ERSENVISAT Symposium Salzburg 6 10 Sep 2004 Recalibrated

ERS-ENVISAT Symposium, Salzburg, 6 -10 Sep 2004 Recalibrated Ozone Profiles from GOME-UV/Vis Nadir spectra Silvia Tellmann, Mark Weber, Vladimir Rozanov, and John Burrows Institute of Environmental Physics/ Institute of Remote Sensing University of Bremen

Overview • O 3 nadir profiling in Bremen • Calibration Problems in FURM V 5 • Improved calibration correction • low latitude approach • global approach • Improved tropospheric retrieval • Conclusion

GOME O 3 profile retrieval in Bremen l Full Retrieval Method (FURM): Optimal estimation and Kozlov information matrix approach & validation with ozone sonde data R. Hoogen, V. V. Rozanov, J. P. Burrows, Ozone Profiles from GOME Satellite Data: Algorithm Description and First Validation, J. Geophys. Res. , 104, 8263 -8280, 1999. R. Hoogen, V. V. Rozanov, K. Bramstedt, K. -U. Eichmann, M. Weber, and J. P. Burrows, Ozone profiles from GOME satellite data-I: Comparison with ozonesonde measurements, Physics and Chemistry of the Earth 24, 447 -452, 1999. Extensive validation of Version 5 with lidar and satellite data K. Bramstedt, K. -U. Eichmann, M. Weber, V. Rozanov, and J. P. Burrows, GOME ozone profiles: A global validation with HALOE measurements, Adv. Space Res. 29, 1637 -1642, 2002. K. Bramstedt, J. Gleason, D. Loyola, W. Thomas, A. Bracher, M. Weber, and J. P. Burrows, Comparison of total ozone from the satellite instruments GOME and TOMS with measurements from the Dobson network 1996 -2000, Atmospheric Chemistry and Physics 3, 1409 -1419, 2003. G. Hansen, K. Bramstedt, V. Rozanov, M. Weber, and J. P. Burrows, Validation of GOME ozone profiles by means of the ALOMAR ozone lidar, Annales Geophysicae 21, 1879 -1886, 2003. A. Bracher, M. Weber, K. Bramstedt, S. Tellmann, J. P. Burrows, Long-term global measurements of ozone profiles by GOME validated with SAGE II considering atmospheric dynamics, J. Geophys. Res. , accepted, 2004.

GOMEO 3 profile retrieval in Bremen (II) Application in scientific studies: Arctic ozone depletion, international field campaigns (OFP, THESEO 2000), ozone miniholes K. -U. Eichmann, K. Bramstedt, M. Weber, R. Hoogen, V. V. Rozanov, and J. P. Burrows, Structure of ozone mini-holes from GOME, European Symposium on Atmospheric Measurements from Space, Proc. ESAMS'99, ESA-WPP-161, 231 -236, 1999. K. -U. Eichmann, K. Bramstedt, M. Weber, V. V. Rozanov, R. Hoogen and J. P. Burrows, O 3 profiles from GOME satellite data - II: Observations in the Arctic spring 1997 and 1998, Physics and Chemistry of the Earth 24, 453 -457, 1999. H. Bremer, M. von König, A. Kleinböhl, H. Küllmann, K. Künzi, K. Bramstedt, J. P. Burrows, K. -U. Eichmann, M. Weber, A. P. H. Goede, Ozone depletion observed by ASUR during the Arctic Winter 1999/2000, J. Geophys. Res. 107, 8277, doi: 10. 1029/2001 JD 000546, 2002. K. -U. Eichmann, M. Weber, K. Bramstedt, and J. P. Burrows, Ozone depletion in the NH winter/spring 1999/2000 as measured by GOME-ERS 2, J. Geophys. Res. 107, 8280, doi: 10. 1029/2001 JD 001148, 2002. U. Klein, I. Wohltmann, K. Lindner, and K. F. Künzi, Ozone depletion and chlorine activation in the Arctic winter 1999/2000 observed in Ny-Ålesund, J. Geophys. Res. , 107 (D 20), 8288, doi: 10. 1029/2001 JD 000543, 2002. Continued development: tropical retrieval (wavelength extension) and cloud/albedo effects, a-priori profile sensitivity S. Tellmann, S. , V. V. Rozanov, M. Weber, and J. P. Burrows, Improvements in the tropical ozone profile retrieval from GOME UV/vis nadir spectra, Adv. Space Res. 34, 739 -743, 2004. L. N. Lamsal, M. Weber, S. Tellmann, and J. P. Burrows, Ozone column classified climatology of ozone and temperature profiles based on ozonesonde and satellite data, J. Geophys. Res. , accepted, 2004.

FURM inversion scheme • calculate ozone on several altitude levels large number of parameters: n 70 altitude [km] O 3 -eigenvectors in altitude range • Develop the profile in a sum of eigenfunctions with proper truncation n 10 O 3 -eigenvectors in wavelength range O 3 -EV [-] • Kozlov-Information-matrix-method combined with optimal estimation: O 3 -EV [-] Wavelength [nm]

Radiometric calibration problems dichroic mirror degradation 1. 0 Ratio [-] ratio of GOME solar spectra direct after launch and 5 years later 0. 8 etalon 0. 6 0. 4 240 600 400 800 Wavelength [nm] • outgassing (mainly optical coatings e. g. dichroic mirror) change optical features • etalon structures: contamination layers on cooled detectors (ice) varies in time spectral modulation • UV degradation of the instrument: scan mirror is exposed to UVradiation • polarization degradation

FURM V 5 empirical calibration Chebychev Polynomials • wavelength range: 290 – 340 nm • broadband calibration correction to allow for UVdegradation Chebychev Polynomials -0. 1 -0. 3 280 320 300 wavelength [nm] 340 Deviations between model and measurement 0. 4 • differential corrections: NOT applied !!!! V 5 restricted to wavelengths residual [-] polynomials [-] 0. 1 Channel 1 0. 2 0. 0 -0. 2 280 290 300 wavelength [nm] 310

Shortcomings in FURM V 5 • lack of information in upper stratosphere (above 35 -40 km) • problems in tropics where ozone maximum is shifted to higher altitudes wavelengths below 290 nm needed to enhance stratospheric information content Chebyshev polynomials inadequate at short wavelengths Altitude [km] 60 Sensitive to measurement 40 20 50 Altitude [km] measurement sensitivity 40 ozone profiles in tropics FURM SAGEII Climatology 30 20 0 0. 2 0. 6 1. 0 Sum of rows AK-Matrix [-] 1. 4 O 3 -concentration

Tropical calibration correction • large differential structures below 290 nm • strong filling-in of Fraunhofer lines Differential sunnormalized radiance y_meas [-] 5 D-4 GOME Model 3 D-4 1 D-4 -1 D-4 280 285 wavelength [nm] 290 New corrections by differential fit: Fit addition error term: a ~1/irradiance 1/irr & polynomial fit before ozone retrieval (prefit in Ch. 1) residual & fit of diff. error y_meas – y_model [-] 275 5 D-4 GOME - Model Diff. error 3 D-4 1 D-4 -1 D-4 275 280 wavelength [nm] 290

High latitude calibration correction Ozone maximum in middle and high latitudes shifted to lower altitudes and longer wavelengths low differential structure of ozone absorption strong correlations between broadband calibration corrections and atmospheric parameter change of broadband calibration correction required below 300 nm statistical investigation of residuals Residual [-] 1. 0 Mean residuals 1997 -2003 in comparison with HALOE 0. 8 0. 6 0. 4 0. 2 0. 0 275 280 285 290 wavelength [nm] 295 300

Results from 1997 Relative deviations between standard version/new version and independent measurement from SAGE II (Stratospheric Aerosol and Gas Experiment II) old results 50 0. 0 -0. 1 -0. 2 Rel. Deviations New 40 40 30 20 new results 50 Rel. Deviations Old 40 30 tropics 20 30 20 -60 -0. 3 Old New altitude [km] 0. 1 Altitude [km] 0. 2 Altitude [km] 0. 3 Comparison of O 3 [ppm] SAGE/FURM 50 -40 -20 0 20 latitude [deg] 40 60 10 -0. 6 -0. 4 -0. 2 0. 0 0. 2 0. 4 0. 6 2(SAGE-GOME)/(SAGE+GOME)

Results from 2003 Relative deviations between SAGE II and FURM 0. 2 0. 1 Altitude [km] 0. 3 New Results 50 40 30 Old New 40 20 35 No of Pixel Altitude [km] -0. 2 50 45 0. 0 -0. 1 Comparison of O 3 [ppm] SAGE/FURM altitude [km] old results 30 new results 50 25 40 20 30 15 -0. 6 20 -0. 2 0. 0 0. 2 0. 4 0. 6 2(SAGE-GOME)/(SAGE+GOME) -0. 3 -60 -0. 4 -40 -20 No of Pixel 0 20 latitude [deg] 40 60

Tropospheric retrieval Ozone Profile Hohenpeissenberg 50 FURM Sonde Climatology Aks for retrieval levels in troposphere 50 30 40 altitude [km] 40 20 10 0 O 3 Conc retrieved tropospheric ozone always quite near to climatological values averaging kernels (AKs) indicate where information in retrieved height level comes from 30 20 10 0 Averaging Kernels [-]

Tropospherical investigations • • • Profile fit: spectral window too large to make accurate tropospheric ozone fit Use just channel 2 to improve tropospheric column fit Two step retrieval: 1. Common ozone profile fit to get information about stratospheric constituents. 2. Use fitted profile and investigate residual in channel 2 (~320 -335 nm) troposphere defined up to ECMWF thermal tropopause Make new diff. fit of tropospherical ozone column, temperature and Ring (inelastic scattering)

Residual & Fit O 3 WF fit [-] Residual [-] Sensitivity in Channel 2 Residual Fit all param. New col. Old col. Temp. Fit Ring WF fit [-] Trop. Col. 0. 04 0. 00 -0. 04 0. 02 0. 00 -0. 02 0. 000 -0. 002 0. 00 -0. 02 320 322 324 326 328 wavelength [nm] 330 332 334

Tropospherical Column Results Relative Deviations between O 3–Sonde and climatology as well as FURM Rel. deviations Sonde-FURM [-] Hohenpeissenberg 1997 troposph. columns from + profile retrieval troposph. columns from differential fit in channel 2 regression line for profile fit regression line for step 2 differential fit in channel 2 • in most cases improvements Rel. deviations Sonde-Climatology [-] • for very small differences to climatological value instabilities in differential fit (noise)

Conclusions Stratospheric information content can be enhanced by combined calibration concept in short wavelength region using • differential „dark current“ approach for tropics/southern anomaly • residual fit for middle and high latitudes. Tropospheric retrieval based upon two step approach • accurate stratospheric ozone fit • accurate fit of Ring & atm. Parameters • realistic climatological ozone profile form Inclusion of new ozone a-priori climatology (Lamsal et al. 2004) Reprocessing of GOME 1995 -2003 in preparation (funding? ) First application to SCIAMACHY (Version 5)

A-priori profile sensitivity Hohenpeissenberg 1997 IUP climatology (Lamsal et al. , JGR, 2004) F&K (Fortuin and Kelder, JGR, 1998) ðImprovement in the lowermost stratosphere with IUP climatology ðsee also talk by Lamsal et al. (Abstract 353) for details on IUP O 3 climatology 23 August 2002

SCIAMACHY nadir application 23 August 2002 limb retrieval V 1. 6 (Savigny et al. , JAS, 2004) FURM nadir retrieval Version 5 ðthere are still outstanding issues regarding calibration ðVisit poster by Bramstedt et al. (3 P 08 -6) for more details