Deep learning models improve imputation of missing tropical cyclone wind data

Researchers have developed deep learning models, including 1DCNNs and LSTMs, to impute missing Radius of Maximum Winds (Rmax) values in tropical cyclone best-track data. The study found that incorporating the radius of 34-knot winds (R34) significantly improved model performance. Temporal models, despite using fewer samples, showed better preservation of Rmax variability, especially when R34 was unavailable, indicating that temporal information can partially compensate for missing storm-size predictors. Transfer learning did not yield improvements, likely due to distributional differences between synthetic and observational datasets. AI

IMPACT Enhances the accuracy of climate hazard assessments by improving the completeness of tropical cyclone data.

RANK_REASON The cluster contains an academic paper detailing a new research methodology using deep learning for data imputation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• 1DCNN
• International Best Track Archive for Climate Stewardship
• long short-term memory
• RAFT
• Rmax
• STORM
• tropical cyclone

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →