Weather2K: A Multivariate Spatio-Temporal Benchmark Dataset for Meteorological Forecasting Based on Real-Time Observation Data from Ground Weather Stations
August 5, 2025 · View on GitHub
Weather2K: A Multivariate Spatio-Temporal Benchmark Dataset for Meteorological Forecasting Based on Real-Time Observation Data from Ground Weather Stations
【Accepted】by the 26th International Conference on Artificial Intelligence and Statistics (AISTATS) 2023
Download the dataset
This is the open-source version (after necessary adjustments and checks) of the Weather2K dataset: https://huggingface.co/datasets/BUPT-PRIS-727/Weather2K
The shape of the numpy file of Weather2K-R is (1866, 13, 13632), which means 1,866 groud weather stations, 3 constants for position information and 10 meteorological factors, and 13,632 time steps with 3-hour time resolution (Time coverage range: January 1, 2017- August 31, 2021).
| Numpy Index | Long Name | Short Name | Unit |
|---|---|---|---|
| 0 | Latitude | lat | (°) |
| 1 | Longitude | lon | (°) |
| 2 | Altitude | alt | (m) |
| 3 | Air pressure | ap | hpa |
| 4 | Air Temperature | t | (°C) |
| 5/6 | Maximum / Minimum temperature | mxt / mnt | (°C) |
| 7 | Relative humidity | rh | (%) |
| 8 | Precipitation in 3h | p3 | (mm) |
| 9 | Wind direction | wd | (°) |
| 10 | Wind speed | ws | (ms-1) |
| 11 | Maximum wind direction | mwd | (°) |
| 12 | Maximum wind speed | mws | (ms-1) |
If you have any quesetions about the data download, please contact wuming@bupt.edu.cn
Archive
Cite
If you are using this dataset please cite
Zhu X, Xiong Y, Wu M, et al. Weather2K: A Multivariate Spatio-Temporal Benchmark Dataset for Meteorological Forecasting Based on Real-Time Observation Data from Ground Weather Stations[C]//International Conference on Artificial Intelligence and Statistics. PMLR, 2023: 2704-2722.