using satellite measurements of stellar scintillation for mapping turbulence in the stratosphere
Post on 29-Jan-2016
42 Views
Preview:
DESCRIPTION
TRANSCRIPT
Using satellite measurements of stellar scintillation for mapping turbulence in the stratosphere
V. F. Sofieva(1), A.S. Gurvich(2), F. Dalaudier(3), and the GOMOS team(4)
(1) Finnish Meteorological Institute, Helsinki, Finland
(2) A.M. Oboukhov Institute of Atmospheric Physics, Moscow, Russia
(3) LATMOS (Service d’Aeronomie du CNRS), France
(4) FMI, Finland; Service d’Aeronomie, France; IASB, Belgium, ACRI-ST, France, ESA/ESRIN, Italy; ESA/ESTEC, Netherlands
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Outlines
• Methodology: using space-borne stellar scintillation measurements for studies of small-scale processes in the atmosphere
• Application: results of analysis of GOMOS/Envisat scintillation measurements
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Atmospheric layers
www.singularvortex.com
Stratosphere
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Introduction• Gravity waves, their generation, propagation
and breaking
• Global effects of (relatively) small-scale gravity waves influence the stratospheric circulation
affect ice cloud formation and polar ozone loss
they play an important role in driving atmospheric circulations (including quasi-biennial oscillation)
play an important role in controlling temperatures in the Antarctic ozone hole
• Satellite observations of stellar scintillations is the new approach for study of small-scale processes in the atmosphere
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Scintillation of stars
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Scintillation of stars: a simplified qualitative explanation
direction of the light
Tw
inkle, twinkle
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Past and present space-borne stellar scintillation measurements• Past:
EFO-2 photometer, Russian MIR station (1996-1999)
– Sampling frequency up to 16 kHz
– Central wavelength 485 nm
– Man-controlled photometer (~100 occultations, within 60° latitudinal band)
• Present: scintillation measurements by GOMOS fast photometers (since March 2002)
First bi-chromatic scintillation measurements
– Blue 470-520 nm
– Red 650-700 nm
Sampling frequency 1 kHz Global coverage Objectives
– Scintillation correction of spectrometer data
– High resolution temperature profile
– Small scale structures of air density, stratospheric dynamics GOMOS fast photometer
EFO-2 fast photometer
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
GOMOS measurements
movie
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
General strategy for retrieval of information about air density irregularities
• 3D distribution from 1D scintillation measurements? The problem is severely ill-posed
• Approach proposed in [Gurvich and Kan, 2003]
1 step. The spectrum of the air density irregularities is parameterized
2 step. 3D-spectrum of air density irregularities 1d-scintillation spectrum at observation point (The theoretical relations are found)
3 step. The parameters of the spectral model are retrieved via fitting experimental scintillation spectra
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Structure of air density irregularities
• Anisotropic irregularities Elongated in horizontal direction Generated by internal gravity
waves
• Isotropic irregularities (turbulence) Result from breaking of internal
gravity waves Dynamic instabilities
• Two-component spectral model of relative fluctuations of air density
KW
anisotropicisotropic
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Spectrum of anisotropic irregularities
• corresponds to anisotropic irregularities generated by a random ensemble of internal gravity waves
CW is the structure characteristic is the anisotropy coefficient 2/0 is the outer scale 2/W is the inner scale
• The associated 1D vertical spectrum for 0 << z << W corresponds to the model of the saturated gravity waves
3
3
2)( zWz CV
32
4 z
BVTT g
AV
WzWW
kηC
2/52
02222
2222zk
222yx
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Spectrum of isotropic irregularities
• Spectrum of locally isotropic turbulence (Kolmogorov model)
CK is the structure characteristic
K is the wavenumber corresponding to inner scale of isotropic irregularities (Kolmogorov scale)
• The corresponding 1D spectrum follows well-known -5/3 power law
• Individual turbulent patches are not resolved, CK represents the effective value in the region of interaction of light wave and the turbulent atmosphere
))/(exp(033.0)( 23/11KKK kkCk
2222zyx kkkk
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Model of 1D vertical spectra of air density irregularities
0
K
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
• Stratospheric distortion of light rays Distortion of optical path (eikonal)
• Frozen field assumption
• Phase screen approximation
• Scintillation spectrum (weak scintillation)
Modelled scintillation spectra
3d spectrum of air density irregularities
kx, ky, kz
2d spectrum of eikonal fluctuations
F (ky, kz)
2d spectrum of scintillation
1d spectrum of scintillation
raydlN r
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Scintillation spectra in phase screen approximation (weak scintillation)
F ,y z
eikonal fluctuations on phase screen plane
22
2
22
2
1exp,,
1),(F
H
Haxd
H
Ψ
z
xEzyx
zyz
2D spectrum of on phase screen plane
qz
yzyzyJ ,F,),(F
F ,J y z 2D spectrum of observed light flux on observation plane
2 22 20
0
4( , ) sin
2z
z y yk L
q k q
(Gurvich, 1984)
V ( ; ) F ( , ) sJ s y z J y z s
s
d d
κ κ
VJ is 1D spatial scintillation spectrum in observation plane
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Scintillation spectra
10-4
10-3
10-2
10-1
100
0
0.5
1
1.5
2
2.5
wavenumber, 1/m
psd/
varia
nce
Iso
0
30
60
75
0
w
Fr
K
i2/
w2 =4
),,()()( 0mod WanisoWisoK kVCkVCkV
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
The main approach to scintillation analysis
• To separate the anisotropic and isotropic components
• To retrieve essential parameters of IGW and turbulence spectra
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Inversion• Vmeas=CK Viso+CW Vaniso(W, 0)+
• Maximum likelihood method, which is equivalent to minimization of
• Combination of linear and non-liner optimization
Non-linear fit (Levenberg-Marquardt) for W and 0
Linear fit (weighted least-squares method) for CK and CW
measWanisoWisoK
TmeasWanisoWisoK CVCCVC VVSVV ),(),( 0
10
2
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Filtering quasi-periodic structures
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Examples of retrievals : 1.photometer data
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Examples of retrievals: moderate turbulence
10-3
10-2
10-1
0
2
4
6
8
10
wavenumber, 1/m
psd
Ck= 1.6e-9 m-2/3
CW
=3.6e-11 m-2
lW
= 4.7 m
L0=0.42 km
10-3
10-2
10-1
0
0.5
1
1.5
2
wavenumber, 1/m
psd
Ck= 2.3e-9 m-2/3
CW
=4e-11 m-2
lW
= 8.7 m
L0=0.76 km
10-3
10-2
10-1
0
0.2
0.4
0.6
0.8
1
wavenumber, 1/m
psd
Ck= 5.7e-9 m-2/3
CW
=4.9e-11 m-2
lW
= 15 m
L0=3.1 km
10-3
10-2
10-1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
wavenumber, 1/m
psd
Ck= 1e-8 m-2/3
CW
=1.3e-10 m-2
lW
= 36 m
L0=2.3 km
measured uncertainty used in 2nd step fit anisotropic isotropic
30 km 35 km
40 km 45 km
R02908/S001 (=67, 35S, 106W, 20 September 2002)
Sofieva et al.,
JGR, 2007
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Examples of retrievals: strong turbulence
10-3
10-2
10-1
0
1
2
3
4
5
6
7
8
9
wavenumber, 1/m
psd
Ck= 3.2e-9 m-2/3
CW
=1.8e-10 m-2
lW
= 14 m
L0=1 km
10-3
10-2
10-1
0
0.5
1
1.5
2
2.5
3
wavenumber, 1/m
psd
Ck= 2e-8 m-2/3
CW
=1.8e-10 m-2
lW
= 16 m
L0=3.5 km
10-3
10-2
10-1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
wavenumber, 1/m
psd
Ck= 7.5e-8 m-2/3
CW
=2e-10 m-2
lW
= 62 m
L0=0.78 km
10-3
10-2
10-1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
wavenumber, 1/m
psd
Ck= 1.4e-7 m-2/3
CW
=1.6e-10 m-2
lW
= 21 m
L0=1.6 km
measured uncertainty used in 2nd step fit anisotropic isotropic
30 km 35 km
40 km 45 km
R07741/S001 (=40, 62S, 7E, 23 August 2003)
Sofieva et al.,
JGR, 2007
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Indication on gravity wave breaking in polar night jet• Qualitative: Sofieva et al. (2007), Global analysis of scintillation variance: Indication of gravity wave
breaking in the polar winter upper stratosphere, Geophys. Res. Lett., 34, L03812, doi:10.1029/2006GL028132
latitude latitude
KT CTC22
Gravity wavestructure characteristic
CW(m-2)
Turbulencestructure characteristic
CT2(K2m-2/3)
Jan-Mar, Dec 2003 June-Sep 2003
Sofieva et al., GRL, 2009
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
CT2 in tropics (at 42 km)
• Turbulence intensity is smaller in tropics than in polar regions; it has a pronounced zonal structure
• Average values of CT2
follow the sub-solar latitude
• Almost all of the local enhancements are over continents
• Many of enhancements correspond well to the typical regions of deep convection
• No systematic increase of CT
2 over large mountain regions is observed
Gurvich et al., 2007, GRL
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Evolution of CT2 at 80 N
during the sudden stratospheric warming in December 2003
altit
ude,
km
Equivalent latitude (deg)
date07/12 14/12 21/12
30
35
40
45
50
altit
ude,
km
Temperature (K)
date07/12 14/12 21/12
25
30
35
40
45
50
55
date
altit
ude,
km
CT
2 ( K2 m-2/3)
07/12 14/12 21/1230
35
40
45
50
0.5
1
1.5
2
x 10-3
200
210
220
230
240
250
260
45
50
55
60
65
70
75
80
Sofieva et al., (2007): Reconstruction of internal gravity wave and turbulence parameters in the stratosphere using GOMOS scintillation measurements, Journal of Geophys. Res., 112, D12113,
doi:10.1029/2006JD007483
Turbulent mixing and beyond, ICTP, Trieste, 3 August 2009
Summary, discussion, future work
• The presented results are the start in studying small-scale air density irregularities
• Strength of scintillation method: It covers the wavenumber range of transition of GW spectrum to
turbulence -> possibility of visualizing GW breaking
Scintillation are affected by small-vertical-scale GW -> importance for GW parameterizations in GCM models
Other measurements at such small vertical resolution are scarce in this altitude range
• The processing of the 2002-Jan 2005 dataset has been recently completed, the data analyses are on-going
top related