Potential impacts of mesoscale assimilation of non-conventional data on rainfall characteristics of Monsoon depressions during Southwest Monsoon

This study deals with the assimilation of satellite radiances in improving the initial conditions of the Advanced Research Weather Research and Forecasting (ARW) model and its impact on the simulation of rainfall and other meteorological features associated with Monsoon depressions (MDs). Eight MDs occurring during 2015-2018 are considered for the study. A set of two numerical experiments: CNTL, which does not consider any data assimilation, and SAT, where satellite radiances are assimilated into the model initial condition, is conducted for each MD case. In this study, satellite radiance data from the sounding instruments Advanced Microwave Sounding Unit (AMSU), Microwave Humidity Sounder (MHS), and Atmospheric Infrared Sounder (AIRS) on polar satellites NOAA 15, NOAA 18, NOAA 19, EOS 2, METOP 1, and METOP 2 are used. The assimilation has been done using the region-specific background error statistics computed for June to August 2017 using the National Meteorological Centre (NMC) method. Improvements in the tracks of MDs after assimilation are seen from 12h to 42h, with a minimum track error of 320 km for SAT runs, in contrast to 400 km for CNTL runs. Winds, mainly at upper tropospheric levels, were simulated well after assimilation, while for lower-level winds, assimilation runs are reliable for a longer range forecast (>24h). The assimilation runs capture both the intensity and spatial spread of precipitation better than the CNTL runs, compared to TRMM precipitation as well as with India Meteorological Department (IMD) station observations. Also, the location and intensity of the maximum precipitation regions simulated in SAT runs are in better agreement with TRMM data than CNTL runs. CNTL runs overestimate the intensity of maximum precipitation with errors from 1.6 cm up to 50.6 cm, while SAT runs underestimate the intensity with errors from -2.6 cm to -19.1 cm. The statistical scores, such as bias, critical success index (CSI), and probability of detection (POD), computed by comparing with GPM data, also highlight the improvement in precipitation forecast of the assimilation runs compared to the CNTL runs. The study presents the significance of satellite radiance assimilation in improving the short-range rainfall prediction.