davidson iwaqr 2009 - earth system research …florida smoke 7 florida, 1/09/08 – dense morning...
TRANSCRIPT
Paula Davidson NOAA/NWS
International Workshop on Air Quality Forecasting Research, Boulder, Colorado, December 2, 2009
2
Outline
1. NAQFC – mission and capabili:es, recent progress
2. Forecast challenges 3. Approach 4. Research needs and opportuni:es 5. Summary/ Way ahead
3
Na1onal Air Quality Forecast Capability
• Poor AQ responsible for losses each year in the US: • > 60,000 premature deaths • > $100B in health costs
• Vision: protect lives and property by providing accurate and :mely AQ predic:ons for the US
• Strategy: work with EPA, State and Local AQ agencies and private sector to develop end‐to‐end AQ forecast capability for the US
• Benefits: hourly predic:ons available to local and state AQ forecasters, decision makers, general public
4
5
2010: smoke
2010: O3 AK,HI 2009: smoke
Ozone performance
6
2009
CONUS, wrt 85ppb Threshold
Operational
2007
Contiguous US (CONUS) wrt 85 ppb threshold Implemented 9/07
Experimental
2008
CONUS, wrt 85ppb Threshold
Operational
Threshold
Threshold Prediction
Obs
erva
tions
Fraction Correct = # green / # all
Florida smoke
7
Florida, 1/09/08 – Dense morning smoke predicted near Orlando – Accident on I‐4 caused 50‐vehicle crash with 3
fatali1es – Evacua1on concerns for PM exposure: senior
ci1zen facili1es
Quan1ta1ve PM performance
8
Model minus observations: regional monthly mean
CB05 AERO4
• Aerosol simula:on using emission inventories: – Show seasonal bias‐‐ winter, overpredic:on; summer, underpredic:on
• Chemical boundary condi:ons/trans‐boundary inputs
• Intermi\ent sources • Impacts of 2‐way chem‐met coupling
Forecast challenges
Approach towards quan1ta1ve PM predic1on
9
PM components Dust
Integrated PM
Smoke
Chemical Boundary Condi1ons
Chemical data assimila1on
Advanced Chemical Mechanism
Emissions Inventory Inputs
Integrated PM
Capability
Progress on NOAA’s PM2.5 Modeling for AQ Predic1ons
• Be\er representa:on of intermi\ent events (dust storms, forest fires)
• Dynamic boundary condi:ons
• Quan:ta:ve predic:on of reac:ve components, especially secondary organic aerosols (SOA)
• Advanced chemical mechanisms: gas phase, aqueous phase
• Wet/dry deposi:on 10
Wildfire smoke
• Smoke forecas:ng system (Rolph et al 2009)
• Fire detec:on
• Plume rise (Stein et al 2009)
• Transport
• Impact on PM2.5 and ozone
• Modeling fire behavior • Surface and profile
observa:ons of smoke 11
h\p://juneauempire.com/stories/072109/reg_466811327.shtml
Dust storms
12
Preliminary testing of dust predictions
Dust storm approaching Stratford, Texas. Dust bowl surveying in Texas NOAA George E. Marsh Album, www.photolib.noaa.gov
• Dust emission – soil type – wetness – roughness – wind speed – par:cle size distribu:on
• Transport (Draxler, in prep)
• Deposi:on • Surface and profile
observa:ons
Chemical boundary condi1ons
• Improved climatology (Hawaii ozone)
• Global models (ozone: Tang et al 2009, dust: Huang et al 2009)
• Global chemical data assimila:on
13
HI surface ozone on 9/11/2009
LBC: static (CONUS)
LBC: monthly mean from O3 sonde
CONUS ozone profiles
Quan1ta1ve predic1on of reac1ve components
14
• Aerosols: 2009 mean of daily developmental predic:ons for CONUS/ AIRNow data (Byun)
• (CB05/Aero4, Daily Max), emissions inventories inputs • Improvement to chem. mechanism & inputs in progress
Chemical data assimila1on
• Methods (3D‐Var, 4‐D Var, ensemble methods) • Observa:ons (in situ, remotely sensed) • Distribu:on of integral increments (e.g. AOD,
total PM2.5) into species, profiles • Observed or retrieved variables • Background and observa:on error covariance
modeling, inclusion of model error • Errors in emissions, non‐random errors • Infrastructure
15
Research needs for AQ Forecas1ng: Near Term
• Improved reac:ve chemical transport of species involved in produc:on, transport, transforma:on, & removal of PM2.5
• Chemical boundary condi:ons • Chemical data assimila:on technology and infrastructure; improved modeling and analyses of emissions inputs
• Advanced chemical mechanisms, wet/dry deposi:on
• Con:nue improvement of atmospheric simula:ons of near‐surface condi:ons
16
Research needs for AQ Forecas1ng: Longer‐Term
• Assimila:on of new chemical observa:ons • Modeling of addi:onal pollutants • Coupling to broader range of environmental models
Addi:onal, related research: • AQ predic:on will benefit from related advances in meteorological/ environmental modeling and ensemble genera:on, post‐processing, atmospheric chemical observa:ons
17
Integrated observa1ons
• Speciated PM – need to be able to verify and constrain PM components
• Profiles – ver:cal distribu:on needed domain‐wide
• Timeliness – for assimila:on and for near‐real‐:me verifica:on
18
Chemical‐meteorological model coupling
19
Atmospheric Model
Ocean
Land Surface
Air Quality
Space
Hydrology
Ecosystem
Ensembles
Etc
Resolu1on Changes, Downscaling Post Processing Bias Correc1on, Sta1s1cal Methods, Ensembles Product Genera1on Verifica1on
4D Data Assimila1on e.g. 4D Var, EnKF, hybrids WIDB
Mul1‐component ensemble +
Stochas1c forcing
Components
Physics Dynamics Chemistry
Couplers
Earth System Models
Weather Industry
20
Real‐:me Weather Obs:
Monitoring, Rem Sensed
Weather Data Assimila:on QC/QA
Customer Feedback
Mesoscale Weather Model
Atm Chemical Data Assimila:on:
3‐D or 4‐D
Emissions Preprocessing
R‐T Boundary Condi1ons, Trans‐boundary transport Especially: Dust, Fires
Real‐:me Chem Observa:ons: In‐Situ
Monitoring, Rem Sensed
Chem Transport Model: PM + Ozone
Verifica:on Product Genera:on
Genera:on
Chem Transport Model: Ozone
Na:onal Air Quality Forecast Capability:
Current Opera:onal Links: Ozone
Con:nuous SCIENCE/TECH Infusion
Research, Development, Tes:ng, Integra:on
?
?
Forecasts for public dissemina1on
Numerical Weather Predic1on
Numerical Chemistry Predic1on
Boundary Condi:ons From Global
Weather Model
Pollutant Emissions Inventory
21
Real‐:me Weather Obs:
Monitoring, Rem Sensed
Weather Data Assimila:on
QC/QA
Customer Feedback
Mesoscale Weather Model
Atm Chemical Data Assimila:on:
3‐D or 4‐D
Emissions Preprocessing
Real‐:me Chem Observa:ons: In‐Situ
Monitoring, Rem Sensed
Chemical mechanism: PM + Ozone
Verifica:on Product Genera:on
Genera:on
Na:onal Air Quality Forecast Capability:
Planned opera:onal links for
chemical‐meteorological model coupling
Con:nuous SCIENCE/TECH Infusion
Research, Development, Tes:ng, Integra:on
Forecasts for public dissemina1on
Atmospheric Predic1on
Real :me boundary condi:ons
weather
Pollutant Emissions Inventory
Real :me boundary condi:ons
chemistry
22
Two-way coupling e.g. radiation, cloud microphysics
Summary
• Opera:onal ozone (CONUS) and smoke (CONUS, AK) predic:on
• Comple:ng na:onwide predic:on coverage
• Forecast challenge: quan:ta:ve PM2.5 predic:on
• Approach: concurrent development of system components towards an integrated capability
23
Summary of Research needs
• Modeling PM, advanced chemical mechanisms • Chemical boundary condi:ons
• Chemical observa:ons and data assimila:on
• Improved emissions inputs
• Coupling to other environmental models
• Modeling of addi:onal pollutants
24