696 
from Eq. (3). 
In Fig. 3 the values measured by Ranson et al. (1985) and the calculated 
values of the direct to total irradiation ratio are plotted versus wavelength with 
the Angstrom turbidity factor estimated from the measured sky radiation ratio at 
À = 560 nm. 
Both the PROSPECT model and the Price functions approximate the respective 
measured spectra rather well. The measured leaf inclination distribution has large 
fluctuations, thus the elliptical distribution is not a very good approximation for 
it. The calculated irradiance ratio increases more rapidly with wavelength than the 
measured one. Obviously, this spectrum at West Lafayette, IN (40*N, 87*W) differs 
from that in Great Britain (53*N, 1*W). 
The most problematic issue is the estimation of the wax refraction index. The 
range of c n values is 0.8 ... 1.0 (1.2). This value significantly affects the shape 
of the angular distribution of canopy reflectance in spectral regions with high 
absorption, especially in the case of low Sun. Hence, the parameter c can be esti 
mated by solving the inverse problem for bidirectional reflectance 'in the red or 
blue spectral region. In the NIR spectral region in nadir the influence of this 
parameter almost disappears and we can keep its value constant. 
There is no change in the representation of the canopy reflectance angular 
distribution in the new model compared with the FOR model. Consequently, the new 
model has the same problems with reflectance angular distribution as its 
predecessors: it works well in the spectral regions where the single scattering of 
radiations dominates, and may have some problems in the case of significant 
multiple scattering. These problems have been discussed by Kuusk (1991b, 1993). In 
the present paper main emphasis is placed on the correspondence of the estimated 
canopy spectra to those measured. A comparison with corn nadir reflectance data 
acquired by Ranson et al. (1985) follows. The canopy and soil reflectances were 
measured with a Barnes Model 12-1000 multiband radiometer in 7 visible and 
reflective infrared channels. The values of model input parameters used are given 
in Table 1. 
Table 1. The input values of model parameters 
LAI 
W 
H 
1.2 (0.3) 
Measured 
0.39 (0.05) 
Measured 
0.56 m (0.09) 
Measured 
°- 1 2 
100 ug/cm 
Estimated inverting the PROSPECT model with 
0.026 cm 
measured leaf reflectance and transmittance 
1.09 
spectra 
0.9 
10.7° 
Estimated fitting the measured leaf inclin 
0.972 
ation distribution 
0.213 
0.100 
Estimated fitting the measured soil 
0.013 
reflectance spectrum 
0.015 
The standard deviation of measured values is in parenthesis. 
Figure 4 compares the measured nadir reflectance values with those calculated 
with the new model. The deviation between the calculated and the measured 
reflectance values does not exceed 3%. However, there is a systematic discrepancy 
between the measured and the calculated data. In those spectral regions where the 
soil is brighter than the vegetation, the estimated values are lower than the 
measured ones and vice versa. The main reason for this discrepancy seems to be the 
overestimation of ground cover in model calculations. If LAI = 1.2, there should