Akademik Veri Yönetim Sistemi
NATURAL HAZARDS, cilt.122, sa.1, 2025 (SCI-Expanded, Scopus)
This study investigates the monitoring of environmental changes in a high-altitude region of the Himalayas, focusing on glaciers, snow cover, and frozen ground as key indicators of climate change. The primary objective is to predict Land Surface Temperature (LST) using machine learning models and remote sensing data. Environmental variables such as precipitation, population density, snow cover, elevation, glaciers, albedo, wind speed, and humidity were analyzed to assess their impact on LST predictions. Data were sourced from MODIS, Sentinel 2 MSI, and NASA FLDAS. Four machine learning models Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting Trees (GBT), and Classification and Regression Trees (CART) were employed. Additionally, ensemble models, combining these individual algorithms, were tested to improve prediction accuracy. Model performance was evaluated using R-2, root mean square error (RMSE) and mean absolute error (MAE). The ensemble RF_GBT model demonstrated the highest predictive performance (R-2 = 0.7983, RMSE = 1.3080, MAE = 0.7964), followed by the standalone GBT model. Among individual models, RF performed well, while SVM showed the lowest predictive capability. The findings underscore the effectiveness of integrating remote sensing data with machine learning for climate change assessments in high-altitude regions and highlight the advantages of ensemble modeling, which combines the strengths of different algorithms to enhance LST prediction accuracy.