Quantification of methane emissions from industrial sites using statistical inversion and a local scale atmospheric transport model parameterized with tracer data

Different methods based on atmospheric measurements and atmospheric transport modeling have been developed and used for the quantification of methane emissions from industrial sites. The principle is to use a tracer or a local scale transport model as a proxy of the atmospheric transport which gives the link between the methane emissions and the measured concentrations, and which can thus be used to derive (through an inversion of the transport) of the estimate of the emissions that is consistent with the measured concentrations.

The tracer release method consists in releasing a tracer gas near a methane source at a well-known rate and measuring both the tracer and methane concentrations in the plumes associated with these emissions. The ratio between measured tracer concentrations and known tracer emissions are then used to convert the measured methane concentrations into methane emission estimates. This method can be hampered by the difficulty to co-locate the tracer and methane source, if the latter is hardly accessible or if it is spread over a significant area, or when investigating a site with multiple sources. This can result into large errors in the estimate of the methane emissions. On the other hand, local scale atmospheric transport model can take into account the spread and multiplicity of sources but they can bear large modeling errors.

We thus present a new concept for the estimation of site scale methane emissions, which combines the tracer release technique, local scale transport modeling and a statistical inversion framework. The information on the atmospheric transport from the tracer is used to control the parameters of the transport model and to derive the statistics of its errors. The model and these statistics are then used as a basis of a statistical atmospheric inversion of the different methane sources within a site using the methane concentration measurements, which accounts for the data and model errors.

In practice, we have developed an inversion system based on Gaussian model whose class of stability is controlled based on the tracer data, and based on a measurement protocol in which acetylene is used as the tracer, and in which concentration are measured along transects across the emission plumes. We apply it and the more traditional tracer release method for a range of experiments. These experiments include an academic test with controlled release of methane and acetylene to test the robustness of the method, and which demonstrates its potential to provide better results than the tracer release technique. They also include numerous campaigns on industrial sites, and in particular on waste water treatment plants and landfills. The presentation of the results will focus on both the estimates of total emissions from the sites but also on the analysis of the ability to quantify individual sources within the sites when using our statistical inversion approach