Abstract. The evaluation of regional air quality models is a challenging task, not only for the intrinsic complexity of the topic but also in view of the difficulties in finding sufficiently abundant, harmonized and time/space-well-distributed measurement data. This study, conducted in the framework of AQMEII (Air Quality Model Evaluation International Initiative), evaluates 4-D model predictions obtained from 15 modelling groups and relating to the air quality of the full year of 2006 over the North American and European continents. The modelled variables are ozone, CO, wind speed and direction, temperature, and relative humidity. Model evaluation is supported by the high quality in-flight measurements collected by instrumented commercial aircrafts in the context of the MOZAIC programme. The models are evaluated at five selected domains positioned around major airports, four in North America (Portland, Philadelphia, Atlanta, Dallas) and one in Europe (Frankfurt). Due to the extraordinary scale of the exercise (number of models and variables, spatial and temporal extent), this study is primarily aimed at illustrating the potential for using MOZAIC data for regional-scale evaluation and the capabilities of models to simulate concentration and meteorological fields in the vertical rather than just at the ground. We apply various approaches, metrics, and methods to analyze this complex dataset. Results of the investigation indicate that, while the observed meteorological fields are modelled with some success, modelling CO in and above the boundary layer remains a challenge and modelling ozone also has room for significant improvement. We note, however, that the high sensitivity of models to height, season, location, and metric makes the results rather difficult to interpret and to generalize. With this work, though, we set the stage for future process-oriented and in-depth diagnostic analyses.
Constraining the uncertainty in emissions over India with a regional air quality model evaluation
