Evaluation of soil salinity by using satellite images and statistical methods: A case study of Chah-Afzal area, Ardakan

Soil salinization is one of the problems in arid and semi-arid areas, and its control requires the adoption of proper management. Using up-to-date soil data and remote sensing technologies is a good way to evaluate soil salinity in different regions. In this research, with the aim of evaluating soil salinity in Chah-Afzal area of Yazd, a comprehensive map of different salinity levels was proposed by combining remote sensing and statistical methods. For this purpose, at first, 50 soil samples from the study area were taken randomly, and their electrical conductivity was measured. Then, Landsat 7 satellite images were prepared at the same time as the field harvest date in August 2007. The pre-processing steps were performed for visualization using salinity indicators. From the 14 variables, six different soil salinity indices and eight satellite image bands were used to obtain the best equation for soil salinity. In order to reduce the number of variables in providing the best salinity estimation equation, factor analysis and backward regression analysis were done, and three equations were introduced to indicate soil salinity. On the basis of the results obtained from the soil salinity maps and the verification of the maps, NDSI salinity indices with the highest correlation coefficient of 0.6 could serve as the best soil salinity model. The study area was classified into five soil salinity classes, which showed that about 90% of the area had a salinity of more than 50 dSm-1. According to the results, preparation of soil salinity maps by combining remote sensing techniques and statistical methods save time and costs and increases the accuracy and speed of studies.