Comparison of northern hemispheric anthropogenic black carbon emissions from global datasets

There is a large interest in the impact of short-lived climate pollutants (SLCPs) for the Arctic climate but the performance of emission inventories that serve as input to climate models remains unknown especially in high latitudes. To assess emissions that are expected to have a more direct impact on the Arctic, a comparison of available SLCP emission inventories was conducted utilizing spatially-distributed global emission datasets downloaded from the ECCAD-GEIA website (http:// eccad. sedoo. fr). In this paper, the comparison was done for black carbon emissions. Differences in both emissions and their locations were addressed. There remains large variation between the emission inventories in northern latitudes. Relatively speaking the variability is larger than at the global level. Total emissions at high latitudes tend to be lower which makes them more sensitive to uncertainties in regionally important source sectors than at the global level. Variations within the sector emission estimates arise most likely from uncertainties in key parameters, i.e. activities and emission factors. The accuracy of the parameters needs further development. However, the differences were unsystematic, so this was not enough to explain the variation. Some of the variation is due to differences in inclusion of relevant source sectors and spatial distributions of the emission data. Notably flaring was included to the full extent only in some inventories, although the emissions are significant in the Arctic region. Another sector omitted in some inventories was international maritime transport. Inclusion of relevant emission sectors is a common improvement suggestion for all models. Another aspect affecting the quality of emission inventories is the location of the emissions. The spatial aspect is especially important in the case of black carbon, since its life-time is relatively short and, therefore, the concentration around the sources is higher. This is highlighted in the Arctic area in particular, since black carbon reduces albedo and thus enhances melting when deposited on snow or ice. There were significant differences between the spatial distributions of the black carbon emissions in the inventories, often with no spatial agreement at all. The differences also varied between source sectors, being sometimes mostly systematic, unsystematic or both. Uncertainty in the spatial distribution of the emissions potentially increases the uncertainty of impact modelling. The differences indicated that the inventories use different spatial proxies for the emissions. One way to develop the quality of the spatial distribution would be to incorporate more data from national or regional emission inventories or models into the global inventories, provided their quality is sufficient.