Bemigisha, Jane 
  
  
if Cij- Cji » 0 (4) 
then alternative i is better than j. 
where : C- is the criteria, i and j are the alternatives 
In the evaluation, the relative importance of criterion 1 (papyrus change ) was assigned to be higher than that of criterion 
2 (land cover change rate), using the option rank order. The corresponding appraisal scores were inverted to express the 
influence ( in terms of risk) of the land cover types to Papyrus thus the lower the coefficient the better. The resulting 
appraisal scores were used as the weights in the final risk modelling involving the two scenarios but after a sensitivity 
analysis. 
24.2 Sensitivity analysis: Sensitivity analysis is the process of ascertaining the consistency of the results of the 
evaluation model. The analysis was carried for effects and weights uncertainty given 20% probability. If found 
consistent, then the results world be suitable for input to the model. 
2.8 Risk potential to papyrus swamp (maps) combining the criteria 
Multi -criteria spatial models mostly aim at deriving the suitability of the choice options (e.g. Eastman et al. 1995). The 
model expression would take the form: 
Sz wx; (5) 
Where S = Suitability to the objective being considered 
w; = Weight of factor i 
x; = Criterion score of factor 
However, the application of the above expression implied suitability than the intended risk. Therefore 
the derived coefficients were inverted by dividing the highest coefficient (score) by each resultant coefficient. 
The final risk potential map for papyrus area was spatially derived by a linear additive function for the weighted 
landcover types in a GIS (ILWIS 2.1). The final expression took the form: 
Risk map = Factor1*P1 +Factor2*P2 +Factor3*P3... (6) 
Where: factor are the modifying factor as distance risk index maps (for the land cover types) 
P is the weighted position (coefficient) suggested by DSS from the rate of change of land cover types and the change in 
area of papyrus by the land cover type (from temporal analysis). 
Thus the expression for Scenario 1 based on equation 6 was as follows: 
Finl=( shrmca*1+ watmca*0.85--Agmca*0.74--Wdmca*0.28--Gramca*0.21--bumca*0.17) (7) 
Where: Finl = Potential risk combining all land cover types, and the three criteria. 
shrma= shrubland, watmca= open water, Agma= Agricultural fields, Wdmca= woodland, Gramca= Grassland, 
Bumca-built up area. (see also appendix 13) 
A similar expression (but excluding open water) was applied for scenario 2 
  
168 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.