Scholarlink

<p>This study unveils a meticulously crafted approach to capture the ever-changing snow cover of the Himalayan Basin in Uttarakhand, Tapovan, Vishnugad The technique that cuts through the clouds to reveal clear, monthly and seasonal maps using the sharp eyes of Sentinel-2 and MODIS (Terra and Aqua) satellites. A combination of cloud masking, temporal gap-filling, and Random Forest (RF) classification was employed to overcome cloud contamination and data loss, those challenges that especially prevail during monsoon seasons. Sentinel-2’s high spatial resolution (10–20&#xa0;m) significantly outperformed MODIS (500&#xa0;m) in complex terrain, demonstrating higher classification accuracy and reliability across all seasons and years (2021–2024), despite Sentinel’s low temporal resolution and MODIS high temporal resolution. Balanced training improved MODIS performance but highlighted a trade-off between precision and recall due to the low spatial resolution. A hierarchical gap-filling strategy and multi-sensor compositing further minimized missing pixels, using a ± 7 to ± 14&#xa0;day temporal window and spatial elevation-based interpolation. Seasonal analysis revealed peak snow cover during winter with substantial melt in monsoon rather than summer, and 2023 exhibited notably lower snow extent, likely linked to climate variability. The results have strong implications for hydrological modelling, glacial melt forecasting, and sustainable water resource management. The methodology proves effective for large-scale, high-altitude snow monitoring in the complex Himalayan terrain and can be extended using future integration with SAR data and deep learning approaches for enhanced cloud detection and classification.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Snow Cover Monitoring: A Modified Cloud Removal Methodology for MODIS Data in the Tapovan Vishnugad Basin

  • 期刊论文

  • 发表日期：

    2026年7月2日

查看全文
  • Gopika Vijayalekshmi,
  • N. K. Goel,
  • Manohar Arora

摘要

This study unveils a meticulously crafted approach to capture the ever-changing snow cover of the Himalayan Basin in Uttarakhand, Tapovan, Vishnugad The technique that cuts through the clouds to reveal clear, monthly and seasonal maps using the sharp eyes of Sentinel-2 and MODIS (Terra and Aqua) satellites. A combination of cloud masking, temporal gap-filling, and Random Forest (RF) classification was employed to overcome cloud contamination and data loss, those challenges that especially prevail during monsoon seasons. Sentinel-2’s high spatial resolution (10–20 m) significantly outperformed MODIS (500 m) in complex terrain, demonstrating higher classification accuracy and reliability across all seasons and years (2021–2024), despite Sentinel’s low temporal resolution and MODIS high temporal resolution. Balanced training improved MODIS performance but highlighted a trade-off between precision and recall due to the low spatial resolution. A hierarchical gap-filling strategy and multi-sensor compositing further minimized missing pixels, using a ± 7 to ± 14 day temporal window and spatial elevation-based interpolation. Seasonal analysis revealed peak snow cover during winter with substantial melt in monsoon rather than summer, and 2023 exhibited notably lower snow extent, likely linked to climate variability. The results have strong implications for hydrological modelling, glacial melt forecasting, and sustainable water resource management. The methodology proves effective for large-scale, high-altitude snow monitoring in the complex Himalayan terrain and can be extended using future integration with SAR data and deep learning approaches for enhanced cloud detection and classification.