The Global Kilometric Tropical Cyclone Wind Database (CyclObs) Enabling High-Resolution Data-Driven Studies

Arthur Avenas, Taiga Tsukada, Théo Cévaër, John A. Knaff, Bertrand Chapron, Alexis Aurelien Mouche

Preprint · 26 December 2025 · Colorado State University (incl. CIRA), NOAA (incl. GFDL, PSL, GSL, NCEP/EMC)

In one sentence. CyclObs provides a global kilometre-scale tropical cyclone surface wind dataset from over 2,500 SAR observations across 500+ storms, intended for machine learning and model validation.

Abstract

Accurate, high-resolution ocean surface wind data is paramount for advancing tropical cyclone (TC) research and reducing uncertainties in operational forecasts. Here we introduce CyclObs, a global kilometer-scale TC ocean surface wind field dataset. CyclObs is derived from more than 2,500 synthetic aperture radar (SAR) observations over at least 500 TCs spanning the period 2016-present and achieving a twice-daily temporal sampling for the most recent years. The dataset is meticulously quality-controlled against independent satellite, aircraft and dropsonde wind measurements. The subsequent SAR wind speed estimates are augmented with computed structural parameters, including metrics for TC intensity (maximum sustained wind), inner-core size (radius of maximum wind), and outer-core characteristics. CyclObs is then leveraged to revisit fundamental questions regarding the variations of the wind structure across different ocean basins and throughout the TC life cycle, or the reexamination of existing wind-pressure relationships. By offering unprecedented spatio-temporal detail on the TC ocean surface wind structure, CyclObs is poised to facilitate the development and validation of next-generation numerical models, support machine learning-based meteorological and climate applications, and enable more robust assessments of the TC destructive potential.

Labels

Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.