Analysis and very short range forecast of cyclone “AILA” with radar data assimilation with rapid intermittent cycle using ARPS 3DVAR and cloud analysis techniques

Abstract

In this study, both reflectivity and radial velocity are assimilated into the Weather Research and Forecasting (WRF) model using ARPS 3DVAR technique and cloud analysis procedure for analysis and very short range forecast of cyclone AILA. Doppler weather radar (DWR) data from Kolkata radar are assimilated for numerical simulation of landfalling tropical cyclone. Results show that the structure of cyclone AILA has significantly improved when radar data is assimilated. Radar reflectivity data assimilation has strong influence on hydrometeor structures of the initial vortex and precipitation pattern and relatively less influence is observed on the wind fields. Divergence/convergence conditions over cyclone inner-core area in the low-to-middle troposphere (600–900 hPa) are significantly improved when wind data are assimilated. However, less impact is observed on the moisture field. Analysed minimum sea level pressure (SLP) is improved significantly when both reflectivity and wind data assimilated simultaneously (RAD-ZVr experiment), using ARPS 3DVAR technique. In this experiment, the centre of cyclone is relocated very close to the observed position and the system maintains its intensity for longer duration. As compared to other experiments track errors are much reduced and predicted track is very much closer to the best track in RAD-ZVr experiment. Rainfall pattern and amount of rainfall are better captured in this experiment. The study also reveals that cyclone structure, intensification, direction of movement, speed and location of cyclone are significantly improved and different stages of system are best captured when both radar reflectivity and wind data are assimilated using ARPS 3DVAR technique and cloud analysis procedure. Thus optimal impact of radar data is realized in RAD-ZVr experiment. The impact of DWR data reduces after 12 h forecast and it is due to the dominance of the flow from large-scale global forecast system model. Successful coupling of data assimilation package ARPS 3DVAR with WRF model for Indian DWR data is also demonstrated.