全球变化知识资源中心

In the last few decades, the northern region of Iraq has been severely impacted by climate change, which has created long-term drought, water shortage, and casual flood events. Dam construction in suitable sites is a primary water management strategy to solve flood and drought. Dam site selection depends on a set of qualitative and quantitative criteria, such as geology, soil type, and altitude. In this study, appropriate areas of dam site selection for water management were identified using remote sensing, geographic information system (GIS), and multi-criteria decision making techniques. In addition, the proposed method of site suitability was evaluated by comparing it with the traditional analytic hierarchy process (AHP). A new validation method to evaluate the accuracy of AHP and fuzzy logic according to the preselected location of a darn was previously suggested in the study area (Bekhme darn) by the local government. Remote sensing data (Landsat imageries and ASTER DEM) and field/reference maps were used to derive 12 conditioning factors needed to produce a suitability map for construction. The criteria considered were geological formation, soil type, fault line, tectonic line, altitude, slope, rainfall data (2000-2011), water discharge, land use/cover, road network, and material used for dam construction. Two mathematical models, namely, AHP and fuzzy logic, were applied in this study. The best model was selected by using location and cluster-based accuracy assessment depending on the proposed dam site (i.e., Bekhme dam). Results showed that fuzzy logic is more accurate than AHP. In terms of the surface area of suitable lands identified, the AHP model determined 13,446,900 m(2) as a highly suitable area, whereas the fuzzy model identified 3,409,200 m(2) from the total area of 69,347,700 m(2). Finally, this study identified four sites for intermediate and large dams in the study area. The total water capacities of proposed Dam 1, Dam 2, Dam 3, and Dam 4 are about 81.5, 239.3, 374.6, and 646.5 million m(3), respectively, when the dam height is 60 m. The proposed methods can be easily replicated to other areas for similar applications. The finding reveals that AHP model is distributed and scattered in the whole study area, whereas the fuzzy model is more suitable for clustered areas.