pr °gram
,Û V
tes, A
Zitier
'«ti»,
fce/haze
*. t W0
LI >9 ail
lenerate
i®S o{
equation
IR) was
al data
ides an
étal.,
, 1987).
nge, but
spectral
o those
raraetric
;ines as
afluence
of the
latively
sort dry
December
:he tine
ip by the
, within
nse is
ions are
■ to be
tructure
die rain
spectral
eas.
along drainageways and/or mature plantations. A
third spectral class, NONFOREST, may include
grasslands (recently burned or areas of senescent
grasses), immature plantations, town centers, and
water courses.
Table 2. Land Cover Class Assignments based on
Spectral Signatures for SITE ONE.
malysìs
SITE
ONE
in
1973
SITE ONE
in
1987
classes
SIGNATORE
% COVER
SIGNATORE
% COVER
ses used
CLOSED
FOREST
21.9
CLOSED
FOREST
13.4
26
17
27
19
15.0
23
19
26
12
5.9
26
16
29
21
5.4
22
18
27
12
5.6
remote
25
16
25
16
1.5
22
18
31
15
1.9
ial and
OPEN FOREST
17.0
OPEN FOREST
31.5
«, and
28
20
28
18
14.7
24
22
25
11
7.4
»station
28
19
31
22
2.3
24
23
27
12
15.2
fry and
24
21
27
12
8.9
finalized
NONFOREST
61.1
NONFOREST
55.1
laluated
28
21
25
16
26.0
24
23
23
10
10.2
city of
27
21
18
10
9.7
24
23
22
8
11.7
/I have
30
26
31
19
0.4
24
23
20
7
7.3
NEVI :
28
21
23
13
16.7
25
25
22
9
25.9
iorption
26
20
16
7
3.1
Isat MSS
27
18
22
12
1.5
equation
29
21
21
11
3.7
Figure 3. Land Cover for SITE 1-A in 1973.
Land Cover/Veqetation Vigor Changes for SITE ONE
Recognizable patterns in spectral signatures and
the geographic distributions of the spectral
classes were used sis the basis for generating the
final classification. Statistics for percent land
cover for the entire area are not considered
accurate because of the extensive cloud/smoke/haze
cover in the 1987 data. Thus, statistics for the
six sub-sites (Table 2) are presented as a means
of indicating representative land cover variations
throughout the study area in haze-free locales.
Land cover maps were generated for each of the
sub-sites for both dates with CLOSED FOREST cover
represented in black, OPEN FOREST cover in dark
gray, and NONFOREST in light gray (with clouds
represented in white). To illustrate land cover
and changes in southwestern sections of the
Central African Republic, sub-sites from within
the closed equatorial rain forest (Figures 3 and
4 of Site 1-A), a location on the boundary between
the closed rain forest and the grasslands (Figures
5 and 6 of Site 1-B), and a sub-site with large
patches of recently burned grasslands (Figures 7
and 8 of Site 1-C) are presented.
Table 3.
Sub-sites
Land Cover Characteristics for the Six
(in percent).
ü
I
%