Assessing the expansion of saline lands through vegetation and wetland loss using remote sensing and GIS

Abstract

Soil is a non-renewable and dynamic source that most of the human food supply comes from the soil, and one of the factors affecting soil degradation is soil salinity. The aim of this study was to predict the salinity expansion of Golpayegan's plain through estimated losses in vegetation cover and wetland area, using indices, and spatial techniques. The first 22 Landsat images were selected from 1985 to 2016. Two sets of maps, Soil Salinity Index (SSI) and Normalized Difference Vegetation Index (NDVI) were extracted from the Landsat images. By using of two indices, the classification of these lands was carried out. These maps with six classes, vegetation, and salinity were entered into the Land Change Modeler (LCM). Smart modeling was continued by Multi-Layer Perceptron (MLP) Neural Network method. After validation of the model, the transition potential probability maps were prepared. By using the Cellular Automata (CA) Markov method, 2025 forecast map was obtained. The results showed that among different salinity indices, SSI6 was the best index. In the study area, the land changing to high salinity has occurred such that the low salinity classes are first obtained and gradually changed to medium and high salinity classes. There is a reverse relationship between vegetation density and soil salinity, that is, with wetland drying up and increasing salinity, the density of vegetation decreases. Decision-makers should be control of drainage networks and water resources exploitation to control the salinity expansion.