pdf_mtg_em_rep11

Cloud Top Pressure from Geostationary Satellite Observations in the Oxygen A-Band : 

Cloud Top Pressure from Geostationary Satellite Observations in the Oxygen A-Band Jürgen Fischer, Rene Preusker Carsten Rathke Free University of Berlin

User requirements : 

User requirements

Outline : 

Outline Review of existing concepts Oxygen A-band principle, history, algorithm, applications CO2 slicing Assessment Sensitivity, error analysis Summary System requirements Spectral, radiometric, auxiliary data Conclusions and recommendations

Review of existing concepts to derive cloud top pressure : 

Review of existing concepts to derive cloud top pressure Part I

Principle (1) : 

Principle (1) radiance ratio between an absorbing and a window channel depends on photon path length photon path length is mainly determined by air-mass above the cloud = cloud top pressure algorithm has to account for: multiple scattering surface reflection multi-layer cloud system

Principle (2) : 

Principle (2)

History : 

History 1961 Yamamoto and Wark: “Determination of Cloud Altitude from Satellite“ 1967 Gemini-5 1996 POLDER (†) 1996 MOS 2002 MERIS (2003 GLI)

Algorithms (1) : 

Algorithms (1) Apparent pressure method: Calculation of air mass from geometry Vicarious calibration Inversion method Radiative transfer calculations Inversion using Regression/LUT/fast RTM /artificial neural networks Consistency

Algorithms (2) : 

Algorithms (2) 1. Radiative transfer: Matrix operator model Mie-theory Spectral optical properties 1D (3D?) - simulations

Algorithms (3) : 

Algorithms (3) 2. Inversion : Regression LUT Fast forward models + Artificial neural networks

Algorithms (4) : 

Algorithms (4) 3. Consistency check fast forward (ann) calculations/estimations from retrieved parameter comparison (inversion of the inversion)

Applications (1) : 

Applications (1) CTP from aircraft measurements ACE 2 campaign 1997

Application (2.1) : 

Application (2.1) CTP from MOS

Application (2.2) : 

Application (2.2) CTP from MOS

Application (2.3) : 

Application (2.3) MOS vs. radiosondes

Application (2.4) : 

Application (2.4) MOS vs. radiosondes

Application (2.5) : 

Application (2.5) MOS vs. radiosondes

Application (2.6) : 

Application (2.6) MOS vs. ATSR2

CO2-sclicing (1) : 

CO2-sclicing (1) Normalized weighting functions for the HIRS-1 channels (from Wielicki and Coakley [1981])

CO2-sclicing (1) : 

CO2-sclicing (1) Wylie and Menzel [1989]

CO2-sclicing (3) : 

CO2-sclicing (3)

Assessment of geo-stationary satellite observations to derive cloud top pressure : 

Assessment of geo-stationary satellite observations to derive cloud top pressure Part II

Sensitivity (1) : 

Sensitivity (1) ffff

Sensitivity (2) : 

Sensitivity (2) fff

Sensitivity (3) : 

Sensitivity (3) ctp=500 hPa

Sensitivity (4) : 

Sensitivity (4) ctp=500 hPa

Sensitivity (5) : 

Sensitivity (5) .

Sensitivity (6) : 

Sensitivity (6) fff

Procedure (1) : 

Procedure (1) Inversion Real world data: Lin Lout albedo geometry RTM ctp

Procedure (2) : 

Procedure (2) . Hidden layer Input layer Output layer L0 R a Jv JS j ctp=s(Wout#s(Win#I))

Error (1) : 

Error (1) .

Error (2) : 

Error (2) .

Error (3) : 

Error (3) .

Summary (1) : 

Summary (1) CTP has been retrieved successfully from satellite measurements in the O2A-Band (POLDER, MOS) and aircraft campaigns The accuracy is in the order of 30hPa (MOS vs. radiosondes; MOS vs. ATSR2) MERIS will give the opportunity to test and improve algorithms within the next 5-? Years.

Summary (2) : 

Summary (2) Lin/Lout = R  ctp x=dR/(R dctp) Sensititivity x is between 0.03 %/hPa and 0.2%/hPa for a bandwitdh between 10nm and 0.7nm Surface influence TOA radiance (even for thick clouds) R is sensitive to cloud optical and geometrical thickness  x is variable with respect to cloud optical thickness and surface albedo

Summary (3) : 

Summary (3) Error is between 20 and 70hPa as a function of cloud optical/geometrical thickness and surface albedo Accuracy of albedo should be better than 0.1 Geometry does not influence accuracy

Summary (L) : 

Summary (L) Wishlist: 2 reference channels (to correct spectral surface albedo) 2 absorption channels (to correct for penetration depth)