Scholarlink

<p>Accurate storm surge estimation is critical for effective coastal disaster management, particularly in cyclone-prone regions. A blended wind field derived from ERA5 reanalysis, India Meteorological Department (IMD) best track and Holland parametric winds is used to simulate the Advanced Circulation (ADCIRC) model, and the resulting surge and cyclone variables, along with astronomical tide obtained from WXTide32 are used as inputs to classical Artificial Neural Network (ANN), Support Vector Regression (SVR) and quantum machine learning models namely Quantum Neural Networks (QNN), Quantum Support Vector Regression (QSVR). The framework is evaluated across five cyclones over Bay of Bengal (BoB) Very Severe Cyclonic Storms (VSCS) Hudhud (2014), Vardah (2016), Gaja (2018), Thane (2011), and the Severe Cyclonic Storm (SCS) Phethai (2018) then compared with ADCIRC numerical model and validated with Indian National Centre for Ocean Information Services (INCOIS) observed storm surge height. Comparative assessment using regression statistics, time-series analysis, and Taylor diagrams demonstrates that the QSVR model consistently provides the most balanced performance across all cyclone cases. For Hudhud (2014), QSVR exhibits a high correlation (CRR <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\sim \)</EquationSource> </InlineEquation> 0.74) with moderate standard deviation and lower RMSE compare to other models. A similar behavior is observed for Phethai and Vardah, where QSVR maintains stable correlations (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\sim \)</EquationSource> </InlineEquation>0.79–0.81) and balanced error characteristics. These findings highlight the potential of quantum-enhanced learning frameworks for next-generation coastal hazard 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

Quantum machine learning framework a novel approach for storm surge height estimation over the Bay of Bengal

  • 期刊论文

  • 发表日期：

    2026年5月2日

查看全文
  • A. Amasarao,
  • P. Sunitha,
  • K. Chandra Sekhar,
  • K. A. Suneetha,
  • B. Mmame

摘要

Accurate storm surge estimation is critical for effective coastal disaster management, particularly in cyclone-prone regions. A blended wind field derived from ERA5 reanalysis, India Meteorological Department (IMD) best track and Holland parametric winds is used to simulate the Advanced Circulation (ADCIRC) model, and the resulting surge and cyclone variables, along with astronomical tide obtained from WXTide32 are used as inputs to classical Artificial Neural Network (ANN), Support Vector Regression (SVR) and quantum machine learning models namely Quantum Neural Networks (QNN), Quantum Support Vector Regression (QSVR). The framework is evaluated across five cyclones over Bay of Bengal (BoB) Very Severe Cyclonic Storms (VSCS) Hudhud (2014), Vardah (2016), Gaja (2018), Thane (2011), and the Severe Cyclonic Storm (SCS) Phethai (2018) then compared with ADCIRC numerical model and validated with Indian National Centre for Ocean Information Services (INCOIS) observed storm surge height. Comparative assessment using regression statistics, time-series analysis, and Taylor diagrams demonstrates that the QSVR model consistently provides the most balanced performance across all cyclone cases. For Hudhud (2014), QSVR exhibits a high correlation (CRR \(\sim \) 0.74) with moderate standard deviation and lower RMSE compare to other models. A similar behavior is observed for Phethai and Vardah, where QSVR maintains stable correlations ( \(\sim \) 0.79–0.81) and balanced error characteristics. These findings highlight the potential of quantum-enhanced learning frameworks for next-generation coastal hazard prediction.