Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and Thermodynamics

Abstract Machine Learning Weather Prediction (MLWP) models have recently demonstrated remarkable potential to rival physics‐based Numerical Weather Prediction (NWP) models, offering global weather forecasts at a fraction of the computational cost. However, thorough evaluations are essential before c...

Full description

Saved in:
Bibliographic Details
Main Authors: Pankaj Lal Sahu, Sukumaran Sandeep, Hariprasad Kodamana
Format: Article
Language:English
Published: Wiley 2025-06-01
Series:Journal of Geophysical Research: Machine Learning and Computation
Subjects:
Online Access:https://doi.org/10.1029/2025JH000594
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849431455931826176
author Pankaj Lal Sahu
Sukumaran Sandeep
Hariprasad Kodamana
author_facet Pankaj Lal Sahu
Sukumaran Sandeep
Hariprasad Kodamana
author_sort Pankaj Lal Sahu
collection DOAJ
description Abstract Machine Learning Weather Prediction (MLWP) models have recently demonstrated remarkable potential to rival physics‐based Numerical Weather Prediction (NWP) models, offering global weather forecasts at a fraction of the computational cost. However, thorough evaluations are essential before considering MLWP models as replacements for NWP models. This study presents a comprehensive evaluation of four leading MLWP models—GraphCast, PanguWeather, Aurora, and FourCastNet—against observations and three state‐of‐the‐art NWP models in predicting tropical cyclones (TCs) across all tropical ocean basins. All MLWP models exhibited strong skill in forecasting TC tracks, achieving an average track error of less than 200 km at a 96‐hr forecast lead time. However, they consistently underestimated maximum sustained wind speeds compared to NWP models and observations. The low bias in TC intensity forecasts by MLWP models is linked to similar bias in their training data, along with the double penalization effect. MLWP models realistically captured the absolute vorticity patterns and their advection, demonstrating their ability to represent the dynamics underlying TC translation. They also captured the low‐level convergence and vertical warm core structure of TCs, although the magnitudes were weaker than observed, highlighting the linkage between dynamical and thermodynamical processes. The consistency in magnitude between various physical fields in the MLWP models suggests that they intuitively learn the interrelationships among different physical fields during the evolution of weather systems, demonstrating their ability to capture complex physical interactions. Among the MLWP models, Aurora showed superior performance, surpassing GraphCast, PanguWeather, and FourCastNet.
format Article
id doaj-art-4cbc09fbf2cf4b78a47c1ca548ade027
institution Kabale University
issn 2993-5210
language English
publishDate 2025-06-01
publisher Wiley
record_format Article
series Journal of Geophysical Research: Machine Learning and Computation
spelling doaj-art-4cbc09fbf2cf4b78a47c1ca548ade0272025-08-20T03:27:37ZengWileyJournal of Geophysical Research: Machine Learning and Computation2993-52102025-06-0122n/an/a10.1029/2025JH000594Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and ThermodynamicsPankaj Lal Sahu0Sukumaran Sandeep1Hariprasad Kodamana2Centre for Atmospheric Sciences Indian Institute of Technology Delhi New Delhi IndiaCentre for Atmospheric Sciences Indian Institute of Technology Delhi New Delhi IndiaYardi School of Artificial Intelligence Indian Institute of Technology Delhi New Delhi IndiaAbstract Machine Learning Weather Prediction (MLWP) models have recently demonstrated remarkable potential to rival physics‐based Numerical Weather Prediction (NWP) models, offering global weather forecasts at a fraction of the computational cost. However, thorough evaluations are essential before considering MLWP models as replacements for NWP models. This study presents a comprehensive evaluation of four leading MLWP models—GraphCast, PanguWeather, Aurora, and FourCastNet—against observations and three state‐of‐the‐art NWP models in predicting tropical cyclones (TCs) across all tropical ocean basins. All MLWP models exhibited strong skill in forecasting TC tracks, achieving an average track error of less than 200 km at a 96‐hr forecast lead time. However, they consistently underestimated maximum sustained wind speeds compared to NWP models and observations. The low bias in TC intensity forecasts by MLWP models is linked to similar bias in their training data, along with the double penalization effect. MLWP models realistically captured the absolute vorticity patterns and their advection, demonstrating their ability to represent the dynamics underlying TC translation. They also captured the low‐level convergence and vertical warm core structure of TCs, although the magnitudes were weaker than observed, highlighting the linkage between dynamical and thermodynamical processes. The consistency in magnitude between various physical fields in the MLWP models suggests that they intuitively learn the interrelationships among different physical fields during the evolution of weather systems, demonstrating their ability to capture complex physical interactions. Among the MLWP models, Aurora showed superior performance, surpassing GraphCast, PanguWeather, and FourCastNet.https://doi.org/10.1029/2025JH000594machine learning based weather predictiontropical cycloneslearning of physics
spellingShingle Pankaj Lal Sahu
Sukumaran Sandeep
Hariprasad Kodamana
Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and Thermodynamics
Journal of Geophysical Research: Machine Learning and Computation
machine learning based weather prediction
tropical cyclones
learning of physics
title Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and Thermodynamics
title_full Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and Thermodynamics
title_fullStr Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and Thermodynamics
title_full_unstemmed Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and Thermodynamics
title_short Evaluating Global Machine Learning Models for Tropical Cyclone Dynamics and Thermodynamics
title_sort evaluating global machine learning models for tropical cyclone dynamics and thermodynamics
topic machine learning based weather prediction
tropical cyclones
learning of physics
url https://doi.org/10.1029/2025JH000594
work_keys_str_mv AT pankajlalsahu evaluatingglobalmachinelearningmodelsfortropicalcyclonedynamicsandthermodynamics
AT sukumaransandeep evaluatingglobalmachinelearningmodelsfortropicalcyclonedynamicsandthermodynamics
AT hariprasadkodamana evaluatingglobalmachinelearningmodelsfortropicalcyclonedynamicsandthermodynamics