Tropical Cyclone-Following Dataset for Limited-Area Machine Learning Applications
Abstract
This work presents a lightweight dataset of tropical cyclone (TC)-following meteorological variables to facilitate the development and application of machine learning (ML) methods in TC research. Using outputs from our previous climate downscaling and idealized simulations, the dataset is constructed as multichannel video sequences that track TC centers throughout their lifetimes. To demonstrate the dataset utility, we present an application in which a regional ML model is trained to predict TC intensity with explicit incorporation of lateral boundary conditions. Our results highlight the importance of lateral boundary information for predicting TC intensity with ML models inside a limited-area domain, underscoring the need to properly account for lateral boundaries when developing regional ML models for TC intensity forecast.
Related
- FAST-ML: A Hybrid Physics-Machine Learning Framework for Tropical Cyclone Intensity Forecasting
- A Hybrid Deep Learning Framework for Tropical Cyclone Intensity Estimation With Spatiotemporal Features and Physical‐Persistence Adjustment
- Skillful Seasonal Tropical Cyclone Prediction with AI Weather--Climate Models Forced by Dynamically Predicted SSTs
- Tibetan Plateau uplift increased western North Pacific tropical cyclone frequency
- When anticipation fails: A forensic analysis of Tropical Cyclone Ditwah reveals whole-system performance limits in a small-island flood-risk system
BibTeX
@misc{luong2026tropical,
title = {{Tropical Cyclone-Following Dataset for Limited-Area Machine Learning Applications}},
author = {Minh-Khanh Luong and Chanh Kieu},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.03925},
eprint = {2611.03925},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.03925},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Tropical Cyclone-Following Dataset for Limited-Area Machine Learning Applications
AU - Minh-Khanh Luong
AU - Chanh Kieu
PY - 2026
DA - 2026/11/02
PB - arXiv
DO - 10.48550/arxiv.2611.03925
UR - https://arxiv.org/abs/2611.03925
AB - This work presents a lightweight dataset of tropical cyclone (TC)-following meteorological variables to facilitate the development and application of machine learning (ML) methods in TC research. Using outputs from our previous climate downscaling and idealized simulations, the dataset is constructed as multichannel video sequences that track TC centers throughout their lifetimes. To demonstrate the dataset utility, we present an application in which a regional ML model is trained to predict TC intensity with explicit incorporation of lateral boundary conditions. Our results highlight the importance of lateral boundary information for predicting TC intensity with ML models inside a limited-area domain, underscoring the need to properly account for lateral boundaries when developing regional ML models for TC intensity forecast.
ER -