401 
Table 3. Colour-Interpretation Key for the Land Units 16 I and 8 III 
IHS Colour 
Land Unit 16 I - Eastern 
Ruaha Lowlands 
Land Unit 8 III - Nduli-Ismani Flats 
Interpretation 
Corresponding 
Deciding 
Symbol 
Interpretation 
Corresponding 
Deciding 
Symbol 
alternatives 
FCC colour 
reflection 
in Fig- 
alternatives 
FCC colour 
reflection 
in Fig- 
factor 
ure 1 
factor 
ure 2 
Yellow 
1) Brown alluvial sand 
White 
Soil colour 
1 
1) Yellow sand 
White 
Soil colour 
A 
2) Dark, saline loam 
White 
Veneer of 
2 
2) Yellow loam with 
White/light pink 
Soil colour 
B 
light sand 
sandy topsoil 
Red 
1) Red, saline loam 
Brownish 
Soil colour 
3 
1) Deep, highly 
(Light) brown to 
Soil colour 
C 
leached red clay 
olive 
2) Shallow, stony red 
Brownish 
Soil colour 
4 
2) Deep, moderately 
(Light) brown to 
Soil colour 
D 
loam 
leached red clay 
olive 
3) Shallow, stony, 
(Light) brown to 
Soil oolour 
E 
moderately leached 
red clay 
olive 
Green 
1) Dark, saline loam 
White/light violet 
Soil colour, 
5 
1) Periodically 
Dark violet, brown- 
Soil colour, 
F 
(vegetation) 
flooded dark 
cracking clay 
ish violet 
drainage 
2) Dark cracking clay 
White, light violet. 
Soil colour, 
6 
2) Extremely hard, 
Light violet 
Soil colour, 
* 
violet 
(surface 
imperfectly 
drainage 
gravel) 
drained dark clay 
Bluish green 
1) Dark saline loam 
Violet 
Soil colour, 
vegetation 
7 
2) Dark cracking clay 
(Brownish) violet 
Soil colour, 
(vegetation) 
8 
Pinkish violet 
1) Brown alluvial sand 
Red 
Vegetation 
9 
and loam 
Dark violet 
1) Rocky hills with 
Red / dark gray / 
Vegetation, 
G 
shallow stony 
soils 
black 
rock 
Black, black- 
1) Dark cracking clay 
Dark grayish violet, 
Soil colour 
H 
ish green 
2) Rocky hills with 
dark brownish gray 
Red / brownish gray 
Vegetation, 
I 
shallow soils 
rock 
3) Burnt area 
Black, dark gray 
Burnt vege 
tation 
J 
* This soil occurs outside of the displayed section of the land unit. 
responds to an alluvial fan. The areas marked with 
"b* are related to recent river floodplains. The 
remaining area belongs to a pediplain. 
The area displayed in Figure 2 consists of a pe 
diplain with some small inselbergs and one big de 
pression (F). An extensive area of rocky hills 
starts in the lower right corner (G). 
9. SUMMARY AND CONCLUSIONS 
The advantages and limitations of aerial photo 
graphs (black & white panchromatic) and two types 
of Landsat MSS imagery (standard false-colour com 
posite and colour-enhanced imagery) were examined 
during a small-scale soil survey in southern Tanza 
nia. 
The three types of remote sensing data proved to 
be of very different use in different agro-ecolo- 
gical zones of the study area. The satellite image 
ry was most useful in the drier parts of the area 
(550 - 900 mm annual rainfall) where colour differ 
ences of the soil surface were well recorded by the 
spectral resolution of the MSS images. In the 
agro-ecological zones with higher rainfall (900 - 
1600 mm) the denser vegetation rendered the satel 
lite imagery much less useful. 
Aerial photography was most helpful in an inter 
mediate agro-ecological zone with a complex relief 
and prevailingly relief-related boundaries. 
Comparing the two types of satellite imagery, the 
colour-enhanced images were characterized by a 
clearly better colour discrimination. The FCC ge 
nerally showed a better differentiation of the 
image texture. 
Colour interpretation keys for the colour-enhanc 
ed imagery were compiled for two selected areas of 
the drier part of the study area. The results de 
monstrate that a definite interpretation of image 
colours is rarely possible. Even within a relative 
ly small climatically homogeneous land unit the 
same soil may cause different colours on the image. 
Vice versa, the same image colour can be related to 
very different soils. 
The following general conclusions can be drawn 
from the present study: 
- Survey accuracy and survey efficiency can be 
considerably increased by the application of 
advanced satellite image enhancement techniques. 
- For small-scale soil mapping aerial photographs 
as well as satellite imagery are of limited 
used in high rainfall areas with dense vegeta 
tion. Thus surveys in this type of area will 
generally be more time-consuming and less ac 
curate than in drier areas. 
- Interpretation keys based on satellite image 
colours can only be established for climatical 
ly homogeneous areas. Even then their use is 
limited without profound field checks. 
- In the past the information from satellite ima 
gery and aerial photographs complemented each 
other due to the different qualities of these 
two remote-sensing materials. The new genera 
tion of satellite imagery (SPOT, Landsat TM) 
now combines high spectral resolution with high 
spatial resolution and (in case of SPOT) with 
stereo capability. It is therefore expected 
that aerial photographs will become more and 
more redundant for small-scale soil and land 
resource surveys. 
REFERENCES 
Allan, J.A. 1978. The contribution of Landsat ima 
gery to reconnaissance soil mapping in north-west 
India. In Remote Sensing Applications in Develop 
ing Countries, 43. Remote Sensing Society. 
Carrol, D.M. 1984. Soil mapping from space - 
achievements of the first decade and future pos 
sibilities. In Satellite Remote Sensing - Reviews 
& Preview, 31. Remote Sensing Society. 
Goosen, D. 1967. Aerial photo interpretation in 
soil survey. FAO, Rome. 
Haydn, R. et al. 1982. Application of the IHS colour 
transform to the processing of multisensor data