468 
2. MATERIALS AND METHODS 2.2. Field spectroscopic data 
2.1. Agronomic data 
The study area was located in the 
Laureinepolder in Watervliet, close to 
the border with the Netherlands. The 
geographical coordinates are 51°17’ N 
and 3°40’ E. The soil is a fertile 
Aquic Udifluvent . 
The winter wheat cultivars included in 
the experiments were 'Granta' in 1984, 
'Castell' in 1985, 'Sarno' in 1986 and 
'Cas tell' in 1988. 
In 1984 and 1985 the experiments were 
carried out on commercial fields, and 
the crops received normal amounts of 
fertilizer. In 1986 and 1988 the 
experiments were conducted on 
experimental test plots which received 
different amounts of N-fertilizer: 0 
kg, 200 kg and 275 kg in 1986, 0 kg, 
100 kg, 150 kg, 250 kg and 325 kg in 
1988. Green (1987) showed that 
increased nitrogen application causes 
an increase of the leaf area without 
changing leaf properties and leaf 
angle. Hence a wider range of biomass 
data were available allowing for more 
accurate reflectance/LAI relationship 
modelling. 
For each day that field measurements 
took place in 1984 and 1985, three 
random colour infrared (CIR) 
photographic and LAI samples were 
taken. Samples measuring 20 cm x 40 cm 
were randomly harvested in the area 
covered by colour infrared photography. 
Harvested green leaf blades were 
spread out on a matt-black painted 
surface (60 cm x 80 cm) and recorded on 
black-and-white film. 
In the dark room, after development of 
the negatives, a transparent dot grid 
(dot dimension 1.13 mm 2 , each dot 
representing .14 cm 2 ) was placed 
directly on the light-sensitive paper 
(18 cm X 24 cm) during exposure. 
The actual dot counting was facilitated 
by an in-house developed 
electro-mechanical counting device. 
Knowing the total amount of dots 
falling on leaves, LAI could be 
computed as dimensions of black 
background and sample plot were known. 
In 1986 and 1988, one (CIR) 
photographic and LAI sample was taken 
of each test, plot per field day. In 
1986 and 1988 leaves of 10 and 25 
plants were photocopied on a well-tuned 
photocopier. The data collected in 1986 
were counted using a transparent dot 
grid overlay. The data collected in 
1988 were digitized with a CCD camera. 
Density slicing of a videodigitized 
photocopy allowed for the calculation 
of white (non-leaf) and black (leaf) 
areas. The LAI could be calculated 
from the average plant density of the 
experimental plots. 
Ground-based multispectral photography 
was chosen as the field spectroscopy 
method to obtain multispectral 
reflectance data of agricultural 
crops. The used method is described in 
detail by De Wulf and Goossens (1988). 
Field spectroscopic data were collected 
10 times during the growing season with 
approximately 10 days intervals, 
starting at the end of April and ending 
at harvest around mid August. Hence the 
phenological stages between 
stem-elongation and maturity were 
covered. 
The camera-to-ground distance was kept 
constant at 2.30 m. The covered ground 
area measured 2.1 m 2 . 
The Kodak neutral test card (20 cm x 
25 cm) was used throughout the 
experiments as reflectance standard. 
For each recording it was positioned at 
the height of the crop canopy and 
covered approximately 2.5 % of the 
photographed area. 
A few exceptions notwithstanding, 
measurements were taken close to solar 
noon in nadir viewing position. 
Due to the latitude of the test area, 
the solar zenith angles ranged between 
28° and 52°. 
To obtain consistency in azimuthal 
position, as recommended by Milton 
(1987), all measurements were taken 
with the sun either left or right of 
the operator (azimuth angles of 90° and 
270°). This symmetrical configuration 
had the additional effect that the 
target and the grey card were never 
shaded by the recording platform. 
The extraction of relative reflectance 
from CIR transparencies has also been 
described in detail by De Wulf and 
Goossens (1988). 
Following vegetation indices were 
calculated from green (G) , red (R) and 
infrared (IR) reflectance. 
1. Simple ratio (SR) (Rouse et al., 
cit. Bariou et al. 1985) 
SR = IR / R 
2. Ratio Infrared/Visible (RIV) 
RIV = IR/VIS 
with VIS = average reflec 
tance in the 
covered visible 
part of the E.M. 
spectrum 
= (G+Rl/2 
3. Normalized Difference (ND) (Rouse et 
al.. cit. Bariou et al. 1985) 
ND = (IR-R)/(IR+R) 
4. Normalized Difference of Infrared 
and Visible (NDIV) 
NDIV = (IR-VIS)/(IR+VIS)