Data Management Plan

Hurricanes and other tropical cyclones pose a high risk to people, property, and ecosystems along the coastline of the United States. The impact of these storms can cascade through the nation's economy and affect communities far from the coast. Understanding the geographic limit of and exposure to winds from tropical cyclones can help citizens, businesses, and government agencies build resilience to these pending dangers.

These data portray wind exposure between 1988 and 2022 in the North Atlantic and Eastern Pacific Ocean basins and between 2001 and 2022 in the Western Pacific Ocean basin. Exposure was quantified using intersecting storm tracks, overlapping wind intensity areas, and calculated return intervals.

Process Steps:

• 2024-02-01 00:00:00 - Grids Grids were created using the ArcGIS Pro Generate Tessellation feature to create hexagon tessellations (100 square kilometers). Grid extents were based generally on the minimum extent of cumulative storm tracks per ocean basin.
• 2024-02-01 00:00:00 - Tracks International Best Track Archive for Climate Stewardship (Ibtracs) storm track features were selected based on the period of interest (1988 to 2022), wind intensity (34-knot wind speed or greater) and storm type (Tropical storm, Extratropical, or Subtropical storm). Tracks that crossed the International Dateline were clipped to avoid issues with buffering later in the process. In order to match the wind radius data, the tracks were dissolved from 3-hour time steps to 6-hour time steps.
• 2024-02-01 00:00:00 - Wind Radii The Extended Best Track (EBT) data contain the radius of winds speeds at 34, 50, and 64 knots across four quadrants - NE, SE, SW, and NW. These data were matched to the individual track line using the storm ID, date, and time values. For 2022 storm tracks, HURDAT2 data was used, as it was not included in the EBT data. To provide wind radius values for tracks that did not have available data, averages of the EBT data for several categories were calculated. EBT data was filtered to include only storms since 1988, tropical, extratropical, or subtropical storms, with wind speeds of 34 knots or above, and with at least one wind speed radius greater than zero in one of the four quadrants. Each location was defined as land or water, based on a 150NM buffer of the ArcGIS online World Countries file, with islands less than 2,000 square kilometers excluded from the buffering process. The same categorization and land/water status was defined for the storm tracks. Based on available data, average values were not required for every combination of category and land/water status.
• 2024-02-01 00:00:00 - Buffers After wind radius values were assigned to each track segment for each appropriate wind speed and directional quadrant, left and right buffer values were calculated based on track segment direction, as follows: 1. Direction greater than 315 or less than or equal to 45 = Left: Maximum of NW and SW, Right: Maximum of NE and SE 2. Direction greater than 45 and less than or equal to 135 = Left: Maximum of NE and NW, Right: Maximum of SE and SW 3. Direction greater than 135 and less than or equal to 225 = Left: Maximum of NE and SE, Right: Maximum of NW and SW 4. Direction greater than 255 and less than or equal to 315 = Left: Maximum of SE and SW, Right: Maximum of NE and NW Using these values, left buffers and right buffers were created for each track segment (using "round" and "geodesic" methodology), using the North America Albers Equal Area Conic projection, for each wind speed (34, 50, and 64 knots). These buffers were merged, and then dissolved by storm ID. The tracks were also dissolved by storm ID.
• 2024-02-01 00:00:00 - Exposure Values In order to assign occurrence and return interval values to the grid cells, a "one to many" spatial join was performed with the grid, and each of the wind speed buffer sets for each ocean basin. The spatial join results were then summarized and joined back to the grid to determine the number of occurrences for each wind speed. The same process was performed to determine the number of intersecting storm tracks. Return intervals were calculated by dividing the number of years (35) in the track dataset by the number of occurrences for each wind speed.