Impact of atmospheric turbulence on the accuracy of point source emission estimates

Greenhouse gases have been extensively studied due to their key role in of Earths climate. Their anthropogenic fluxes are of particular interest for policies targeting the mitigation of climate change, as their long lifetimes, especially in case of most abundant CO2, will have an impact over several centuries. Monitoring of emissions is a critical part of climate mitigation, as without timely, accurate and precise information on the implementation progress, any potential diversions from the plan cannot be identified and acted upon in sufficient time. The scientific community has developed multiple observation-based emission estimation methods, among which application of space- and airborne state-of-the-art instrumentation hold much promise out thanks to their ability to provide relevant data on a global scale. Modern remote-sensing instruments have already demonstrated the ability to estimate emission from the strongest sources of greenhouse gases, like coal power plants, megacities or industrial sites. However, due to inherent technical difficulties as well as basic atmospheric physics, the accuracy of any single measurement is limited. Here, using the high-resolution atmospheric model WRF-GHG set over the largest point-like CO2 emitter in Europe, namely the Bechatw Power plant, we demonstrate how atmospheric dynamics limit the potential accuracy of emission estimation using the cross-sectional mass-flux method. We show how atmospheric turbulence affects the plume structure, and how that translates into emission estimates. We demonstrate how assumptions about well-mixedness can cause inaccuracies in emission estimates. Furthermore, through a novel application of temporally-tagged tracers, we also show that part of the CO2 plume variability is projected from the emission point across distances considerably longer than PBL turbulent scales, larger than was previously assumed. Unless the discussed effects can be taken into the account when planning, executing and interpreting measurements, the discussed effects can have potentially detrimental consequences for the accuracy of estimated flux values. It is worth noting that the presented results are of general nature and will affect attempts to quantify emissions of any pollutant for which similar estimation techniques are applied, including CO2 and CH4, NOx and others.