Improved NAVGEM Tropical Cyclone Predictions from Upscaled COAMPS Dynamic Initialization Analyses that have Incorporated Rapid Scan Atmospheric Motion Vectors During the Tropical Cyclone Intensity (TCI-15) Field Experiment

Our research team has developed a dynamic initialization utilizing rapid-scan Atmospheric Motion Vectors (AMVs) that have been reprocessed at 15-minute intervals during the lifecycle of Hurricane Joaquin as part of the ONR Tropical Cyclone Intensity (TCI-15) field experiment. Zonal and meridional wind increments relative to a COAMPS-TC wind field forecast are calculated on the 45 km, 15 km, and 5 km grids with a novel SAMURAI-COAMPS Dynamic Initialization (SCDI) technique. Utilizing a SCDI analysis at 1800 UTC 4 October 2015 as the initial conditions for a 72-h COAMPS-TC model forecast resulted in an accurate track prediction, and the interruption of a rapid decay followed by a period of constant intensity was well predicted. Upscaling a similar SCDI analysis on the 15 km grid based on 15-minute interval AMVs between 1800 UTC 29 September to 0000 UTC 30 September provided a more realistic intensity and structure of Tropical Storm Joaquin for the initial conditions of the NAVGEM global model than the synthetic TC vortex used operationally.

The hypothesis is that improving the parent global model forecast is anticipated to produce better initial and lateral boundary conditions for the next SCDI analysis, and thus impact the end-to-end cycling through the next COAMPS-TC forecast and the next NAVGEM forecast. To validate this hypothesis, the full sequence of 40 six-hourly SCDI analyses with AMVs at 15-minute intervals throughout the lifecycle of Hurricane Joaquin are upscaled to the NAVGEM grid. The NAVGEM forecasts each six hours are compared with the operational NAVGEM that utilizes synthetic vortices to represent Hurricane Joaquin from the National Hurricane Center warning messages. A special objective is to improve the NAVGEM track forecasts after Hurricane Joaquin had struck the Bahamas and reversed course to move northeastward, because the operational NAVGEM forecasts had forecast Joaquin would move to the east coast of the United States. These poor NAVGEM forecasts are attributed to the synthetic vortex not representing the vortex structure and outflow structure above 400 mb. Therefore, incorporating the full three-dimensional vortex structure from the SCDI analyses based on the rapid scan AMVs is expected to greatly improve the track forecasts as the outflow from Hurricane Joaquin interacts with the midlatitude trough to the west over northern Florida.