عنوان مقاله [English]
Background and objectives: Identifying the saline soils and preparing digital maps of soil salinity, is an effective step in correct management of saline lands. Since vast areas of Iran are covered by saline soils, so these maps are very important. Soil salinity is one of the stages of land degradation that eventually leads to decrease in soil productivity. Soil salinity could be caused by natural processes or human activities. However, soil salinity is a major environmental hazard. So, providing a soil salinity map for these regions, can improve the level of management. Soil salinity maps are prepared by using satellite images as easily as possible. Considering the difficulty of mapping salinity from satellite data, in this study, two approaches for modeling and classification of soil salinity maps were evaluated. The purpose of this study is to evaluate the modeling method and supervised classification of soil salinity mapping using ASTER and ETM+ images in the East of Semnan plain.
Materials and methods: After site selection and spreading a net over the image of area, we determined the location of sampling points. The soil salinity map was prepared After the following steps: measuring EC of soil samples, geometrical and radiometric modification of satellite data, applying some processing such as principal components analysis, fusion of ASTER multispectral bands with ETM+ panchromatic band, transformation of tasseled cap, filtering, producing the salinity indexes, assessment of spectral, and also using supervised classification method.
Results: the salinity map was obtained using modeling method from the eighth band of Aster satellite. The results show that the component that is obtained from integration of an ETM+ panchromatic band and band 5 of ASTER, and a component of salinity index (Salinity2) have a significant relationship. The model validation by the MAE, RMSE and R showed that the selected model has good performance. The accuracy of the salinity map which was produced by Supervised Classification method has been estimated as 84% based on maximum likelihood method and 74% based on minimum distance method. This represents that the accuracy obtained by the above mentioned methods is lower than modeling method for preparing the salinity map.
Conclusion: According to the results of the study, adjusting the salinity indicators resulted in obtaining new indicators for mapping soil salinity. A better diagnosis of soil salinity was resulted from the use of band 3 of Aster image. . So it can be suggested that a part of the electromagnetic spectrum, including (0.52 – 0.86, 2.145-2.185 and 2.295-2.365 micrometers) can be useful in mapping soil salinity in different areas.