摘要：Soil salinization is one of the significant phenomena accelerating land degradation processes, which in turn cause loss of soil productivity, reduction in biomass production and human health implications. Due to high spatial and temporal variability of soil salinity, mapping and tracking its changes is an essential issue for anticipating natural disasters like desertification and for mitigating severe economic and social consequences in especially arid and semi-arid regions of the world. The main objective of this study is to review the most recent remote sensing applications for soil salinity assessment via various multispectral data such as LANDS AT, Moderate Resolution Imaging Spectroradiometer (MODIS), SPOT, IKONOS, Quick Bird and the Indian Remote Sensing (IRS) series of satellites, as well as hyperspectral data such as HuanJing-Hyper Spectral Imager (HJ-HSI) and Hyperion Earth-Observing 1 (EO-1). Furthermore, utility of recent remote sensing analysing techniques and methods including linear spectral unmixing, Decision-Tree Analysis (DTA), supervised maximum likelihood classification, principal components analysis, inverted Gaussian function, partial least square regression technique and regression-kriging approaches are discussed. Examples of spatio-temporal changes in salt-affected Mediterranean soils are referred. In addition, various spectral bands which are sensitive to soil salinity for each of the multispectral data are introduced. Besides, details are given on the generation of salinity and vegetation indices to achieve the most significant correlation with ground electrical conductivity values by means of combining appropriate spectral bands with special focus on the Mediterranean countries.
