Phase: RAMI4PILPS
The AddingS model (AddingS) of Kallel et al. [2008] combines the formalism of the SAIL canopy reflectance model [Verhoef, 1984] and the “adding” method [van de Hulst, 1957].
Given a thick vegetation layer, AddingS divides it into thin sublayers of LAI = 0.01.
Then, the adding operators (reflectance and transmittance) of a thin sublayer are estimated based on the SAIL model’s bidirectional reflectance distribution function (BRDF) and bidirectional transmission distribution function (BTDF).
Although SAIL underestimates the reflectance due to its semi-isotropic diffuse flux assumption, the estimation is accurate since for a thin layer the contribution of the diffuse fluxes to the BRDF and BTDF is very small.
Indeed, BRDF and BTDF depend mainly on the direct flux scattering and extinction which are accurately computed in SAIL.
Finally, to compute the original thick layer operators, AddingS uses the adding-doubling method allowing sublayer concatenation.
Energy closure is thus not enforced.
In practice the operators are matrices of size 200 × 200.
They correspond to a sphere tessellation dividing it into 400 samples (20 samples for both zenithal and azimuthal angles).
The AddingS model is written in MATLAB and, when run on a DELL OPT.
GX 620 workstation with a PENTIUM 4 processor with 1 Gb of RAM, it takes about 40 s to do the closed forest canopy (3-D) or grasslands (1-D) simulations.
