More than ten state-of-the-art regional air quality models have been applied as part of the Air Quality Model Evaluation International Initiative (AQMEII). These models were run by twenty independent groups in Europe and North America. Standardised modelling outputs over a full year (2006) from each group have been shared on the web-distributed ENSEMBLE system, which allows for statistical and ensemble analyses to be performed by each group. The estimated ground-level ozone mixing ratios from the models are collectively examined in an ensemble fashion and evaluated against a large set of observations from both continents. The scale of the exercise is unprecedented and offers a unique opportunity to investigate methodologies for generating skilful ensembles of regional air quality models outputs. Despite the remarkable progress of ensemble air quality modelling over the past decade, there are still outstanding questions regarding this technique. Among them, what is the best and most beneficial way to build an ensemble of members? And how should the optimum size of the ensemble be determined in order to capture data variability as well as keeping the error low? These questions are addressed here by looking at optimal ensemble size and quality of the members. The analysis carried out is based on systematic minimization of the model error and is important for performing diagnostic/probabilistic model evaluation. It is shown that the most commonly used multi-model approach, namely the average over all available members, can be outperformed by subsets of members optimally selected in terms of bias, error, and correlation. More importantly, this result does not strictly depend on the skill of the individual members, but may require the inclusion of low-ranking skill-score members. A clustering methodology is applied to discern among members and to build a skilful ensemble based on model association and data clustering, which makes no use of priori knowledge of model skill. Results show that, while the methodology needs further refinement, by optimally selecting the cluster distance and association criteria, this approach can be useful for model applications beyond those strictly related to model evaluation, such as air quality forecasting

Abstract. Vertical aerosol profiles were directly measured over the city of Milan during three years (2005)(2006)(2007)(2008) of field campaigns. An optical particle counter, a portable meteorological station and a miniaturized cascade impactor were deployed on a tethered balloon. More than 300 vertical profiles were measured, both in winter and summer, mainly in conditions of clear, dry skies.The mixing height was determined from the observed vertical aerosol concentration gradient, and from potential temperature and relative humidity profiles. Results show that inter-consistent mixing heights can be retrieved highlighting good correlations between particle dispersion in the atmosphere and meteorological parameters. Mixing height growth speed was calculated for both winter and summer showing the low potential atmospheric dispersion in winter.Aerosol number size distribution and chemical composition profiles allowed us to investigate particle behaviour along height. Aerosol measurements showed changes in size distribution according to mixing height. Coarse particle profiles (d p >1.6 µm) were distributed differently than the fine ones (d p <1.6 µm) were, at different heights of the mixing layer. The sedimentation process influenced the coarse particle profiles, and led to a reduction in mean particle diameter for those particles observed by comparing data above the mixing height with ground data (−14.9±0.6% in winter and −10.7±1.0% in summer). Conversely, the mean particle diameter of fine particles increased above the mixing height under stable atmospheric conditions; the average inCorrespondence to: L. Ferrero (luca.ferrero@unimib.it) crease, observed by comparing data above the mixing height with ground data, was +2.1±0.1% in winter and +3.9±0.3% in summer. A hierarchical statistical model was created to describe the changes in the size distribution of fine particles along height. The proposed model can be used to estimate the typical vertical profile characterising launches within prespecified groups starting from: aerosol size and meteorological conditions measured at ground-level, and a mixing height estimation. The average increase of fine particle diameter, estimated on the basis of the model, was +1.9±0.5% in winter and +6.1±1.2% in summer, in keeping with experimental findings.

The Kuroshio Extension (KE) flow in the North Pacific Ocean displays a very distinctive decadal variability of bimodal character involving two completely different states (a large-meander ''elongated'' state and a small-meander ''contracted'' state) connected by very asymmetric temporal transitions. Although such a flow has been widely studied by means of a suite of mathematical models and by using several observational platforms, a satisfactory theoretical framework answering quite elementary questions is still lacking, the main question being whether such variability is induced by a time-varying wind forcing or, rather, by intrinsic oceanic mechanisms. In this context, the chaotic relaxation oscillation produced by a process-oriented model of the KE low-frequency variability, with steady climatological wind forcing, was recently recognized to be in substantial agreement with altimeter data. Here those model results are further compared with a comprehensive altimeter dataset. The positive result of such a comparison allows the conclusion that a minimal model for the KE bimodality has been identified and that, consequently, nonlinear intrinsic oceanic mechanisms are likely to be the main cause of the observed variability. By applying the methods of nonlinear dynamical systems theory, relevant dynamical features of the modeled flow are then explained, such as the origin of the relaxation oscillation as a consequence of a homoclinic bifurcation, the spatiotemporal character of the bimodal behavior, and the degree of predictability of the flow in the different stages of the oscillation (evaluated through a field of finite-time Lyapunov exponents and the corresponding Lagrangian time series).

This work tackles the problem of the automated detection of the atmospheric boundary layer (BL) height h, from aerosol lidar/ceilometer observations. A new method, the Bayesian selective method (BSM), is presented. It implements a Bayesian statistical inference procedure which combines in a statistically optimal way different sources of information. Firstly, atmospheric stratification boundaries are located from discontinuities in the ceilometer backscattered signal. The BSM then identifies the discontinuity edge that has the highest probability to effectively mark the BL height. Information from the contemporaneous physical boundary layer model simulations and a climatological dataset of BL height evolution are combined in the assimilation framework to assist this choice.The BSM algorithm has been tested for 4 months of continuous ceilometer measurements collected during the BASE:ALFA project, and is shown to realistically diagnose the BL depth evolution in many different weather conditions. A standard one-dimensional processing of the ceilometer signal without the a priori support of the dynamical and climatological BL models often fails to correctly detect h, with the greatest inaccuracies occurring at night-time when residual layers can generate very strong signals, which are then classified by an automated application of the gradient or of the wavelet analysis as the most probable BL height. The BSM approach instead carries information on the low climatological probability to find elevated BL depths at night and penalizes the selection of these points. Moreover, this method is able to correctly convey information along the temporal dimension, thus filling data gaps using earlier and subsequent ceilometer information for the retrieval.

Abstract. In this paper we present an approach for the statistical analysis of multi-model ensemble results. The models considered here are operational long-range transport and dispersion models, also used for the real-time simulation of pollutant dispersion or the accidental release of radioactive nuclides.We first introduce the theoretical basis (with its roots sinking into the Bayes theorem) and then apply this approach to the analysis of model results obtained during the ETEX-1 exercise. We recover some interesting results, supporting the heuristic approach called "median model", originally introduced in Galmarini et al. (2004a, b).This approach also provides a way to systematically reduce (and quantify) model uncertainties, thus supporting the decision-making process and/or regulatory-purpose activities in a very effective manner.