Jhuma Biswas and Bijit Kumar Banerjee
Abstract
This study presents a high-resolution analysis of a severe nocturnal thunderstorm that occurred over Guwahati, Northeast India, on 10 May 2025, using the Weather Research and Forecasting (WRF) model along with multiple observational datasets. The simulation employs nested domains at 9 km and 3 km resolution, initialized using Global Forecast System-Final Analysis (GFS-FNL) data. For model validation, observational datasets from ERA5 (European Reanalysis version 5), IMERG (Integrated Multi-satellite Retrievals for Global Precipitation Measurement) precipitation estimates, radiosonde soundings, and high frequency ground-based meteorological data from the Central Pollution Control Board (CPCB) are used. The results reveal intense low-level moisture transport from the Bay of Bengal (BoB), high convective instability with Convective Available Potential Energy (CAPE) values exceeding 800 J/kg, and elevated K Index values (>30), all of which contributed to the initiation and intensification of the thunderstorm. Vertical cross-sections reveal organized updrafts, enhanced mid-tropospheric relative humidity, and pronounced wind shear conditions favourable for deep convection and lightning generation. While the WRF model effectively captures the storm structure and temporal evolution, it tends to underestimate near-surface humidity and overestimate wind speed. Overall, the study demonstrates the effectiveness of high-resolution WRF simulations, particularly when supported by dense observational networks, in improving the understanding of thunderstorm dynamics over complex terrain. The findings hold important implications for improving convective storm forecasting and strengthening disaster preparedness efforts in the vulnerable northeastern region of India.