448 
Photo 1 . 
M1 
Photo 3. 
E 
Photo 4. 
S1 
Photo 5. 
S2 
The escarpment (E) was not subdivided, al 
though different land-use units could be ob 
served at this stage. The Muguga and Embakasi 
forests were typical escarpment features, and 
therefore a part of the overall land-cover 
unit and not a separate one (Photo 3). 
The mottled pattern of photo-tones of the 
slope area (S) has been grouped into five 
different zones S1-S5. The main land-use is 
agriculture, however each zone has a diffe 
rent character. S3 and S4 are characterized 
by large-scale plantation agriculture, S1, S2 
and S5 by small-scale activities associated 
with more built up areas. Density of settle 
ment, forested areas and the presence of ir 
ritation ponds, all contribute to the mottled 
appearance of the satellite image classified 
into zones according to the mixture of these 
items (Photos 4-9). 
Photo 6. 
S3 
Photo 7. 
S4 
Photo 8. 
S4 
Photo 9. 
S5 
In the plains (P) four zones were distin 
guished, P1 representing open rangelands, 
P2 the expansion of Nairobi, P3 managed range- 
lands and P4 plantation agriculture (Ph.10-13) 
The resulting land-cover map derived from 
the LANDSAT image provides a useful frame 
work of nine zones from which a more detai 
led field study is required to define land- 
use completely. The landscape planner requi 
res a land-use plan for detailed work. This 
land-cover map is an intermediate step inte 
grating the broad physiographic zones with 
the broad pattern of human activity in the 
area. Once this is mapped, it becomes possi 
ble to concentrate fieldwork in appropriate 
areas and hence proceed to a more detailed 
land-use map in a more efficient manner. 
4 DISCUSSION AND CONCLUSIONS 
The LANDSAT data at 1:1M scale in a FCC 
transparency proved to be easily interpre-