Compilation of a database on the composition of anthropogenic VOC emissions for atmospheric modeling in Europe

To analyse and generate air pollution control strategies and policies, e.g. efficient abatement strategies or action plans that lead to a fulfilment of air quality aims, atmospheric dispersion models (CTMs) have to be used. These models include a chemical model, where the numerous volatile organic compounds (VOCs) species are lumped together in classes. On the other hand, emission inventories usually report only total non-methane VOC (NMVOC), but not a subdivision into these classes. Thus, VOC species profiles are needed that resolve total NMVOC emission data. The objective of this publication is to present the results of a compilation of VOC species profiles that dissolve total VOC into single-species profiles for all relevant anthropogenic emission source categories and the European situation. As in atmospheric dispersion models usually modules for generating biogenic emissions are directly included, only anthropogenic emissions are addressed. VOC species profiles for 87 emission source categories have been developed. The underlying data base can be used to generate the data for all chemical mechanisms. The species profiles have been generated using recent measurements and studies on VOC species resolution and thus represent the current state of knowledge in this area. The results can be used to create input data for atmospheric dispersion models in Europe.The profiles, especially those for solvent use, still show large uncertainties. There is still an enormous need for further measurements to achieve an improved species resolution. In addition, the solvent use directive and the DECOPAINT directive of the European Commission will result in a change of the composition of paints; more water-based and high-solid paints will be used; thus the species resolution will change drastically in the next years. Of course, the species resolution for combustion and production processes also requires further improvement.