نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانش‌آموخته کارشناسی ارشد سنجش از دور و سیستم اطلاعات جغرافیایی، دانشکده علوم اجتماعی، دانشگاه محقق اردبیلی، اردبیل، ایران
2 استادیار سنجش از دور، گروه جغرافیای طبیعی، دانشگاه محقق اردبیلی، اردبیل
3 استادیار گروه آموزشی علوم خاک و مهندسی خاک، دانشگاه محقق اردبیلی، اردبیل، ایران
4 دانشیار دانشکده کشاورزی، گروه علوم و مهندسی خاک، دانشگاه مراغه، مراغه، ایران
چکیده
کلیدواژه‌ها
عنوان مقاله [English]
نویسندگان [English]
Background and objective: Preventing soil salinization and managing agricultural irrigation heavily depend on accurate soil salinity estimation. Soil salinity represents a prevalent form of land degradation, characterized by its temporal and spatial evolution. Traditional methods like laboratory analysis and field surveys are inadequate for monitoring soil salinity due to their inability to keep pace with the rapid changes of this phenomenon and their associated high costs. Additionally, the methods used in assessing spatial changes should have the capability to respond to new questions and developments occurring in this field. To address this challenge, satellite imagery emerges as a valuable tool for continuous monitoring, given the sensitivity of electromagnetic signals to soil parameters, particularly in the surface layer directly linked to soil salt content. Numerous studies have been conducted on soil salinity, yielding different results based on ground samples and satellite imagery. Therefore, the attention of soil mappers is drawn to employing data and techniques capable of ensuring sufficient accuracy and reliability by eliminating image errors. Given the significance of this issue, the objective of this study is the evaluating and establishing a relationship between ground data and spectral indices extracted from Landsat satellite images in Bonab County.
Material and methods:
In this study, three types of data were used: Landsat 7 and 8 satellite images with a 15-year interval, DEM imagery as auxiliary data for classification operations, and soil salinity samples collected from 74 different points at 500-meter intervals. These samples were collected from a 40-square-kilometer area in the fall of 2014. To assess the significance of ground samples with satellite images, 12 remote sensing spectral indices were utilized, and after necessary preprocessing (atmospheric, radiometric corrections, and applying a 3*3 filter), corresponding values to EC were extracted. Pre- and post-filter images were examined using regression methods. Subsequently, stepwise regression was employed to examine the relationship between independent variables and the dependent variable. All spectral indices were included as independent variables in the model. The results indicated that among these indices, NDWI and NDSI had the most significant correlation with ground samples. To create the soil salinity change map for the years 1999 to 2014, ground samples and the NDSI index were used. Additionally, using DEM data, ground data, and Landsat 8 imagery, a maximum likelihood classification map for 2014 was generated.
Results and discussion: Regression analysis between EC samples and spectral indices revealed that NDVI (0.45), NDWI (0.37), SI-T (0.43), and NDSI (0.41) had a more significant correlation with soil salinity compared to other indices. The use of filters improved the coefficient of determination for these correlations. Additionally, VSSI and BI indices showed the least significant correlation with ground samples. The soil salinity change chart indicates that in an area of approximately 40 square kilometers, the most significant soil salinity changes, covering 35.3 square kilometers, occurred from saline to highly saline land. The maximum likelihood classification map for 2014 shows that with the drying of Lake Urmia, the trend of increasing salinity in the region has intensified.
Conclusion: In this study, Landsat 7 (1999) and Landsat 8 (2014) imagery was utilized to assess the significant relationship and produce a soil salinity map between ground data and remote sensing spectral indices in Bonab County. The results demonstrated that all extracted indices had significant correlations with soil salinity data, with NDVI, NDWI, SI-T, and NDSI showing stronger correlations compared to other indices. Furthermore, the results of filtering showed that applying a filter to the index could improve research outcomes. The study emphasizes using satellite imagery for ongoing soil salinity monitoring due to its sensitivity and adaptability, outperforming traditional methods. Significant correlations between ground data and spectral indices like NDVI, NDWI, SI-T, and NDSI underscore their effectiveness in analyzing soil salinity dynamics. These findings provide valuable guidance for future research, advocating for filtering techniques to improve accuracy in assessing spatial changes in soil salinity. The findings of this study can serve as a useful guide for selecting data and satellite images in similar studies related to spatial changes in soil salinity.
کلیدواژه‌ها [English]
)