|
|
:
|
In recent years, floods in rivers in coastal areas have caused intermittent losses of billions of rials. Morphometric factors of the river and its urrounding environments and predicting future conditions can be effective and essential in planning and managing coastal plains. The Mandestan Plain is located in Dashti County, Bushehr Province, in the extreme southwest of Iran. Severe and destructive floods in recent years have caused extensive damage to agricultural lands, facilities, buildings, and residents of this region. In this research, Landsat 5, 7, and 8 satellite images from 2000, 1987, and 2020, field surveys, and Envi 5.3, ArcGIS 10.4.1, and Idrisi TerrSet software were used as research tools. First, the probability values of land use changes in 2020 were obtained based on Markov chains. The purpose of this research was to predict changes in the Mend River by 2030. For this purpose, first, land use satellite images from 1987, 2000, and 2020 were prepared using images from Landsat satellite sensors. In the next step, using a Markov chain, land use was predicted based on the land use of 1987 and 2000 and the land use of 2020. Then, the actual land use map was compared with the projected land use map, and in the next step, a map was prepared for the year 2030 using images from the years 2000 and 2020. The results showed that there are three major land uses in the research area, including water, uncovered areas, and covered lands. The maps prepared for the years 1987, 2000, and 2020 have an overall accuracy of 99 percent, 95 percent, and 99 percent, respectively, and the kappa coefficients of the maps for the years 1987, 2000, and 2020 are 0.98, 0.89, and 0.85, respectively. Overall, the results showed that using satellite images and Markov chains has a relatively good ability to predict changes in the Mand River.
|