Pr

1
From dust microphysical properties to dust
hyperspectral infrared remote sensing
Clémence Pierangelo(1), Michael Mishchenko(2),
Alain Chédin(1) (1) Laboratoire de Météorologie
Dynamique - Institut Pierre Simon Laplace, Ecole
Polytechnique (2) Goddard Institute for Space
Sciences NASA clemence.pierangelo_at_lmd.polytechni
que.fr
Why studying dust in the infrared?

• Advantages
• night and day detection
• sensitivity to dust vertical distribution
  altitude retrieval
• sensitivity to dust mineralogical composition
  might be retrieved
• over deserts

• Limitations
• Spatial resolution (20km)
• Aerosol refractive indices poorly known high
  spectral dependency
• High sensitivity to temperature and gas profiles
• No direct validation

IPCC 2001 dust radiative forcing poorly known.
Since then, most studies focus on the visible
wavelength, whereas the closure of the Earth
radiative balance also needs knowledge of the
dust effect on terrestrial and atmospheric
infrared radiation (3.5 to 15 µm), and
computations of IR forcing from visible or
near-IR measurements are not reliable enough.
Satellite observation
Dust optical properties at infrared
wavelengths
Purpose modelling of the maximum effect of
shape spheroids with aspect ratio2 (effect
stronger than a mixture of spheroids with several
aspect ratios)
For realistic values of re (1 to 3µm), moderate
impact on ?, g and Qext normalized (max 30).
The impact of size is greater in the visible than
in the IR. The ratio of IR to visible extinction
increases with dust size.
Optical properties are greater for oblate or
prolate spheroids than spheres. The maximum
impact is about 10.

• Weak impact of the aspect ratio on the phase
  function because the size parameter is relatively
  small.
• Note that the impact of the phase function on the
  radiance at satellite level is not as crucial as
  in the visible (no reflected solar radiation).

and big variability of optical properties with
data set too! (SHADE model probably not
realistic)
Huge variability of refractive indices with
wavelength with data set
? Refractive index is a more problematic issue
than size or shape
Radiative transfer computations with aerosols at
infrared wavelengths
Simulations of AIRS (Advanced Infrared
Sounder-AQUA) brightness temperatures for 324
channels with a code coupling the line-by-line
 4A  and DISORT
BT 177 (8.14µm) BT 165 (9.33µm)(K.) Simulations
for 100 tropical atmospheric situations (Same
aerosol properties) The aerosol impact itself
depends on the atmospheric situation
First component to the signal the
temperature and the water vapor profiles.
Aerosol optical depth (AOD) and altitude impact
a few K.
Aerosol size and shape impact a few tenth of K.
Remote sensing of dust with hyperspectral
infrared sounders Application to
AIRS
2. Validation of the LUT approach
Validation of the LUT retrieved atmospheric
situation is performed comparing its surface
temperature (SST) and water vapor content (WVC)
to MODIS / SSMI observations.
Method Look-Up-Tables (LUT) built for 8 AIRS
channels, several dust altitudes and 10 µm AOD,
one aerosol model (OPAC) and almost 600
atmospheric situations (Pierangelo et al., ACP,
in press)
Robustness of the retrieval to dust microphysical
properties
The error in the retrieval caused by the use of
one fixed size distribution is below 10 for the
AOD, and still lower for the altitude
Method Channel 165 (1072cm-1) sensitive to size,
not to shape
The error in the retrieval caused by the use of
spherical particles is below 10 for the AOD, and
still lower for the altitude
Very strong effect of the refractive index on the
AOD!!! But, with the data sets dust, mineral
or SHADE, the retrieved AOD could not be
greater than 0.5? way to exclude these models
The effective radius of Saharan dust decreases
from 2.8 µm to 1.2 µm with transport.