Atmospheric Chemistry and Transport Modelling Introduction and current

Atmospheric Chemistry and Transport Modelling: Introduction and current activities at CETEMPS (also involving satellite data) Gabriele Curci CETEMPS – Dip. Fisica, Università degli Studi dell’Aquila gabriele. curci@aquila. infn. it 28 Jan. 2010 Università Tor Vergata, Roma Dipartimento di Informatica, Sistemi e Produzione

OUR MULTI-DISCIPLINARY INTERACTIVE CLIMATE SYSTEM: A REMARKABLE “Play. Station” FOR SCIENTISTS! [IPCC, 2007]

Understanding of atmospheric composition is key to understanding of climate change … [IPCC, 2007]

Year 2007 … air quality, and much more! [EEA, 2008]

HOW TO MODEL ATMOSPHERIC COMPOSITION? Solve continuity equation for chemical mixing ratios Ci(x, t) U = wind vector Eulerian form: Pi = local source Lightning of chemical i Lagrangian form: Transport Li = local sink Chemistry Aerosol microphysics Deposition Volcanoes Fires Human Land biosphere activity Ocean [adapted from D. J. Jacob, Harvard]

EULERIAN MODELS PARTITION ATMOSPHERIC DOMAIN INTO GRIDBOXES This discretizes the continuity equation in space Solve continuity equation for individual gridboxes • Detailed chemical/aerosol models can presently afford -106 gridboxes • In global models, this implies a horizontal resolution of ~ 1 o (~100 km) in horizontal and ~ 1 km in vertical • Chemical Transport Models (CTMs) use external meteorological data as input • General Circulation Models (GCMs) compute their own meteorological fields [D. J. Jacob, Harvard]

OPERATOR SPLITTING IN EULERIAN MODELS Reduces dimensionality of problem • Split the continuity equation into contributions from transport and local terms: … and integrate each process separately over discrete time steps: These operators can be split further: • split transport into 1 -D advective and turbulent transport for x, y, z (usually necessary) • split local into chemistry, emissions, deposition (usually not necessary) [D. J. Jacob, Harvard]

SPLITTING THE TRANSPORT OPERATOR Must account for sub-grid turbulence • Wind velocity U has turbulent fluctuations over time step Dt: Time-averaged component (resolved) Fluctuating component (stochastic) • Split transport into advection (mean wind) and turbulent components: advection turbulence (1 st-order closure) • Further split transport in x, y, and z to reduce dimensionality. In x direction: advection operator turbulent operator [D. J. Jacob, Harvard]

VERTICAL TURBULENT TRANSPORT (BUOYANCY) • generally dominates over mean vertical advection • K-diffusion OK for dry convection in boundary layer (small eddies) • Deeper (wet) convection requires non-local convective parameterization Convective cloud (0. 1 -100 km) detrainment Model vertical levels updraft downdraft entrainment Model grid scale Wet convection is subgrid scale in global models and must be treated as a vertical mass exchange separate from transport by grid-scale winds. Need info on convective mass fluxes from the model meteorological driver. [D. J. Jacob, Harvard]

LOCAL (CHEMISTRY) OPERATOR: solves ODE system for n interacting species For each species System is typically “stiff” (lifetimes range over many orders of magnitude) → implicit solution method is necessary. • Simplest method: backward Euler. Transform into system of n algebraic equations with n unknowns Solve e. g. , by Newton’s method. Backward Euler is stable, mass-conserving, flexible (can use other constraints such as steady-state, chemical family closure, etc… in lieu of DC/Dt ). But it is expensive. Most 3 -D models use higher-order implicit schemes such as the Gear method. [D. J. Jacob, Harvard]

TROPOSHERIC OZONE-NOx-HOx-CO-HC CHEMISTRY O 2 O 3 TROPOSPHERE O 3 NO 2 OH Dry deposition HOx family NOx family HO 2 RO 2 NO HNO 3 STRATOSPHERE H 2 O 2 CH 4 VOC CO

ATMOSPHERIC COMPOSITION MODELS @ CETEMPS MM 5 http: //www. mmm. ucar. edu/mm 5/ Chimere http: //euler. lmd. polytechnique. fr/chimere/ GEOS-Chem http: //www. as. harvard. edu: 16080/chemistr y/trop/geos/ WRF/Chem http: //ruc. fsl. noaa. gov/wrf/WG 11/ Regional Scale Meteorological Model Regional Scale Chemistry Transport Model Global Scale Chemistry Transport Model Regional Scale Meteorological-Chemistry model

Fore. Chem: Experimental “Chemical Weather” Forecast http: //pumpkin. aquila. infn. it/forechem/ CURRENT VERSION • European Domain, 0. 5°x 0. 5° • Forecast 2 days ahead (D-1 D+2) • Maps of max and mean of PM 10, PM 2. 5, O 3, NO 2, CO, SO 2 • Animations 72 -h UNDER DEVELOPMENT • Italian nested domain, 10 x 10 km • Graphics • Historical archive • NRT comparison with observations

DATA FLOW IN A CHEMISTRY-TRANSPORT MODEL EMISSIONS Anthropogenic and Biogenic/Natural sources of gas and aerosols LANDUSE INFO BOUNDARY CONDITIONS METEO FIELDS Global or regional model (e. g. ECMWF, MM 5, WRF) From larger scale CTM simulations Model Core simulates transport, chemical and deposition processes and solves continuity equation for chemical species

METEOROLOGICAL FIELDS DRIVE ADVECTION, TURBULENT VERTICAL DIFFUSION, REACTION RATES, BIOGENIC EMISSIONS AND DEPOSITION Vertical LIDAR profile over Milan Vertical model particulate profile Model particulate chemical composition LIDAR data by ISAC-RM [Stocchi et al. , in prep. ] Freshly emitted pollutants mix up to the PBL top Mixing and photochemical formation Advection of Saharan dust

MM 5 METEOROLOGICAL MODEL: NESTED DOMAINS TO INCREASE RESOLUTION 12 km USGS 4 km Ice Tundra/Bare Wetland Water Forests Grass & Shrubs Irrigated Crops Dryland Crops Urban

AT URBAN SITES MODEL UNDERESTIMATES TEMPERATURE AND OVERESTIMATES WIND MM 5 simulation vs. DEXTER observations (June 2007) TEMPERATURE (°C) T is underpredicted at night WIND SPEED (m/s) BOLOGNA/urban Wind Speed is overestimated S. PIETRO CAPOFIUME/rural

MM 5 “TUNABLE” LANDUSE PARAMETERS SUMMER Name Description Urban Dry Crops Wet Crops Grass Forest ALBD Albedo 18 17 18 20 13 SLMO Soil Moisture 10 30 30 15 35 SFEM Surface Emissivity 88 92 92 91 94 SFZ 0 Roughness length 50 15 16 12 50 THERIN Thermal Inertia 3 4 4 3 4 SCFX ? 0. 52 0. 60 0. 52 SFHC Heat Capacity 18. 9 e 5 25 e 5 20 e 5 30 e 5 The highlighted parameters are very different between urban to dry crops. Since dry crops category corresponds to mostly urbanized areas we try to modify these parameters toward urban-like values

WIND SPEED IS SENSITIVE TO CHANGES TO SURFACE ROUGHNESS, WHILE TEMPERATURE IS INSESITIVE TO ALL PARAMETERS 1 -25 April 2005 15 km resolution SYNOP data TEMPERATURE (°C) Daily cycle WIND SPEED (m/s) Accurate landuse info e. g. from satellite observations may be very important to improve meteorological simulation!

MODEL OF EMISSIONS OF GASES AND AEROSOLS FROM NATURE (MEGAN, GUENTHER ET AL. , ACP 2006) Base Emission Factor [mg/m 2/h] MM 5 Shortwave Radiation [W/m 2] MEGAN Isoprene Emission Rate [µg/m 2/h] STATIC HOURLY MONTHLY HOURLY Temporal resolution 1 h Spatial resolution 0. 5°x 0. 5° MODIS Leaf Area Index [m 2/m 2] Can be increased up to 1 km MM 5 2 -m Temperature [K] Satellite info: LANDUSE, VEGETATION DENSITY, SOIL MOISTURE

SATELLITE OBSERVATIONS MAY CONSTRAIN NOx AND VOC EMISSIONS THE CASE OF BIOGENIC ISOPRENE EMISSIONS FROM HCHO COLUMN WIN D CITY FOREST WIN D HCHO VOC NOx HCHO VOC HCHO = Formaldehyde VOC = Volatile Organic Compound THIS HCHO IS WELL CORRELATED TO ITS PARENT VOC!

TOP-DOWN CONSTRAINT OF EMISSIONS FROM SATELLITES RELIES ON BAYESIAN APPROACH Maximum a posteriori (MAP) solution for scalar EISOP: A posteriori solution: Ω [1016 molec cm-2] (Forward model) Ω (HCHO column) : EISOPRENE r = 0. 81 with gain: EISOP [1012 molec cm-2 s-1] A posteriori uncertainty: K is fitted from the EISOP: HCHO scatter plot calculated with CTM [Curci et al. , in prep. ]

MAP SOLUTION REDUCES THE A-PRIORI ERROR BECAUSE IT ADDS “PIECES OF INFORMATION” FROM OBSERVATIONS A = g · K = averaging kernel ds = tr(A) = degrees of freedom of signal or pieces of info Monthly mean map of “pieces of information” in OMI HCHO observations (Jul 2005) [Curci et al. , in prep. ]

MAP SOLUTION REDUCES THE A-PRIORI ERROR BECAUSE IT ADDS “PIECES OF INFORMATION” FROM OBSERVATIONS -15% [Curci et al. , in prep. ]

MODEL HCHO BIAS QUANTITATIVELY TRANSLATED INTO CORRECTION TO UNDERLYING ISOPRENE EMISSIONS OMI corrects model bias over Balkans and Spain [Curci et al. , in prep. ]

CTM MAY THEN BE USED TO EVALUATE THE IMPACT OF EMISSIONS ON AIR POLLUTANT LEVELS ! Large episodic contribution from BVOC emissions to ozone throughout the Mediterranean basin Up to 100 µg/m 3 in one extreme case in Spain! Observations from EMEP and Air. Base databases [Curci et al. , 2009]

SATELLITE DATA MAY BE INTEGRATED IN CTMs ALSO IN “DATA ASSIMILATION” PROCESS Data assimilation (DA) technique allows correction of model concentrations of observed species and those related to it. OMI/NO 2 Column 30 August 2007 Application of DA of OMI/NO 2 to simulation of North African fires. MTC Influence of fires at the end of August 2007 was detected in ozone data at Monte Cimone (MTC: 44 N, 11 E, 2165 m s. l. m) Fire region: OMI NO 2 column is assimilated as a new source of NOx during 28 -30 August 2007 NO 2 plume from fires [Grassi et al. , in prep. ]

NO 2 COLUMN ASSIMILATION STRONGLY AFFECTS NO 2 AND OZONE FIELDS Difference between simulation with and witout NO 2 DA over North Africa O 3 NO 2 28 Aug 2007 30 Aug 2007 NO 2 x 5! O 3 +20% [Grassi et al. , in prep. ]

OZONE CONCENTRATION AT MONTE CIMONE ARE BETTER SIMULATED WITH OMI/NO 2 DATA ASSIMILATION Obs O 3 With DA No DA [Grassi et al. , in prep. ]

BOUNDARY CONDITIONS ARE ALSO AN IMPORTANT INPUT TO REGIONAL CTM TOP BCs are implemented as concentrations specified at domain edges and trasported inside by winds Altitude NORTH WEST SOUTH Longitude Domain of the regional model (e. g. Chimere) EAST

BCs ARE TYPICALLY STATIC (MONTHLY MEANS) IN CONTINENTAL SCALE REGIONAL CTMs Hourly BC of dust in CHIMERE (regional) from GEOS-Chem (global)

SAHARAN DUST EVENT JULY 27 -29, 2005 MODIS AOT 550 nm 24/07 25/07 26/07 27/07 28/07 29/07

SAHARAN DUST EVENT JULY 27 -29, 2005 SEVIRI/MSG Sequence of RGB images composite with Brightness Temperatures Differences using Infra. Red SEVIRI channels (IR 8. 7, IR 10. 8, IR 12. 0) : DUST appears Magenta Thanks to W. Di Nicolantonio e A. Cacciari (CGS)

DRASTIC IMPROVEMENT OF PARTICULATE MATTER SIMULATION WITH REFINED BCs CORRECTED THROUGH COMPARISON WITH OBSERVED AOT Comparison of CHIMERE PM 10 with measurements at EMEP ground sites EMEP average Chimere w/ std BCs Chimere w/ Daily BCs Several dust events are captured with updated BCs BIAS decrease by 40% -4. 4 -2. 5 µg/m 3

A-PRIORI INFORMATION FROM MODEL IS USED TO RETRIEVE GROUND CONCENTRATIONS OF FINE PARTICULATE MATTER [Di Nicolantonio et al. , 2009]

CLEAN CHARACTERIZATION OF AEROSOL PHYSICAL AND CHEMICAL PROPERTIES IS THE HOT TOPIC IN ATMOSPHERIC CHEMISTRY POLLUTED … more rain and clouds later less rain first … [Rosenfeld et al. , Science 2008]

THE NEW WRF/CHEM MODEL SIMULATES AEROSOL-CLOUDS FEEDBACK AT UNPRECEDENT HIGH RESOLUTION The model is under development also at CETEMPS. In a first sensitivity simulation we tested model sensitivity to European anthropogenic aerosol emissions. W/out aerosol Total precipitation increases by only 2% Aerosols delay onset of precipitation that is recovered later. W/ aerosol [Tuccella et al. , in prep. ]

FURTHER READING • • • Di Nicolantonio, W. , A. Cacciari, A. Petritoli, C. Carnevale, E. Pisoni, M. L. Volta, P. Stocchi, G. Curci, E. Bolzacchini, L. Ferrero, C. Ananasso, C. Tomasi (2009), MODIS and OMI satellite observations supporting air quality monitoring, Radiation Protection Dosimetry, doi: 10. 1093/rpd/ncp 231 Hodzic, A. , Jimenez, J. L. , Madronich, S. , Aiken, A. C. , Bessagnet, B. , Curci, G. , Fast, J. , Lamarque, J. -F. , Onasch, T. B. , Roux, G. , Schauer, J. J. , Stone, E. A. , and Ulbrich, I. M. (2009), Modeling organic aerosols during MILAGRO: importance of biogenic secondary organic aerosols, Atmos. Chem. Phys. , 9, 6949 -6981 Curci, G. , Beekmann, M. , Vautard, R. , Smiatek, G. , Steinbrecher, R. , Theloke, J. , Friedrich, R. (2009), Modelling study of the impact of isoprene and terpene biogenic emissions on European ozone levels, Atmospheric Environment, 43, 1444 -1455, doi: 10. 1016/j. atmosenv. 2008. 02. 070 Steinbrecher, R. , Smiatek, G. , Koble, R. , Seufert, G. , Theloke, J. , Hauff, K. , Ciccioli, P. , Vautard, R. , Curci, G. (2009), Intra- and inter-annual variability of VOC emissions from natural and seminatural vegetation in Europe and neighbouring countries, Atmospheric Environment, 43, 1380– 1391, doi: 10. 1016/j. atmosenv. 2008. 09. 072 Bessagnet, B. , L. Menut, G. Curci, A. Hodzic, B. Guillaume, C. Liousse, S. Moukhtar, B. Pun, C. Seigneur and M. Schulz (2008), Regional modeling of carbonaceous aerosols over Europe Focus on Secondary Organic Aerosols, Journal of of Atmospheric Chemistry, 61, 175 -202. Curci, G. Visconti, D. J. Jacob and M. J. Evans (2004), Tropospheric fate of Tunguska generated nitrogen oxides, Geophys. Res. Let. , 31, L 06123, doi: 10. 1029/2003 GL 019184

THANKS FOR YOUR ATTENTION!

EXTRAS If you haven’t had enough!

SPECIFIC ISSUES FOR AEROSOL CONCENTRATIONS • A given aerosol particle is characterized by its size, shape, phases, and chemical composition – large number of variables! • Measures of aerosol concentrations must be given in some integral form, by summing over all particles present in a given air volume that have a certain property • If evolution of the size distribution is not resolved, continuity equation for aerosol species can be applied in same way as for gases • Simulating the evolution of the aerosol size distribution requires inclusion of nucleation/growth/coagulation terms in Pi and Li, and size characterization either through size bins or moments. condensation coagulation Typical aerosol size distributions by volume nucleation [D. J. Jacob, Harvard]

EMISSION INVENTORY FOR ANTHROPOGENIC EMISSIONS 1. Total annual emissions (e. g. EMEP, European Monitoring and Evaluation of Pollution) of: CO, NH 3, SO 2, NOx, VOC, PM 2. Speciation of VOC [Passant, 2002] VOC 1 … VOC 2 VOC 350 3. Corrispondence of emitted and modelled species CO NOx … PM 20% PM fine, 80% PM coarse VOCi ? ? ? 350 VOC: too many! 1. Chemical degradation of many is unknown 2. Computational limits AGGREGATION

CTM MAY THEN BE USED TO EVALUATE THE IMPACT OF EMISSIONS ON AIR POLLUTANT LEVELS Average daily ozone maximum only with anthropic emissions (summer 2000) Increase of ozone max due to biogenic VOC emissions [Curci et al. , Atmo Env 2009]