• Ackerman, S. A., , W. L. Smith, , H. E. Revercomb, , and J. D. Spinhirne, 1990: The 27–28 October 1986 FIRE IFO cirrus case study: Spectral properties of cirrus clouds in the 8–12 μm window. Mon. Wea. Rev, 118 , 23772388.

    • Search Google Scholar
    • Export Citation
  • Banta, R. M., , L. S. Darby, , P. Kaufmann, , D. H. Levinson, , and C. J. Zhu, 1999: Wind-flow patterns in the Grand Canyon as revealed by Doppler lidar. J. Appl. Meteor, 38 , 10691083.

    • Search Google Scholar
    • Export Citation
  • Baum, B. A., , D. P. Kratz, , P. Yang, , S. C. Ou, , Y. Hu, , P. F. Soulen, , and S. C. Tsay, 2000: Remote sensing of cloud properties using MODIS Airborne Simulator imagery during SUCCESS. I. Data and models. J. Geophys. Res, 105 , 1176711780.

    • Search Google Scholar
    • Export Citation
  • Chepfer, H., , J. Pelon, , G. Brogniez, , C. Flamant, , V. Trouillet, , and P. H. Flamant, 1999: Impact of cirrus cloud ice crystal shape and size on multiple scattering effects: Application to spaceborne and airborne backscatter lidar measurements during LITE mission and E LITE campaign. Geophys. Res. Lett, 26 , 22032206.

    • Search Google Scholar
    • Export Citation
  • Chiriaco, M., , H. Chepfer, , V. Noel, , A. Delaval, , M. Haeffelin, , P. Dubuisson, , and P. Yang, 2004: Coupling infrared radiometer and lidar observations for retrieving semi-transparent cirrus clouds particle size. Mon. Wea. Rev, 132 , 16841700.

    • Search Google Scholar
    • Export Citation
  • Collis, R. T. H., , and P. B. Russell, 1976: Lidar measurements of particles and gases by elastic backscattering and differential absorption. Laser Monitoring of the Atmosphere. E. D. Hinkley, Ed. Springer-Verlag, 71–151.

    • Search Google Scholar
    • Export Citation
  • Dabas, A. M., , P. Drobinski, , and P. H. Flamant, 1998: Chirp-induced bias in velocity measurements by a coherent Doppler CO2 lidar. J. Atmos. Oceanic Technol, 15 , 407415.

    • Search Google Scholar
    • Export Citation
  • Dabas, A. M., , P. Drobinski, , and P. H. Flamant, 2000: Velocity biases of adaptive filter estimates in heterodyne Doppler lidar measurements. J. Atmos. Oceanic Technol, 17 , 11891202.

    • Search Google Scholar
    • Export Citation
  • Dabas, A. M., , P. Drobinski, , O. Reitebuch, , E. Richard, , P. Delville, , P. H. Flamant, , and C. Werner, 2003: Multiscale analysis of a straight jet streak using numerical analyses and airborne Doppler lidar measurements. Geophys. Res. Lett, 30 .1049, doi:10.1029/2001GL014040.

    • Search Google Scholar
    • Export Citation
  • Drobinski, P., , R. A. Brown, , P. H. Flamant, , and J. Pelon, 1998: Evidence of organised large eddies by ground-based Doppler lidar, sonic anemometer and sodar. Bound.-Layer Meteor, 88 , 343361.

    • Search Google Scholar
    • Export Citation
  • Drobinski, P., , C. Haeberli, , E. Richard, , M. Lothon, , A. M. Dabas, , P. H. Flamant, , M. Furger, , and R. Steinacker, 2003: Scale interaction processes during MAP-IOP 12 south foehn event in the Rhine valley. Quart. J. Roy. Meteor. Soc, 129 , 729754.

    • Search Google Scholar
    • Export Citation
  • Dubuisson, P., , J. C. Buriez, , and Y. Fouquart, 1996: High spectral resolution solar radiative transfer in absorbing and scattering media: Application to the satellite simulation. J. Quant. Spectrosc. Radiat. Transfer, 55 , 103126.

    • Search Google Scholar
    • Export Citation
  • Elouragini, S., 1995: Etude des nuages par télédétection radar laser (lidar). Ph.D. thesis, Faculté des Sciences de Tunis, 75 pp.

  • Goodman, J. W., 1984: Statistical properties of laser speckles patterns. Laser Speckles and Related Phenomena, J. C. Dainty, Ed., Topics in Applied Physics, Vol. 9, Springer Verlag, 51–58.

    • Search Google Scholar
    • Export Citation
  • Guérit, G., , P. Drobinski, , P. H. Flamant, , and J. P. Cariou, 2002: Effectiveness of simultaneous independent realizations at low carrier-to-noise ratio to improve heterodyne Doppler lidar performance. Part I: Theory and numerical simulations. Appl. Opt, 41 , 22322239.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., 1975: Cirrus uncinus generating cells and the evolution of cirriform clouds. Part I: Aircraft observations of the growth of the ice phase. J. Atmos. Sci, 32 , 799807.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., 1993: Microphysical structure of stratiform and cirrus clouds. Aerosol Cloud Climate Interactions, P. V. Hobbs, Ed., International Geophysics Series, Vol. 54, Academic Press, 97–121.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., , and C. M. R. Platt, 1984: A parameterization of the particle size spectrum of ice clouds in terms of the ambient temperature and the ice water content. J. Atmos. Sci, 41 , 846855.

    • Search Google Scholar
    • Export Citation
  • Inoue, T., 1985: On the temperature and effective emissivity determination of semi-transparent cirrus clouds by bispectral measurements in the 10 microns window region. J. Meteor. Soc. Japan, 63 , 8899.

    • Search Google Scholar
    • Export Citation
  • Intrieri, J. M., , G. L. Stephens, , W. L. Eberhard, , and T. Uttal, 1993: A method for determining cirrus cloud particle sizes using lidar and radar backscatter technique. J. Appl. Meteor, 32 , 10741082.

    • Search Google Scholar
    • Export Citation
  • Intrieri, J. M., , W. L. Eberhard, , T. Uttal, , J. A. Shaw, , J. B. Snider, , Y. Han, , and B. W. Orr, 1995: Multi-wavelength observations of a developing cloud system: The FIRE II 26 November 1991 case study. J. Atmos. Sci, 52 , 40794093.

    • Search Google Scholar
    • Export Citation
  • King, M. D., and Coauthors, 2003: Cloud and aerosol properties, precipitable water, and profiles of temperature and humidity from MODIS. IEEE Trans. Geosci. Remote Sens, 41 , 442458.

    • Search Google Scholar
    • Export Citation
  • Krupp, C., 1991: Holographic measurements of ice crystals in cirrus clouds during the International Cloud Experiment ICE 1989. Report of the Fourth ICE/EUCREX Workshop, Lille, France, Laboratoire d'Optique Atmosphérique, USTL.

  • Liou, K. N., 1986: Influence of cirrus clouds on weather and climate processes: A global perspective. Mon. Wea. Rev, 114 , 11671199.

  • Lynch, D. K., , K. Sassen, , D. O'C. Starr, , and G. Stephens, 2002: Cirrus. Oxford University Press, 504 pp.

  • McClatchey, R. A., , R. W. Fenn, , J. E. A. Selby, , F. E. Volz, , and J. S. Garing, 1972: Optical properties of the atmosphere. 3d ed. Air Force Cambridge Research Laboratory, Hanscom Air Force Base Publication AFCRL-72-0497, 108 pp.

  • Miloshevich, L. M., , and A. J. Heymsfield, 1996: A balloon-borne cloud particle replicator for measuring vertical profiles of cloud microphysics: Instrument design and performance. Proc. 12th Int. Conf. on Clouds and Precipitation, Zurich, Switzerland.

  • Minnis, P., , D. P. Garber, , and D. F. Young, 1998: Parameterizations of reflectance and effective emittance for satellite remote sensing of cloud properties. J. Atmos. Sci, 55 , 33133339.

    • Search Google Scholar
    • Export Citation
  • Mishchenko, M. I., 1991: Light scattering by randomly oriented axially symmetric particles. J. Opt. Soc. Amer, 79 , 871882.

  • Noel, V., , H. Chepfer, , G. Ledanois, , A. Delaval, , and P. H. Flamant, 2002: Classification of effective shape ratios in cirrus clouds based on lidar depolarization ratio. Appl. Opt, 41 , 42454257.

    • Search Google Scholar
    • Export Citation
  • Platnick, S., , M. D. King, , H. Gerber, , and P. V. Hobbs, 2001: A solar reflectance technique for cloud retrievals over snow and ice surfaces. J. Geophys. Res, 106 , 1518515199.

    • Search Google Scholar
    • Export Citation
  • Platt, C. M. R., 1973: Lidar and radiometric observations of cirrus clouds. J. Atmos. Sci, 30 , 11911204.

  • Platt, C. M. R., , S. A. Young, , P. J. Manson, , G. R. Patterson, , S. C. Marsden, , and T. Austin, 1998: The optical properties of equatorial cirrus from observations in the ARM pilot Radiation Observation Experiment. J. Atmos. Sci, 55 , 19771996.

    • Search Google Scholar
    • Export Citation
  • Post, M. J., , and W. D. Neff, 1986: Doppler lidar measurements of winds in a narrow mountain valley. Bull. Amer. Meteor. Soc, 67 , 274281.

    • Search Google Scholar
    • Export Citation
  • Sassen, K., 1991: The polarization lidar technique for cloud research: A review and current assessment. Bull. Amer. Meteor. Soc, 72 , 18481866.

    • Search Google Scholar
    • Export Citation
  • Sassen, K., , and K. N. Liou, 1979: Scattering of polarized laser light by water droplet, mixed-phase and ice crystal clouds. Part II: Angular depolarizing and multiple-scattering behavior. J. Atmos. Sci, 36 , 852861.

    • Search Google Scholar
    • Export Citation
  • Spinhirne, J. D., 1982: Lidar clear atmosphere multiple scattering dependence on receiver range. Appl. Opt, 21 , 24672468.

  • Stephens, G. L., , S. C. Tsay, , P. W. Stackhouse, , and P. J. Flatau, 1990: The relevance of the microphysical and radiative properties of cirrus clouds to climate and climate feedback. J. Atmos. Sci, 47 , 17421753.

    • Search Google Scholar
    • Export Citation
  • Stubenrauch, C. J., , W. B. Rossow, , F. Cheruy, , A. Chedin, , and N. A. Scott, 1999: Cloud as seen by satellite sounders (3I) and imagers (ISCCP). Part I: Evaluation of cloud parameters. J. Climate, 12 , 21892213.

    • Search Google Scholar
    • Export Citation
  • Ulaby, F. T., , R. K. Moore, , and A. K. Fung, 1986: Microwave Remote Sensing: Active and Passive. Vol. 1, Microwave Remote Sensing Fundamentals and Radiometry, Artech House, 456 pp.

    • Search Google Scholar
    • Export Citation
  • Uttal, T., , E. E. Clothiaux, , T. P. Ackerman, , J. M. Intrieri, , and W. L. Eberhard, 1995: Cloud boundary statistics during FIRE II. J. Atmos. Sci, 52 , 42764284.

    • Search Google Scholar
    • Export Citation
  • Winker, D. M., , J. Pelon, , and M. P. McCormick, 2003: The CALIPSO mission: Spaceborne lidar for observation of aerosols and clouds. Lidar Remote Sensing for Industry and Environmental Monitoring III, U. N. Singh, T. Itabe, and Z. Liu, Eds., International Society for Optical Engineering, (SPIE Proceedings Vol. 4893), 1–11.

  • Yang, P., , B. Gao, , B. A. Baum, , Y. X. Hu, , W. J. Wiscombe, , S. Tsay, , and D. M. Winker, 2001: Radiative properties of cirrus couds in the infrared (8-13μm) spectral region. J. Quant. Spectrosc. Radiat. Transfer, 70 , 473504.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    From the Yang et al. (2001) calculation: (a) Qabs,part,10.6 = f (Qsca,part,10.6) for crystals in class 1, where the line is the linear fit between Qabs,part,10.6 and Qsca,part,10.6; (b) same as (a), but for class 2; (c) same as (a), but for class 3; (d) same as (a), but for class 4; (e) k10.6/k532 = f (r) for crystals in class 1, where the horizontal line is the constant value of γ that is taken into account in the retrieval method; (f) same as (e), but for class 2; (g) same as (e), but for class 3; (h) same as (e), but for class 4.

  • View in gallery

    Theoretical profiles of (a) IWC, (b) ice particle effective radius for two microphysics, (c) absorption efficiency at 10.6 μm, (d) concentration n, (e) N, (f) absorption coefficient at 10.6 μm, (g) lidar profiles at 10.6 μm, and (h) lidar profiles at 532 nm.

  • View in gallery

    Evolution of the logarithm of lidar backscattered signal as a function of time for the 6 Nov 2002 (a) 532-nm and (b) 10.6-μm backscattered signal.

  • View in gallery

    For 8 Oct 2002 with values averaged between 1210 and 1220 UTC: (a) 532-nm lidar profile, (b) 10.6-μm lidar profile for the four classes of microphysics, and (c) particle absorption coefficient profile at 10.6 μm αabs,part,10.6(z) for the four classes of microphysics.

  • View in gallery

    Same as Fig. 4, but for 14 Oct 2002 averaged between 1210 and 1220 UTC.

  • View in gallery

    Same as Fig. 4, but for 6 Nov 2002 averaged between 1005 and 1015 UTC.

  • View in gallery

    For 14 Oct 2002, αabs,part,10.6(z) retrieved using microphysics in class 3 and for values of Qsca,part,10.6(zbase) = 0.4 (solid line), 0.6 (dashed line), and 0.9 (dotted line).

  • View in gallery

    For 14 Oct 2002, αabs,part,10.6(z) retrieved using three different profiles of Qabs,part,10.6(z): profile retrieved using the described method for microphysics in class 1 (solid line), profile is constant in the cloud and equal to 0.7 (dashed line), and profile is constant in the cloud and equal to 0.9 (dotted line).

  • View in gallery

    Calculation of the atmospheric transmission at 532 nm (solid line), of its first first-order Taylor approximation (dotted–dashed lines), of the atmospheric transmission at 10.6 μm (dashed lines), and of its first first-order Taylor approximation (dotted line) for 14 Oct 2002, from the obtained results using microphysics in class 3.

  • View in gallery

    For the 14 Oct 2002 case using microphysics in class 3: (a) 10.6-μm lidar signal (solid line) and its variability (dashed line), and (b) αabs,part,10.6(z) profile (solid line) and its variability calculated from lidar variability (dashed line).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 11 11 4
PDF Downloads 5 5 2

Dual Lidar Observations at 10.6 μm and 532 nm for Retrieving Semitransparent Cirrus Cloud Properties

View More View Less
  • 1 IPSL/LMD, Université Pierre et Marie Curie, Paris, France
  • | 2 AS&M, Hampton, Virginia
  • | 3 IPSL/LMD, Université Pierre et Marie Curie, Paris, France
  • | 4 IPSL/SA, Université Pierre et Marie Curie, Paris, France
© Get Permissions
Full access

Abstract

To improve the estimation of the infrared radiances in cirrus clouds, one needs to consider the vertical inhomogeneities of the cloud properties. The position of the maximum of absorption within an ice cloud is potentially important to the improvement of the split-window techniques for retrieving particle size and for understanding the radiative effect of the cloud in the infrared spectrum. Current remote sensing techniques used for inferring ice clouds hardly reach the level of accuracy required to resolve the vertical inhomogeneities of a cloud and to determine the position of absorption. This study explores the possibility of retrieving the vertical structures of ice clouds by combining data from two lidar measurements acquired at the wavelengths of 532 nm and 10.6 μm. A method is proposed to retrieve the variability of ice crystal absorption efficiency at 10.6 μm, the particle concentration weighted by the crystal area, and the attenuation by absorption at 10.6 μm. The method is tested against observations collected at Site Instrumental de Recherche en Télédétection Atmosphérique (SIRTA) in Palaiseau, France. Observations and simulations both show that lidar observations collected simultaneously at those two wavelengths can be used to determine the level within the ice cloud where maximum attenuation of infrared radiation occurs. The maximum attenuation may occur near the cloud base or the cloud top, depending on the case studied.

Corresponding author address: Marjolaine Chiriaco, LMD, Tour 45-55 3e étage, BP 99, 4 place Jussieu, 75252 Paris Cedex 05, France. Email: chiriaco@lmd.jussieu.fr

Abstract

To improve the estimation of the infrared radiances in cirrus clouds, one needs to consider the vertical inhomogeneities of the cloud properties. The position of the maximum of absorption within an ice cloud is potentially important to the improvement of the split-window techniques for retrieving particle size and for understanding the radiative effect of the cloud in the infrared spectrum. Current remote sensing techniques used for inferring ice clouds hardly reach the level of accuracy required to resolve the vertical inhomogeneities of a cloud and to determine the position of absorption. This study explores the possibility of retrieving the vertical structures of ice clouds by combining data from two lidar measurements acquired at the wavelengths of 532 nm and 10.6 μm. A method is proposed to retrieve the variability of ice crystal absorption efficiency at 10.6 μm, the particle concentration weighted by the crystal area, and the attenuation by absorption at 10.6 μm. The method is tested against observations collected at Site Instrumental de Recherche en Télédétection Atmosphérique (SIRTA) in Palaiseau, France. Observations and simulations both show that lidar observations collected simultaneously at those two wavelengths can be used to determine the level within the ice cloud where maximum attenuation of infrared radiation occurs. The maximum attenuation may occur near the cloud base or the cloud top, depending on the case studied.

Corresponding author address: Marjolaine Chiriaco, LMD, Tour 45-55 3e étage, BP 99, 4 place Jussieu, 75252 Paris Cedex 05, France. Email: chiriaco@lmd.jussieu.fr

1. Introduction

Cirrus clouds play a major role in the radiative energy budget of the earth–atmosphere system (Liou 1986; Stephens et al. 1990; Lynch et al. 2002). Their net effect is mainly governed by the competition between their albedo and greenhouse effects. Both macrophysical and microphysical properties of cirrus clouds regulate these two effects. The global coverage, as well as the altitude, temperature, vertical structure, and spatial inhomogeneities of these clouds should be properly taken into account to quantify their effects. Additionally, ice water content and its spatial distribution play a significant role in the global radiative effect of cirrus clouds. Moreover, one of the main uncertainties in their radiative impact is because of the poor knowledge of the natural variability of their microphysical properties, such as ice crystal concentration, particle effective radius, shape, and orientation of ice crystals in space. This lack of information is especially important for semitransparent ice clouds, which are potentially very efficient for producing greenhouse warming.

Quantifying the role of these clouds in the far-infrared domain is a particularly challenging problem, because in this part of the spectrum, semitransparent ice clouds contribute to warming of the atmosphere. At these wavelengths, absorption and emission processes are predominant; however, recent studies (Yang et al. 2001) have quantified the slight, but nonnegligible effect of scattering phenomena by ice crystals. Moreover, a correct estimation of the outgoing infrared radiances above and below an ice cloud requires a good knowledge of its internal structure (i.e., the vertical variability of particle concentration, effective radius, and optical properties). For example, Chiriaco et al. (2004) have shown that considering different positions of the maximum of absorption within a cloud (cloud top, base, and middle) can induce a 100% difference on the crystal size retrieval. Even if some advanced methods (Stubenrauch et al. 1999) based on infrared sounders have significantly contributed to improved observations of the vertical structure of the atmosphere in the infrared domain, most observations at these wavelengths do not allow documenting the vertical distribution of crystals in cirrus clouds.

The current study explores the potential for coupling data from two lidars to infer the vertical distribution of ice and particles within midlatitude semitransparent cirrus clouds. The first lidar operates in the visible domain (532 nm), and has been used for studying semitransparent ice clouds for years (Sassen and Liou 1979; Spinhirne 1982; Sassen 1991; Noel et al. 2002). The second operates in the far infrared (10.6 μm), and has been principally used to study the dynamics of the lower atmosphere (Post and Neff 1986; Drobinski et al. 1998; Banta et al. 1999) or cirrus clouds (Intrieri et al. 1993, 1995; Uttal et al. 1995; Dabas et al. 2003). The vertical resolution of the two lidars allows a precise observation of the different layers inside the cloud, using their two complementary wavelengths.

This paper proceeds as follows: presented in section 2 is a theoretical study regarding the sensitivities of lidar signals (at both 532-nm and 10.6-μm wavelengths) to the vertical distribution of the ice crystal optical properties. Based on these results, a method is proposed (section 3) for retrieving the vertical variations of three different variables characterizing the ice cloud: the particle absorption efficiency at 10.6 μm, the particle concentration weighted by ice crystal surface area, and the attenuation by absorption at 10.6 μm. Section 4 presents the cloud observations at the Site Instrumental de Recherche en Télédétection Atmosphérique (SIRTA) in Palaiseau, France and the characteristics of the two lidar instruments. Section 5 describes the retrieval of the cloud variables for three different cirrus cloud cases. The theoretical and observed results are discussed in section 6. The conclusions of this study are summarized in section 7.

2. Theoretical study

The sensitivity of 10.6-μm and 532-nm lidar signals to various ice cloud properties is examined to estimate the potential for combining observations from both instruments to improve the cloud property retrieval method. Actually, the maximum of absorption at 10.6 μm within the cloud is not necessarily at the same altitude as the maximum of the lidar backscatter signal at 532 nm, because the pure scattering effect at 532 nm is not simply replaced by a pure absorbing effect at 10.6 μm.

a. Properties of ice crystals and cirrus clouds

The natural variability of ice crystals size and shape is very large (Heymsfield 1975, 1993; Heymsfield and Platt 1984; Krupp 1991; Miloshevich and Heymsfield 1996) and cannot be fully taken into account in the current theoretical study. We choose to describe ice clouds with 18 types of randomly oriented particles presented in Table 1. Each ice crystal is defined by the particle habit and the radius of a volume-equivalent sphere (r). The following optical properties are required for the current study: (i) the scattering efficiency at 532 nm Qsca,part,532; (ii) the extinction (Qext,part,10.6), scattering (Qscat,part,10.6), and absorption (Qabs,part,10.6), efficiencies at 10.6 μm, linked by Qext,part,10.6 = Qabs,part,10.6 + Qscat,part,10.6; and (iii) the lidar particle backscatter-to-extinction ratio k532 and k10.6. The method developed in this study does not require the knowledge of k532 and k10.6, but only the ratio γ = k10.6/k532.

At 532 nm, optical properties are calculated by the geometric optics method (GOM). At 10.6 μm, they are calculated by GOM for large effective radius (r2 and r3 in Table 1), and by the finite-difference time domain (FDTD) method for small particles (r1). A linear relation between Qsca,part,10.6 and Qabs,part,10.6 is deduced from values given in Table 1 for each crystal habit (Figs. 1a–d):
i1558-8432-45-4-537-e1
The a and b coefficients are different for each habit, but they are sufficiently close for some habits to associate them into four classes (given in Table 1); for example, a = 0.32 and b = 0.59 for hollow columns, and a = 0.29 and b = 0.60 for plates; hence, hollow columns and plates are associated into a similar class. Figures 1e–h show that ratios γ = k10.6/k532 may be associated into the same classes, in which the ones value of the γ coefficient can be considered as constant. Nevertheless, Fig. 1f shows that this approximation is not valuable for aggregates when r < 15 μm, introducing uncertainty on the following retrieval method (section 3) when considering crystal in class 2. Habits are regrouped into classes in order to allow the crystal habit to vary within the cloud, because in the following retrieval method this class is fixed in the cloud.

b. Lidar equations at 532 nm and 10.6 μm

1) General lidar equation

At all wavelengths, the general lidar equation can be written as
i1558-8432-45-4-537-e2
where Sλ(z) is the backscattered signal, Kλ is a calibration constant, and η is a multiple scattering correction parameter. The lidar backscattering coefficients by particles βsca,part,λ(z) can be written as functions of kλ(z) and the scattering coefficient αsca,part,λ(z) or the scattering efficiency Qsca,part,532(z), if one considers that there is not any particle size distribution but one single particle size:
i1558-8432-45-4-537-e3
where N(z) is the particle concentration n(z) weighted by the ice crystal surface area: N(z) = n(z)πr2. Molecules contribute to the lidar backscattering coefficients [βsca,mol,λ(z)] and to the attenuation [αext,mol,λ(z)], and kλ is the lidar backscatter-to-extinction ratio. Extinction by aerosols is not considered in the following study, because lidar signals are normalized just under the cloud base in order to not consider the lower aerosols layers.

2) Simplified lidar equation at 532 nm

At 532 nm, there is no absorption by ice particles [αabs,part,532(z) = 0] and molecules [αabs,mol,532(z) = 0]; hence, αext,part,532(z) = αsca,part,532(z) and αext,mol,532(z) = αsca,mol,532(z).

Moreover, the scattering efficiency Qsca,part,532(z) is equal to 2 (Ulaby et al. 1986) for particles larger than 1 μm. The η coefficient is dependant on temperature and for the clouds considered in this study (that have temperature between 255 and 214 K), it varies between 0.36 and 0.54 (Platt et al. 1998). Nevertheless, other studies (Platt 1973; Chepfer et al. 1999) show that for the particular case of a visible lidar located at ground and detecting cirrus clouds, this coefficient could be considered as being 0.5.

From (2) and (3), the lidar equation at 532 nm is
i1558-8432-45-4-537-e4
The “N” index indicates that the signal is normalized to the molecular contribution. This allows for removal of the constant Kλ. The contribution of molecules to scattering phenomena βsca,mol,532(z) and αsca,mol,532(z) depend on the temperature T(z) and pressure P(z) (Collis and Russell 1976):
i1558-8432-45-4-537-e5
and
i1558-8432-45-4-537-e6
where K is the Boltzmann constant (1.38 × 10−23 J K−1), and wavelength Λ is given in micrometers.

3) Simplified lidar equation at 10.6 μm

At 10.6 μm, the molecules contribute only to absorption [βsca,mol,10.6(z) = 0], hence: αext,mol,10.6(z) = αabs,mol,10.6(z). Furthermore, η is neglected at 10.6 μm because of the large contribution of absorption by particles and the weak value of the scattering size parameter. The lidar equation in (2) can be simplified as
i1558-8432-45-4-537-e7

In clear-sky conditions, water vapor dominates molecular absorption at 10.6 μm in the troposphere, but there is also absorption by CO2. Here, αabs,mol,10.6(z) is computed by considering both water vapor and CO2 on the basis of a line-by-line radiative transfer code (Dubuisson et al. 1996) and by assuming the atmospheric profile of a midlatitude standard atmosphere (McClatchey et al. 1972).

c. Example of lidar signals with two different cloud microphysics

The sensitivity of the lidar signals to variations in cloud properties are examined theoretically by simulating the lidar returns.

1) Clouds description

Two different clouds (A and B) are defined. Both clouds have the same ice water content (IWC) profile (Fig. 2a) with a maximum at 8000 m. Both clouds are composed of hollow columns with r ranging between 0 and 55 μm. The effective radius profiles differ for each cloud (Fig. 2b). For cloud A, the maximum of r is 1000 m below the peak in IWC, and for cloud B, it is located 700 m above the IWC maximum. Figure 2c shows the resulting Qabs,part,10.6(z). The maximum Qabs,part,10.6(z) is naturally located at the same altitude as the maximum r(z), because the largest particles are the most efficient absorbers.

The corresponding n(z) profiles are shown in Fig. 2d. Both clouds have two n(z) maxima located close to the cloud top and base, because at these locations r is small and IWC(z) is not zero. The principal maximum n(z) is at 8800 m for cloud A, that is, 800 m above the peak IWC(z), and at 7000 m for cloud B, that is, 1000 m under the IWC maximum, because of the r(z) differences between the two (Fig. 2b). In Fig. 2e N(z) is plotted. For both clouds, multiplying n(z) by the crystal surface area shifts the greatest concentrations closer to the maximum IWC: 200 m above for cloud A and 200 m below for cloud B.

2) Simulated lidar profiles at 532 nm

The simulated lidar profiles at 532 nm are shown in Fig. 2h. For clouds A and B, the maximum lidar backscattered signals occurs at the same altitudes as the maxima in N(z). The difference between n(z) (Fig. 2d) and the lidar profile is because of the vertical distribution of the particle size and the attenuation by particles and the contribution of scattering by molecules.

3) Simulated lidar profiles and absorption attenuation profiles at 10.6 μm

Figure 2f shows αabs,part,10.6(z). For both clouds, the maxima of αabs,part,10.6(z) are found at the same altitudes as the N(z) maxima. This correspondence is consistent with the definition of αabs,part,10.6(z) and the orders of magnitude of N(z) and Qabs,part,10.6(z). The variability of αabs,part,10.6(z) can be expressed as
i1558-8432-45-4-537-e8
From typical values of n(z) (from 1 to 104 m−3; Heymsfield and Platt 1984) and Qabs,part,10.6(z) (from 0.5 to 1; Yang et al. 2001), it is possible to retrieve their variability:
i1558-8432-45-4-537-e9
i1558-8432-45-4-537-e10
To first order, the variation in αabs,part,10.6(z) is governed by the variation in N(z) (Figs. 2e and 2f). The variation of αabs,part,10.6(z) is secondarily influenced by the absorption efficiency [Qabs,part,10.6(z)]; this very slight influence corresponds to the weak differences between Figs. 2e and 2f.

The simulated lidar profile at 10.6 μm is shown in Fig. 2g. For cloud A, the maximum lidar backscattered signal at 10.6 μm occurs at 8300 m, whereas the maximum αabs,part,10.6(z) is at 8200 m. For cloud B, the maximum 10.6-μm signal is at 7700, which is 100 m under the maximum of αabs,part,10.6(z).

3. Retrieval method

A method coupling the 532-nm and 10.6-μm lidar observations is proposed to retrieve Qabs,part,10.6(z), N(z), and αabs,part,10.6 (z). To retrieve these variables, an expression of N(z) is extracted from (4). This expression is then introduced in (7), leading to (11) where the γ term appears because k10.6/k532= γ. In the 10.6-μm lidar equation in (7), N(z) is replaced by this expression, which allows the introduction of the γ term in (11):
i1558-8432-45-4-537-e11

In midlatitudes, cirrus clouds are frequently optically thin. Hence, for those specific clouds, the arguments of the exponential parts are small so this part can be approximated with a Taylor series. The dependencies of the lidar signals as discussed for Fig. 2 suggest that a rough estimation of Qabs,part,10.6(z) is sufficient for correctly determining the variability of N(z) and αabs,part,10.6(z) within the cloud. Hence, in a primary stage, only the zeroth order of the series is used. In the second step, this estimation of Qabs,part,10.6(z) is used to obtain N(z), considering the first order of the Taylor series. For the third step, αabs,part,10.6(z) is deduced from Qabs,part,10.6(z) and N(z).

a. Retrieval of the ice crystal absorption efficiency at 10.6 μm

By only considering the zeroth order of the Taylor series for estimating Qabs,part,10.6(z), the transmission term is neglected. Even if rigorously valid for subvisible clouds only, this approximation will be applied to the whole semitransparent cloud in order to make a rough estimation of Qabs,part,10.6(z). The error induced by this approximation will be examined in section 6.

Under those conditions, Qsca,part,10.6(z) drops out of (11):
i1558-8432-45-4-537-e12
With this expression, values of Qabs,part,10.6 and Qext,part,10.6 can be deduced from (1).

b. Particle concentration retrieval

To retrieve N(z), a first-order Taylor series is applied to (9), and the integrands are written as discrete summations. It is then possible to write (11) at a level i, with cloud base at b:
i1558-8432-45-4-537-e13
The zeroth order of (13) is zero because of the Qsca,part,10.6(z) expression in (12):
i1558-8432-45-4-537-e14
To isolate N at a given level i, the contribution of each term at level i is dropped out of the sum terms. Equation (14) then leads to a second-degree polynomial form, A(i)N(i)2 + B(i)N(i) + C(i) = 0, where
i1558-8432-45-4-537-e15
i1558-8432-45-4-537-e16
i1558-8432-45-4-537-e17
where N(i) at each level zi can be retrieved by iterating step by step from the cloud base (zb) to the cloud top (zt). To initiate this procedure, the concentration in the first cloud-base layer N(b) is deduced from the value of Qsca,part,10.6(b) and the lidar equations at cloud base (zb):
i1558-8432-45-4-537-e18
i1558-8432-45-4-537-e19
The values of k532(b) and k10.6(b) are deduced from the Yang et al. (2001) database, consistent with the value of Qsca,part,10.6(b).

Equations (18) and (19) lead to two independent determinations of N(b). To have the best estimation, the average between those two relations is calculated. Values of N(b) are used to initialize (16) and (17) and to calculate N(z) at each level of the cloud.

c. Absorption coefficient retrieval at 10.6 μm

The particle absorption coefficient at 10.6 μm is obtained by multiplying the particle absorption efficiency (12) with the particle concentration:
i1558-8432-45-4-537-e20

4. Observations

The method proposed to retrieve the cloud properties using both infrared and visible lidar is applied to midlatitude cirrus clouds observed at the SIRTA ground-based site.

a. Instruments

1) lidar instruments

The SIRTA 532-nm lidar is a classical Nd-Yag zenith-viewing lidar, with a temporal resolution of 1 min and a vertical resolution of 15 m. It measures both the backscattered signal and linear depolarization ratio and is operated 4 days per week from 0800 to 2000 LT. The cloud-base and -top altitudes are derived from the lidar profile in the parallel channel with a threshold method.

The SIRTA 10.6-μm lidar is a heterodyne Doppler lidar (HDL) with scanning capabilities. It has been operated in field campaigns to study atmospheric flow dynamics (Drobinski et al. 2003). The HDL main characteristics are a single-mode TE-CO2 laser transmitter operating at 10.6 μm. The TE-CO2 laser pulse has a duration of about 3 μs and is characterized by a gain-switched spike, an intrapulse frequency chirp of 1–1.5 MHz, and a pulse spectral width of 0.2–0.3 MHz. The gain-switched spike lasts about 300 ns and contains most of the energy, so the spatial resolution for heterodyne intensity measurement (which is the relevant measurement in this study) is about 15 m. The average output energy is 170 mJ, and the pulse repetition frequency is 2 Hz. For the current study, there are 100 shots by line of sight averaged over a 1-min interval. Depending on the case, there were four or five lines of sight, which lead to a temporal resolution of 4 or 5 min. When there are four lines of sight, the directions are 5° at each cardinal point from the zenith angle. When there are five lines of sight, a zenith line of sight is added to the original four. The sampling resolution along each line of sight is 15 m.

The accuracy of S(z)10.6 is estimated in the appendix, and the uncertainty will be discussed in section 6.

2) lidar calibration

Outside cloud and aerosol layers, the 532-nm lidar signal is backscattered only by molecules. The 532-nm lidar signal is normalized with respect to the molecular backscattering profile of a particle-free zone below the cloud. The molecular backscattering coefficient βsca,mol,532(z) is computed with (5) using pressure and temperature profiles obtained from Météo-France radiosonde data that were taken daily at 1200 and 0000 UTC at Trappes, which is 15 km from SIRTA. The 532-nm lidar signal is then imposed to be equal to molecular one (that is computed) just under the base of the cloud, and in a range of 200 m, which allows the complete profile to be normalized.

The 10.6-μm lidar signal cannot be normalized to molecular signal, because at this wavelength, there is mainly absorption by molecules and no scattering, leading to a very low signal-to-noise ratio (SNR). Usual approaches consist in calibrating the 10.6-μm signal against a solid reflective target or against a low and thick cloud composed of liquid water (Elouragini 1995). Those approaches cannot be applied to calibrate the current dataset because such a target was not available, and no low, thick liquid water clouds were measured during the campaign. Hence, the 532-nm lidar information is used to normalize the 10.6-μm lidar signal by neglecting the cloud transmission at the base of the cloud at 532 nm as at 10.6 μm, permitting the coupling of the two lidar equations as follows. The calibration altitude is chosen as the first cloud layer for both lidars, so it has a very large SNR.

All of the parameters necessary for the normalization are known, except for the scattering efficiency for the particles at 10.6 μm. For effective radii larger than 15 μm, Qsca,part,10.6(z) ranges between 0.5 and 1. An arbitrary value of 0.6 is used to normalize the signal. The impact of this value will be discussed in section 6a. Then, the 10.6-μm lidar signals at other levels are normalized as follows:
i1558-8432-45-4-537-e21

b. Cloud cases

A multilidar minicampaign was conducted at SIRTA in October and November 2002 for testing the proposed method. Both lidars were operated simultaneously in nonoperational modes for the 10.6-μm lidar during a 2-h period whenever the right conditions were present—thin to thick cirrus clouds without lower-altitude cloudiness. As shown by Mishchenko (1991), oriented particles in cirrus clouds are associated with weak values of the lidar depolarization ratio and large values of the lidar-backscattered signal. This is not the case for the three studied clouds; hence, they do not correspond to particle with specific orientation. Table 2 summarizes the dates and time periods of the observations for both lidars.

Figure 3 shows an example of the temporal evolution of the lidar-backscattered signals at 532 nm (Fig. 3a) and 10.6 μm (Fig. 3b) for the 6 November case averaged over 5-min intervals because of the constraints imposed by the 10.6-μm measurement cycles. The two lidars were separated by 150 m. The two figures do not show exactly the same cloud boundaries, particularly the cloud top, which differs by ∼2000 m. Moreover, because the 10.6-μm signal is not sensitive to molecules, the 10.6-μm SNR is very weak outside of the cloud (Fig. 3b), whereas the 532-nm lidar signal becomes noisy 3500 m higher than the cloud top (Fig. 3a).

Those differences can be explained by the following three reasons: First, the 532-nm lidar signal is backscattered by both molecules and particles, and this allows for a significant SNR at the cloud edges, and so too at the cloud-top and -base altitudes. The 10.6-μm lidar signal, because of its larger wavelength, is not backscattered by molecules, only by cloud particles. Therefore, its SNR is very weak out of the cloud. Hence, the cloud top cannot be estimated from the 10.6-μm signal only. The synergy between the 10.6-μm and the 532-nm lidars removes the ambiguity. Second, for both lidars to avoid specular reflections resulting from eventual ice crystals horizontally oriented in space, the vertical line of sight has a small zenith angle. The real direction of the line of sight can fluctuate within a few degrees, and an offset between both lidars could exist, leading to discrepancies between the cloud base seen by the 10.6-μm lidar and the 532-nm lidar. For application of the retrieval method, the possible altitude difference between both lidars is corrected by aligning cloud bases for both lidar signals. Third, absorption occurs at 10.6 μm, but not at 532 nm. Hence, a part of the 10.6-μm signal is absorbed within the cloud, and the signal is attenuated quickly as compared with the 532-nm beam.

5. Results: Application to three case studies

a. 8 October 2002

Figures 4a and 4b depict both lidar signal profiles averaged between 1210 and 1220 UTC 8 October 2002, zoomed in the cloudy area. For this case, the line of sight is 5° to the north of the zenith, necessitating an angular correction of the 10.6-μm profile. The 532-nm lidar profile (Fig. 4a) is normalized using a molecular profile, as explained in section 4a. It shows a cloud between 6.35 and 9 km, that is, between 251 and 214 K. The 10.6-μm profile normalization depends on the choice of the microphysical model (Table 1). Hence, there are four different normalized 10.6-μm lidar profiles with similar shapes but different values. A cloud composed of ice crystals in class 2 (aggregates and bullet rosette) has the largest backscattered signal at 10.6 μm, and the smallest signal is for a cloud composed of crystals in class 3 (columns). For this cloud, whatever the microphysics, the maximum 10.6-μm lidar signal occurs 200 m lower than that of the 532-nm one (summarized in Table 3).

The method developed in section 3 is applied to this case and allows for the retrieval of Qabs,part,10.6(z), N(z), and αabs,part,10.6(z) (Fig. 4c). The αabs,part,10.6(z) profiles have the same shape whatever the cloud microphysics, but are maximal at the same altitude as the 10.6-μm lidar signal (i.e., 6.9 km), or 300 m lower, depending on the cloud microphysics (Table 3). The αabs,part,10.6(z) profiles start diverging at 7.3 km, certainly because the 10.6-μm SNR starts being too weak at this altitude. A cloud composed of crystals in class 4 (droxtals) absorbs the most at 10.6 μm. For a cloud composed of crystals in class 1 (hollow columns and plates) or 2 (aggregates and bullet rosettes), there is a large absorbing layer near the cloud base (at 6.6 km). Hence, the most absorbing layer in the infrared is located 200–500 m under the altitude where the 532-nm lidar signal is maximal, depending on the cloud microphysics.

b. 14 October 2002

The cloud observed during 14 October 2002 is located at a lower altitude than the previous case. The temperatures range between 255 and 240 K. This cloud was relatively warm, but the average lidar depolarization ratio is 30%, which shows that it is composed of ice. Furthermore, the 532-nm lidar signal is strongly attenuated, indicating that the true altitude of the cloud top is not known. Nevertheless, the retrieval method can still be applied to the low cloud layers.

The bispectral lidar-normalized profiles (Figs. 5a and 5b) indicate a cloud between 5.4 and 8 km. Both profiles are averaged between 1210 and 1220 UTC, and the line of sight is the same as for the 8 October 2002 case, requiring an angular correction. As for the 8 October case, the 10.6-μm lidar profiles are normalized for the four microphysics classes (Fig. 5b). The four resulting profiles have a similar shape, with a maximum located 200 m under the maximum 532-nm lidar signal. The αabs,part,10.6(z) restitution is shown in Fig. 5c. The four αabs,part,10.6(z) profiles have the same shape, with a similar tendency as that for 8 October 2002: a cloud composed of droxtals absorbs the most in the infrared. Depending on the cloud microphysics among the four proposed, the most absorbent layer in the infrared is located between 5.5 and 5.8 km, that is, between the 532-nm and the 10.6-μm lidar signal maximum (also shown in Table 3).

c. 6 November 2002

For 6 November 2002, the 532-nm lidar profile (Fig. 6a) shows a cloud located between 7 and 9 km, that is, between 247 and 225 K, when the profiles are averaged between 1005 and 1015 UTC. For this case, the lines of sight for both lidars are vertical, so there is no correction necessary. As for the two other cases, the normalization of the 10.6-μm lidar signal leads to four possible profiles, with similar shapes but different values (Fig. 6b). For the four classes of microphysics, the signal is maximal at 8.3 km, that is, 200 m above the altitude where the 532-nm lidar signal is maximal. As for the two previous cases, four profiles of αabs,part,10.6(z) (Fig. 6c) are retrieved. For classes 1 and 2, the 10.6-μm absorption is maximal at 7.6 km, that is, under the 532-nm and 10.6-μm lidar signal maxima. For a cloud composed of class 3 or 4, the curves are off scale, but the most absorbent layer in the infrared is located at an altitude above 8.25 km, which is higher than the maxima of 532-nm lidar signal.

6. Discussion

a. Impact of the assumptions

The assumptions used in the retrieval method can affect the results at some level.

1) The impact of arbitrarily choosing Qsca,part,10.6(b) for normalizing the 10.6-μm signal

To evaluate the impact of the choice of Qsca,part,10.6(b) = 0.6, αabsnpart,10.6(z) was also retrieved for 0.4 and 0.9, which are other possible values of Qsca,part,10.6(z). Figure 7 shows that the choice of this arbitrary value has only a weak consequence on the absolute values of the retrieved parameter, and no consequence concerning its variability. For the three cases, the maxima and minima are exactly at the same altitudes. The absolute values of all retrieved parameters are linked to this arbitrary value. Hence, only the variability of the different parameters can be used and not their absolute values.

2) The impact of Qabs,part,10.6(z) accuracy on αabs,part,10.6(z) retrieval

As shown in section 3a, only the zeroth order of the Taylor series approximation is considered for the Qsca,part,10.6(z) and then Qabs,part,10.6(z) retrievals. It is clear that the accuracy of Qabs,part,10.6(z) is not good, but as shown in section 2c(3), the variation of αabs,part,10.6(z) is governed by the variation of N(z) and not by the variation of Qabs,part,10.6(z). Hence, only a rough estimation of Qabs,part,10.6(z) is sufficient for correctly determining the variability of αabs,part,10.6(z). This is confirmed in Fig. 8, where αabs,part,10.6(z) has been retrieved for three different profiles of Qabs,part,10.6(z) (for 14 October 2002). The first Qabs,part,10.6(z) profile is from the retrieval method, as explained in section 3, for microphysics in class 1; the second one is for the Qabs,part,10.6(z) that is constant within the cloud and equal to 0.7 [the minimal value of Qabs,part,10.6(z) in class 1], with a consistent constant value of Qsca,part,10.6(z) = 0.33; the third profile is similar to the first, but for Qabs,part,10.6(z) = 0.9 [the maximal value of Qabs,part,10.6(z) in class 1] and Qsca,part,10.6(z) = 0.97. The three αabs,part,10.6(z) profiles obtained are very similar, with different values but same variability. Hence, whatever the shape of the Qabs,part,10.6(z) profile, αabs,part,10.6(z) has maxima at the same altitudes.

3) The impact of the first-order Taylor series approximation

The transmission terms T2λ of the lidar equations were examined by computing the full transmission term using the values of Qabs,part,10.6(z), N(z), and αabs,part,10.6(z) deduced from the retrieval method:
i1558-8432-45-4-537-e22
i1558-8432-45-4-537-e23
Figure 9 shows the transmission terms within the cloud for 14 October, and the values of the first-order Taylor Series of these terms. The transmission exceeds 0.99 at both wavelengths with values sufficiently close to 1 to apply a Taylor series for the retrieval method. Furthermore, the difference between the complete term and its Taylor series is so small that it cannot be seen in the figure. It shows that the part of error on the retrieval, which is because of the assumption of the Taylor series, is very weak.

4) Impact of the 10.6-μm lidar signal accuracy

The accuracy of S(z)10.6 is estimated in the appendix. The resulting uncertainty on the 10.6-μm lidar profile for 14 October 2002 is shown in Fig. 10a, and the retrieved αabs,part,10.6(z) is in Fig. 10b. Those figures show that the three profiles of absorption are very close. The part of error resulting from this variability is negligible.

5) Choice of optical properties

Another assumption concerns the choice of optical properties. As mentioned in section 2a, the current results have been obtained in considering 18 different ice crystal models randomly oriented in space. Those models do not pretend to reproduce the natural real variability, which is much more complex and can include particles preferentially oriented in the horizontal plane. Future studies would require other particle shapes, sizes, and orientations in space to be tested; this cannot be done, for instance, because optical properties are not consistently computed at 532 nm and 10.6 μm for others particle types.

b. Discussion of the results

The results of observations and simulations are compared to verify their consistency, and by doing so, check the validity of the proposed retrieval method. Because of the uncertainties in the absolute values discussed earlier, only the variability of the different parameters is compared for simulations and observations. Clouds can be separated into two classes: one with the maximum 532-nm signal located higher in the cloud than the maximum 10.6-μm signal, and the second defined by the contrary situation. These differences can be interpreted as the different partitioning of r within the ice cloud (Fig. 2b).

For clouds in the first class (simulated cloud A and the 8 and 14 October observed cloud cases), Fig. 2b shows that the larger particles seem to be located in the lower layer of the cloud. Moreover, for those cases the simulations and observations consistently show that the altitude of the maximum αabs,part,10.6(z) is located below the altitudes of the 532-nm lidar signal maximum. This difference in altitude is about 100–500 m.

On the contrary, for clouds in the second class (simulated cloud B and the 6 November case when the cloud is composed of columns or droxtals), Fig. 2b shows that the large particles may be located in the upper cloud layers. For those cases, observations and simulations are also consistent in that the maxima in αabsnpart,10.6(z) are located above the altitude of the 532-nm lidar signal maximum.

In situ data are not available to verify the results. However, the retrieved parameters have a realistic order of magnitude in regards to Heymsfield and Platt (1984) who summarized different ice cloud in situ measurements of n(z), and Yang et al. (2001) who computed relevant theoretical crystal properties. The retrieved αabs,part,10.6(z) are inferior to 10−3 m−1. Even if this is not very constraining, this is consistent with the range obtained by Heymsfield and Platt (1984) (4.10−9 and 8.10−5 m−1 for r = 50 μm, and 0.75 × 10−7 and 1.5 × 10−3 m−1 for r = 200 μm).

c. Consequences for passive remote sensing application in the infrared domain

The above results, related to the position of the maximum of attenuation by absorption at 10.6 μm [αabs,part,10.6(z)], may be useful for passive remote sensing techniques that are using infrared wavelengths for studying ice clouds. These methods are often used to retrieve ice crystal effective radius and IWC (Inoue 1985; Ackerman et al. 1990; Minnis et al. 1998; Baum et al. 2000; King et al. 2003), and they require the knowledge of the level of the maximum of absorption within the ice cloud to get precise results. As soon as the ice cloud is semitransparent, this level can hardly be determined from passive remote sensing observations, leading to very large uncertainties on the particle effective radius estimation.

In the framework of the preparation of the future Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) lidar-in-space missions (Winker et al. 2003), a first improvement (Chiriaco et al. 2004) consists of using the cloud base and top as derived from the Nd-Yag lidar, and considering that the most absorbent layer is located somewhere between the two. Unfortunately, the temperature difference between the cloud top and base can reach 10 K or more, and so bias the effective radius retrieval.

The 6 November 2002 case is used to quantify this bias (Table 4). The brightness temperature that would be seen by a passive radiometer if the cloud contained an opaque layer is computed for three different configurations wherein the opaque layer is located 1) at the cloud top, 2) at the altitude of the maximum of 532-nm lidar backscattered signal, and 3) at the altitude of the maximum attenuation by absorption at 10.6 μm. The brightness temperatures TB at 8.65 μm are computed with the Dubuisson et al. (1996) radiative transfer code and give the following successively: 1) TB = 225 K, 2) TB = 237 K, and 3) TB = 244 K.

Those strong differences in brightness temperatures can lead to very different retrievals of the particle effective radius. As an example, in applying the Chiriaco et al. (2004) technique to the radiances observed with the Moderate Resolution Imaging Spectroradiometer (MODIS) on board the satellite Terra, in the far infrared (8.5, 11.15, 12 μm) above the SIRTA, one can retrieve the following values of effective radius: 1) r ranging between 19 and 57 μm, 2) 10 μm < r < 32 μm, and 3) 6 μm < r < 21 μm. The most realistic result of particle effective radius is the third one, when the opaque cloud layer is situated at the altitude of the most absorbent layer within the cloud. The range of other retrievals illustrates the uncertainty on the effective radius when no information on the level of the most absorbent layer is available.

7. Conclusions and perspective

This paper studies the potential for combining 532-nm and 10.6-μm lidar observations to retrieve vertical profiles of four cloud properties: the profile of particle absorption efficiency at 10.6 μm, the profile of particle concentration weighted by the crystal area, the profile of the attenuation by absorption at 10.6 μm, and the ice water content profile. A new method for retrieving those variables was presented and applied to three cases of midlatitude semitransparent cirrus clouds observed at the SIRTA ground base site at Palaiseau, France, during the autumn season.

This study shows that absolute values of the different variables cannot be retrieved with the proposed approach, but their variability within the cloud can be correctly retrieved for three of the considered variables.

Moreover, relationships between the level of maximum lidar signals at 10.6 μm and 532 nm and the level of the maximum of the different variables have been studied, using both observations and simulations. When the large particles are located in the lower part of the cloud (and the small ones below), the lidar signal maxima are located in the upper layer, and the maximum of the infrared absorption is in the lower layers. On the contrary, when the small particles are located in the lower part of the cloud, the lidar maxima are located in the lower layer, and the maximum of the infrared absorption is in the upper layers.

Moreover, the proposed method gives access to the level of maximum attenuation by absorption at 10.6 μm within the ice cloud; in two cases it is located close to the cloud base, and in the third case close to the cloud top. It is never collocated with the maximum of the 532-nm lidar signal, but it is always closer to it than to the cloud base or top. In the future, a statistical study based on a significant amount of cirrus cloud observations may allow highlighting tendencies concerning the position of the maximum of attenuation by absorption within midlatitude cirrus clouds.

The method proposed in this paper is new and shows, for the first time, the high-resolution vertical variability of attenuation by absorption at 10.6 μm [αabs,part,10.6(z)], information that is typically very hard to obtain in ice clouds. As such, it may have significant consequences for ice cloud studies using remote sensing observations in the infrared domain that could be used to calibrate the infrared lidar signal. Even if it gives only the variability of cloud absorption in the infrared and not its absolute value, it could reduce errors on two important ice cloud parameters: the particle effective size retrievals within ice clouds and the radiative upward and downward fluxes emitted by ice clouds. The estimation of the particle size as well as the cloud-emitted fluxes are based on remote sensing techniques that are highly sensitive to the vertical variability of the absorption within the cloud, and particularly the temperature of the cloud layer that absorbs/emits the most radiation. This paper shows that, considering the level of absorption maximum within the cloud, the top of the cloud can induce a 30-K difference on the computed brightness temperature and a 100% error on the retrieval of the effective radius. The dual-lidar method can help find this level, which is hard from passive remote sensing observations, in particular when the cloud is semitransparent.

Based on the results shown in the current paper, future work will consist, in a first step, of comparing the effective radius obtained for semitransparent cirrus clouds with different techniques (Minnis et al. 1998; King et al. 2003; Platnick et al. 2001; Chiriaco et al. 2004) applied to MODIS observations during SIRTA overpasses, deriving the level of maximum absorption within the cloud from dual-lidar observations and comparing it with the derived cloud temperatures. The method itself could be improved by adding a complementary constraint like a very narrow field-of-view infrared radiometer collocated with 10.6-μm lidar observations to constrain the cloud properties retrieval. In a second step, more cases of dual-lidar observations should be analyzed in order to relate the maximum of the absorption to the 532-nm backscatter signal in the framework of the preparation of the future CALIPSO lidar that would fly in formation with several infrared radiometers in space missions.

Acknowledgments

The authors are very grateful to P. Yang for having computed ice crystal optical properties for this specific study. They are also very grateful to the SIRTA ground-based site for providing lidar observations, to P. Dubuisson for the water vapor absorption at 10.6 μm, and to P. Minnis for comments and English language corrections.

REFERENCES

  • Ackerman, S. A., , W. L. Smith, , H. E. Revercomb, , and J. D. Spinhirne, 1990: The 27–28 October 1986 FIRE IFO cirrus case study: Spectral properties of cirrus clouds in the 8–12 μm window. Mon. Wea. Rev, 118 , 23772388.

    • Search Google Scholar
    • Export Citation
  • Banta, R. M., , L. S. Darby, , P. Kaufmann, , D. H. Levinson, , and C. J. Zhu, 1999: Wind-flow patterns in the Grand Canyon as revealed by Doppler lidar. J. Appl. Meteor, 38 , 10691083.

    • Search Google Scholar
    • Export Citation
  • Baum, B. A., , D. P. Kratz, , P. Yang, , S. C. Ou, , Y. Hu, , P. F. Soulen, , and S. C. Tsay, 2000: Remote sensing of cloud properties using MODIS Airborne Simulator imagery during SUCCESS. I. Data and models. J. Geophys. Res, 105 , 1176711780.

    • Search Google Scholar
    • Export Citation
  • Chepfer, H., , J. Pelon, , G. Brogniez, , C. Flamant, , V. Trouillet, , and P. H. Flamant, 1999: Impact of cirrus cloud ice crystal shape and size on multiple scattering effects: Application to spaceborne and airborne backscatter lidar measurements during LITE mission and E LITE campaign. Geophys. Res. Lett, 26 , 22032206.

    • Search Google Scholar
    • Export Citation
  • Chiriaco, M., , H. Chepfer, , V. Noel, , A. Delaval, , M. Haeffelin, , P. Dubuisson, , and P. Yang, 2004: Coupling infrared radiometer and lidar observations for retrieving semi-transparent cirrus clouds particle size. Mon. Wea. Rev, 132 , 16841700.

    • Search Google Scholar
    • Export Citation
  • Collis, R. T. H., , and P. B. Russell, 1976: Lidar measurements of particles and gases by elastic backscattering and differential absorption. Laser Monitoring of the Atmosphere. E. D. Hinkley, Ed. Springer-Verlag, 71–151.

    • Search Google Scholar
    • Export Citation
  • Dabas, A. M., , P. Drobinski, , and P. H. Flamant, 1998: Chirp-induced bias in velocity measurements by a coherent Doppler CO2 lidar. J. Atmos. Oceanic Technol, 15 , 407415.

    • Search Google Scholar
    • Export Citation
  • Dabas, A. M., , P. Drobinski, , and P. H. Flamant, 2000: Velocity biases of adaptive filter estimates in heterodyne Doppler lidar measurements. J. Atmos. Oceanic Technol, 17 , 11891202.

    • Search Google Scholar
    • Export Citation
  • Dabas, A. M., , P. Drobinski, , O. Reitebuch, , E. Richard, , P. Delville, , P. H. Flamant, , and C. Werner, 2003: Multiscale analysis of a straight jet streak using numerical analyses and airborne Doppler lidar measurements. Geophys. Res. Lett, 30 .1049, doi:10.1029/2001GL014040.

    • Search Google Scholar
    • Export Citation
  • Drobinski, P., , R. A. Brown, , P. H. Flamant, , and J. Pelon, 1998: Evidence of organised large eddies by ground-based Doppler lidar, sonic anemometer and sodar. Bound.-Layer Meteor, 88 , 343361.

    • Search Google Scholar
    • Export Citation
  • Drobinski, P., , C. Haeberli, , E. Richard, , M. Lothon, , A. M. Dabas, , P. H. Flamant, , M. Furger, , and R. Steinacker, 2003: Scale interaction processes during MAP-IOP 12 south foehn event in the Rhine valley. Quart. J. Roy. Meteor. Soc, 129 , 729754.

    • Search Google Scholar
    • Export Citation
  • Dubuisson, P., , J. C. Buriez, , and Y. Fouquart, 1996: High spectral resolution solar radiative transfer in absorbing and scattering media: Application to the satellite simulation. J. Quant. Spectrosc. Radiat. Transfer, 55 , 103126.

    • Search Google Scholar
    • Export Citation
  • Elouragini, S., 1995: Etude des nuages par télédétection radar laser (lidar). Ph.D. thesis, Faculté des Sciences de Tunis, 75 pp.

  • Goodman, J. W., 1984: Statistical properties of laser speckles patterns. Laser Speckles and Related Phenomena, J. C. Dainty, Ed., Topics in Applied Physics, Vol. 9, Springer Verlag, 51–58.

    • Search Google Scholar
    • Export Citation
  • Guérit, G., , P. Drobinski, , P. H. Flamant, , and J. P. Cariou, 2002: Effectiveness of simultaneous independent realizations at low carrier-to-noise ratio to improve heterodyne Doppler lidar performance. Part I: Theory and numerical simulations. Appl. Opt, 41 , 22322239.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., 1975: Cirrus uncinus generating cells and the evolution of cirriform clouds. Part I: Aircraft observations of the growth of the ice phase. J. Atmos. Sci, 32 , 799807.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., 1993: Microphysical structure of stratiform and cirrus clouds. Aerosol Cloud Climate Interactions, P. V. Hobbs, Ed., International Geophysics Series, Vol. 54, Academic Press, 97–121.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., , and C. M. R. Platt, 1984: A parameterization of the particle size spectrum of ice clouds in terms of the ambient temperature and the ice water content. J. Atmos. Sci, 41 , 846855.

    • Search Google Scholar
    • Export Citation
  • Inoue, T., 1985: On the temperature and effective emissivity determination of semi-transparent cirrus clouds by bispectral measurements in the 10 microns window region. J. Meteor. Soc. Japan, 63 , 8899.

    • Search Google Scholar
    • Export Citation
  • Intrieri, J. M., , G. L. Stephens, , W. L. Eberhard, , and T. Uttal, 1993: A method for determining cirrus cloud particle sizes using lidar and radar backscatter technique. J. Appl. Meteor, 32 , 10741082.

    • Search Google Scholar
    • Export Citation
  • Intrieri, J. M., , W. L. Eberhard, , T. Uttal, , J. A. Shaw, , J. B. Snider, , Y. Han, , and B. W. Orr, 1995: Multi-wavelength observations of a developing cloud system: The FIRE II 26 November 1991 case study. J. Atmos. Sci, 52 , 40794093.

    • Search Google Scholar
    • Export Citation
  • King, M. D., and Coauthors, 2003: Cloud and aerosol properties, precipitable water, and profiles of temperature and humidity from MODIS. IEEE Trans. Geosci. Remote Sens, 41 , 442458.

    • Search Google Scholar
    • Export Citation
  • Krupp, C., 1991: Holographic measurements of ice crystals in cirrus clouds during the International Cloud Experiment ICE 1989. Report of the Fourth ICE/EUCREX Workshop, Lille, France, Laboratoire d'Optique Atmosphérique, USTL.

  • Liou, K. N., 1986: Influence of cirrus clouds on weather and climate processes: A global perspective. Mon. Wea. Rev, 114 , 11671199.

  • Lynch, D. K., , K. Sassen, , D. O'C. Starr, , and G. Stephens, 2002: Cirrus. Oxford University Press, 504 pp.

  • McClatchey, R. A., , R. W. Fenn, , J. E. A. Selby, , F. E. Volz, , and J. S. Garing, 1972: Optical properties of the atmosphere. 3d ed. Air Force Cambridge Research Laboratory, Hanscom Air Force Base Publication AFCRL-72-0497, 108 pp.

  • Miloshevich, L. M., , and A. J. Heymsfield, 1996: A balloon-borne cloud particle replicator for measuring vertical profiles of cloud microphysics: Instrument design and performance. Proc. 12th Int. Conf. on Clouds and Precipitation, Zurich, Switzerland.

  • Minnis, P., , D. P. Garber, , and D. F. Young, 1998: Parameterizations of reflectance and effective emittance for satellite remote sensing of cloud properties. J. Atmos. Sci, 55 , 33133339.

    • Search Google Scholar
    • Export Citation
  • Mishchenko, M. I., 1991: Light scattering by randomly oriented axially symmetric particles. J. Opt. Soc. Amer, 79 , 871882.

  • Noel, V., , H. Chepfer, , G. Ledanois, , A. Delaval, , and P. H. Flamant, 2002: Classification of effective shape ratios in cirrus clouds based on lidar depolarization ratio. Appl. Opt, 41 , 42454257.

    • Search Google Scholar
    • Export Citation
  • Platnick, S., , M. D. King, , H. Gerber, , and P. V. Hobbs, 2001: A solar reflectance technique for cloud retrievals over snow and ice surfaces. J. Geophys. Res, 106 , 1518515199.

    • Search Google Scholar
    • Export Citation
  • Platt, C. M. R., 1973: Lidar and radiometric observations of cirrus clouds. J. Atmos. Sci, 30 , 11911204.

  • Platt, C. M. R., , S. A. Young, , P. J. Manson, , G. R. Patterson, , S. C. Marsden, , and T. Austin, 1998: The optical properties of equatorial cirrus from observations in the ARM pilot Radiation Observation Experiment. J. Atmos. Sci, 55 , 19771996.

    • Search Google Scholar
    • Export Citation
  • Post, M. J., , and W. D. Neff, 1986: Doppler lidar measurements of winds in a narrow mountain valley. Bull. Amer. Meteor. Soc, 67 , 274281.

    • Search Google Scholar
    • Export Citation
  • Sassen, K., 1991: The polarization lidar technique for cloud research: A review and current assessment. Bull. Amer. Meteor. Soc, 72 , 18481866.

    • Search Google Scholar
    • Export Citation
  • Sassen, K., , and K. N. Liou, 1979: Scattering of polarized laser light by water droplet, mixed-phase and ice crystal clouds. Part II: Angular depolarizing and multiple-scattering behavior. J. Atmos. Sci, 36 , 852861.

    • Search Google Scholar
    • Export Citation
  • Spinhirne, J. D., 1982: Lidar clear atmosphere multiple scattering dependence on receiver range. Appl. Opt, 21 , 24672468.

  • Stephens, G. L., , S. C. Tsay, , P. W. Stackhouse, , and P. J. Flatau, 1990: The relevance of the microphysical and radiative properties of cirrus clouds to climate and climate feedback. J. Atmos. Sci, 47 , 17421753.

    • Search Google Scholar
    • Export Citation
  • Stubenrauch, C. J., , W. B. Rossow, , F. Cheruy, , A. Chedin, , and N. A. Scott, 1999: Cloud as seen by satellite sounders (3I) and imagers (ISCCP). Part I: Evaluation of cloud parameters. J. Climate, 12 , 21892213.

    • Search Google Scholar
    • Export Citation
  • Ulaby, F. T., , R. K. Moore, , and A. K. Fung, 1986: Microwave Remote Sensing: Active and Passive. Vol. 1, Microwave Remote Sensing Fundamentals and Radiometry, Artech House, 456 pp.

    • Search Google Scholar
    • Export Citation
  • Uttal, T., , E. E. Clothiaux, , T. P. Ackerman, , J. M. Intrieri, , and W. L. Eberhard, 1995: Cloud boundary statistics during FIRE II. J. Atmos. Sci, 52 , 42764284.

    • Search Google Scholar
    • Export Citation
  • Winker, D. M., , J. Pelon, , and M. P. McCormick, 2003: The CALIPSO mission: Spaceborne lidar for observation of aerosols and clouds. Lidar Remote Sensing for Industry and Environmental Monitoring III, U. N. Singh, T. Itabe, and Z. Liu, Eds., International Society for Optical Engineering, (SPIE Proceedings Vol. 4893), 1–11.

  • Yang, P., , B. Gao, , B. A. Baum, , Y. X. Hu, , W. J. Wiscombe, , S. Tsay, , and D. M. Winker, 2001: Radiative properties of cirrus couds in the infrared (8-13μm) spectral region. J. Quant. Spectrosc. Radiat. Transfer, 70 , 473504.

    • Search Google Scholar
    • Export Citation

APPENDIX

Accuracy of S(z)10.6 Estimate

The vertical variability of the 10.6-μm signal power S(z)10.6_N, averaged over a range gate Δz = NcTS (where N is the total number of samples within the range gate and TS is the sampling period) is related to the 10.6-μm signal statistical properties, driven by the number of independent samples M in the range gate. The quantity M is defined as (Goodman 1984):
i1558-8432-45-4-537-ea1
This quantity is the statistical mean of the average lidar signal power S(z)10.6_N, and the variance of S(z)10.6_N; M1/2 measures the relative amplitude of the statistical fluctuations of the signal. This parameter has to be computed in order to estimate the statistical uncertainty on the lidar signal.
Here M1/2 includes the random fluctuations of the return signal resulting from the speckle effect and the random fluctuations of the noise. It can be shown that M is given by (Guérit et al. 2002)
i1558-8432-45-4-537-ea2
where SNR is the signal-to-noise ratio, also called in the lidar community carrier-to-noise ratio (CNR); M is the number of independent time speckles in the atmospheric return. Guérit et al. (2002) have shown that M can be approximated by
i1558-8432-45-4-537-ea3
where wS is the spectral width of the signal. In the nacc case signals are accumulated, then MnaccM (Dabas et al. 1998).

In the present experiment, N = 300, nacc = 10, TS = 10−2 μs, and wS has been estimated to be about 0.3 MHz (Dabas et al. 2000). This leads to Mnacc × 3.4 = 34. The evaluation of M to evaluate the variability of S(z)10.6_N requires the estimation of the SNR, which assumes that the noise power can be measured accurately a priori. It is not an issue for this 10.6-μm lidar because the bulk of the noise is the detection shot noise controlled by the local oscillator, which is constant with time.

In the cirrus cloud, SNR → ∞ so MM ≈ 34, that is, σS10.6_N = /5.8, whereas outside the cirrus, the aerosol content is very low so that SNR → 0 so MN = nacc × 300 = 3000, that is, σS10.6_N = /54.8.

We neglect the shot-to-shot laser pulse energy fluctuations: at high SNR the variability is dominated by speckle effect, and at low SNR the variability is dominated by the local oscillator shot noise, which is constant with time.

Fig. 1.
Fig. 1.

From the Yang et al. (2001) calculation: (a) Qabs,part,10.6 = f (Qsca,part,10.6) for crystals in class 1, where the line is the linear fit between Qabs,part,10.6 and Qsca,part,10.6; (b) same as (a), but for class 2; (c) same as (a), but for class 3; (d) same as (a), but for class 4; (e) k10.6/k532 = f (r) for crystals in class 1, where the horizontal line is the constant value of γ that is taken into account in the retrieval method; (f) same as (e), but for class 2; (g) same as (e), but for class 3; (h) same as (e), but for class 4.

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 2.
Fig. 2.

Theoretical profiles of (a) IWC, (b) ice particle effective radius for two microphysics, (c) absorption efficiency at 10.6 μm, (d) concentration n, (e) N, (f) absorption coefficient at 10.6 μm, (g) lidar profiles at 10.6 μm, and (h) lidar profiles at 532 nm.

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 3.
Fig. 3.

Evolution of the logarithm of lidar backscattered signal as a function of time for the 6 Nov 2002 (a) 532-nm and (b) 10.6-μm backscattered signal.

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 4.
Fig. 4.

For 8 Oct 2002 with values averaged between 1210 and 1220 UTC: (a) 532-nm lidar profile, (b) 10.6-μm lidar profile for the four classes of microphysics, and (c) particle absorption coefficient profile at 10.6 μm αabs,part,10.6(z) for the four classes of microphysics.

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 5.
Fig. 5.

Same as Fig. 4, but for 14 Oct 2002 averaged between 1210 and 1220 UTC.

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 6.
Fig. 6.

Same as Fig. 4, but for 6 Nov 2002 averaged between 1005 and 1015 UTC.

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 7.
Fig. 7.

For 14 Oct 2002, αabs,part,10.6(z) retrieved using microphysics in class 3 and for values of Qsca,part,10.6(zbase) = 0.4 (solid line), 0.6 (dashed line), and 0.9 (dotted line).

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 8.
Fig. 8.

For 14 Oct 2002, αabs,part,10.6(z) retrieved using three different profiles of Qabs,part,10.6(z): profile retrieved using the described method for microphysics in class 1 (solid line), profile is constant in the cloud and equal to 0.7 (dashed line), and profile is constant in the cloud and equal to 0.9 (dotted line).

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 9.
Fig. 9.

Calculation of the atmospheric transmission at 532 nm (solid line), of its first first-order Taylor approximation (dotted–dashed lines), of the atmospheric transmission at 10.6 μm (dashed lines), and of its first first-order Taylor approximation (dotted line) for 14 Oct 2002, from the obtained results using microphysics in class 3.

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Fig. 10.
Fig. 10.

For the 14 Oct 2002 case using microphysics in class 3: (a) 10.6-μm lidar signal (solid line) and its variability (dashed line), and (b) αabs,part,10.6(z) profile (solid line) and its variability calculated from lidar variability (dashed line).

Citation: Journal of Applied Meteorology and Climatology 45, 4; 10.1175/JAM2355.1

Table 1.

Description of the particle models used in the study to describe the cirrus clouds.

Table 1.
Table 2.

Description of three ice cloud cases observed at SIRTA.

Table 2.
Table 3.

Altitudes at which the different variables are at a maximum, for all cases.

Table 3.
Table 4.

Simulation of brightness temperatures for different cloud altitudes in a radiative transfer code, in comparison with measured brightness temperature from MODIS data. Impact on the particle radius obtained with the split-window technique is shown.

Table 4.
Save