Mesoscale adaptive observing and high-resolution tropical cyclone forecasts using COAMPS-TC during T-PARC/TCS08

In support of the recent THORPEX Pacific Asian Regional Campaign (T-PARC) and the ONR Tropical Cyclone Structure-08 (TCS08) experiments, the newly developed Coupled Ocean/Atmosphere Mesoscale Prediction System – Tropical Cyclone (COAMPS®-TC) designed to predict tropical cyclone track, structure and intensity has been applied. The COAMPS-TC system was configured with three nested grids with 45, 15, 5 km horizontal resolution, with the inner two most meshes moving with the tropical cyclone center. The COAMPS-TC system was applied to 12 systems during T-PARC/TCS08, comprising a total of 96 forecasts. Statistical verification indicates that the COAMPS-TC system provided superior track guidance relative to a number of other modeling systems including the operational COAMPS region over the Western Pacific. Verification of the structure and intensity forecasts will be presented as well.

Adaptive observing guidance for tropical cyclones during T-PARC/TCS08 were provided from a number of operational, academic and research institutions all over the world. At the Naval Research Laboratory, mesoscale model guidance was produced twice daily using the recently developed adjoint and tangent linear models for the atmospheric portion of the nonhydrostatic COAMPS-TC over the Northwestern Pacific. A unique aspect of this system is that an exact adjoint to the explicit microphysics has been developed. An adaptive response function region was used to target favorable areas for tropical cyclogenesis and development. Real-time COAMPS-adjoint forecasts with lead times of 36 h and 48 h were executed twice daily during T-PARC/TCS08 using a horizontal resolution of 40 km. The characteristics of the COAMPS sensitivity patterns such as preferred altitudes, comparison of wind field to temperature and moisture field sensitivities, and sensitivity to the sea surface temperature will be summarized. Preliminary data-denial experiment results will be presented and the implications for observing system design on the mesoscale will be discussed. Implications of the adjoint-based sensitivity fields for the predictability of tropical cyclone formation also will be addressed.