Scholarlink

<p>Understanding and predicting future climate hazard patterns, especially drought, is crucial for effective water resource management. Global climate models (GCMs) are tailored to provide improved simulations for multiple climate indicators. This study evaluates the performance of 22 Global Climate Models (GCMs) in simulating historical precipitation (1950–2014) across 21 grid points in Sindh, Pakistan, for subsequent Multi-Model Ensemble (MME) and future drought assessment at regional level. Using three ranking frameworks, with the ability to address multicollinearity and nonlinear related discrepancies, provide a robust ranking at each grid station. Furthermore, the unified ranking at each grid point is obtained using a comprehensive rating metric and top five high-performing GCMs are selected using the Majority Judgment Majority Rule for the study region. Additionally, a novel framework for regional aggregation is also proposed to enhance ensemble accuracy. Furthermore, uncertainties are reduced by utilizing six MMEs under the umbrella of geometric, regression, and machine learning frameworks. The Lp-norm ensemble emerges as the most suitable MME, effectively capturing observed precipitation trends under the Kling-Gupta Efficiency with knowable moments (KGEkm; 0.547). Additionally, a novel drought index, named the Hybrid Framework-Gaussian Climate Drought Index, is also proposed for future drought assessments under three Shared Socioeconomic Pathways (SSPs): SSP1-2.6, SSP2-4.5, and SSP5-8.5. We have utilized a K-Component Gaussian Mixture Model to characterize drought trends, enabling more reliable predictions compare to traditional univariate probability models. Analysis of the long-term trend under steady-state probabilities demonstrates the likelihood of severe and prolonged droughts under higher emission scenarios (SSP5-8.5) over an extended period. These findings emphasize the importance of careful GCM selection, regional aggregation, robust ensemble modeling, and advanced drought indices for improving regional climate risk assessments.</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

Advancing climate modeling: a multi-framework methodology for evaluating GCMs and predicting drought trends

  • 期刊论文

  • 发表日期：

    2025年10月2日

查看全文
  • Muhammad Shakeel,
  • Hussnain Abbas,
  • Zulfiqar Ali,
  • Maysaa Elmahi Abd Elwahab,
  • Ehtesham Asharf,
  • Amna Nazeer

摘要

Understanding and predicting future climate hazard patterns, especially drought, is crucial for effective water resource management. Global climate models (GCMs) are tailored to provide improved simulations for multiple climate indicators. This study evaluates the performance of 22 Global Climate Models (GCMs) in simulating historical precipitation (1950–2014) across 21 grid points in Sindh, Pakistan, for subsequent Multi-Model Ensemble (MME) and future drought assessment at regional level. Using three ranking frameworks, with the ability to address multicollinearity and nonlinear related discrepancies, provide a robust ranking at each grid station. Furthermore, the unified ranking at each grid point is obtained using a comprehensive rating metric and top five high-performing GCMs are selected using the Majority Judgment Majority Rule for the study region. Additionally, a novel framework for regional aggregation is also proposed to enhance ensemble accuracy. Furthermore, uncertainties are reduced by utilizing six MMEs under the umbrella of geometric, regression, and machine learning frameworks. The Lp-norm ensemble emerges as the most suitable MME, effectively capturing observed precipitation trends under the Kling-Gupta Efficiency with knowable moments (KGEkm; 0.547). Additionally, a novel drought index, named the Hybrid Framework-Gaussian Climate Drought Index, is also proposed for future drought assessments under three Shared Socioeconomic Pathways (SSPs): SSP1-2.6, SSP2-4.5, and SSP5-8.5. We have utilized a K-Component Gaussian Mixture Model to characterize drought trends, enabling more reliable predictions compare to traditional univariate probability models. Analysis of the long-term trend under steady-state probabilities demonstrates the likelihood of severe and prolonged droughts under higher emission scenarios (SSP5-8.5) over an extended period. These findings emphasize the importance of careful GCM selection, regional aggregation, robust ensemble modeling, and advanced drought indices for improving regional climate risk assessments.