destructive effects of the thunderstorms between the 
years 1965 and 1975. Studies are now performed to 
protect the New El-Menia town against the run-off 
caused by the thunderstorms (Salem et al 1982). 
3 INTERPRETATION 
followed in more details by using a 10 times magnifi 
cation of band 7. Some representative photomorphic 
units have been choosen to be studied in detail on 
aerial photographs. The aerial photographs were stu 
died with a mirror stereoscope and the zoom-transfer 
scope. 
3.1 Multistage analysis 
4. RESULTS AND DISCUSSION 
The main idea of a multistage analysis is to make a 
maximum use of the different remote sensing documents, 
in view of organizing but also of minimizing the 
field observations. This approach is referred by 
Lillesand and Kiefer (1979) for geology mapping. It 
has been recommended for the assessment of deserti 
fication by Rapp (1974/ and was successfully used by 
El-Hag (1984') . 
Landsat MSS imagery, conventional aerial photo 
graphs and ground truth have been used in the current 
study. Also, different tools have been employed to 
zoom on different features (e.g. over head projec 
tor and transparant screen, mirror stereoscope and 
zoom-transfer scope). It resulted progressivelly 
in more accurate information for corresponding smal 
ler units of the study area. The detailed informa 
tions, obtained from the aerial photographs and from 
the field observations have been fed in the small 
scale interpretation of the MSS landsat imagery. 
That allowed us to establish a map of the soil con 
ditions by comparing the image characteristics with 
the ground conditions. 
3.2 Diazo processing technique 
This technique is based upon the idea that each fea 
ture has its own specific reflectance characteristics. 
Thus, the grey tone of different features in diffe 
rent bands will differ according to these characte 
ristics. It was noticed from the original negative 
images that the vegtation is dark in band 7 and light 
in band 5, whereas water is dark in band 4 and 5 and 
light in band 7. Light sand is dark in all bands. 
Consequently vegetation and light sand are indis 
tinguishable on band 7 but different in band 5, 
whereas vegetation and water can only he distingui 
shed by examining both band 5 and band 7. This 
comparison between two or more black and white ima 
ges is inconvenient. Producing the colour composites 
by superimposing two or three bands, each in a dif 
ferent colour, is a more convenient method. The in 
formation contained in different spectral bands are 
combined into a single colour composite. 
The preliminary documents used were neaative hlack 
and white frames of three spectral bands (4, 5 and 7) 
on a scale 1:1,000,000. A photographic enchancement 
was performed by using Agfa-Gevaert 081P-Graphic 
Gevalith ortho film. It resulted in two different 
enchanced images for each band, one has the maximal 
contrast within the desert area and the other within 
the cultivated Nile Valley. The obtained images 
were then colour-coded using the colour diazo tech 
nique (Gad and Daels 1985). 
For the desert area the colour composite was for 
med by magenta (band 7), cyan (band 5) and green 
(band 4). For the cultivated area the following 
combination was used : cyan (band 7), magenta (band 
5) and green (band 4). 
3.3 Visual interpretation 
The different landscape elements are represented by 
the image characteristics. The interpretation of the 
landsat images was performed first for the colour 
composites on Macro and micro photomorphic unit 
levels. The used image characteristics were : the 
colour classes, the, texture, the quality of the unit 
boundaries, homogeneity of shape and size and rela 
tive size. Some features have been choosen to be 
The delineation of photomorphic units in the desert 
area, supplemented with field observations and the 
existing base maps, reveals the different physieqtra- 
phic regions. 
4.1 Western desert 
It was possible to distinguish the different erosio- 
nal and depositional features from the different docu 
ments. It was obvious that the wind erosion process 
is dominant obscuring the drainage pattern. Multista 
ge interpretation enabled us to localize the belts of 
sand sheet, sand dunes and the eroded limestone pla 
teau. 
@ 5(263) cm 263 
l~n~n 135/134-263,®, Y.H.P (5~51 3/4(263),®,X,h,P/ti 
A 
¡’TH 12(263) 
(ft Pi 4(263)(189) 
B 
in nl 4/189,®-c,Y,h,P/b 
I I 263,f,Y,H,b 
lo3 2/263-134, ®,Y,h,b 
c 
Multistage analysis of a sand sheet zone. 
A. Macro photomorphic units 
8. Micro photomorphic units 
C. Aerial photo interpretation 
A. Plateau and Hills, 3. Sloping area (sandy and graveily material). :. Low elevated area (sandy 
rsaterltl}. Guillas * avin « V } /1 Nile Valley 
(Z2 16 íNlle Va 
03 • *(3) 
f 7 r 71 (Nile /alle 
i6,f-m.X,H, 
iTTI 263/189,c,Y 
FF»=t3 3/189,c.Y.h 
4.1.1 San< 
This land 
just west 
part of t 
(fig. 2a) 
patches. 
parts, wi 
pebble de 
also that 
nation of 
gravel. 1 
the micro 
differeno 
deposits 
The aer 
area of s 
that it i 
It was al: 
tern in t! 
SE direct 
SE direct 
1962). 
4.1.2 Sam 
This phys: 
tomorphic 
composite 
bluish wh: 
microphoti 
smaller ui 
The inter] 
(fig. 3C) 
and the ii 
its morph< 
the movinc 
wind direc 
sat images 
it was no 1 
tudinal d\ 
colour in 
detailed : 
vealed th< 
tudinal d)