DART radiative transfer model
DART is a 3D radiative transfer model, designed for scientific research, in particular remote sensing. Developed at CESBIO since 1992, DART model was patented in 2003. It is freeware for scientific activities.
DART model simulates, simultaneously in several wavelengths of the optical domain (e.g., visible and thermal infrared), the radiative budget and remotely sensed images of any Earth scene (natural / urban with /without relief), for any sun direction, any atmosphere, any view direction and any sensor FTM. It was designed to be precise, easy to use and adapted for operational use. For that, it simulates:
- Terrestrial landscape.
- The atmosphere (optional simulation).
- The space or airborne radiometric sensor (optional simulation).
It simulates any landscape as a 3D matrice of cells that contain turbid material and triangles. Turbid material is used for simulating vegetation (e.g., tree crowns, grass, agricultural crops,…) and the atmosphere. Triangles are used for simulating translucent and opaque surfaces that makes up topography, urban elements and 3D vegetation. DART can use structural and spectral data bases (atmosphere, vegetation, soil,…). It includes a LIDAR simulation mode.
General Information On Radiative Transfer
The approaches used to simulate radiative transfer differ on 2 levels: mathematical method of resolution and mode of representation of the propagation medium. These two levels are in general dependent. The models of radiative transfer are often divided into 2 categories associated with the 2 principal modes of representation of the landscape: homogeneous or heterogeneous representation. For the models known as homogeneous (Idso and of Wit, 1970; Ross, 1981; Verhoef, 1984; Myneni et al., 1989), the landscape is represented by a constant horizontal distribution of absorbing and scattering elements (sheets, branches, etc...). On the other hand, for the models known as heterogeneous, the landscape is represented by a no uniform space distribution of unspecified elements of the landscape (North, 1996; Govaerts, 1998).
Simulation of the "Earth - Atmosphere" scene
DART simulates radiative transfer in the "Earth-Atmosphere" system, for any wavelength in the optical domain (shortwaves : visible, thermal infrared,…). Its approach combines the ray tracing and the discrete ordinate methods. It works with natural and urban landscapes (forests with different types of trees, buildings, rivers,…), with topography and atmosphere above and within the landscape. It simulates light propagation from solar irradiance (Top of Atmosphere) and/or thermal emission within the scene.
The study of the functioning of Continental surfaces requires the understanding of the various energetic and physiologic mechanisms that influence these surfaces. For example, the radiation absorbed in the visible spectral domain is the major energy source for vegetation photosynthesis. Moreover, energy and mass fluxes at the "Earth - Atmosphere" interface affect surface functioning, and consequently climatology.
In this context, Earth observation from space (i.e., space remote sensing) is an indispensable tool, due to its unique potential to provide synoptic and continuous surveys of the Earth, at different time and space scales.
The difficulty in studying continental surfaces arises from the complexity of the energetic and physiologic processes involved and also from the different time and space scales concerned. It comes also from the complexity of satellite remote sensing space and from its links to quantities that characterize Earth functioning. These remarks underline the need of models, because only these can couple and gather within a single scheme all concerned processes.
- modelling radiative transfer in heterogeneous 3-D vegetation canopies, 1996, Gastellu-Etchegorry JP, Demarez V, Pinel V, Zagolski F, Rem. Sens. Env., 58:131-156.
- Radiative transfer model for simulating high-resolution satellite images, Gascon F., 2001, Gastellu-Etchegorry J.P. et Lefèvre M.J., IEEE, 39(9), 1922-1926.
- The radiation transfer model intercomparison (RAMI) exercise, 2001, Pinty B., Gascon F., Gastellu-Etchegorry et al., Journal of Geophysical Research, Vol. 106, No. D11, June 16, 2001.