6
Figure 2. Landscape typology of the planregion of Flanders. 1 : complex and heterogeneous modern landscapes ;
2 : re-allotments ; 3 : traditional landscapes affected by 1 or more modern impacts ; 4 : traditio
nal rural landscapes (P1, Z1.. different types} ; 5 : townscapes. Ze,SJ,Ba,JMP,SM,Ti indicate
the test sites (see table 2-4).
Table 2. Number of landscape elements in different
regions based upon 80 samples using a simulated
pixelsize of 30mx30m. The landscape components are :
biotic volumes (BM), abiotic volumes (AM), biotic
space (BR), abiotic space (AR), biotic screens (S).
Code
î Site
landscape
elements/
sq.km
for
BM
each
AM
component
BR AR
S
Ti
Ti el t
1861
8
2
102
19
2
Ze
Zerkegem
2133
7
7
110
8
12
Wd
Waterland
1818
0
3
76
9
2
JP
St.Jan (P)
2169
1
2
31
4
3
Aa
Aalter
1980
8
27
100
16
11
SJ
St.Joris
2277
16
10
102
16
20
JM
St.Jan (M)
2553
2
5
111
4
9
LM
Lovendegem
( M ) 219 6
5
9
55
16
3
LR
Lovendegem
(R)1732
0
2
47
4
0
Wa
Waarschoot
2362
12
21
106
17
8
Se
Semmerzake
2383
10
11
92
28
22
Ma
Massemen
2833
8
24
147
13
21
Gi
Gijzegem
2722
22
18
107
26
22
Le
Lede
3486
27
26
131
21
24
Ba
Bazel
2430
8
12
112
18
27
Ni
Nieuwkerken
2805
17
12
135
12
36
Me
Merkplas
1816
18
3
84
18
4
VB
VIiermaal
(B)
2777
41
4
90
9
16
VO
VIiermaal
(0)
1825
5
3
15
0
0
WO
Wellen (0)
1746
1
1
20
0
0
WB
Wellen (B)
3402
55
16
91
14
10
For all these samples in the rural landscapes, the
category of the agricultural land (biotic space BR)
occupies at least 70 % of the area, but clearly
(table 2) has also the largest number of landscape
elements, i.e. land-use fields. Important to note is
that land use fields group already a lot of cadastral
fields (i.e. adjacent cadastral fields with the same
crop) which are used for the agricultural census.
The probability of having a pure pixel of the TM
size in the agricultural space (regardingsless the
crop !) varies between about 50 % to 90 %.
Table 3. Proportion of pure pixels and probability
of having one in a given landscape component (based
upon80 sampleswith a simulated pixelsize of 30 m x
30 m). Further explanation : see text.
Code Site % pure % probability of pure
pixel in pixel in the component :
the
site
BM
AM
BR
AR
S
Ti
Tielt
53
6
1
76
14
3
Ze
Zerkegem
47
5
5
76
6
8
Wd
Waterland
60
0
3
84
10
3
JP
St.Jan (P)
43
2
5
76
10
7
Aa
Aalter
46
5
17
62
10
7
SJ
St.Joris
41
10
6
62
10
12
JM
St.Jan (M)
14
1
4
85
3
i 7
LM
Lovendegem (M)
34
5
10
63
18
4
LR
Lovendegem (R)
56
0
3
89
8
0
Wa
Waarschoot
34
7
13
63
11
5
Se
Semmerzake
43
6
7
56
17
14
Ma
Massemen
18
3
11
72
5
9
Gi
Gijzegem
18
11
9
55
13
12
Le
Lede
13
12
11
57
10
10
Ba
Bazel
35
5
7
63
10
15
Ni
Nieuwerkerken
24
8
6
63
6
17
Me
Merksplas
58
8
3
71
15
3
VB
VIiermaal (B)
31
26
3
55
6
10
VO
Vliermaal (0)
79
22
13
65
0
0
WO
Wellen (0)
71
4
5
91
0
0
WB
Wellen (8)
27
30
9
48
8
5
Again the open landscapes (polders, openfields and
re-allotments) give the highest probability,. Low pro
babilités occur in all landscapes (with large AND
small fields) having a high density of biotic
screens (tree- and hedgerows). A further analysis
of the different landscape units, based upon these
indicators, clearly show two main groups : open
landscape (regardingsless the field sizes) and
enclosed landscapes.
4.3. The size of the fields and the compactness of
their shapes
A shape analysis of the land use fields allowed the
comparison o1
with the pixi
sical geograp
are always ba
not very sign
mote sensing
use. Table 4
ve landscape
ries between
occur. Most f
gular, the co
the ratio bet 1
tion of the w
length. Althoi
same for all :
the pixel size
(L/B ratio) dc
block fields -
Bazel (fig.3)
Table 4. Compa
six
typical
1c
dev
iation,
CV
Code
Length
site
L (in
m)
m
s
CV
Ze
29
11
39
Wa
38
28
73
Se
24
10
42
JM
54
25
47
Ba
22
8
36
SJ
32
13
43
Figure 3. Field
Bazel (L = lengt
Nevertheless, thi
these sites rema
size (especially
biotic screens a:
there are the exl
the sites WaarscP
are both represer
Flanders (fin.4).
larger and the la
