Improved COAMPS-TC Tropical Cyclone Predictions Using a Dynamic Initialization Scheme With Rapid-Scan Atmospheric Motion Vectors and ONR TCI datasets

A new dynamic initialization scheme has been developed that uses rapid-scan Atmospheric Motion Vectors (AMVs) that have been reprocessed at 15-minute intervals during the lifecycle of Hurricane Joaquin as part of the Office of Naval Research (ONR) Tropical Cyclone Intensity (TCI-15) field experiment. Zonal and meridional wind increments relative to a triple-nested Coupled Ocean-Atmosphere Mesoscale Prediction System-Tropical Cyclones (COAMPS-TC) wind field forecast are calculated on the 45 km, 15 km, and 5 km grids with this novel Spline Analysis at Mesoscale Utilizing Radar and Aircraft Instrumentation (SAMURAI)-COAMPS Dynamic Initialization (SCDI) technique. The hypothesis is that the input of high-temporal and spatial representation of the 3D flow field in the tropical cyclone (TC) and near-environment in this dynamic initialization framework will accomplish the following TC initialization goals in the numerical model: (i) correct symmetric and asymmetric diabatic heating (ii), improved vortex balance, (iii) improved balance/linkage between vortex and environment, (iv) improved vortex structure (wind field, asymmetries, heating, vertical tilt), (v) improved near-environment (vertical wind shear, outflow jets), and (vi) model physics spin-up (PBL moisture, hydrometeors, physics-dynamics balance). In order to robustly test this hypothesis, the SCDI scheme is executed on 40 forecasts (from Sept. 26, 2015 to October 6, 2015) using a special dataset with AMVs at 15-minute intervals throughout this entire period, and compared to the standard COAMPS-TC forecasts that do not use these datasets and use a static bogus vortex initialization. It is demonstrated that the SCDI scheme produces improved structure and intensity forecasts in comparison to observations over the conventional bogus vortex method.