Engineering topics
Publications and source records attributed to Kattawar, G. W..
Seven models for computing underwater radiances and irradianes by numberical solution of the radiative transfer equation are compared. The models are applied to the solution of several problems drawn from optical oceanography.
Some 40 publications that appeared in scientific journals from 1973 to 1981 as well as 45 scientific reports issued during the grant period are listed by title. Topics cover the development of a matrix operator theory of radiative transfer which made possible the exact model calculations of the radiance as a function of height in planetary atmospheres; calculation of the Mie phase matrix for various types of particles as well as for radiance and polarization in planetary atmospheres; analysis of high dispersion spectroscopic observations of Venus; calculation of curves of growth for Venus; the development of a theory for calculating radiative transfer in spherical shell atmospheres; investigations of zonal winds on Venus; and examination of Rayleigh scattering.
The contribution of inelastic molecular scattering (Rayleigh-Brillouin and rotational Raman scattering) to the filling-in of Fraunhofer lines in the light of the blue sky is studied. Aerosol fluorescence is shown to be negligible, and aerosol scattering is ignored. The angular and polarization dependences of the filling-in detail for single scattering are discussed. An approximate treatment of multiple scattering, using a backward Monte Carlo technique, makes it possible to investigate the effects of the ground albedo. As the molecular scatterings alone produce more line-filling than is observed, it seems likely that aerosols dilute the effect by contributing unaltered sunlight to the observed spectra.
The multipole expansion obtained by Morita et al. (1968) of the Gaussian laser beam used to levitate an aerosol particle in order that its complete phase matrix may be measured is compared with that of Tsai and Pogorzelski (1975) in order to demonstrate the effect of the incorrect expansion used by Morita. Errors incurred by the use of an equation in which one side satisfies the scalar wave equation while the other side does not and can be reduced to a plane wave amplitude are calculated as functions of the inverse of the wave number times the beam waist, the wave number times the radial spherical coordinate and the angular spherical coordinate. Errors on the order of a few percent, considered undetectable are obtained in the squared-field amplitudes due to the expansion, however, they are found to become significant (several tens of percent) when the angle is zero. It is concluded that the expansion of Morita should only be used in the regions where the spherical angle is less than 0.01 and its product with the wave number and the radial spherical coordinate is less than unity.
In an attempt to gain some insight into the problem of how close particles have to be on the average before phase effects become important, the cross section and phase matrices for two identical particles as a function of the distance between their centers for both broadside and end-on illumination have been calculated. The point dipole approximation is employed to build two pseudospherical particles each consisting of 32 dipoles with an effective size parameter of 0.9283 and refractive index of 1.54-0.0i. In regard to cross sections, it is found that in certain cases several percent deviation from the converged value, which is twice the single particle cross section, could be found up to distances of separation of 25 single particle diameters. The maxima and minima in the cross section curves can be explained in terms of simple interference effects. Also the element P11 of the phase matrix showed a great deal of structure, and the position of maxima and minima could be explained by phase shift analysis.
A method for partitioning the radiance emerging from a planetary atmosphere in proportion to the average number of scatterings in each atmospheric layer is applied to Venus in order to account for the various atmospheric levels which contribute to the radiance. A backward Monte Carlo technique is used to calculate core and continuum radiances for the P(16) line of the 8689-A CO2 band for various phase angles and colatitudes in a realistic model of the Venus atmosphere. Effective temperatures for the radiance are defined in two ways, both of which are shown to vary little with phase angles up to 120 deg, and fall rapidly beyond this point, in accordance with spectroscopic results. Temperature increases on the order of 7 K and 10 K are found going from the core of the line to the continuum for colatitudes of 90 and 20 deg, respectively, and the effective temperature is calculated to decrease from equator to pole. An effective optical depth is also defined, which, together with the concepts of effective temperature and effective height, is considered to bring a more fundamental understanding of the concept of level of line formation.
This paper investigates the effects of sphericity on the radiation reflected from a planet with a homogeneous conservative-scattering atmosphere of optical thicknesses of 0.25 and 1.0. A Henyey-Greenstein phase function with asymmetry factors of 0.5 and 0.7 was considered. Significant differences were found when these results were compared with the plane-parallel calculations. Also, large violations of the reciprocity theorem, which is only true for plane-parallel calculations, were noted. Results are presented for the radiance versus height distributions as a function of planetary phase angle. These results will be useful to researchers in the field of remote sensing and planetary spectroscopy.
The plane-parallel approximation and the more realistic spherical shell approximation for the radiance reflected from a planetary atmosphere are compared and are applied to the study of a planet the size of the earth with a homogeneous conservative Rayleigh scattering atmosphere extending to a height of 100 km. Inadequacies of the approximations are considered. Radiance versus height distributions for both single and multiple scattering are presented, as are results for the fractional radiance from altitudes in the atmosphere which contribute to the total unidirectional reflected radiance at the top of the atmosphere. The data can be used for remote sensing applications and planetary spectroscopy.
Contrary to published reports, the limb-darkening observed by Mariner 10 is consistent with the sulfuric-acid aerosol scattering that explains the polarization and spectroscopic features of Venus. A combination of both geometric and photometric systematic errors explains the previously reported discrepancy. Although the aerosol model cannot be distinguished from isotropic scattering at the Mariner 10 phase angle, Venera 9 data at larger phases clearly favor the aerosol model. There is no evidence for isotropic scattering in the clouds of Venus.
The effects are investigated of sphericity on the radiation reflected from a planet with a homogeneous, conservative scattering atmosphere of optical thicknesses of 0.25 and 1.0. A Henyey-Greenstein phase function with asymmetry factors of 0.5 and 0.7 is considered. Significant differences were found when these results were compared with the plane-parallel calculations. Also large violations of the reciprocity theorem, which is only true for plane-parallel calculations, were noted. Results are presented for the radiance versus height distributions as a function of planetary phase angle.
Spectral line profiles, curves of growth, and curves for the equivalent width of a line as a function of Venus phase angle have been computed for a Rayleigh scattering cloud and compared with those for a cloud of isotropic scatterers. The results are very similar for the two kinds of scattering, with the exception of the curves of equivalent width as a function of Venus phase angle. These latter curves exhibit the 'inverse phase effect' and rule out a scale height of the clouds much less than half the scale height of the gas. The optical depth of the clouds is approximately 100.
A study was made to see how much information could be extracted from the Mariner 9 Mars isophotes taken at phase angle roughly 60 deg. Minnaert functions and both isotropic and Rayleigh scattering could easily be ruled out, and it was essential to use forward-peaked phase functions computed from Mie theory. Isophotes similar to those observed are plotted assuming a semi-infinite dust cloud with a considerable variation in particle properties and size distribution, so long as the ratio of the multiply to singly scattered light is held within certain limits. These conditions are met by micron-sized, moderately absorbing mineral grains whose mean size should not be much larger than a micron. It was also found that a dust cloud of finite optical thickness bounded from below by a Lambert ground would fit the isophote data.
Line profiles and curves of growth were computed for both reflected and transmitted radiation for typical lines in CO2 bands (in the photographic infrared) which occur in the spectrum of Venus. In the model, the pressure variation with altitude was considered, and the base of the cloud deck was set at the 2 bar level. The temperature was held constant at 250 K, and a Voigt profile was used for the line shape. It was also assumed that the scale height of the cloud particles was equal to the scale height of the gas. The calculations were made for four values of the scattering optical thickness (0.1, 1.0, and 100) using a continuum single-scattering albedo of 0.9975 (which gives a Bond albedo of 0.896 for a scattering optical thickness of 100, the value observed for Venus at these wavelengths). Curves of growth are also presented for reflected radiation which has been averaged over the visible disk for three values of the Venus phase angle (0, 86, and 166 deg).
A matrix operator method for a model of the real atmosphere is used to calculate the complete radiation field of the atmosphere. The radiance, direction and amount of polarization, and ellipticity are obtained at the top and bottom of the atmosphere for three values of the surface albedo (0, 0.15, 0.90) and five solar zenith angles, taking account of molecular absorption and scattering (including by ozone) together with the variation of the density of these molecules with height.
The volume averaged cross sections for extinction, scattering, absorption, and radiation pressure along with the elements of the phase matrix were computed for a spherical polydispersion of water coated carbon particles. The wavelength used was 4910 A. The thickness of the water coating was held constant over the polydispersion, while the ratio of core radius to total radius was varied, or the ratio was held constant over the polydispersion, while the thickness of the coating varied. The results show large differences in both the intensity and degree of polarization between dry and wet aerosols. Also interference phenomena can occur, which demonstrate that the assignment of an average refractive index to aerosols can be misleading and can lead to erroneous results.
The ellipticity and the direction of polarization are calculated for radiation that has undergone multiple scattering from plane-parallel layers. Both the radiation emerging from the top of the layer and that transmitted through the bottom are considered. Two different phase functions are used for the scattering layer: Rayleigh and haze L. The direction of polarization of the reflected radiation shows little variation as the optical depth of the layer increases, but there is a much larger variation for the transmitted radiation. When the optical thickness is small, the direction of polarization for haze L varies rapidly with zenith angle near those angles at which the single-scattered polarization is zero. The ellipticity of the radiation from haze L layers increases at first in direct proportion to the optical thickness of the layer. In general, the ellipticity of the transmitted radiation is considerably greater than that of the reflected due to the greater average number of photon collisions in the former case. When the ellipticity is small, it is shown that the product of the polarization and the ellipticity is closely equal to -V/2I, where I and V are the first and fourth components of the Stokes vector.
Spectral line profiles, curves of growth, and curves for the equivalent width of a line as a function of Venus phase angle are computed for a Rayleigh scattering cloud and compared with those for a cloud of isotropic scatterers. The results are similar for the two kinds of scattering, with the exception for the curves of equivalent width as a function of Venus phase angle. These latter curves exhibit the inverse phase effect and rule out the possibility that the scale height of the clouds can be much less than half the scale height of the gas.
The matrix operator method was used to calculate the polarization of radiation scattered on layers of various optical thicknesses, with results compared for Rayleigh scattering and for scattering from a continental haze. In both cases, there are neutral points arising from the zeros of the polarization of single scattered photons at scattering angles of zero and 180 degrees. The angular position of these Rayleigh-like neutral points (RNP) in the sky shows appreciable variation with the optical thickness of the scattering layer for a Rayleigh phase matrix, but only a small variation for haze L phase matrix. Another type of neutral point exists for non-Rayleigh phase functions that is associated with the zeros of the polarization for single scattering which occurs between the end points of the curve. A comparison of radiances calculated from the complete theory of radiative transfer using Stokes vectors with those obtained from the scalar theory shows that differences of the order of 23% may be obtained for Rayleigh scattering, while the largest difference found for a haze L phase function was of the order of 0.1%.