Annual upscaling of methane emission field measurements from two Danish landfills, using empirical emission models

Abstract

An empirical model was developed and employed to estimate annual methane (CH4) emissions from two Danish landfills (Skellingsted and AV Miljø). The overall aim was to provide accurate annual CH4 emission estimates based on discrete emission field measurements and to address temporal variability caused by the impact of barometric pressure. Four non-linear regression models were developed, corresponding to the two landfills as well as to the western and eastern waste sections of AV Miljø. A comparison of model predictions with on-site eddy covariance fluxes showed that the models can accurately predict short-term emission variability. Predicted annual CH4 emissions for the Skellingsted and AV Miljø landfills were 69 ± 4 and 80 ± 4 tonnes, respectively, whereas for the western and eastern sections of the AV Miljø landfill, emissions were estimated at 63 ± 3 and 19 ± 1 tonnes, respectively. The results demonstrate that even though maximum emissions from Skellingsted were approximately threefold compared to AV Miljø, annual predicted CH4 emissions for Skellingsted were lower. This was because during the most frequently occurring pressure change events, emission rates were higher at AV Miljø in comparison to Skellingsted. An optimised sampling strategy was proposed, targeting the determination of an empirical emission model though the effective use of discrete field measurements. Analysis of annual emission estimates, based on the number of the tracer dispersion method (TDM) measurements, showed that both the number as well as the distribution of performed TDM measurements across the range of expected dP/dt influence the uncertainty.