International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B-YF. Istanbul 2004
TVI
1
0,8 er
= 06 [m "— -—
P o4 "m
0,2
0
1987 1993 2000 2003
Years
“Sunflower = * Wheat "=== Barley
Figure 8. Transformed Vegetation Index results.
PVI
100
50 AR gro
PVI
1987 1993 2000 2003
Years
Sunflower = * Wheat "-»» Barley
Figure 9. Perpendicular Vegetation Index results.
In all vegetation indexes, it was observed that there was a
decrease in the year of 2000. As expressed in many references,
the drought effect observed by the precipitation measurements
(Figure 10) in a year of 2000 was the main cause (Caldag et.al.
2003).
[I December
D November
October |
Bi September |
DO August
B July
El June
Wi vay
CO April
O March
Wi February
El January
Monthly total precipatation (mm)
Figure 10. Monthly total precipitation in Kirklareli
in between 1985-2003 years.
4.2 Vegetation biomass classification
For this application, only NDVI dataset was chosen. Three
training areas representing highly vegetated (mean value greater
than 0.7), low vegetated (mean value of 0.4 — 0.0), and non-
vegetated (mean value smaller than 0.0) areas were chosen
(Figure 11) and the supervised classification method was
applied.
184
a
BUE, Mor AO Ia Rm
4 sie, Ab 5005 aam. Wr T
y * SES 4
a
i Lf ON P
3 A ;
AD a D =
WS 2 Ve ST IN
Figure 11. Selected training areas for the supervised
classification.
The classification result is given in Figure 12. As can be seen,
the green areas represent the highly vegetated areas, brown
areas represent the non-vegetated areas and the dark orange
areas represent the low-vegetated areas. So this can be
concluded that the area is mainly highly vegetated area. The
areal extents of each class were also caleulated and shown
graphically in Figure 35. The effect of drought in year 2000 was
also confirmed once more with this figure.
a
5,
N
‘
-1
s ; ^ EN eun
, uo wt F : s % QM Qo 2
“> 5 as d
Figure 12. Classified image of 1993 NDVI.
A
^3 a
Ze AD
