On the basis of these legends and final field survey to verify 
boundaries, a series of thematic maps at 1:250,000 were prepared from 
the Landsat imagery. The fully verified information on the aerial 
photographs were transferred onto a photomosaic base to produce 
detailed maps at 1:100,000. 
RESULTS 
The full implications of these map results as yet have not been fully 
assessed. However, the legends derived for theme mapping have been 
found to be consistently applicable and operational. The 
geomorphological mapping legend supplies a useful integrative frame 
work for resource mapping as most of the geomorphological features can 
be recognised directly as photomorphic elements. The use of the 
USDA soil taxonomy system allow the soils to be classified according 
to their important practical properties and facilitate maps in terms 
of their land potential. The maps reveal a close correspondence 
between soils and vegetation association. 
Regarding image evaluation each type of imagery has advantages and 
concomitant disadvantages with respect to levels of detail and accuracy. 
These are summarised in Table 3- 
Imagery Scale 
Geomorphology 
Soil 
Vegetation 
Land Use 
Aerial Photos 1:60. 
,000 
* * * * 
* * * 
* * 
Band 7 1:250,000 
* 
* * 
* 
p 
♦ * 
Standard FCC 
H 
* * * 
* * 
* * i - 
* * 
R.B.V. 
fl 
* * 
* * 
* 
* * * * 
Band 7 destriped 
It 
* * 
♦ * 
* 
* * * 
Band 7 H.P.F. 
It 
* 
* 
* 
* * 
Boxcar FCC 
It 
* * * * 
* * * 
* * * 
* * 
Edge Enhancement 
If 
* 
* 
* * * 
* * * 
**** Excellent *** Good ** Moderate * Poor 
(1. Obsolete 2. February scene) 
Table 3* Summary of Performance of Image Types 
The Landsat imagery most suitable to manual interpretation are the 
standard FCC and the 'boxcar' filtered scenes. These both provide 
good discrimatory information on geomorphological features, being 
particularly revealing on the raised marine surfaces, which are zones 
of significant development. 
The larger scale aerial photography makes it obviously superior in 
supplying a finer level of detail for geomorphological mapping. This 
is obviously important for site specific development, but is not so 
important for regional mapping. This also must be balanced against 
the considerable amount of extra time required for the photo 
int erpre t ation. 
The scale and monochromatic nature of the aerial photographs make them 
generally unsuitable for direct discrimination of vegetation type in 
xeric environments. The advantage of Landsat is the extra dimension 
of colour contrast. 
The historical nature of the photographs also renders them obsolete 
for contemporary land cover mapping, but RBV is shown to be excellent 
210