Minimizing Common Errors When Projecting Geospatial Data Onto a Vortex-Centered Space

Abstract

The problem of transforming geospatial data into polar coordinates, a common task in the analysis of data centered on geophysical phenomena, is examined. The azimuthal equidistant projection is shown to be the optimal transform for this purpose. A mathematical and observational analysis of the errors incurred by using a common alternative transform is conducted and reveals that such errors can be significant at radii as small as a few hundred kilometers. When evaluating observed 200-hPa wind fields near Atlantic tropical cyclones (TCs), median azimuthal maximum errors in total and radial wind range up to 10% within 700 km of the TC and up to 50% within 1,600 km of the TC. Errors are shown to depend strongly on the statistical characteristics of the data set and increase nonlinearly with latitude and radius.