February 25, 2003

1
Level 2 Status

• Sung-Yung Lee and Edward Olsen
• Chris Barnet and John Blaisdell
• February 25, 2003

2
Summary

• V2.7 Delivered to DAAC
• Functioning level 2
• Good yield over ocean surface
• Bias only tuning
• Coefficient update planned
• No regression step without coefficient update
• Convergence of GSFC research version and Team
  software.
• Many minor updates to both software programs
• Version 2.8 has all the fixes for convergence
• Convergence effort will continue in the
  background
• Work on V3.0 starts
• AIRS RTA V7 implemented
• Regression/Tuning/AngCorr coefficients update
• Sidelobe Correction Algorithm delayed
• March 11 deadline for software updates
• Data processing for validation report
• Bug fixes will continue

3
Team Software for Level 2

• Unified Team software
• Various Algorithms and software modules are
  supplied and maintained by science team members
• Each team member develops algorithm/software at
  their home institution and ports the software to
  the team software.
• This causes time lag between research versions
  and team software.
• Unique among EOS instrument teams
• Roles
• JPL team is responsible for data flow, I/O
  routines, bullet proofing, etc
• GSFC team is responsible for cloud clearing,
  final algorithm
• NESDIS members are responsible for regression
  guess, tuning, angle correction
• MIT members are responsible for MW only
  algorithms including precipitation
• UMBC is responsible for RTA

4
PGE V2.7 Retrieval

• Delivered to DAAC
• To go operational in early March
• No public release of level 2 data
• Retrieval software is functioning reasonably well
• High quality retrieval for clear ocean cases
• Global average yield of 60 over ocean
• New option to use GSFC tuning in brightness
  temperature
• Regression Coefficients
• Valid over ocean
• New version based on cloud cleared radiances is
  expected soon
• MW Tuning available
• GSFC bias only tuning with respect to ECMWF
• NESDIS Tuning with respect to NCEP
• IR Tuning
• GSFC bias only tuning in br temp
• NESDIS tuning in radiance unit
• Bias only tuning works
• UMBC bias only tuning

5
Status of Convergence

• Convergence is essentially achieved in V2.8
• Subtle bugs were found and fixed both at JPL and
  GSFC
• Missed updates were completed on both sides
• Choices in processing were made more equal
• Lingering small differences still to be checked
  as time permits
• Would have been impossible to find these things
  without offline GSFC system for comparison.
• One remaining issue
• use of quality flags to determine which channels
  to use in processing (what is bad?)
• Must repeat convergence with V3.0 build to check
  updates and test additional conditions

6
QA Flags for MW instruments

• Team software does not check the MW QA flags
• QA_scanline
• QA_reciever_a11
• QA_channel
• Currently 7th bit (noisy bit) is turned on few
  times a granule
• All of the fields that denote serious problems
  will cause BTs to be set to -9999.
• All other bits diagnose situations which are
  usually harmless.
• Recommendation is to ignore the QA flags, but
  check the radiances for -9999.

7
Level 2 PGE Version 2.8

• Version 2.7 with convergence fixes
• Intermediate version for validation exercise
• Preview of the main validation run scheduled for
  mid-March
• Reports at this team meeting are based on this
  run
• 3 days (Sept 6, Sept 29, Nov 16) of normal
  retrieval
• Matchup retrievals for Ra Obs, Fixed sites, and
  SurfMarine
• Sept 6, Sept 26 - Oct 2, Nov 16
• In /archive/AIRSOps/test/ValRun1/

8
Retrieval Statistics vs ECMWF

• RMS difference to ECMWF forecast
• Granule 50 of Sept 6, a night granule over
  tropical Atlantic
• Blue Curve
• 76 Clear Cases (HHA)
• Black Curve
• All cases (1200)
• Version 2.7.12, with regression.

9
Global ECMWF Comparison

• Ocean cases
• abs(lat) lt 40
• zen angle lt 40
• Black curve is bias
• Blue curve is RMS difference
• Regression will improve statistics

10
Browse ProductsTotal Precipitable Water Vapor
11
Browse ProductsSurface Skin Temperature
12
Plan for V3.0

• Converged software is the starting point
• Scheduled delivery of May, 2003
• First build of v3.0 will be during the second
  week of March
• Validation report will be based on this build
• Bug fixes continue
• Validation of this version will focus on near
  nadir ocean cases
• AIRS RTA version 7
• Tuning, angle correction need new coefficients
• Use MW antenna temperature instead of brightness
  temperature
• Sidelobe correction delayed until 3.x
• New tuning algorithm expected from McMillin
• New Regression algorithm expected from Goldberg
• UARS Climatology update - fix water/ozone profile