Impact of the assimilation of DWR-derived precipitation rates through latent heat nudging on simulation of rainfall events over Indian region using NCUM-R

Abstract

Radar estimate of precipitation is an important observation which can be used to initialize the high-resolution numerical weather prediction models for improved forecast of high impact weather. Impact of Indian Doppler Weather Radar (DWR) derived precipitation rates on short-range weather forecasts is studied here. The present work is one of the first-of-its-kind studies to assimilate the DWR derived rain rate into a high resolution model over Indian region. Chennai DWR derived surface precipitation rates are included in the high resolution regional version of NCMRWF Unified Model (NCUM-R) through latent heat nudging to study its impact on the simulations of two convective rainfall events that occurred over east-peninsular India. The high resolution initial condition for NCUM-R is prepared using 4D-VAR data assimilation system in which various conventional and satellite observations are used. In this study, two sets of numerical experiments are carried out, one experiment named as control experiment (CNTL) uses only the analysis created by 4D-VAR and in the second (LHN), latent heat nudging is also performed with DWR derived precipitation rates in addition to the use of the 4D-VAR analysis.The results of the study show that the assimilation of DWR derived precipitation rates has a positive impact on the simulation of convective rainfall events. The root mean square errors (RMSE) of the analysed meteorological fields, mainly the mass fields are improved in the LHN compared to CNTL in both cases. The error of Convective Available Potential Energy (CAPE) in the LHN analyses is considerably reduced compared to the CNTL analyses for both cases. The stability indices also suggested that the condition for convection is properly reproduced in the LHN analysis in both cases. When evaluated against rainfall observations, the amount and orientation of the main rainfall pattern showed noteworthy improvements in the LHN relative to CNTL. The rainfall forecasts are further verified based on the object oriented Contiguous Rain Area (CRA) technique. The vector displacement errors and total mean square error are reduced considerably in LHN simulations compared to CNTL. The vertical structure and time evolution of moisture flux and moisture convergence are also well simulated in the LHN. This preliminary study revealed encouraging results which shows that the assimilation of Indian DWR derived rain rates in the high resolution regional NWP system can improve the forecast of convective rainfall events over Indian region.