Doppler weather radar data assimilation at convective-allowing grid spacing for predicting an extreme weather event in Southern India

Abstract

ABSTRACT Thunderstorms and heavy rainfall are the most devastating and dangerous weather phenomena over the tropical region. Southern Peninsula India in general and regions close to Western Ghats in particular are very prone to such severe weather activity. To improve the short-range forecasting of such extreme weather hazards, tremendous efforts are made to develop Doppler Weather Radar (DWR) network over Indian region for detection and continuous monitoring of weather activities. In this study, the Weather Research and Forecasting (WRF) model and its three-dimensional variational (3D-Var) data assimilation system are used to evaluate the importance of the frequently available DWR observations at convective-allowing grid spacing for predicting an extreme weather event. Various assimilation experiments are performed using background error covariance statistics generated from the National Meteorological Center (NMC) method and Ensemble method, quality control in radial winds, and with/without Digital Filter Initialization (DFI) schemes for DWR data assimilation. Results suggest that with the assimilation of DWR observations, the WRF model is able to predict selected rainfall events with less temporal, spatial and intensity error. Moreover, synoptic conditions are more favourable after DWR assimilation. Due to lack of precise initial conditions, control experiments are not able to capture this extreme weather activity. Results suggest that better skill is seen when the DFI method is used in combination with ensemble background error covariance as compared to other experiments. The maximum rainfall prediction skill is enhanced with DWR data assimilation experiment with DFI method and ensemble background error covariance statistics.