What I do


The main research interests of Claudio Carnevale are related to the implementation of nonlinear deterministic and nondeterministic models for air quality control. In particular, the activity can be split in 4 different fields:

Modelling and control of deterministic nonlinear systems

This field is devoted to the analysis and control of natural (air quality) nonlinear systems, through the modellisation, implementation and simulation of deterministic three-dimensional air quality models. The area is characterized by important challenges as 3D models have to describe complex, nonlinear and time-varying dynamics, with hundreds of state variable. Therefore, the study is focused on:

      1. State and input variable selection;

      2. Mathematical Modellisation of dynamical laws driving state and input variable, describing the physical and chemical phenomena involving pollutant in atmosphere.

      3. Study of the robustness and accuracy of the numerical schemes implemented for the solution of the partial differential equation (time-space) system.

      4. Definition and collection of the input data.

      5. Definition of benchmarking techniques.

The main objective of this research area is the implementation/upgrade, validation and application of the TCAM (Transport and Chemical Aerosol Model) 3D model and its integration in an air quality Decision Support System. The model allows the users to evaluate the impact of different emission control scenarios involving the main source activity and sector. The model has been applied and validated in a number of International Projects (CityDelta, Escompte, POMI) and it is one of the models suggested in the Model Documentation System (MDS) database (European Topic Centre on Air and Climate Change).

Online and offline Data Assimilation techniques for air quality applications

This research fulfills the requirements of the 2008/50 Directive, which allows member states and regional authorities to use a combination of measurement and modeling to monitor air pollution concentration. In particular, the aim of this research area is the study, development and implementation of the approaches allowing the integration of model data from (1) and pollutant levels measured by air quality stations and/or satellites. A range of techniques as the Kalman Filtering, Optimal Interpolation, Kriging and Co-Kriging methods and Inverse Distance Weighting (IDW) can be used to combine data from different sources to create spatial concentration fields. These techniques can be split on two categories: approaches that do not consider data source uncertainty (kriging, co-kriging, IDW) and approaches that explicitly take into account a description of statistical properties of model and measurement errors (Optimal Interpolation and Kalman Filter). A second classification concerns the way these techniques are integrated in the modeling systems. The easiest way to perform this integration is usually called “off-line” data assimilation (or data fusion). In this case, the assimilation procedure uses measurements and model output, and merges them to produce the final results representing the best estimate of (one or more) atmospheric features. Another set of approaches is the so-called “online” assimilation techniques, which actively interact with the model. These are usually used to improve initial conditions or model parameters, and include the variational methods of 3D- and 4D-var as well as ensemble methods such as Ensemble Kalman Filters.

Identification of nonlinear systems

The deterministic 3D models described in (1) are characterized by high computational costs, making unsuitable their integration in complex control/management problems. For this reason the development of computational efficient data-driven models can be studied. Such models can be identified starting from measured pollutants level time series or from the results of different scenario simulated by (1). The research area consists to the application of linear and nonlinear model (arx, time-varying arx, neural network, neuro-fuzzy systems) estimation techniques to solve two main problems:

      1. Concentration forecasting: in this case, the models are identified starting from measured pollutant time series and are used to forecast the dynamic of secondary pollutants (mainly PM10 and Ozone) and the occurrences of directive threshold exceedances. Usually these models are integrated with off-line data assimilation techniques in order to interpolate the pointwise forecast all over a domain.

      2. Surrogate model identification: in this case, the results of simulation performed by (1) are used to identify simplified, surrogate models describing a (partial and simplified) version of the relationship between emission source and air quality concentration over a certain domain in order to be integrated in optimal control problems (4).

Decisional models in multi-objective optimization problems

The development of Decisional Models is a key issue in the control and management of air quality systems. The key point is that the surrogate models identified in (3), can be used to define efficient emission reduction techniques through the development and solution of a multi-objective optimal control problem considering the impact of the decision on air quality, human and environmental health and socio-economics aspects related to them. The research activity is the core of a number of international project (APPRAISAL, Air Pollution Policies foR Assessment of Integrated Strategies At regional Local scales, FP7, 2012-2015; OPERA, the Operational Pollution Emission Reduction Assessment, LIFE+, 2010-2013; RIAT (the Regional Integrated Assessment Tool, Joint Research Centre EC founds, 2009-2010).

Research Projects

    • APPRAISAL (Air Pollution Policies foR Assessment of Integrated Strategies At regional Local scales, FP7 308395, 1.0MEuro, 2012-2015), WP Leader.

    • SINOPIAE (Sistema prototipale multi-sorgente integrante tecniche di osservazione multi-spettrale da satellite, aeromobile e a terra per il monitoraggio multi-scala della variazione di indicatori ambientali legata ai costituenti atmosferici e dispersione energetica, 31.000 Euro Bando Regione Lombardia 2012), University of Brescia Unit Leader.

    • OPERA (the Operational Pollution Emission Reduction Assessment, LIFE09 ENV/IT/092, 2.3MEuro, 2010-2013).

    • RIAT (the Regional Integrated Assessment Tool, funded by Joint Research Centre EC - 384364, 200KEuro, 2009-2010).

    • REMS (Rete lombarda di Eccellenza per la Meccanica Strumentale e laboratorio esteso, Bando Regione Lombardia, 1MEuro, 2011-2013).

    • QUITSAT (Qualità dell'aria mediante l'Integrazione di misure a Terra, da Satellite e di modellistica chimica multifase e di Trasporto, funded by Agenzia Spaziale Italiana, 4.5MEuro, 2006-2010), WP Leader.

    • PAR-TCAM (Parallelizzazione del modello di chimica e trasporto TCAM, per applicazioni di pianificazione della qualità dell'aria, Bando congiunto Regione Lombardia/CILEA – Iniziativa LISA), Project Leader.

    • POMI (the PO Valley Model Intercomparison, coordinato da Joint Research Centre, EC, 2009-2010).

    • CityDelta (An intercomparison of model reponses to urban emission scenarios, coordinato da Joint Research Centre, 2004-2007).