|
چکیده
|
Assessing and predating changes of the natural ecosystems are required for the management and sustainable usage of land and water resources in arid and semiarid regions in which, the exploitation of land is rapidly changing. This research is mainly aimed to detect and simulate the changes of land use/cover changes in relation to groundwater levels. For this purpose, Markov chain model and artificial neural networks (ANNs) were applied to predict land-use and groundwater level changes, respectively. Moreover, TM and EMT + Landsat satellite images of 1988, 2001, and 2014 were utilized as one of the data sources of this study. Supervised classification methods were used for change detection analysis. Subsequently, Markov chain model was applied to predict the land-use changes in 2027 after finding an acceptable accuracy of about 82%. The results of change detection show that agricultural lands and artificial forests have increased during 1988–2014, while rangeland and residential and barren lands have decreased in area. It is predicted that the region will experience an increasing trend in terms of artificial forest and agricultural lands, while rangeland, residential and barren lands will continue to follow a decreasing trend in 2027. Furthermore, the groundwater levels have dropped down about 6.1 m during the period of 2000–2014. The results of prediction analysis by ANNs show that the groundwater level will continue decreasing about 16.8 m during 2014–2027. These changes could be related to conversion of rangelands to agricultural lands and artificial forests, which needs more water for irrigation.
|