ارزیابی شدت بیابان‎زایی با استفاده از شاخص‌های طیفی منتج از تصاویر ماهواره‌ای مطالعه موردی: شهرستان بندر ماهشهر

ABSTRACT
Desertification is one of the factors in the destruction of natural ecosystems in arid regions of the world. Knowing the areas exposed to desertification is very important to combat this phenomenon. Remote sensing is a practical tool for evaluating and monitoring land degradation and desertification. The current research aims at the desertification intensity evaluation in Bandar Mahshahr County based on the spectral indices derived from satellite images. To begin with, utilized the ENVI software to extract several indices, such as NDVI, SAVI, RVI, TGSI, and Albedo, from the satellite image captured by the Landsat 8 OLI in the region. Then, Linear regression was utilized to determine correlations of spectral indices in the region, and the desertification intensity in the region was classified. The results showed that the correlation coefficient between NDVI and Albedo indices was -0.83, between SAVI and Albedo indices was -0.78, and between RVI and Albedo indices was -0.77. The correlation coefficient between TGSI and Albedo indices was 0.86. The higher correlation between TGSI and Albedo indicates that the Albedo-TGSI model is more appropriate for evaluating the desertification intensity in the region. The desertification map of the Albedo-TGSI model showed that the areas with less desertification intensity are located mainly in the northern and eastern parts, and the areas with more desertification intensity were situated in the southern and southwestern parts of the region
Extended abstract
Introduction
Many arid and semi-arid regions of the world are affected by land degradation and desertification. Climate changes, environmental hazards, and human activities cause desertification. Desertification causes a decrease in land potential due to factors such as loss of vegetation and destruction of soil resources. Controlling desertification is one of the necessities and priorities of natural resources management. Due to spatial and temporal information, remote sensing (RS) and satellite images play an essential role in evaluating and monitoring land degradation and desertification at local, regional, and global scales. Over the last few years, spectral indices have been increasingly utilized to determine land cover. These indicators are particularly beneficial in identifying areas susceptible to environmental hazards. Using spectral indices in creating desertification intensity maps can be an effective tool. By visualizing the areas susceptible to desertification, decision-makers and land managers can prioritize their efforts and resources more effectively. The detailed information provided by these intensity maps allows for targeted interventions and the implementation of appropriate land management and conservation practices to mitigate the effects of desertification. Additionally, by utilizing spectral indices to create intensity maps, stakeholders can better understand the spatial distribution and severity of desertification, leading to more informed decision-making in natural resources management. This, in turn, can facilitate the development and implementation of sustainable land use policies and programs aimed at controlling and reversing the process of desertification. Therefore, these maps serve as effective tools for reducing the impact of land degradation and implementing strategic desertification control measures. This research aims to assess and classify the severity of desertification in Bandar Mahshahr County, located in the southwest of Iran and south of Khuzestan province, by utilizing spectral indices derived from satellite images.
 
Materials and Methods
In this research, all the processes were performed on the OLI sensor image of the Landsat satellite 8 of the region on June 18, 2021, in row 39 and pass 165. The dark Subtraction method was used for the atmospheric corrections of the image. Then, spectral indices of NDVI, SAVI, RVI, TGSI, and Albedo were extracted from the region's image using ENVI 5.6 software. SPSS 22 software was used for statistical analysis, and ArcGIS 10.8 software was used to prepare desertification intensity maps. After extracting the spectral indices, the correlation between them was evaluated. To investigate the relationship between the four indices NDVI, SAVI, RVI, and TGSI with the Albedo index, a linear regression model based on 40 random pixels was used. In order to obtain desertification intensity equations, the slope coefficient of the regression line between the spectral indices was calculated. The natural breaks (Jenks) method in ArcGIS software was used to classify the data value into five degrees of desertification (areas without impact, low intensity, medium intensity, high intensity, and very high intensity). The map of spectral indices was validated using the error matrix and two parameters as Overall Accuracy and Kappa Coefficient.
 
Results and Discussion
Numerical values for the NDVI index, -0.45 to 0.51; for the SAVI index, from -0.91 to 1.03; for the RVI index, from 0.36 to 3.14; and for the TGSI index, from -0.09 to 0.17 were obtained. An Albedo index map was created to assess the relationship between the NDVI, SAVI, RVI, and TGSI indices and the Albedo index. Based on the obtained results, the minimum and maximum values of the Albedo index were 0.127 and 0.415, respectively. The lowest values of the Albedo index were estimated in the northern and eastern regions, and the highest values were estimated in the southern and southwestern regions. The results showed that with an increase in vegetation in the region, the number of the Albedo decreases. The linear regression model results between the indices showed that the three indices, NDVI, SAVI, and RVI, have a negative correlation with the Albedo index. Thus, the Albedo index decreases as the NDVI, SAVI, and RVI indices increase. The correlation coefficient between the two indices NDVI and Albedo is -0.83, between SAVI and Albedo, is .78, and between RVI and Albedo is -0.77. The linear regression model results between the TGSI and Albedo indices showed that these indices have a strong correlation relationship. The correlation coefficient between the TGSI and Albedo indices was 0.86. The study findings indicated that as the TGSI index increases, the Albedo also increases. Previous studies have also shown a significant relationship between desertification processes and Albedo and TGSI indices. Thus, the amount of Albedo is a function of the size of the surface soil particles, and with an increase in the size of the surface soil particles, the amount of Albedo increases. The study of desertification intensity maps in this region showed that the areas with less desertification intensity are located mainly in the northern and eastern parts, and the areas with higher desertification intensity are situated in the southern and southwestern parts of the region. For spectral index map validation, 231 pixels were selected as the ground reality of the study area. More samples were taken from the classes that had more desertified lands. Validation results of the spectral indices showed that the NDVI index had the least accuracy, and the TGSI index had the most accuracy in zoning the desertification intensity in the region.
 
Conclusion
This research used Landsat satellite images to extract spectral indices and prepare a desertification intensity map in Bandar Mahshahr County. The overall accuracy criteria and Kappa coefficient of the produced maps show the reliability of the desertification intensity zoning results. The TGSI index map has been the most accurate in zoning the desertification intensity in the region. The linear regression model results showed that the three spectral indices NDVI, SAVI, and RVI have a negative correlation with the Albedo index, and the TGSI index has a positive and strong correlation with the Albedo index. The strong correlation between TGSI and Albedo indices showed that the Albedo-TGSI model is suitable for evaluating the desertification intensity in the study area according to its climatic conditions. This model can be used in regions with similar climates to determine the desertification intensity. According to the obtained maps of desertification, the southern and southwestern parts of the region have the highest intensity of desertification.
 
Funding
There is no funding support.
 
Authors’ Contribution
All of the authors approved the content of the manuscript and agreed on all aspects of the work.
 
Conflict of Interest
Authors declared no conflict of interest.
 
Acknowledgments
We are grateful to all the scientific consultants of this paper.