Scholarlink

<p>Accurate precipitation forecasting is critical for detecting climate change and mitigating flood risks. This study proposes a hybrid model based on Variational Mode Decomposition (VMD), which decomposes precipitation data into distinct modal components for trend analysis. The Particle Swarm Optimization (PSO) algorithm is used to optimize model parameters, combining fast convergence with high solution accuracy. The Bi-directional Long Short-Term Memory (BiLSTM) network improves prediction accuracy by capturing long-term dependencies in time-series data. We use monthly precipitation data from Handan City, Hebei Province (2001–2020), for model testing. The VMD-PSO-BiLSTM model is compared with other forecasting models. Results show that the hybrid model’s predictions align closely with actual precipitation trends, with a minimum error of 0.12%, a maximum relative error of 8.71%, and a prediction accuracy (R²) of 0.97. This model enhances forecasting accuracy and offers valuable support for climate change research and flood prevention efforts.</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

Monthly rainfall prediction model based on VMD-PSO-BiLSTM-case study: Handan City, China

  • 期刊论文

  • 发表日期：

    2025年2月1日

查看全文
  • ShaoLei Guo,
  • Yuehan Zhang,
  • Xianqi Zhang,
  • Wanhui Cheng,
  • He Ren

摘要

Accurate precipitation forecasting is critical for detecting climate change and mitigating flood risks. This study proposes a hybrid model based on Variational Mode Decomposition (VMD), which decomposes precipitation data into distinct modal components for trend analysis. The Particle Swarm Optimization (PSO) algorithm is used to optimize model parameters, combining fast convergence with high solution accuracy. The Bi-directional Long Short-Term Memory (BiLSTM) network improves prediction accuracy by capturing long-term dependencies in time-series data. We use monthly precipitation data from Handan City, Hebei Province (2001–2020), for model testing. The VMD-PSO-BiLSTM model is compared with other forecasting models. Results show that the hybrid model’s predictions align closely with actual precipitation trends, with a minimum error of 0.12%, a maximum relative error of 8.71%, and a prediction accuracy (R²) of 0.97. This model enhances forecasting accuracy and offers valuable support for climate change research and flood prevention efforts.