Willis Research Network Global Tropical Cyclone Wind Footprint Dataset

This dataset consists of 1046 global historical tropical cyclone (TC) wind footprints generated using the modeling approach described in Done et al. (2020). The dataset contains all TCs that made landfall with maximum wind speed of at least 40mph. Near-miss TCs are also included, defined as TCs that track within 50km of a coastline, or within 250km of the coastline with maximum wind speeds greater or equal to 58mph.

Input track data sources are the Extended Best Track Dataset for the East Pacific and North Atlantic basins, and the Joint Typhoon Warning Center for all other TC basins. The data record length extends from 2019 as far back as the required input data are available. Archived radius of maximum wind (rmax) data extend back to 1988 for the North Atlantic and the East Pacific, and only as far as the early 2000s for the other basins. For all basins, landfalling track points are identified using the landfall flag in the IBTrACS dataset. Near-miss storms are identified using the ‘distance to land’ data in IBTrACS. The modeling approach uses the default Willoughby wind field model to create the gradient level winds that are then brought down to the surface using the Kepert and Wang boundary layer model. A correction factor of 1.225 is applied to all urban locations to correct for a low wind speed bias in urban locations.

Each footprint is defined as the storm lifetime maximum 1-minute average wind at 10meters above Earth’s surface. The units are meters per second. Each footprint is on a latitude longitude grid with a grid spacing of 4km for the Caribbean and US regions and 3km for all other regions. Each footprint is a single text file named using the storm name with the word 'urban' appended to indicate the bias correction has been applied. Each text file has three columns of latitude, longitude and wind speed.

Any missing storms are due to missing data from the default data sources described above. The most common reason for missing storms is missing rmax data near the start or end of tracks.

The parent dataset tar file contains 17 gzipped tar files, one for each global region.