Retrieval of BrO vertical distributions from SCIAMACHY limb measurements: Data quality assessment an

1
Retrieval of BrO vertical distributions from
SCIAMACHY limb measurements Data quality
assessment and algorithm improvements
A. Rozanov1, S. Kühl2, C. Sioris3, H.
Bovensmann1, J. P. Burrows1, M. Chipperfield9,
F. Goutail7, F. Hendrick4, M. von Hobe8, S.
Hrechanyy8, C. McLinden5, K. Pfeilsticker2, J.
Pukite2, M. van Roozendael4, R. Salawitch6,
B.-M. Sinnhuber1, F. Stroh8, T. Wagner2
1Instutute of Environmental Physics, University
of Bremen, Germany 2 Instutute of Environmental
Physics, University of Heidelberg,
Germany 3Harvard-Smithsonian Center for
Astrophysics, Cambridge, USA 4Belgian Institute
for Space Aeronomy (IASB-BIRA), Brussels, Belgium
5Meteorological Service of Canada, Toronto,
Canada 6Jet Propulsion Laboratory, Pasadena,
California, USA 7Service dAeronomie -
CNRS, France 8Juelich Research
Center, Germany 9University of Leeds, UK
2
BOOST A joint intercomparison project

• Bromine Oxide in the lOwer STratosphere (BOOST)
• Project objectives
• Comparison of BrO vertical distributions
  retrieved from SCIAMACHY limb measurements using
  different retrieval algorithms
• Investigation of possible reasons for the
  disagreement between the retrievals identified in
  previous studies
• Improvement of the existing retrieval algorithms
  (especially w.r.t. the retrieval quality in the
  lower stratosphere and the upper troposphere)
• Investigation of the sensitivity of the retrieved
  profiles to the retrieval parameters such as
  spectral range, initial profiles, cross
  sections, spectral corrections

3
BOOST participating groups

• SCIAMACHY retrieval groups
• IUP, University of Bremen Alexei Rozanov
• IUP, University of Heidelberg Sven Kühl
• Harvard-Smithsonian Center for Astrophysics
  (SAO) Chris Sioris

• Balloon-borne and ground-based measurements
• IUP, University of Heidelberg DOAS, Klaus
  Pfeilsticker
• Juelich Research Center TRIPLE, Serhiy Hrechanyy
• Service dAeronomie CNRS SAOZ, Florence
  Goutail
• Belgian Institute for Space Aeronomy
  ground-based, François Hendrick

• Modeling groups
• IUP, University of Bremen Björn-Martin Sinnhuber
  
• Jet Propulsion Laboratory Ross Salawitch
• Meteorological Service of Canada Chris McLinden
• University of Leeds Martyn Chipperfield

4
Retrieval algorithm of the University of Bremen

• Forward modeling (SCIATRAN)
• Fully spherical treatment for SS
• Approximation for MS

Simulated limb spectra Weighting functions w.r.t.
concentrations
Measured and simulated limb spectra Vertically
integrated WF

• Pre-processing (DOAS-like fit) at each tangent
  height
• Shift and squeeze correction
• Spectral corrections

Correction parameters
Measured and simulated limb spectra with all
corrections applied Weighting functions A priori
constraints

• Main inversion procedure
• Solution Information Operator or Optimal
  Estimation
• Measurement vector differential signal in all
  spectral points at all selected tangent heights
• State vector trace gas number densities at
  altitude levels

Vertical distributions of trace gas number
densities
5
Retrieval algorithm of the University of
Heidelberg

• Forward modeling (TRACY)
• Fully spherical Monte Carlo

Block Air Mass Factors

• DOAS fit for both measured and simulated spectra
  at each tangent height
• Shift
• Spectral corrections

One iteration only

• Inversion procedure
• Solution Optimal Estimation (maximum a
  posteriori)
• Measurement vector fitted slant columns
• State vector trace gas number densities at
  altitude levels

Fitted slant columns Block Air Mass Factors A
priori constraints
6
Retrieval algorithm of SAO

• Forward modeling (VECTOR)
• Fully spherical treatment for SS
• Approximation for MS

Simulated limb spectra
Measured or simulated limb spectra Cross sections

• DOAS fit for both measured and simulated spectra
  at each tangent height for a set of temperatures
• No shift squeeze correction
• Spectral corrections

Slant columns as a function of tangent height
Inversion procedure (Chahine-like)
Measured slant columns Simulated slant columns
interpolated to an appropriate temperature
First iteration
Subsequent iterations
Vertical distributions of trace gas number
densities
7
Retrieval settings
8
Initial set of the limb states to be compared

• Selection criteria (based on results from Dorf at
  al., 2006)
• At the current stage of the project comparisons
  will be performed for the limb states collocated
  with
• Three balloon-borne DOAS measurements
  (photochemically corrected)
• Three balloon-borne TRIPLE measurements
  (uncorrected)
• Selected ground-based zenith-sky twilight
  measurements (photochemically corrected)
• For each balloon flight air mass trajectory
  calculations were done identifying the forward
  (being in the future w.r.t. the balloon flight)
  and the backward (being in the past w.r.t. the
  balloon flight) match with SCIAMACHY limb
  observations

9
Selected balloon flights

• List of DOAS flights
• March 23rd, 2003 Kiruna (67.9oN, 21.1oE) 1519
  - 1609
• October 9th, 2003 Air sur lAdour (43.7oN,
  0.3oW) 1539 - 1709
• March 24th, 2004 Kiruna (67.9oN, 21.1oE) 1355
  - 1735

• List of TRIPLE flights
• September 24th, 2002 Air sur l'Adour (43.7oN,
  0.3oW) balloon ascent 820 - 918 UTC
• March 6th, 2003 Kiruna (67.9oN, 21.1oE) balloon
  ascent 630 - 756 UTC balloon descent 833 -
  944 UTC
• June 9th, 2003 Kiruna (67.9oN, 21.1oE) balloon
  descent 756 - 909 UTC

• Ground-based zenith-sky twilight measurements
• Selected days, Harestua, Norway (60.2oN, 10.8oE)

10
Comparisons to balloon-born DOAS (1)
Balloon flight March 23rd, 2003 Kiruna
(67.9oN, 21.1oE) 1519 - 1609 Ground based
same day sunrise at Harestua, Norway (60.2oN,
10.8oE)
Backward match March 23rd, 2003 1107 UT
Orbit 5545 State 7 75oN,16oE
Forward match March 24th, 2003 901 UT Orbit
5558 State 10 56oN, 26oE
11
Comparisons to balloon-born DOAS (2)
Balloon flight October 9th, 2003 Air sur
lAdour (43.7oN, 0.3oW) 1539 - 1709
Backward match October 9th, 2003 951 UT
Orbit 8407 State 9 41oN,8oE
Forward match October 10th, 2003 920 UT
Orbit 8421 State 9 41oN, 16oE
12
Comparisons to balloon-born DOAS (3)
Balloon flight March 24th, 2004 Kiruna
(67.9oN, 21.1oE) 1355 - 1735 Ground based
March 24th, 2004 sunset at Harestua, Norway
(60.2oN, 10.8oE)
Backward match March 24th, 2004 1036 UT
Orbit 10798 State 9 66oN,9oE
Forward match March 25th, 2004 825 UT Orbit
10811 State 9 62oN, 38oE
13
Cross sections comparison Fleischmann vs.
Wilmouth
Convolved cross sections (FWHM 0.2 nm)
Absolute difference after scaling, shift and
squeeze
Scaling factors (Wimouth/Fleischmann) 1.03 _at_
298 K, 1.11 _at_ 228/223 K Relative shift 0.004
0.006 nm _at_298 K 0.009 nm 0.02 nm _at_ 228/223 K
14
Effect of BrO cross section Fleischmann vs.
Wilmouth
Balloon flight March 24th, 2004 Kiruna
(67.9oN, 21.1oE) 1355 - 1735
Backward match March 24th, 2004 1036 UT
Orbit 10798 State 9 66oN,9oE
Forward match March 25th, 2004 825 UT Orbit
10811 State 9 62oN, 38oE
15
Comparisons to TRIPLE (1)
Balloon flight September 24th, 2002 Air sur
l'Adour (43.7oN, 0.3oW) balloon ascent 820 -
918 UTC SZA 65o - 56o
Forward match SZA 50o September 24th, 2002
1032 UT Orbit 2968 State 10 47oN, 2oE
Backward match SZA 44.5o September 23rd,
2002 1106 UT Orbit 2954 State 11 40oN,10oW
16
Comparisons to TRIPLE (2)
Balloon flight March 6th, 2003 Kiruna (67.9oN,
21.1oE) balloon ascent 630 - 756 UTC SZA
85o - 78o balloon descent 833 - 944 UTC SZA
76o 73o
Forward match SZA 72.5o March 6th, 2003
1004 UT Orbit 5301 State 8 66oN, 17oE
Backward match SZA 82.6o March 5th, 2003
1213 UT Orbit 5288 State 6 76.5oN,4oE
17
Comparisons to TRIPLE (3)
Balloon flight June 9th, 2003 Kiruna (67.9oN,
21.1oE) balloon descent 756 - 909 UTC SZA
49o - 45o
Forward match SZA 45.6o June 9th, 2003 1018
UT Orbit 6661 State 11 68oN, 16oE
Backward match SZA 81.5o June 8th, 2003
1901 UT Orbit 6652 State 5 68oN,21oE
18
Comparison to ground-based measurements
Zenith-sky twilight measurements at Harestua,
Norway (60.2oN, 10.8oE)
19
Dependence on a priori information (1)
University of Bremen retrievals
Balloon flight October 9th, 2003 1539 -
1709 Air sur lAdour 43.7oN, 0.3oW SCIAMACHY
limb October 10th, 2003 920 UT Orbit 8421,
State 9 41oN, 15oE
20
Dependence on a priori information (2)
SAO retrievals
Balloon flight October 9th, 2003 1539 -
1709 Air sur lAdour 43.7oN, 0.3oW SCIAMACHY
limb October 10th, 2003 920 UT Orbit 8421,
State 9 41oN, 15oE
21
Conclusions (1)

• All retrievals agree within error bars
• Error bars of SAO retrievals are very large in
  the lower layers ? 60 at 17 km increasing
  downwards, often about 100 below 16 km
• Below 20 km SAO retrieval tend to produce higher
  values as compared to Bremen and Heidelberg
  results
• Good agreement between Uni Bremen and SAO
  retrievals in the upper layers above 22 25 km
• Retrievals of Uni Heidelberg result in slightly
  higher values compared to Uni Bremen and SAO
  retrievals in the upper layers above 20 - 22 km,
  which is in a better agreement with ground-based
  measurements but in worse agreement with
  balloon-born DOAS results
• Mostly good agreement between Bremen and
  Heidelberg in the lower layers below 20 - 22 km.

22
Conclusions (2)

• Overall good agreement with balloon-born DOAS
• Variable degree of agreement with TRIPLE
• Reasonable agreement with exception of the lowest
  point for the flight on September 24th, 2002
• Very good agreement for the forward match on June
  9th, 2003, a photochemical correction is required
  for the backward match
• Disagreement for the flight on March 6th, 2003,
  further investigations necessary
• No clear conclusion for ground-based measurements
• Dependence on a priori information
• Uni Bremen retrievals are independent of a priori
  information down to 17 km and show a weak
  dependence on the form of a priori profile below.
  Although retrieval results obtained using
  different a priori profiles are still agree
  within error bars.
• SAO retrievals are found to be nearly independent
  of the initial profile

23
Outlook

• Additional comparisons are needed
• Additional DOAS flight June 17th, 2005
• Photochemical correction of the backward match
  for TRIPLE flight on March 6th, 2003
• Balloon born SAOZ measurements
• Ground-based measurements

• Model simulations and retrievals
• SCIATRAN (Uni Bremen)
• VECTOR (SAO)
• TRACY (Uni Heidelberg)

Further investigation of the influence of the
spectral range, reference tangent height, and
retrieval initialisation parameters