The Usefulness of High-Resolution Observational Data for Verification within the United States Air Force

Traditional “grid point” to “grid point” evaluations of weather event forecast fields at convection permitting horizontal grid spacing are inherently problematic. Thus, spatial verification techniques which avoid the double penalty problem attributed to small position and time displacement errors will be applied to convection-allowing deterministic models/ensembles available within the 16^th Weather Squadron. In order to highlight the usefulness of high-resolution observational data, predictions of tornadoes using severe weather proxy variables (e.g., updraft helicity; vertical vorticity) will be compared to observations of rotation tracks provided from the Multi-Radar Multi-Sensor (MRMS) system. Using “neighborhood” approaches to account for spatial and temporal discontinuities in weather features, probabilities generated for the forecast and observed grids will be directly compared by computing the Fractions Skill Score (FSS). In addition to measuring any possible sensitivities of the FSS to variations in the radius of influence, other verification metrics could also be explored. Insights gained from our initial MRMS verification will be expanded to new, higher resolution satellite data sets such as GOES-16 and Himawari to expand this capability across the globe.