Document Type : Research
Authors
1
1- MSc. graduate, Combat Desertification Department, Desert Studies Faculty, Semnan University, Semnan, Iran
2
2- Assistant Professor, Combat Desertification Department, Desert Studies Faculty, Semnan University, Semnan, Iran
3
Associate Professor, Soil Conservation and Watershed Management Research Institute, Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran
10.22092/wmrj.2024.365469.1581
Abstract
The process of identifying landforms is a topic that has been researched by many researchers. All definitions of geomorphology emphasize the study and identification of landforms. Understanding landforms and how they are distributed is an essential requirement in applied geomorphology and other environmental sciences. Among the different geomorphological types of the desert, the pavement cover is one of the most prominent features of the desert. Examination of cross-sections in these lands indicates that angular or rounded pebbles are concentrated on the surface of a thin layer of coarse-grained destructive materials, which usually have the same thickness. Desert pavements, as one of the landforms of dry areas, play an essential role in the processes of dry areas of the Earth. In this research, using ETM+ sensor data of Landsat satellite according to the desired criteria to identify and classify the desert pavements of Semnan city, the characteristics of desert pavements were identified. For this purpose, through field visit and sampling of the study area, the percentage density of stone cover was measured and the position of all 17 samples was recorded with GPS. In order to classify the types of desert pavements, the methods of support vector machine, neural network, spectral angle map, spectral information divergence and fuzzy art amp were used, and the classification accuracy of each method using overall accuracy coefficients, kappa, user accuracy, the accuracy of the manufacturer, the assigned error and the omitted error, as well as the presentation of the error matrix table, were compared and investigated. Also, the area and the percentage of the area of the maps produced by each classification method were also calculated in the Arc GIS 10.2 environment. The results show that the images obtained from the ETM+ sensor have a significant ability to distinguish the desert pavement classes. Also, based on Kappa coefficient, support vector machine methods (85/05), fuzzy art amp (81/44), neural network (55/17), spectral angle map (53/89) and spectral information divergence (22/2) 50), respectively, have the greatest ability in spectral separation of different pavement layers of the south Semnan desert. Also, the amount of coefficients obtained for different classes is different in different classification methods, and changing the classification method does not cause a drastic change in the relative resolution of the classes. Preparing a map of desert pavements and their types provides a basis for evaluating the region in terms of structural and geomorphological features, which is beneficial in many environmental planning and management issues.
Keywords
Main Subjects
)