مهرداد سلجوقی

The history of life on Earth suggests that humans have always been exposed to a variety of natural disasters (Mayahi, 2021). Some of these disasters are related to climatic factors and fluctuations such as droughts and some are related to human factors (Obiahu, 2020). Land use change and land cover are among the most important environmental issues that have caused global concern (Ota, 2018). Such changes are usually caused by human activities such as deforestation, urbanization, agricultural intensification, overgrazing, and subsequent land degradation. In addition  to, natural factors can also lead to these changes. Factors such as intensive agriculture and overgrazing are major causes of land degradation in arid areas. Changes due to human intervention can lead to the destruction of natural resources. Currently, land use changes from natural lands such as forests and savannas to other uses such as agricultural lands, pastures and settlements have intensified (Chen, 2014.(

2.1 Data used The satellite data used in the present study include two satellite images of Landsat 5TM sensors dated 06/17/2010 and Landsat 8, OLI sensors dated 10/06/2019 with row and transit numbers 159 and 41-42. The digital elevation model (DEM) of ASTER sensor with a resolution of 30 square meters was also used. Satellite imagery and digital elevation models were obtained from the archives of the USGS Web site. 1: 25000 topographic maps from Iran Mapping Organization were used as well. Furthermore, layer construction and composition were performed using ERDAS IMAGINE 2014, ENVI and ArcGIS software.
2-2 Preprocessing of satellite images
In order to control the satellite image in terms of accurate ground recording, geometric correction was performed on the image. To achieve this, images from TM sensors and the OLI sensors were referenced. Because changes in lighting conditions affect the actual radiation reaching to a pixel, atmospheric correction must be made on the images. In the present study, in order to perform atmospheric correction, ATCORE plugin in ERDAS IMAGINE 2014 software and metadata file along with satellite images were used (Iranmehr, 2014).
2-3 Processing satellite images and preparing land cover maps
In order to select educational samples in order to perform supervised classification, aerial photographs, Google Earth images and GPS captured points were used in field operations. The distribution of educational samples in the study area was homogeneous and with proper distribution to the extent possible. The number of pixels selected in each training sample should be at least ten times the number of spectral bands used in the image (Sharma, 2011), which was observed in the present study. The appropriate band composition for classification was determined from the Evaluate command in the Signature Editor based on the Best Average Separability. Based on this, a band composition was used to classify the TM sensor image and a band composition of 2357 for the OLI sensor image.
2-4 Assessing the accuracy of land cover maps
In the present study, after classifying the satellite images, the accuracy of the classified image was evaluated using educational samples that were not involved in the classification process. For this purpose, the classification accuracy was evaluated by using the error matrix and calculating the overall accuracy coefficients and Kappa coefficient.
In order to determine the effect of land use change on soil erosion, the amount of erosion per year in each land use was obtained and compared with each other.

3.1 Classification accuracy assessment
In the present study, the accuracy of image classification was performed by using educational samples and the error matrix and calculating the statistical indicators of overall accuracy and kappa coefficient (Table 4). Using the obtained results, the overall accuracy and kappa coefficient in 2010 are 0.92 and 0.90, respectively, and for 2019 are 0.93 and 0.91, respectively.
3-2 Land use assessment
Examination of the results of land use changes in Figures 3 and 4 as well as Table 5 shows that the land use of the region has changed significantly to the extent that in this 9-year period agricultural land has increased by 54.5 percent from 4.24 percent. The percentage in 2010 has reached 9.78 percent in 2019. Bare and saline lands and residential areas also increased by 3.27 and 0.23 percent, respectively, while forests, pastures and aquifers decreased by 0.01, 7.34 and 1.61 percent, respectively.

Soil erosion is one of the environmental problems that is a threat to natural resources, agriculture and the environment and it is considered one of the main land degradation processes in different parts of the world, including Iran (Patil, 2013 & Brath, 2002). According to the results of  Miahi et al. (2021), basins that have heavy rainfall and showers with a direct impact on rain erosion index will increase the potential for soil erosion (Mohammadi, 2018 & Mayahi, 2021). The results showed that, the erosion class showed a very high increase of nearly 5%. Also, these changes in erosion in the use of pastures and bare and saline lands have been more evident because, in this period, rangelands have decreased and bare and saline lands have increased. According to the findings, the decrease in vegetation due to vegetation degradation, especially in arid and semi-arid areas, will
increase soil loss because vegetation as a protective factor of soil against direct rainfall reduces the erosive force of rain and loss (Obiahu, 2020). Therefore, with the loss of vegetation and land use changes, the possibility of direct rainfall colliding with the soil surface provides the soil to be harvested by water and transported in the direction of the slope. Consequently, soil losses become significant (Descheemaeker, 2008). A similar study by Ota et al. (2018) showed that changes in slope cause differences in chemical and physical properties of soil that lead to changes in soil nutrients and, thus, increase soil loss (Ota, 2018).