Development of Algorithm for the Development of Algorithm for the Retrieval of Atmospheric Profiles from Retrieval of Atmospheric Profiles from Infrared Sounder onboard INSAT Infrared Sounder onboard INSAT - - 3D 3D Pradeep K Thapliyal M V Shukla, P C Joshi and P K Pal M V Shukla, P C Joshi and P K Pal Meteorology & Oceanography Group (MOG) Meteorology & Oceanography Group (MOG) Space Applications Centre (SAC) Space Applications Centre (SAC) Indian Space Research Indian Space Research Organisation Organisation (ISRO) (ISRO) Ahmedabad Ahmedabad - - 380015 (INDIA) 380015 (INDIA) International TOVS Study Conference (ITSC-17), Monterey, CA (USA) 14-20 April 2010 [email protected][email protected]
13
Embed
Development of Algorithm for the Retrieval of Atmospheric ...Development of Algorithm for the Retrieval of Atmospheric Profiles from Infrared Sounder onboard INSAT-3D Pradeep K Thapliyal.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Development of Algorithm for the Development of Algorithm for the Retrieval of Atmospheric Profiles from Retrieval of Atmospheric Profiles from Infrared Sounder onboard INSATInfrared Sounder onboard INSAT--3D3D
Pradeep K ThapliyalM V Shukla, P C Joshi and P K PalM V Shukla, P C Joshi and P K Pal
Meteorology & Oceanography Group (MOG)Meteorology & Oceanography Group (MOG)Space Applications Centre (SAC)Space Applications Centre (SAC)
Indian Space Research Indian Space Research OrganisationOrganisation (ISRO)(ISRO)AhmedabadAhmedabad--380015 (INDIA)380015 (INDIA)
International TOVS Study Conference (ITSC-17), Monterey, CA (USA) 14-20 April 2010
is the first guess profile, is the first guess profile, YYmm
radiance measurements, and radiance measurements, and Y(X)Y(X)
is forward model.is forward model.HH
is a priori matrix that constrains the solution (e.g. first gueis a priori matrix that constrains the solution (e.g. first guess error ss error covcov. matrix).. matrix).EE is expected radiance error covariance matrix.is expected radiance error covariance matrix.
Physical RetrievalPhysical Retrieval
••
Minimization of the cost function using nonlinear Newtonian iterMinimization of the cost function using nonlinear Newtonian iteration yields ation yields the following iterative solution:the following iterative solution:
XX
n+1n+1
= X= X
00
+ (+ (KK
nn
TT
EE--11
KK
nn
+ + γγ
nn
I)I)--11. {. {KK
nn
TT
EE--11
[[δδYY
nn
mm
+ + KK
nn
((XX
nn
––
XX
00
)] + )] + γγ
nn
(X(X
nn
––
XX
00
)})}
••
Iterative solution in terms of eigenvectors is:Iterative solution in terms of eigenvectors is:
ff
n+1n+1
= (= (ξξ
nn
TT
EE--11
ξξ
nn
+ + γγ
nn
I)I)--11. {. {ξξ
nn
TT
EE--11
[[δδYY
nn
mm
+ + ξξ
nn
.f.f
nn
] + ] + γγ
nn
.f.f
nn
}}
wherewhere,, ξξ
= K.V= K.V, , and and VV
is eigenvector matrix, and is eigenvector matrix, and ff
is is coefficient vectorcoefficient vectorVV contains 5 contains 5 EOFsEOFs
for temperature profile, 3 for temperature profile, 3 EOFsEOFs
Iteration stops when Iteration stops when (d(dn+1n+1 –– ddnn)) < threshold< threshold (~ 0.1). (~ 0.1). If If ddn+1n+1 > > ddnn thenthen γγnn is increased.is increased.
––
Brightness temperature residual testBrightness temperature residual test::
RMS radiance residual is defined as: RMS radiance residual is defined as:
If If rrn+1n+1 ≤≤ rrnn the iteration continues until the iteration continues until rr
n+1n+1
is acceptably small (less then NEDT)is acceptably small (less then NEDT)
∑=
++ −=nch
knkkn nchxyYr
1
21
21 /)]([
Convergence TestConvergence Test
Sample ResultSample Result
SUMMARYSUMMARY
INSATINSAT--3D having 19 channel Sounder is scheduled for 3D having 19 channel Sounder is scheduled for launch in Dec 2010launch in Dec 2010
Retrieval package based on twoRetrieval package based on two--step algorithm ready for step algorithm ready for INSATINSAT--3D3D
Accuracies of the retrieved products are comparable to the Accuracies of the retrieved products are comparable to the similar products from other missions (Simulation study).similar products from other missions (Simulation study).
Algorithm to be developed for cloud property Algorithm to be developed for cloud property retrieval/cloudy sky retrievals.retrieval/cloudy sky retrievals.
Further improvements/developments under Further improvements/developments under ISROISRO--SSEC SSEC MoUMoU
Special Thanks toSpecial Thanks to……..ITWG for financial support to attend ITSCITWG for financial support to attend ITSC--1717SSEC/UWSSEC/UW for providing visiting fellowship (2005for providing visiting fellowship (2005--06)06)Allen HuangAllen Huang and and Jun LiJun Li for providing guidance and for providing guidance and support in the retrieval algorithmsupport in the retrieval algorithmHaalHaal WoolfWoolf for the RT Modelfor the RT ModelS S SuzaneSuzane and and E E BorbasBorbas for the for the SeeBorSeeBor dataset and dataset and spectral spectral emissivityemissivity datasetdataset