Simulating rainfall during thunderstorm events: Insights into cloud-to-rain microphysical processes over the Indian subcontinent

Abstract

We investigated the performance of mass transfer from cloud water to rainwater, commonly referred to as ‘autoconversion’, in cloud microphysical schemes within high-resolution numerical models, such as the Weather Research and Forecasting Model (WRF). The proper choice of autoconversion rate in the numerical model is crucial for droplet growth and conversion rates to form precipitation. The liquid water content and cloud properties, varying from shallow to convective clouds, are highly sensitive to autoconversion rates, and their sensitivity is closely linked with model biases. A parcel-bin model guided by observations can provide significant insights for selecting a more suitable suite of autoconversion rates for the forecasting model. The parcel-bin model is important for the explicit representation of hydrometeors population and calculating microphysical process rates. We calculated drop size distribution, relative dispersion, diffusion growth rate coefficient, and autoconversion rates for two environmental conditions using a small-scale model based on in situ airborne measurement data from the Cloud-Aerosol Interaction and Precipitation Enhancement Experiment (CAIPEEX) over the Indian subcontinent. We evaluated four different (Kessler, KES; Liu-Daum, LD; Khairoutdinov-Kogan, KK; Lee-Baik, LB) autoconversion rates in the WRF model for simulating two thunderstorm events over India. The LD, LB and KK autoconversion rates exhibited closer performance and demonstrated a better probability distribution of raindrop size and precipitation compared to KES. The present study highlights the importance of proper choice of autoconversion rates in numerical weather prediction models.