Scholarlink

<p>In this study, we conducted an experiment to construct multi-model ensemble (MME) predictions for the El Niño-Southern Oscillation (ENSO) using a neural network, based on hindcast data released from five coupled ocean-atmosphere models, which exhibit varying levels of complexity. This nonlinear approach demonstrated extraordinary superiority and effectiveness in constructing ENSO MME. Subsequently, we employed the leave-one-out cross-validation and the moving base methods to further validate the robustness of the neural network model in the formulation of ENSO MME. In conclusion, the neural network algorithm outperforms the conventional approach of assigning a uniform weight to all models. This is evidenced by an enhancement in correlation coefficients and reduction in prediction errors, which have the potential to provide a more accurate ENSO forecast.</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

Construction of multi-model ensemble prediction for ENSO based on neural network

  • 期刊论文

  • 发表日期：

    2025年8月1日

查看全文
  • Yuan Ou,
  • Ting Liu,
  • Tao Lian

摘要

In this study, we conducted an experiment to construct multi-model ensemble (MME) predictions for the El Niño-Southern Oscillation (ENSO) using a neural network, based on hindcast data released from five coupled ocean-atmosphere models, which exhibit varying levels of complexity. This nonlinear approach demonstrated extraordinary superiority and effectiveness in constructing ENSO MME. Subsequently, we employed the leave-one-out cross-validation and the moving base methods to further validate the robustness of the neural network model in the formulation of ENSO MME. In conclusion, the neural network algorithm outperforms the conventional approach of assigning a uniform weight to all models. This is evidenced by an enhancement in correlation coefficients and reduction in prediction errors, which have the potential to provide a more accurate ENSO forecast.