The semianalytical cloud retrieval algorithm for SCIAMACHY II. The application to MERIS and SCIAMACHY data A. A. Kokhanovsky, W. von Hoyningen-Huene, V. V. Rozanov, S. Noël, K. Gerilowski, H. Bovensmann, K. Bramstedt, M. Buchwitz, J. P. Burrows
To cite this version:
A. A. Kokhanovsky, W. von Hoyningen-Huene, V. V. Rozanov, S. Noël, K. Gerilowski, et al.. The semianalytical cloud retrieval algorithm for SCIAMACHY II. The application to MERIS and SCIA- MACHY data. Atmospheric Chemistry and Physics Discussions, European Geosciences Union, 2006, 6 (2), pp.1813-1840. hal-00303889
HAL Id: hal-00303889 https://hal.archives-ouvertes.fr/hal-00303889 Submitted on 13 Mar 2006
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Atmos. Chem. Phys. Discuss., 6, 1813–1840, 2006 Atmospheric www.atmos-chem-phys-discuss.net/6/1813/2006/ Chemistry ACPD © Author(s) 2006. This work is licensed and Physics 6, 1813–1840, 2006 under a Creative Commons License. Discussions
Semianalytical cloud retrieval algorithm
A. Kokhanovsky et al. The semianalytical cloud retrieval Title Page algorithm for SCIAMACHY Abstract Introduction
II. The application to MERIS and Conclusions References SCIAMACHY data Tables Figures
A. A. Kokhanovsky, W. von Hoyningen-Huene, V. V. Rozanov, S. Noel,¨ J I K. Gerilowski, H. Bovensmann, K. Bramstedt, M. Buchwitz, and J. P. Burrows J I Institute of Remote Sensing, University of Bremen, Germany Back Close Received: 15 April 2004 – Accepted: 20 February 2006 – Published: 13 March 2006 Full Screen / Esc Correspondence to: A. Kokhanovsky ([email protected])
Printer-friendly Version
Interactive Discussion
EGU
1813 Abstract ACPD The Semianalytical CloUd Retrieval Algorithm (SACURA) is applied to the SCanning Imaging Absorption spectroMeter for Atmospheric CHartographY (SCIAMACHY) data. 6, 1813–1840, 2006 In particular, for the first time we derive simultaneously cloud optical thickness (COT), 5 liquid water path (LWP), cloud top height (CTH), and cloud droplet radius (CDR) us- Semianalytical cloud ing SCIAMACHY measurements in the visible (442 nm, COT), in the oxygen A-band retrieval algorithm (755–775 nm, CTH) and in the infrared (1550 nm, CDR). Some of the results obtained are compared with those derived from the Medium Resolution Imaging Spectrometer A. Kokhanovsky et al. (MERIS), which has better spatial resolution and observe almost the same scene as 10 SCIAMACHY. Title Page
1 Introduction Abstract Introduction
Conclusions References In addition to being one of the most significant components of the hydrological cycle, clouds scatter and absorb electromagnetic radiation in the atmosphere. Their amount Tables Figures and distribution are of a great significance for radiative transfer and the atmospheric ra-
15 diative balance. Clouds have a direct role in climate and global climate change. There J I are several important issues in the cloud research area (for example, the coupling of gases and clouds via gas-aerosol-cloud interactions and the ability of clouds to precip- J I
itate; see Crutzen and Ramanathan, 1996; Raschke, 1996). For both climate research Back Close and numerical weather prediction, it is necessary to have knowledge about cloud pa- 20 rameters on a global scale. This is only feasible from spaceborne remote sensing Full Screen / Esc instrumentation. A variety of techniques have been developed for the remote sensing of clouds. They Printer-friendly Version are mostly based on the look-up-table approach (LUT). Then spectral reflectances are calculated for multiple illumination and observation conditions and stored in special Interactive Discussion 25 tables. Subsequently, the measured spectral reflectances are compared with pre- calculated values in an iterative manner until the minimum deviation is found. It is EGU 1814 believed that cloud parameters found for this minimum represent characteristics of a cloud field under study. ACPD The Semianalytical CloUd Retrieval Algorithm (SACURA), used in this study, does 6, 1813–1840, 2006 not rely on LUTs at all. Instead we use semianalytical solutions of the radiative transfer 5 equation valid for weakly absorbing clouds having large optical thickness (Kokhanovsky et al., 2003; Kokhanovsky and Rozanov, 2003, 2004; Rozanov and Kokhanovsky, Semianalytical cloud 2004). This algorithm is flexible and is readily applied to the data from different in- retrieval algorithm struments. The manuscript describes the application of the algorithm to the data obtained from A. Kokhanovsky et al. 10 the SCanning Imaging Absorption spectroMeter for Atmospheric CHartographY (SCIA- MACHY) on board of the ENVIronmental SATellite (ENVISAT). Technical characteris- Title Page tics of the SCIAMACHY are reported in Bovensmann et al. (1999, 2004) and references therein. Details of the SACURA are described elsewhere by Kokhanovsky et al. (2003) Abstract Introduction and Rozanov and Kokhanovsky (2004). The validation of SACURA is presented in a Conclusions References 15 separate publication (Kokhanovsky et al., 2005).
The following cloud parameters from SCIAMACHY measurements are retrieved by Tables Figures SACURA: the cloud optical thickness, the cloud droplet radius, the liquid water content, the cloud thermodynamic state and the cloud top height. The determination of such J I a long list of parameters is possible due to the broad spectral range, i.e. simultane- 20 ous measurements from 220 till 2380 nm, coupled with high spectral resolution of this J I advanced optical instrument. Back Close However, the spatial resolution of SCIAMACHY is limited as it requires high signal to noise ratio to retrieve its primary objective, trace gas abundances. The spatial resolu- Full Screen / Esc tion is variable from 30×240 km2 to 30×30 km2. Extended cloud fields cover many 25 millions of square kilometers and for such cloud systems the use of SCIAMACHY Printer-friendly Version measurements for monitoring cloud properties is justified. Similarly the synergistic combination of SCIAMACHY data and the simultaneous high spatial resolution data of Interactive Discussion MERIS for clouds having dimensions smaller than the ground scene of SCIAMACHY is of a great interest. EGU
1815 The main emphasis of Sect. 2 of this study is on the investigation of the cloud op- tical thickness (COT) as retrieved by SCIAMACHY at the wavelength 442 nm. The ACPD retrievals are compared with those obtained from the Medium Resolution Imaging 6, 1813–1840, 2006 Spectrometer (MERIS), whose measurements have significantly superior spatial res- 2 2 5 olution (0.3×0.3 km or 1.1×1.1 km , dependent on the measurement mode). There- after Sect. 3 is devoted to the determination of the cloud thermodynamic state using the Semianalytical cloud measurements of SCIAMACHY at wavelengths 1500 and 1670 nm. The cloud phase retrieval algorithm index (CPI), which is defined to be the ratio of reflectances at these two wavelengths is introduced. We show that this ratio is sensitive to the cloud thermodynamic state. The A. Kokhanovsky et al. 10 cloud altitude determination using measurements in the oxygen A-band are explained and discussed in Sect. 4. The last section of the paper is devoted to the discussion of Title Page SCIAMACHY calibration issues. Abstract Introduction
2 The cloud optical thickness Conclusions References
The geographical location of seven studied SCIAMACHY measurement states from Tables Figures 15 the orbit 02223 (08:57, 3 August 2002) is shown in Fig. 1. Retrievals using MERIS data are performed in the quadrangle. The MERIS browse image of the area at the J I moment of measurements is shown in Fig. 2. The region above the line AB roughly corresponds to that of a MERIS scene in Fig. 1. The letter W in Fig. 2 shows the J I position of Warsaw. The MERIS browse image indicates that many clouds present Back Close 20 in the scene. In particular, note an extended cloud area covering parts of Sweden, Poland, and Belarus. Full Screen / Esc The cloud reflection function as measured by SCIAMACHY versus the same function but obtained by the MERIS at the wavelength 442 nm, where the contribution from the Printer-friendly Version surface is relatively low is shown in Fig. 3. The definition of the reflection function is Interactive Discussion 25 given, e.g., by Kokhanovsky et al. (2003). The MERIS reflectances were averaged over large SCIAMACHY pixels. Both re- flectances are highly correlated but they are not equal. On average, SCIAMACHY re- EGU 1816 flectances are a factor of 0.89 smaller than those obtained by the averaging of MERIS pixels at the wavelength 442 nm. ACPD If MERIS data are considered to be well calibrated, then the factor C=1.12 should be 6, 1813–1840, 2006 applied to SCIAMACHY data at 442 nm for correct retrievals of cloud parameters using 5 absolute radiances. This is a major conclusion of this work. Note that von Hoyningen- Huene et al. (2006) find the same factor but analysing only carefully selected clear Semianalytical cloud SCIAMACHY pixels over water. We present relative differences of both reflectances (in retrieval algorithm percent) in Fig. 4. It follows from this Figure that SCIAMACHY reflectances are in the average 10 percent smaller than that of MERIS at the wavelength 442 nm. However, A. Kokhanovsky et al. 10 values between 0 and 20% also occur. The dispersion of data around its average value of 10 percent is due to the problem with the instruments, atmospheric effects or both. Title Page Similar results were found comparing Advanced Along Track Scanning Spectrometer (AATSR) data with that of SCIAMACHY (B. Kerridge and B. Latter, personal communi- Abstract Introduction cation). So, clearly, SCIAMACHY Level1 data should be further improved with respect Conclusions References 15 to the calibration correction. Some discussion on this topic is given in the last section
of this paper. Tables Figures Let us estimate what could be the result of the influence of the calibration error =12% (see above) on the retrieval of the cloud optical thickness. For this we will use J I a simple relationship between the COT τ and the reflection function R for a cloud above 20 a black surface given by Kokhanovsky et al. (2003): J I
DK (µ)K (µ ) Back Close τ = 0 − Db, (1) R − R ∞ Full Screen / Esc D 4 b . g R where = 3(1−g) , =1 072. Here the value of is the asymmetry parameter, ∞ is the Printer-friendly Version reflection function of a semi-infinite cloud, K (µ) is the escape function, µ and µ0 are cosines of the observation and incidence angles, respectively. Convenient expressions Interactive Discussion 25 for the functions K (µ), R∞, and g are given by Kokhanovsky et al. (2003). Note that we have approximately for the case studied: K (µ)K (µ0)=1, R∞=1 and g=0.85. Then EGU
1817 one obtains the following approximate equation: 9R ACPD τ = . (2) 1 − R 6, 1813–1840, 2006
This means that the ratio A of the SCIAMACHY derived optical thickness τs to the MERIS derived optical thickness τm is given by the following expression: Semianalytical cloud 1 − R retrieval algorithm A = m , (3) 5 C − Rm A. Kokhanovsky et al.
where C=1.12 as specified above and Rm is the MERIS TOA reflection function. We present the dependence A(Rm) in Fig. 5. Title Page It follows from Fig. 3 that Rm<0.8 in most cases. Then A changes in the range 0.6−0.9. So we can take as an average value: A=0.75. Therefore, we should have as Abstract Introduction 10 a consequence of Rm=CRs: τs=0.75 τm. Here Rs is the reflection function measured by SCIAMACHY. To check this we prepared a scatter plot similar to that shown in Fig. 3 Conclusions References
but for the retrieved cloud optical thickness. This is given in Fig. 6. Our estimation, Tables Figures therefore, is confirmed. It means that if one does not account for the calibration coeffi- cient (C=1.12), then the retrieval gives lower values of the cloud optical thickness (by J I 15 approximately 25%) as compared to those obtained from MERIS measurements. This is a major conclusion of this work. J I Similar biases must appear in the retrieved values of the liquid water path w and Back Close the cloud droplet effective radius aef . However, it is difficult to correct them because the calibration coefficient is not spectrally neutral and the values of w and aef can be Full Screen / Esc 20 obtained only using SCIAMACHY measurements at 1500 nm. Remind, that the largest
wavelength of the MERIS is 900 nm. So C can not be found at larger wavelengths Printer-friendly Version using MERIS. Interactive Discussion
EGU
1818 3 The cloud thermodynamic state ACPD The cloud thermodynamic state can not be estimated using MERIS data due to lack of infrared measurements. This can be done using SCIAMACHY, however. Let us 6, 1813–1840, 2006 introduce the so-called cloud phase index (CPI): R(1550 nm) Semianalytical cloud α = . (4) retrieval algorithm 5 R(1670 nm) This index is similar but not identical to that proposed by Acarreta et al. (2004). We A. Kokhanovsky et al. find using radiative transfer calculations that the value of α is in the range 0.7–1.0 for water clouds and it is in the range 0.5–0.7 for ice clouds (see Fig. 7). Calculations shown in Fig. 7 were performed using the modified asymptotic equations given by Title Page
10 Kokhanovsky and Rozanov (2003) in the assumptions that both crystals and droplets Abstract Introduction have the spherical shape. It was assumed that the cloud optical thickness is varied in the range 5–30 for both water and ice clouds. The effective radius of water clouds Conclusions References was changed in the range 5–30 microns. For ice clouds we assumed two ranges of Tables Figures the effective radius change: 5–30 µm (red colour in Fig. 7) and 31−50 µm (black colour 15 in Fig. 7). The analysis shows that the increase of the size of particles increases the separation of both cloud types. Note that ice crystals have highly irregular shape J I
and also the size is often in the range 100−500 µ. This will enhance differences even J I further. Calculations as shown in Fig. 7 correspond to the nadir observation and the solar zenith angle equal to 60 degrees. We found that the change of the illumination Back Close
20 angle does not influence the positions of regions with characteristic values of CPI for Full Screen / Esc water/ice cloud significantly. Therefore, CPI indeed can be used as an indicator of the cloud phase. Printer-friendly Version Obviously, we have for clear sky over a black surface: R(1550 nm)>R(1670 nm), and, therefore, α>1. Some highly reflecting soils could have values of α<1 similar to water Interactive Discussion 25 clouds (see Fig. 8). However, these pixels could be screened using information on the derived cloud top pressure, cloud optical thickness or both. EGU
1819 The probability distribution of CPI for the whole SCIAMACHY orbit analysed is shown in Fig. 9. There are similarities between Figs. 7 and 9. Interestingly, data with α>1 ACPD appear in Fig. 9. These data are due to the clear pixels over water. Note that not all 6, 1813–1840, 2006 pixels with α<1 correspond to clouds. Some of them are due to soil spectral features 5 clearly seen through the cloudless atmosphere, which is almost transparent at the wavelengths 1500 and 1670 nm. Atmospheric scattering and gaseous absorption is Semianalytical cloud low at these channels. We stress that the CPI could be also biased due to the partially retrieval algorithm cloudy pixels. It follows that α=0.9 for the soil spectrum given in Fig. 8. Therefore, we repro- A. Kokhanovsky et al. 10 cessed data given in Fig. 9 to select only cloudy pixels. For this we used the condition: R(442 nm)>0.3. The albedo of soil at this wavelength is below 0.1. Then the histogram Title Page has the form as shown in Fig. 10. We also found that with a rare exclusion, we deal with predominantly water clouds in the scene studied. Abstract Introduction
Conclusions References 4 The cloud top height Tables Figures
15 The cloud top height can be found by analysing SCIAMACHY measurements in the oxygen A-band. One example of the SCIAMACHY spectrum together with the fit using J I the radiative transfer theory (Rozanov and Kokhanovsky, 2004) is shown in Fig. 11. It follows that the experimental spectrum can be fitted using a cloud having the top height J I H=6.3 km, and the cloud geometrical thickness (CGT) equal to 5.9 km in the case stud- Back Close 20 ied. The optical thickness obtained from the fit is equal to 15.5. Note that the obtained value of τ could be biased due to the calibration errors as discussed above. Also the Full Screen / Esc retrieved cloud geometrical thickness corresponds to the case of a single cloud layer. For multi-layered cloud systems, which often exist in the terrestrial atmosphere, the Printer-friendly Version retrieved cloud geometrical thickness will di er from the total geometrical thickness of ff Interactive Discussion 25 all layers. However, the cloud top height obtained with SCIAMACHY is least biased among all other cloud products derived due to the fact that the SACURA uses the rel- ative reflection function for retrievals of H as discussed by Rozanov and Kokhanovsky EGU 1820 (2004). Then the issue of calibration is only of a minor importance. The same is true for CPI. ACPD The cloud top height as retrieved for 4 selected states of SCIAMACHY is shown in 6, 1813–1840, 2006 Fig. 12. Clear sky pixels are represented by values of CTH equal or below zero. We 5 see a high variability of cloud altitudes in the SCIAMACHY states studied. Also the most northern state is characterized by quite high and extended cloudiness. This also Semianalytical cloud points to the fact that crystals may exist in these clouds. retrieval algorithm
A. Kokhanovsky et al. 5 The calibration of SCIAMACHY
First comparisons of radiances and irradiances measured by SCIAMACHY with inde- Title Page 10 pendent sources indicate an error in the present absolute radiometric calibration, which has a strong impact on the quality of most level-2 data products. So results presented Abstract Introduction
above are necessarily biased. To overcome this problem, an extensive analysis of the Conclusions References radiometric ground calibration measurements of SCIAMACHY has been performed, and a new procedure has been developed to recalculate some of the radiometric key Tables Figures 15 data from existing end-to-end measurements. These calculations were primarily based on a subset of NASA sphere measurements, performed for SCIAMACHY’s radiance J I and irradiance verification during the OPTEC 5 period in 1999/2000. This integrating sphere is a 20-inch diameter internally illuminated sphere coated J I with BaSO4. It has a long history of providing accurate absolute radiances for NASA’s Back Close 20 SBUV2 and TOMS programs and has also been used for the validation of the GOME absolute radiance calibration. The derived new SCIAMACHY key data show a signif- Full Screen / Esc icant difference to the on-ground ambient measured and calculated bidirectional re- flectance distribution function (BRDF) keydata of SCIAMACHY’s Elevation Scan Mirror Printer-friendly Version Diffuser (ESM diffuser). Together with a re-determined nadir and limb sensitivity of the Interactive Discussion 25 instrument, this leads to correction factors for both solar irradiance and reflectance. First tests with in-flight measurements show a significant improvement of the quality of the level-1 data products when using these new key data. EGU 1821 New calibration results will be discussed in a separate publication. We only note that the derived calibration correction factor is nearly independent on the wavelength ACPD in the region 420–500 nm and is equal to 1.07. It is equal to 1.1 in the oxygen A- 6, 1813–1840, 2006 band. It slightly varies with the wavelength in 1550–1670 nm band as shown in Fig. 13. 5 Therefore, it may influence the retrievals of the CPI. Let us discuss the improvements in the retrieved cloud products as obtained after Semianalytical cloud application of newly derived calibration coefficient k. We define this coefficient as a retrieval algorithm multiplier, which should be applied to the reflection function specified above to account for improved calibration. Note that this multiplier is larger than one. This allows to state A. Kokhanovsky et al. 10 that in reality the reflection function is larger than given by the operational level 1 data. This leads to the underestimation of the cloud optical thickness and overestimation the Title Page size of scatterers if the coefficient k is ignored. In particular, it follows that the ratio A=τs/τm is given by Eq. (3) with C=1.05, if the coefficient k is applied to operational Abstract Introduction level-1 data. This means that A becomes closer to one as compared to data without Conclusions References 15 any additional calibration, thus increasing the consistency between SCIAMACHY and
MERIS reflectances. We compare retrievals of cloud effective radius, cloud liquid water Tables Figures path and cloud optical thickness with our new calibration and without it in Figs. 14–16 . It follows that the account for the calibration correction factor allows to reduce aef and J I increase w moving them to results usually obtained for cloud fields in situations similar 20 to that given in Fig. 2. J I
Back Close 6 Conclusions Full Screen / Esc We show that the SACURA is capable to derive important cloud parameters using SCIAMACHY data such as the cloud optical thickness, the cloud top height, the liquid Printer-friendly Version water path, the effective radius, and the cloud thermodynamic state. The results of Interactive Discussion 25 the retrievals may be biased due to the problems with the calibration of SCIAMACHY, if uncorrected operational level 1 data are used. In this paper we have used the Pro- cessor V3.51 L1 data. When the calibration coefficients will be finally approved they EGU 1822 will be applied to all SCIAMACHY data. The SACURA may be a suitable candidate as the SCIAMACHY operational cloud retrieval algorithm then. However, for this two ACPD other problems should be solved, namely, the speed of the retrieval should be further 6, 1813–1840, 2006 enhanced and also the partly cloudy scenes must be taken into account. In conclusion, 5 we underline that the SCIAMACHY provides an atmospheric research community with a wealth of new and important data reflecting global atmospheric conditions (includ- Semianalytical cloud ing not only cloud properties but also trace gases abundances (Bovensmann et al., retrieval algorithm 2004) and also atmospheric aerosol characteristics (Hoyningen-Huene et al., 2006)). This is of a great importance for the atmospheric research in general and for the cli- A. Kokhanovsky et al. 10 mate change problems in particular. One can find the cloud products as generated by SACURA on day-to-day basis at the website of the Institute of Remote Sensing Title Page (Bremen University) (http://www.iup.physik.uni-bremen.de/scia-arc). Abstract Introduction Acknowledgements. This work has been funded by the DFG Project BU 6888/8-1 and also by BMBF via GSF/PT-UKF. We are grateful to the AVIRIS team for the spectrum shown in Conclusions References 15 Fig. 8 and to S. Janz (NASA/GSFC) and TPD-TNO for providing the SCIAMACHY calibration data. MERIS and SCIAMACHY data were supplied by ESA. Discussions with P. Stammes and Tables Figures E. P. Zege are highly appreciated. We thank B. J. Kerridge and B. Latter for providing results of their comparisons of the top of atmosphere reflectance measured by AATSR with that obtained by SCIAMACHY. J I J I
20 References Back Close
Acarreta, J. R., Stammes, P., and Knap, W. H.: First retrieval of cloud phase from SCIAMACHY Full Screen / Esc spectra around 1.6 µm, Atmos. Res., 72, 89–105, 2004. Bovensmann, H., Burrows, J. P., Buchwitz, M., Frerick, J., Noel,¨ S., and Rozanov, V. V.: SCIA- Printer-friendly Version MACHY: Mission objectives and measurement method, J. Atmos. Sci., 56, 127–150, 1999. 25 Bovensmann, H., Buchwitz, M., Frerick, J., et al.: SCIAMACHY on ENVISAT: In-flight optical Interactive Discussion performance and first results, in: Remote Sensing of Clouds and the Atmosphere, edited by: Schafer,¨ K. P., Comeron, A., Carleer, M. R., and Picard, R. H., SPIE Proc., 5235, 160–173, 2004. EGU 1823 Crutzen , P. J. and Ramanathan, V. J. (Eds.): Clouds, Chemistry, and Climate, Springer, Berlin, 1996. ACPD Kokhanovsky, A. A. and Rozanov, V. V.: The reflection function of optically thick weakly ab- sorbing turbid layers: a simple approximation, J. Quant. Spectr. Rad. Transfer, 77, 165–175, 6, 1813–1840, 2006 5 2003. Kokhanovsky, A. A. and Rozanov, V. V.: The physical parameterization of the top-of-atmosphere reflection function for a cloudy atmosphere-underlying surface system: the oxygen A-band Semianalytical cloud case study, J. Quant. Spectr. Rad. Transfer, 85, 35–55, 2004. retrieval algorithm Kokhanovsky, A. A., Rozanov, V. V., Nauss, T., Reudenbach, C., Daniel, J. S., Miller, H. L., and A. Kokhanovsky et al. 10 Burrows, J. P.: The semianalytical cloud retrieval algorithm for SCIAMACHY I. The validation, Atmos. Chem. Phys. Discuss., 5, 1995–2015, 2005. Kokhanovsky, A. A., Rozanov, V. V., Zege, E. P.,Bovensmann, H., and Burrows, J. P.: A semian- alytical cloud retrieval algorithm using backscattered radiation in 0.4–2.4 µm spectral region, Title Page J. Geophys. Res., 103, doi:10.1029/2001JD001543, 2003. 15 Raschke, E. (Ed.): Radiation and water in the Climate System, Springer, Berlin, 1996. Abstract Introduction Rozanov, V. V. and Kokhanovsky, A. A.: Semianalytical cloud retrieval algorithm as applied to the cloud top altitude and the cloud geometrical thickness determination from top-of- Conclusions References atmosphere reflectance measurements in the oxygen A band, J. Geophys. Res., 104, Tables Figures doi:10.1029/2003JD004104, 2004. 20 von Hoyningen-Huene, W., Kokhanovsky, A. A., Wuttke, M. W., et al.: Validation of SCIA- MACHY top-of-atmosphere reflectance for aerosol remote sensing using MERIS L1 data, J I Atmos. Chem. Phys. Discuss., 6, 673–699, 2006. J I
Back Close
Full Screen / Esc
Printer-friendly Version
Interactive Discussion
EGU
1824 ACPD 6, 1813–1840, 2006 80 Semianalytical cloud retrieval algorithm 70 N 1 A. Kokhanovsky et al. N 2
60 Title Page
Abstract Introduction
S 50 1 S 2 Conclusions References
latitude, degrees latitude, Tables Figures 40
J I
30 J I
Back Close
20 Full Screen / Esc 0 102030405060 longitude, degrees Printer-friendly Version
Fig. 1. The geographical position of the MERIS (red quadrangle) and SCIAMACHY (black Interactive Discussion lines) ground scenes analysed for the orbit 02223 (3 August 2002).
EGU
1825 ACPD 6, 1813–1840, 2006
Semianalytical cloud retrieval algorithm
A. Kokhanovsky et al.
Title Page
Abstract Introduction
Conclusions References
Tables Figures
J I
J I
Back Close
Full Screen / Esc
Printer-friendly Version
Interactive Discussion Fig. 2. The MERIS browse image for the same date and orbit as in Fig. 1.
EGU
1826 ACPD 1.0 6, 1813–1840, 2006 R =0.89R s m 0.9 Semianalytical cloud
0.8 retrieval algorithm A. Kokhanovsky et al. 0.7
0.6 Title Page
Abstract Introduction 0.5
Conclusions References 0.4 Tables Figures
0.3 J I reflection function (SCIAMACHY) function reflection 0.2 J I
Back Close 0.1 Full Screen / Esc 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Printer-friendly Version reflection function (MERIS) Interactive Discussion Fig. 3. Scatter plot of the SCIAMACHY reflectance via that of MERIS. EGU
1827 ACPD 40 6, 1813–1840, 2006
Semianalytical cloud retrieval algorithm
20 A. Kokhanovsky et al.
Title Page
Abstract Introduction 0
Conclusions References
Tables Figures
J I -20 J I
Back Close SCIA/MERIS relative difference in reflectance, %
Full Screen / Esc -40 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Printer-friendly Version SCIAMACHY TOA reflectance Interactive Discussion Fig. 4. The relative difference of the SCIAMACHY reflectance and that of MERIS (in percent). EGU
1828 ACPD
1.0 6, 1813–1840, 2006
Semianalytical cloud
0.8 retrieval algorithm A. Kokhanovsky et al.
0.6 Title Page
Abstract Introduction
A Conclusions References 0.4
Tables Figures
J I 0.2
J I
Back Close
0.0 Full Screen / Esc 0.0 0.2 0.4 0.6 0.8 1.0 R m Printer-friendly Version
Fig. 5. The dependence of the parameter A (see Eq. 3) on the reflection function (the solid Interactive Discussion line).
EGU
1829
50 ACPD 45 6, 1813–1840, 2006
40 Semianalytical cloud retrieval algorithm τ =3τ /4 35 s m A. Kokhanovsky et al. 30
Title Page 25
Abstract Introduction 20 Conclusions References
15 Tables Figures optical thickness(SCIAMACHY) optical 10 J I
J I 5 Back Close 0 0 5 10 15 20 25 30 35 40 45 50 Full Screen / Esc optical thickness (MERIS) Printer-friendly Version Fig. 6. Scatter plot of the SCIAMACHY-retrieved cloud optical thickness via that of MERIS- retrieved. Interactive Discussion
EGU
1830 ACPD 180 6, 1813–1840, 2006 water 160 ice Semianalytical cloud retrieval algorithm 140 A. Kokhanovsky et al.
120
Title Page 100 Abstract Introduction
80 Conclusions References frequency Tables Figures 60
J I 40 J I
Back Close 20
Full Screen / Esc 0 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Printer-friendly Version CPI Interactive Discussion Fig. 7. The modeled CPI frequency distribution. EGU
1831 ACPD
6, 1813–1840, 2006 3400 soil Semianalytical cloud 3300 retrieval algorithm
A. Kokhanovsky et al. 3200
Title Page 3100
Abstract Introduction
3000 Conclusions References
Tables Figures 2900 reflected light intensity, a.u. intensity, light reflected J I
2800 J I
Back Close 2700 1.45 1.50 1.55 1.60 1.65 1.70 Full Screen / Esc wavelength, microns Printer-friendly Version Fig. 8. The spectral reflectance of soil as measured by the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS). Interactive Discussion
EGU
1832 ACPD 500 6, 1813–1840, 2006
Semianalytical cloud retrieval algorithm 400 A. Kokhanovsky et al.
300 Title Page
Abstract Introduction
Conclusions References
frequency 200 Tables Figures
J I
100 J I
Back Close
Full Screen / Esc 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Printer-friendly Version CPI Interactive Discussion Fig. 9. The measured CPI frequency distribution. EGU
1833 ACPD 60 6, 1813–1840, 2006
Semianalytical cloud retrieval algorithm
A. Kokhanovsky et al.
40
Title Page
Abstract Introduction
Conclusions References frequency Tables Figures 20
J I
J I
Back Close
Full Screen / Esc 0 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Printer-friendly Version CPI Interactive Discussion Fig. 10. The measured CPI (similar to Fig. 11 but pixels with R(442 nm)≤0.3 are excluded). EGU
1834 ACPD 1.0 6, 1813–1840, 2006
0.9 experiment fit(H=6.3km,δH=5.9km, τ=15.5) Semianalytical cloud 0.8 retrieval algorithm A. Kokhanovsky et al. 0.7
0.6 Title Page
Abstract Introduction 0.5
Conclusions References 0.4 Tables Figures reflection function reflection 0.3 J I
0.2 J I
Back Close 0.1 Full Screen / Esc 0.0 756 758 760 762 764 766 768 770 772 Printer-friendly Version wavelength, nm Interactive Discussion Fig. 11. The measured oxygen A-band spectrum and the correspondent fit. EGU
1835 ACPD 6, 1813–1840, 2006
Semianalytical cloud
12 retrieval algorithm m
k 2 1
,
t 10 A. Kokhanovsky et al.
h g
i 8
e h
6 Title Page
p o
t 4
Abstract Introduction
d u 2
Conclusions References
o l
c 0 Tables Figures -2 -4 60 J I 45 50 es 50 40 re J I 55 30 eg la 60 , d titu 65 20 e Back Close de ud , d 70 10 it e 75 g Full Screen / Esc gre on es 80 l Printer-friendly Version Fig. 12. The retrieved cloud top height for several SCIAMACHY states. Interactive Discussion
EGU
1836 ACPD
1.20 6, 1813–1840, 2006
1.18 Semianalytical cloud retrieval algorithm 1.16 A. Kokhanovsky et al. 1.14
1.12 Title Page
Abstract Introduction 1.10
Conclusions References 1.08 Tables Figures 1.06
calibration correction factor J I 1.04 J I
1.02 Back Close
1.00 Full Screen / Esc 1560 1580 1600 1620 1640 1660 wavelength, nm Printer-friendly Version
Fig. 13. The calibration correction factor for the reflection function. Interactive Discussion
EGU
1837 ACPD 6, 1813–1840, 2006
120 Semianalytical cloud retrieval algorithm 100 A. Kokhanovsky et al.
80 Title Page
Abstract Introduction 60 Conclusions References frequency
Tables Figures 40
J I
20 J I
Back Close
0 Full Screen / Esc 5 1015202530 effective radius, µm Printer-friendly Version Fig. 14. The effective radius histogram. More dense bars give results of the retrieval before the Interactive Discussion application of the new calibration correction factor (see Fig. 13).
EGU
1838 ACPD 1000 6, 1813–1840, 2006
Semianalytical cloud retrieval algorithm 800 A. Kokhanovsky et al.
600 Title Page
Abstract Introduction
Conclusions References
frequency 400 Tables Figures
J I 200 J I
Back Close
Full Screen / Esc 0 0 100 200 300 400 500 600 700 liquid water path, g/m2 Printer-friendly Version
Interactive Discussion Fig. 15. The same as in previous figure but for w.
EGU
1839 ACPD 500 6, 1813–1840, 2006
Semianalytical cloud
400 retrieval algorithm A. Kokhanovsky et al.
300 Title Page
Abstract Introduction
Conclusions References
frequency 200 Tables Figures
J I
100 J I
Back Close
Full Screen / Esc 0 10 20 30 40 50 Printer-friendly Version cloud optical thickness Interactive Discussion Fig. 16. The same as in previous figure but for τ. EGU
1840