Polyline averaging using distance surfaces: A spatial hurricane climatology

Abstract

The US Gulf states are frequently hit by hurricanes, causing widespread damage resulting in economic loss and occasional human fatalities. Current hurricane climatologies and predictive models frequently omit information on the spatial characteristics of hurricane movement—their linear tracks. We investigate the construction of a spatial hurricane climatology that condenses linear tracks to one-dimensional polylines. With the aid of distance surfaces, an average hurricane track is calculated by summing polylines as part of a grid-based algorithm. We demonstrate the procedure on a particularly vulnerable coastline around the city of Galveston in Texas, where the tracks of the closest storms to Galveston are also weighted by an inverse distance function. Track averaging is also applied as a means of interpolating possible paths of historical storms where records are sporadic observations, and sometimes anecdotal. We offer the average track as a convenient regional summary of expected hurricane movement. The average track, together with other hurricane attributes, also provides a means to assess the expected local vulnerability of property and environmental damage.