Highlight

Algorithm

Radiative transfer solver in MCARaTS package uses Monte Carlo photon-tracing method. Trajectories are simulated based on physics with taking into accout for cloud and aerosol particles, hydrometeors, gases, and underlying surface.

To solve 3-D radiative transfer, one of four schemes can be used:

  • Fully-3-D radiative transfer (F3D);

  • Partially-3-D radiative transfer (P3D);

  • Independent pixel (column) approximation (IPA or ICA);

Input

Visualization of the cloudy atmosphere

MCARaTS v0.10 has a visualization mode, in which approximate, pseudo-radiances are calculated using a volume rendering technique. Here is a sunset scene over ocean. There are thin clouds. Typical aerosol profile was assumed.

Scattering and absorbing properties of atmospheric medium can be arbitrily configured by users. Three-dimensional distribution of extinction coefficients, single scattering albedo, scattering phase function, and gaseous absorption coefficients is defined by the user. The following figure shows an example of artificial cloud realization generated with a fractal model. Underlying surface is assumed macroscopically flat, but its properties can vary two-dimensionaly. The surface is modeled by a mixed model of diffuse and specular reflectors, and their fractions are variable. The diffuse reflection is modeled by a modified Lommel-Seeliger model, and the specular reflection is modeled by Fresnell reflection by ocean-like rough micro-facet model. Six variables defines the surface properties and each of them can vary spatialy.

Radiation sources can be solar incidence from top and/or thermal emission, or localized artificial source (e.g., laser beam, isotropically-emitting lamp, etc).

Here are other examples, looking down a stratocumulus cloud deck.

Finally, here is a movie (19 MB) of flight simulation.

Flux computation example

Area-averaged radiative fluxes for a case of stratocumulus clouds over surface reflectance distribution as a Japanese charater that means "surface".

High-resolution radiance example

The following figures show high-resolution area-averaged nadir radiance with explicit 3D radiative transfer and ICA: 256x256 pixels, 16 km domain width. The cloud input data were artificialy generated with a modified Gaussian model, with inhomogeneity of 15.625 m width and about 20 m height. The computation was made by using a parallel computer (Compaq Alpha 833 MHz 64 CPUs). The accuracy of pixel-by-pixel radiance is about 1% (relative unit). These computations were achieved by our improved computational power and improved argorithms.

Wavelength = 0.66 µm, solar zenith angle = 20 deg.

Wavelength = 0.66 µm, solar zenith angle = 60 deg.

Wavelength = 2.13 µm, solar zenith angle = 20 deg.

Wavelength = 2.13 µm, solar zenith angle = 60 deg.

The differece of 3D radiance from IPA comes from net horizontal radiative convergence/divergence, which cause smoothing and roughening phenomena, as illustrated in the above figures. The smoothing and roughening can be also found in amplitude spectra of 3D and IPA radiances, as in the foolowings.

Solar zenith angle = 20 deg.

Solar zenith angle = 60 deg.