Optical Characterization Of Complex Aerosol And Cloud .

1m ago
9 Views
0 Downloads
1.38 MB
137 Pages
Transcription

Optical Characterization of Complex Aerosol and Cloud Particles:Remote Sensing and Climatological ImplicationsLI LIUSubmitted in partial fulfillment of therequirements for the degreeof Doctor of Philosophyin the Graduate School of Arts and SciencesCOLUMBIA UNIVERSITY2004

2004Li LiuAll Rights Reserved

ABSTRACTOptical Characterization of Complex Aerosol and Cloud Particles:Remote Sensing and Climatological ImplicationsLi LiuOptical characterization of aerosol and cloud particles has been a challenge toresearchers involved in a wide range of disciplines including remote sensing and climatestudies. This thesis addresses several important atmospheric radiation problems involvingcloud and aerosol particles with complex structure. We solve these problems by (i)extensively using state-of-the-art theoretical techniques to compute radiative properties ofnonspherical and composite atmospheric particulates; (ii) combining theoretical and highquality laboratory data for single scattering by irregular dust-like aerosols; (iii) applyingadvanced retrieval algorithms to analyze satellite observations of tropospheric aerosols;and (iv) validating satellite retrievals with high-quality ground-based data.In Chapter 1, the superposition T-matrix method is used to compute electromagneticscattering by semi-external aerosol mixtures in the form of polydisperse, randomlyoriented two-particle clusters with touching components. The results are compared withthose for composition-equivalent external aerosol mixtures. It is concluded thataggregation had a relatively weak effect on radiative properties of composite aerosols.In Chapter 2, scattering and absorption characteristics of water cloud dropletscontaining black carbon (BC) inclusions are calculated in the visible spectral range by acombination of ray-tracing and Monte Carlo techniques.In addition, Lorenz-Miecalculations are performed assuming that the same amount of BC particles are mixedwith water droplets externally. It is concluded that under normal conditions the effect ofBC inclusions on the radiative properties of cloud droplets is weak.In Chapters 3 and 4, we compare and combine the results of laboratorymeasurements of the Stokes scattering matrix for nonspherical quartz aerosols at a visiblewavelength in the scattering angle range 5 –173 and the results of Lorenz-Mie

computations for projected-area-equivalent spheres with the refractive index of quartz. Asynthetic normalized phase function is constructed and then used to analyze the potentialeffect of particle nonsphericity on the results of retrievals of mineral troposphericaerosols based on radiance observations from Advanced Very High ResolutionRadiometer (AVHRR).Chapter 5 presents the validation results of the aerosol optical thickness retrievedfrom AVHRR channel 1 and 2 radiances. The satellite retrieved optical thickness iscompared with the accumulated historical ship-borne sun-photometer measurements.Comparisons of single-scattering albedo and Ångström exponent values retrieved fromthe AVHRR data and those measured in situ at Sable Island indicate that the currentlyadopted value 0.003 can be a reasonable choice for the imaginary part of the aerosolrefractive index in global satellite retrievals.In chapter 6, we analyze existing lidar observations of polar stratospheric clouds(PSCs) and derive several constraints on PSC particle miscrophysical properties based onextensive T-matrix computations of light scattering by polydispersions of randomlyoriented, rotationally symmetric nonspherical particles.

TABLE OF CONTENTSLIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivLIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .viiACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiINTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1Chapter 1 Scattering and radiative properties of semi-external versusexternal mixtures of different aerosol types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.1Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16161.2 Computations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21Chapter 2 The effect of black carbon on scattering and absorption ofsolar radiation by cloud droplets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29302.1Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.2Optical properties of black carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.3 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322.4Ray-tracing/Monte Carlo model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332.5Model computations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332.5.1 Internal mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.5.234External mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.5.3 Numerical results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .i40

Chapter 3 Scattering matrix of quartz aerosols: comparison andsynthesis of laboratory and Lorenz-Mie results . . . . . . . . . . . . . . . . . . . . . . . . . . . 463.1Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.2Measurements and Lorenz-Mie computations . . . . . . . . . . . . . . . . . . . . 49473.3 Synthetic phase function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .523.4 Discussion and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55Chapter 4 Investigation of the effects of particle nonsphericity on aerosolretrievals from AVHRR observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634.1Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2Effect of particle shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6370Chapter 5 Global validation of the operational two-channel AVHRRretrieval product over the ocean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5.177Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .785.2 Validation of satellite aerosol optical thickness retrievals. . . . . . . . . . .805.35.2.1Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805.2.2Ship data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 815.2.3Primary validation results . . . . . . . . . . . . . . . . . . . . . . . . . . . .81Validation of the aerosol single-scattering albedo . . . . . . . . . . . . . . . . . 895.4 Discussion and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96Chapter 6 Constraints on PSC particle microphysics derived from lidarobservations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1036.1Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104ii

6.2 T-matrix computations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1056.3Observation data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1086.4 Analysis results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1106.5 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118iii

LIST OF FIGURES1.1(a) External, (b) semi-external, and (c) internal particle mixtures . . . . . . . . . . . 171.2Phase function versus scattering angle for dust-sulfate semi-external (solidcurves) and external (dotted curves) mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . 221.3As in Fig. 1.2, but for sulfate-soot mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231.4As in Fig. 1.2, but for scattering-matrix element ratios . . . . . . . . . . . . . . . . . . . 251.5As in Fig. 1.3, but for scattering-matrix element ratios . . . . . . . . . . . . . . . . . . . 262.1External (a) and internal (b) mixing of large cloud droplets and smalleraerosol particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.232Single-scattering co-albedo 1 ϖ and asymmetry parameter g for mixturesof cloud droplets and BC particles versus BC particle effective radius at awavelength of 0.55 µm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362.3Single-scattering co-albedo 1 ϖ and asymmetry parameter g for mixturesof cloud droplets and BC particles versus BC mass fraction . . . . . . . . . . . . . . . 382.4Relative differences (in %) between the single-scattering co-albedo andasymmetry parameter for external and internal mixtures of cloud droplets andBC particles versus the BC mass fraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.139Normalized distribution of the average area of the particle projection forrandomly oriented quartz aerosols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493.2Laboratory data for nonspherical quartz aerosols and results of Lorenz-Miecomputations for projected-area-equivalent quartz spheres . . . . . . . . . . . . . . . . 503.3The pattern of the differences between the Lorenz-Mie phase function forspherical quartz particles (solid curve) and the phase function fornonspherical quartz aerosols depends on the vertical position of theexperimental a 1 (Θ) profile (dashed, dotted, and dot-dashed curves) . . . . . . . .iv53

3.4Synthetic and Lorenz-Mie phase functions for nonspherical quartz aerosolsand projected-area-equivalent quartz spheres, respectively . . . . . . . . . . . . . . .564.1 Synthetic and Lorenz-Mie phase functions for nonspherical quartz aerosolsand projected-area-equivalent quartz spheres, respectively, used in the onechannel retrieval algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .674.2 Monthly averages of the ratio τN/τS and the respective scattering angle versuslongitude . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .695.1 AVHRR retrieved aerosol optical thickness τSAT versus ship measurementsτSP at λ 0.55 µm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 855.2 τSAT versus τSP for different AVHRR instrument (NOAA-7, 9, 11, 14) . . . . . .875.3 Comparison of τSAT and τSP at λ 0.55 µm for three increasing values ofdiffuse component of surface reflection S 0.002, 0.004, 0.005 and thecorresponding linear regression lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 885.4 Aerosol single-scattering albedo versus Ångström exponent for Re(m) 1.5and four increasing values of Im(m) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .915.5 Monthly averages of the Ångström exponent and single-scattering albedo forJuly 1999 derived from two-channel AVHRR data assuming a fixed aerosolrefractive index m 1.5 0.003i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 925.6 The annual cycle of the aerosol single-scattering albedo measured in situ atSable Island . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .935.7 The annual cycle of the aerosol optical thickness retrieved from channel–1and –2 AVHRR data over Sable Island during the period November 1994–December 1999 assuming that the imaginary part of the aerosol refractiveindex is fixed at 0.001, 0.002, 0.003, and 0.005 . . . . . . . . . . . . . . . . . . . . . . . .945.8 As in Fig. 5.7, but for the constrained Ångström exponent. . . . . . . . . . . . . . . .956.1 Linear depolarization ratio δ(603 nm) , backscatter color index α , andv

depolarization color index β versus effective equal-volume-sphere radiusreff for polydisperse, randomly oriented spheroids with a refractive index ofm 1.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1116.2 As in Fig. 6.1, but for polydisperse, randomly oriented cylinders . . . . . . . . . . . 1126.3As in Fig. 6.1, but for the refractive index m 1.308 typical of water ice atvisible wavelengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1136.4 As in Fig. 6.3, but for polydisperse, randomly oriented circular cylinders . . . . 1146.5 The bars depict the respective ranges of the effective radius that reproduce thevalues of δ , α , and β observed for type Ia PSCs, as shown in Figs. 6.1 and6.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1156.6The bars depict the respective ranges of the effective radius that reproduce thevalues of δ and α observed for type Ib PSCs, as shown in Figs. 6.1 and 6.2 . 1166.7 The bars depict the respective ranges of the effective radius that reproduce thevalues of δ , α , and β observed for type II PSCs, as shown in Figs. 6.3 and6.4 .