trial took place at the UK Meteorological Office site in Cardington, Bedfordshire, UK, between 31 October and 4 November 2016.
A total of seven UAS flights were analysed. These sampled methane concentrations from a UAS downwind of the controlled emission source. The calculations of the methane fluxes were conducted without knowing the emission rate of the controlled source.
The UAS validation experiments successfully characterised the methane releases. The measured methane flux, taking into account the measurement uncertainty, always overlapped with the controlled methane emission rate. For the 7 flights, the mean percentage difference between the measured and emitted methane flux was an overestimate of 6% with a mean absolute difference an overestimate of 0.4 kg/h.
Other experiments undertaken as part of the field trial have demonstrated that the method can also detect very small methane fluxes (down to 0.15 kg/h) with comparable relative uncertainties to those calculated for flux rates over an order of magnitude higher. These small flux rates are similar in magnitude to the small point source emissions that may be expected from fugitive emissions in natural gas infrastructure. The method developed here may therefore have significant utility in the monitoring and measurement of fugitive methane emission flux rates from other UK industrial infrastructure.
The following conclusions have been reached regarding the expected performance of the method:
 The flux derived using mass balancing can be considered to be accurate to within one standard deviation, when all sources of variability and error are known or measured
 Repeated flights (or increased sampling time) can significantly reduce the uncertainty in the measured methane flux
 Sampling when the wind speeds, wind directions, and background concentrations are constant would lead to reduced uncertainty
 Further improvements to the accuracy of flux calculation could be made by appropriate measurement of wind speed and direction on board the UAS platform
5 of 80
 A nearby wind measurement on an elevated tower (preferably at 10 m above local ground level) remains a good substitute as long as the tower is placed in an environment representative of the intended UAS sampling
The future use of the UAS mass balance approach should always consider the following:
 Appropriate zoning of downwind areas to ensure that the sampling captures the landfill plume
 The positions of obstacles to air flow (for example any buildings) and site topography between the site and measurement location should be noted and considered when planning UAS sampling to optimise the sampling zone
 The locations of any other nearby methane emission sources must be noted. Ideally, these should not be upwind of the site of interest as this would affect background variability and could result in much larger systematic errors. If this is unavoidable, additional care may be needed to ensure good background measurements are recorded to better remove the extraneous source
 When establishing the regular grid pattern for sampling across the flux plane, the appropriate size of the cells in the grid should be defined by the instrument response rate and the wind speed
 The randomised sampling in the flux plane must ensure that at least 50 independent methane concentration measurements are taken within each individual grid cell
 Sampling in non-stagnant wind speeds (greater than around 2 metres per second) to reduce flux uncertainty (with the upper wind speed limit defined by the safe operating conditions of the UAS – around 10 metres per second)
In parallel with the UAS measurements, complementary measurements of the known methane releases were undertaken using a tracer gas dispersion method. This method is based on the assumption that a tracer gas released at a methane emission source will disperse in the atmosphere in the same way as the emitted methane. Assuming the air is well mixed, the methane emission rate can be calculated as a function of the ratio of the downwind measurements of the methane and tracer gas concentrations.
Using a constant release of an acetylene tracer, two separated teams undertook a total of 132 downwind plume measurement sets over five methane releases. The methane fluxes measured by the different teams were comparable and within experimental error. The tracer gas dispersion method was able to determine actual release rates to within the measurement uncertainties for all the tests other than one. For that test, the difference between the actual and measured methane rates was only 1kg/hour.
Both methane measurement techniques were successful in matching the known methane releases. The UAS and the tracer gas dispersion method have different operational constraints so together they represent options that allow methane emissions from landfills and other facilities to be quantified within a known level of uncertainty.

More information

Identifiers

ORCID for Tristan Rees-White: orcid.org/0000-0001-9009-8432

ORCID for Richard Beaven: orcid.org/0000-0002-1387-8299

Catalogue record

Export record