Temporal and Spatial Variation of Soil Salinity Variations in a 24-year Period in Abadan District Using Satellite Images

Soil Salinity caused by poor water and soil resources management and non- sustainable agriculture, lead to the decrease of soil productivity. In traditional soil salinity measurements, soil samples are collected from field and analyzed for electrical conductivity. Remote Sensing data and techniques are more effective and frugal for soil salinity monitoring. Surface salts spectral reflectance are studied widely as a direct evidence of saline soils. This study aims to distinguish spatial and temporal variation of saline soils in Abadan district. For this purpose, Landsat images with a 24 –year interval were used in conjunction with field data. Three indices including B/G, NDSI and SI extracted from images. Linear regression used for assessment of indices and results shows that SI index is more compatible with field data (R2=0.573). In the period of this study, soil salinity in some parts of the district has been decreased and in some other parts has been increased. The indices are not acceptable in the absence of field data. This study shows that soil salinity in Abadan district has decreased in the period of the study. However, the soil salinity is above the standard levels, except in few parts that soil salinity has been increased. Satellite images and salinity indices are useful to make an overview of soil salinity, but data uncertainty should be indicated.
Extended Abstract
1- In the traditional soil salinity measurement methods, samples are collected from field and analyzed in the laboratory to determine the concentration of soluble substances or electrical conductivity (EC). However, these methods are time consuming and costly, because they need large amount of sampling to determine the spatial variation of soil salinity in an area. Unlike traditional methods, remote sensing techniques are more effective and frugal method to monitor the soil salinity and salinity map preparation. Halophyte plants on the soil surface can be a sign of soil salinity, which enables indirect detection of saline soils based on the spectral signature of the vegetation. Normally, unhealthy or salt-affected plants have less photosynthesis activity, which lead to more reflection in the visible and less reflection in near-infrared (NIR) spectrum which is referred as salt stress in plants. Several vegetation indices such as NDVI, SAVI are used as indirect indices of soil salinity. Besides, some researchers have introduced salinity indices that are directly applicable to salinity detection, such as NDSI and SI.
The purpose of this study was to investigate the spatial and temporal changes in salinity of surface soils in Abadan district. For this purpose, Landsat satellite imagery has been used for a 24-year period from 1990 to 2014. Three soil salinity indices have been applied on satellite images and compared to illustrate changes in soil salinity and shows which indices are more suitable for the study area?
2- The study area is Abadan district, located at Iran-Iraq political border in the southwest of Iran. The annual precipitation of the study area is less than 250 mm per year. This area has been affected by saline water in recent years and soil productivity had been decreased.
The satellite image of TM Landsat 5 (1990) and OLI Landsat 8 (2014) were used. The usual radiometric and atmospheric corrections were done for these images. The regression was done based on the Landsat 8 data that was simultaneous with soil sampling.
The salinity indices used in this study includes the Normalized Differential Salinity Index (NDSI), the ratio of blue to green bands and Salinity Index (SI):
Band ratio= B/G


3- Soil salinity indices have been created primarily to detect the spectral reflectance of salts and to distinguish the difference in spectral reflectance of saline soils in satellite images.
Using different indices of soil salinity shows that there are various preferences to estimate soil salinity. The linear regression model was used to evaluate different soil salinity indices and model parameters were obtained. Based on the performed regression, the SI index shows the best fit with the field truth data.
Because the salt’s tendency to accumulate in separate parts of a pixel in a satellite image, a variety of spectral reflectance of salts are present in a given pixel. In addition, salt layers have a considerable difference in terms of structure (from soft to hard) and color (from white to dark). So the spectral reflectance is too complex. Thus, the reflections recorded in a pixel may include several different salt minerals and several different salt structures.
Urban areas show higher levels of salinity than other parts of the study area, but it is different from soil. Therefore it is considered as an interference with surface soil reflectance. The soil salinity of large parts of the study area is high and the areas with low or moderate soil salinity are very limited.
4- According to the regression results of the three studied indices, the SI index has the highest correlation with field truth data. Of course, it should be noted that none of the indices is highly acceptable and can’t be used without field data. Saline soils in arid, desert and marginal regions have a lot of morphological and chemical diversity. This makes a variety of reflections from the surface. High evaporation and the low quality of groundwater and high water-table levels lead to salt accumulation in soil. In terms of human factors, water supply projects, cut off the hydraulic connection between the river and the flood plain, which leads to an increase in groundwater levels which may lead to soil salinity. This breaks the water-salt balance, so the salt is deposited in the soil. Dam construction on the upstream of the river also causes lowland soil salinity. This study showed that salinity in most parts of Abadan district has been decreased in the 24-year period of this study. In some areas, where soil salinity has been increased, it may be because of preparing the land for farming. It should be noted that despite the decrease in the soil salinity during the period of the study, soil salinity is still more than usual and non-saline soils in the region are rarely seen. Even though the soil salinity has been decreased, most of the district area has been covered with saline soils