CAFS ozone data for validation of the Aura ozone products'

1
CAFS ozone data for validation of the Aura ozone
products.

• I. Petropavlovskikh (CIRES)
• R. Shetter, S. Halls (NCAR)
• P.K. Bhartia (NASA/Goddard)
• L. Froidevaux (JPL)

2
Barry Leffer (Univ. of Houston)
Sam Halls (NCAR)
Rick Shetter (NCAR)
3
CAFS instrument actinic flux measurements

• Custom multi-shell frosted quartz optical
  collectors on the top and bottom of the aircraft
  with fiber optic couplings to a monolithic
  monochromators with CCD detectors.
• AVE2004 and PAVE2005 collects photons reaching a
  point in the atmosphere from the 2p steradians
  above or below the aircraft.
• AVE2005 CAFS collector design is modified to
  block photons coming from directions near the
  horizon.

4
Down and Up-Welling Actinic Flux Data
5
AURA validation campaigns.

• Fall AVE 2004, Texas, mid-latitude flights
• First employment
• 7 flights
• Winter PAVE 2005, Pease Tradeport (NH), mid- and
  high-latitude polar flights
• Low sun conditions
• Large ozone variability
• 8 flights
• Summer AVE 2005, Texas, mid-latitude flights
• Modified design to reduce scattered light
• Hurricane over-flights
• 7 flights

6
CAFS and MLS match

• Use latitude matches as selection criteria
• Integrate MLS profiles above CAFS pressure level
• Using OMI TO data estimate spatial ozone
  variability across the 250 km MLS footprint
• Estimate measurement noise in CAFS data across
  the MLS footprint (1200 sec)
• Estimate CAFS errors due to profile shape using
  AK and difference between MLS and standard ozone
  profiles

7
CAFS and MLS latitude match MLS profiles
integrated above CAFS pressure level
8
Errors in CAFS and MLS data

• MLS errors combination of 10 precision and
  spatial ozone variability
• Use OMI TO data to estimate spatial ozone
  variability across the 250 km MLS footprint
• CAFS errors measurement noise and profile
  sensitivity
• Measurement noise in CAFS data across the MLS
  footprint (1200 sec)
• Use AK and difference between MLS and standard
  ozone profiles to estimate CAFS profile
  sensitivity

9
CAFS AK and errors at 12 km
Error, 0 1 2 3
72 76 80
84 SZA
Altitude, km
AK
10
CAFS ozone retrievals are insensitive to the
underlying surface albedo (340 nm)
11
AVE 2004
12
AVE 2004, Nov 3 flight
18km
8km
MLS clear MLS cloudy CAFS data CAFS error
Altitude
13
Summary of AVE 2005 comparisonsNov 3, 5, 7, 9,
10, and 12
MLS CAFS
14
Mean offset is 5.5 , RMSD is 4.5 , slope is 0.71
MLS CAFS
15
PAVE 2005
16
Polar AVE campaigns 2005
12 km
MLS clear MLS cloudy CAFS data CAFS error
Altitude
17
Summary of PAVE 2005 comparisonsJan 27, 31 and
Feb 5
MLS CAFS
18
Mean offset is 2 , RMSD is 9 , slope is 0.9
MLS CAFS
19
Conclusions

• For SZAlt 70º CAFS have errors less than 1
• For SZA gt 70º CAFS errors increase to 3
• AVE 2004 (high altitude flights)
• MLS/CAFS agreement within 5 , with 5 offset
• small latitude dependence between 20 and 40 N
• PAVE 2005 (middle altitude results)
• MLS/CAFS agreement within 9 , small 2 offset
• significant longitude dependence
• MLS/CAFS offset does not show any particular
  altitude dependence
• Continue analysis with AVE 2005 data