Shahraki, Masoud (2021) Spatial Variability Estimation and Mapping of Soil Salinity Using Data and Remote Sensing Geostatistical Analysis: A Case of Hamoun International Wetland. Masters thesis, University of Zabol.
|
Text
Spatial Variability Estimation and Mapping of Soil Salinity Using Data and Remote Sensing Geostatistical Analysis A Case of Hamoun International Wetland.pdf Download (142kB) | Preview |
Abstract
Soil salinity is one of the serious problems in arid and semi-arid regions of the world, and therefore the study of its spatial and temporal changes is important for land management and rehabilitation, both agriculturally and environmentally. In recent years, the bed of Hamoun International Wetland has become one of the important centers for soil harvesting and dust production. By decreasing soil moisture and increasing soil salinity, the sensitivity of the wetland bed to wind erosion and the occurrence of the phenomenon has increased. It is necessary to prepare an updated map of how soil salinity is distributed in order to provide appropriate methods for soil stabilization and reduction of wind erosion. In recent decades, the use of geostatistical techniques and remote sensing data to estimate, map and analysis the spatial and temporal changes in soil physical and chemical properties, especially in large-scale areas and in the presence of limited data, has developed greatly. The purpose of this study is to investigate and map the spatial distribution of soil salinity and alkalinity using remote sensing techniques and geostatistics in a part of Hamoun International Wetland. For this purpose, a total of 192 samples were collected from 96 points from two surface (0-15 cm) and subsurface (15-50 cm) depths from parts of Hamoon International Wetland by a supervised random sampling method in fall of 1398. Electrical conductivity, soil pH and sodium adsorption ratio of soil samples were measured by the standard methods. Multispectral satellite images (Sentinel 2 and Landsat 8 sensors) and radar (Sentinel 1) were obtained from the study area. Data analysis was performed by the statistical and graphical software GS +, ArcGIS, ERDAS and SPSS. Soil characteristics maps were prepared using geostatistical methods and remote sensing indicators. The results of measured soil properties showed that the soil in the study area was saline and alkaline. For an autocorrelation analayis, an appropriate semivariogram model was fitted for each property. The results showed a poor to moderate spatial correlation for the investigated parameters. The accuracy of the method used for mapping soil properties was evaluated through some statistical measures. It was found that soil pH has the highest estimation accuracy for both surface and subsurface soil depths with the RSME, MAE and R2 equal to 0.243 , 0.2 and 0.05, and 0.217, 0.17 and 0.04 respectively for surface and subsurface data. In order to evaluate the effect of auxiliary variable (i.e. DEM) on the accuracy of estimatiomn, statistical analysis was perfomed first and it was found that the auxiliary variable has little effect on increasing the estimation accuracy. Remote sensing investigation has also shown that radar images of Senitenel 1, unlike other images, have the ability to penetrate into the subsoil. Based on the results, soil pH did not have a significant correlation with any of the Sentinel 2 and Landsat 8 images, but it had the highest correlation with VV polarization in radar images. The correlation between electrical conductivity and remote sensing indices with GNDVI index showed the highest correlation between Landsat and Sentinel 2 images and sentinel 1 VV polarization . SAR also has a high correlation with Landsat NDVI index, Sentinel 2 b5 / b9 index and Sentinel 1 VH polarization. the results show that Sentinel 1 radar images have the ability to penetrate the lower layers in cases where the vegetation prevents penetration to the lower layers compared to the images of the Sentinel 2 and Landsat 8 satellites, which can be used as an advantage in similar cases. also, according to the maps obtained from geostatistical methods and remote sensing, it was found that reducing moisture and increasing salinity is directly related to the rate of soil erosion and wind from the study area, this is effective in increasing dust and dust and air pollution.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Electrical conductivity, Vegetation index, Spectral reflectance, Remote sensing, Geostatistic |
| Subjects: | S Agriculture > S Agriculture (General) |
| Depositing User: | Mrs najmeh khajeh |
| Date Deposited: | 02 Oct 2022 06:40 |
| Last Modified: | 02 Oct 2022 06:40 |
| URI: | http://eprints.uoz.ac.ir/id/eprint/3108 |
Actions (login required)
 |
View Item |
