ADMAeolus Spinoff Products and Retrieval Algorithms

1
ADM-Aeolus Spin-off Productsand Retrieval
Algorithms

• Marie-Laure Denneulin
• Pierre Flamant, Juan Cuesta
• Alain Dabas, Dorit Huber

2
Overview

• ADM-Aeolus signals and L2A products
• Dual Channel Receiver Mie channel and Rayleigh
  channel
• Inherent Cross-Talk needs to be corrected
• Bin-accumulated signal cannot be process with
  common techniques
• L2A Products for aerosols and clouds to be
  retrieved by range bin
• L2A algorithm Overview
• Feature Finder Algorithm
• Processing Algorithms Rayleigh and Mie Channel
  Algorithms
• Scene Classification Algorithm
• Rayleigh Channel Processing Algorithm
• Accumulated Lidar Signal
• Retrieval of particle extinction and location of
  particle layer inside the bin
• Numerical Testing
• Mie Channel Processing Algorithm
• Accumulated Lidar signal
• Optical Depth Retrieval Algorithm

3
ADM-Aeolus signals and L2A products

• ADM-Aeolus is a High Spectral Resolution Lidar
  implementing two channels for detection of
  particule and molecule signals
• Mie Channel Fizeau Interferometer
• Rayleigh Channel Dual Fabry-Perot
  Interferometer
• Cross-Talk between Mie Channel and Rayleigh
  channel due to the receiver configuration
• Calibration allow to correct cross-talk.
• Need for matching bins for the two channels

4
ADM-Aeolus signals and L2A products

• ADM-Aeolus signals are accumulated over range
  bin
• Accumulated signals can not be processed using
  common inversion techniques.
• A new solution is implemented in the L2A
  processor.
• Aerosols and clouds products to be retrieved by
  range bin
• Altitude, layer geometrical depths better than
  range bin
• Local Optical depths
• Backscatter ratio
• Scattering ratio

5
L2A processor
Inputs
FEATURE FINDER Occurrence of particle layers
based on Mie Channel Particle in range bin i ?
NO YES
ADM Aeolus L1B data Useful Signal, SNR
Outputs

• Particle layer height and geometrical thickness
• Cloud/Aerosol fraction over granularities
• Scattering ratio (first guess)

Auxiliary Data
Climatology of kp for different kinds of particles
PROCESSING ALGORITHM OpticalProperties
To be used when Mie Channel stands alone
SCENE CLASSIFICATION ALGORITHM
Meteorological information
Temperature, pressure, water vapor, cloud cover,
water and ice content,
At scale of an observation (50 km)
6
Rayleigh Channel Processing Algorithm (1/3)

• Bin-accumulated signal
• A solution does not require auxiliary data on
• Auxiliary information about molecule extinction
  and backscatter coefficient is required
• Normalized Integrated two-way transmission
• Geophysical auxiliary data (temperature,
  pressure) are required to compute a molecular
  signal without particle extinction
• Normalization

Molecular Transmission from the satellite to
the top of the bin i
Particle Transmission from the satellite to the
top of the bin i
Instrumental transmission
7
Rayleigh Channel Processing Algorithm (2/3)

• Simple solution (1st order Taylor expansion)
• 3 approximations constant particle extinction,
  full filled bin, small LOD
• New Solution for a single layer between za and zb
  
• 1 approximation constant particle extinction

8
Rayleigh Channel Processing Algorithm

• At present, 7 filling cases for a bin are
  considered in processing
• Validation of processing using a Credibility
  Criterion defined as
• To validate processing from bin 1 to bin i, the
  credibility criterion is computed for the bin i1
• CCi1 lt1 Continuation, go to bin i1 to process
  SLOD
• CCi1 1 Acceptation, All processing above are
  accepted as correct
• CCi1 gt1 Stop, try other filling cases

7 different equations for the processing of each
bin
9
Rayleigh Channel Numerical testing
A cloud between 5.35 km and 6.85 km with an
Optical Depth 0.3 (SLOD0.38), partially
filling 2 bins and filling completely one bin,
cross-talk corrected
Retrieval Total SLOD 0.3836 (Total LOD
0.3038)
10
Mie Channel Processing Algorithm

• Bin-accumulated signal
• A solution requires auxiliary data on
• Particle backscatter-to-extinction ratio
• Molecular extinction coefficient
• The SLOD is retrieve with

Molecular Transmission from the satellite to
the top of the bin i
Particle Transmission from the satellite to the
top of the bin i
Instrumental transmission
Mean Molecular Transmission for the bin i
11
Mie Channel Processing Algorithm
A cloud between 5.35 km and 6.85 km with an
Optical depth 0.3 (SLOD0.38), partially
filling 2 bins and filling completely one bin,
cross-talk corrected
Retrieval Total SLOD0.3732 (Total
LOD0.2955)
12
Conclusion

• ADM-Aeolus is a High Spectral Resolution Lidar
  implementing two receivers to sample molecule and
  particle backscatter signal.
• The inherent cross-talk needs to be corrected. A
  good calibration of the Mie and the Rayleigh
  channels is needed. Matching bins are required to
  correct cross-talk.
• Rayleigh channel a new processing algorithm has
  been proposed and tested to retrieve
• the Local Optical Depths
• the location of the particle layer inside a bin
  at a resolution better than the bin depth.
• The retrievals are validated using a Credibility
  Criteria
• Mie Channel a standard solution based on
  backscatter-to-extinction ratio i.e. kp auxiliary
  data set, has been implemented to retrieve LOD
  for each bin.