Taking an area of about 2.3×10~4 km~2 of southeastern Iran, this study aims to detect and predict regional-scale salt-affected lands. Three sets of Landsat images, each set containing 4 images for 1986, 2000, and...Taking an area of about 2.3×10~4 km~2 of southeastern Iran, this study aims to detect and predict regional-scale salt-affected lands. Three sets of Landsat images, each set containing 4 images for 1986, 2000, and 2015 were acquired as the main source of data. Radiometric, atmospheric and cutline blending methods were used to improve the quality of images and help better classify salinized land areas under the support vector machine method. A set of landscape metrics was also employed to detect the spatial pattern of salinized land expansion from 1986 to 2015. Four factors including distance to sea, distance to sea water channels, slope, and elevation were identified as the main contributing factors to land salinization. These factors were then integrated using the multi-criteria evaluation (MCE) procedure to generate land sensitivity map to salinization and also to calibrate the cellular-automata (CA) Markov chain (CA-Markov) model for simulation of salt-affected lands up to 2030, 2040 and 2050. The results of this study showed a dramatic dispersive expansion of salinized land from 7.7 % to 12.7% of the total study area from 1986 to 2015. The majority of areas prone to salinization and the highest sensitivity of land to salinization was found to be in the southeastern parts of the region. The result of the MCE-informed CA-Markov model revealed that 20.3% of the study area is likely to be converted to salinized lands by 2050. The findings of this research provided a view of the magnitude and direction of salinized land expansion in a past-to-future time period which should be considered in future land development strategies.展开更多
文摘Taking an area of about 2.3×10~4 km~2 of southeastern Iran, this study aims to detect and predict regional-scale salt-affected lands. Three sets of Landsat images, each set containing 4 images for 1986, 2000, and 2015 were acquired as the main source of data. Radiometric, atmospheric and cutline blending methods were used to improve the quality of images and help better classify salinized land areas under the support vector machine method. A set of landscape metrics was also employed to detect the spatial pattern of salinized land expansion from 1986 to 2015. Four factors including distance to sea, distance to sea water channels, slope, and elevation were identified as the main contributing factors to land salinization. These factors were then integrated using the multi-criteria evaluation (MCE) procedure to generate land sensitivity map to salinization and also to calibrate the cellular-automata (CA) Markov chain (CA-Markov) model for simulation of salt-affected lands up to 2030, 2040 and 2050. The results of this study showed a dramatic dispersive expansion of salinized land from 7.7 % to 12.7% of the total study area from 1986 to 2015. The majority of areas prone to salinization and the highest sensitivity of land to salinization was found to be in the southeastern parts of the region. The result of the MCE-informed CA-Markov model revealed that 20.3% of the study area is likely to be converted to salinized lands by 2050. The findings of this research provided a view of the magnitude and direction of salinized land expansion in a past-to-future time period which should be considered in future land development strategies.
