Measurements of Stratospheric Volcanic Aerosol Optical Depth
from NOAA/TOVS Observations
Clémence
Clémence PIERANGELO
PIERANGELO11,, Alain
Alain CHEDIN
CHEDIN11,, Patrick
Patrick CHAZETTE
CHAZETTE22
11Laboratoire de Météorologie Dynamique, Institut Pierre-Simon Laplace – Ecole Polytechnique, 91128 Palaiseau cedex, FRANCE
Laboratoire de Météorologie Dynamique, Institut Pierre-Simon Laplace – Ecole Polytechnique, 91128 Palaiseau cedex, FRANCE
22Laboratoire des Sciences du Climat et de l’Environnement, Institut Pierre-Simon Laplace – CEA, 91191 Gif sur Yvette Cedex, FRANCE
Laboratoire des Sciences du Climat et de l’Environnement, Institut Pierre-Simon Laplace – CEA, 91191 Gif sur Yvette Cedex, FRANCE
Introduction :
In June 1991, the eruption of Mount Pinatubo (Philippines) injected about 20 Mt of sulfur dioxide into the stratosphere. The rapid conversion of sulfur dioxide into sulfuric acid (H2SO4)
droplets caused a significant extinction, both in the visible and infrared. Here, we present a method to retrieve the Pinatubo aerosol optical depth at three infrared wavelengths (4.0 µm, 8.3 µm
and 11.1 µm) from HIRS-2 measurements onboard NOAA-10 and NOAA-12 and collocated radiosonde measurements. No assumption on aerosol size distribution or refractive index is used,
which is a major advantage of our method.
2. Simulation of the impact of volcanic aerosol on observations
1. Signature of Pinatubo aerosol on CM-3 and CM-12 time series
Radiative transfer simulations with absorbing aerosols
Experimental signature
Colocations radiosonde measurements / HIRS observation
∆Iν (sat) = εν (surf)τ ν (surf)(1− τ νa(surf))Bν (Tsurf)
m
sat
(
+ ∑ Bν [T(k)] τ νm (k + 1)(1− τ νa(k + 1 )) − τ νm (k)(1− τ νa(k))
H2O, T profiles
Clear-sky Radiative
Transfer Model
3R-N
k =0
Theoretica l signature
Variability of the simulated brightness temperature difference, ∆BT, between
an aerosol-free and an aerosol-loaded atmosphere at 8.3 µm (channel 10),
with total water vapor content, for the 2311 atmospheric situations of the
TIGR data base.
→ the aerosol effect highly depends on the atmospheric situation :
→ careful selection of TIGR atmospheric situations for the retrieval scheme
Deltac = BTcalc. - BTobs. (« experimental signature »)
3- or 12-month running mean → CM3 or CM12
NOAA 12 – CM3
0
-0.6
8707
8807
8907
9007 9101 9107
date (yymm)
8807
1.2
20S 20N
20N 60N
20S 60S
b)
1
0.6
0.4
0.2
0
-0.2
9207
9307
9407
-1
9107
9507
9207
9307
9407
date (yymm)
date (yymm)
channel 8 - CM-3
channel 5 - CM-3
20S 20N
20N 60N
20S 60S
d)
0.8
0.6
0.4
0.2
0
-0.2
1.5
b)
1.5
20S 20N
20N 60N
20S 60S
1
0.5
0
-0.5
d)
9507
20S 20N
20N 60N
20S 60S
1
0.5
0
-0.5
-0.4
8907
-0.6
8707
9007 9101 9107
8807
date (yymm)
8907
-1
9107
9007 9101 9107
9207
date (yymm)
top pannels : CM3 : 3-month centered running mean
deviation (calculated – observed) for three latitude
bands (20N-60N, 20S-20N, 20S-60S), for NOAA10 (left
pannel) and NOAA12 (right pannel).
9407
-1
9107
9507
9207
date (yymm)
a)
0.5
0
-0.5
-1
9107
9207
9307
9407
c)
4. Validation of the results
0
-0.5
channel 10
-1
9107
9507
20S 20N
20N 60N
20S 60S
1
9207
Mean dev. (K)
b)
0.5
d)
-0.5
9207
9307
9407
9507
20S 20N
20N 60N
20S 60S
1
1.8
9507
channel 10
1.8
lidar data + model
CM3 - sea
CM3 - sea and night
CM3 - total
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
lidar data + model
CM12 - sea
CM12 - sea and night
CM12 - total
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
9107
0.5
• seasonal water vapour cycle
• NOAA12 drift since early 1994
• greater diurnal cycle of temperature and
date (yymm)
NOAA 12 – CM12
humidity over land: →only sea collocations
• solar contamination of channel 18 : →only sea-night collocations for channel 18
0
-1
9107
9407
channel 5 - CM-12
1.5
20S 20N
20N 60N
20S 60S
1
9307
date (yymm)
channel 8 - CM-12
1.5
9507
0.5
date (yymm)
(a): channel 10 (8.3 µm), (b): channel 8 (11.1 µm), (c):
channel 18 (4.0 µm), (d): channel 5 (14.0 µm)
9407
channel 18 - CM-12
1.5
20S 20N
20N 60N
20S 60S
1
9307
date (yymm)
channel 10 - CM-12
1.5
bottom pannel : CM12 : 12-month centered running
mean deviation for NOAA12.
Potential sources of error :
9307
Variability of ∆BT with the optical depth and the altitude of the aerosol layer for channels 8, 10 and 18
→Linearity between ∆BT and the aerosol optical depth
Eq.(1) becomes : ∆I = ∑ akδ k
(2)
→ no effect of the altitude
∆
I
=
a
δ
(3)
Eq.(2) becomes :
a depends on the atmospheric and surface situation, on the wavelength, but not on the aerosol properties
mean deviation (K)
8807
Mean dev. (K)
-0.6
8707
Mean dev. (K)
-0.4
9007 9101 9107
0
-0.5
date (yymm)
Mean dev. (K)
Mean dev. (K)
1
0.8
Mean dev. (K)
1.2
8907
-1
9107
channel 5 - CM-3
Pinatubo eruption
channel 8 - CM-3
0
-0.5
20S 20N
20N 60N
20S 60S
0.5
mean deviation (K)
0.2
-0.2
-0.4
-0.6
8707
0.5
c)
1
9201
9207
9301
9307
9107
9201
-0.5
-1
9107
1.8
9207
9307
9407
9507
date (yymm)
channel 8
1.8
lidar data + model
CM3 - sea
CM3 - sea and night
CM3 - total
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
lidar data + model
CM12 - sea
CM12 - sea and night
CM12 - total
1.6
1.4
1.2
1
9201
9207
9301
3. Retrieval of aerosol optical depth
0.6
0.4
0.2
9307
9107
9201
0.02
0.01
0
-0.01
8707 8807 8907 9007 9107 9207 9307 9407 9507
0.05
0.03
0.02
0.01
0
-0.01
8707 8807 8907 9007 9107 9207 9307 9407 9507
date (yymm)
date (yymm)
20N 60N - 3 month running mean
0.05
11.10
8.30
4.00
0.04
0.03
0.02
0.01
0
20N 60N - 12 month running mean
0.06
aerosol optical depth
aerosol optical depth
0.06
-0.01
8707 8807 8907 9007 9107 9207 9307 9407 9507
0.05
0.05
11.10
8.30
4.00
0.04
0.03
0.01
0
date (yymm)
date (yymm)
20S 60S - 3 month running mean
20S 60S - 12 month running mean
11.10
8.30
4.00
0.04
0.03
0.02
0.01
0
-0.01
8707 8807 8907 9007 9107 9207 9307 9407 9507
date (yymm)
0.06
0.05
Deltac
a
No assumption on the aerosol size distribution or
refractive index
9307
1.2
1
0.8
0.6
0.4
0.2
0
lidar data + model
CM12 - sea
CM12 - sea and night
CM12 - total
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
9107
9201
9207
9301
date (yymm)
9307
9107
9201
9207
9301
date (yymm)
9307
Spatial and time behaviour of the aerosol
loading :
- transport faster to southern latitudes (peak in
November 1991 for 20S-60S, in April/May
1992 for 20N-60N) [Lambert et al., 1993]
- aerosol removal faster in the tropics than in
midlatitude zones
Comparison between the experimental CM-3 (left hand-side) or CM-12 (right hand-side) time series from
NOAA12, and their simulated ∆BT (in red), using TIGR climatological atmospheric profiles and lidar
aerosol profiles at Observatoire de Haute Provence (France) (3 month and 12 month running means)
[Chazette et al., 1995].
δ=
0.02
-0.01
8707 8807 8907 9007 9107 9207 9307 9407 9507
aerosol optical depth
aerosol optical depth
0.06
11.10
8.30
4.00
0.04
mean deviation (K)
0.06
mean deviation (K)
11.10
8.30
4.00
0.03
aerosol optical depth
aerosol optical depth
0.05
0.04
1.4
9301
channel 18
1.8
lidar data + model
CM3 - sea
CM3 - sea and night
CM3 - total
1.6
9207
date (yymm)
channel 18
1.8
Inversion is based on Eq.(3).
For each collocation, and each channel :
¾ Deltac (BTcalc-BTobs cf 1.)
¾ ∆BT calculated from the closest TIGR situation for
a reference optical depth (0.01), cf 2.
→ Computation of the corresponding coefficient a,
for each wavelength (a = ∆BT / 0.01)
Computation of the maximum particle
number :
- as expected, does not depend on the channel
- for August-July 1991 in the northern
midlatitudes, we retrieve 10 to 20 particles/cm3,
in agreement with in-situ measurements of
Deshler et al.,[1992, 1993]
0.8
date (yymm)
0.06
9307
0
9107
20S 20N - 12 month running mean
9301
date (yymm)
channel 8
0
20S 20N - 3 month running mean
9207
date (yymm)
0
mean deviation (K)
0
-0.4
channel 18 - CM-3
1.5
20S 20N
20N 60N
20S 60S
Mean dev. (K)
0.2
-0.2
0.4
a)
1
Mean dev. (K)
0.4
0.6
channel 10 - CM-3
1.5
Mean dev. (K)
0.6
20S 20N
20N 60N
20S 60S
c)
0.8
Mean dev. (K)
1
Mean dev. (K)
Mean dev. (K)
channel 18 - CM-3
1.2
20S 20N
20N 60N
20S 60S
a)
mean deviation (K)
NOAA 10 – CM3
channel 10 - CM-3
1
0.8
∆I : difference between the radiance emerging at the satellite
level for an aerosol free and an aerosol loaded atmosphere
ε : surface emissivity
τm(k): molecular transmittance between level k and satellite
τa(k): aerosol transmittance between level k and satellite
Sensitivity study
Calculated BT (clear sky)
1.2
(1)
Observed BT (with aerosols)
)
11.10
8.30
4.00
0.04
0.03
0.02
0.01
0
-0.01
8707 8807 8907 9007 9107 9207 9307 9407 9507
date (yymm)
Retrieved optical depth at three infrared wavelengths and for
three latitude zones. Left hand-side pannels: 3-month running mean,
right hand-side pannels : 12-month running mean.
From top to bottom : 20N-20S, 20N-60N, 20S-60S. Missing data
correspond to month when the number of collocations is lower than 50.
Comparison between ISAMS aerosol optical depth at 12.1 µm and HIRS-2
aerosol optical depth at 12.1 µm retrieved from channels 8, 10, and 18.
Conclusions :
Covering the time period July 1987- September 1995, this analysis of differences between HIRS-2 measurements and collocated temperature and water vapor radiosonde profiles clearly reveals
the signature of Mount Pinatubo aerosols. Simulations of their effect on brightness temperature using the climatological data base TIGR shows that, for a given channel, it depends on the
atospheric situation and linearly on the aerosol optical depth, and not on their microphysical properties. 3- or 12- month running mean aerosol optical depth at 4 µm, 8.3 µm and 11.1 µm are
retrieved from channels 18, 10 and 8 measurements, for three latitude zone. Validation with lidar measurements and comparison with ISAMS measurements show good agreements.