Title: Operational model evaluation for particulate matter in Europe and North America in the context of AQMEII
Authors: SOLAZZO EFISIOBIANCONI RobertoPIROVANO GuidoVAUTARD R.BESSAGNET B.Chemel CharlesCOLL I.FERREIRA J.FORKEL R.FRANCIS X.v.MIRANDA A.i.MORAN M.d.NOPMONGCOL U.SARTELET K.n.Sokhi Ranjeet S.WERHAHN J.WOLKE R.YARWOOD G.ZHANG J.RAO S. TrivikramaGALMARINI StefanoVOLKER M.APPEL W.BRANDT J.CHRISTENSEN J.SILVER J.d.SCHAAP MartijnGROSSI P.HANSEN A.VIRA J.PRANK M.GRELL G.
Citation: ATMOSPHERIC ENVIRONMENT vol. 53 p. 75-92
Publisher: PERGAMON-ELSEVIER SCIENCE LTD
Publication Year: 2012
JRC N°: JRC64810
ISSN: 1352-2310
URI: http://adsabs.harvard.edu/abs/2012AtmEn..53...75S
http://publications.jrc.ec.europa.eu/repository/handle/JRC64810
DOI: 10.1016/j.atmosenv.2012.02.045
Type: Articles in Journals
Abstract: More than ten state-of-the-art regional air quality models have participated in the Air Quality Model Evaluation International Initiative (AQMEII), run by twenty independent groups in Europe and North America. Standardised modelling outputs from each group have been shared on the web distributed ENSEMBLE system, which allows statistical and ensemble analyses to be performed. In this study, the simulations issued from the models are inter-compared and evaluated with a large set of observations for ground level Ozone in both the continents. To facilitate the discussion and interpretation of the results, each continental domain is divided into four sub-regions and analyses focused on hourly data for the summer months of June, July and August. The analysis presented in this paper moves towards helping to gain more insight into ensemble modelling in terms of ensemble size and quality of the members. The analysis is based on systematically minimization of the model error and it is of direct relevance for diagnostic/probabilistic model valuation. We show that the most commonly used multi-model approach, that is the average over all available members, can be outperformed by subsets of members opportunely selected in terms of bias, error, and correlation. More importantly found this result not to depend strictly on the single member’s skills but requiring necessarily the presence of low ranking-skill members to produce. We apply a clustering methodology for discerning among members and build a skilful ensemble based on model associativity and data agglomeration, which makes no use of priori knowledge of model skill. Dendrogram representation has been used better understanding the level of repetition of the information produced by the models and to a certain extent their independence. Results show that, whilst this methodology needs further refinements, by opportunely selecting the cluster distance and association criteria, this approach can be of direct usefulness for applications not only strictly related to model evaluation, but also for example, to air quality forecasting.
JRC Institute:Institute for Environment and Sustainability

Files in This Item:
There are no files associated with this item.


Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.