Comparison between two multicriteria methods for assessing land suitability for agriculture: application in the area of Mleta in western part of Algeria

In order to facilitate the decision making in spatial applications, several methods are available to aggregate the performance of decision-makers. The multitude of these methods makes it difficult to choose one method over another. Two main types of methods are used, the outranking methods which are based on bottom-up approach and the hierarchical analysis methods which are based on top-down approach. These methods can be adapted to this kind of applications and can be integrated into GIS which is an essential tool in spatial management and decision making. Methods developed according to the first approach are different from those developed in accordance with the second one. This study aims to present a comparison between two multicriteria decision-making methods ELECTRE (Elimination Et Choix Traduisant la REalite) Tri and AHP (Analytic Hierarchy Process) to assess land suitability for agriculture. Twelve criteria [water reserve easily utilizable, drainage, permeability, potential of hydrogen (pH), electrical conductivity, active limestone, cation exchange capacity, soil texture, soil useful depth, slope, availability of labor, proximity (roads)] grouped in five factors (agronomic, planning, land enhancement and improvement, soil conservation and environment protection, and socioeconomic conditions) were used in this study. An application has been carried out on the area of Mleta in Algeria. Two land suitability maps for durum wheat have been produced by the two procedures (optimistic and pessimistic) available in ELECTRE Tri, and another map has been obtained by AHP. After the comparison between the obtained results, it was apparent that ELECTRE Tri is more appropriate for the land suitability assessment for agriculture despite its complexity given the large number of required parameters. The subjective parameters available in ELECTRE Tri make it possible to take into account imprecise data. Another advantage of this method lies in the way of the decision-makers' preferences aggregation and the assignment of the land units to classes of predefined categories.