Storm-Relative Observations in Tropical Cyclone Data Assimilation with an Ensemble Kalman Filter

Abstract

Abstract A storm-relative data assimilation method for tropical cyclones is introduced for the ensemble Kalman filter, using the Hurricane Weather Research and Forecasting (HWRF) Ensemble Data Assimilation System (HEDAS) developed at the Hurricane Research Division of the Atlantic Oceanographic and Meteorological Laboratory at the National Oceanic and Atmospheric Administration. The method entails translating tropical cyclone observations to storm-relative coordinates and requires the assumption of simultaneity of all observations. The observations are then randomly redistributed to assimilation cycles to achieve a more homogeneous spatial distribution. A proof-of-concept study is carried out in an observing system simulation experiment in which airborne Doppler radar radial wind observations are simulated from a higher-resolution (4.5/1.5 km) version of the same model. The results here are compared to the earth-relative version of HEDAS. When storm-relative observations are assimilated using the original HEDAS configuration, improvements are observed in the kinematic representation of the tropical cyclone vortex in analyses. The use of the storm-relative observations with a more homogeneous spatial distribution also reveals that a reduction of the covariance localization horizontal length scale by ½ to ~120 km provides the greatest incremental improvements. Potential positive impact is also seen in the slower cycle-to-cycle error growth. Spatially smoother analyses are obtained in the horizontal, and the evolution of the azimuthally averaged wind structure during short-range forecasts demonstrates better consistency with the nature run.