(2) 
, and 
nents 
vinter 
green 
y and 
meter 
were 
> Was 
rœa 
ance. 
n the 
rable 
ather 
ntial, 
siven 
t site 
POT 
ween 
Were 
jlved 
falfa. 
vely, 
ilure 
Was 
y for 
were 
data 
man, 
1 see 
time 
sing 
nine 
Clevers, Jan 
  
the total above-ground fresh and dry biomass and its decomposition between organs (green and yellow leaves, stems, 
ears). An under-sample of 6 plants was collected in the immediate vicinity of each of these 6 samples. The total leaf 
surface of the 6*6 plants was measured with a planimeter, as well as the fresh and dry biomass. The LAI was then 
determined by extrapolating the leaf surfaces obtained on the 6 under-samples to the 6 samples, using the ratio of the 
biomass of the samples and the under-samples. Additional information was also described such as the crop calendar, 
canopy height, row direction and stand density. Final yield figures were obtained too. All this information is available in 
the ReSeDA database. 
3.2 Remote Sensing Data 
For this study six SPOT images were acquired. Four of them were within the 1997 growing period of the wheat crops 
(namely February 1%, March 25" May 2™ and July 7"). The SPOT images were first corrected for the sensor's 
modulation transfer function (MTF) and for atmospheric effects. The latter was based on actual atmospheric optical 
thickness measurements during satellite overpass, using a sun photometer installed on the test site, and applying the 6S 
software. Subsequently, the images were geometrically corrected to the Lambert II extended projection system. Finally, 
the mean and standard deviation of the reflectances in the three SPOT spectral bands was derived for the relevant fields 
for all recording dates (present in the ReSeDA database). 
4 RESULTS AND DISCUSSION 
4.1 Destructive Measurement of Leaf Area Index 
From the destructive biomass measurements and the specific leaf area the LAI was ascertained. The results for the 
wheat fields are depicted in figure 1. Subsequently, curve fitting was applied for interpolation purposes. Continuous 
logistic functions, including a growth period and a senescence period, were fit to the green LAI measurements: 
LAI(t) = M x (GRO(t) - SEN(t)) (3) 
with: GRO(t) = 1/(1 + exp(-a.(t-t;))) (4) 
SEN(t) = 1/(1 + exp(-b.(t-t,))) (5) 
where t is the day of year in 1997 and M, a, t;, b and t, are parameters estimated by non-linear fitting (nlmin function 
under the Splus software). 
Table 1 gives the obtained values of the fitted parameters for the various wheat fields, along with the residual root mean 
square errors (rmse) in m?/n? units (note: for field 208 no curve was fitted because measurements at the beginning of 
the growing season are missing). The fitted curves are presented in figure 2. 
4.0 Estimating Leaf Area Index from Remote Sensing 
First of all, the WDVI values for the five wheat fields identified in section 3.1 were calculated according to equation 
(1). The slope of the soil line (C) was calculated by using reflectances derived from the SPOT images for a bare field 
during the first half of 1997 and for the harvested wheat fields on August 30^, 1997. The estimate for C (based on NIR 
and red reflectances) yielded a value of 1.26. Using this estimate, the WDVI was calculated from the four SPOT images 
within the growing season of wheat. The result is illustrated in figure 3. 
Subsequently, the parameter estimates obtained from Clevers (1991), and described in section 2.1, were used for 
deriving estimated LAI values from these WDVI values according to equation (2). A subdivision into the vegetative 
growth stage (first two SPOT dates) and the generative growth stage (last two SPOT dates) was made. The result is 
shown in figure 4. 
Finally, the LAI estimates for the SPOT dates for wheat were compared with the destructive field measurements of 
LAL The fitted logistic growth curves (figure 2) were used to obtain interpolated values for the SPOT recording dates. 
Results of the comparison between SPOT derived LAI estimates and field measurements are illustrated in figure 5. 
Since the SPOT recording of July 7" occurred just after harvesting the wheat fields, this date was not included in the 
comparison with field measurements. A good correspondence was obtained (rmse = 0.57). 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000. 275