1 2004
cm
0, 20-
m).
from
OC in
The
epend
ses in
en for
same
ation,
on to
SOC.
lation
linear
terest
te its
varies
1" for
on)
efers
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
3.4 Regression Analysis between LAI and SOC
LAI 2 a * b SOC
estimate standard p<05
of the equation error
a - 1,336 0,367 0,068 :
b 0,045 0,008 0,032; *
r 0,97 0,033; *
Table 2. Linear regression between SOC and LAI
where a = intersection with the y-axe
b = gradient
r = correlation coefficient
* = significant correlation with p < 0,5
The correlation coefficient r (0,97) indicates a significant
correlation between SOC and LAI, with reliability over 95%.
3.5 Pasture reflectance
3.5.1 Conversion of DNs into reflectance
Band Exo-atmospheric reflectance (p) (96)
. -Monjelada Barreiro R. —Bondade Descalvado
Blue 9131 10,0 + 10,46 + 10,55 +
0,38 0,36 0,37 0,31
Green 8,31 + 9,48 + 9,55 + 9,30 +
0,61 0,48 0,50 0,34
Red 6,11 + 8,19 + 9,61 + 10,33 +
0,87 1,08 0,80 0,75
NIR 25,51 + 30,13 + 22,66 + 20,17 +
2,10 1.87 1,60 1.02
SWIR | 17,92 + 23,18 + 27,16 + 28,38 +
2,39 2.83 2,01 2,44
SWIR II 6,17 + 8,81 + 14,42 + 14,58 +
1,46 1,92 2,06 1,90
Table 3. Conversion of DNs into planetary reflectance
Observation: the value + refers to the standard deviation
3.5.2 Regression analysis between SOC and pasture
reflectance The following table shows the results of the
linear regression analysis between SOC and six spectral
bands.
Reflectance = a + b : SOC r
Blue (12.8 € 0.9) * (- 0.06 + 0.02) - SOC 0.91
Green (4.4 £ 0.6) * (-0.02 £ 0.01) - SOC 0.64
Red (16.8 + 1.8) + (- 0.19 + 0.04) - SOC 0.96
NIR {11,1 x 10.4) + (0:31 + 0.25), - SOC 0.68
SWIRI (44.9 + 5.2) + (-0.47 + 0.12) - SOC 0.95
_SWIR IL (28.7 + 5.8) + (-0.41 + 0.13) - SOC 0.91
Table 4. Linear regression between SOC and
pasture reflectance
where a = intersection with the y-axe
b = gradient, r = correlation coefficient
Two correlation coefficients (r) were equal or higher than
0.95, indicating a significant correlation with a probability of
at least 95%; therefore these two correlations (red band/SOC
and SWIR I/SOC) were plotted and discussed in detail.
799
3.5.3 Correlation between SOC and pasture reflectance in
the red spectrum
Memes: - Desclavado
10 Bondade
i | Barreiro Rico
|
©
Reflectance (%)
oo
~
r Monijelada
Reflectance red spectrum =
o
(16,8 x 1,8) - (0,19 320,04) - C |
25 30 35 40 45 50 55 60
Soil organic carbon 0-50cm (My ha!)
Figure 4. Correlation between SOC and reflectance in the red
spectrum of the studied pastures
C
[-
Observations: The average reflectance of each pasture is
acquired by following pixel number: (217 Monjelada, 495
Barreiro Rico, 89 Bondade e 89 Descalvado).
The SOC is considered a strong factor of influence in terms
of soil reflection. The higher the SOC, the fewer the
reflectance in the wavelength spectra visible to short wave
(Major et al, 1992), as can be seen in figure x. The soil
reflectance of the studied pastures is likely to be the same, as
the studied pastures feature the same soil type, climate,
topography and a closed vegetation cover (except Bondade).
This leads to the assumption, that the in figure x plotted
differences in pasture reflectance are mainly linked to
differences in plant reflectance. The red spectra suffer
radiative absorption by the photosynthetically active
pigments chlorophyll a (absorption max. 675nm) and b
(absorption max. 480nm) in the green leaves. Therefore, the
photosynthetically absorption pattern are responsible for the
different reflectance of the studied pastures in the red
spectrum.
3.5.4 Correlation between SOC and pasture reflectance in
the shortwave infrared I spectrum
Descalvado
D
|
28 | Bondade
|
23 PE
© | 7 Barreiro Rico
Eu | |
S tan
= 2) |
©
X | |
201 E Monjelada
| Reflectance SWIR 1 spectrum =
18 | ;
| (44.9 + 5.2) - (0.47 + 0.12) - C
16!
25 D 35 40 45 50 55 i90]
Sail ageric cation 0-50cm (Mg hai)
Figure 5. Correlation between SOC and reflectance in the
shortwave I
infrared spectrum (1550 - 1750nm) of the studied pastures
