September 26, 2013


The inversion of a vegetation model by a neural network

The Leaf Area Index (LAI) is defined as half the total intercepting green foliage area per unit ground surface area (Chen and Black, 1992). The ADEOS-2/POLDER-2 algorithm computes the LAI using a neural network, which inverts the radiative transfer model of Kuusk (1995) considering the vegetation as a turbid medium of leaves with spherical orientation. This model simulates the simple scattering in the canopy (in particular the hot spot phenomenon quantified by the parameter l*, ratio of the leaf size to the canopy height) following the Nilson and Kuusk (1989)' approach, and the multiple scattering according the SAIL model (Verhoef, 1984). Furthermore, the leaf optical properties are described by the PROSPECT model (Jacquemoud et al., 1996) whereas the spectral and angular properties of soil are reproduced by the coupling of the functions of Price (1990) and the directional Walthall (1985)' model, respectively. For the specific application to the POLDER data, parameters of the PROSPECT model have been adjusted to account for the chlorophyll concentration, Cab, the senescent pigments concentration, Cs, the dry matter content, Cdm, the water equivalent thickness, Cw, and the effective number of layers inside a leaf, N. More, only director factors, a1 and a2, of the 2 first functions of Price (1990), which have been optimized, are considered; others are set to 0. The learning base of the neural network has been produced by sampling LAI [0 - 6.5], Cab [15 - 80µg/cm²], N [1 - 4.5], Cs [0 - 2], a1 [0.1- 0.8], whereas other parameters are fixed (Cw = 0.01g/cm², Cdm = 0.015g/cm², l* = 0.1 and a2 = 1). The network inputs are a single orbit of 11 directional reflectances in 3 spectral bands (565nm, 670nm, and 865nm) and their angular configurations. The output is the LAI estimated for each POLDER track. Then, a simple merging algorithm, with gaussian temporal weighting, averages these retrieved LAI over 30 days to get a monthly value mainly characteristic of the central 10 days period.
Then, the FVC is derived following the relationship:
FVC = 1 - exp(- 0.5 x LAI)


Chen, J. M. and T. A. Black, Defining leaf area index for non-flat leaves, Plant, Cell and Environment, 15, 421-429, 1992.

Jacquemoud, S., S. L. Ustin, J. Verdebout, G. Scmuck, G. Andreoli and B. Hosgood, Estimating leaf biochemistry using the PROSPECT leaf optical properties model, Remote Sensing of Environment, 56, 194-202, 1996.

Kuusk, A., A fast invertible canopy reflectance model, Remote Sensing of Environment, 51: 342-350, 1995.

Nilson, T. and A. Kuusk, A reflectance model for the homogeneous plant canopy and its inversion, Remote Sensing of Environment, 27: 157-167, 1989.

Price, J. C., On the information content of soil reflectance spectra, Remote Sensing of Environment, 33: 113-121, 1990.

Verhoef, W., Light scattering by leaf layers with application to canopy reflectance modeling: the SAIL model, Remote Sensing of Environment, 16: 125-141, 1984.

Walthall, C. L., J. M. Norman, J. M. Welles, G. Campbell and B. L. Blad, Simple equation to approximate the bidirectional reflectance from vegetative canopies and bare soil surfaces, Applied Optics, 24, 383-387, 1985.

© CNES Copyrights 2002