Scholarlink

Abstract <p>With ongoing debate on whether to use multi-model ensembles or a single optimal model for projecting future climate, choosing appropriate climate model is a challenge in climate change related research. This paper discusses various algorithms for integrating climate model simulations by evaluating their efficiencies using a new metric combining three statistical scores: Nash–Sutcliffe efficiency, percentage bias, and the root-mean-square error to observations standard deviation ratio. It analyzes future climate, which has been derived from 31 global climate models (GCMs) under two socioeconomic pathways scenarios, alongside historical climate data from five stations over the period 1980–2014. The approach based on selecting a single climate model using a Kullback–Leibler divergence of 0.016 and a Bhattacharyya distance of 0.091 has shown the least disparity, indicating the superiority of a single GCM over ensembles for climate model selection. This study proposes a simple procedure for climate model selection, which contributes to more efficient decision-making process within climate adaptation and mitigation strategies.</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

Can Multi-model Ensemble Reduce the Uncertainty in Future Climate Projections?

  • 期刊论文

  • 发表日期：

    2025年10月1日

查看全文
  • S. Nepal,
  • V. P. Pandey,
  • B. R. Thapa

摘要

Abstract

With ongoing debate on whether to use multi-model ensembles or a single optimal model for projecting future climate, choosing appropriate climate model is a challenge in climate change related research. This paper discusses various algorithms for integrating climate model simulations by evaluating their efficiencies using a new metric combining three statistical scores: Nash–Sutcliffe efficiency, percentage bias, and the root-mean-square error to observations standard deviation ratio. It analyzes future climate, which has been derived from 31 global climate models (GCMs) under two socioeconomic pathways scenarios, alongside historical climate data from five stations over the period 1980–2014. The approach based on selecting a single climate model using a Kullback–Leibler divergence of 0.016 and a Bhattacharyya distance of 0.091 has shown the least disparity, indicating the superiority of a single GCM over ensembles for climate model selection. This study proposes a simple procedure for climate model selection, which contributes to more efficient decision-making process within climate adaptation and mitigation strategies.