12.07.2015 Views

VLIDORT User's Guide

VLIDORT User's Guide

VLIDORT User's Guide

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

m⎛ P⎞l( μ)0 0 0⎜⎟mmm ⎜ 0 Rl( μ)− Tl( μ)0 ⎟Pl( μ)= ⎜mm⎟ . (2.25)⎜0 − Tl( μ)Rl( μ)0⎟m⎝ 0 0 0 Pl( μ)⎠The “Greek matrices” Β l for 0 ≤ l ≤ LM contain the sets of expansion coefficients that define themmscattering law. The Pl(μ)matrices contain entries of normalized Legendre functions Pl(μ)andfunctions R ml(μ)and mlTl(μ)which are related to Pmn(μ)(for details, see for example [Siewert,2000b]).2.1.3. Boundary conditionsDiscrete ordinate RT is pure scattering theory: in a multilayer medium, it is only necessary tospecify the layer total optical thickness values Δ n , the layer total single scatter albedo ω n , and thelayer 4 x 4 matrices Β nl of expansion coefficients (l being the moment number) for the totalscattering. To complete the calculation of the radiation field in a stratified multilayer medium,we have the following boundary conditions:(Ι) No diffuse downwelling radiation at TOA. Thus for the first layer we have:+In( 0, μ,φ)= 0 . (n = 1) (2.26)(ΙΙ) Continuity of the upwelling and downwelling radiation fields at intermediateboundaries. If N TOTAL is the number of layers in the medium, then:±( ) ±I n − 1Δ n −1 = I n(0) . (n = 2,…N TOTAL ) (2.27)(ΙΙΙ) A surface reflection condition relating the upwelling and downwelling radiation fieldsat the bottom of the atmosphere:−+I ( , μ , φ)= ( μ,φ;μ′, φ′) ( , μ′, φ′nΔnR InΔn) . (n = N TOTAL ) (2.28)Here, reflection matrix R relates incident and reflected directions.The convention adopted here is to use a “+” suffix for downwelling solutions, and a “−” suffixfor upwelling radiation. Conditions (I) and (II) are obeyed by all Fourier components in theazimuthal series. For condition (III), it is necessary to construct a Fourier decomposition of theBRDF operator R to separate the azimuth dependence; we return to this issue in section 2.5.4.The Lambertian case (isotropic reflectance) only applies for Fourier component m = 0 and Eq.(2.28) then becomes:1⎡⎤−I = ⎢ ( − ) + ∫ + ′ ′ ′n( Δn,μ)2δm, 0R0E1μ0I0Tn−1exp λnΔnIn( Δn,μ ) μ dμ⎥ . (2.29)⎣0⎦Here, R 0 is the Lambertian albedo, E 1 = diag{1,0,0,0}, and T n− exp( − λnΔn)1 is the wholeatmosphereslant path optical depth for the solar beam.18

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!