For the past few years the Radar Operations Center has been testing and refining a 2-dimensional velocity dealiasing algorithm (2D VDEAL) as a candidate replacement algorithm for the current baseline Velocity Dealiasing Algorithm (VDA) on the Weather Surveillance Radar-1988D (WSR-88D). Unlike the VDA which relies on radial continuity, averages of nearby neighbors, or an environmental wind table to help resolve winds exceeding the maximum unambiguous velocity (Nyquist Velocity, VN), the new algorithm simultaneously dealiases all gates in an elevation scan using a least-squares approach to minimize the discontinuity caused by aliasing. Greater weight is given to velocity differences near zero or at multiples of 2VN as well as velocity differences where the corresponding spectrum width values are low. The 2D VDEAL can be used by all volume coverage patterns (VCPs) except VCP 121 which uses the Multiple PRF Dealiasing Algorithm) with no perceptible delay in product availability.
The 2D VDEAL has been found to be adept at dealiasing hurricanes, especially those with widely-spaced outer bands. During Hurricane Irene (August 2011) velocity dealiasing errors were tabulated for each algorithm for base velocity products from 8 WSR-88Ds along the United States Eastern Seaboard for a total of 1690 volume scans. Sites included in the study were Miami, FL, Melbourne, FL, Charleston, SC, Wilmington, NC, Morehead City, NC, Wakefield, VA, Sterling, VA, and Upton, NY. For the ½ deg resolution velocity products at or below 1.5 deg elevation, the legacy VDA had an error rate of 11.98% while the new 2D VDEAL had an error rate of 0.85%. For the 1 deg resolution velocity products, the legacy VDA had an error rate of 11.15% while the 2D VDEAL had an error rate of 1.28%. This paper summarizes the velocity dealiasing results between the algorithms for data from Hurricane Irene and provides examples.