International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 
  
The criterion weights were calculated as 0.26, 0.21, 0.17, 
0.16, 0.10, 0.06, and 0.04 respectively for annual rainfall, size 
of watershed, basin slope, gradient of main drainage channel, 
drainage density, land use and type of the soil. With the input 
values in pairwise comparison and weights calculated, 
consistency ratio (CR) was found as 0,042. This indicated a 
reasonable level of consistency in the pairwise comparison of 
the factors. 
Three composite maps showing the flood vulnerable areas 
were created using multicriteria evaluation methods with 
GIS, namely Boolean Method and two WLC Methods are 
presented in Figures 3.3 - 3.5. The user interface was 
designed such that a decision maker could step through a 
weighting process with pairwise comparison that would result 
in the calculation. 
£ BOOLEAN METHOD 
  
  
  
ww 
— 
Figure 3.3: The final map that is created with Boolean 
Method. 
  
  
  
  
  
  
  
X: RANKING METHOD 10x 
  
+f RANKING = 
f 
M i 
      
   
  
  
- 
  
  
  
  
  
  
Figure 3.4: The final map that is created with Ranking 
Method. 
LM C crai er aua [-[n[x] 
| PAIRWISE Le) 
HISH i 
m MEDIU M-HIGH| 
[J MEDIUM | 
LOW-MED IUM | 
ow | 
Ca No D ata i 
  
  
  
  
  
  
hd 
  
  
This interface and the processes were presented in Figure 3.6, 
It involves: Viewing all criteria in a list box and each criteria 
pair in a matrix, Entering his or her preferences and ranks for 
cach criteria pair, Calculating weights from the input ranks 
and preferences with Visual Basic Application VBA program 
embedded in GIS environment, Calculating consistency ratio 
on weights and input ranks, Modifving the ranks or 
preferences according to the user's choice. 
  
Factors 
  
Precipitation 
Size of Watershed 
        
Default Factors | 
  
  
  
  
EE EEN ER ER EE 
BFE Ess [5 E 
i ENTE 
  
Figure 3.6: All calculations with weights and consistency 
ratio. 
3.4 Fuzzy Measure Application in MCE 
Zheng and Kainz (1999) have stated that uncertainty is 
endemic in GIS and it is best to draw attention to it because 
of its complexity and potentially damaging effects on 
decision making. The OWA decision rule is based on the 
principles of the fuzzy set aggregation (Malezewski. 1999a; 
Malczewski, 1999b). In GIS and decision making context, 
vulnerability was considered as a fuzzy concept expressed as 
a fuzzy set membership. Jiang and Eastman (2000) presented 
an example paper about suitability as a fuzzy concept. To 
rescale the range to a common numerical basis is unclear and 
involves fuzziness. The criteria were standardized to a 
common numeric range and then combined by weighted 
averaging. In WCL, criterion weights and ordered weights 
were used. Finally the illustrations gave an evaluation of 
vulnerability of flood hazard. The result of OWA method for 
MAX showed the most vulnerable areas and MIN showed the 
least vulnerable areas. AVERAGE fell midway between two 
extreme cases of fuzzy MIN and MAX operation and had a 
full tradeoff. Many other solutions were possible. The 
solutions were the effects of the distribution of risk and 
tradeoff. With this method, the aim was not to ignore the poor 
qualities, but they could be compensated for. Different 
illustrations give important idea to planners for location 
solutions. By compensating one criterion, decision maker can 
create different solutions. 
3.5 Sensitivity Analysis 
Sensitivity analysis depends on the error in the input data as 
criterion weights and criterion attributes. The sensitivity 
between the criteria basin slope and main channel gradient 
was analyzed. With addition of small perturbations to the 
weights for 10 cells, the ranking has changed for only one 
cell. This result indicates that those two criteria are not so 
sensitive for this analysis. When the sensitivity of criterion 
values was analyzed, it was also seen that there were no 
significant changes in the ranking with small perturbations on 
the gradient of the main channel criterion. 
Inté 
The 
as 
vul 
wit 
Pai 
per 
hig 
per 
bas 
zon 
the 
(20 
the 
was 
ana 
vuli 
cou 
obt: 
pair 
floc 
pair 
By 
clas 
is a 
on 
bet 
and 
crite 
app! 
inp 
Aft 
risk 
the 
Det 
wari 
repr 
crea 
parc 
ente 
- The 
in f 
exan 
reali 
chan 
subj: 
crite 
char; 
sens 
that 
care) 
attril 
sensi 
of tl 
that : 
insur 
flooc 
one 
shou 
alteri