A full year evaluation of the CALIOPE-EU air quality modeling system over Europe for 2004

作者：Pay M T; Piot M; Jorba O; Gasso S; Goncalves M; Basart S; Dabdub D; Jimenez Guerrero P; Baldasano J M^*

来源：Atmospheric Environment, 2010, 44(27): 3322-3342.

DOI：10.1016/j.atmosenv.2010.05.040

• Air quality
• Model evaluation
• Europe
• High resolution
• Ozone
• Particulate matter

The CALIOPE-EU high-resolution air quality modeling system, namely WRF-ARW/HERMES-EMEP/CMAQ/BSC-DREAM8b, is developed and applied to Europe (12 km x 12 km, 1 h). The model performances are tested in terms of air quality levels and dynamics reproducibility on a yearly basis. The present work describes a quantitative evaluation of gas phase species (O(3), NO(2) and SO(2)) and particulate matter (PM2.5 and PM10) against ground-based measurements from the EMEP (European Monitoring and Evaluation Programme) network for the year 2004. The evaluation is based on statistics. Simulated O(3) achieves satisfactory performances for both daily mean and daily maximum concentrations, especially in summer, with annual mean correlations of 0.66 and 0.69, respectively. Mean normalized errors are comprised within the recommendations proposed by the United States Environmental Protection Agency (US-EPA). The general trends and daily variations of primary pollutants (NO(2) and SO(2)) are satisfactory. Daily mean concentrations of NO(2) correlate well with observations (annual correlation r = 0.67) but tend to be underestimated. For SO(2), mean concentrations are well simulated (mean bias = 0.5 mu g m(-3)) with relatively high annual mean correlation (r = 0.60), although peaks are generally overestimated. The dynamics of PM2.5 and PM10 is well reproduced (0.49 < r < 0.62), but mean concentrations remain systematically underestimated. Deficiencies in particulate matter source characterization are discussed. Also, the spatially distributed statistics and the general patterns for each pollutant over Europe are examined. The model performances are compared with other European studies. While O(3) statistics generally remain lower than those obtained by the other considered studies, statistics for NO(2), SO(2), PM2.5 and PM10 present higher scores than most models.

• 出版日期2010-9