Scholarlink

<p>Reliable forecasts of tropical cyclone tracks are essential for anticipating the environmental and societal impacts of these extreme weather events. In this study, we apply the long short-term memory (LSTM) model to better represent the spatiotemporal correlations of typhoon track forecast errors and to provide situation-dependent Cone of Uncertainty (CoU). The datasets used include the Central Weather Administration’s (CWA) official typhoon track forecasts, as well as deterministic and ensemble forecasts from the European Centre for Medium-Range Weather Forecasts and the National Centers for Environmental Prediction models. In most situations, the LSTM-calibrated track forecasts generally outperform the CWA official forecasts. More importantly, the estimated CoU can reasonably cover the observed tracks. Relationships between CoUs and factors such as typhoon translation speed and direction are also examined. Incorporating global numerical model information further reduces the radius of forecast uncertainty while maintaining reliable coverage. Overall, this study demonstrates that the LSTM-based approach offers a reliable representation of typhoon track forecast uncertainty, which provides valuable insights into situation-dependent uncertainty quantification for extreme weather prediction.</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

Quantifying situation-dependent uncertainty in tropical cyclone track forecasts with a recurrent neural network approach

  • 期刊论文

  • 发表日期：

    2026年2月26日

查看全文
  • Hsiao-Chung Tsai,
  • Fang-Yi Lin,
  • Yung-Lan Lin,
  • Nai-Ning Hsu,
  • Treng-Shi Huang,
  • Russell L. Elsberry

摘要

Reliable forecasts of tropical cyclone tracks are essential for anticipating the environmental and societal impacts of these extreme weather events. In this study, we apply the long short-term memory (LSTM) model to better represent the spatiotemporal correlations of typhoon track forecast errors and to provide situation-dependent Cone of Uncertainty (CoU). The datasets used include the Central Weather Administration’s (CWA) official typhoon track forecasts, as well as deterministic and ensemble forecasts from the European Centre for Medium-Range Weather Forecasts and the National Centers for Environmental Prediction models. In most situations, the LSTM-calibrated track forecasts generally outperform the CWA official forecasts. More importantly, the estimated CoU can reasonably cover the observed tracks. Relationships between CoUs and factors such as typhoon translation speed and direction are also examined. Incorporating global numerical model information further reduces the radius of forecast uncertainty while maintaining reliable coverage. Overall, this study demonstrates that the LSTM-based approach offers a reliable representation of typhoon track forecast uncertainty, which provides valuable insights into situation-dependent uncertainty quantification for extreme weather prediction.