نوع مقاله : مقاله علمی پژوهشی
نویسندگان
1 دانشگاه شهید چمران اهواز، دانشکده علوم زمین، گروه سنجش از دور و GIS
2 گروه سنجش از دور و GIS، دانشکده علوم زمین، دانشگاه شهید چمران اهواز؛ اهواز؛ ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Objective: Considering the high uncertainty of Snow Cover Area (SCA) in mountainous regions with complex topography, this study aimed to improve snow discrimination from other land surfaces through the combination of spectral indices and metaheuristic optimization algorithms.
Methods: Landsat 8 satellite images from Alborz and Chaharmahal and Bakhtiari provinces, acquired in January and March, were used. Spectral indices, including NDSI, NDSII-1, S3, SWI, and the second principal component (PCAM2), were extracted. To model snow as a continuous phenomenon, an Adaptive Neuro-Fuzzy Inference System (ANFIS) and its optimized versions using metaheuristic algorithms, including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Whale Optimization Algorithm (WOA), Differential Evolution (DE), and Ant Colony Optimization (ACOR), were employed. Model performance in training and testing phases was evaluated using RMSE and MSE.
Results: The results indicated that algorithm performance depends on the spatial and temporal conditions of the study area. In the Karun North watershed, PSO achieved the lowest RMSE (0.077) on 2022/11/20, and together with ANFIS, it showed the minimum error on 2023/03/20. In contrast, WOA exhibited the highest error and instability in both regions. In the Tehran–Karaj watershed, ANFIS consistently provided the most accurate and stable performance. Overall, ANFIS-based models and metaheuristic algorithms, particularly PSO, demonstrated higher capability in discriminating snow from other land surfaces, while WOA performed relatively poorly.
Conclusions: The study demonstrates that combining multiple spectral indices with metaheuristic-optimized computational intelligence models can significantly reduce uncertainty in snow cover estimation in mountainous regions. Among the evaluated models, PSO and ANFIS are recommended as the most efficient options for snow cover monitoring in mountainous watersheds.
کلیدواژهها [English]
)