Dust sources location: combining TOMS and MODIS data

1
Dust sources location combining TOMS and MODIS
data

• Paul Ginoux
• NOAA-GFDL
• Princeton, NJ

Paul Ginoux NOAA-GFDL Princeton, NJ
Dust Workshop Boulder, October 22-23, 2004
2
Mineral Dust Aerosols TOMS Aerosol Index

• TOMS Aerosol Index

 ?, shorter wavelength (331 or 340 nm) ?0, longer
wavelength (360 or 380 nm)
http//toms.gsfc.nasa.gov
3
Mineral Dust Aerosols TOMS Aerosol Index
The maxima of TOMS AI are persistently (over 20
years of observations) at the same locations
(Prospero et al., RG, 2002).
Niger river basin (Mali)
Bodele depression (Chad)
Cott El Jerid (Tunisia)
Wadis in Libya
Prospero, Ginoux, Torres, Nicholson, and Gill,
Rev. Geophys., 40(1), 2002.
4
Mineral Dust Aerosols TOMS Aerosol Index
Most TOMS AI maxima are associated with
topographic depressions, dry lakes or river beds
(Prospero et al., 2002)
5
Mineral Dust Aerosols Global dust sources
Ginoux et al. (2001) prescribed the dust sources
based on topography and vegetation (bare
surface from AVHRR 1x1 degree)
GOCART, GFDL models, KARMA, GEOS-CHEM, US Army,
several regional models, IPCC 2001, AEROCOM.
6
Mineral Dust Aerosols Global dust emission
7
Mineral Dust Aerosols Other global dust
inventories

• Zender et al. (2003) Geo-morphological approach
  comparison with Ginoux et al. (2001)
• Tegen et al. (2004) sources are associated with
  Paleo-lakes and anthropogenic sources are less
  than 10.
• Mahowald et al. (2004) TOMS does not allow to
  characterize all sources and anthropogenic
  sources could be as high as 50.

8
What can we learn from MODIS 500m RGB?
MODIS over Bodele
ONC map over Bodele
The dust plumes are associated with ephemeral
lakes
9
What can we learn from MODIS 500m RGB?
Bodele Clean day
Bodle dusty day
10
What can we learn from MODIS 500m RGB?
Afghanistan
Mauritania
The dust plumes are associated with ephemeral
lakes
11
What can we learn from MODIS 500m RGB?
Some dust plumes are not associated with
ephemeral lakes and appear to be of anthropogenic
origin vehicles or dry lake due to excessive
irrigation.
Morocco
Aral sea
Laayoune
Es Semara
Dust plumes NOT always associated with ephemeral
lakes
12
What can we learn from MODIS 500m RGB?
Nevada
Eastern Taklamakan
Dust plumes NOT always associated with ephemeral
lakes
13
Mineral Dust Aerosols TOMS AI frequencies and
ephemeral lakes
Some ephemeral lakes are active others not. TOMS
AI can help determine which one are active, but
if one wants to simulate past or future climate,
or perform regional studies we need to understand
why.
14
Mineral Dust Aerosols TOMS AI frequencies and
ephemeral lakes
15
Conclusions

• Use of TOMS dust sources are essentially
  associated with topographic depressions (Prospero
  et al., 2002). Based on this finding several
  global inventories are available (e.g. G-2001
  Z-2003 T-2004) which reproduce essential global
  characteristics of dust observations. These
  global inventories offer the advantage to by-pass
  complex physical models which needs highly
  variable surface soil characteristics.
• TOMS AI drawbacks no physical meaning, poor
  resolution
• Use MODIS Are the dust sources associated with
  ephemeral lakes? Some lakes are active others are
  not. Anthropogenic sources can be detected.
• MODIS drawbacks no retrieval of aerosol
  properties over bright surface (NB. Ch. Hsu and
  Deep blue algorithm alleviates the problem),
  false color can be misinterpret due to Rayleigh
  scattering.
• Use of TOMS and MODIS classification of active
  and non-active ephemeral lakes. But painstaking
  job and impossible to automize.
• Future OMI (TOMS successor) provides nUV and
  visible radiances at higher resolution. AIRS
  astonishing possibilities to retrieve 3D dust!!!

16
Acknowledgements

• J. M. Prospero and his dust newsletter
• J. Descloitres (MODIS)
• TOMS Science team for Aerosol Index
• MODIS Science team for level1B radiances