the phenomenon of desertification. With the aide of a supervised 
classification using the method of maximum of likelihood, we 
obtained the map of land cover which includes 7 classes which 
represent the following table: 
C) the map of the slopes 
We obtenaid the slopes from a digital terrain model (see figure 
6), this map includes 4 classes which are: 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Classify | Themes | Corteges floristiques state of classify slope in (%) | surface in (ha) | surface in (%) 
deterioration class 1 0-3 4415211 77.40 
class 1 Cereal class 2 3-6 9084,33 15,92 
class 2 course | Artemisia compestris FD’ class 3 6-12 3147,93 5,52 
class 3 course Stipa tenacissima FD class 4 >12 662.93 1.16 
class 4 course Noaea mucronata VD" Tb 3: classes of slopes 
class 5 Sand 
class 6 structure 4 - Realisation of the map of sensitivity to the desertification 
class 7 Bare soil by crossing of the different information layers. 
  
  
  
  
  
  
  
  
  
Tb. 1: classes of land cover 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
O.c | 1% | 2% | 3% | 4% | 5% | 6% | 7% |pixels| total 
l 100 25 | 100% 
2 95.65] 4.35 46 | 100% 
3 8.82 |91.18 34 | 100% 
4 100 61 100% 
5 100 33 | 100% 
6 100 13 | 100% 
7 100| 12 | 100% 
total 224 | 100% 
  
  
  
  
  
Tb.2: The classification Matrix of confusion 
In order to determine the precision of the classification we have 
evaluated the indices (G and K) which are respectively 0.97 and 
0.98, these two meaningful values show that the classification is 
good. 
B) the pedopaysagic map 
The pedopaysagic map was established using an approach based 
settled on a photo interpretation of the satellite pictures 
LANDSAT TM of 08 April 1995. This interpretation allowed us 
to point out some radiometric units in relation with the couple 
soil morphology and soil land cover (see figure 5). 
  
( remote sensing images ) 
  
  
C colored composition ) 
  
  
fü photo interpretation ] 
  
  
É terrain samples 3 
  
  
( interpretation of the big land units ) 
  
  
(^ results verification 9 
i 
( pedopaysagic map - 
Fig 4: Methodical organigram for the pedopaysagic map 
  
realisation 
  
" fairly degraded 
^ very degraded 
768 
  
  
  
       
  
remote sensing images ) 
| land cover map | pedopaysagic slopes map 
map 
[ information coding ] 
  
(in decision matrix ud 
map 
crossing 
  
( synthesis map ) 
  
Fig 7.: General organisation chart of the methodology for the 
realisation of the map of sensitivity to the desertification 
Once we have built the basis maps we proceeded with a numeric 
coding of the thematic information in order to introduce theme in 
a GIS, in order to make the operations of seizing, of storage and 
make the combinations. On the other hand we defined and 
evaluated the potential sensitivity to desertification of each one 
of the physical parameters, their reactions to the processes of 
deterioration are different and depend on their own features. The 
potential sensitivity of the vegetation parameter depends on the 
physiognomic type, of its biology of its rate of recovery and 
especially of the appeal that its presents for man. for the 
morphology and the pedopaysagic are examined in their opinion 
carriers ( the relief, the texture and the blinding). The final 
crossing different layers of information according to a very 
definite matrix us allowed to get 4 classes: very sensitive class, 
sensitive class, class fairly sensitive, class little or most sensitive 
(see figure 8). 
CONCLUSION 
In this study the objective was to develop a methodology 
combining multisources data in order to apprehend the 
phenomenon of desertification. The integration of the 
cartographic data in a GIS and their crossing according to a very 
definite matrix, allowed to get some results indicating the most 
exposed zones to the problem of desertification 
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998