The objective, hence, is to illustrate how the 
integration of ELECTRE methods into a GIS does 
contribute in land suitability evaluation. 
First of all, we process the "classical" procedure 
then we apply ELECTRE approach. 
The images showing proximity to roads and 
hydrographic network were created by rasterizing the 
roads and hydrographic vectors and then running 
DISTANCE. 
The image illustrating the proximity to habitations 
and wells were created by using the cost module 
using the COSTGROW algorithm. 
Running SURFACE on a digital elevation model for 
the study area, the image representing the slope 
factor was created. The land cover was obtained 
from the pedologic map 1/25000. The main types of 
soils are: 
isohumic, 
calcimagnesic, 
immature alluvial deposited, 
unweathered mineral alluvial deposited, 
fersiallitic. 
3.1 IDRISI approach 
Each of these images expresses the basic nature of 
the factors identified in this study. However, using 
the weighted linear combination procedure (MCE 
module), the factor maps should be standardised to a 
consistent range of byte binary integers. In addition, 
it requires that each be constructed such that higher 
values indicate areas that are more suitable on that 
factor. 
To develop a set of weights for the factors such that 
they add up to 1, the pairwise comparison method 
was used (Saaty's technique, WEIGHT module). Weights can 
be derived by taking the principal eigenvector of a square 
reciprocal matrix of pairwise comparisons between the criteria. 
The comparison concern the relative importance of the two 
criteria involved in determining suitability for the stated 
objective. Ratings are provided on a 9-point continuous scale 
[1(equal), 3 (moderate), 5 (strong), 7 (very strong), 9 
(extreme)]. 
The MCE (MultiCriteria Evaluation) process combine the 
criteria into a single decision image called "suitability map" 
(in the sense that it shows varying degrees of suitability for the 
objective under question) in the form of a weighted linear 
combination (Voogd, 1983): 
ie, $ - Sw, , Where: S=suitability, x; =criterion score of 
factor i and w; =weight of factor i (see fig.2). 
This procedure is not unfamiliar in GIS and has a form very 
similar to the nature of a regression equation. 
3.2 ELECTRE approach 
Using ELECTRE Tri method we do not have to standardise 
the factor maps. However, the image illustrating the 
soils should be in a quantitative scale (see fig.3). 
Figure 2: IDRISI approach results. 
  
  
  
  
  
  
  
  
  
Figure 3: an example of soil scale quantification. 
isohumic 4... mme ee ae ps 
fersiallitic 
calci- 
magnesic 
immature 
alluvial 
deposited 
unweathered 
mineral 
alluvial 
deposited 
  
  
— 
a 
t» 
pot ge 
a 
Ao 
uM A 
o 
oo "T 
o 
ELECTRE Tri builds an outranking relation S, i.e., validates or 
invalidates the assertion aSb, (and b,Sa), whose meaning is "a 
is at least as good as b,"; where F denote the set of the indices 
of the criteria 2.82, …— Zn (F={1,2, ..., m}) and B the set of 
indices of the profiles defining p+ categories (B(1, 2, ..., p}). 
b, being the upper limit of category C, and the lower limit of 
category C4«,, h- 1, 2, ..., p (see table 1). 
Table 1: profiles evaluations. 
  
  
Criteria Scale b, b; 
Habitation -4234.67/0(m)  -1500  -200 
road -3471.31/0(m)  -800  -200 
well -3832.75/0 (m)  -1000  -500 
hydrography - 3429.65 / 0 (m) -600 . -300 
Slope -32 / 0 (96) -20 À 
Soil 0-1-5-6-7-9 +4 +6 
  
Preferences restricted to the significance axis of each criterion 
are defined through pseudo-criteria. The indifference and 
preference thresholds (q,(by) and pj(b) constitute the intra- 
criterion preferential information. 
At the comprehensive level of preferences, in order to validate 
the assertion aSb, (or bpSa), the conditions of concordance and 
non-discordance should be verified. 
88 International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
Two 
the c 
e 
t Cu. > pe qe mo oc rp 
The 
follo 
ELE( 
that 1 
(bsSa 
"cutti 
Two 
qa" m» "m gg m rd 
assig 
  
  
LEUR)