Models Play a Critical Role in Linking Emissions to Aerosol and Trace Gas Distributions and Subsequent Effects
Modeled
Observed
07/29/09
1
Models are an Integral Part of Air Quality Studies • Field experiment planning • Provide 4-Dimensional context of the observations • Facilitate the integration of the different measurement platforms • Evaluate processes (e.g., role of biomass burning, heterogeneous chemistry….) • Evaluate emission estimates (bottom-up as well as top-down) • Emission control strategies testing • Air quality forecasting 07/29/09 • Measurement site selection
2
07/29/09
3
Air Quality Modeling: Improving Predictions of Air Quality (analysis and forecasting perspectives) Met model
Chemical, Aerosol, Removal modules
Predicted Quantity: e.g., ozone AQ violation
CTM
How confident are we in the models & predictions?
Emissions
07/29/09
Observations
4
Chemical Transport Model • 3D atmospheric transport-chemistry model (STEM-III)
∂ci 1 = −u ⋅ ∇ci + ∇ ⋅ ( ρK∇ci ) + f i (c) + Ei ∂t ρ where chemical reactions are modeled by nonlinear stiff terms
f i (c) = Pi (c) − Di (c)ci • Use operator splitting to solve CTM
M[Δ] t t 07/29/09
t
t TXΔ2Δ2Δ2ΔΔ2Δ2Δ2 TY t TZ t C t TZ t TY t TX t 5
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 ~ 0.51o (~50 to 100 km) in horizontal and ~ 0.51 km in vertical • Chemical Transport Models (CTMs) use external meteorological data as input (or run online) 07/29/09 6 • General Circulation Models (GCMs) compute their own meteorological fields From D. Jacob
07/29/09
7
07/29/09
8
07/29/09
9
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 07/29/09
10
From D. Jacob
LAGRANGIAN APPROACH: TRACK TRANSPORT OF POINTS IN MODEL DOMAIN (NO GRID)
• Transport large number of points with trajectories from input meteorological data base (U) + random turbulent component (U’) over time steps ∆t • Points have mass but no volume
position to+∆t U’∆t position to
U∆t
07/29/09
• Determine local concentrations as the number of points within a given volume • Nonlinear chemistry requires Eulerian mapping at every time step (semiLagrangian) PROS over Eulerian models: • no Courant number restrictions • no numerical diffusion/dispersion • easily track air parcel histories • invertible with respect to time CONS: • need very large # points for statistics • inhomogeneous representation of domain • convection is poorly represented 11 From D. Jacob • nonlinear chemistry is problematic
LAGRANGIAN RECEPTOR-ORIENTED MODELING Run Lagrangian model backward from receptor location, with points released at receptor location only
• Efficient costeffective quantification of source influence distribution on receptor (“footprint”) • Enables inversion of source influences by the adjoint method (backward model is the adjoint of the Lagrangian forward model) 07/29/09
12
From D. Jacob
Air Quality Prediction: A Challenge of Scales and Integration Megacity Impacts
Air quality Analysis Regional Prediction
Global Assimilation
Satellite Products
07/29/09
Modified after Pierce NASA/Langley
13
Integrated Science Studies: Impacts of Global Composition on Regional Air Quality Global-Regional-Urban nesting of CTMs Effects of Boundary Conditions are significant and improve predictions (Tang et al., JGR 2007). Alaskan BB Impacts Northern Boundary
07/29/09
Assessment of continental inflow/outflow requires unified modeling/measurement strategy to accurately characterize coupling between the continental boundary layer, free troposphere, and longrange transport. 14
Model Resolution, Transport and Removal also Contribute to Differences Emissions
Monthly mean concentrations
Temporal variation in concentrations
BC 07/29/09
15
Removal Processes Remain Poorly Characterized in Models Impact of Wet Removal on Predicted BC Progress limited by lack of understanding and observations
07/29/09
16
An Air Quality Modelling System Site data Topography
Land cover
Weather forecast and local data Model results
Observations
Site data processing
Emission inventories Industries Vehicular traffic
Chemical data
3D meteo modelling
Emissions modelling
Models interface
Area sources
3D chemical-transport modelling
Chemical 3D IC / BC processing
Monitoring data Model results
3D meteo with on-line chemistry Analysis of concentrations Comparison with regulatory norms and objectives
Concentrations: 3D & ground
Post-processing & graphic
07/29/09
17
Need to Estimate Emissions at Appropriate Scales
07/29/09
18
Detail: a matter of scale (3) SO2 emissions in the vicinity of Shanghai, China
1º x 1º
.5º x .5º
10’ x 10’
30’’ x 30’’
07/29/09
19
Emissions processing for AQM Inventories
Thematic data
“reference raw emission data” (point / line / area) SPACE DISAGGREGATION
TIME MODULATION
Modulation profiles
Residential
(hourly, daily, monthly) 2.5
Highways
Somma di FACTOR
NMVOC & PM SPECIATION & SIZE
2
DAY NOM_PROFIL
1.5
1 - Leggeri_Extraurbane 2 - Leggeri_Extraurbane
Deciduous
3 - Leggeri_Extraurbane 4 - Leggeri_Extraurbane 5 - Leggeri_Extraurbane 1
6 - Leggeri_Extraurbane 7 - Leggeri_Extraurbane
0.5
3
NonIndComb
Speciation & dimensional profiles
Formaldeide
gg_0_150 0 0
1
2
3
4
5
6
2.5 7
gg_150_300 8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Acetaldeide
gg_300_500
23
gg_500_1000
HOUR
2
Aldeidi superiori
gg_1000_2000 P rodInd
1.5
Chetoni
SmalB usi DomA ctiv A irtraf
1
Colture VdA_auto_EA
0.5
VdA_comm_EA Decid
0
Conif
1
2
3
4
5
6
7
mese
8
9
10 11 12
ThP owP lants TotElP owReq Veicoli_VdA
Etilene Alcani (B.R.)
Latifoglie
Alcani (A.R.)
Conifere Auto
Aromatici (B.R.) Aromatici (A.R.) Olefine (B.R.) Olefine (A.R.) Isoprene Monoterpeni 0%
20%
40%
60%
80%
100%
Model-ready input 07/29/09
(hourly, gridded, speciated emissions)
20
Other types of emission data • Emissions estimated “on-line”, using current meteorology – desert dust – sea salt – forest fires
• Data from plants emissions monitoring systems • Real-time traffic
07/29/09
21
Other factors • Periodical update of inventory data (changes in activity levels, fuels, control equipment …)
• Special situations (traffic bans, events, …)
• Future scenarios …examples later
07/29/09
22
OUR MODELING SYSTEM FRAMEWORK
Assimilation Observations
Observations
07/29/09
23
STEM
Non reactive tracers: Anthropogenic CO, Biomass Burning CO, BC, OC, Sulfate aerosol, Dust, Anthropogenic Hg, Volcano SO2 Simple decaying rates Dry deposition Wet scavenging Horizontal Resolution: 60x60 km Vertical Resolution: WRF's half-sigma coordinate Time Resolution: 6hr Forecast length: +120hr where 00hr=00UTC Full Chemistry: SAPRC99 mechanism + on-line four-bin SCAPE II module on-line TUV for photolysis rates computation BC/IC: RAQMS global chemical transport model Horizontal Resolution: 60x60 km - North America Vertical Resolution: WRF's half-sigma coordinate Time Resolution: 6hr 07/29/09 24 Forecast length: +48hr where 00hr=00UTC 7th International Conference on Air Quality – Science and Application – 2009 Istanbu
07/29/09
25
Mission A strong Overview outflow event will appear from Saturday to Sunday
July 1 to 25 Model CO
Midwest Ohio etc NY-MA-MD TX-NM Southeast California Canada
2km wind field
07/29/09
26
BETTER UNDERSTANDING OF ARCTIC ATMOSPHERIC COMPOSITION AND CLIMATE Transport of mid-latitudes pollution to the Arctic Transport pathways for different pollutants - Contributions from different source regions - Source-receptor relationships. Boreal forest fires Compositions evolution of the fire plumes - aerosol optical properties and evolution - implications for regional and global atmospheric composition. Aerosol radiative forcing Regional radiative forcing from Arctic haze, fire plumes major sources of soot to the Arctic - soot deposition effect on ice albedo. Chemical processes HOx-NOx chemistry - regional extent of halogen radical chemistry - regional implications for ozone, aerosols, mercury - stratosphere-troposphere exchange effect on 07/29/09 27 tropospheric ozone in and theApplication Arctic. – 2009 Istanbul 7th International Conference on Air Quality – Science
ARCTAS Arctic Research of the Composition of the Troposphere from Aircraft and Satellites Part of the international IPY/POLARCAT Arctic field program for atmospheric composition NASA TROPOSPHERIC CHEMISTRY PROGRAM ARCTAS Partners: NASA, NOAA, DOE, NSF, Canada, CNRS, DLR MEASURE
SATELLITE
07/29/09
VALIDATION ASSIMILATION ANALYSIS
MODEL 28
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
ARCTAS AIRCRAFT FLEET DC-8 in situ chemistry and aerosol Ceiling 37 kft, range 4000 nmi, endurance 9 h Payload: O3, H2O, CO, CO2, CH4, NOx and Hox hemistry, BrO, mercury, NMVOCs, halocarbons, SO2. HCN/CH3CN, actinic fluxes, aerosol composition, aerosol mass and number concentrations, aerosol physical and optical properties, remote ozone and aerosol.
P-3B radiation and in situ aerosol Ceiling 30 kft, range 3800 nmi, endurance 8 h Payload: optical depth, radiative flux, radiance spectra, aerosol composition, black carbon 29
07/29/09 7th International Conference on Air Quality – Science and Application – 2009 Istanbul
SATELLITE TEAMS Satellite: CALIPSO (CALIOP); MetOp(GOME-2); AQUA (MODIS, AIRS); TERRA (MODIS, MISR, MOPITT); AURA (MLS, TES, HIRDLS, OMI); ENVISAT (SCIAMACHY) Capabilities: Aerosol Optical Depth properties, CO, Ozone, BrO, NO2, HCHO, Methan
MODEL
Forecasts/analyses: GEOS-5, GOCART, GEOS-Chem, STEM, MOZART, RAQMS
GROUND TEAMS
Ground Teams: UAF, NATIVE, ARC-IONS 07/29/09
30
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
MISSION DEPLOIMENTS Spring deployment: April 2008 Fairbanks and Barrow, Alaska Thule, Greenland Iqaluit, Canada
Summer deployment: June – July 2008 California, Cold Lake and Yellowknife, Canada Thule Greenland 07/29/09
31
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
FLIGHT PLANNING AIMS TO DEFINE THE BEST FLIGHT MENU ACCORDING TO: weather conditions sampling opportunities progress in meeting mission objectives UNIVERSITY OF IOWA Forecasting Team Purpose: Provide Gridded Data Emissions for all chemical transport models. Daily-bases High Resolution 3-D Meteorological and Chemical Forecast
07/29/09
32
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
07/29/09
33
07/29/09
34
07/29/09
35
07/29/09
36
07/29/09
37
07/29/09
38
07/29/09
39
07/29/09
40
07/29/09
41
07/29/09
42
07/29/09
43
07/29/09
44
07/29/09
45
07/29/09
46
07/29/09
47
07/29/09
48
07/29/09
49
07/29/09
50
07/29/09
51
07/29/09
52
07/29/09
53
How did we do?
07/29/09
54
DC-8 SPRING FLIGHT 17th of April European Plume
Anthropogenic CO 5.5 km by S. Kulkarni, CGRER
Asian Plume Biomass CO 5.5km
07/29/09
Biomass CO curtain plot 55
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
DC-8 SPRING FLIGHT 17th of April Black Carbon
Speed
Temp
Missed Polar Pick European Asian Plume Plume
Biomass CO tracer vs. measured Hydrogen Cyanide (HCN )
RH
07/29/09
56
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
th DC-8 SUMMER FLIGHT 09 of July Siberian-Asian
Plume
AOD
by S. Kulkarni, CGRER
BC 07/29/09
5.5km 57
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
DC-8 SUMMER FLIGHT 09th of July Speed
Direction
Composite satellite image http://amrc.ssec.wisc.edu/
Temp
RH 5.5km 07/29/09
58
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
DC-8 SUMMER FLIGHT 09th of July Biomass CO tracer vs. measured Hydrogen Cyanide (HCN ) Siberian Asian Plum
STEM FULL CHEMISTRY NORTH AMERICA DOMAIN OZONE
07/29/09
59
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
CONCLUSIONS •For the selected flights, the major targets highlighted during the pre-flight meetings have been reasonably met. •The WRF model was able to capture the important features of synoptic conditions (fronts positioning, clouds cover, general circulation) and Boundary Layer flow. •Computed-measured comparisons showed the System’s capability to describe the trends seen for the observations •The preliminary results coming out from the present study represent an encouraging starting point for further analysis based on the feedback of 07/29/09 60 measurements into the models in order to better 7th International Conference on Air Quality – Science and Application – 2009 Istanbul understand the Arctic Environment and to improve
ACKNOWLEDGEMENTS I would like to thank the ARCTAS Measurement Team for permission in using the measurements, CGRER and the University of Iowa and the Space Science and Engineering Center, University of WisconsinMadison, for the satellite imagine.
07/29/09
61
7th International Conference on Air Quality – Science and Application – 2009 Istanbul
07/29/09
62
OPERATOR SPLITTING IN EULERIAN MODELS Reduces dimensionality of problem
• Split the continuity equation into contributions from transport and local terms:
Ci Ci t t
TRANSPORT
dCi dt
LOCAL
dCi Transport advection, convection: U Ci dt TRANSPORT Local chemistry, emission, deposition, aerosol processes: dCi dt
Pi (C) Li (C) LOCAL
… and integrate each process separately over discrete time steps:
Ci (to t ) (Local)•(Transport) Ci (to ) These operators can be split further: • split transport into 1D advective and turbulent transport for x, y, z (usually necessary) split local into chemistry, emissions, deposition (usually not necessary)63 •07/29/09
SPLITTING THE TRANSPORT OPERATOR
• Wind velocity U has turbulent fluctuations over time step ∆ t:
U(t ) U U '(t )
Timeaveraged component (resolved)
Fluctuating component (stochastic)
• Split transport into advection (mean wind) and turbulent components:
Ci 1 U Ci KCi t
air density K turbulent diffusion matrix
advection turbulence (1storder closure)
• Further split transport in x, y, and z to reduce dimensionality. In x direction:
Ci Ci 1 Ci u ( K xx ) t x x x 07/29/09
advection operator
turbulent operator
U (u , v, w)
64
SOLVING THE EULERIAN ADVECTION EQUATION Ci Ci u t x
• Equation is conservative: need to avoid diffusion or dispersion of features. Also need mass conservation, stability, positivity… • All schemes involve finite difference approximation of derivatives : order of approximation → accuracy of solution • Classic schemes: leapfrog, LaxWendroff, CrankNicholson, upwind, moments… • Stability requires Courant number u∆ t/∆ x < 1 … limits size of time step • Addressing other requirements (e.g., positivity) introduces nonlinearity in advection scheme 07/29/09
65
VERTICAL TURBULENT TRANSPORT (BUOYANCY) • generally dominates over mean vertical advection
• Kdiffusion OK for dry convection in boundary layer (small eddies) • Deeper (wet) convection requires nonlocal convective parameterization Convective cloud (0.1100 km)
detrainment Model vertical levels
updraft
downdraft
entrainment
07/29/09
Model grid scale
Wet convection is subgrid scale in global models and must be treated as a vertical mass exchange separate from transport by gridscale winds. Need info on convective mass fluxes from the model meteorological driver.
66
LOCAL (CHEMISTRY) OPERATOR: solves ODE system for n interacting species For each species i 1, n dC i Pi (C) Li (C) dt
C (C1 ,...Cn )
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 C(to t )
Ci (to t ) Ci (to ) Pi (C(to t )) Li (C(to t )) t
i 1, n
Solve e.g., by Newton’s method. Backward Euler is stable, massconserving, flexible (can use other constraints such as steadystate, chemical family closure, etc… in lieu of ∆ C/∆t ). But it is expensive. Most 3D models use higherorder implicit schemes such as the Gear method. 07/29/09
67
COMPOSITION? Solve continuity equation for chemical mixing ratios Ci(x, t) Lightning
Eulerian form:
U = wind vector
Ci U Ci Pi Li t
Pi = local source
Transport
Lagrangian form:
Chemistry Aerosol microphysics
dCi Pi Li dt
Volcanoes Fires
Human Land biosphere activity
of chemical i
Li = local sink
Ocean
Courtesy John Reilly, MIT 07/29/09
69