1. Anna Serdyuchenko, Victor Gorshelev, Wissam Chehade, Mark Weber, John P. Burrows University of Bremen, Institute for Environmental Physics

2. 2 Inspired by satellites, performed on Earth…. Atmospheric species detection:  Satellite spectrometers  Available databases for the reference data Challenges of cross-section measurements in laboratory:  Modern demands on the O3 absorption cross-section quality  Experimental equipment  Absolute calibration. Results:  Ozone spectra in UV /VIS/IR;  Comparison with existing datasets. MolSpec Lab IUP University of Bremen21 June 2011

3. 3 MolSpec Lab IUP University of Bremen

4. 4 Satellite/airborne/ground spectrometers monitor: Stratosphere: ozone chemistry, volcanic events, solar proton events Troposphere: biomass burning, pollution, arctic haze, forest fires, dust storms, industrial plumes Clouds, aerosols, UV index Detected: the solar radiation transmitted, backscattered and reflected from the Earth atmosphere and surface, also direct sunlight Absorption spectra cover: O3 O2 NO2 N2O BrO OClO SO2 H2CO2 CO CO2 CH4 H2O GOME, GOME2 (Global Ozone Monitoring Experiment): scanning 4 channels grating spectrometer with nadir-view SCIAMACHY (Scanning Imaging Absorption Spectrometer for Atmospheric Chartography): scanning 8 channels grating spectrometer with nadir/limb-view MolSpec Lab IUP University of Bremen21 June 2011

5. Long-term time-series of O 3 is important for air quality study, Montreal Protocol monitoring of ozone depleting substances, ozone-climate interaction etc Long-term global data sets covering several decades are only available by combining datasets from multiple sensors (satellite spectrometers and ground based instruments: Brewer, Dobson spectrophotometers ). At least two decades of observations with global coverage in several daysSpectrometerSatelliteLaunch GOMEERS-2April 1995 SCIAMACHYENVISATMarch 2002 GOME-2MetOp -AOctober 2006 MetOp - B~ 2012 MetOp - C~ 2016 5MolSpec Lab IUP University of Bremen21 June 2011

6. 6 Absorption cross-sections can introduce an important error source For GOME, SCIAMACHY, and GOME2 flight models were used to measure absorption cross-section prior to launch. Advantage: exact match of spectral resolution between satellite radiance and cross-sections; Knowledge for instrumental slit function is not neededSpectrometerSatelliteLaunch Pre-launch O3 cross-sections GOMEERS-2April 1995GOME FM (Burrows et al. 1999) SCIAMACHYENVISATMarch 2002SCIAMACHY FM (Bogumil et al. 2003) GOME-2MetOp -AOctober 2006GOME2 FM3 (Gür et al., 2005) MetOp - B~ June 2012GOME2 FM21 (Gür et al., 2005) MetOp - C~ 2016- MolSpec Lab IUP University of Bremen21 June 2011

7. Spurious instrumental trends in multiple instruments time series SCIAMACHY total O 3 retrieval (using SCIAMACHY FM reference spectra) were 5% higher than GOME (with GOME FM reference spectra) in the range 325-335 nm GOME2 total O 3 retrieval (using GOME2 FM) is 9% higher than calculated with resolution adjusted GOME FM Harmonisation of O 3 FM cross-sections from GOME and SCIAMACHY for a consistent retrieval State of the Climate in 2008, Bull. Amer. Meteor. Soc., submitted, 2009 721 June 2011MolSpec Lab IUP University of Bremen

8. The committee "ACSO" ("Absorption Cross Sections of Ozone") was established in spring 2009 by the World Meteorological Organization (WMO) and the International Association of Meteorology and Atmospheric Sciences (IAMAS) to  Review the presently available ozone absorption cross sections  Determine the impact of changing the reference ozone absorption cross sections for all of the commonly used atmospheric ozone monitoring instruments.  Recommend whether a change needs to be made.  The recommendations are to be discussed with the community of the involved experts. The work will be finished within two years after the first meeting in May 2009. 821 June 2011MolSpec Lab IUP University of Bremen

9. 9 ‘Low’ resolution (0.2 nm): SCIAMACHY ‘High’ resolution (0.02 nm): Brion et al, currently recommended by HITRAN IUP MolSpec Lab homepage: http://www.iup.uni-bremen.de/gruppen/molspec/databases/sciamachydata/index.html IGACO (Integrated Global Atmospheric Chemistry Observations) homepage: http://igaco-o3.fmi.fi/ACSO/cross_sections.html 21 June 2011MolSpec Lab IUP University of Bremen

10. 10 wavelength coverage UV/VIS/NIR; vacuum wavelength accuracy better than 0.001 nm; spectral resolution ~ 0.02 nm; absolute intensities accuracy: 2%; temperature: 190K-300K, step 10K sufficient accuracy to detect a 1% pro decade trend! 21 June 2011MolSpec Lab IUP University of Bremen

11. 1121 June 2011MolSpec Lab IUP University of Bremen

12. Echelle spectrometer coverage Echelle spectrometer coverage FT spectrometer coverage 1221 June 2011MolSpec Lab IUP University of Bremen

13. 13MolSpec Lab IUP University of Bremen21 June 2011

14. 1421 June 2011MolSpec Lab IUP University of Bremen

15. 15MolSpec Lab IUP University of Bremen21 June 2011 Error originAccuracy/precision  absorption cross section 1-2 % ODOptical densityReproducibility of spectra intensitiesBetter than 1 % for most of the spectral regions AScaling factorAbsolute calibration parametersBetter than 2 % for most of the spectral regions I / I 0 transmitted intensity with /without ozone L absorption path length (1 – 20 m) N ozone density p Ozone pressure (1- 50 mbar) k Boltzmann constant T Temperature (200 – 300 K)

16. 16 Instrument: Echelle, FTS Averaged: over 2000 (100) scans Time: ~ 30 minutes Source: Xe and deuterium lamps 21 June 2011MolSpec Lab IUP University of Bremen

17. 21 June 201117MolSpec Lab IUP University of Bremen

18. 18 Measurements: Each spectrum consists of 8 pieces11 temperatures3 – 5 approachesAbout 300 measurements Processing: 16 spectra edges2 parameters: OD and SD values11 temperaturesAbsolute calibration parameters 21 June 2011MolSpec Lab IUP University of Bremen

19.  Ozone quantity retrieval  Polynomial dependence on temperature:  So-called Bass-Paur parameterization (2 nd order polynomial)  For selected wavelengths in Huggins band  Comparison with existing datasets  High resolution datasets ▪ Bass –Paur (BP) ▪ Brion et al (BMD)  Low resolution datasets: ▪ GOME ▪ GOME-2 ▪ SCIAMACHY 1921 June 2011MolSpec Lab IUP University of Bremen

20.  Band-integrated cross-sections as functions of temperature:  Insensitive to resolution  Direct comparison:  resolution matching  relative shift in wavelength  absolute scaling factor  regions: ▪ UV (220 – 320 nm); ▪ DOAS (320 – 340 nm); ▪ Bottom (350 – 450 nm); ▪ Visible (450 - 700 nm); ▪ IR (700 – 1000 nm) 2021 June 2011MolSpec Lab IUP University of Bremen

21. 21 No resolution matching ‘Bass-Paur ’parametrisation 328 nm “hill”, 330 nm “valley” Black– SCIAMACHY Blue – BMD Red – New Data 21 June 2011MolSpec Lab IUP University of Bremen

22. 22 325 – 340 nm contains DOAS window used for retrievals. All datasets cover this region T, KMean ValueNew data, x 10 -20 2035.47 ± 3.3%5.46 2235.72 ± 2.6%5.74 2436.23 ±1.9%6.30 2737.26 ± 1.3%7.35 293 8.23 ± 1% 8.30 21 June 2011MolSpec Lab IUP University of Bremen

23. BP – Bass Paur, BMD – Brion et al  quadratic polynomial inter/extrapolation 193-293K with step of 10K  inter/extrapolation on the common grid with step of 0.01 nm  no resolution matching needed  difference mean over 20/25 nm Green – comparison BP versus new experiment (325-340 nm); Red – comparison BMD versus new experiment (325-345 nm); Black – comparison BP versus BMD; Comparison in Huggins band region: new data versus high resolution datasets, relative difference and shifts 23 BP- BMDBMD- New Data BP –New Data Shift, nm0.02 0.005 nm0.02 Mean relative difference 3%1-2%2-3% 21 June 2011MolSpec Lab IUP University of Bremen

24. 24  interpolation on the common grid (step 0.01 nm)  resolution matching: Gaussian, FWHM = 0.2 nm  difference mean over 20 nm  Shift: 0.004 nm, 0.016@293K  mean relative difference: 2% Comparison in Huggins band region: new data versus SCIAMACHY, relative difference and shifts 21 June 2011MolSpec Lab IUP University of Bremen

25. Direct comparison in VIS and IR regions: new data versus SCIAMACHY, GOME and GOME-2, 203 K 2521 June 2011MolSpec Lab IUP University of Bremen

26. Direct comparison in VIS and IR regions: new data versus SCIAMACHY at different temperatures 2621 June 2011MolSpec Lab IUP University of Bremen

27. Total ozone retrieval of GOME-2 and SCIAMACHY. The data have to be convoluted to GOME-2 and SCIAMACHY slit function. The optimum wavelength shifts are found at - 0.007nm and +0.007nm for GOME-2 and SCIAMACHY respectively. 2721 June 2011MolSpec Lab IUP University of Bremen The total ozone columns retrieved from GOME-3 and SCIAMACHY satellites using the new experimental cross-section data are in good agreement with the amounts retrieved by the current data and are +1% and +2% respectively

28. 28 Summer 2009: Start of FTS measurements VIS/IR Autumn 2009: New spectrometer (Echelle) Spring 2010: Relative measurements UV/VIS Summer 2010: Absolute calibration methods Winter 2010/11: absolute measurements Spring 2011: Tests on retrievals Summer 2011: Publication 21 June 2011MolSpec Lab IUP University of Bremen

29. Summary  Full measurements set  11 temperatures,  UV-VIS-IR,  Absolute scaling using O 3 trap Future Steps  Additional IR region – around 1.6 mkm and 10 mkm  Test on retrievals – spring/summer 2011  Publication – summer 2011 2921 June 2011MolSpec Lab IUP University of Bremen

30. 30 Thank you for attention ! 21 June 2011

31. MolSpec Lab IUP University of Bremen3121 June 2011

32. 32 Data setTemperatures, KResolution, nmWavelength, nm GOME FM Burrows et al., 1999 ‏ 202 221 241 273 2930.17 @ 330 nm230 - 800 SCIAMACHY FM Bogumil et al., 2003 203 223 243 273 2930.20 @ 330 nm230 - 1070 GOME2 FM Spietz et al., 2005 203 223 243 273 2930.29 @ 330nm240 – 790 Paur and Bass, 1985203 218 228 243 273 298 <0.025 nm(?) ‏ 245 - 345 Malicet et al. 1995 Brion et al., 1993 Daumont et al., 1992 218 228 243 273 2950.01-0.02 nm 195-345 UV-FTS Voigt et al., 2001 203 223 246 280 293 0.03 @ 230 nm230 - 850 Introduction and motivation Experimental set-up Analysis of sources of uncertainty Results and Outlook http:/ http://www.iup.uni-bremen.de/gruppen/molspec/databases/index.html MolSpec Lab IUP University of Bremen21 June 2011

33. MolSpec Lab IUP University of Bremen33 Dataset Resolution, nm (FWHM) Wavelength region, nm Current work FTS0.02 – 0.20450 – 1000 Echellet0.02211 – 830 SCIAMACHY0.32 – 1.45230 – 1070 GOME, GOME-20.2 – 0.4231 – 794 High Resolution Brion et al.0.02195 – 830 External datasets: GOME, GOME2, SCIAMACHY, Brion et al (high resolution) Resolution matching Introduction and motivation Experimental set-up Analysis of sources of uncertainty Results and Outlook 21 June 2011

34. MolSpec Lab IUP University of Bremen34 Upgraded cooling system Max possible cooling: down to 193 K Temperature stabilization at intermediate points with step of 10 K Reliable temperature determination (better than 5% accuracy) : Pt sensors, spectroscopic method Upgraded gas pre-cooler features 10 meter Cu pipe bound to fit cryostat bath guaranteed cooling down to cryostat vessel temperature; ozone-friendly internal coating minimal heat gain between precooler ant test cell Sensors: Precooler:  Goals and strategy  Re-analysis  Experimental set-up  Serial measurements and preliminary results 21 June 2011

35. MolSpec Lab IUP University of Bremen35 Cooling: double jacket (vacuum/ethanol) cell, cryostat with pre-cooling: 10 m Cu tube coil with inert coating Control: O2-A band at 760 nm, experimental spectrum: from FTS at 0.5 cm, model spectrum: using HITRAN line parameters accuracy: 5 K or better Introduction and motivation Experimental set-up Analysis of sources of uncertainty Results and Outlook 21 June 2011