Scholarlink

Forecasting precipitation accurately at local scale especially in mountainous regions like North-west Himalayas (NWH) remains a major challenge. Although Machine Learning (ML) techniques offers potential advantages over traditional Numerical Weather Prediction (NWP) models but with a condition that the available input meteorological variables can effectively differentiate between various weather and precipitation events. This study utilizes near surface meteorological variables (air temperature, wind speed, water vapor mixing ratio and sea level pressure) of the High Asia Refined Analysis, version 2 (HARv2) data for the past 41 years (1980–2020) at 13 locations in the Western Himalaya (WH) and Central Himalaya (CH), India. The results are statistically analyzed with the help of Spearman rank correlation and the Mann–Whitney U-test along with the comparison of means, variances, and daily tendencies of input meteorological variables across binary (precipitation/no precipitation) and percentile-based precipitation categories. Results confirm the availability of significant discriminatory capability. The findings of this study support the use of ML models for real-time, local-scale weather forecasting in Himalayan regions by utilizing local meteorological observations for improved predictive accuracy.

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

Utility of Local Scale Surface Meteorological Variables for Developing Machine Learning Techniques for Real Time Weather Forecasting

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Navdeep Batolar,
  • Dan Singh,
  • Mukesh Kumar,
  • Himanshu Pandey

摘要

Forecasting precipitation accurately at local scale especially in mountainous regions like North-west Himalayas (NWH) remains a major challenge. Although Machine Learning (ML) techniques offers potential advantages over traditional Numerical Weather Prediction (NWP) models but with a condition that the available input meteorological variables can effectively differentiate between various weather and precipitation events. This study utilizes near surface meteorological variables (air temperature, wind speed, water vapor mixing ratio and sea level pressure) of the High Asia Refined Analysis, version 2 (HARv2) data for the past 41 years (1980–2020) at 13 locations in the Western Himalaya (WH) and Central Himalaya (CH), India. The results are statistically analyzed with the help of Spearman rank correlation and the Mann–Whitney U-test along with the comparison of means, variances, and daily tendencies of input meteorological variables across binary (precipitation/no precipitation) and percentile-based precipitation categories. Results confirm the availability of significant discriminatory capability. The findings of this study support the use of ML models for real-time, local-scale weather forecasting in Himalayan regions by utilizing local meteorological observations for improved predictive accuracy.