Scholarlink

<p>India, in addition to its varied emission regimes ranging from urban pollution to desert dust, is also a country with contrasting precipitation climatologies with varying wet and dry spell durations. Recent decades have witnessed a substantial increase in anthropogenic aerosol emissions coinciding with altered monsoon precipitation patterns, yet regime-dependent aerosol-cloud interactions across India’s climatic gradient are yet to be quantified. This study examines aerosol-cloud relationships across seven distinct precipitation regimes using 20 years (2003–2022) of CAMS reanalysis aerosol data and ERA5 meteorological fields, spanning from arid Jodhpur in Thar desert of western India (mean monsoon rainfall: 2.83&#xa0;mm day<sup>-1</sup>, 1631 dry spell days) to humid Cochin in Malabar coast of south India (mean monsoon rainfall: 16.47&#xa0;mm day<sup>-1</sup>, 276 dry spell days). Correlation analyses stratified by persistent wet and dry spells (≥ 3 consecutive days) reveal that the nature of aerosol-cloud interactions varies markedly across sites, reflecting spatially heterogeneous aerosol cloud coupling. The most unique finding that this study brings out is the positive correlation observed between sea salt aerosols and cloud properties, even at inland sites as remote as Jodhpur. Unlike the nearly uniform association of sea salt with columnar water vapour, the relationship of sulphate exhibits specificity with respect to individual study locations (environmental regimes). Over arid (Jodhpur) and semi-arid (Ahmedabad) regions, sulphate shows a statistically significant positive relationship with cloud cover, indicating the role of heterogeneous nucleation. The challenges of deciphering the aerosol dependencies of cloud parameters through machine learning algorithms were evident in the complexity and diversity of aerosol-cloud interactions shaped by the availability of water vapor, sea salt, sulphate, and conducive meteorological conditions. A feature ablation study indicated that ridge regression demonstrated the most consistent improvement across 6 out of 7 cities, whereas tree-based models captured region-specific non-linearities more effectively, with Cochin recording the highest gains (+ 30.1% for ridge regression), while Raipur and Hyderabad showed notable improvements with Random Forest (+ 18.7% and + 15.8%, respectively) underscoring the need for context-sensitive model choice.</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

Investigating the impact of aerosol-cloud interactions on precipitation patterns in India

  • 期刊论文

  • 发表日期：

    2026年5月13日

查看全文
  • Krishnendhu S,
  • Chinmay Mallik,
  • K. K. Mohbey,
  • Malika Acharya,
  • S. Ramachandran

摘要

India, in addition to its varied emission regimes ranging from urban pollution to desert dust, is also a country with contrasting precipitation climatologies with varying wet and dry spell durations. Recent decades have witnessed a substantial increase in anthropogenic aerosol emissions coinciding with altered monsoon precipitation patterns, yet regime-dependent aerosol-cloud interactions across India’s climatic gradient are yet to be quantified. This study examines aerosol-cloud relationships across seven distinct precipitation regimes using 20 years (2003–2022) of CAMS reanalysis aerosol data and ERA5 meteorological fields, spanning from arid Jodhpur in Thar desert of western India (mean monsoon rainfall: 2.83 mm day-1, 1631 dry spell days) to humid Cochin in Malabar coast of south India (mean monsoon rainfall: 16.47 mm day-1, 276 dry spell days). Correlation analyses stratified by persistent wet and dry spells (≥ 3 consecutive days) reveal that the nature of aerosol-cloud interactions varies markedly across sites, reflecting spatially heterogeneous aerosol cloud coupling. The most unique finding that this study brings out is the positive correlation observed between sea salt aerosols and cloud properties, even at inland sites as remote as Jodhpur. Unlike the nearly uniform association of sea salt with columnar water vapour, the relationship of sulphate exhibits specificity with respect to individual study locations (environmental regimes). Over arid (Jodhpur) and semi-arid (Ahmedabad) regions, sulphate shows a statistically significant positive relationship with cloud cover, indicating the role of heterogeneous nucleation. The challenges of deciphering the aerosol dependencies of cloud parameters through machine learning algorithms were evident in the complexity and diversity of aerosol-cloud interactions shaped by the availability of water vapor, sea salt, sulphate, and conducive meteorological conditions. A feature ablation study indicated that ridge regression demonstrated the most consistent improvement across 6 out of 7 cities, whereas tree-based models captured region-specific non-linearities more effectively, with Cochin recording the highest gains (+ 30.1% for ridge regression), while Raipur and Hyderabad showed notable improvements with Random Forest (+ 18.7% and + 15.8%, respectively) underscoring the need for context-sensitive model choice.