The POLARCAT Model Intercomparison Project (POLMIP): overview and evaluation with observations

A model intercomparison activity was inspired by the large suite of atmospheric chemistry observations made during the International Polar Year (2008) in the Arctic. Nine global and two regional chemical transport models have performed simulations for 2008 using a common emissions inventory to quantify the differences in model chemistry and transport schemes. This paper summarizes the models and compares their simulations of ozone and its precursors, and presents an evaluation of the simulations using a variety of surface, balloon, aircraft and satellite observations. Despite using the same emissions, large differences are seen among the models. Differences in a number of model parameters are identified as contributing to differences in the modelled chemical species, including cloud fields and photolysis rates. The largest differences among models, and between models and observations, are in NOy partitioning (PAN vs. HNO3) and in oxygenated volatile organic compounds (VOCs) such as acetaldehyde and acetone. Comparisons to surface site measurements of ethane and propane indicate that the emissions of these species are significantly underestimated. While limited in spatial and temporal coverage, the aircraft measurements provide a simultaneous evaluation of many species. Satellite observations of NO2 from OMI have been used to evaluate the models over source regions, indicating anthropogenic emissions are underestimated in East Asia, but fire emissions are generally overestimated. The emission factors for wildfires in Canada are evaluated using the correlations of VOCs to CO in the model output in comparison to enhancement factors derived from aircraft observations, showing reasonable agreement for methanol and acetaldehyde, but underestimate of ethanol, propane and acetone, while overestimating ethane emission factors.