Evaluation of sabkha changes caused by Houralimiz wetland during 20 years using Landsat images

Background and objectives: The drying up of Hur al-Azim wetland has caused turn in it to Sabkha. The increase in Sabkhas has caused the salinization of the fertile lands of Karkheh and the destruction of agricultural lands in this region, It also creates puffy soils that act as a fine dust center. Sabkhas changes rate of Hur al-Azim wetland margins have not been studied so far. Because of the socioeconomic and environmental importance of this wetland, such research is severely needed. The aim of this study is to evaluate the changes rate of Sabkhas that caused by the drying of Hur al-Azim wetland in a period of 20 years, which is the first research of its kind. In this paper, for the first time, the rate of change of the Sabkhas in different classes was examined and determined. Also, the rate of increase of the Sabkhas has been compared with long-term climate change. Monitoring the changing process of wetlands and their surrounding lands can be helpful in managing these ecosystems, and preparing sabkhas Destitution maps and their changes maps can help environmental and natural resource managers in making more sensible decisions, better land use planning, and improving resource management.
Materials and methods: In this study, Landsat 8 images in 2017 and Landstat 5 in 1997 were used to study the Sabkhas changes and mapping caused by Hur al-Azim Wetland over a 20-year period. The study area was divided into 8 classes, which include: wetland, forming Sabkhas, dried wetland, salt ponds, dark Sabkhas, light Sabkhas, abandoned agricultural lands and agricultural lands. Then, the supervised classification method and the support vector machine method were used to prepare the map of the area. The post-classification comparison method has also been used to investigate changes in the 20-year period. Also, to check the changes in the 20-year period, comparison after classification method was used.
Results: The results of classification accuracy showed that the backup vector machine method was obtained in Landsat 8 image with kappa coefficient of 0.79 and overall accuracy of 84.70 and in Landsat 5 image with kappa coefficient of 0.74 and overall accuracy of 81.42. It turns out that the support vector machine algorithm is a good way to classify this area. Also, the results of the 20-year period changing review, using the comparison after classification methods were showed that classes of dried-up wetlands, salt ponds, and abandoned agricultural land have changed dramatically over the past 20 years. Decreased changes were observed in classes of wetland, agricultural fields, dark Sabkhas and light Sabkhas. Examination of the average long-term rainfall-temperature data in the region shows that the temperature has increased and the amount of rainfall has decreased in the region, which has been one of the affective factors on the region changes.The only class that has not changed during the period of this study was the class of forming Sabkhas.
Conclusion: The results show that the change rate is high in the study area during this 20-year period. One of the reasons is the changes in the wetlands area in recent years. The results also show that by using remote sensing methods with fewer number of points, in addition to saving time and money relatively accurate maps can be prepared and a useful method for examining changes in the amount of Sabkhas is to compare past and present classified maps.