Scholarlink

<p>This study investigates precipitation patterns across two topographically contrasting districts of Uttarakhand, mountainous Chamoli and plains-dominated Udham Singh Nagar. Using IMDAA reanalysis data and Random Forest (RF) classification, we examined the influence of physiographic characteristics on precipitation variability. The RF model incorporated temperature, relative humidity, wind speed, and pressure as predictor variables, achieving classification accuracies between 0.66 and 0.78. Elevation emerged as the dominant predictor, accounting for 68% of precipitation variance. Notably, peak precipitation (78%) occurred at elevations between 1500 and 2500m, demonstrating a non-linear relationship between elevation and rainfall patterns. Monsoon precipitation analysis revealed significant inter-district variation, with maximum rainfall during JJAS showing a 25mm difference (37mm in Chamoli versus 62mm in Udham Singh Nagar). The wettest period in Udham Singh Nagar extended from July 1 to September 15, while Chamoli’s concentrated between July 15 and August 15, highlighting topography’s influence on precipitation distribution. Spatial analysis demonstrated that Chamoli exhibited more grid points with high classification accuracy compared to Udham Singh Nagar, where only single grid point reached 78% accuracy. This disparity reflects the challenges in modeling precipitation across varying physiographic conditions. This research advances our understanding of precipitation classification in complex terrain, with significant implications for regional climate modeling and water resource management in mountainous regions. Future studies should integrate additional environmental parameters to enhance model performance and address persisting classification challenges in microclimates.</p> Graphical abstract <p></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

Assessment of precipitation extremes and risk factors in the Himalayan foothills: a machine learning approach for hydro-meteorological hazard analysis

  • 期刊论文

  • 发表日期：

    2026年2月2日

查看全文
  • Aayushi Tandon,
  • Kunal Kishor,
  • Amit Awasthi,
  • Kanhu Charan Pattnayak

摘要

This study investigates precipitation patterns across two topographically contrasting districts of Uttarakhand, mountainous Chamoli and plains-dominated Udham Singh Nagar. Using IMDAA reanalysis data and Random Forest (RF) classification, we examined the influence of physiographic characteristics on precipitation variability. The RF model incorporated temperature, relative humidity, wind speed, and pressure as predictor variables, achieving classification accuracies between 0.66 and 0.78. Elevation emerged as the dominant predictor, accounting for 68% of precipitation variance. Notably, peak precipitation (78%) occurred at elevations between 1500 and 2500m, demonstrating a non-linear relationship between elevation and rainfall patterns. Monsoon precipitation analysis revealed significant inter-district variation, with maximum rainfall during JJAS showing a 25mm difference (37mm in Chamoli versus 62mm in Udham Singh Nagar). The wettest period in Udham Singh Nagar extended from July 1 to September 15, while Chamoli’s concentrated between July 15 and August 15, highlighting topography’s influence on precipitation distribution. Spatial analysis demonstrated that Chamoli exhibited more grid points with high classification accuracy compared to Udham Singh Nagar, where only single grid point reached 78% accuracy. This disparity reflects the challenges in modeling precipitation across varying physiographic conditions. This research advances our understanding of precipitation classification in complex terrain, with significant implications for regional climate modeling and water resource management in mountainous regions. Future studies should integrate additional environmental parameters to enhance model performance and address persisting classification challenges in microclimates.

Graphical abstract