aerosol size distribution performance based on changes to particle emissions and nucleation
DESCRIPTION
Aerosol Size Distribution Performance Based on Changes to Particle Emissions and Nucleation. Robert A. Elleman & David S. Covert Department of Atmospheric Sciences University of Washington. Importance of Size Distribution Modeling. Health Effects in Ultrafine. Size Distribution. PM 2.5. - PowerPoint PPT PresentationTRANSCRIPT
Aerosol Size Distribution Performance Based on Changes to Particle Emissions and Nucleation
Robert A. Elleman&
David S. Covert
Department of Atmospheric SciencesUniversity of Washington
Drivers.com
PM2.5Size Distribution
www.eere.energy.gov
Importance of Size Distribution Modeling
• Health Effects in Ultrafine
CMAQ Simulation• CMAQ v4.4 (September 2004)
• 4 km horizontal resolution
• 22 levels -- lowest layer is ~30 meters
• 4 day spin-up
• 3 day simulation– 00 UTC 08/26 – 00 UTC 08/29, 2001
PNW 2001 Pacific 2001
Observational Datasets
From: http://www.pnl.gov/pnw2001/From: http://www.pnl.gov/pnw2001/
Cassier Tunnel
Slocan Park
Sumas Mtn.Langley Ecole
Golden Ears ParkCassier Tunnel
Slocan Park
Sumas Mtn.Langley Ecole
Golden Ears Park
From http://www.msc.ec.gc.ca/projects/pacific2001/study_sites_e.html
Portland
Seattle
Vancouver
Portland
Seattle
Vancouver
CMAQ v4.4 Number Concentrations
Flight track
Pacific Standard Time = UTC - 8
Ternary Number Concentrations
Flight track
Pacific Standard Time = UTC - 8
• Emission Size Distribution– Can cause consistent number underprediction
– Outdated science in model biases fewer, larger particles
• Nucleation mode– Not all nucleated particles survive to be observed and
included in Aitken mode
Process Nano-mode• Few 1 nm particles survive to Aitken mode
Kerminen, V.-M., and M. Kulmala, 2002: Analytical formulae connecting the “real” and the “apparent” nucleation rate and the nuclei number concentration for atmospheric nucleation events. J. Aerosol Sci., 33, 609-622.
CS
GR
“Condensation Sink”
H2SO4 Growth Rate=
Nucleation Model # Concentrations
Ternary
obs
Binary/None
Ternary
Binary/None
Ternary w/Proc
Ternary w/Proc
0826PM 0827AM 0827PM Langley Sumas Average
Binary -6,469 -3,961 -14,782 -11,696 -14,147 -10,211
None -6,480 -3,962 -14,784 -11,699 -14,157 -10,216
Ternary 47,944 9,272 15,647 8,583 17,772 19,844
Ternary w/Processing
6,360 -3,521 -13,767 -9,136 -1,982 -4,409
Nucleation Model Size Distributions
• Ternary nucleation with processing to 10 nm best simulates the shape of the size distribution and the Aitken mode diameter
• Emission Size Distribution– Can cause consistent number underprediction
– Outdated science in model biases fewer, larger particles
• Nucleation mode– Not all nucleated particles survive to be observed and
included in Aitken mode
Improve Emissions Size Distributions
• Scale to 4 km grid– All processes < 4 km scale must be
parameterized
• Use “urban background” measurements to constrain model– Mostly from traffic
Emissions Sensitivity Matrix
4 km
Organics All Others
Name % Mass in Ait % Mass in Ait Aitken Dgv Aitken sg Acc Dgv Acc sgBase 0.1 0 30 nm 1.7 300 nm 2
Best Guess 10 10 60 nm 1.7 280 nm 1.7
Upper Bound 20 20 50 nm 1.7 280 nm 1.7
All Species
Effect of Emission Size Distribution
Number Enhancement
Dg, aits
g,ait Dg, accs
g,acc
Base 45 2.3 80 2.1
Best Guess 35 2.1 110 2.0
Upper Bound 30 1.8 105 1.9
Size Parameters
Emissions CMAQ
Base 1 1
Best Guess 4 2.4
Upper Bound 13 6.4
CMAQ v4.4
Terna
ryTernary w/Processing
Emission Size Distribution
Conclusion
Contact Information:
Robert Elleman [email protected] of Washington (206) 543-9144Department of Atmospheric SciencesSeattle, WA, USA