NCEP Global Dust Modeling

1
NCEP Global Dust Modeling

• Ho-Chun Huang, Dongchul Kim, Youhua Tang, Sarah
  Lu, Pius Lee2, Marina Tsidulko, Jeff McQueen,
  Shrinivas Moorthi, Jongil Han, Mark Iredell,
  Geoff DiMego, William Lapenta, Stephen Lord,
  Paula Davidson3, Ivanka Stajner4, Arlindo
  daSilva5, Mian Chin5, and Thomas Diehl6
• NOAA/NWS/NCEP/EMC 2NOAA/ARL 3NOAA/NWS/OST
  4Noblis 5NASA/GSFC 6UMBC
• Dust Emissions Workshop, Silver Springs, March 25
  2009

2
GOALS

• To serve as an initial benchmark for a planned
  inline implementation
• To provide aerosol lateral boundary conditions
  for regional air quality forecasting capability
  (AQFC) as well as regional dust modeling system
• To provide modeled aerosol fields for
  assimilation of satellite and in-situ data
• To meet NWS and WMO global dust forecasting goals

3
Approach

• Requirement NCEP GFS Driven Meteorology
• NASA/GSFC GOCART
• Community
• Available and widely used global dust model
• Build on existing collaboration with NASA GSFC
• Leverage several existing project
• JCSDA
• ESMF

4
Dust Emissions

• Ginoux et al. 2001
• surface topographic depression
• surface wetness
• surface wind speed

5
Quantitative Verification

• June-August 2006
• MODIS collection 4 AOD
• AERONET Level 2 AOD
• AIRNOW PM2.5 (via CMAQ 12 km, TEXAQS)
• CALIPSO vertical profiles (in progress)
• Spring 2007
• Mt. Batchelor PM1
• CALIPSO vertical imagery
• Calibration Experiments
• Horizontal Resolution T126 vs. T62
• Physics SAS vs. RAS convection
• Dust Source1x1 vs. 2.5x2 erodable fraction
  maps

6
2006 Atlantic Transports of Saharan Dust - TEXAQS
7
(No Transcript)
8
(No Transcript)
9
Dust storm entered CMAQ domain Distribution
pattern is still good but model underestimates
the intensity
Courtesy of Dongchul Kim
10
Comparison for surface stations over Texas
11
Submicron aerosol
Mt. Bachelor data from D. Jaffe, U.
of Washington-Bothell
Mt. Bachelor, Oregon GFS-GOCART dust
µg/m3

• When submicron dust is enhanced in GOCART then
  measured scattering is typically enhanced at Mt.
  Bachelor
• For a couple of cases of enhanced dust at Mt.
  Bachelor, qualitative comparison of CALIPSO and
  GOCART vertical sections reveal similar altitude
  of aerosol/dust layers

Courtesy of Ivanka Stajner
12
Surface
T62 (2x2) vs. T126 (1x1) 12/27/2007 to
01/29/2008
T62
T126
500mb
13
GFS RAS vs. SAS

• 09/21/2008 to 10/15/2008
• Relax Arakawa-Schubert convective scheme (RAS)
• Simplified Arakawa-Schubert convective scheme
  (SAS)

SAS-RAS AOD
14
Courtesy of Youhua Tang
15
Mian DU1 factor 0.1 DU2-5 factor 0.25 Paul DU1
factor 0.1 DU2-5 factor 0.225
16
Summary

• Initial Global Dust Forecasting System Developed
  Leveraged existing models
• Initial Configuration
• Resolution T126, Limited by production
  resources
• Impact of Convection SAS
• Impact of dust source map under testing
• Captures large scale patterns and timing
• MODIS AOD and AERONET AOD
• Source Strength under-estimated
• Positive impact on CMAQ TEXAQS simulations

17
Questions to dust emissions

• What is the target spatial and temporal scale of
  your emissions algorithm
• Regional scale and daily to episodic mean
• Global scale and weekly, monthly, or seasonal
  mean
• When large scale meteorology can not pick up
  meso- or small scale meteorological features that
  stir the airborne dust, e.g., moist convection,
  what is your strategy to account for the missing
  events (episodes)