DeepTornado: A Tornado Radar Dataset over China

DeepTornado is a tornado radar dataset constructed from operational S-band Doppler weather radar observations of the China New Generation Weather Radar (CINRAD) network, with reference to confirmed tornado cases over China during the period 2017–2024. The dataset will be updated as more confirmed tornado cases become available. The dataset consists of radar data patches extracted from low-elevation scans and includes three radar variables: reflectivity factor, radial velocity, and spectrum width, each provided at the lowest three elevation angles. All radar data are interpolated to a Cartesian grid with a spatial resolution of 250 m, and each sample is stored as an individual compressed NumPy (.npz) file. An empirical labeling scheme is adopted that jointly considers radar signatures and confirmed tornado reports. Samples are categorized into four classes: confirmed supercell tornadoes (TOR), warning tornado-like mesocyclones without confirmed tornado reports (WRN), confirmed tornadoes with weak or ambiguous radar signatures (WEK), and non-tornadic background scenes (NUL). The dataset contains 2,064 labeled samples derived from 143 confirmed tornado cases, along with 5,314 NUL samples representing diverse non-tornadic conditions. For confidentiality and operational considerations, absolute geographic coordinates, timestamps, and radar site identifiers are not included in the released data. Each sample is provided as a standalone radar-centered data patch without direct event-level metadata. DeepTornado is intended to support statistical analyses of tornado-related radar features, data-driven and machine learning–based methodological studies, and related research on severe convective storms under realistic observational constraints. Users are encouraged to consider the inherent limitations of weather radar observations and to construct additional background samples as needed for specific applications. ---------------- Administrative and Usage Notice This dataset has been officially registered as a meteorological data resource with the China Meteorological Administration (CMA). The dataset is made publicly accessible via Zenodo exclusively for academic research and scientific exchange purposes. It is intended for use by meteorological professionals and researchers with relevant domain expertise. Public accessibility does not imply unrestricted or unconditional use. Use of the dataset for operational deployment is restricted to authorized meteorological authorities within China and is subject to the relevant administrative regulations of the China Meteorological Administration. Commercial use of the dataset is not recommended and may require prior approval, registration, or licensing from the China Meteorological Administration or other relevant authorities. Users are solely responsible for determining whether their intended use falls within permitted academic research use, authorized operational use, or regulated commercial use, and for completing any required registration, reporting, authorization, or licensing procedures prior to use where applicable. The authors and contributors act solely as data compilers and publishers for academic dissemination purposes. They do not represent the China Meteorological Administration or the National Meteorological Center, do not provide legal, administrative, or commercial authorization for data use, and do not guarantee that any specific use of the dataset complies with applicable laws or regulations. To the maximum extent permitted by applicable law, the authors and contributors disclaim any responsibility or liability arising from the use, misuse, interpretation, redistribution, or derivative use of the dataset. Any consequences resulting from non-compliant, unauthorized, or non-research use are the sole responsibility of the user.