Knowledge of the radiative behavior and the energy budget of land surfaces is essential for studying the functioning of natural and urban surfaces with remotely acquired information. Account of their 3D nature is often essential because in most cases these surfaces are not isotropic. For example, it has long been known that the albedo of a canopy with anisotropic Bi-directional Reflectance Factors (BRF) may be underestimated by as much as 45% if it is computed with nadir reflectance only (Kimes and Sellers, 1985). Radiative transfer (R.T.) models have the potential for correcting this type of error provided they account for the three dimensional (3D) nature of Earth surfaces. Neglect of the 3D structure of canopies can lead to large errors on the 3D radiation budget and remote sensing measurements. For example, for vegetation BRF and directional brightness temperature (DTDF) distribution functions, errors can be as large as 50%, depending on instrumental (e.g., view and sun directions) and experimental (e.g., vegetation heterogeneity) conditions (Gastellu-Etchegorry et al., 1999). The problem is similar for urban canopies due to their strong spatial heterogeneity. The application of R.T. modeling to urban surfaces is important in the context of the advent of satellite sensors with spatial and spectral resolutions that are more and more adapted to urban characteristics such as building dimensions and temperature spatial variability. It explains the numerous works conducted in the field of remote sensing of urban surfaces (Soux et al., 2004; Voogt and Oke, 1998). The use of descriptions with qualitatively based land use data instead of more fundamental surface descriptors is a source of inaccuracy for modeling BRFs and DTDFs (Voogt and Oke, 2003). R.T. models are essential tools for assessing accurately radiative quantities such as the exitance, the irradiance and remote sensing measurements in the optical and thermal domains. However, in order to meet this objective, models must account for the three dimensional (3D) nature of Earth surfaces. Here, we consider vegetation canopies and urban canopies. This consideration of the 3D architecture of Earth surfaces is possible with the socalled 3D models. Generally speaking, the latter ones are intended to be accurate, robust and more comprehensive than other models. Ideally, they should be used in place of other models. However, they are often more difficult to manage, both in terms of computation www.intechopen.com Modeling and Simulation in Engineering 30 time and landscape description. Moreover, when dealing with specific situations, one needs that the model be accurate and robust, but one does not necessarily need that the model be comprehensive. This explains that in many cases, the objective of 3D models is to calibrate models that are simpler to manage in terms of landscape description, computation time, etc. Once calibrated, these models can meet the required accuracy levels. These remarks stress the usefulness of 3D R.T. models. A number of 3D models is being d...
