×

You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Title
Remote sensing for resources development and environmental management
Author
Damen, M. C. J.

241
a 35mm KODAK
4 RESULTS AND DISCUSSION
> or minus 2
solar zenith
rallel to the
ground level
lm by 1.5m
~ey card with
included in
1 was judged
etween camera
i relation to
ght meters do
tter speed
Lsions. Hence
ons were made
0 ASA. It was
suitable for
-ding to NASA
a strip of
wedge in a
ras used for
points were
sion densito-
WRATTEN No 94
red) filters
ties relating
jspectively,
jrements were
ensities were
unit spectral
reen, red and
1 number of
it ions were
re determined
ison (1982).
Tasseled Cap
tained:
for further
dp to green
(Vegetation
ence or ND
Perpendicular
s to the Soil
»nd 1977) was
ines for the
ogy, applied
i used for
.ides can be
Lew).
placing a dot
;he amount of
1884 r ;.95
fifi r=.«3
28 S
K
ent.
4.1. Green Leaf Area Index (GLAI)
Development of GLAI was is markedly different for
both growing seasons (Figure 2). It is supposed that
significantly more rainfall, - April and June
featured almost the double amount as usual caused
stronger leaf development in 1985. In general the
measured values agree with data given by other
authors for sugar beets (Evans 1972, Thorne 1971).
Figure 2. Sugar beet cover percentage (full line) and
GLAI (dashed line) in function of time in 1984 (x)
and 1985 (+).
Linear relationships of low GLAI values with VI have
been reported by Hinzmann et al (1986) for wheat and
Badhwar et al (1986) for aspen forest.
Asymptotic behaviour of retained ratios and
combinations of reflectance could be observed for
GLAI values larger than 2 (Figure 3), which is in
agreement with results reported by Tucker (1977). Due
to saturating reflectance large GLAI values cannot
be estimated very accurately by spectral parameters.
This puts severe constraints on yield prediction
procedures as photosynthesis levels off at much
higher GLAI values.
Figure 3. Asymptotic behaviour of ND in function of
sugar beet GLAI.
Best fit regression analysis was run on the GLAI data
and spectral parameters obtained in 1984 (Table 1).
The indicated regressions were used to calculate
predicted values based on 1985 spectral parameters
and plotted against averaged measured GLAI values
(Figure 4). Dispersion around the 1:1 line increases
for high GLAI values, and PVI appears to be most
powerful predictor for this biomass parameter. It
should be noted that regression correlation
coefficients were not indicative of predictive
power.
Table 1. Regressions of sugar beet GLAI and spectral
parameters in 1984.
Parameter Regression type
Corr. coeff.
VI
power fit
.77
PVI
power fit
.79
ND
power fit
.90
GREENNESS
logarithmic fit
.81
Figure 4. Relationship between measured (1985) and
predicted GLAI for sugar beets. Dashes represent 1:1
line.
4.2. Cover percentage
Complete closure of sugar beet canopy occurs in a
relatively short time span (Figure 2). Hence few
experimental data are available for low cover
percentage. It is precisely for these low values that
linear relationships with certain spectral parameters
are reported in litterature (Colwell et al 1977 for
wheat using VI, McDaniel 1982 for grasslands using
ND).
As can be seen in Figure 5 100 % coverage can yield
varying spectral values.
Figure 5. Asymptotic behaviour of ND for cover
percentage of sugar beets.