Cumulus parameterization accounting for differences in heavy rainfall mechanisms over Korea and the central US

A reduction in the contribution of cumulus parameterization has been found to enhance the precipitation forecast skill over Korea in the Korean Integrated Model (KIM), a global numerical weather prediction model operational at the Korea Meteorological Administration (KMA). However, in some other regions, weakening cumulus convection not only induces spurious grid-point storms by failing to eliminate instability in the atmospheric column, but also degrades the forecast skill for both precipitation and large-scale fields. These regional differences arise from differences in the mechanisms driving heavy rainfall. Over the Korean Peninsula, heavy rainfall during summer is more frequently associated with warm-type convection under strong moisture convergence conditions, whereas cold-type rain mechanisms dominate over the central U.S. under convectively unstable synoptic conditions. To address this, the cumulus parameteriztion scheme in KIM is revised to reduce convective strength when the column-integrated moisture flux convergence exceeds a certain threshold, allowing the contribution of cumulus parameterization to be selectively reduced when heavy rainfall is driven by warm-type convection. The revised scheme is evaluated for three heavy rainfall events over Korea―on July 31, August 8, and August 15, 2022―associated with Typhoon Songda, a quasi-stationary Changma front, and a low-pressure system, respectively, as well as through medium-range forecasts conducted at a horizontal resolution of approximately 8 km (NE576NP3). The revised scheme improves the representation of the pattern and intensity of the precipitation core for heavy rainfall events over Korea, as confirmed by higher skill scores for heavier precipitation categories. This improvement is achieved without degrading the performance in other regions