GGWS-PCNN: A global gridded wind speed dataset (1973/01-2021/12; Ongoing Update)

<p><strong>Profile of the dataset</strong></p> <ul> <li>The GGWS-PCNN&nbsp;is a global gridded monthly dataset of 10-m wind speed based on an artificial intelligence algorithm (the partial convolutional neural network), observations from weather stations (the HadISD dataset), and 34 climate models from CMIP6.</li> <li>It has&nbsp;a resolution of 1.25&deg; &times; 2.5&deg; (latitude &times; longitude).&nbsp; We will update this dataset as soon as the new HadISD version is accessible.</li> <li>For more details about the dataset and its reconstructed processes, please see our paper &quot;<strong>An artificial intelligence reconstruction of global gridded surface winds</strong>&quot; published in the <em>Science Bulletin</em>.</li> </ul> <p><strong>Notice</strong></p> <ul> <li>The HadISD discovered an issue in the wind data after 2013. So in their version&nbsp;3.3.0.202201p and later, they fixed this issue. Find the website<strong>&nbsp;</strong><a href="https://www.metoffice.gov.uk/hadobs/hadisd/">Met Office Hadley Centre observations datasets</a>&nbsp;for more details.</li> <li>Due to the limitations of existing AI algorithms in reconstructing data with many missing values, our product has a small number of outliers (e.g. wind speeds less than zero or very high), most of which are located in the Antarctic region. We recommend you remove these outliers&nbsp;before using this dataset.</li> </ul> <p><strong>Reference</strong></p> <p>Lihong Zhou, Haofeng Liu, Xin Jiang, et al. (2022). <a href="https://www.researchgate.net/publication/363806982_An_artificial_intelligence_reconstruction_of_global_gridded_surface_winds">An artificial intelligence reconstruction of global gridded surface winds</a>. Science Bulletin.</p>