Diagnosis of phase errors in high-resolution NWP model forecasts of precipitation and application to improved aviation weather forecasts

Accurate, high-resolution forecasts of convective storms are needed to more efficiently and safely route air traffic across the National Airspace System (NAS). Toward addressing this need, a rapidly updating forecast system that blends observation and model-based forecasts called the Consolidated Storm Prediction for Aviation (CoSPA) project has been developed. To achieve a seamless blend from observations to model-based forecasts requires the forecasts of precipitation in the two forecast methods be synchronized. This paper discusses potential approaches for synchronizing the two forecasts. Phase errors (in space) are determined using both a least-squares and a variational approach. Both approaches determine the phase error by comparing the MIT-LL vertically integrated liquid (VIL) mosaic to that obtained with the 3-km version of the High Resolution Rapid Refresh (HRRR). A statistical assessment of how the phase errors vary as a function of time of day and forecast lead time (determined from the two techniques) is given. This information can be used to enhance the model forecasts prior to blending with the observation-based forecasts and result in much improved 0-6 hour forecasts of storms.