Comparison of forward and backward Lagrangian transport modelling to determine methane emissions from anaerobic digestion facilities

Abstract

In biogas plants, where biological treatment of organic matter by anaerobic digestion (AD) is performed, as well as in plants for upgrading of biogas to biomethane, there might be emissions to air from different parts of the plants. Precise and comparable methods to quantify methane (CH 4 ) emissions from biogas plants are necessary to evaluate mitigation strategies and to generate more accurate emissions factors for national emission inventories in the context of the United Nations Framework Convention on Climate Change (UNFCCC) reporting. The main objective of this study was to provide recommendations for the measurement and modelling procedures of inverse dispersion modelling (IDM) to assure comparability of CH 4 emissions among diverse approaches and AD facilities. Two Lagrangian stochastic dispersion models were used operating in forward (LASAT) and backward (WindTrax) mode based on open-path concentration and meteorological data acquired by three international measurement teams at two biogas plants. The sensitivity analysis of the modelling systems regarding input parameters and setups (e.g. source configuration, terrain, sampling distance from the site, different combinations of wind data) showed that in WindTrax (backward mode) the measurement speed of meteorological parameters (via three-dimensional ultrasonic anemometer) had the largest influence on the calculated emission estimates, while in LASAT (forward mode) the assumption of the source configuration greatly affected the flux estimates. The used concentration data (e.g. location of observations, measurement mode of background concentration) was critically important to the success of IDM. The CH 4 emission rates from the AD facility in flat terrain (plant 1) calculated with both Lagrangian models by IDM showed good agreement with the DIAL (Differential Absorption Light Detection And Ranging) method and methane release tests (deviation in the range of 10–20%) when using best fit configurations (e.g. volume source in LASAT, area source in WindTrax, optimal measurement fetch). In moderately complex terrain (plant 2), the retrieval rates of controlled CH 4 releases, and thus the certainty of flux estimates of both dispersion models tended to decrease without the use of a terrain model. • First comparison of emission results inferred with LASAT and WindTrax. • Concentration and meteorological data obtained by three independent measurement teams. • Evaluation of sensitivity of the modelling system to input parameters and setups. • Recommendations for measurement and modelling procedure. • Uncertainty of ±10–20% for both models in flat terrains using best fit configurations.