Scholarlink

<p>Global climate model simulations from the Coupled Model Intercomparison Project Phase 6 (CMIP6) dataset are widely used to produce climate service products for assessing heat stress across Africa. While several studies have identified biases in key variables within CMIP6 and recommended the use of the NASA Earth Exchange Global Daily Downscaled CMIP6 Projections (NEX-GDDP) dataset (where some biases have been corrected), there is little information on how accurately CMIP6 simulations capture heat stress characteristics across Africa and to what extent the NEX-GDDP removes these biases. This study assesses the ability of both the CMIP6 and NEX-GDDP datasets to reproduce heat stress characteristics across continents, with a focus on heat stress hazard. The Global Meteorological Forcing Data (GMFD) dataset served as a reference. The heat index (HI) was used to quantify heat stress and identify the degree of heat stress severity, along with the associated annual hazard, which was defined as the cumulative magnitude of the HI exceeding a specified threshold. A self-organising map (SOM) technique was used to cluster simulations from the CMIP6 and NEX-GDDP datasets based on their HI biases and to identify dominant patterns of annual heat stress hazard. The results show that both datasets are in close agreement with the GMFD in capturing the spatial and seasonal patterns of heat stress across Africa. However, CMIP6 simulations exhibit substantial biases in heat stress coverage, frequency, and hazard. NEX-GDDP corrects many of these biases, although it sometimes overcorrects, resulting in positive biases in heat stress frequency south of 20°N. Self-organising maps reveal the dominant patterns in the observed and simulated interannual hazard over Africa, but one of these patterns appears to be an artefact common to both CMIP6 and NEX-GDDP datasets. This study highlights the limitations of global climate simulations in accurately capturing heat stress variability across Africa. </p> Graphical abstract <p></p> <p>Based on the graphical snapshot, this study investigates the extent to which bias correction of CMIP6 simulations—implemented through the NEX-GDDP framework—enhances the representation of heat stress hazards over Africa. The heat index (HI), which combines temperature and relative humidity to characterize heat stress, is employed. Heat stress hazard is defined as the annual cumulative magnitude of heat stress for all days with at least “Caution”-level stress (HI ≥ 27&#xa0;°C). Using GMFD observational data as a reference, biases in heat stress hazards are quantified for both CMIP6 and NEX-GDDP datasets. To analyze spatial and temporal bias patterns, Self-Organizing Map (SOM) clustering is applied, revealing six dominant bias configurations as shown above. The results show distinct bias patterns across CMIP6 simulations and group these simulations based on their performance. Notably, all NEX-GDDP simulations cluster within the SOM group that features the lowest bias, indicating that bias correction significantly improves the quality of heat stress hazard representation across the continent. These findings offer valuable insights for climate scientists working on model evaluation and regional hazard assessment. For policymakers, the improved accuracy in heat stress projections supports more informed decision-making in climate adaptation planning, public health preparedness, and infrastructure resilience across Africa.</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 Heat Stress Hazards in Africa Using CMIP6 and NEX-GDDP Datasets

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月18日

查看全文
  • Olumuyiwa A. Oloniyo,
  • Babatunde J. Abiodun,
  • Romaric C. Odoulami,
  • Akintunde I. Makinde,
  • Abayomi A. Abatan,
  • Gbenga J. Abiodun

摘要

Global climate model simulations from the Coupled Model Intercomparison Project Phase 6 (CMIP6) dataset are widely used to produce climate service products for assessing heat stress across Africa. While several studies have identified biases in key variables within CMIP6 and recommended the use of the NASA Earth Exchange Global Daily Downscaled CMIP6 Projections (NEX-GDDP) dataset (where some biases have been corrected), there is little information on how accurately CMIP6 simulations capture heat stress characteristics across Africa and to what extent the NEX-GDDP removes these biases. This study assesses the ability of both the CMIP6 and NEX-GDDP datasets to reproduce heat stress characteristics across continents, with a focus on heat stress hazard. The Global Meteorological Forcing Data (GMFD) dataset served as a reference. The heat index (HI) was used to quantify heat stress and identify the degree of heat stress severity, along with the associated annual hazard, which was defined as the cumulative magnitude of the HI exceeding a specified threshold. A self-organising map (SOM) technique was used to cluster simulations from the CMIP6 and NEX-GDDP datasets based on their HI biases and to identify dominant patterns of annual heat stress hazard. The results show that both datasets are in close agreement with the GMFD in capturing the spatial and seasonal patterns of heat stress across Africa. However, CMIP6 simulations exhibit substantial biases in heat stress coverage, frequency, and hazard. NEX-GDDP corrects many of these biases, although it sometimes overcorrects, resulting in positive biases in heat stress frequency south of 20°N. Self-organising maps reveal the dominant patterns in the observed and simulated interannual hazard over Africa, but one of these patterns appears to be an artefact common to both CMIP6 and NEX-GDDP datasets. This study highlights the limitations of global climate simulations in accurately capturing heat stress variability across Africa.

Graphical abstract

Based on the graphical snapshot, this study investigates the extent to which bias correction of CMIP6 simulations—implemented through the NEX-GDDP framework—enhances the representation of heat stress hazards over Africa. The heat index (HI), which combines temperature and relative humidity to characterize heat stress, is employed. Heat stress hazard is defined as the annual cumulative magnitude of heat stress for all days with at least “Caution”-level stress (HI ≥ 27 °C). Using GMFD observational data as a reference, biases in heat stress hazards are quantified for both CMIP6 and NEX-GDDP datasets. To analyze spatial and temporal bias patterns, Self-Organizing Map (SOM) clustering is applied, revealing six dominant bias configurations as shown above. The results show distinct bias patterns across CMIP6 simulations and group these simulations based on their performance. Notably, all NEX-GDDP simulations cluster within the SOM group that features the lowest bias, indicating that bias correction significantly improves the quality of heat stress hazard representation across the continent. These findings offer valuable insights for climate scientists working on model evaluation and regional hazard assessment. For policymakers, the improved accuracy in heat stress projections supports more informed decision-making in climate adaptation planning, public health preparedness, and infrastructure resilience across Africa.