Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4388552 | Ecological Engineering | 2016 | 13 Pages |
Allocation of forest in a river basin is a complex management problem, with conditions that may encourage conflict, in particular among groups with different interests, due to the diversity of the objectives. Aiming at the conservation of water resources, the main objective of this study was to assess the performance of Ordered Weighted Averaging (OWA), a multicriteria decision analysis, in the prioritization of areas for forest restoration. The study area is a Brazilian river basin, originally covered by Atlantic Forest. The relevant criteria to this decision-making process were identified as land-use suitability, soil erodibility, erosivity, proximity to roads, and proximity to surface water. We proposed three order weight sets, based on the decision strategy space, with medium-to-high risk-taking and medium tradeoff among the criteria (OWA1); medium risk-taking and total tradeoff (WLC), representing the traditional Weighted Linear Combination method; and medium-to-low risk-taking and medium tradeoff among the criteria (OWA2). Then, using an OWA operator (as a GIS routine), we produced maps for the three solutions, which were reclassified in five priority levels (very low, low, medium, high and very high). We did the cross tabulation among the solutions, for areas classified as high and very-high priority, and some environmental characteristics, intrinsically related to the criteria: land-use/land-cover, soil erodibility, distance from the surface water, soils and slope. WLC resulted in 54% of the Corumbataí river basin with high priority to forest restoration; 37% with very-high priority; and 9% with medium priority, and the spatial distribution of classes was intrinsically related to two main criteria: proximity to surface water and soil erodibility. OWA1 resulted in one alternative with 80% of the basin classified as very-high priority to forest restoration and 20% as high priority. Such classes’ distribution is the result of the distribution of order weights, to obtain the desired risk level. In OWA2, due to the distribution of order weights among the criteria, the prioritization of areas occurred differently from the previous alternatives, with no predominance of one or more criteria. Under this scenario, the river basin presented 1% of this area with very low priority for forest restoration; 17% with low; 63% with medium; 10% with high; and 9% with very-high priority. OWA is suitable for the prioritization of areas, once it allows us to control the criteria influence on the final solution, through the tradeoff. This implies, however, the method’s ability to normalize the criteria to a continuous scale, since the categorization in the criteria maps affects the spatial distribution of the alternatives.