The challenge of estimating regional trace gas emissions from atmospheric observations

This paper discusses some of the major issues that surround estimating regional emissions of trace gases from atmospheric observations through inversion modelling. Inversion methods use modelled knowledge of how emissions dilute in the atmosphere as they travel from their source to an observation point, together with the observations, to calculate a grid of emissions. The problem is one of minimizing the mismatch between a modelled and observed time series of concentration. There are many methods of comparing time series, some involving a priori knowledge others without. The location, terrain and height of the observation station can also be very significant in determining how well a model can represent the dilution from emission source to receptor. The inversion solution (emission map) will assign some of the sources incorrectly for a variety of reasons, e.g. local sources, intermittent releases, errors in the modelled transport or observation, and the choice of the spatial and temporal resolution of the emission map. The reasons for uncertainty in the modelled emissions are discussed along with suggestions as to how some of these can be minimized. Using multiple stations to further constrain the inversion should reduce the uncertainty; however, care is needed if the potential improvements are to be realized.