446 
The 1:1M road map of Kenya provided the geo 
graphical orientation. The multispectral 
additive viewer was used to study various 
conbinations of bands 4,5,6 and 7 separately. 
Band 5 is used for the black and white repro 
duction in this paper, showing maximum detail 
2.2 Method 
Yellow and Green: (E) Even less vegetation 
and very dry condition, resulting in in 
crease of reflection. 
Yellow, Blue and 
^rown: (P) Vegetation cover is 
further diminished and the grey soil gives 
an even higher reflection than in zones E 
or S. 
Visual interpretation of the LANDSAT image 
identified four photo-tonal zones, resulting 
in the physiographic map (Figure 2). 
Two traverses were chosen which crossed all 
of the zoned types, and fieldwork was carried 
out along these cross-sections (Figure 1, 3) 
to identify the existing land-cover. A data 
sheet was designed to record field informa 
tion at intervals of approximately 5 km. 
The variables recorded included altitude, 
landform, drainage, vegetation and land-use. 
The land-cover/land-use classification system 
of the United States Geological Survey was 
applied (Anderson 1976). 
Because of the construction of the FCC, these 
zones are mostly identified by vegetation.. 
However, there are more sub-divisions of vege 
tation possible, and the physiography is more 
properly identified as follows: 
Mountains (M) uniformly red 
Slopes (S) mottled red and green 
Escarpment (E) yellowish 
Plains 
(P) 
yellow-blue 
Figure 2. Physiographic map. 
SK-N 
E 
Compilation of the physiographic data veri 
fied the physiographic map initially compiled 
from photo-tones and colours. The additional 
field data led to the production of a land- 
cover map at 1:1M (Figure 4), the same scale 
as the initial LANDSAT image. This landcover 
map is more detailed than the physiographic 
map, and represents the limit of interpreta 
tion at this scale. 
3 RESULTS 
3.1 Physiographic map 1:1M (Figure 2) 
From an initial inspection of the LANDSAT 
data in FCC transparency at 1:1M scale, the 
boundaries shown in Figure 2 were drawn. 
These boundaries were drawn around areas 
which had similarity in colour, texture and 
tone, each representing a physiographic zone, 
identified as mountain (M), slope (S), 
escarpment (E),and plains (P). The colour 
analysis shows the classes to be roughly 
identified as follows (Table 1): 
Dark Red: (M) Heavy vegetation or 
forest 
Red and Green: (S) Vegetation is less 
dense and probably includes bare soil. 
The January image presents the end of the 
dry season, when certain crops have already 
been harvested. 
Figure 3. Physiographic cross-sections. 
Table 1. Physiographic elements. 
Mapping unit 
Phctntone 
Altitude 
Drainage 
(M) Mpuntain 
dark red 
2100 m 
parallel 
(S) Slppes 
red and green 
1500 m 
1800-2100 
m 
parallel 
parallel 
(E) Escarpment 
yellow and green 
1800-2100 
m 
internal 
(P) Plain 
yellow, blue and 
brown 
1500 m 
1800 m 
dendritic 
dendritic 
These boundaries were also 
presence in the individual 
examined 
black and 
for their 
white 
transparencies of the LANDSAT data. Figure 2 
shows the relationship between the physio 
graphic boundaries and the band 5 monochrome 
data. The physiographic regions can be iden 
tified by tone and texture on the monochrome 
data of the other bands. Field checking of 
these boundaries and reference to available 
topographic maps showed that the four physio 
graphic regions had the following characte 
ristics (Table 1):