Emission verification from national to local levels using top-down approaches: sampling strategy and modeling performances

The actual deployment of GHG mixing ratio observation networks near and in urbanized areas introduces several key issues for atmospheric transport models. Whereas data density increases with model resolution, the co-localization problem between fluxes and atmospheric observations, the impact of the urban landscape on the local dynamics, the boundary inflow in limited domain simulations, and the level of accuracy to capture correctly the structures of emission signals (e.g., narrow urban plumes mixed with natural vegetation fluxes), become increasingly important at very high resolutions. We present here several case studies from national to local levels, in the context of rapidly changing urban landscapes, isolated emission sources, and heavily populated areas, comparing model performances with surface stations and satellite measurements. The different simulations aim at verifying emissions at different scales, using single tracer simulations of atmospheric carbon dioxide, or multiple species as carbon monoxide, methane, and emission component decomposition (traffic, residential, point sources). We provide first order analysis of model performances at these different scales, in the context of natural fluxes and changing inflow conditions, and sampling strategies to characterize the emissions at national and local levels. Transport error characterization is investigated for both column-integrated and surface measurements.