University of Tehran Press Journal of Ecohydrology 2423-6098 9 3 2022 09 23 Accuracy assessment of remote sensing methods for extraction and monitoring of Zrebar Lake, Iran Accuracy assessment of remote sensing methods for extraction and monitoring of Zrebar Lake, Iran 505 516 91068 10.22059/ije.2023.342056.1632 FA Karim Solaimani Professor, Remote Sensing Centre, Sari Agricultural Sciences and Natural Resources University, Sari, Iran Shadman Darvishi M.Sc. Faculty of Environmental Sciences, Aban Haraz Institute of Higher Education, Amol, Iran Fatemeh Shokrian Assistant Professor, Dept. of Watershed Management, Sari Agricultural Sciences and Natural Resources University, Sari, Iran Journal Article 2021 12 22 Water resources play a very important role in the life of humans, plants and animals. Extracting and monitoring the extent of changes in water areas can be effective in predicting many problems. Today, many methods and algorithms have been introduced using remote sensing data to monitor water changes, so it is very necessary to study the accuracy of these methods. In this regard, the aim of the present study is to evaluate the accuracy of remote sensing methods for monitoring the water of Zarebar Lake over a period of 33 years (1984 to 2017). Therefore, after images preprocessing, water body of Zaribar Lake was extracted using maximum likelihood algorithms, minimum distance, Mahalanobis distance, support vector machine and NDWI, MNDWI and AWEI indices in ENVI5.3 and ArcGIS10.4 softwares, then these method was validated using ground control points. According to the results, all methods have an accuracy of 80%, but the support vector machine and maximum likelihood algorithms have a higher accuracy with a kappa coefficient above 85%. Also, the results of the study of lake water changes show that during the period 1984 to 2017, the water body based on support vector machine and maximum likelihood algorithms has decreased by 738.46/ha (46.83 percent) and 613.06/ha (42.45 percent), respectively. Also, in the same period, based on support vector machine algorithms and maximum likelihood algorithm 47.35% and 36.22% have been added to the reed bed on the lake. Due to the decreasing trend of lake and invasion of vegetation, it is necessary to consider a plane to prevent the lake water body for future. Water resources play a very important role in the life of humans, plants and animals. Extracting and monitoring the extent of changes in water areas can be effective in predicting many problems. Today, many methods and algorithms have been introduced using remote sensing data to monitor water changes, so it is very necessary to study the accuracy of these methods. In this regard, the aim of the present study is to evaluate the accuracy of remote sensing methods for monitoring the water of Zarebar Lake over a period of 33 years (1984 to 2017). Therefore, after images preprocessing, water body of Zaribar Lake was extracted using maximum likelihood algorithms, minimum distance, Mahalanobis distance, support vector machine and NDWI, MNDWI and AWEI indices in ENVI5.3 and ArcGIS10.4 softwares, then these method was validated using ground control points. According to the results, all methods have an accuracy of 80%, but the support vector machine and maximum likelihood algorithms have a higher accuracy with a kappa coefficient above 85%. Also, the results of the study of lake water changes show that during the period 1984 to 2017, the water body based on support vector machine and maximum likelihood algorithms has decreased by 738.46/ha (46.83 percent) and 613.06/ha (42.45 percent), respectively. Also, in the same period, based on support vector machine algorithms and maximum likelihood algorithm 47.35% and 36.22% have been added to the reed bed on the lake. Due to the decreasing trend of lake and invasion of vegetation, it is necessary to consider a plane to prevent the lake water body for future. https://ije.ut.ac.ir/article_91068_58f22c9a19b296e7e5fbfa7d7a059c79.pdf