Do increasing horizontal resolution and downscaling approaches produce a skillful thunderstorm forecast?

The timely prediction of thunderstorms (TS) is always a challenging task for operational and research community. The present study is aimed to address the credibility of the high grid-spacing and downscaling approach for improved simulation of TS. Fourteen TS are simulated with different domain configurations using weather research and forecasting (WRF) model. Two nested domains with 9–3 km (known as DD3), and 6–2 km (DD2), and 3 km single domain (SD3) are considered for simulations. Results indicate that the high-resolution DD2 has improved 2-m temperature (T2), 2-m relative humidity (RH2), and 10-m wind speed (WS10) at different stages of TS. The average mean error of T2 and RH2 in the DD2 experiment is 0.7 °C, − 6% during the mature stage, and 0.2 °C, − 4% at dissipating stage. The error in SD3 and DD3 is relatively higher (9–17% for T2 and 20–60% for RH2). Better horizontal and vertical representation of thermodynamic variables in DD2 run reinforces the atmosphere to initiate and intensify the convection in the right place. The DD2 could show slightly higher instability (convective available potential energy, CAPE, 3188 J kg−1) as compared with DD3 (3164 J kg−1) and SD3 (3020 J kg−1). The model is biased to simulate early TS activity. DD2 run could simulate the formation, mature and dissipation stages with fewer timing errors (− 1.35 h, − 1.5 h, and − 2.6 h, respectively) than other experiments. The critical success index of the DD2 run is higher for all the rainfall thresholds; however, it is more than 0.2 up to 2.5 mm h−1. The results highlight that high resolution nested configuration yields better simulation skills than the single domain configuration.