SCIAMACHY Limb Measurements in the UV/Vis Spectral Region: First Results

Advances in Space Research 34 (2004) 775–779 www.elsevier.com/locate/asr

SCIAMACHY limb measurements in the UV/Vis spectral region: ﬁrst results

K.-U. Eichmann a,*, J.W. Kaiser b, C. von Savigny a, A. Rozanov a, V.V. Rozanov a, H. Bovensmann a, M. von Konig€ a, J.P. Burrows a

a Institute of Environmental Physics, University of Bremen, FB 1, D-28334 Bremen, Germany b Remote Sensing Laboratories, University of Zurich, Switzerland

Received 1 December 2002; received in revised form 15 May 2003; accepted 23 May 2003

Abstract

Stratospheric density profile measurements of ozone and nitrogen dioxide derived from limb scattered radiance spectra measurements of the SCanning Imaging Absorption spectroMeter for Atmospheric CHartographY (SCIAMACHY) are presented in this paper. SCIAMACHY is part of the ENVISAT satellite instrumentation, which has been successfully launched in March 2002. SCIAMACHY determines both the total amount and vertical density distributions of a large number of atmospheric constituents by measuring the Earthshine radiance, simultaneously from the ultraviolet (UV) to the near infrared (NIR), in the three viewing ge- ometries nadir, limb, and occultation. The results of the trace gas retrievals of ozone and nitrogen dioxide are compared with nearby POAM III measurements as a first step of verification. Ó 2004 COSPAR. Published by Elsevier Ltd. All rights reserved.

Keywords: SCIAMACHY; Stratospheric density proﬁle measures; Ozone; Stratospheric nitrogen diozide

1. Introduction limited horizontal coverage of occultation experiments like HALogen Occultation Experiment (HALOE) The SCanning Imaging Absorption SpectroMeter for (Russell et al., 1993), Polar Ozone and Aerosol Mea- Atmospheric CHartographY (SCIAMACHY) instru- surement (POAM) (Lucke et al., 1999), and Strato- ment is a space-based spectrometer (Burrows and spheric Aerosols and Gas Experiment (SAGE) Chance, 1991) launched on board ENVIronmental (McCormick et al., 1989). The limb scatter technique SATellite (ENVISAT) in March 2002 (see e.g. Bovens- has previously been used by the Limb Ozone Retrieval mann et al., 1999). SCIAMACHY employs three dis- Experiment/Shuttle Ozone Limb Sounding Experiment tinct viewing modes during each orbit cycle to measure (LORE/SOLSE) instruments flown in 1997 on NASA’s the sun light scattered and absorbed by the Earth’s at- space shuttle mission STS-87 (Flittner et al., 2000). It mosphere: nadir, limb, and occultation. has been also applied to the measurements of the Op- The limb viewing mode of SCIAMACHY is used to tical Spectrograph and Infrared Imager System (OSI- overcome the limited vertical resolution of retrieved RIS) (Llewellyn et al., 1997) aboard the Odin satellite trace gas profiles from nadir viewing spectrometers like (Murtagh et al., 2002) launched in February 2001. the Global Ozone Monitoring Experiment (GOME) This paper, presents first results of both ozone and (Burrows et al., 1999; Hoogen et al., 1999) and the nitrogen dioxide density profiles retrieved from SCIAMACHY limb scatter measurements. To verify the limb retrieval algorithm, first comparisons with trace gas * Corresponding author. Tel.: +49-427-218-4352; fax: +49-421-218- 4555. profiles of the POAM III occultation instrument on E-mail address: [email protected] board the French SPOT-4 satellite (Lucke et al., 1999) (K.-U. Eichmann). were made.

2. The SCIAMACHY instrument which takes both the sphericity of the atmosphere and two orders of scattering into account. Radiance and SCIAMACHY is a passive remote-sensing spec- weighting functions are calculated from analytical for- trometer detecting light in the wavelength range from mulae (Kaiser, 2001). the ultraviolet (240 nm) to the near infrared (2380 nm) A non-linear optimal estimation approach was used with moderate resolution (0.24–1.48 nm). This spectral to derive the ozone profiles. One method (Savigny) is range is divided into eight channels, each science chan- based on the comparison of normalized and paired limb nel comprises 1024 pixel. Thus approximately 8000 radiances for each tangent height both in the center and spectral points are recorded simultaneously. For most of in the wings of the Chappuis–Wulf ozone absorption the orbit, SCIAMACHY makes alternating nadir and bands (525, 600, 675 nm, averaged over 2 nm) (Flittner limb measurements. et al., 2000; McPeters et al., 2000; von Savigny, 2002). In limb viewing mode, SCIAMACHY is looking The other retrieval method (Rozanov) is based on the forward into the flight direction, scanning the atmo- simultaneous fitting of wavelengths from 520 to 670 nm. sphere horizontally and vertically. The instantaneous Employing the Chappuis bands for O3 profile retrievals field-of-view (IFOV) for the limb mode is 0.045° verti- the altitude range between about 15 and 40 can be cally and 1.8° horizontally, corresponding to a geomet- sensed. Below 15 km the line of sight optical depth be- rical field-of-view at the tangent point of 2.6 and 110 comes so large, that retrievals from SCIAMACHY km, respectively. One limb cycle has a total duration of measurements are not reliable down to the troposphere. 60 s. Each horizontal (azimuth) scan has a duration of Above 40 km the absorption signature is too weak for 1.5 s and is divided into up to four individual scans retrievals at this wavelength range. Using an optimal depending on the intensity of the radiance as a function estimation approach the retrieval results are mostly in- of the wavelength. Thus the spatial resolution in across fluenced by the a priori profiles at these heights. Using track (azimuth) direction is typically 240 km, deter- the UV absorption features of ozone around 260 nm, a mined by the integration time. A total of 35 vertical scan retrieval of ozone up to about 75 km will be possible. steps are made. Starting at about )3 km tangent height, Temperature dependent SCIAMACHY Flight Model the tangent height step size is typically 3.3 km. 34 hor- ozone cross-sections and a latitude and time dependent izontal scan steps up to 100 km altitude are performed MPI climatology atmosphere were used as input pa- and one dark signal measurement at about 150 km al- rameter for the forward models, the ground albedo was titude. The horizontal scans are covering 960 km in taken to be A ¼ 0.3. Systematic errors of less than 4% across-track direction. The light path in along track can be expected due to an incorrect albedo (von Savigny direction is approximately 400 km. More details on the et al., 2003). Stratospheric background aerosol profiles characteristics of the SCIAMACHY instrument and the and optical properties are taken from the LOWTRAN-7 data products can be found in e.g. Bovensmann et al. aerosol model (Kneizys et al., 1988). About 2% percent (1999). difference in the retrieved ozone density in the stratosphere can be expected between using the single scattering approximation of SCIARAYS and the full 3. Data and methods multiple scattering forward model SCIATRAN-CDIPI. Below the ozone peak the relative errors will be bigger. For this paper, the raw level zero SCIAMACHY data The NO2 retrieval is performed in the spectral were used, which is the signal (counts) integrated over range 420–455 nm using a ratio of limb measurements the whole azimuth scan of 960 km to get a better signal- at different tangent heights. A measurement at a tan- to-noise ratio. The data are divided by the integration gent height of 40 km is used as a reference spectrum. times (s) and a preliminary dark current correction is The model spectra are calculated using the CDI radi- performed by subtracting the spectrum at 150 km tan- ative transfer model and the optimal estimation gent height. To eliminate the effects of the solar Fra- method is used for the retrieval. Only measurements at unhofer structure and multiplicative instrumental tangent heights between 10 and 40 km are taken into artifacts, the signal is divided by the spectrum at about account. 50 km tangent height. The polarization sensitivity of the instrument is not taken into account. The tangent height information is taken from the operational data product. 4. Results Two forward models were used to simulate the limb scattered radiances. The spherical version of SCIA- Limb radiance profiles were recorded by SCIAM- TRAN (Rozanov et al., 2001) is based on a new Com- ACHY on April 25, 2002 over North America to derive bining Differential-Integral approach using the Picard the vertical ozone distribution. In Fig. 1 the flight path Iterative approximation (CDIPI) (Rozanov et al., 1998, of ENVISAT and the points of limb measurements are 2001). SCIARAYS is a fast radiative transfer model plotted over a total ozone map from the ERS-2 GOME K.-U. Eichmann et al. / Advances in Space Research 34 (2004) 775–779 777

ozone number densities (cm3) as a function of altitude (km) between 10 and 40 km are also compared to the measurements of the POAM III occultation sensor. POAM III has a slightly better vertical resolution of 1–2 km in the stratosphere (Lucke et al., 1999) and is vali- dated against ozone sondes and other satellite measurements. All three profiles agree fairly well (within 10%) taken the different vertical resolution and the dif- ferences in viewing geometry, geolocation and measurement time into account. POAM has a three times higher vertical resolution leading to more features in the profile. It also has completely different viewing angles, so even if the tangent points of both instruments are at the same place and time, they will see different air masses. The result of this effect is hard to quantify from one measurement alone and has to be studied in a comprehensive validation campaign which is beyond the scope of this paper. The integrated subcolumns between 10 and 40 km are 378 and 384 DU for SCIAM- Fig. 1. Spatial distribution of GOME total ozone amounts (DU) measured on April 25, 2002 in the Northern hemisphere. Overplotted ACHY and POAM, respectively. The total ozone as grey boxes are the SCIAMACHY geolocations of the tangent point column for the geolocation of the SCIAMACHY mea- field of view at ground of the limb measurements for the orbit 00798. surement from GOME and TOMS are 459 and 456 DU, The circle and the square depicts the locations of the profiles shown in respectively. Fig. 2 from SCIAMACHY and POAM III, respectively. A latitude–altitude cross-section of the ozone density (1012 cm3) from 70°Nto60°S is shown in the upper instrument (Burrows et al., 1999), which is flying in the part of Fig. 3. High values of ozone at high altitudes can same orbit with a time difference of half an hour. be found in the tropics. Towards the high latitudes the Ozone profiles derived from the two methods devel- ozone maxima decrease in altitude. In the Northern oped at the University of Bremen using two different hemisphere the maximum can be found below 20 km. forward models are shown in Fig. 2. The SCIAMACHY This corresponds to the increase in total subcolumn

Fig. 2. Ozone number density (cm3) proﬁle for Orbit No. 00798 on April 25, 2002. The diamonds and the quadrates depicts the two retrieval methods for SCIAMACHY, while the circles depicts the POAM III measurements. 778 K.-U. Eichmann et al. / Advances in Space Research 34 (2004) 775–779

Fig. 3. Upper panel: stratospheric ozone concentrations (1012 cm3) as a function of latitude (deg) from North to South and altitude between 10 and 40 km along an ENVISAT orbit (No. 00798, April 25, 2002). Lower panel: ozone stratospheric subcolumn (DU) from 10 to 50 km for this orbit.

ozone from 10 to 50 km shown in the lower part of sunset, and NO2 has a rather strong diurnal variability, Fig. 3, where we find more than 400 DU north of 55°N. these measurements were reduced to the SCIAMACHY This is what can be expected for this region and part of solar zenith angle using the SLIMCAT 1D photo- the year (see Fig. 1). chemical model (Chipperfield, 1999). The reaction Fig. 4 shows a comparison of NO2 number density photolysis rates are from the JPL 2000 database (Sander profiles retrieved from SCIAMACHY limb spectra et al., 2000). This was done by scaling the model NOx so (black line) measured on August 8, 2002 at 61°N, 160°W that values at sunset correspond to the POAM mea- (orbit 2302, start time 21:50) to a nearby measurement surement. However, this scaling procedure introduces made by the POAM III occultation instrument. As an uncertainty due to an effective solar zenith angle POAM measurements were performed during local which is assigned to the POAM measurement. Different

Fig. 4. Number density profiles of NO2 from SCIAMACHY measurements on August 8, 2002 at 61°N, 160°W (orbit 2302, start time 21:50). K.-U. Eichmann et al. / Advances in Space Research 34 (2004) 775–779 779 grey lines in the Figure show the dependence of the Burrows, J.P., Weber, M., Buchwitz, M., et al. The Global Ozone POAM-measurement converted to the solar zenith angle Monitoring Experiment (GOME): mission concept and first corresponding to SCIAMACHY measurement on the scientific results. J. Atmos. Sci. 56, 151–175, 1999. Chipperfield, M. Multiannual simulations with a three-dimensional effective solar zenith angle assigned to the POAM chemical transport model. J. Geophys. Res. 104 (D1), 1781–1805, measurement. 1999.

Flittner, D.E., Bhartia, P.K., Herman, B.M. O3 profiles retrieved from limb scatter measurements: theory. Geophys. Res. Lett. 27, 2601– 2604, 2000. 5. Conclusions Hoogen, R., Rozanov, V.V., Burrows, J.P. Ozone profiles from GOME satellite data: algorithm description and first validation. J. First results of the SCIAMACHY limb retrieval in Geophys. Res. 104, 8263–8280, 1999. Kaiser, J. Atmospheric parameter retrieval from UV–Vis–NIR limb the UV/Vis region for ozone and nitrogen dioxide are scattering measurements. Ph.D. Thesis, University of Bremen, presented. Vertical profiles of ozone density between 10 2001. and 50 km and of nitrogen dioxide between 10 and 40 Kneizys, F., Sbettle, E., Abreu, J.C.L.W., et al. Users Guide to km, both with a vertical resolution of about 3 km, have LOWTRAN7, Technical Report AFGL-TR-88-0177, Air Force been derived. First verification of the ozone vertical Geophysics Laboratory, Hanscom AFB, MA 01736, 1988. Llewellyn, E.J., Degenstein, D.A., McDade, I.C., et al. OSIRIS – an profile with POAM III data shows promising results. application of tomography for absorbed emissions in remote Using a 1D photochemical model to modulate the sensing. Appl. Photon. Technol. 2, 627–632, 1997. POAM nitrogen dioxide profile for the time SCIAM- Lucke, R.L., Korwan, D., Bevilacqua, R.M., et al. The polar ozone ACHY made the measurements, good agreement can be and aerosol measurement (POAM) III instrument and early found. Despite problems with SCIAMACHY level 1 validation results. J. Geophys. Res. 104, 18785–18799, 1999. McCormick, M.P., Zawodny, J.M., Veiga, R.E., et al. An overview of calibration the usage of SCIAMACHY level zero data SAGE I and II ozone measurements. Planet. Space Sci. 37, 1567– leads to good quality in profiles when using an internal 1586, 1989. calibration procedure. We were able to retrieve ozone McPeters, R., Janz, S., Hilsenrath, E., et al. The retrieval of O3 profiles profiles for whole orbits. from limb scatter measurements: results from the shuttle ozone limb sounding experiment. Geophys. Res. Lett. 27, 2597–2600, 2000. Murtagh, D.P., Frisk, U., Merino, F., et al. An overview of the Odin Acknowledgements atmospheric mission. Can. J. Phys. 80, 309–318, 2002. Rozanov, A., Rozanov, V., Burrows, J.P. A numerical radiative The authors thank the POAM, GOME, and the transfer model for a spherical planetary atmosphere: combined TOMS teams for providing the input data. We also want differential-integral approach involving the picard iterative approximation. J. Quant. Spectrosc. Radiat. Transfer 69 (4), 491–512, to thank the whole SCIAMACHY team for their sup- 2001. port. SCIAMACHY is a national contribution to the Rozanov, V., Kurosu, T., Burrows, J. Retrieval of atmospheric ESA ENVISAT project, funded by Germany, The constituents in the UV–Visible: a new quasi-analytical approach Netherlands, and Belgium. SCIAMACHY Level zero for the calculation of weighting functions. J. Quant. Spectrosc. data have been provided by ESA. This work has been Radiat. Transfer 60, 277–299, 1998. Russell, J.M., Gordley, L.L., Park, J.H., et al. The halogen occultation funded by the BMBF via GSF/PT-UKF, by DLR- experiment. J. Geophys. Res. 98, 10777–10797, 1993. Bonn, and by the University of Bremen. Sander, S., Friedl, R.R., DeMore, W.B., et al. Chemical kinetics and photochemical data for use in stratospheric modeling, Supplement to Evaluation 12: update of key reactions, JPL publication 00-3, Jet Propulsion Laboratory, California Institute of Technology, Pasa- References dena, CA, 2000. von Savigny, C. Retrieval of stratospheric ozone density profiles from Bovensmann, H., Burrows, J.P., Buchwitz, M., et al. SCIAMACHY – OSIRIS scattered sunlight observations. Ph.D. Thesis, York Mission objectives and measurement modes. J. Atmos. Sci. 56, University, Toronto, Canada, 2002. 125–150, 1999. von Savigny, C., Haley, C., Sioris, C., et al. Stratospheric ozone Burrows, J.P., Chance, K.V. Scanning imaging absorption spectrom- profiles retrieved from limb scattered sunlight radiance spectra eter for atmospheric chartography. Proc. SPIE 1490, 146–155, measured by the OSIRIS instrument on the Odin satellite. 1991. Geophys. Res. Lett. 30, 1755–1758, 2003.