A Few GIS Clarifications on Tornado Density Mapping

A geographical information system (GIS) perspective is taken to examine conceptual and methodological complications present in tornado density and probability mapping. Tornado density is defined as the inverse-distance-weighted count of tornado touchdown points or tornado-affected cells within a neighborhood area. The paper first adds a few geographic elements into the tornado definition and then characterizes tornado density as a density field in GIS that depends on predefined, modifiable areas to exist. Tornado density is therefore conceptually distinguished from both individual tornadoes and tornado probability. Three factors are identified to be vital in tornado density mapping: the neighborhood size, the distance decay function, and the choice of tornado properties. Correspondingly, 12 neighborhood sizes ranging from 20 to 360 km are tested, four distance decay functions are compared, and two tornado properties—tornado touchdown locations and pathlengths—are separately incorporated in mapping. GIS interpretations, clarifications, and demonstrations are provided for these factors to reach a thorough understanding of how the factors function and affect the resultant tornado density maps. Historical tornado data of the eastern half of the United States from 1973 to 2013 are used in these demonstrations. Uncertainty and propagation analyses are recommended for future tornado density and probability mapping, and a Monte Carlo simulation using tornado pathlength data is conducted as an example of uncertainty modeling. In all, tornado density mapping is diagnosed as a largely subjective activity, and the mapper needs to make multiple choices according to the mapping purpose, scale, and the involved tornado record data.