Assimilation of radar radial velocity data with the WRF hybrid 4DEnVar system for the prediction of hurricane Ike (2008)

Abstract

Four dimensional ensemble-variation data assimilation (4DEnVar) is the method that considers the flow dependent background error covariance (BEC) and asynchronous observations throughout the assimilation window, which avoids the maintenance of the adjoint model. The impacts of assimilation of radial velocity (Vr) data using hybrid-4DEnVar for the analyses and forecasts of hurricane Ike are investigated using Weather Research and Forecasting and Data Assimilation model (WRFDA). 4DEnVar is coupled with Ensemble Transform Kalman Filter (ETKF) by updating the ensemble mean by the hybrid scheme and the ensemble perturbations are updated by the ETKF. Single observation tests for typical Jet cast and tropical cyclone (TC) case are conducted before the real hurricane Ike (2008) case. It is found that the analysis increment moves downstream by the end of the assimilation window. The linear propagation represented by the 4DEnVar method is close to the full nonlinear model integration. For the real IKE case, it is found that positive and spiral temperature increments, best track and intensity forecast are found in 4DEnVar experiment, indicating a more realistic thermal structure of hurricane Ike. 3DEnVar and 3DVar-FGAT are limited due to the lack of the BEC description spatially and temporally. 3DVar experiment produces much smoother and weaker increments with cold temperature increments at the hurricane vortex center at lower levels.