Mihaela Mircea 1, Massimo D'Isidoro 1, Alberto Maurizi 1, Maria Gabriella Villani 1, Andrea Buzzi 1, Sandro Fuzzi 1, Francesco Tampieri 1 Gabriele Zanini 2, Fabio Monforti 2, Lina Vitale 2 1 Istituto di Scienze dell’Atmosfera e del Clima, CNR, Bologna, Italy 2 ENEA,Italian Agency for New Technologies, Energy and the Environment, Via Martiri di Monte Sole 4, Bologna, Italy Ozone modeling over Italy: a sensitivity analysis to precursors using BOLCHEM air quality model

Objectives to investigate the sensitivity of ozone concentration to the reduction of NOx and VOC for few periods during the years 1999 and 2003 over the whole Italy to asses the relative importance of precursors in reducing the ozone levels identifying the regions of Italy where local emissions strategies could not be effective

Motivation  there is no such study – only studies over Northern Italy  in some areas, the ozone concentrations can be controlled by transboundary processes and that it cannot be ruled out by simulating a small area.  the topography of Italy is very complex and leads to very complicate circulations features.  Italy is in a geographic position (from ca. 37 o to 47 o lat North) that lead to various climate features.  The ozone episodes to simulate were chosen for the year 1999 because it is used as a reference year in meteorological studies and for the summer 2003 because it was characterized by very high temperatures for a many days.

BOLCHEM flow chart Gas Chemistry (SAPRC90/CB4) EmissionsDry Deposition Meteorological Model (BOLAM) Winds, T, q Transport/dispersion Chemistry&transport model (CTM)

Photochemical mechanisms CB-IV (Gery et al., 1989) : lumped-structure condensed mechanism -85 reactions and 30 chemical species -organics are grouped according to bond type (for example, as carbon single bonds, carbon double bounds or carbonyl bounds) -organic species are treated explicitly (e.g. formaldehyde, ethene, isoprene), represented by carbon bond (PAR – single bonded one carbon atom, OLE –two carbon atoms) or molecular (TOL and XYL aromatic hydrocarbons) surrogates according with their chemistry or importance in the environment. SAPRC90 (Carter, 1990) : lumped-molecular condensed mechanism -131 reactions with 35 chemical species -calculate the kinetic and mechanistic parameters for lumped species in the mechanism created for representative emissions profile (mole-weighted approach) -organics species are treated explicitly (e.g. formaldehyde, acetaldehyde, etc) or represented by molecules as alkane, alkenes, aromatics, etc.

Chemistry is coupled “online” with meteorology Meteorology and chemistry use:  same transport scheme (WAF (Weighted Average Flux) 3-d advection scheme)  same grid: horizontal and vertical components (the vertical coordinate system is terrain-following (  ), with variables distributed on a non-uniformly spaced staggered Lorenz grid. the horizontal discretization uses geographical coordinates on an Arakawa C-grid)  same physics for the subgrid-scale transport (for example vertical diffusion in surface layer and PBL parameterization depend on the Richardson number)  same time step

Emissions, initial and boundary conditions The chemical fields are driven by hourly surface emissions and 3 hourly lateral boundary conditions after the initialisation. Emissions, initial and boundary conditions are produced by Thoscane model (Zanini et al., 2004). The emission inventory includes the ship emissions and the point source emissions are also considered in the simulations.

Model configuration and meteorological inputs  The model domain extends between: NW (20.77, 47.55); NE ( ); SW ( ); SE (19.42 – 35.79).  The horizontal resolution used in the simulations is of 20 km. The vertical resolution includes 33 sigma vertical layers from surface to the tropopause. The lower layer is approximately 20m thick above the surface.  The meteorological fields are supplied by ECMWF. The lateral boundary conditions are updated every 6 hours. The weather fields were re-initialized every 48 hours with the analyses in order to avoid an excessive error growth in the meteorological forecast.

Evaluation of model performance The model runs are on an grid between fine and coarse, therefore the model results are only compared with observations at rural and semi-rural locations. 1999: 4 clear sky periods selected based on Meteosat Images of Europe: January: June: 1-4 July: 1-5 August: 5-8

Time series of observed and predicted O3 mixing ratio of gases at sites (1) Atmospheric Chemistry

Time series of observed and predicted O3 mixing ratio of gases at sites (2)

Quantitative performance statistics for hourly concentrations of O3

NOx reduced(-35%) VOC reduced(-35%) Base case

Differences in ozone concentrations:  O3=O3(65%VOC)-O3(65%NOx) CB4  O3 > 0 NOx limited area  O3 <0 VOC limited area SAPRC90 ppb Chemical regimes over Italy ppb

Areas selected for the analysis Center of the area LatLonSize (km X km) Milano 45 o 28’9 o 10’ 160x160 Genova 44 o 25’8 o 54’ 80x80 Venezia 45 o 26’12 o 19’ 240x240 Roma 41 o 54’12 o 28’ 80x80 Napoli 40 o 51’14 o 16’ 80x80 Taranto 40 o 25’17 o 14’ 80x80 Pachino 36 o 15’15 o 05’ 100x100

NOx reduced(-35%) VOC reduced(-35%)  O3=local-global reductions  O3(ppb) CB4 SAPRC90

Extreme summer 2003  O3=O3(65%VOC)-O3(65%NOx)

Fraction of VOC or NOx limited area

O3 increase due to the increase in isoprene emissions ppb

Preliminary conclusions  The differences in the predicted ozone concentrations due to the photochemical mechanisms are comparable to those obtained by reducing the emissions of NOx or VOC  The distribution and the “intensity” (differences in ozone concentration) of VOC or NOx limited areas depend on the photochemical mechanism. For example, in the same meteorological and environmental conditions, a region can be VOC or NOx sensitive according with the photochemical mechanism used.  The local reduction of VOC was efficient for Milano and Venice areas. In the other regions, significant increase in ozone concentration was observed by reducing locally both the NOx or VOC emissions.  The increase of isoprene leads to substantial increase in the concentration of ozone at some locations (up to 25%), therefore, the uncertainties in isoprene emissions can bias the air quality design.

Atmospheric Chemistry – ISAC (CNR) ACKNOWLEDGEMENTS  GEMS and ACCENT: EU-projects  ICTP Programme for Training and Research in Italian Laboratories”, Trieste, Italy